Ethical statement
Patients were enrolled in the HEMODIAG_2020 cohort (ID-RCB: 2013-A00260-45, NCT02134574, CHU Montpellier) and provided written informed consent. Umbilical cord blood units (UCBs) were sourced from the Biological Resource Center Collection of the University Hospital of Montpellier - (BIOBANQUES Identifier - BB-0033-00031, CHU Montpellier). Samples from Healthy donor (HD) were obtained from the French national blood center “Etablissement Français du Sang”.
Collection of peripheral blood mononuclear cells (PBMCs) and UCBMCs
Cells were collected through differential centrifugation using Ficoll-Paque™ PLUS, (Cytiva Europe GMBH, Freiburg im Breisgau, Germany)40, washed and frozen in a 10/90 DMSO/FBS mix at -80°C employing a gradual decrease rate of 1°C/min and stored in liquid nitrogen. Patient’s PBMCs were cultured for 4-8 hours in complete culture medium: RPMI1640 glutamax™ (Thermo Fisher Scientific France, Illkirch-Graffenstaden, France) with 10% decomplemented FBS (Thermo Fisher Scientific) prior to use.
Cell lines and cell culture
We used the PLH Epstein-Barr Virus (EBV) transformed lymphoblastoid cell line to produce eNK as described40. Raji and Daudi (B cell lines derived from Burkitt lymphomas), K-562 (chronic myelogenous leukemia) and U266-84 (multiple myeloma) were obtained from ATCC. The MOLM-13 (AML-M5a) cell line was obtained from Dr. J.E. Sarry, (UMR1037, INSERM, France). These cell lines were cultured in RPMI supplemented with 10% FBS without antibiotics, and routinely tested for mycoplasma contamination.
Materials
Privigen® (CSL Berhing, Paris, France) a human polyvalent IgG mix, RTX and OBI were obtained from the University Hospital of Montpellier. S239D/H268F/S324T/I332E modified RTX (SDH-RTX) was produced in CHO cells and purified by protein A (Evitria, Zürich, Switzerland). These mutations have been previously described8 and increased the binding to CD16a between 33 (F158) and 57 (V158) times.
CD16a F158V genotyping
After DNA extraction, 15ng of DNA were used to determine CD16 F158V polymorphism as described in Supplemental figure 2 and42.
NK cell expansion and phenotyping
NK cell isolation and expansion were conducted as previously described40,43–46. CD3+ cells were eliminated using EasySep™ Human CD3 Positive Selection Kit II (Stemcell technologies, Saint-Egrève, France), according to provider’s instructions. CD3- cells were counted and the percentage of CD56+ cells determined. Cells were then cultured at 1×106 cells/mL in eNK medium, consisting of RPMI1640 Glutamax™+10% FBS, 100IU/mL rhIL-2 (Thermo Fisher Scientific France) and 5ng/mL hrIL-15 (Miltenyi Biotec, Bergisch Glabach, Germany).
From day 5-7 until day 21-22 the medium was changed every 2-3 days, to reach 0.6×106 cells/mL and 70Gy-irradiated PLH feeder cells were added, as previously described40. PBMCs from patients or eNK between days 14 and 22 of expansion were analyzed using the NK phenotyping staining mix (Supplemental methods). The samples were incubated at 4°C for 20min, washed twice with PBS+2%FBS before analysis.
CD20 and CD16 antigen quantification
We incubated 2×105 cells (eNK, tumor cell lines, PBMCs from HD or patients) with 5µL of anti-CD20-PE (BD, clone B27) plus 1:1000 eFluor™780 (thermos fisher scientific, France) in 95µL of PBS+2% FBS, or 5µL of highly concentrated (50µg/mL) anti-CD16 (BD biosciences France, Le Pont de Claix, France), clone 3G8 known to differentially bind 158F and 158V CD16 alleles at low concentration47, plus 95µL of CD16 quantification staining mix (Supplemental methods). Cells were incubated at 4°C, for 20min, before washing and analysis. We used Quantibrite™ Beads (BD Biosciences France) to quantify CD20 or CD16 expression.
eNK arming
In order to load antibodies on the CD16a, eNK were incubated at 2×106 cells/mL with 10µg/mL of mAbs (WT-RTX, SDH-RTX or OBI) in RPMI1640 Glutamax™ for 1 hour (37°C, 5% CO2), followed by 3 washing steps using complete culture medium (RPMI1640 Glutamax™ +10% FBS). Viability post-arming was quantified as described in supplemental methods.
Armed eNK were incubated (37°C, 5% CO2) for 1 or 8 hours in medium with or without 5mg/mL of Privigen® before washing. Arming efficiency was assessed using the “arming efficiency staining mix” (Supplemental methods), containing mouse anti-human IgG Fab antibody. To measure the impact of Privigen® on arming we used the “hIgG staining mix” (Supplemental methods) containing an anti RTX idiotype recognizing both WT-RTX and SDH-RTX.
Cytotoxicity assay
Cells lines at 1.4×106 cells/mL were stained with CellTrace™ Far Red (CTFR) (Thermo Fisher scientific France), 1:2500° for target cells, and CellTrace™ CFSE (carboxyfluorescein succinimidyl ester, here called “CTG”) (Thermo Fisher scientific France), 1:5000° for control MOLM-13 cells following provider’s instructions. We placed 2×105 of each target and control cells and eNK in a 96 wells U bottom plate and incubated them for 8 hours (37°C, 5% CO2). After co-culture, cells were stained using 20 µL freshly prepared Viobility™ 405/452 and eNK were stained using 20µL of 1:200 in PBS+2%FBS anti-CD56 (Miltenyi Biotec, clone: REA 196). Normalized survival ratio was calculated by comparing numbers of surviving MOLM-13 and target cells, as described previously48.
Concerning patient’s PBMCs, 2×105 cells were seeded in 96U well plates and CTFR (1/2500°) stained eNK were added at a ratio effector/PBMC of 3:1 and incubated in 100µL of complete culture medium with or without 10µL/mL mAbs. After 8 hours, cells were washed and stained for cell death (Viobility™ 405/520) and “tumor B-cells assessment staining mix” (Supplemental methods). Cells were washed and resuspended in PBS+2%FBS (180µL) and precision counting beads (20µL).
Cytometry acquisition
Cytometry analyses were conducted using a Gallios 3 Lasers (Beckman Coulter, Pasadena, CA, USA) or a Symphony A3 5 lasers flow cytometer (BD biosciences France). Acquisition was carried out using Kaluza acquisition software V1.3 (Beckman Coulter) or Facs Diva (BD biosciences France).
Statistical analysis
Statistics were performed using Prism V7.04 software. Each sample value represents the average of, at least, a technical duplicate. Statistical tests are detailed in each figure legend.