Raji, Raji-2R, and Raji-4RH cells were generously provided by Matthew Barth, from Roswell Park Cancer Institute, Buffalo, New York, USA.11 K526-mbIL21-41BBL cells were generously provided by Dean A. Lee, from Nationwide Children’s Hospital, Columbus, Ohio, USA.23 Ramos, Daudi, and RS4;11 cells were purchased from ATCC. DOHH-2 cells were ordered from DSMZ. Raji, Raji-2R, Raji-4RH, and K526-mbIL21-41BBL cells were cultured in RPMI 1640 medium with 10% heat inactivated fetal bovine serum (FBS) at 37°C with 5% CO2. Obinutuzumab was generously provided by Christian Klein, from Roche, Zurich, Switzerland. N-820 and N-803 were generously provided by Hing Wong, Peter R. Rhode, John H. Lee, and Jeffrey T. Safrit, from ImmunityBio/Altor Bioscience, Culver City, California, USA. All fetal bovine serum used for cell culture and assays was heat-inactivated (heating to 56°C for 30 min) before use in order to inactivate complement. Leucocytes were obtained after informed consent from healthy donors at the New York Blood Center, New York, New York, USA. PBMCs were obtained by Ficoll gradient (Amersham Biosciences, Piscataway, New Jersey, USA) separation as we previously described.25

PBMCs were stimulated with irradiated genetically modified K562-mbIL21-41BBL cells as previously described.23 Expanded PBNK cells were isolated by negative selection using Miltenyi NK cell isolation kit (Miltenyi Biotec, Cambridge, Massachusetts, USA).25

Tumor targets were labeled with carboxyfluorescein succinimidyl ester (CFSE) (Thermofisher Scientific, USA, cat #C34570) or engineered to express luciferase. Briefly, exPBNK cells were incubated with target cells at different effector-to-target (E:T) ratios with IgG, N-803, N-820, rituximab, or rituximab+N-803 at 37°C for 2–3 days. For CFSE-labeled target cells, dead cells were stained with propidium iodide (Thermofisher Scientific, USA, cat #3566). Cytolytic activity was evaluated by flow cytometry analysis. Target cells were gated on CFSE⁺ and the live/dead cell percentage was visualized by propidium iodide exclusion. For target cells that express luciferase, living cells were monitored by measuring luciferase activity with a luminometer (Molecular Devices Multifilter F5 plate reader) as relative light units (RLU). Cells were treated with 1% Triton X-100 as a measure of maximal killing. Target cells incubated without effector cells were used to measure spontaneous death RLU. Percent lysis was calculated from the data with the following equation: % specific lysis = 100×(spontaneous death RLU–test RLU)/(spontaneous death RLU–maximal killing RLU). All tests were run in quadruplicate.

Cell culture supernatants were collected after 3 days culture and were stored at −80 °C. The concentrations of cytokines/chemokines/growth factors were measured by the Bio-Plex Pro human cytokines screening panel 48 cytokines Assay (Bio-Rad Laboratories, Hercules, California, USA) according to the manufacturer’s instructions. The beads were read on a Luminex System (Bioplex 200, Bio-Rad) and the data were analyzed using Bioplex Manager Software.

Granzyme B (eBioscience, San Diego, California, USA), IFN-γ (eBioscience), GM-CSF (Raybiotech, Peachtree Corners, Georgia, USA), and chemokine ligand 22 (CCL22) (Raybiotech) concentrations were analyzed by ELISA according to the manufacturers’ instructions. The absorbance at 450 nm was measured on a Molecular Devices Multifilter F5 plate reader.

The exPBNK cells under different conditions were analyzed for phenotypic expression of activating and inhibitory NK receptors by flow cytometry as we previous described.25

Fixed and permeabilized cells were stained with mouse anti-human primary antibodies Ki67-PE (Thermofisher, cat #12-5698-80), phospho-STAT5-PE (Tyr694) (Thermofisher, cat #12-9010-41), phospho-Akt-APC (Ser473) (Thermofisher, cat #17-9715-42). Cells were analyzed using MACSQuant Analyzer (Miltenyi Biotec). No stain, or isotype controls were used for gating as we previously described.12

Raji, Raji-2R, or Raji-4RH were incubated with 10 nM IgG, 10 nM N-803, 10 nM N-820, 10 nM rituximab, or 10 nM N-803+10 nM rituximab for 30 min. The samples were washed three times in phosphate buffered saline (PBS). The binding of N-820 or rituximab to CD20 on the surface of BL cells was measured by staining the cells with phycoerythrin (PE) conjugated anti-CD20 antibody (Miltenyi Biotec, cat #130-113-374).

ExPBNK cells were incubated with 10 nM IgG, 10 nM N-803, 10 nM N-820, 10 nM rituximab, or 10 nM N-803+10 nM rituximabfor 3 days. The binding of N-820 or rituximab on the surface of exPBNK cells was measured by fluorescein isothiocyanate (FITC) conjugated goat anti-mouse IgG F(ab’)2 specific antibody (Jackson ImmunoResearch, cat #115-096-003).

ExPBNK cells were co-cultured with Raji-2R at E:T=1:1 plus 10 nM N-820 or Raji-2R at E:T=1:1 plus 10 nM rituximab+10 nM N-803 for 3 days. Harvested mixed cells were stained for CD56-PE-Vio770 (Miltenyi, cat# 130-113-313). CD56⁺ cells were gated and isolated using a MoFlo XDP high-speed cell sorter (Beckman Coulter Inc., Miami, Florida, USA). A minimum of 1–10×10⁵ sorted cells with a purity >90% were collected into cold tissue culture media.

The Human Cancer Inflammation & Immunity Crosstalk RT2 Profiler PCR Arrays (PAHS-181ZA, Qiagen, Frederick, Maryland, USA) were used to profile the expression of 84 genes involved in mediating communication between tumor cells and exPBNK cells mediated by N-820 compared with rituximab+N-803. Total RNA was isolated from exPBNK cells using Qiagen RNeasy Mini Kit by following manufacturer’s protocol. The first-strand cDNA was mixed with 2×RT2 SYBR Green qPCR Master Mix and ddH2O. The qPCR was performed on an Applied Biosystems (ABI) QuantStudio 5 PCR system according to the RT2 Profiler PCR Array instructions under the following conditions: 95°C for 10 min, then 40 cycles at 95°C for 15 s and 60°C for 1 min. Data were normalized against the house keeping genes by calculating the ΔCt for each gene of interest in the plate. Fold changes of gene expression, scatterplot and heatmap were analyzed and generated using Qiagen Gene Globe data analysis web port.

STRING database in Cytoscape software V.3.7.2 was used to assess PPI information and converted the results visually by using Cytoscape.29 A confidence score >0.7 was set as significant. CytoHubba, a plug-in Cytoscape software, was used to identify the top 10 hub genes in PPI networks through the degree method.30

Six to eight week old NOD/SCID/γ-chain-/- (NSG) mice were purchased from the Jackson Laboratory (Bar Harbor, Maine, USA) and were bred, treated, and maintained in the animal facility of New York Medical College (NYMC) with NYMC Institutional Animal Care and USE Committee (IACUC)-approved protocols. The animal experiments were conducted in accordance with the recommendations of the Guide for Care and Use of Laboratory Animals.

Luciferase-expressing tumor cells (Raji-2R-Luc and Raji-4RH-Luc) were generated as we previously described.25 Tumor cells were retro-orbitally or intraperitoneally administered into the NSG mice. Tumor engraftment and progression were evaluated using the Xenogen IVIS-200 system (Caliper Life Sciences, Hopkinton, Massachusetts, USA) as we have previously described.25 Mice were followed until death.

Statistical analyses were performed using the INSTAT statistical program (GraphPad, San Diego, California, USA). Average values are reported as the mean±SEM. Results were compared using the one-tailed unpaired Student’s t-test with p<0.05 considered as significant. Probability of survival in animal studies was determined by the Kaplan-Meier method using the Prism program 8.0 (GraphPad Software).

To investigate if N-820 affects the expression of receptors on exPBNK cells and stimulates the proliferation of exPBNK cells, we cultured purified exPBNK cells in medium with 10 nM N-820 for 3 days. 10 nM IgG and 10 nM N-803 were used as controls. The activating and inhibitory receptors on exPBNK cells were examined by flow cytometry. N-820 significantly enhanced the expression of NK activating receptors such as NKp44, CD16, NKp30, and NKG2D (figure 1A) as compared with IgG. N-820 showed similar activity to N-803 in stimulating NK receptor expression (figure 1A) except for NKp46 and KIR2D2/3. The proliferation of exPBNK cells was examined using CellTiter 96 AQueous One solution cell proliferation assay after 3 days culture with PBS, 10 nM IgG, 10 nM N-820, or 10 nM rituximab+10 nM N-803. N-820 significantly stimulated exPBNK proliferation as compared with IgG (p<0.001), N-803 (p<0.001), and rituximab+N-803 (p<0.002) (figure 1B). Consistent with the enhanced exPBNK proliferation, N-820 significantly stimulated Ki67 expression in exPBNK cells as compared with IgG (p<0.001), N-803 (p<0.001), and rituximab+N-803 (p<0.01) (figure 1C). IL-15 is known to rapidly induce STAT5 phosphorylation, which plays important roles in regulating NK-cell survival, proliferation, cytotoxicity, and maturation.31 As expected, N-820 significantly enhanced STAT5 phosphorylation in exPBNK cells as compared with IgG (p<0.001), N-803 (p<0.02), and rituximab+N-803 (p<0.02) (figure 1D). In addition to the IL-15–STAT5 pathway, accumulating data demonstrate that the PI3K–AKT–mTOR pathway is essential for the development, differentiation, and activation of NK cells.32 We examined AKT phosphorylation in exPBNK cells stimulated by IgG, N-820, N-803, or rituximab+N-803. There was no significantly difference of AKT phosphorylation in exPBNK cells among N-820, N-803, or rituximab+N-803 but N-820 significantly stimulated AKT phosphorylation in exPBNK cells as compared with IgG (p<0.001) (figure 1E,F).

Previous studies have shown that IL-15 stimulation results in transient (hours to days) CD25 expression, a key component required to form the high-affinity heterotrimeric IL-2Rαβγ on the majority of NK cells correlated with STAT5 phosphorylation.33 Therefore, we examined CD25 expression stimulated by N-820. We found that N-820 with or without BL tumor cells significantly stimulated CD25 expression on exPBNK cells (p<0.001) as compared with IgG controls. N-820 significantly stimulated CD25 expression on exPBNK cells as compared with rituximab+N-803 without co-culturing with BL tumor cells (p<0.001) or with Raji-2R (p<0.0012) or Raji-4RH (p<0.0016) cells (figure 1G).

To investigate the cytokines and growth factors that are significantly secreted by exPBNK cells stimulated by N-820 and compare with the cells stimulated by rituximab+N-803, exPBNK cells were cultured with 10 nM IgG, 10 nM N-820, or 10 nM rituximab+10 nM N-803 for 3 days. The concentrations of cytokines/chemokines/growth factors in the supernatants were measured by the Bio-Plex Pro human cytokines screening panel 48 cytokines Assay. We found that the levels of eotaxin, FGF, GM-CSF, HGF, INF-A2, INF-G, IL-1RA, IL-2RA, IL-8, IL-12P40, IL-13, IL-17, CXCL9, SCF, TNF-A, TNF-B, LIF, PDGF-BB, CXCL1, RANTES, MIP-1B were significantly increased from exPBNK cells by N-820 as compared with rituximab+N-803 (figure 2A,B), while CXCL12 and SCGF-B were significantly reduced by N-820 as compared with rituximab+N-803 (figure 2A).

Since exPBNK cells express high level of CD16 (figure 1A) and N-820 stimulated the ADCC of primary NK cells,27 we investigated whether N-820 significantly stimulates the ADCC activity of exPBNK cells against rituximab-sensitive Raji and rituximab-resistant Raji-2R and Raji-4RH cell lines. Raji, Raji-2R, and Raji-4RH cells were treated with 10 nM IgG, 10 nM N-820, 10 nM N-820+exPBNK cells, or 10 nM IgG+exPBNK cells at E:T ratio=0.1:1, 1:1, and 3:1. We found that the combination of N-820 and exPBNK significantly killed Raji, Raji-2R, and Raji-4RH cells (p<0.001) as compared with N-820 alone or IgG alone (figure 3A) in an E:T ratio dependent manner. And N-820 significantly enhanced the ADCC activity of exPBNK cells (p<0.001) as compared with exPBNK+IgG against Raji, Raji-2R, and Raji-4RH (figure 3B). We further compared the activity of N-820 with the same doses of N-803, rituximab, or rituximab+N-803 on the cytotoxicity of exPBNK cells. We found that N-820 significantly enhanced the in vitro cytotoxicity of exPBNK as compared with N-803 (p<0.001), rituximab (p<0.001), and rituximab+N-803 (p<0.05 at low E:T ratios and p<0.001 at higher E:T ratios) in an E:T ratio dependent manner (figure 3B). Obinutuzumab is a humanized, type II anti-CD20 monoclonal antibody glycoengineered to enhance Fc receptor affinity.34 Considering the enhanced ADCC of obinutuzumab as compared with rituximab in the setting of rituximab-sensitive and rituximab-resistant BL,34–36 we compared the effect of N-820 on the ADCC activity of exPBNK with the same dose of obinutuzumab+N-803 against Raji, Raji-2R, and Raji-4RH cells. ExPBNK cells+10 nM N-820 showed similar killing activity as exPBNK cells+10 nM obinutuzumab+10 nM N-803 against Raji cells (p=0.26) (figure 3C). These studies demonstrated N-820 had enhanced cytolytic activity of exPBNK cells against Raji-2R cells (p=0.0611) and significantly enhanced cytolytic activity against Raji-4RH cells as compared with obinutuzumab+N-803 (p<0.01) (figure 3C). We also examined if N-820 enhances the anti-tumor effects of exPBNK against other CD20+B-NHL cells. We found that the combination of exPBNK cells with N-820 significantly killed CD20⁺ DOHH-2 (follicular lymphoma), Ramos (BL), or Daudi (BL) and significantly enhanced granzyme B release as compared with other control groups (online supplemental figure 2A). A CD20− acute leukemia cell line RS4;11 cells were used as a negative control. The combination of exPBNK cells with N-820 did not significantly killed RS4;11 and did not significantly enhanced granzyme B release as compared with the control groups (online supplemental figure 2B).

Previously, we reported the expression of CD20 was significantly reduced in Raji-2R and Raji-4RH as compared with Raji.25 The enhanced cytolytic activity of exPBNK cells+N-820 as compared with exPBNK cells+obinutuzumab+ N-803 against rituximab-resistant BL cells may be due to the higher affinity of N-820 to the CD20 target on the rituximab-resistant BL surface. We further confirmed that N-820 significantly enhanced exPBNK in vitro cytotoxicity as compared with controls in part by significantly enhancing the secretion of both granzyme B and IFN-γ. We found that 10 nM N-820 significantly enhanced granzyme B (figure 4A) and IFN-γ (figure 4B) release from exPBNK cells as compared with 10 nM IgG, 10 nM N-803, 10 nM rituximab, 10 nM rituximab+10 nM N-803, and 10 nM obinutuzumab against Raji, Raji-2R, and Raji-4RH (p<0.001) at E:T=1:1.

We investigated the binding ability of N-820 to CD20 on Raji, Raji-2R, or Raji-4RH cells. Raji, Raji-2R, or Raji-4RH cells were incubated with 10 nM IgG, 10 nM N-803, 10 nM N-820, 10 nM rituximab, or 10 nM N-803+10 nM rituximab for 30 min. After PBS wash, the cells were stained with PE conjugated anti-CD20 antibody (clone LT20, which recognizes the similar epitope of CD20 as rituximab).37 N-820 significantly blocked the accessibility of CD20 to another anti-CD20 antibody LT20 on the surface of Raji as compared with the IgG (p<0.001) and rituximab (p<0.01) (figure 5A, left panels). N-820 significantly blocked the accessibility of CD20 to LT20 on the surface of Raji-2R (figure 5A, middle panels) and Raji-4RH (figure 5A, right panels) as compared with the IgG (p<0.001) and rituximab (p<0.001), indicating N-820 may have a significant higher binding ability to CD20 in rituximab-sensitive and rituximab-resistant BL cells.

N-820, rituximab, and N-803 all contain IgG Fc region (online supplemental figure 1), which can bind to CD16. We further examined the binding ability of N-820 to exPBNK cells. ExPBNK cells were incubated with 10 nM IgG, 10 nM N-803, 10 nM N-820, 10 nM rituximab, 10 nM N-803+10 nM rituximab, or 10 nM obinutuzumab. After incubation for 3 days, N-820 can still be detected and significantly binds to exPBNK cells compared with controls (p<0.001) (figure 5B).

Expanded NK cells from three different donors were co-cultured with Raji-2R cells at E:T ratio=1:1 with IgG, N-820, or rituximab+N-803. After 3 days, NK cells were sorted with 93%–95% purity (figure 6A). The important genes involved in the effect of N-820+BL on exPBNK cells were identified using the Human Cancer Inflammation & Immunity Crosstalk RT2 Profiler PCR Array. The present study found that 60 genes had at least a twofold increase in exPBNK cells stimulated by N-820+Raji-2R as compared with the gene expression levels in exPBNK cells stimulated by rituximab+N-803+Raji-2R and three genes (CCL22, CXCR4, and CXCR5) were downregulated (figure 6B). Consistent with the results of cytokines and growth factors screened in figures 2 and 4, the transcription levels of IL-17, CXCL9, CXCL1, CSF2, CSF3, GZMB, and IFNG were enhanced in exPBNK cells stimulated by N-820+Raji-2R as compared with those stimulated by rituximab+N-803+Raji-2R. A heat map provides a graphical representation of fold expression in exPBNK cells stimulated by N-820+Raji-2R against exPBNK cells stimulated by rituximab+N-803+Raji-2R, red and green color represent increasing or decreasing genes, respectively (figure 6C). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis shows that the differentially expressed genes (DEGs) are related to apoptosis, NK-mediated cytoxicity, HIF-1, NK-kappa B, Jak-STAT, TNF, PI3K-Akt, Toll-like receptor, chemokine signal pathways, and cytokine–cytokine receptor interaction (figure 6D). PPI network was constructed with fourfolds upregulated DEGs using the STRING database imported in Cytoscape software (V.3.7.2) and revealed 25 nodes (figure 6E). Using the plug-in CytoHubba app in Cytoscape software, we evaluated the degree and betweenness in the PPI network and screened the hub genes. The top 10 hub genes showing significant interaction were cytokine and chemokine genes: IL-6, IL-10, IL-4, IL-17a, CXCL10, CXCL11, CXCL9, CXCL1, CXCL2, and CXCL5 (figure 6E), which are consistent with the KEGG pathway analysis results (figure 6D) that chemokine signal pathways and cytokine–cytokine receptor interaction are the two top enhanced events.

CSF2 encodes an immune regulator GM-CSF.38 Using ELISA, we further confirmed that GM-CSF was significantly released from exPBNK cells stimulated by N-820 as compared with rituximab+ALT803 (p<0.001) (figure 6F), which is consistent with the results from the cytokine screening (figure 2). GM-CSF was also significantly released from exPBNK cells stimulated by Raji-2R cells as compared with exPBNK cells alone. More importantly, GM-CSF was significantly released from exPBNK cells stimulated by N-820+Raji-2R as compared with rituximab+N-803+Raji-2R (p<0.001) (figure 6F). We further confirmed these findings using Raji-4RH cells as tumor targets (data not shown).

CCL22 induces regulatory T cell migration to malignant cells through the CCR4 receptor, leading to inhibition of antitumor immunity and tumor progression.39 We found that the CCL22 secretion level was significantly enhanced when co-culturing exPBNK cells with Raji-2R cells as compared with exPBNK or Raji-2R alone (figure 6G) in part suggesting that the crosstalk between exPBNK cells and Raji-2R cells potentially inhibits antitumor immunity in the tumor microenvironment. The addition of N-820 (p=0.0018) or rituximab+N-803 (p=0.011) to the exPBNK+Raji-2R co-culture significantly reduced the secreted CCL22 levels (figure 6G). Consistent with reduced CCL2 transcription, N-820 significantly reduced CCL22 protein expression compared with rituximab+N-803 when added to the exPBNK+Raji-2R co-culture (p=0.0128) (figure 6G).

We further confirmed the anti-tumor effects of N-820 combined with exPBNK cells in human rituximab-resistant BL xenografted NSG mice. 1×10⁶of Raji-2R-Luc cells were injected in NSG mice on day 0. After confirming tumor engraftment at day 7, 1×10⁷ exPBNK cells+10 mg/kg IgG (n=12), 1×10⁷ exPBNK cells+5 mg/kg N-820 (n=8), 1×10⁷ exPBNK cells+10 mg/kg N-820 (n=12), 10 mg/kg N-820 (n=8), or PBS (n=12) was intraperitoneally injected to each mouse. NK cells were administered once a week for 3 weeks. IgG and N-820 were injected two times a week for 3 weeks. We found that the combination of exPBNK+10 mg/kg N-820 significantly reduced the tumor burden measured by bioluminescence signal intensity as compared with the groups treated with PBS (p=0.0141), exPBNK (p=0.0005), 10 mg/kg N-820 (p=0.0212) in Raji-2R xenografted mice (figure 7A–C). More importantly, the combination of exPBNK+10 mg/kg N-820 significantly extended the survival of mice as compared with the group treated with PBS (p=0.0029), exPBNK (p=0.0406), exPBNK +5 mg/kg N-820 (p=0.0421), 10 mg/kg N-820 (p=0.0088) (figure 7D,E). We further used NSG mice xenografted with another rituximab-resistant cell line, Raji-4RH, to confirm these results. We found that 1×10⁷ exPBNK cells+10 mg/kg N-820 (n=8) decreased tumor growth as compared with the tumor burden in the control groups (figure 7F) and significantly enhanced the survival of Raji-4RH xenografted NSG mice as compared with the group treated with exPBNK cells+10 mg/kg rituximab+0.2 mg/kg N-803 (n=7, p=0.0159), the group treated with exPBNK cells+10 mg/kg IgG (n=5, p<0.001), and the group treated with PBS (n=5, p<0.001) (figure 7G).