Isolation of NK cells

NK cells were isolated from leukapheresis, acquired from healthy volunteer donors. Subsequently, automated NK cell separation was performed by immunomagnetic selection via CD3 depletion using CliniMACS Prodigy® LP-3-56 followed by an CD56 enrichment using CliniMACS Prodigy® PD-56 Engineering. NK cell purity was determined as described earlier [18]. This project was approved by the Ethics Committee of the Goethe University Frankfurt, Germany (approval no. 274/18 and 329/10). Peripheral blood apheresis products from healthy donors were obtained from the Institute for Transfusion Medicine of Hannover Medical School (MHH) after donors’ written informed consent. The protocol was approved by the ethics committee of Hannover Medical School.

CAR construction and lentiviral vector production

A second-generation CD33-targeting CAR which incorporates the My96 scFv sequence, which has been used in the immunoconjugate AVE9633, was constructed as described earlier [26]. In short, the My96 scFv was combined in frame with CD8 hinge and transmembrane domain, 4-1BB/CD137 co-stimulatory domain, and CD3ζ activation domain. A leader peptide derived from granulocyte-macrophage colony-stimulating factor receptor unit α (GM-CSFRα) was included to facilitate CAR cell surface expression. Third-generation self-inactivating baboon envelope-pseudotyped lentiviral vectors (BaEV-LVs) were produced by transient transfection into HEK293T cells using MACSfectin™ (Miltenyi Biotec, Bergisch Gladbach, Germany) [18, 27].

CAR-NK cell manufacturing

Two days after isolation, NK cells were transduced with lentiviral particles either as described earlier for the small-scale production [18] or using the CliniMACS Prodigy® device. Briefly, purified NK cells were cultivated with NK MACS GMP Medium supplemented with 5% human AB serum, MACS GMP recombinant human IL-2, MACS GMP recombinant human IL-15 and additionally IL-1β for activation. On day 2 of culture, transduction was performed by addition of pre-mixed lentiviral vector particles with 1.25 µg/mL Vectofusin-1 to NK cells followed by spinoculation for 2 h at 32 °C and 400 g. NK cells were washed at day 3 to remove excess lentiviral vector. Subsequently, NK cells were cultured until day 14 using NK MACS GMP Medium supplemented with 5% human AB serum, as well as MACS GMP recombinant human IL-2 and IL-15. The NK cells in small scale were cultured according to the same scheme. After the first culture week, small-scale cells were split in a 1:2 ratio on day 9, 10 and 13.

Flow cytometry analysis of transduced NK cells

During culture, the NK cell population was monitored by flow cytometry using a standardized in process control (IPC) and quality control (QC) strategy. The NK cell purity was verified using the NK cell composition panel including the fluorochrome-conjugated antibodies anti-CD45-VioBlue (clone REA747, Miltenyi Biotec), anti-CD14-VioGreen (clone REA599, Miltenyi Biotec), anti-TCRαβ-FITC (clone REA652, Miltenyi Biotec), anti CD3-PE (clone SK7, BioLegend, San Diego, CA, USA), anti-CD56-APC (clone REA196, Miltenyi Biotec) and anti-CD19-APC-Vio770 (clone REA675, Miltenyi Biotec).

CAR expression was analyzed by using a CD33-CAR Detection Reagent which contained a recombinantly expressed fusion protein that consisted of the extracellular domains of human CD33 and a specifically mutated human IgG1-Fc region connected to biotin, and secondary a biotin-targeting antibody, anti-biotin PE (clone REA746) (both Miltenyi Biotec).

NK cell transduction efficacy was determined through the NK cell transduction panel containing anti-CD45-VioBlue (clone REA747, Miltenyi Biotec), anti-CD14-VioGreen (clone REA599, Miltenyi Biotec), anti-CD33-CAR Detection Reagent-Biotin (Miltenyi Biotec), anti-Biotin-FITC (clone Bio3-18E7, Miltenyi Biotec), anti CD3-PE (clone SK7, BioLegend) and anti-CD56-APC (clone REA196, Miltenyi Biotec).

For the phenotypical analysis of untransduced (UTD)- and CAR-NK cells, fluorochrome-conjugated antibodies anti-NKG2A-PE (clone Z199, Beckman Coulter, Brea, CA, USA), anti-NKG2C-AF488 (clone 134591, R&D Systems, Minneapolis, MN, USA), anti-CD3-PerCP (clone UCHT-1, BioLegend), anti-CD57-BV421 (clone NK-1), anti-CD69-BV605 (clone FN50) and anti-CD16-PE-CF594 (clone 3G8) (all BD Biosciences, Franklin Lakes, NJ, USA) were used.

All stainings were performed using FcR Blocking Reagent (Miltenyi Biotec) or unspecific hIgG (Kiovig or Intratect®, University Hospital Frankfurt, Frankfurt am Main, Germany) to increase the specificity of immunofluorescent staining and 7-amino-actinomycin D (7AAD) (Miltenyi Biotec) for the exclusion of dead and apoptotic cells from flow cytometric analysis.

In vitro cytotoxicity assay

SS- and PROD-produced UTD-NK cells as well as SS-/PROD-produced CAR-NK cells were co-incubated for four hours with OCI-AML2 cells at an effector-to-target (E:T) ratio of 1:5. OCI-AML2 cells were previously labeled using Cell Trace CFSE proliferation kit (Thermo Fisher Scientific, Waltham, MA, USA). Subsequently, cells were stained with 4′,6-diamidino-2-phenylindole (DAPI) (AppliChem, Darmstadt, Germany) and the viability of target cells was analyzed using a BD FACSCelestaTM device (BD Biosciences).

In vivo analysis of PROD-CAR-NK cells in an OCI-AML2-xenograft NSG-SGM3 mouse model

In vivo functional comparison of SS- and PROD-produced CAR-NK cells was performed in our established OCI-AML2-xenograft NSG-SGM3 mouse model as described earlier [18]. OD.Cg-Prkdcscid Il2rgtm1Wjl Tg(CMV-IL3,CSF2,KITLG)1Eav/MloySzJ (NSG-SGM3) mice were obtained from The Jackson Laboratory, Bar Harbor, ME, USA (JAX stock No.: #013062 (NSG-SGM3)) [28]. Mice were held under standardized pathogen free conditions with adequate access to food and water. Experiments were approved by the Regierungspräsidium Darmstadt, Germany.

Bone marrow sectioning and analysis

Mouse femurs and tibiae were isolated and incubated in a series of 4% PFA, EDTA- and sucrose-solutions before they were embedded in gelatin. 7 µm-sections of bones were generated on the CryoStar™ NX70 cryostat (Thermo Fisher Scientific). OCI-AML2 cells were detected by their GFP-expression. For nuclear counterstaining DAPI (Sigma-Aldrich, St. Louis, MO, USA) was used.

In vivo cytokine release

Cytokine levels in sera of mice were determined three days pre OCI-AML2 cell injection and one day post therapy initiation (day 4 post AML cell injection) using a BD™ Cytometric Bead Array (CBA; BD Biosciences) according to manufacturer’s instructions. BD™ CBA Flex Sets were used to measure the cytokine concentrations. Data were obtained using a BD FACSVerse™ Bioanalyzer. For data analysis the FCAP Array software (v3.0.1; BD Biosciences) was used.

Statistical analysis

For statistical analysis, a normal distribution for NK cell functionality was assumed due to the deployment of healthy donors as NK cell source. For the animal experiments, also a normal distribution could be assumed. Accordingly, data were analyzed by two-tailed, unpaired Student’s t test or Mann–Whitney-test and defined as significant when p < 0.05. Statistical analysis was performed using GraphPad PRISM version 9 (GraphPad Software, Inc., San Diego, CA, USA).