Directly reprogrammed NK cells driven by BCL11B depletion enhance targeted immunotherapy against pancreatic ductal adenocarcinoma

Han-Seop Kim; Jae Yun Kim; Ji-Young Lee; Binna Seol; Ji Eun Choi; Yee Sook Cho

doi:10.1186/s13045-025-01730-1

. 2025 Nov 13;18:100. doi: 10.1186/s13045-025-01730-1

Directly reprogrammed NK cells driven by BCL11B depletion enhance targeted immunotherapy against pancreatic ductal adenocarcinoma

Han-Seop Kim ¹, Jae Yun Kim ¹, Ji-Young Lee ¹, Binna Seol ¹, Ji Eun Choi ^1,², Yee Sook Cho ^1,^2,^✉

PMCID: PMC12613755 PMID: 41233896

Abstract

Background

Pancreatic ductal adenocarcinoma (PDAC) is a lethal malignancy characterized by desmoplastic stroma, immunosuppressive tumor microenvironment (TME), and resistance to standard therapies. Natural killer (NK) cell-based immunotherapies have shown limited efficacy due to impaired persistence, infiltration, and function in PDAC.

Methods

We established a direct reprogramming strategy to generate cytotoxic NK cells (1 F-NKs) by targeting BCL11B, a transcription factor essential for T cell lineage commitment, using shRNA or CRISPR/Cas9 in peripheral blood mononuclear cells (PBMCs). A genome-wide CRISPR/Cas9 screen identified tumor-intrinsic modulators of NK resistance. Functional and in vivo studies assesses the efficacy of 1 F-NKs alone and in combination with mesothelin (MSLN)-CAR engineering and PKMYT1 inhibition.

Results

BCL11B depletion enabled the generation of CD56^brightCD16^bright 1 F-NKs with potent cytotoxicity and elevated NKG2D and CX3CR1 expression. Site-specific integration of a mesothelin (MSLN)-CAR into BCL11B locus generated MSLN-1 F-NKs with stable antigen specific activity. A genome-wide screen identified PKMYT1 as a modulator of tumor resistance to NK cell-mediated killing; its inhibition by RP6306 upregulated NKG2D ligands (MICA/B) and CX3CL1, sensitizing PDACs to 1 F-NK cytotoxicity. In PDAC xenograft models, 1 F-NKs alone or combined with CAR engineering and RP6306 significantly reduced tumor growth and prolonged survival. Notably, this triple combination elicited a synergistic antitumor effect, outperforming each monotherapy or dual combination.

Conclusions

This study presents a synergistic immunotherapy platform that integrates NK reprogramming, CAR engineering, and tumor sensitization. The combinatorial approach significantly enhances antitumor efficacy in PDAC and offers a promising strategy for overcoming immune resistance in solid tumors.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13045-025-01730-1.

Keywords: Natural killer cells, BCL11B, Direct reprogramming, Pancreatic ductal adenocarcinoma, Chimeric antigen receptor, CRISPR-Cas9, PKMYT1, NKG2D

Background

Pancreatic ductal adenocarcinoma (PDAC) remains one of the most common and lethal malignancies, accounting for over 90% of pancreatic cancers and characterized by a five-year survival rate below 10% [1, 2]. Its dismal prognosis is largely attributed to late-stage diagnosis, high metastatic potential, and resistance to conventional therapies. In particular, the desmoplastic and immunosuppressive tumor microenvironment (TME) plays a pivotal role in limiting immune surveillance and dampening responses to both chemotherapy and immunotherapy [3–5]. The PDAC TME is composed of various stromal and immunosuppressive cellular elements, including tumor-associated macrophages (TAMs), myeloid-derived suppressor cells (MDSCs), cancer-associated fibroblasts (CAFs), and regulatory T cells (Tregs) [6–8], alongside high levels of inhibitory solublemediators such as transforming growth factor-β (TGF-β), interleukin-10 (IL-10), and vascular endothelial growth factor (VEGF) [7, 8]. These components collectively establish a hostile landscape that restrains effector lymphocytes, including natural killer (NK) cells, thereby limiting the efficacy of immune-based therapies.

NK cells are innate cytotoxic lymphocytes capable of targeting malignant cells in an antigen-independent manner. Their ability to mediate allogeneic responses without inducing graft-versus-host disease (GVHD) makes them attractive candidates for off-the-shelf cell-based therapies [9]. However, the application of NK cell immunotherapy in solid tumors such as PDAC has been hindered by several challenges, including suboptimal in vivo persistence, inefficient tumor infiltration, and susceptibility to functional suppression by the TME [7, 10, 11]. In addition, current NK platforms, including donor-derived NK cells, chimeric antigen receptor-modified NK cells (CAR-NKs), and NK cells differentiated from induced pluripotent stem cells (iPSC-NKs), face practical and biological limitations such as product inconsistency, exhaustion, and inadequate performance in solid tumor settings [12].

In response to these limitations, transcription factor-mediated reprogramming has emerged as a promising strategy to generate NK-like effector cells with enhanced functionality. Prior work demonstrated that forced expression of the Yamanaka factors OCT4, SOX2, KLF4, and MYC (OSKM) can reprogram peripheral blood cells into NK cells (4 F-NKs) with potent cytolytic capacity [13]. Nonetheless, this approach involves induction of pluripotency-associated networks, raising translational concerns related to oncogene activation and genomic instability. As a potentially safer alternative, direct lineage reprogramming using immune identity factors has been proposed. In particular, BCL11B, a transcriptional repressor critical for maintaining T cell fate, has been shown to suppress NK cell programs. Its downregulation permits T-to-NK lineage conversion and acquisition of cytotoxic properties [14–16], possibly without activation of pluripotency-associated pathways.

Beyond effector cell generation, tumor-specific targeting remains a critical component of effective immunotherapy in PDAC. Mesothelin (MSLN), a cell-surface glycoprotein highly expressed in over 90% of PDAC tumors, has emerged as a clinically relevant antigen and is currently being evaluated in CAR-T and CAR-NK platforms [17, 18]. Engineering NK cells with MSLN-specific CARs offers offers the potential to enhance tumor recognition and improve selective cytotoxicity. Simultaneously, targeting tumor-intrinsic mechanisms that confer resistance to immune attack represents a complementary strategy. In particular, cell cycle regulators such as PKMYT1, a G2/M checkpoint kinase frequently upregulated in PDAC [19], have been implicated in modulating immune susceptibility. Pharmacological inhibition of PKMYT1 has been linked to the upregulation of ligands such as MICA/B and chemokines including CX3CL1, which play key roles in NK cell activation and recruitment [20, 21].

Taken together, these insights guided the rationale for a multifaceted approach to enhance NK cell immunotherapy for PDAC. Specifically, we explored the potential of BCL11B knockdown to reprogram peripheral lymphocytes into NK cells, the addition of MSLN-targeted CARs to confer antigen specificity, and pharmacologic modulation of tumor sensitivity through inhibition of PKMYT1. This integrated strategy was designed to address key barriers in NK cell-based immunotherapy, namely effector generation, antigen-specific tumor targeting, and resistance mechanisms, within the immunosuppressive PDAC microenvironment.

Materials and methods

Primary cell preparation

Peripheral blood from healthy donors were obtained from the Korea Red Cross blood center, following ethical approval by the Institutional Review Board (IRB) of KRIBB (IRB# P01-201812-31-010). Peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation using Ficoll-Hypaque PLUS (GE Healthcare Life Sciences, Uppsala, Sweden). Isolated PBMCs were cultured in StemPro-34 serum-free medium (SFM) supplemented with 2.5% StemPro-34 nutrient supplement, 1% penicillin/streptomycin (P/S), 2 mM GlutaMAX I (all from Thermo Fisher Scientific, Waltham, MA, USA), 20 ng/ml human interleukin-3 (IL-3), 20 ng/ml human IL-6, 100 ng/ml human stem cell factor (SCF), and 100 ng/ml human Flt3 ligand (Flt3L) (all from PeproTech, Cranbury, NJ, USA).

Natural killer (NK) cells were purified from PBMCs (pNKs) or human umbilical cord blood mononuclear cells (CB-NKs; STEMCELL Technologies, Vancouver, BC, Canada) using a magnetic-activated cell sorting (MACS) NK isolation kit (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s protocol. Briefly, 1 × 10⁷ PBMCs or UCBMCs were resuspended in 40 µl of binding buffer (2% bovine serum albumin [BSA; Sigma-Aldrich, St. Louis, Mo, USA] and 1 mM EDTA [Thermo Fisher Scientific]), followed by incubation with 10 µl of biotinylation reagent for 5 min at 4 °C. After washing, 20 µl of magnetic microbeads was added and incubated for 10 min at 4 °C.

Purified NK cells were centrifuged at 300 x g for 5 min and cultured in RPMI 1640 medium (Thermo Fisher Scientific) supplemented with 200 IU IL-2 (PeproTech), 1% P/S and 10% fetal bovine serum (FBS; Thermo Fisher Scientific). For extended culture, NK cells were maintained in RPMI 1640 medium containing 200 IU/mL IL-2, with or without 10 ng/mL IL-15 and 10% FBS. NK-92 cells (ATCC CRL-2407; Manassas, VA, USA) were maintained in RPMI 1640 medium containing 10% FBS and 200 IU IL-2, with media refreshed every 2–3 days.

CD3⁺ T cells, CD14⁺ monocytes, and CD19⁺ B cells were isolated from PBMCs using Pan T Cell, Pan Monocyte, and B Cell Isolation Kits (all from Miltenyi Biotec), respectively. Each subset was isolated via negative selection using biotin-conjugated antibody cocktails and MACS-based magnetic separation following manufacturer protocols, analogous to the NK cell isolation procedure.

Human iPSC-NK cell generation

Human induced pluripotent stem cells (hiPSCs), previously reprogrammed from newborn foreskin fibroblast, were maintained on growth factor-reduced Matrigel-coated dishes (Thermo Fisher Scientific) in mTeSR1 medium (STEMCELL Technologies) as previously described [22]. For NK cell differentiation, hiPSCs were subjected to spin embryoid body (EB) formation following established protocols [23]. Briefly, cells were dissociated into single-cell suspensions using TrypLE Express (Thermo Fisher Scientific) at 37 °C and passed through a 70 μm cell strainer. A total of 8 × 10³ cells were seeded into each well of a round-bottom 96-well plate, centrifuged at 1,500 rpm for 4 min at 8 °C, and cultured in STEMdiff™ APEL™2 Medium (STEMCELL Technologies) supplemented with 40 ng/mL SCF, vascular endothelial growth factor (VEGF; R&D Systems), and 20 ng/mL bone morphogenetic protein 4 (BMP4; PeproTech) at 37 °C for 6 days.

To initiate NK lineage differentiation, 6–8 spin EBs were transferred into each well of a 2% gelatin-coated 24-well plate and cultured in NK cell differentiation medium (NKM), composed of 56.6% DMEM, 28.3% Ham’s F-12, 15% human AB serum (Sigma), 5 ng/mL sodium selenite (Sigma), 50 µM ethanolamine (Sigma), 20 µg/mL ascorbic acid, 25 µM β-mercaptoethanol, 2 mM L-glutamine, and 1% P/S. The medium was further supplemented with a cytokine cocktail containing 5 ng/mL interleukin-3 (IL-3), 10 ng/mL IL-15, 20 ng/mL IL-7, 20 ng/mL SCF, and 10 ng/mL Flt3L. From day 6 onward, IL-3 was excluded, and cytokines were replenished every 5–7 days. On day 28, differentiated NK cells were enriched using a magnetic-activated cell sorting (MACS) NK Cell Isolation Kit (Miltenyi Biotec) according to the manufacturer’s instructions.

Generation of directly reprogrammed 4 F- and 1 F-NKs

Directly reprogrammed NK cells (drNKs) were generated from PBMCs by ectopic expression of OCT4, SOX2, KLF4, and MYC (OSKM), referred to as four-factor-induced NK cells (4 F-NKs), as previously described [13]. For 4 F-NK generation, PBMCs transduced with OSKM reprogramming factors (CytoTune™ 2.0 Sendai Reprogramming Kit; Thermo Fisher Scientific) at various multiplicities of infection (MOI: 0–10), using an OSK: K:M ratio of 5:3:5. The transduced cells were cultured in 4 F reprogramming medium I (4 F-RM I), composed of StemSpan™ SFEM II (STEMCELL Technologies) supplemented with 10% fetal bovine serum (FBS), 1% P/S, 20 ng/mL each of human IL-3, IL-6, FLT3L, stem cell factor (SCF), and thrombopoietin (TPO) (all from PeproTech), as well as 5 µM CHIR99021 (Tocris Bioscience). After 5 days, cells were transferred to 4 F-reprogramming medium II (4 F-RM II), consisting of StemSpanSFEM II supplemented with 10% FBS, 1% P/S, 200 IU IL-2, 20 ng/mL each of human IL-7 and IL-15, 25 ng/mL each of human SCF and FLT3L, and 2 µM StemRegenin 1 (SR1; Cellagen Technology). Cells were maintained in 1 F-RM II for an additional 12 days.

Lentiviral plasmids encoding short hairpin RNAs (shRNAs) targeting human BCL11B or a non-targeting control were obtained from Origene (BCL11B Human shRNA Plasmid Kit, #TL306424). The target sequences used for BCL11B knockdown were as follows: shBCL11B-#1 (5′-CAG ATC GGC AAG GAG GTG TA-3′), shBCL11B-#2 (5′-CCT AAC CTG TGT CTG CGA AG-3′), and shBCL11B-#3 (5′-GAA ACT AGC GGT GTT CTT T-3′). Lentiviral particles were produced by transfecting HEK293T cells (0.8 × 10⁷ cells per 90 mm dish) with shRNA plasmids in DMEM supplemented with 10% FBS and transfected using 293 Expresso™ Transfection Reagent (Excellgen, Waltham, MA, USA) according to the manufacturer’s instructions. After 72 h, supernatants containing lentiviral particles were collected, filtered through a 0.45 μm membrane, and concentrated by ultracentrifugation at 25,000 rpm for 150 min. Viral pellets were resuspended in RPMI 1640 medium, and titers were determined by limiting dilution in HEK293T cells, calculating transducing units (TU/mL) based on GFP expression using the formula: TU/mL = (number of seeded cells × % GFP⁺ cells) / (0.1 × dilution factor).

To generate BCL11B single-factor-induced NK cells (1 F-NKs), PBMCs or PBMC subsets (1–3 × 10⁵ cells) were first incubated for 3 days in starting cell medium (SCM), composed of StemPro-34 serum-free medium (SFM) supplemented with 2.5% StemPro-34 nutrient supplement, 1% P/S, 2 mM GlutaMAX I, 2 µM 5-azacytidine (Tocris Bioscience, Bristol, UK), 20 ng/ml human IL-3, 20 ng/ml human IL-6, 100 ng/ml human SCF, and 100 ng/ml human FLT-3 L. On day 0, cells were transduced with lentiviral vectors expressing shBCL11B (variants #1, #2, or #3) at a multiplicity of infection (MOI) of 10. On day 1, transduced cells were transferred into 1 F-reprogramming medium (1 F-RM I), consisting of StemSpan™ SFEM II supplemented with 10% FBS, 1% P/S, 20 ng/mL each of human IL-3, IL-6, IL-7, IL-15, FLT3L, SCF, and TPO, along with 3 µM CHIR99021 and 2 µM SR1. Medium was changed daily. After 5 days, cells were transferred to reprograming medium II (1 F-RM II), compose of StemSpan™ SFEM II supplemented with 10% FBS, 1% P/S, 200 IU IL-2, 20 ng/ml human IL-7, 20 ng/ml human IL-15, 20 ng/ml human SCF, 20 ng/ml human FLT-3, 20 ng/ml human IL-21, and 2 µM SR1. Cells were maintained in 1 F-RM II for an additional 12 days.

Reprogrammed NK cells were expanded in NK medium consisting of RPMI 1640 supplemented with 10% FBS, 200 IU IL-2, and 10 ng/mL human IL-15. Culture medium was refreshed every 3 days.

Generation of MSLN-CAR-expressing NKs

To generate mesothelin (MSLN)-specific CAR-expressing BCL11B single-factor-induced NK cells (MSLN-1 F-NKs), a CRISPR/Cas9-mediated knock-in strategy was employed, combining gene disruption of BCL11B with site-specific integration of the MSLN-CAR transgene via adeno-associated virus serotype 6 (AAV6)-mediated homology-directed repair (HDR). CRISPR/Cas9 genome editing was performed using two single-guide RNAs (sgRNAs) targeting exon 1 of BCL11B. The sgRNA sequences (sgRNA#1: 5′-CGC CCG GAG AGC TGC ACT GAT GG-3′; sgRNA#2: 5′-GCA TCT ATT CTG GCA TCG CCC GG-3′) were cloned into the pSpCas9(BB)-2 A-Puro (PX459) plasmid (Addgene #62988). A donor plasmid encoding the MSLN-CAR construct was designed with homology arms flanking the Cas9 cleavage site in the BCL11B locus: a left homology arm (LHA) spanning chr14:99,271,219–99,271,818 and a right homology arm (RHA) spanning chr14:99,270,561–99,271,160 (reference genome: NC_000014.9).

The donor construct, driven by the spleen focus-forming virus (SFFV) promoter and terminated by a polyadenylation signal, was cloned into an AAV6 transfer vector (pJEP300-pAAV-CMV-MCS2-pA, Addgene #112138). Recombinant AAV6 particles were produced by triple transfection of HEK293T cells with the MSLN-CAR transfer vector, a serotype 6 rep-cap plasmid, and an adenoviral helper plasmid (E2A, E4, and VA). Virus was harvested after 72 h and concentrated via polyethylene glycol (PEG) precipitation. Titers were quantified by TaqMan-based qPCR using SFFV-specific primers (Thermo Fisher Scientific).

For genome editing, PBMCs (5 × 10⁶ cells per 100 µL) were electroporated on day 0 using the Neon™ Transfection System (Thermo Fisher Scientific) with 8 µg of sgRNA#1 or #2-containing PX459 plasmid (1700 V, 20 ms, 1 pulse). Immediately after electroporation, cells were co-transduced with the AAV6-MSLN-CAR virus and cultured in starting cell medium (SCM) for 1 day. Subsequently, cells were sequentially cultured in reprogramming medium I (1 F-RMI) for 5 days and reprogramming medium II (1 F-RMII) for 12 days, following the 1 F-NK reprogramming protocol.

To confirm site-specific integration of the MSLN-CAR construct at the BCL11B locus, genomic DNA was extracted using the QIAamp DNA Blood Maxi Kit (Qiagen), and PCR was performed using primers flanking the left integration junction: pre-LHA (5′-ACC GAA CCG GGG CAG TTT TA-3′) and the SFFV promoter (5′-TTT TCA TGT ACC CGC CCT TGA T-3′). The CAR transgene consisted of an extracellular single-chain variable fragment (scFv) targeting human mesothelin, a CD8α hinge and transmembrane domain, and intracellular signaling domains comprising the CD28 costimulatory and CD3ζ activation motifs. The full coding sequence is provided in Table S2.

To generate MSLN-CAR-expressing NK-92 cells (MSLN-NK-92), a second-generation lentiviral CAR construct was developed using the pHR CD19-empty CAR backbone (Addgene plasmid #113015). In this construct, the intracellular signaling domains from human CD28 (NCBI Reference Sequence: NM_006139.3) and CD3ζ (CD3Z; NM_000734.3) were retained to mediate co-stimulatory and activation signaling, respectively. The CD19-specific scFv originally present in the CD19-CAR was replaced with a nucleotide sequence encoding MSLN-specific scFv to confer target specificity. Each domain of the CAR construct, including the MSLN scFv, CD8α hinge, CD28 transmembrane and intracellular domains, and CD3ζ domain, was synthesized and assembled using Gibson Assembly-based overlap PCR. The finalized MSLN-CAR cassette was cloned into a lentiviral expression vector, followed by production of replication-incompetent lentiviral particles via transient transfection of HEK293T cells using packaging (psPAX2) and envelope (pMD2.G) plasmids. NK-92 cells were transduced with the resulting lentivirus and selected to establish stable MSLN-CAR-expressing NK-92 cell lines.

To evaluate the surface-binding capability of the mesothelin-targeting scFv used in the CAR construct, we performed a flow cytometry-based binding assay. The enhanced green fluorescent protein (EGFP)-tagged MSLN-CAR vector was generated by inserting an internal ribosome entry site (IRES) and EGFP downstream of the CAR cassette using Gibson Assembly-based overlap PCR (Fig. S6A). HEK293T cells (1 × 10⁶) were transduced with the EGFP-MSLN-CAR lentiviral vector at multiplicities of infection (MOIs) ranging from 0 to 2. After 48 h of culture, EGFP expression was confirmed. To assess binding specificity, transduced HEK293T cells and MSLN-1 F-NKs were incubated with recombinant human mesothelin conjugated to allophycocyanin (APC-MSLN; Cat# MSN-HA2H6, ACROBiosystems) at 2 µL per 1 × 10⁶ cells in 100 µL for 30 min at 4 °C in the dark. Following washing with PBS containing 2% FBS, cells were analyzed using a BD Accuri C6 flow cytometer (BD Biosciences, San Jose, CA, USA). Data were analyzed with FlowJo v10.4.2 (Treestar Inc, Ashland, OR, USA). For HEK293T cells, EGFP⁺MSLN⁺ populations were quantified to assess CAR-mediated binding. For MSLN-1 F-NKs, CD56⁺MSLN⁺ double-positive cells were evaluated to confirm surface mesothelin engagement by CAR-engineered NKs.

Quantitative RT-PCR

Total RNA was extracted using the RNeasy Mini kit (Qiagen, Valencia, CA, USA) and reverse-transcribed into cDNA using the SuperScript VILO™ cDNA synthesis kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. Quantitative polymerase chain reaction (qPCR) was performed using SYBR Green-based detection with the following primers: BCL11B (forward 5’- TCC AGC TAC ATT TGC ACA ACA − 3’; reverse 5’- GCT CCA GGT AGA TGC GGA AG -3’) and GAPDH (forward 5’- GTC TCC TCT GAC TTC AAC AGC G-3’; reverse 5’- ACC ACC CTG TTG CTG TAG CCA A -3’). All primers were synthesized by Macrogen. Gene expression levels were normalized to GAPDH using the ΔΔCt method.

RNA-Seq analysis

Total RNA was isolated from 1 F-NKs, 4 F-NKs, iPSC-NKs, pNKs, and Pan T cells using the RNeasy Mini Kit (Qiagen, Valencia, CA, USA), following the manufacturer’s instructions. RNA quality was assessed using an Agilent 2100 Bioanalyzer equipped with an RNA 6000 Nano Chip (Agilent Technologies, Amstelveen, The Netherlands). RNA concentration was measured with an ND-2000 spectrophotometer (Thermo Fisher Scientific). RNA sequencing libraries were prepared and sequenced on the Illumina NovaSeq 2000 platform (Illumina, CA, USA) to generate 150 bp paired-end reads. Raw sequencing reads were aligned to the human reference genome (GRCh38/hg38) using STAR (Spliced Transcripts Alignment to a Reference). Gene expression was quantified and normalized as counts per million (CPM) using the trimmed mean of M-values (TMM) method. All RNA sequencing and bioinformatic analyses were performed by Ebiogen Inc. (Seoul, Korea).

Flow cytometric analysis

For flow cytometric analysis, live cells were blocked in PBS containing 0.5% BSA and 1 mM EDTA for 20 min at 4 °C. Cells were then stained with FITC-, APC, or PE-conjugated primary antibodies or corresponding isotype controls diluted in PBS containing 1% BSA and 1 mM EDTA for 20 min at 4 °C. After washing with PBS, stained cells were analyzed using a BD Accuri C6 flow cytometer (BD Biosciences). Data were processed and visualized using FlowJo software Version 10.4.2. (Treestar) and Prism 9.0 (GraphPad Software Inc, San Diego, CA, USA). A full list of antibodies used is provided in Table S1.

Cancer cells

Human cancer cell lines were obtained from the American Type Culture Collection (ATCC) and the Korea Cell Line Bank (KCLB, South Korea) and maintained under appropriate culture conditions. Cell lines derived from bladder (253 J), bone (A-673), breast (MCF-7), colon (HCT116), liver (HepG2), prostate (PC-3), skin (SK-MEL-3), and stomach (KATO III) cancers were cultured in DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (P/S). Blood (Raji) and pancreatic (BxPC-3) cancer cell lines from ATCC were maintained in RPMI 1640 with 10% FBS and 1% P/S. Human pancreatic ductal adenocarcinoma (PDAC) cell lines, including AsPC-1, Capan-1, Capan-2, CFPAC-1, MIAPaCa-2, and PANC-1 (KCLB), were also cultured in RPMI 1640 supplemented with 10% FBS and 1% P/S. All cells were incubated at 37 °C in a humidified atmosphere with 5% CO₂ and were routinely tested for mycoplasma contamination.

Calcein-AM- based cytolytic assay

The cytolytic activity of NK cells against various cancer cell lines was assessed using a calcein-AM release assay. Target cells were suspended in RPMI 1640 containing 10% FBS and labeled with calcein-AM (Thermo Fisher Scientific) according to the manufacturer’s instructions. Labeled cells were seeded in 96-well plates and incubated at 37 °C with 5% CO₂ for 1 h. NK effector cells were added at the indicated effector-to-target (E: T) ratios, and plates were centrifuged at 400 × g for 1 min before further incubation for the specified durations. After incubation, supernatants were collected, and fluorescence was measured using a microplate reader (excitation: 485 nm; emission: 535 nm). Specific lysis was calculated using the formula: Lysis (%) = [(measured value - minimum value)/(maximum value - minimum value)] ×100 where the minimum value corresponds to the fluorescence intensity of wells containing only calcein-labeled target cells, and the maximum value corresponds to the fluorescence intensity of wells in which target cells were completely lysed using 2% Triton X-100 (Sigma).

For NKG2D-blocking experiments, NK cells were pre-incubated with 10 µg/mL of either control IgG (clone MOPC-21; BioLegend, San Diego, CA, USA) or anti-NKG2D antibody (clone 1D11; BioLegend) for 30 min before co-culture with PDAC target cells. To evaluate the effects of kinase inhibition on target susceptibility, PANC-1 cells were pre-treated for 48 h with indicated concentrations (0, 10, 100, or 1000 nM) of kinase inhibitors targeting PKMYT1 (RP6306), EIF2AK3 (AMG PERK 44), PRKAA1 (SU6656), IRAK (IRK3 Degrader), NEK2 (NCL00017509), or CHEK1 (PF47736) (all from Tocris Bioscience). A full list of antibodies is provided in Table S1.

Antibody-dependent cell cytotoxicity (ADCC) assay

To assess antibody-mediated cytotoxicity, target cells were labeled with calcein-AM and pre-incubated with serial dilutions of Misitatug, a human anti-MSLN monoclonal antibody, for 30 min at room temperature (RT). NK effector cells were added at the indicated E: T ratios and co-cultured for 2 h at 37 °C in a humidified incubator with 5% CO₂. Supernatants were collected, and calcein release was quantified as described above. ADCC activity was calculated using the formula: ADCC (%) = [(measured value with antibody- measured value without antibody)/(maximum value - minimum value)]×100.

Annexin V/PI-based apoptotic cell death assay

To assess apoptotic cell death, PDAC cells were pretreated with the indicated kinase inhibitors for 48 h at the same concentrations as described above. Effector NK cells were pre-incubated with 10 µg/mL of either control IgG or anti-NKG2D antibody for 20–30 min prior co-culture with pre-treated or untreated target cells. Following a 4 h co-culture, both suspended and adherent cells were harvested (using trypsinization if necessary), washed with PBS, and stained using the Annexin V-APC/PI Apoptosis Detection kit (BioLegend). Cells were incubated with Annexin V-APC and propidium iodide (PI) in binding buffer for 15 min at RT in the dark. Samples were analyzed by flow cytometry (BD Accuri C6) and Annexin V⁺ or PI⁺ populations were quantified as total cell death.

PI staining-based cell death assay

In selected experiments, cell death was assessed by single PI staining. After co-culture with NK cells, target cells were harvested and incubated with 1 µg/mL of PI in PBS for 5–10 min at RT. PI⁺ (non-viable) cells were quantified by flow cytometry, and results were presented as NK-mediated lysis (%). This assay was primarily used for rapid endpoint quantification of necrotic target cells in cytotoxicity assays.

Adherent cell assay

To assess NK adhesion and cytotoxicity against PDAC cells, a live-cell imaging assay was performed with modifications from a previously described method [24]. CFPAC-1 and PANC-1 cells were pre-treated with or without 100 nM of the PKMYT1 inhibitor RP6306 for 24 h and stained with CellTracker™ Green CMFDA (Thermo Fisher Scientific) following the manufacturer’s protocol. Labeled PDAC cells were seeded in 6-well plates at a density of 5 × 10⁵ cells per well and incubated at 37 °C with 5% CO₂ for 6 h to allow for adhesion.

NK cells were pre-stained with CD56-PE and pre-incubated with either isotype IgG or anti-NKG2D antibody (for blocking assays), then added to the adherent PDAC cells at an E: T ratio of 2:1. Co-cultures were incubated for 4 h at 37 °C with 5% CO₂. Microscopic images were captured using an Axio Vert.A1 microscope (Carl Zeiss, Jena, Germany). The number of adherent, viable PDAC cells was quantified by analyzing five randomly selected fields per sample. A full list of antibodies is provided in Table S1.

Conjugate formation assay

To assess NK–PDAC cell conjugation, 1 F-NKs were labeled with CD56-APC and co-incubated with CellTracker™ Green-labeled CFPAC-1 or PANC-1 cells at an E: T ratio of 1:1. PDAC cells were pre-treated with or without 100 nM RP6306 for 24 h. For blocking assays, NK cells were pre-incubated with 50 µg/mL of isotype IgG or anti-NKG2D antibody for 30 min before co-culture. After 1 h of incubation at 37 °C, conjugate formation was evaluated by flow cytometry. Double-positive events (CD56-APC⁺/CellTracker Green⁺) were quantified, and conjugation efficiency was expressed as a percentage of total events. All experiments were performed in triplicate. Antibodies used are listed in Table S1.

Transwell migration assay

Cell migration was assessed using the QCM™ Chemotaxis 24-Well Colorimetric Assay (5 μm pore size; Sigma-Aldrich). The lower chambers were filled with 650 µL of either negative control medium, conditioned medium (CM), or medium supplemented with 100 ng/mL CX3CL1 (clone L393H11, BioLegend). NK cells (1 × 10⁶ cells/well) were suspended in 150 µL of RPMI 1640 containing 1% FBS and 100 IU/mL IL-2 (PeproTech) and added to the upper chambers (inserts). Cells were incubated at 37 °C for 3 h. Migrated cells in the lower chamber were quantified using a hemocytometer. All conditions were tested in triplicate. CM was prepared by seeding PANC-1 or MIAPaCa-2 cells in 6-well plates at 1 × 10⁶ cells per well and incubating overnight to allow for cytokine accumulation in the supernatant.

In vitro tumor environment condition assay

To mimic tumor-associated immunosuppressive conditions, 1 F-NK cells or pNKs were cultured in RPMI 1640 medium supplemented with 200 IU/mL IL-2, 10 ng/mL TGF-β1, 10 ng/mL human IL-15, and 10% FBS at a density of 1 × 10⁶ cells/mL for 48 h. Following incubation, the expression levels of cell surface receptors were assessed by flow cytometry. For cytotoxicity comparison, TGF-β1-conditioned or untreated NK cells were co-cultured with PDAC cells at an E: T ratio of 1:1 for 4 h. Target cell lysis was measured using PI staining followed by flow cytometric analysis.

For 3D spheroid cultures, PANC-1 cells and pancreatic stellate cells (PSCs; ScienCell Research Laboratories, Carlsbad, USA) were co-cultured on poly(2-hydroxyethyl methacrylate) (poly-HEMA)-coated plates. Poly-HEMA was dissolved in 95% ethanol at a final concentration of 2% and stirred overnight at 60 °C. For coating, 60 µL of the prepared poly-HEMA solution was added to each well of a 96-well plate and allowed to dry at RT for 24 h in a laminar flow hood.

Prior to seeding, PANC-1 and PSCs were labeled with CellTracker™ Red CMTPX or CellTracker™ Green CMFDA dyes, respectively. For spheroid formation, cells were seeded in poly-HEMA-coated 96-well plates at a total density of 8 × 10⁴ cells per well. Co-culture spheroids were generated by mixing PANC-1 and PSCs in a 1:1 ratio (4 × 10⁴ cells per cell type). The plates were incubated at 37 °C for 2–4 days to allow spheroid formation. NK cells were then added at an E: T ratio of 1:1 and co-cultured with the spheroids for an additional 4 h before analysis.

Genome-wide CRISPR screening

Genome-wide CRISPR screening was performed to identify genes conferring resistance to NK-mediated cytotoxicity (Fig. 1A). PANC-1 cells stably expressing Cas9 (PANC-1/Cas9) were generated via lentiviral transduction using lentiCas9-Blast (Addgene #52962), followed by blasticidin selection (30 µg/mL; InvivoGen, San Diego, CA, USA) for 28 days. A total of 2 × 10⁸ PANC-1/Cas9 cells were transduced with the Human Whole-Genome Dual-gRNA Library (VectorBuilder Inc., Chicago, IL, USA) at a MOI of 0.5, ensuring a 1,000-fold representation per gRNA. Following puromycin selection (4 µg/mL) for 7 days, 4 × 10⁷ transduced cells per condition were co-cultured with or without NK cells at an E: T ratio of 2:1 for 16 h. Surviving (NK-resistant) and control cells were harvested and stored for genomic analysis. Genomic DNA was extracted using the QIAamp DNA Blood Maxi Kit (Qiagen Inc., Valencia, CA, USA). gRNA sequences were amplified via PCR using the following primers: Forward primer P5 (5’-AAT GAT ACG GCG ACC ACC GAG ATC TAC ACT CTT TCC CTA CAC GAC GCT CTT CCG ATC TCT TGT GGA AAG GAC GAA ACA-3’) and Reverse primer P7 (5’-CAA GCA GAA GAC GGC ATA CGA GAT CGT GAT GTG ACT GGA GTT CAG ACG TGT GCT CTT CCG ATC TTT AAC TTG CTA TTT CTA GCT CTA A-3’). Paired-end 150 bp sequencing was performed using the Illumina NovaSeq platform. After demultiplexing, FASTQ sequence alignment to the library reference file was performed using MAGeCK (Model-based Analysis of Genome-wide CRISPR-Cas9 Knockout) software (version 0.5.9) [25]. Rank plots were visualized in R using MAGeCKFlute (version 2.2.0).

Generation of PKMYT1 knockout cell lines

To validate screen hits, PKMYT1 knockout (KO) PANC-1 cells were generated. sgRNAs targeting PKMYT1 were obtained from the same dual-gRNA library and cloned into a lentiviral dual-sgRNA/Cas9 expression vector (VectorBuilder, ID: VB250427-1315mdb; Chicago, IL, USA). The sgRNA sequences were: sgRNA sequences: sgRNA#1- GGA GGA CGG CCG GCT CTA TG; sgRNA#2- CCG GGC CCG CAA GTT GGC CG. Target cells were transduced with the lentiviral constructs, selected with puromycin (4 µg/mL) for 3 days, and subsequently maintained for an additional 7 days before use in downstream experiments.

Western blot analysis

Cells were lysed, and total protein was separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore, Burlington, NJ, USA). Membranes were blocked and then incubated overnight at 4 °C with primary antibodies, followed by a 1 h incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies (Thermo Fisher Scientific) at RT. Protein bands were visualized using the SuperSignal™ West Pico PLUS chemiluminescent substrate (Thermo Fisher Scientific) and imaged with the Amersham™ Imager 600 system (GE Healthcare). Details of all primary antibodies used are listed in Table S1.

Immunocytochemistry

For immunocytochemistry, cells were fixed with 4% paraformaldehyde in PBS for 15 min and washed four times with PBS. Fixed cells were then blocked and permeabilized in PBS containing 0.2% Triton X-100, 10% FBS, and 1% BSA for 1 h at RT. Cells were incubated with primary antibodies diluted in PBS containing 2% BSA overnight at 4 °C. After washing, cells were incubated with Alexa Fluor 546-conjugated anti-rabbit secondary antibody (1:500; Thermo Fisher Scientific) for 1 h at RT. Fluorescent images were acquired using an Axio Vert.A1 fluorescence microscope (Carl Zeiss). Antibodies used are listed in Table S1.

Enzyme-linked immunosorbent assay (ELISA)

CX3CL1 secretion from PDAC cells was quantified using a commercial ELISA kit (RayBiotech, Peachtree Corners, GA, USA) according to the manufacturer’s protocol. Briefly, PANC-1 cells (2.5 × 10⁵) and MIAPaCa-2 cells (5.0 × 10⁵) were seeded in 6-well plates and treated with either 100 nM PKMYT1 inhibitor RP6306 or DMSO as a control for 24 h. Culture supernatants were collected and filtered through a 0.22 μm syringe filter to remove cellular debris. CX3CL1 concentrations were then measured using the ELISA kit.

Subcutaneous and orthotopic xenograft models

All animal experiments were conducted in accordance with the guidelines of the KRIBB Institutional Animal Care and Use Committee (Animal Welfare Assurance Number: KRIBB-KRIBB-AEC-24105). Male BALB/c-nude mice (6–8 weeks old) were purchased from Orient Bio Co. (Seongnam, South Korea). GFP-luciferase-expressing PANC-1 and CFPAC-1 cells (GFP/Luc-PANC-1 and GFP/Luc-CFPAC-1) were generated via lentiviral transduction using a vector encoding both GFP and luciferase (LL310PA-1; System Biosciences, Palo Alto, CA, USA), with polybrene, following the manufacturer’s instructions.

For the subcutaneous xenograft model, mice were anesthetized with 2% isoflurane and 2 × 10⁶ GFP/Luc-PANC-1 or GFP/Luc-CFPAC-1 cells in 50 µL of Matrigel: PBS (1:1, v/v) were subcutaneously injected into both flanks of each mouse (n = 3 per group) on day 0. NK cells (1 × 10⁷ in 150 µL PBS) were administered intravenously (i.v.) via the tail vein on days 4 and 7. RP6306 (5 mg/kg) was administered intraperitoneally (i.p.) on days 1 and 7. An anti-NKG2D blocking antibody (clone 1D11, BioXCell; 150 µg in 100 µL PBS) was also injected i.p. on days 1 and 7, as needed.

For the orthotopic xenograft model, a 1.5 cm incision was made in the left abdominal flank under 2% isoflurane anesthesia to access the pancreas. A total of 1 × 10⁶ GFP/Luc-PANC-1 or GFP/Luc-CFPAC-1 cells cells in 50 µl of Matrigel: PBS (1:1, v/v) were injected into the pancreatic tail using a 27-gauge syringe. After injection, the pancreas was returned to the abdominal cavity, and the incision was sutured. Mice (n = 4–6 per group) were allowed to recover under observation. NK cells (1.0 × 10⁷ in 150 µl PBS) were injected i.v. on days 4 and 7 in the PANC-1 tumor model, and on days 4, 7, and 14 in the CFPAC-1 tumor model. RP6306 (5 mg/kg) was administered i.p. on days 1 and 7.

To monitor tumor growth, mice were anesthetized with 2% isoflurane and administered D-luciferin (150 mg/kg, Promega) via i.p. injection. Imaging was performed using an IVIS Lumina III system (Perkin Elmer, Waltham, MA, USA), and tumor burden was quantified by bioluminescent intensity. The tumor volume was calculated using the formula: L × W² × 0.52 where L represents the longest tumor diameter and W represents the shortest tumor diameter. Mouse survival was monitored for up to 180 days, at which all mice in the experimental and control groups had expired. Kaplan-Meier analysis was used to evaluate survival, and tissues were harvested post-mortem for histological and molecular assessments.

Blood sample collection

Mice were randomly assigned to five experimental groups: negative control, pNKs, NK-92, 1 F-NK, and MSLN-1 F-NK. Each group received an i.v. injection of NK cells (1.5 × 10⁷ cells in 150 µL PBS via tail vein). Blood was collected via retro-orbital bleeding under 2% isoflurane anesthesia using sterile, heparinized capillary tubes (Cat. No. 2501; Kimble Chase Life Science, Rockwood, TN, USA). Approximately 200 µL of blood was collected per mouse into EDTA-coated tubes and kept on ice. Red blood cells (RBCs) were lysed using ACK lysis buffer (Thermo Fisher Scientific) for 10 min at RT. The remaining cells were washed three times with PBS and resuspended in PBS containing 0.5% BSA and 1 mM EDTA for downstream analysis.

Statistical analysis

All data are presented as mean ± standard deviation (SD). Statistical comparisons between two groups were conducted using unpaired two-tailed Student’s t-test. For comparisons among multiple groups, one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test was used. Kaplan-Meier survival curves were analyzed using the log-rank (Mantel-Cox) test. A P value < 0.05 was considered statistically significant. Significance was denoted as follows: P < 0.001 was marked as “*”, and individual P values between 0.001 and 0.05 were reported explicitly.

Results

Generation of 1 F-NKs and MSLN-1 F-NKs through BCL11B depletion

To streamline NK cell reprogramming, we investigated whether knockdown (KD) of a single transcription factor, BCL11B, could induce NK lineage conversion, thereby enabling the generation of directly reprogrammed NK cells (termed 1 F-NKs) from peripheral blood mononuclear cells (PBMCs) (Fig. 1A). Lentiviral delivery of three independent small hairpin RNAs (shRNAs) targeting BCL11B achieved efficient gene silencing (Figs. 1A-B). In previous work, we established direct reprograming of PBMCs into NK cells (termed 4 F-NKs) using four transcription factors (OCT4, SOX2, KLF4, and cMYC; OSKM) under defined reprogramming conditions using 4 F-NK reprogramming medium (4 F-RM) [13]. When BCL11B KD was applied under these same conditions, CD56⁺CD3⁻ NK cell induction was limited, yielding a low-frequency population (15.66 ± 7.41%; fold change: 17.66 ± 3.30), suggesting that 4 F-RM is suboptimal for single-factor reprogramming (Fig. 1B).

To enhance the efficiency of BCL11B-directed reprogramming, we systematically optimized cytokine and growth factor conditions known to support NK lineage specification. Based on these optimizations, we established a novel single-factor reprogramming medium (1 F-RM), composed of a reprogramming initiation medium (RM I) followed by a reprogramming maturation medium (RM II). This protocol incorporated IL-7, IL-15, and an aryl hydrocarbon receptor (AHR) antagonist, which have been shown to promote NK lineage commitment and maturation.

Using 1 F-RM conditions, BCL11B KD yielded a time-dependent increase in CD56⁺CD3⁻ 1 F-NK cell output, reaching high purity (85.10 ± 9.05%; fold change: 21.50 ± 6.45) by day 18 (Figs. 1B and S1A-D). These cells retained phenotypic stability and proliferative capacity for 5–6 weeks in culture, with over 100-fold expansion observed by day 35 (Fig. 1B, right). Lineage tracing using purified PBMC subsets revealed that CD3 T cells were the only population capable of generating CD56⁺ NK-like cells following BCL11B KD, whereas CD56⁺ NK cells, CD19⁺ B cells, and CD14⁺ monocytes failed to expand or survive (Figs. S1E-G). The resulting 1 F-NKs consistently exhibited a CD56^brightCD16^bright phenotype distinct from conventional CD56^dimCD16⁺ PBMC-derived NK cells (pNKs), suggesting that 1 F-NKs are not derived from pre-existing NK populations. Interestingly, reprogramming efficiency was higher in whole PBMC cultures than in isolated T cells alone, implying a potential role for accessory cells in facilitating lineage conversion via paracrine signaling or microenvironmental cues (Fig. S1F).

To enable antigen-specific targeting, we engineered mesothelin (MSLN)-specific chimeric antigen receptor (CAR)-expressing 1F-NKs (MSLN-1F-NKs) by integrating an MSLN-CAR construct into the BCL11B locus during reprogramming (Figs. 1C-E). CRISPR-Cas9-mediated targeted integration was achieved by co-delivery of BCL11B-specific single guide RNAs (sgRNAs #1/#2) and an AAV vector encoding the MSLN-CAR transgene (Fig. 1D). PCR analysis confirmed successful integration at the BCL11B locus compared to the control (WT-control) (Fig. 1F). Flow cytometry at day 18 showed efficient generation of CD56⁺MSLN⁺ NK cells (45.68 ± 8.01%) in the sgRNA + MSLN-CAR group, compared to negligible expression (~ 0.1%) at day 0 and in the sgRNA-only control (Fig. 1E; P < 0.001), confirming effective CAR knock-in and stable expression during reprogramming.

Characterization of 1 F-NKs and MSLN-1 F-NKs

To evaluate the transcriptional identity and lineage fidelity of 1 F-NKs, we performed transcriptomic profiling in comparison with 4 F-NKs, induced pluripotent stem cells (iPSCs)-derived NKs (iPSC-NKs), pNKs, and primary T cells (Pan T) (Figs. S2-S3). Principal component analysis (PCA) revealed that 1 F-NK cells exhibit a transcriptionally consistent and lineage-restricted identity, characterized by strong intra-group similarity and minimal overlap with other immune lineages (Fig. S2A). Pearson correlation and hierarchical clustering further confirmed high intra-group similarity and close transcriptional proximity between 1 F-NKs and 4 F-NKs, while remaining transcriptionally distinct from T-lineage cells (Figs. S2B-C). While these results support inter-replicate consistency, further investigation using single-cell RNA sequencing will be necessary to fully assess potential intra-population heterogeneity.

Gene ontology (GO) and KEGG pathway analyses of differentially expressed genes (DEGs) showed significant enrichment in immune effector processes and NK-mediated cytotoxic pathways in 1 F-NKs (Fig. S2D). These functional annotations were supported by the expression of NK lineage-defining transcription factors (EOMES and TBX21), activating receptors (FCGR3A [CD16A] and NCAM1), and cytotoxic effectors (GZMA, GZMK, GZMM, and PRF1) (Figs. S3A, D). T cell-associated transcripts (CD3E, CD4, and CD8A) were nearly undetectable (Fig. S3B), further confirming successful lineage redirection. In addition, 1 F-NKs expressed high levels of NK-activating receptors such as KLRK1 (NKG2D), NCR1 (NKp46), and NCR3 (NKp30) (Fig. S3C), along with chemokine receptors involved in tumor homing, including CX3CR1, CXCR1, CXCR2, CCR2, and CCR5 (Fig. S3E).

Flow cytometry confirmed that both 1 F-NKs and MSLN-1 F-NKs shared a CD56^brightCD16^bright phenotype, distinguishing them from the CD56^dimCD16⁺ subset in pNKs and the CD56^brightCD16^dim phenotype in iPSC-NKs, and closely resembling 4 F-NKs (Fig. 1G). Both cell types also expressed high levels of activation markers (CD69, DNAM-1, NKp30, NKp46) and CD16, a receptor critical for antibody-dependent cellular cytotoxicity (ADCC), while showing minimal expression of inhibitory receptors such as KIR2DL1, KIR2DL2, and KIR2DL3, suggesting a poised effector phenotype (Figs. 1H-I). Broad chemokine receptor profiling demonstrated expression of multiple trafficking-associated receptors, including CCR1-10, CXCR1-6, and CX3CR1, with particularly high expression of CCR5, CXCR3, and CX3CR1, supporting potential tumor-homing capabilities (Fig. S4).

Functional assays demonstrated that 1 F-NKs mediated strong cytotoxicity against a diverse panel of human cancer cell lines, including bladder (253j), hematologic (Raji), bone (A-673), breast (MCF-7), colon (HCT116), liver (HepG2), prostate (PC-3), skin (SK-MEL-3), and gastric (KATO III) cancer cells (Fig. 2A). Notably, they exhibited significantly enhanced cytotoxicity against multiple pancreatic ductal adenocarcinoma (PDAC) cell lines, including AsPC-1, BxPC-3, Capan-1, Capan-2, CFPAC-1, MIAPaCa-2, and PANC-1, showing potency comparable to that of 4 F-NKs and superior to both pNKs and NK-92 cells (Fig. 2B).

Fig. 2 — Antitumor cytotoxicity of 1 F-NKs and MSLN-1 F-NKs. **(A)** Cytolytic activity of 1 F-NKs against diverse cancer cell lines measured using a 4 h calcein AM-based assay. Effector-to-target (E: T) ratios tested were 0.25:1, 1:1, and 2.5:1. **(B)** Comparative cytotoxicity 1 F-NKs against pancreatic ductal adenocarcinoma (PDAC) cell lines (AsPC1, Capan-1, Capan-2, CFPAC-1, MIAPaCa-2, and PANC-1), relative to pNKs, NK-92, and 4 F-NKs. NK cells were co-cultured with target cells for 4 h at E: T ratios of 1:1, 2.5:1, or 5:1. Calcein-release assays were used to quantify NK-mediated lysis. Data are presented as mean ± SD (n = 3–5). Two-tailed Student’s t-test; *P < 0.01 vs. pNKs. **(C)** Representative flow cytometry histogram plots showing surface mesothelin (MSLN) expression in a panel of PDAC cell lines: AsPC-1, BxPC-3, Capan-1, Capan-2, CFPAC-1, MIAPaCa-2, and PANC-1). **(D)** Quantification of MSLN surface expression (mean fluorescence intensity; MFI) from (C). Data are mean ± SD (n = 3). Two-tailed Student’s t-test; *P < 0.01 vs. Capan-1. **(E)** Enhanced cytotoxicity of MSLN-1 F-NKs, relative to parental 1 F-NKs. Cells were co-cultured with PDAC cell lines for 4 h at the indicated E: T ratios (1:1, 2.5:1, and 5:1). Lysis was measured using a calcein AM-based assay. Data are mean ± SD (n = 3–5). Two-tailed Student’s t-test; **P <* 0.001 vs. 1 F-NKs

To evaluate antigen-specific activity, we assessed MSLN expression across PDAC cell lines and confirmed positive expression in most lines, with the exception of Capan-1 and BxPC-3 (Figs. 2C-D). MSLN-1 F-NKs exhibited significantly increased cytotoxicity against MSLN⁺ PDAC cells (PANC-1, CFPAC-1, and MIAPaCa-2) relative to non-engineered 1 F-NKs (Fig. 2E). In contrast, no difference in cytolytic activity was observed against MSLN^low/− Capan-1 cells, confirming target specificity. For comparison, we generated MSLN-CAR-engineered NK-92 cells (NK92-CAR) (Figs. S5A-B). Although NK92-CAR cells exhibited enhanced cytotoxicity relative to parental NK-92 cells, MSLN-1 F-NKs showed superior tumor cell lysis under matched conditions (Fig. S5C).

We evaluated antigen recognition by the MSLN-CAR construct using a flow cytometry-based surface binding assay with APC-conjugated recombinant human mesothelin (APC-MSLN). In HEK293T cells transduced with the EGFP-MSLN-CAR construct, we observed a dose-dependent increase in EGFP⁺MSLN⁺ populations, indicating specific surface interaction between the scFv and its cognate antigen (Fig. S6A-B). Similarly, MSLN-1 F-NKs incubated with APC-MSLN demonstrated a marked increase in CD56⁺MSLN⁺ populations compared to untransduced controls, confirming surface binding by CAR-expressing NK cells (Fig. S6C). These findings validate specific CAR-mediated engagement of mesothelin but do not yet define quantitative binding affinity or downstream functional responses, which remain subjects for future investigation.

To determine whether the elevated CD16 expression on 1 F-NKs contribute to enhanced ADCC, we performed calcein-AM-based cytotoxicity assays using MSLN⁺ PDAC cell lines (AsPC-1 and PANC-1) in the presence of anti-MSLN monoclonal antibodies (Fig. S7A). Combined treatment with 1 F-NKs and antibody significantly increased target cell lysis compared to either treatment alone, confirming a CD16-dependent ADCC effect. Moreover, 1 F-NKs induced significantly higher ADCC responses than pNKs under identical conditions (Fig. S7B), indicating that the CD16^high phenotype translates into functionally superior Fc-mediated killing. These results support the potential of 1 F-NKs as a versatile NK platform suitable for integration with monoclonal antibody-based cancer therapies.

NKG2D-NKG2DL axis as a key driver of 1 F-NK cytotoxicity

We hypothesized that the enhanced cytotoxic potential of 1 F-NKs may be attributed, in part, to their high expression of NKG2D, a well-characterized activating receptor essential for tumor recognition (Fig. 3A). Flow cytometry and representative histograms demonstrated significantly elevated NKG2D expression in directly reprogrammed NKs, including 1 F-NKs, MSLN-1 F-NKs, and 4 F-NKs, compared to conventional pNKs, cord blood-derived NK cells (CB-NKs), and iPSC-NKs (Fig. 3B), indicating that transcription factor-mediated reprogramming promotes a more activated phenotype. Analysis of multiple PDAC cell lines (AsPC-1, Capan-1, Capan-2, CFPAC-1, MIAPaCa-2, and PANC-1) revealed heterogeneous expression of NKG2D ligands (MICA/B and ULBP1-6), suggesting variable susceptibility to NKG2D-mediated recognition by NK cells (Fig. 3C).

Fig. 3 — Role of NKG2D signaling in mediating 1 F-NK cytotoxicity against PDAC cells. **(A)** Schematic overview of the NKG2D-ligand interaction axis involved in NK cell-mediated cytotoxicity against PDAC cells. **(B)** Flow cytometry analysis of NKG2D surface expression. Representative histograms (left) and summary bar graphs (right) show NKG2D⁺ cell frequencies and MFI across CB-NKs, pNKs, iPSC-NKs, 4 F-NKs, 1 F-NKs, and MSLN-1 F-NKs. Data are shown as mean ± SD (n = 3). Two-tailed unpaired Student’s t-test; *P < 0.001 **(C)** Flow cytometry analysis of surface expression of NKG2D ligands (MICA, MICB, ULBP1-ULBP6) across PDAC cell lines. **(D-E)** Functional effects of NKG2D blockade in 1 F-NKs co-cultured with PDAC cells (E: T = 2:1, 4 h). **(D)** Cytotoxicity assessed by calcein-AM release. **(E)** CD107a degranulation measured by flow cytometry. NK cells were pretreated for 30 min with no antibody (Control), isotype IgG, or anti-NKG2D antibody (10 µg/mL). Data are normalized to the no-antibody control (set as 100%). **(F)** Representative dot plots showing conjugate formation between CD56⁺ 1 F-NKs and CellTracker Green-labeled PDAC cells after 1 h of co-culture (E: T = 1:1). **(G)** Quantification of conjugate frequency (%), normalized to isotype IgG control. **(H-I)** Analysis of 1 F-NK adherence to PDAC cells following NKG2D blockade. **(H)** Representative fluorescence images showing 1 F-NK attachment to PDAC monolayers in the presence or absence of anti-NKG2D antibody. **(I)** Quantification of adherent live PDAC cells after 4 h of co-culture with 1 F-NKs (E: T = 2:1), normalized to IgG control. Data are presented as mean ± SD; *P < 0.001. **(J)** Effect of marimastat (1 µM, 24 h) on MICA/B shedding. PDAC cells (CFPAC-1 and PANC-1) were stained with anti-MICA/B antibodies, and surface expression was assessed *via* flow cytometry. **(K)** Assessment of NK cytotoxicity against marimastat-treated PDAC cells (E: T = 1:1, 4 h), as determined by propidium iodide (PI) staining. Data represent mean ± SD; *P < 0.001 vs. control. **(L)** Schematic summary illustrating the role of the NKG2D-ligand axis in mediating NK cell cytotoxicity against PDACs

To directly assess the contribution of NKG2D to 1 F-NK function, we performed blocking experiments using a neutralizing anti-NKG2D antibody. NKG2D blockade significantly reduced 1 F-NK-mediated cytotoxicity (Figs. 3D and S8A) and degranulation, as evidenced by decreased CD107a expression (Figs. 3E and S8B). Moreover, disruption of NKG2D signaling impaired immune synapse formation, as shown by reduced NK-PDAC conjugate formation (Figs. 3F-G) and increased PDAC cell viability in co-culture assays (Figs. 3H-I and S8C), reinforcing the functional relevance of this axis in tumor killing.

Given that PDAC tumors frequently evade immune detection by shedding MICA/B through matrix metalloproteinase (MMP) activity, we explored whether inhibition of ligand shedding could restore tumor sensitivity. Treatment with marimastat, a broad-spectrum MMP inhibitor, significantly increased surface MICA/B levels on PDAC cells (Figs. 3J and S8D), enhancing recognition by 1 F-NKs and augmenting cytotoxic responses (Fig. 3K). Collectively, these results establish the NKG2D-NKG2DL interaction as a central mechanism underlying 1 F-NK cytotoxicity and support the rationale for combining NK cell therapy with strategies that prevent NKG2D ligand shedding to overcome tumor immune evasion (Fig. 3L).

PKMYT1 inhibition enhances tumor susceptibility and NK cell infiltration

To identify tumor-intrinsic vulnerabilities that sensitize PDAC cells to 1 F-NK-mediated cytotoxicity, we conducted a genome-wide CRISPR-Cas9 loss-of-function screen in PANC-1 cells (Fig. 4A). The screen identified several immune evasion and survival-related genes, including CD58, ULBP2, ULBP5, NCR3LG1, and HLA family members (Fig. 4B). Among kinase-related hits, PKMYT1 emerged as a promising target alongside ELF2AK3, PRKAA1, IRAK3, NEK2, and CHEK1 (Figs. 4C-D). In functional validation assays, pharmacologic inhibition of PKMYT1 using RP6306 significantly reduced viability and increased apoptosis in PDAC cell lines (CFPAC-1, MIAPaCa-2, and PANC-1), accompanied by increased γ-H2AX expression (Figs. S9A-D).

Fig. 4 — PKMYT1 inhibition enhances 1 F-NK-mediated cytotoxicity against PDAC cells. **(A)** Schematic of genome-wide CRISPR-Cas9 screening in PDAC cells for regulators of NK cytotoxicity sensitivity. **(B)** Rank plot of gene hits based on MAGeCK scores; activating regulators are indicated in red and negative regulators in blue. **(C)** Workflow for selecting druggable kinase targets. **(D)** MAGeCK log₂ fold changes and significance scores for six prioritized kinase candidates. **(E)** PANC-1 cells were pretreated for 48 h with kinase inhibitors at 10, 100, or 1000 nM, then co-cultured with 1 F-NKs (E: T = 1:2) for 4 h; PI-based lysis assay was performed. **(F–G)** PANC-1 cells were pretreated with DMSO or PKMYT1 inhibitor (RP6306) for 48 h and then co-cultured with pNKs or 1 F-NKs (E: T = 2:1) for 4 h. **(F)** Apoptotic cell death was quantified by Annexin V/PI staining. **(G)** CD107a surface expression was analyzed by flow cytometry. **(H–I)** 1 F-NKs were co-cultured with PANC-1 cells ± RP6306 pretreatment (100 nM). **(H)** Representative fluorescence images show CellTracker Green-labeled PANC-1 cells and 1 F-NK adherence. **(I)** Quantification of adherent viable PANC-1 cells across five random fields following 4 h co-culture (E: T = 2:1). **(J)** Surface expression of MICA/B in PANC-1 cells after 24 h treatment with 100 nM RP6306 was measured by flow cytometry; MFI was normalized to control. **(K)** CFPAC-1 and PANC-1 cells were pretreated with DMSO or RP6306 (100 nM, 24 h), then co-cultured with 1 F-NKs (E: T = 1:1, 4 h). 1 F-NKs were pretreated with either IgG or anti-NKG2D antibody (10 µg/mL, 30 min). Cytotoxicity was measured by PI staining. **(L–M)** 1 F-NKs were co-cultured with CellTracker Green-labeled, RP6306-treated PANC-1 cells (E: T = 2:1) for 1 h. **(L)** Representative dot plots show CD56⁺/CellTracker⁺ conjugates. **(M)** Frequency of conjugate formation was quantified. **(N)** Soluble CX3CL1 levels were measured by ELISA in conditioned media (CM) from DMSO- or RP6306-treated PDAC cells. **(O)** CM was used as a chemoattractant in Transwell assays. 1 F-NK migration was measured ± anti-CX3CL1 antibody. **(P)** 1 F-NKs were pretreated with anti-CX3CR1 or isotype control antibody and co-cultured with PANC-1 cells (E: T = 2:1, 4 h); PI staining was used to assess lysis. **(Q)** pNKs or 1 F-NKs were cultured ± TGF-β1. Surface levels of NKG2D and NKp30 were assessed by flow cytometry and normalized to untreated controls. **(R)** pNKs and 1 F-NKs were co-cultured with PANC-1 cells (E: T = 2:1, 4 h) ± TGF-β1; PI staining quantified lysis. **(S)** 3D spheroids of PANC-1 and PSCs were labeled (green/red), co-cultured with NKs (blue, Hoechst-labeled) for 4 h (E: T = 2:1), and stained for CD107a. **(T)** Schematic summarizing the effects of PKMYT1 inhibition on enhancing 1 F-NK cytotoxicity and overcoming TME resistance. Note: PI-based lysis assays were used in panels E, K, P, and R. Two-tailed Student’s t-test; *P < 0.001 vs. control

RP6306 pretreatment dose-dependently enhanced the cytotoxic activity of both 1 F-NKs and MSLN-1 F-NKs (Figs. 4E-F, S10A-B), and increased CD107a degranulation in co-cultured 1 F-NKs, indicating improved effector activation (Fig. 4G). Furthermore, PKMYT1 knockout in PANC-1 and MIAPaCa-2 cells significantly enhanced their sensitivity to 1 F-NK-mediated killing (Figs. S11A-B), suggesting that PKMYT1 regulates immune evasion across heterogeneous PDAC subtypes.

Adhesion assays demonstrated reduced viable, adherent PANC-1 cells after co-culture with 1 F-NKs and RP6306, consistent with enhanced NK cytotoxicity (Figs. 4H-I). This was associated with increased MICA/B surface expression (Figs. 4J and S12A). Importantly, NKG2D blockade abrogated the cytotoxic enhancement, confirming that the NKG2D-MICA/B axis mediates this sensitization (Figs. 4K and S12B). Conjugate formation between 1 F-NKs and RP6306-treated PANC-1 cells was also significantly increased (Figs. 4L-M), supporting improved immune synapse formation via MICA/B upregulation.

Beyond modulating tumor recognition, RP6306 treatment upregulated CX3CL1 secretion in PDAC cells (MIAPaCa-2, and PANC-1) (Fig. 4N), enhancing 1 F-NK migration in transwell assays (Fig. 4O). Supporting this, 1 F-NKs expressed higher levels of CX3CR1 than pNKs (Fig. S13A) and migrated more efficiently toward CX3CL1-containing medium (Fig. S13B). Flow cytometry further confirmed increased CX3CL1 expression in multiple RP6306-treated PDAC cell lines (Figs. S13C-D). Blocking CX3CR1 on 1 F-NKs reduced their cytotoxicity against RP6306-treated tumor cells (Fig. 4P and S12C), demonstrating the importance of the CX3CL1-CX3CR1 axis in mediating NK infiltration and function.

The TME in PDAC is known to suppress NK cytotoxicity, primarily via TGF-β-induced downregulation of activating receptors such as NKG2D and NKp30 [26, 27]. Notably, 1 F-NKs retained higher NKG2D and NKp30 expression than pNKs in the presence of TGF-β (Figs. 4Q and S12D) and maintained superior cytotoxicity under immunosuppressive conditions (Figs. 4R and S12E). To assess cytotoxicity in a more physiologically relevant model, we utilized a 3D co-culture spheroid model of PDAC by embedding PANC-1 cells with pancreatic stellate cells (PSCs), thereby recapitulating the fibrotic TME (Fig. 4S, left). In this setting, 1 F-NKs significantly reduced the frequency of PANC-1 cells within the spheroids compared to pNKs, confirming their superior cytotoxicity in a TME-like context (Figs. 4S, left; S14A). Additionally, CD107a degranulation was significantly higher in 1 F-NKs than in pNK-treated spheroids, indicating enhanced NK activation and cytotoxic function within the TME (Figs. 4S, right; S14B).

Taken together, these findings demonstrate that PKMYT1 inhibition enhances tumor susceptibility to 1 F-NK-mediated cytotoxicity by improving both immune recognition and NK cell infiltration. These effects are mediated through NKG2D-MICA/B and CX3CL1-CX3CR1 axes and remain functional even in suppressive microenvironments, supporting the clinical utility of combining PKMYT1i with reprogrammed NK-based therapies (Fig. 4T).

In vivo antitumor efficacy of 1 F-NKs, MSLN-1 F-NKs, and PKMYT1i in PDAC subcutaneous models

To evaluate the therapeutic efficacy of 1 F-NKs and MSLN-1 F-NKs, we established subcutaneous xenograft models using luciferase-expressing CFPAC-1 (CFPAC1-Luc) (Fig. 5A) and PANC-1 (PANC-1-Luc) human PDAC cells (Fig. 5G) in Balb/c nude mice. Mice were received intravenously (i.v.) administered either of PBS or 1.0 × 10⁷ NKs (pNKs, NK-92, 1 F-NKs, or MSLN-1 F-NKs) on days 4 and 7 post tumor implantation (Fig. 5A).

Fig. 5 — Antitumor activity of 1 F-NKs and MSLN-1 F-NKs in a subcutaneous xenograft model. **(A)** Experimental timeline evaluating the in vivo antitumor effects of 1 F-NKs and MSLN-1 F-NKs. Mice bearing subcutaneous CFPAC-1-Luc xenografts were intravenously (*i.v.*) injected with PBS (Control) or 1.0 × 10⁷ of NK cells (pNKs, NK-92, 1 F-NKs, or MSLN-1 F-NKs) on days 4 and 7 post-tumor inoculation (n = 3 *per* group). **(B)** Representative bioluminescence images collected at the indicated time points. **(C)** Quantification of bioluminescence signals on day 35. **(D)** Representative images of harvested tumors collected on day 35. **(E)** Tumor volume quantification from panel (D). **(F)** Flow cytometry analysis of hCD45⁺ NK cells in mouse peripheral blood on day 14 following NK administration. **(G)** Experimental design for testing the combination of 1 F-NKs and PKMYT1i in PANC-1-Luc subcutaneous xenografts. Mice were treated with PBS, 1.0 × 10⁷ 1 F-NKs, or 1 F-NKs with intraperitoneal PKMYT1i or vehicle (DMSO) (n = 3 *per* group). **(H)** Representative bioluminescence images at indicated time points. **(I)** Quantification of bioluminescence on day 28. **(J)** Representative images of tumors collected on day 28. **(K)** Tumor volume quantification from panel (J). **(L)** Ex-vivo imaging of isolated tumors on day 3 post-inoculation. Tumors were assessed for NK infiltration using DiR-labeled NKs and luciferase signal from PANC-1-Luc cells. **(M)** Quantification of the relative infiltration based on DiR and Luc signal ratios from panel (L). **(N)** Immunofluorescence staining of tumor tissues collected on day 28. MICA and CX3CL1 were labelled in green, and nuclei were counterstained with DAPI (blue). **(O)** Quantitative analysis of MICA/B and CX3CL1 staining intensity from panel (N). All data are presented as mean ± SD. Statistical significance was assessed by one-way ANOVA followed by post hoc multiple comparisons. *P < 0.001 vs. control

Bioluminescence imaging (BLI) revealed a significant tumor suppression in the 1 F-NK group, compared to PBS, pNKs, and NK-92 controls (Figs. 5B-C and S15A). MSLN-1 F-NKs demonstrated superior efficacy in CFPAC-1 xenografts, underscoring enhanced tumor targeting by CAR modification (Figs. 5B-C and S15A). By day 35, 1 F-NKs reduced tumor volume to 882.95 ± 39.93 mm³ (n = 4), compared to PBS (2,169 ± 220.80 mm³, n = 4), NK-92 (1527.44 ± 68.04 mm³, n = 4), and pNKs (1,270.01 ± 30.85 mm³, n = 4) (Figs. 5D-E). MSLN-1 F-NKs achieved the most pronounced tumor regression (550.89 ± 18.49 mm³, n = 4), confirming superior antitumor potency (Figs. 5D-E; S16A, left). Peripheral blood analysis showed higher persistence of human CD45⁺ NK cells in mice treated with 1 F-NKs (1.22 ± 0.09%) or MSLN-1 F-NKs (1.21 ± 0.17%) compared to pNKs (0.59 ± 0.18%) or NK-92 (0.52 ± 0.08%) at day 7, sustained through day 21 (Figs. 5F and S17). No significant body weight changes were observed (Fig. S17B, left), indicating favorable tolerability. To evaluate NK cell infiltration into tumors, DiR-labeled 1 F-NKs and MSLN-1 F-NKs were tracked using ex vivo fluorescence imaging (Fig. S18A). Quantitative analysis revealed that both NK cell types exhibited tumor localization, with MSLN-1 F-NKs demonstrating significantly higher accumulation (1.60 ± 0.09) compared to 1 F-NKs (0.71 ± 0.02) and pNKs (0.33 ± 0.01) (Fig. S18B). These findings indicate enhanced tumor trafficking and retention of MSLN-CAR-engineered NK cells, supporting their improved therapeutic efficacy in targeting PDAC.

To interrogate the in vivo relevance of NKG2D signaling, we employed an anti-NKG2D blocking antibody in PANC-1-Luc xenografts (Fig. S19A). 1 F-NK monotherapy resulted in robust tumor suppression (6.73 × 10⁹ ± 9.2 × 10⁸ radiance), which was significantly attenuated by NKG2D blockade (1.52 × 10¹⁰ ± 2.6 × 10⁹), restoring tumor burden to PBS (2.59 × 10¹⁰ ± 6.3 × 10⁹) and anti-NKG2D-only (2.12 × 10¹⁰ ± 3.2 × 10⁹) levels (Figs. S19B-C). Correspondingly, tumor volume was smallest in the 1 F-NK group (326.49 ± 43.64 mm³), with partial rescue by NKG2D blockade (654.00 ± 43.63 mm³), and minimal effect from controls (Figs. S19D-E). These results confirm that NKG2D signaling is essential for 1 F-NK-mediated tumor control in vivo.

We next examined the combinatorial efficacy of 1 F-NKs and PKMYT1 inhibition using RP6306 in PANC-1-Luc-bearing mice (Fig. 5G). Mice were treated with PBS, 1 F-NKs (1.0 × 10⁷ cells, i.v. on days 4 and 7), PKMYT1i [5 mg/kg, intraperitoneally (i.p.) on days 1 and 7], or a combination of both (Fig. 5G). Monotherapies with 1 F-NKs or PKMYT1i partially suppressed tumor growth, whereas combination therapy significantly reduced tumor burden (Figs. 5H-I and S15B). On day 28, BLI signal intensity in the combination group was 1.75 × 10⁸ ± 1.4 × 10⁷, compared to PBS (2.19 × 10¹⁰ ± 7.1 × 10⁸), 1 F-NKs alone (3.81 × 10⁸ ± 1.9 × 10⁷), and PKMYT1i alone (4.41 × 10⁸ ± 1.8 × 10⁷) (Figs. 5H-I). Final tumor volumes confirmed synergistic efficacy: 155.09 ± 12.31 mm³ in the combination group vs. 383.5 ± 61.57 mm³ (1 F-NKs), 617.19 ± 95.02 mm³ (PKMYT1i), and 1,686.92 ± 61.83 mm³ (PBS) (Figs. 5J-K; S16A, right). Treatment was well tolerated, with no weight loss observed (Fig. S16, left).

To assess NK infiltration, DiR-labeled 1 F-NKs were tracked via ex vivo fluorescence imaging (Fig. 5L). Tumor accumulation was significantly higher in the combination group (2.36 ± 0.4) vs. 1 F-NKs alone (0.73 ± 0.07) or PBS (0.003 ± 0.001) (Fig. 5M). Immunohistochemistry (IHC) staining revealed increased MICA and CX3CL1 expression in tumors from the combination group (62.28 ± 2.52% and 78.13 ± 2.17%, respectively), compared to 1 F-NKs alone (17.88 ± 1.68% and 24.78 ± 2.13%) and PBS (10.10 ± 1.22% and 9.55 ± 0.77%) (Figs. 5N-O). These findings support that PKMYT1 inhibition augments tumor susceptibility and 1 F-NK infiltration by upregulating MICA and CX3CL1, enhancing NKG2D- and CX3CR1-mediated tumor targeting.

In vivo antitumor efficacy of 1 F-NKs, MSLN-1 F-NKs, and PKMYT1i in PDAC orthotopic models

To assess the therapeutic efficacy of 1 F-NKs and MSLN-1 F-NKs in a physiologically relevant tumor setting, we established orthotopic xenograft models using luciferase-expressing CFPAC-1 (CFPAC-1-Luc) cells in Balb/c nude mice (Fig. 6A). Mice received intravenous administration of PBS, 1 F-NKs, or MSLN-1 F-NKs (1.0 × 10⁷ cells) on days 4, 7, and 14 post-implantation. BLI revealed significantly reduced tumor burden in the MSLN-1 F-NK group compared to PBS and 1 F-NK monotherapy (Figs. 6B-C). By day 35, average BLI signal intensity in the MSLN-1 F-NK group (5.34 × 10⁸ ± 5.8 × 10⁷ radiance, n = 6) was significantly lower than in the PBS (1.44 × 10¹⁰ ± 1.7 × 10⁹) and 1 F-NK (3.24 × 10⁹ ± 1.8 × 10⁸) groups (Figs. 6B-C and S15C).

Ex vivo BLI imaging of liver, spleen, and kidney revealed significantly reduced metastatic dissemination in MSLN-1 F-NK-treated mice (Figs. 6D-E). Quantification showed that metastatic burden in the MSLN-1 F-NK group was markedly lower in the liver (19.78 ± 0.51%), spleen (3.55 ± 1.32%), and kidney (4.95 ± 1.35%) compared to PBS (liver: 73.49 ± 2.75%; spleen: 39.54 ± 4.22%; kidney: 25.38 ± 1.52%) and 1 F-NK-treated mice (liver: 29.31 ± 2.26%; spleen: 9.12 ± 1.27%; kidney: 19.60 ± 1.01%) (Fig. 6F). Kaplan-Meier survival analysis showed a trend toward prolonged survival in the MSLN-1 F-NK group compared to 1 F-NKs, though the p-value did not reach statistical significance (p = 0.0625) (Fig. 6G). No significant body weight changes were observed (Fig. S16B, right), indicating good tolerability.

To evaluate the combinatorial efficacy of 1 F-NKs with PKMYT1 inhibition, we employed an orthotopic PANC-1-Luc model (Fig. 6H). Mice were treated with PBS, pNKs, 1 F-NKs, PKMYT1i (5 mg/kg, i.p. on days 1 and 7), or combination therapies (PKMYT1i + pNKs or PKMYT1i + 1 F-NKs). BLI analysis showed that the PKMYT1i + 1 F-NK group exhibited the most significant tumor suppression among all groups (Figs. 6I-J). By day 28, BLI intensity in the combination group (5.32 × 10⁸ ± 3.7 × 10⁷ radiance, n = 6) was markedly reduced compared to PBS (1.44 × 10¹⁰ ± 2 × 10⁹), PKMYT1i alone (8.9 × 10⁸ ± 6.1 × 10⁷), pNKs (5.47 × 10⁹ ± 8.1 × 10⁸), 1 F-NKs (2.96 × 10⁹ ± 6 × 10⁸), and PKMYT1i + pNKs (8.09 × 10⁸ ± 3.4 × 10⁷) (Figs. 6J, S15D). Body weight remained stable in all groups (Fig. S16C, right).

We further evaluated the therapeutic synergy of MSLN-1 F-NKs with PKMYT1i in the same orthotopic model (Fig. S20A). Mice were treated with PBS, 1 F-NKs, MSLN-1 F-NKs (1.0 × 10⁷ cells, i.v. on days 4 and 7), PKMYT1i (5 mg/kg, i.p. on days 1 and 7), or combination regimens (PKMYT1i + 1 F-NKs and PKMYT1i + MSLN-1 F-NKs). The PKMYT1i + MSLN-1 F-NK group displayed the most pronounced antitumor effect, with BLI signals reduced to 4.15 × 10⁸ ± 2.2 × 10⁷ radiance (n = 4) by day 28, significantly lower than PBS (2.41 × 10¹⁰ ± 6.2 × 10⁹), PKMYT1i (6.97 × 10⁹ ± 1.5 × 10⁹), 1 F-NKs (4.82 × 10⁹ ± 9.4 × 10⁸), MSLN-1 F-NKs (8.49 × 10⁸ ± 1.9 × 10⁸), and PKMYT1i + 1 F-NKs (6.38 × 10⁸ ± 1.7 × 10⁸) (Figs. S20B-C).

Ex vivo BLI imaging confirmed significantly reduced metastatic lesions in the liver, spleen, and kidney in the PKMYT1i + 1 F-NK group (Fig. 6K). Quantitative analysis revealed the lowest metastatic burden in this group, liver: 10.61 ± 0.38%; spleen: 8.21 ± 0.92%; kidney: 9.10 ± 1.40%, compared to PBS (liver: 92.23 ± 3.21%), and other treatments, including pNKs, 1 F-NKs, PKMYT1i alone, and PKMYT1i + pNKs (Fig. 6L). Notably, Kaplan-Meier survival analysis demonstrated significantly prolonged survival in the PKMYT1i + 1 F-NK group (p = 0.0013) (Fig. 6M). Mechanistically, the enhanced antitumor activity observed with combination therapy is attributed to PKMYT1i-mediated upregulation of NKG2D ligands (MICA/B), which potentiates NKG2D-NKG2DL axis signaling, augments NK infiltration via the CX3CR1-CX3CL1 pathway, and enhances cytolytic activity against tumor cells.

Discussion

This study presents a single-transcription-factor-directed strategy for generating cytotoxic NK cells, termed single-factor-induced NK cells (1 F-NKs), through targeted depletion of BCL11B. By avoiding the use of pluripotency-inducing factors, this approach provides a simplified and defined route to reprogram peripheral lymphocytes into functional NK effectors. In doing so, it addresses critical limitations of current NK cell platforms, including reliance on complex culture systems, inconsistent cytotoxicity, and functional impairment within immunosuppressive tumor microenvironments (TME), particularly in highly refractory cancers such as pancreatic ductal adenocarcinoma (PDAC).

BCL11B is a well-established lineage-specifying transcription factor that represses NK-associated gene expression while maintaining T cell identity [28, 29]. Previous studies have shown that BCL11B downregulation can induce partial T-to-NK conversion in ITNK models [14, 29–31]; however, these efforts often resulted in heterogeneous populations retaining T cell features, complicating translational potential [14, 31, 32]. In contrast, the 1 F-NK approach generated a highly consistent CD56⁺CD16⁺ NK population without feeder layers [33, 34] or pluripotency factor engagement [13], supporting its scalability, standardization, and clinical feasibility.

Compared to NK cells reprogrammed via OSKM (OCT4, SOX2, KLF4, and MYC) (4 F-NKs) [13], which transiently engage pluripotency programs and raise tumorigenic safety concerns, 1 F-NKs are generated through selective BCL11B suppression and demonstrated stable acquisition of NK lineage identity. Transcriptomic and phenotypic analyses revealed preservation of key effector functions, cytotoxicity, activating receptor expression, and CD16-mediated ADCC, at levels comparable to 4 F-NKs. Building on this foundation, the use of clinically validated, non-integrating gene delivery methods (e.g., synthetic mRNAs, episomal vectors, or Sendai virus) offers a promising route for scalable and safe reprogramming. Although AAV-based systems have demonstrated clinical utility, their suitability depends on the specific application, and careful evaluation of genomic stability and insertional mutagenesis remains essential. In parallel, ongoing efforts to enhance donor-to-donor consistency, batch-to-batch reproducibility, and GMP-compliant manufacturing are expected to facilitate robust and standardized production pipelines.

Phenotypically, 1 F-NKs exhibited a distinct CD56^brightCD16^bright phenotype, differing from canonical peripheral subsets such as the CD56^dimCD16^bright cytotoxic and the CD56^brightCD16^dim immunoregulatory subset [35–37]. This hybrid phenotype, alongside the upregulation of chemokine receptors (CXCR3, CCR5, CX3CR1) and activating receptors (NKG2D, DNAM-1, NKp30, NKp44, NKp46), may suggest a potential for enhanced tissue migration and cytotoxic engagement [38, 39]. Notably, 1 F-NKs retained robust expression of CD16 (FcγRIIIa), a receptor critical for mediating antibody-dependent cellular cytotoxicity (ADCC) [40], which is often downregulated in tumor-infiltrating NK cells due to chronic TME exposure [41, 42]. When combined with anti-mesothelin antibodies, 1 F-NKs demonstrated superior ADCC activity against mesothelin-positive PDAC cells compared to peripheral NK cells, underscoring their potential for antibody-guided immunotherapy. Moreover, 1 F-NKs retained functional resilience under TGF-β–rich conditions, a hallmark of solid tumor immunosuppression [26, 27, 43], supporting their potential to function effectively within hostile TMEs. Future work aimed at characterizing the molecular basis of this TME resistance, and extending testing across diverse antigen-antibody pairs and donor sources, is expected to further expand their therapeutic relevance.

To confer tumor specificity, a targeted knock-in strategy was used to introduce mesothelin-targeted CAR constructs into the BCL11B locus during reprogramming, generating MSLN-1 F-NKs. This dual-engineering approach allowed for simultaneous reprogramming and CAR expression within a unified manufacturing workflow. Stable CAR expression and improved cytotoxicity in MSLN⁺ PDAC models support the utility of this method. The flexibility of this system allows for substitution with other tumor-specific antigens (e.g., CD70 [44], PSCA [45]), and future evaluation in patient-derived xenografts or organoid models will help assess its broad applicability across PDAC subtypes.

In parallel, we explored tumor-intrinsic resistance mechanisms to NK cell-mediated cytotoxicity. PDAC cells are known to shed MICA/B ligands via matrix metalloproteinases (MMPs), reducing ligand availability for NKG2D and impairing NK cell recognition [46]. Restoration of MICA/B surface expression with marimastat, an MMP inhibitor, enhanced 1 F-NK-mediated cytotoxicity. Additionally, a genome-wide CRISPR-Cas9 screen identified PKMYT1, a G2/M checkpoint kinase frequently upregulated in PDAC, as a modulator of NK resistance [19, 47–49]. Pharmacologic inhibition of PKMYT1 using RP6306 induced DNA damage responses, upregulated MICA/B and CX3CL1, and enhanced tumor susceptibility to 1 F-NK cytotoxicity. NKG2D blockade significantly impaired this sensitization, confirming its role in mediating RP6306-induced tumor susceptibility [20, 21]. CRISPR-mediated PKMYT1 knockout further validated these findings across PDAC cell lines (PANC-1 and MIAPaCa-2). In vivo, RP6306 treatment enhanced tumor susceptibility to 1 F-NKs, and combination of MSLN-1 F-NKs with RP6306 achieved improved anti-tumor efficacy relative to either treatment alone. Notably, while NKG2D blockade partially impaired efficacy, this finding also suggests the presence of additional effector mechanisms that contribute to tumor control. As PKMYT1 is a cell cycle regulator, safety profiling and tumor-selective delivery strategies will be essential for future clinical application.

Collectively, this study outlines a modular strategy that integrates transcription factor-guided reprogramming, CAR-based tumor targeting, and pharmacologic sensitization to address key bottlenecks in NK cell therapy, including effector generation, immune escape, and tumor specificity. The adaptability of the platform to different tumor antigens and resistance mechanisms provides a strong foundation for extending this approach to other hard-to-treat malignancies. Future work will focus on refining manufacturing protocols, conducting toxicologic and biodistribution studies, and evaluating efficacy in immunocompetent and patient-derived models. Rather than relying on a single modality, our results highlight the potential of a combinatorial immunotherapy framework that coordinates effector engineering, antigen targeting, and tumor vulnerability modulation to achieve meaningful clinical outcomes in refractory cancers such as PDAC.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1^{(3.3MB, docx)}

Acknowledgements

The authors thank the Korean Fund for Regenerative Medicine (23A0106L1) for its administrative and collaborative support.

Author contributions

Y.S.C. and H.S.K. designed the study; H.S.K., J. Y. K., J. L., B. S. and J. E. C. performed experiments. H.S.K., J. Y. K., J. L., B. S. and J. E. C. analysed the data. Y.S.C. supervised the work. Y.S.C. and H.S.K. wrote the manuscript with input from all authors.

Funding

This work was supported by the Korean Fund for Regenerative Medicine (23A0106L1) and, the KRIBB research initiative program (KGM1232511), and the National Research Foundation of Korea (NRF) grant (RS-2023-00272798).

Data availability

MAGeCK analysis https://figshare.com/account/items/28588673/edit.

Declarations

Competing interests

The authors declare no competing interests.

Ethics declarations

The other authors declare no competing interests.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1.Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin. 2024;74(1):12–49. [DOI] [PubMed] [Google Scholar]
2.Park W, Chawla A, O’Reilly EM. Pancreatic cancer: A review. JAMA. 2021;326(9):851–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Neesse A, Algul H, Tuveson DA, Gress TM. Stromal biology and therapy in pancreatic cancer: a changing paradigm. Gut. 2015;64(9):1476–84. [DOI] [PubMed] [Google Scholar]
4.Cannon A, Thompson C, Hall BR, Jain M, Kumar S, Batra SK. Desmoplasia in pancreatic ductal adenocarcinoma: insight into pathological function and therapeutic potential. Genes Cancer. 2018;9(3–4):78–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Mascharak S, Guo JL, Foster DS, Khan A, Davitt MF, Nguyen AT, Burcham AR, Chinta MS, Guardino NJ, Griffin M, et al. Desmoplastic stromal signatures predict patient outcomes in pancreatic ductal adenocarcinoma. Cell Rep Med. 2023;4(11):101248. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Huber M, Brehm CU, Gress TM, Buchholz M, Alashkar Alhamwe B, von Strandmann EP, Slater EP, Bartsch JW, Bauer C, Lauth M. The Immune Microenvironment in Pancreatic Cancer. Int J Mol Sci 2020, 21(19). [DOI] [PMC free article] [PubMed]
7.Carlomagno S, Setti C, Ortolani F, Sivori S. Pancreatic ductal adenocarcinoma microenvironment: soluble factors and cancer associated fibroblasts as modulators of NK cell functions. Immunol Lett. 2024;269:106898. [DOI] [PubMed] [Google Scholar]
8.Pitt JM, Marabelle A, Eggermont A, Soria JC, Kroemer G, Zitvogel L. Targeting the tumor microenvironment: removing obstruction to anticancer immune responses and immunotherapy. Ann Oncol. 2016;27(8):1482–92. [DOI] [PubMed] [Google Scholar]
9.Laskowski TJ, Biederstadt A, Rezvani K. Natural killer cells in antitumour adoptive cell immunotherapy. Nat Rev Cancer. 2022;22(10):557–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Wang L, Chen Z, Liu G, Pan Y. Functional crosstalk and regulation of natural killer cells in tumor microenvironment: significance and potential therapeutic strategies. Genes Dis. 2023;10(3):990–1004. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Van Audenaerde JRM, De Waele J, Marcq E, Van Loenhout J, Lion E, Van den Bergh JMJ, Jesenofsky R, Masamune A, Roeyen G, Pauwels P, et al. Interleukin-15 stimulates natural killer cell-mediated killing of both human pancreatic cancer and stellate cells. Oncotarget. 2017;8(34):56968–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Li W, Wang X, Zhang X, Aziz AUR, Wang D. CAR-NK cell therapy: A transformative approach to overcoming oncological challenges. Biomolecules 2024, 14(8). [DOI] [PMC free article] [PubMed]
13.Kim HS, Kim JY, Seol B, Song CL, Jeong JE, Cho YS. Directly reprogrammed natural killer cells for cancer immunotherapy. Nat Biomed Eng. 2021;5(11):1360–76. [DOI] [PubMed] [Google Scholar]
14.Li P, Burke S, Wang J, Chen X, Ortiz M, Lee SC, Lu D, Campos L, Goulding D, Ng BL, et al. Reprogramming of T cells to natural killer-like cells upon Bcl11b deletion. Science. 2010;329(5987):85–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Liu P, Li P, Burke S. Critical roles of Bcl11b in T-cell development and maintenance of T-cell identity. Immunol Rev. 2010;238(1):138–49. [DOI] [PubMed] [Google Scholar]
16.Korchagina AA, Shein SA, Koroleva E, Tumanov AV. Transcriptional control of ILC identity. Front Immunol. 2023;14:1146077. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Avula LR, Rudloff M, El-Behaedi S, Arons D, Albalawy R, Chen X, Zhang X, Alewine C. Mesothelin enhances tumor vascularity in newly forming pancreatic peritoneal metastases. Mol Cancer Res. 2020;18(2):229–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Le K, Wang J, Zhang T, Guo Y, Chang H, Wang S, Zhu B. Overexpression of mesothelin in pancreatic ductal adenocarcinoma (PDAC). Int J Med Sci. 2020;17(4):422–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Tomovic Pavlovic K, Kocic G, Smelcerovic A. Myt1 kinase inhibitors - Insight into structural features, offering potential frameworks. Chem Biol Interact. 2024;391:110901. [DOI] [PubMed] [Google Scholar]
20.Korbecki J, Siminska D, Kojder K, Grochans S, Gutowska I, Chlubek D, Baranowska-Bosiacka I. Fractalkine/CX3CL1 in neoplastic processes. Int J Mol Sci 2020, 21(10). [DOI] [PMC free article] [PubMed]
21.Lu X. Structure and function of ligand CX3CL1 and its receptor CX3CR1 in cancer. Curr Med Chem. 2022;29(41):6228–46. [DOI] [PubMed] [Google Scholar]
22.Kim HS, Lee J, Lee DY, Kim YD, Kim JY, Lim HJ, Lim S, Cho YS. Schwann cell precursors from human pluripotent stem cells as a potential therapeutic target for Myelin repair. Stem Cell Rep. 2017;8(6):1714–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Zhu H, Kaufman DS. An improved method to produce Clinical-Scale natural killer cells from human pluripotent stem cells. Methods Mol Biol. 2019;2048:107–19. [DOI] [PubMed] [Google Scholar]
24.Ruscetti M, Leibold J, Bott MJ, Fennell M, Kulick A, Salgado NR, Chen CC, Ho YJ, Sanchez-Rivera FJ, Feucht J, et al. NK cell-mediated cytotoxicity contributes to tumor control by a cytostatic drug combination. Science. 2018;362(6421):1416–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Li W, Xu H, Xiao T, Cong L, Love MI, Zhang F, Irizarry RA, Liu JS, Brown M, Liu XS. MAGeCK enables robust identification of essential genes from genome-scale CRISPR/Cas9 knockout screens. Genome Biol. 2014;15(12):554. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Ren Y, Xue M, Hui X, Liu X, Farooq MA, Chen Y, Ji Y, Duan Y, Ajmal I, Yao J, et al. Chimeric cytokine receptor TGF-beta RII/IL-21R improves CAR-NK cell function by reversing the immunosuppressive tumor microenvironment of gastric cancer. Pharmacol Res. 2025;212:107637. [DOI] [PubMed] [Google Scholar]
27.Wu X, Xiao Y, Guo D, Zhang Z, Liu M. Reduced NK cell cytotoxicity by Papillomatosis-Derived TGF-beta contributing to Low-Risk HPV persistence in JORRP patients. Front Immunol. 2022;13:849493. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Liao R, Wu Y, Qin L, Jiang Z, Gou S, Zhou L, Hong Q, Li Y, Shi J, Yao Y, et al. BCL11B and the NuRD complex cooperatively guard T-cell fate and inhibit OPA1-mediated mitochondrial fusion in T cells. EMBO J. 2023;42(21):e113448. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Forkel H, Grabarczyk P, Depke M, Troschke-Meurer S, Simm S, Hammer E, Michalik S, Hentschker C, Corleis B, Loyal L, et al. BCL11B depletion induces the development of highly cytotoxic innate T cells out of IL-15 stimulated peripheral blood alphabeta CD8 + T cells. Oncoimmunology. 2022;11(1):2148850. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Maluski M, Ghosh A, Herbst J, Scholl V, Baumann R, Huehn J, Geffers R, Meyer J, Maul H, Eiz-Vesper B, et al. Chimeric antigen receptor-induced BCL11B suppression propagates NK-like cell development. J Clin Invest. 2019;129(12):5108–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Sottile R, Panjwani MK, Lau CM, Daniyan AF, Tanaka K, Barker JN, Brentjens RJ, Sun JC, Le Luduec JB, Hsu KC. Human cytomegalovirus expands a CD8(+) T cell population with loss of BCL11B expression and gain of NK cell identity. Sci Immunol. 2021;6(63):eabe6968. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Jiang Z, Qin L, Tang Y, Liao R, Shi J, He B, Li S, Zheng D, Cui Y, Wu Q, et al. Human induced-T-to-natural killer cells have potent anti-tumour activities. Biomark Res. 2022;10(1):13. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Silla L. Double-bright (CD56bright/CD16bright) natural killer cell adoptive immunotherapy for SARS-CoV-2. Br J Haematol. 2020;190(6):e322–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Silla L, Valim V, Pezzi A, da Silva M, Wilke I, Nobrega J, Vargas A, Amorin B, Correa B, Zambonato B, et al. Adoptive immunotherapy with double-bright (CD56(bright) /CD16(bright)) expanded natural killer cells in patients with relapsed or refractory acute myeloid leukaemia: a proof-of-concept study. Br J Haematol. 2021;195(5):710–21. [DOI] [PubMed] [Google Scholar]
35.Amand M, Iserentant G, Poli A, Sleiman M, Fievez V, Sanchez IP, Sauvageot N, Michel T, Aouali N, Janji B, et al. Human CD56(dim)CD16(dim) cells as an individualized natural killer cell subset. Front Immunol. 2017;8:699. [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Saito LM, Ortiz RC, Amor NG, Lopes NM, Buzo RF, Garlet GP, Rodini CO. NK cells and the profile of inflammatory cytokines in the peripheral blood of patients with advanced carcinomas. Cytokine. 2024;174:156455. [DOI] [PubMed] [Google Scholar]
37.Poli A, Michel T, Theresine M, Andres E, Hentges F, Zimmer J. CD56bright natural killer (NK) cells: an important NK cell subset. Immunology. 2009;126(4):458–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Dogra P, Rancan C, Ma W, Toth M, Senda T, Carpenter DJ, Kubota M, Matsumoto R, Thapa P, Szabo PA, et al. Tissue determinants of human NK cell development, function, and residence. Cell. 2020;180(4):749–e763713. [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Corvino D, Kumar A, Bald T. Plasticity of NK cells in cancer. Front Immunol. 2022;13:888313. [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Temming AR, de Taeye SW, de Graaf EL, de Neef LA, Dekkers G, Bruggeman CW, Koers J, Ligthart P, Nagelkerke SQ, Zimring JC, et al. Functional attributes of antibodies, effector cells, and target cells affecting NK Cell-Mediated Antibody-Dependent cellular cytotoxicity. J Immunol. 2019;203(12):3126–35. [DOI] [PubMed] [Google Scholar]
41.Brownlie D, von Kries A, Valenzano G, Wild N, Yilmaz E, Safholm J, Al-Ameri M, Alici E, Ljunggren HG, Schliemann I, et al. Accumulation of tissue-resident natural killer cells, innate lymphoid cells, and CD8(+) T cells towards the center of human lung tumors. Oncoimmunology. 2023;12(1):2233402. [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Wong JKM, McCulloch TR, Alim L, Omer N, Mehdi AM, Tuong ZK, Bonfim-Melo A, Chung E, Nicol A, Simpson F, et al. TGF-beta signalling limits effector function capacity of NK cell anti-tumour immunity in human bladder cancer. EBioMedicine. 2024;104:105176. [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Shin SH, Lee YE, Yoon HN, Yuk CM, An JY, Seo M, Yoon S, Oh MS, Shin SC, Kim JH, et al. An innovative strategy Harnessing self-activating CAR-NK cells to mitigate TGF-beta1-driven immune suppression. Biomaterials. 2025;314:122888. [DOI] [PubMed] [Google Scholar]
44.Van den Eynde A, Gehrcken L, Verhezen T, Lau HW, Hermans C, Lambrechts H, Flieswasser T, Quatannens D, Roex G, Zwaenepoel K, et al. IL-15-secreting CAR natural killer cells directed toward the pan-cancer target CD70 eliminate both cancer cells and cancer-associated fibroblasts. J Hematol Oncol. 2024;17(1):8. [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Teng KY, Mansour AG, Zhu Z, Li Z, Tian L, Ma S, Xu B, Lu T, Chen H, Hou D, et al. Off-the-Shelf prostate stem cell antigen-Directed chimeric antigen receptor natural killer cell therapy to treat pancreatic cancer. Gastroenterology. 2022;162(4):1319–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Groh V, Wu J, Yee C, Spies T. Tumour-derived soluble MIC ligands impair expression of NKG2D and T-cell activation. Nature. 2002;419(6908):734–8. [DOI] [PubMed] [Google Scholar]
47.Esposito F, Giuffrida R, Raciti G, Puglisi C, Forte S. Wee1 Kinase: A Potential Target to Overcome Tumor Resistance to Therapy. Int J Mol Sci 2021, 22(19). [DOI] [PMC free article] [PubMed]
48.Wang M, Zhang C, Ying Y, Hua M, Meng F, Wang Z, Liu A, Zeng S, Zhang Z, Xu C. PKMYT1 induced by YAP/TEAD1 gives rise to the progression and worse prognosis of bladder cancer. Mol Carcinog. 2024;63(1):160–72. [DOI] [PubMed] [Google Scholar]
49.Gao W, Guo X, Sun L, Gai J, Cao Y, Zhang S. PKMYT1 knockdown inhibits cholesterol biosynthesis and promotes the drug sensitivity of triple-negative breast cancer cells to Atorvastatin. PeerJ. 2024;12:e17749. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1^{(3.3MB, docx)}

Data Availability Statement

MAGeCK analysis https://figshare.com/account/items/28588673/edit.

[CR1] 1.Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin. 2024;74(1):12–49. [DOI] [PubMed] [Google Scholar]

[CR2] 2.Park W, Chawla A, O’Reilly EM. Pancreatic cancer: A review. JAMA. 2021;326(9):851–62. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Neesse A, Algul H, Tuveson DA, Gress TM. Stromal biology and therapy in pancreatic cancer: a changing paradigm. Gut. 2015;64(9):1476–84. [DOI] [PubMed] [Google Scholar]

[CR4] 4.Cannon A, Thompson C, Hall BR, Jain M, Kumar S, Batra SK. Desmoplasia in pancreatic ductal adenocarcinoma: insight into pathological function and therapeutic potential. Genes Cancer. 2018;9(3–4):78–86. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR5] 5.Mascharak S, Guo JL, Foster DS, Khan A, Davitt MF, Nguyen AT, Burcham AR, Chinta MS, Guardino NJ, Griffin M, et al. Desmoplastic stromal signatures predict patient outcomes in pancreatic ductal adenocarcinoma. Cell Rep Med. 2023;4(11):101248. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Huber M, Brehm CU, Gress TM, Buchholz M, Alashkar Alhamwe B, von Strandmann EP, Slater EP, Bartsch JW, Bauer C, Lauth M. The Immune Microenvironment in Pancreatic Cancer. Int J Mol Sci 2020, 21(19). [DOI] [PMC free article] [PubMed]

[CR7] 7.Carlomagno S, Setti C, Ortolani F, Sivori S. Pancreatic ductal adenocarcinoma microenvironment: soluble factors and cancer associated fibroblasts as modulators of NK cell functions. Immunol Lett. 2024;269:106898. [DOI] [PubMed] [Google Scholar]

[CR8] 8.Pitt JM, Marabelle A, Eggermont A, Soria JC, Kroemer G, Zitvogel L. Targeting the tumor microenvironment: removing obstruction to anticancer immune responses and immunotherapy. Ann Oncol. 2016;27(8):1482–92. [DOI] [PubMed] [Google Scholar]

[CR9] 9.Laskowski TJ, Biederstadt A, Rezvani K. Natural killer cells in antitumour adoptive cell immunotherapy. Nat Rev Cancer. 2022;22(10):557–75. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Wang L, Chen Z, Liu G, Pan Y. Functional crosstalk and regulation of natural killer cells in tumor microenvironment: significance and potential therapeutic strategies. Genes Dis. 2023;10(3):990–1004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR11] 11.Van Audenaerde JRM, De Waele J, Marcq E, Van Loenhout J, Lion E, Van den Bergh JMJ, Jesenofsky R, Masamune A, Roeyen G, Pauwels P, et al. Interleukin-15 stimulates natural killer cell-mediated killing of both human pancreatic cancer and stellate cells. Oncotarget. 2017;8(34):56968–79. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR12] 12.Li W, Wang X, Zhang X, Aziz AUR, Wang D. CAR-NK cell therapy: A transformative approach to overcoming oncological challenges. Biomolecules 2024, 14(8). [DOI] [PMC free article] [PubMed]

[CR13] 13.Kim HS, Kim JY, Seol B, Song CL, Jeong JE, Cho YS. Directly reprogrammed natural killer cells for cancer immunotherapy. Nat Biomed Eng. 2021;5(11):1360–76. [DOI] [PubMed] [Google Scholar]

[CR14] 14.Li P, Burke S, Wang J, Chen X, Ortiz M, Lee SC, Lu D, Campos L, Goulding D, Ng BL, et al. Reprogramming of T cells to natural killer-like cells upon Bcl11b deletion. Science. 2010;329(5987):85–9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] 15.Liu P, Li P, Burke S. Critical roles of Bcl11b in T-cell development and maintenance of T-cell identity. Immunol Rev. 2010;238(1):138–49. [DOI] [PubMed] [Google Scholar]

[CR16] 16.Korchagina AA, Shein SA, Koroleva E, Tumanov AV. Transcriptional control of ILC identity. Front Immunol. 2023;14:1146077. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Avula LR, Rudloff M, El-Behaedi S, Arons D, Albalawy R, Chen X, Zhang X, Alewine C. Mesothelin enhances tumor vascularity in newly forming pancreatic peritoneal metastases. Mol Cancer Res. 2020;18(2):229–39. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Le K, Wang J, Zhang T, Guo Y, Chang H, Wang S, Zhu B. Overexpression of mesothelin in pancreatic ductal adenocarcinoma (PDAC). Int J Med Sci. 2020;17(4):422–7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Tomovic Pavlovic K, Kocic G, Smelcerovic A. Myt1 kinase inhibitors - Insight into structural features, offering potential frameworks. Chem Biol Interact. 2024;391:110901. [DOI] [PubMed] [Google Scholar]

[CR20] 20.Korbecki J, Siminska D, Kojder K, Grochans S, Gutowska I, Chlubek D, Baranowska-Bosiacka I. Fractalkine/CX3CL1 in neoplastic processes. Int J Mol Sci 2020, 21(10). [DOI] [PMC free article] [PubMed]

[CR21] 21.Lu X. Structure and function of ligand CX3CL1 and its receptor CX3CR1 in cancer. Curr Med Chem. 2022;29(41):6228–46. [DOI] [PubMed] [Google Scholar]

[CR22] 22.Kim HS, Lee J, Lee DY, Kim YD, Kim JY, Lim HJ, Lim S, Cho YS. Schwann cell precursors from human pluripotent stem cells as a potential therapeutic target for Myelin repair. Stem Cell Rep. 2017;8(6):1714–26. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR23] 23.Zhu H, Kaufman DS. An improved method to produce Clinical-Scale natural killer cells from human pluripotent stem cells. Methods Mol Biol. 2019;2048:107–19. [DOI] [PubMed] [Google Scholar]

[CR24] 24.Ruscetti M, Leibold J, Bott MJ, Fennell M, Kulick A, Salgado NR, Chen CC, Ho YJ, Sanchez-Rivera FJ, Feucht J, et al. NK cell-mediated cytotoxicity contributes to tumor control by a cytostatic drug combination. Science. 2018;362(6421):1416–22. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR25] 25.Li W, Xu H, Xiao T, Cong L, Love MI, Zhang F, Irizarry RA, Liu JS, Brown M, Liu XS. MAGeCK enables robust identification of essential genes from genome-scale CRISPR/Cas9 knockout screens. Genome Biol. 2014;15(12):554. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR26] 26.Ren Y, Xue M, Hui X, Liu X, Farooq MA, Chen Y, Ji Y, Duan Y, Ajmal I, Yao J, et al. Chimeric cytokine receptor TGF-beta RII/IL-21R improves CAR-NK cell function by reversing the immunosuppressive tumor microenvironment of gastric cancer. Pharmacol Res. 2025;212:107637. [DOI] [PubMed] [Google Scholar]

[CR27] 27.Wu X, Xiao Y, Guo D, Zhang Z, Liu M. Reduced NK cell cytotoxicity by Papillomatosis-Derived TGF-beta contributing to Low-Risk HPV persistence in JORRP patients. Front Immunol. 2022;13:849493. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR28] 28.Liao R, Wu Y, Qin L, Jiang Z, Gou S, Zhou L, Hong Q, Li Y, Shi J, Yao Y, et al. BCL11B and the NuRD complex cooperatively guard T-cell fate and inhibit OPA1-mediated mitochondrial fusion in T cells. EMBO J. 2023;42(21):e113448. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Forkel H, Grabarczyk P, Depke M, Troschke-Meurer S, Simm S, Hammer E, Michalik S, Hentschker C, Corleis B, Loyal L, et al. BCL11B depletion induces the development of highly cytotoxic innate T cells out of IL-15 stimulated peripheral blood alphabeta CD8 + T cells. Oncoimmunology. 2022;11(1):2148850. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Maluski M, Ghosh A, Herbst J, Scholl V, Baumann R, Huehn J, Geffers R, Meyer J, Maul H, Eiz-Vesper B, et al. Chimeric antigen receptor-induced BCL11B suppression propagates NK-like cell development. J Clin Invest. 2019;129(12):5108–22. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR31] 31.Sottile R, Panjwani MK, Lau CM, Daniyan AF, Tanaka K, Barker JN, Brentjens RJ, Sun JC, Le Luduec JB, Hsu KC. Human cytomegalovirus expands a CD8(+) T cell population with loss of BCL11B expression and gain of NK cell identity. Sci Immunol. 2021;6(63):eabe6968. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Jiang Z, Qin L, Tang Y, Liao R, Shi J, He B, Li S, Zheng D, Cui Y, Wu Q, et al. Human induced-T-to-natural killer cells have potent anti-tumour activities. Biomark Res. 2022;10(1):13. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.Silla L. Double-bright (CD56bright/CD16bright) natural killer cell adoptive immunotherapy for SARS-CoV-2. Br J Haematol. 2020;190(6):e322–3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR34] 34.Silla L, Valim V, Pezzi A, da Silva M, Wilke I, Nobrega J, Vargas A, Amorin B, Correa B, Zambonato B, et al. Adoptive immunotherapy with double-bright (CD56(bright) /CD16(bright)) expanded natural killer cells in patients with relapsed or refractory acute myeloid leukaemia: a proof-of-concept study. Br J Haematol. 2021;195(5):710–21. [DOI] [PubMed] [Google Scholar]

[CR35] 35.Amand M, Iserentant G, Poli A, Sleiman M, Fievez V, Sanchez IP, Sauvageot N, Michel T, Aouali N, Janji B, et al. Human CD56(dim)CD16(dim) cells as an individualized natural killer cell subset. Front Immunol. 2017;8:699. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR36] 36.Saito LM, Ortiz RC, Amor NG, Lopes NM, Buzo RF, Garlet GP, Rodini CO. NK cells and the profile of inflammatory cytokines in the peripheral blood of patients with advanced carcinomas. Cytokine. 2024;174:156455. [DOI] [PubMed] [Google Scholar]

[CR37] 37.Poli A, Michel T, Theresine M, Andres E, Hentges F, Zimmer J. CD56bright natural killer (NK) cells: an important NK cell subset. Immunology. 2009;126(4):458–65. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR38] 38.Dogra P, Rancan C, Ma W, Toth M, Senda T, Carpenter DJ, Kubota M, Matsumoto R, Thapa P, Szabo PA, et al. Tissue determinants of human NK cell development, function, and residence. Cell. 2020;180(4):749–e763713. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Corvino D, Kumar A, Bald T. Plasticity of NK cells in cancer. Front Immunol. 2022;13:888313. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR40] 40.Temming AR, de Taeye SW, de Graaf EL, de Neef LA, Dekkers G, Bruggeman CW, Koers J, Ligthart P, Nagelkerke SQ, Zimring JC, et al. Functional attributes of antibodies, effector cells, and target cells affecting NK Cell-Mediated Antibody-Dependent cellular cytotoxicity. J Immunol. 2019;203(12):3126–35. [DOI] [PubMed] [Google Scholar]

[CR41] 41.Brownlie D, von Kries A, Valenzano G, Wild N, Yilmaz E, Safholm J, Al-Ameri M, Alici E, Ljunggren HG, Schliemann I, et al. Accumulation of tissue-resident natural killer cells, innate lymphoid cells, and CD8(+) T cells towards the center of human lung tumors. Oncoimmunology. 2023;12(1):2233402. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR42] 42.Wong JKM, McCulloch TR, Alim L, Omer N, Mehdi AM, Tuong ZK, Bonfim-Melo A, Chung E, Nicol A, Simpson F, et al. TGF-beta signalling limits effector function capacity of NK cell anti-tumour immunity in human bladder cancer. EBioMedicine. 2024;104:105176. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Shin SH, Lee YE, Yoon HN, Yuk CM, An JY, Seo M, Yoon S, Oh MS, Shin SC, Kim JH, et al. An innovative strategy Harnessing self-activating CAR-NK cells to mitigate TGF-beta1-driven immune suppression. Biomaterials. 2025;314:122888. [DOI] [PubMed] [Google Scholar]

[CR44] 44.Van den Eynde A, Gehrcken L, Verhezen T, Lau HW, Hermans C, Lambrechts H, Flieswasser T, Quatannens D, Roex G, Zwaenepoel K, et al. IL-15-secreting CAR natural killer cells directed toward the pan-cancer target CD70 eliminate both cancer cells and cancer-associated fibroblasts. J Hematol Oncol. 2024;17(1):8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR45] 45.Teng KY, Mansour AG, Zhu Z, Li Z, Tian L, Ma S, Xu B, Lu T, Chen H, Hou D, et al. Off-the-Shelf prostate stem cell antigen-Directed chimeric antigen receptor natural killer cell therapy to treat pancreatic cancer. Gastroenterology. 2022;162(4):1319–33. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR46] 46.Groh V, Wu J, Yee C, Spies T. Tumour-derived soluble MIC ligands impair expression of NKG2D and T-cell activation. Nature. 2002;419(6908):734–8. [DOI] [PubMed] [Google Scholar]

[CR47] 47.Esposito F, Giuffrida R, Raciti G, Puglisi C, Forte S. Wee1 Kinase: A Potential Target to Overcome Tumor Resistance to Therapy. Int J Mol Sci 2021, 22(19). [DOI] [PMC free article] [PubMed]

[CR48] 48.Wang M, Zhang C, Ying Y, Hua M, Meng F, Wang Z, Liu A, Zeng S, Zhang Z, Xu C. PKMYT1 induced by YAP/TEAD1 gives rise to the progression and worse prognosis of bladder cancer. Mol Carcinog. 2024;63(1):160–72. [DOI] [PubMed] [Google Scholar]

[CR49] 49.Gao W, Guo X, Sun L, Gai J, Cao Y, Zhang S. PKMYT1 knockdown inhibits cholesterol biosynthesis and promotes the drug sensitivity of triple-negative breast cancer cells to Atorvastatin. PeerJ. 2024;12:e17749. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Directly reprogrammed NK cells driven by BCL11B depletion enhance targeted immunotherapy against pancreatic ductal adenocarcinoma

Han-Seop Kim

Jae Yun Kim

Ji-Young Lee

Binna Seol

Ji Eun Choi

Yee Sook Cho

Abstract

Background

Methods

Results

Conclusions

Supplementary Information

Background

Materials and methods

Primary cell preparation

Human iPSC-NK cell generation

Generation of directly reprogrammed 4 F- and 1 F-NKs

Generation of MSLN-CAR-expressing NKs

Quantitative RT-PCR

RNA-Seq analysis

Flow cytometric analysis

Cancer cells

Calcein-AM- based cytolytic assay

Antibody-dependent cell cytotoxicity (ADCC) assay

Annexin V/PI-based apoptotic cell death assay

PI staining-based cell death assay

Adherent cell assay

Conjugate formation assay

Transwell migration assay

In vitro tumor environment condition assay

Genome-wide CRISPR screening

Fig. 1.

Generation of PKMYT1 knockout cell lines

Western blot analysis

Immunocytochemistry

Enzyme-linked immunosorbent assay (ELISA)

Subcutaneous and orthotopic xenograft models

Blood sample collection

Statistical analysis

Results

Generation of 1 F-NKs and MSLN-1 F-NKs through BCL11B depletion

Characterization of 1 F-NKs and MSLN-1 F-NKs

Fig. 2.

NKG2D-NKG2DL axis as a key driver of 1 F-NK cytotoxicity

Fig. 3.

PKMYT1 inhibition enhances tumor susceptibility and NK cell infiltration

Fig. 4.

In vivo antitumor efficacy of 1 F-NKs, MSLN-1 F-NKs, and PKMYT1i in PDAC subcutaneous models

Fig. 5.

In vivo antitumor efficacy of 1 F-NKs, MSLN-1 F-NKs, and PKMYT1i in PDAC orthotopic models

Fig. 6.

Discussion

Supplementary Information

Acknowledgements

Author contributions

Funding

Data availability

Declarations

Competing interests

Ethics declarations

Footnotes

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases