Hanne Verswyvel

Chapter 4 │ Page 127 e ects at 24 hours. In short, cells were washed with phosphate bu er saline (PBS), detached with accutase (Sigma-Aldrich), spun down and resuspended in FACS bu er (sheath bu er (BD Biosciences), with 0.1% BSA and 0.05% NaN3). Cell density of untreated controls and NTP-treated samples was set at 5x105 cells/ml. Samples were incubated with 1 µL of LIVE/DEAD™ Fixable Near-IR (L10119, ThermoFisher Scientific), and a single stain for a target ligand: anti-HLA-C (5 µL, Cat. No. 566372, BD), anti-HLA-E (5 µL, 12-9953-42, Invitrogen), anti-CD155 (5 µL, 566718, BD), anti-CD112 (5 µL, 337410, BioLegend), anti-CD73 (5 µL, 12-0739-42, Invitrogen), or anti-MICA/B (20 µL, 558352, BD). Fluorescence Minus One (FMO) gating controls were included for all cell lines and treatment conditions. After 30 minutes of incubation in the dark at 4°C, cells were washed twice with FACS bu er, and resuspended in 100 µL FACS bu er for read-outs. Sample acquisition was performed on the NovoCyte Quanteon (Agilent Technologies). Experimental data was gated and analyzed using the FlowJo software version 10.8.1 (FlowJo LLC, Ashland, OR, USA) (Supplementary Figure 1). 2.7. Statistical Analysis Prior to statistical calculations, the Grubbs’ Test was performed to detect significant outliers in the experimental data. Afterwards, the linear mixed model in JMP Pro17 (SAS Software, Tervuren, Belgium) was used to analyze statistical di erences between treatment conditions. NTP exposure was designated as the fixed e ect, while the di erent experimental repeats and the interaction between experiments and the date they were performed were considered as random e ects. The random slope model was only retained when the treatment-date interaction was significant (p ≤ 0.05). Statistical di erence by the fixed e ect was determined, and the post-hoc Dunnett’s test was used to calculate adjusted p values of the treatments compared to untreated controls. P values equal or less than 0.05 were

RkJQdWJsaXNoZXIy MTk4NDMw