Assessment of phenotype and function of mbIL21-expanded NK-cell infusion product (green) or PB NK cells patients. Mononuclear cells were isolated from PBMCs of patients on protocol 2009-0266 (without NK cells [red]) or 2012-0708 (with NK cells [green]) ∼28 days after stem cell transplant. Samples from patients on protocol 2012-0708 were obtained prior to receiving the third dose of NK cells (∼3 weeks after receiving the second dose). (A) Cytotoxicity against 721.221 targets. Cells were applied to the cytotoxicity assay at a 10:1 NK-to-target ratio (according to NK-cell content determined by immune phenotyping). (B) NK-cell responses to stimulation with 721.221 targets at a 2:1 NK-to-target ratio for 3 hours. Degranulation (CD107a), cytokine production, and CD62L cleavage were determined by mass cytometry. (C) Cytotoxicity and IFNγ production from panels A and B stratified according to cell dose received (low, ≤10⁶/kg per dose; high, ≥10⁷/kg per dose). P values shown for unpaired Student t test. (D) Differences in phenotype of NK cells at day +28 posttransplant as determined by mass cytometry. Shown are surface markers on NK cells from supplemental Figure 1 that were significantly different between 2009-0266 (no NK-cell infusions) and 2012-0708 (with NK-cell infusions) using multiple unpaired Student t-test comparisons followed by the 2-stage false discovery approach. Corrected P values are indicated as *P < .01, **P < .001, ***P < .0001, or ****P < .00001. (E) Heatmap with unsupervised clustering analysis of the NK-cell phenotypes from supplemental Figure 1 according to relative expression of each receptor in each sample. Sample origin is indicated along the top row: day +28 PB-NK samples from protocol 2009-0266 (red) or 2012-0708 (blue), or NK-cell product (green). (F) SPADE trees of NK-cell subsets present in the NK-cell product or peripheral blood at day +28 posttransplant. Given the small number of samples in each group, KIR expression was excluded from clustering to avoid bias from individual KIR and HLA genotypes. (G) Heatmap with unsupervised clustering analysis of the NK-cell subsets from each sample according to those identified in panel F (using only nodes constituting at least 1% of any 1 sample). Sample origin color coding is as in panel E. (H) Principal components analysis based on the percentage of NK cells in each node of each sample as in panel G. X-axis and Y-axis show principal component 1 and principal component 2 that explain 31.9% and 24.1% of the total variance, respectively. Prediction ellipses indicate 95% probability that a new observation from the same group will fall inside the ellipse. N = 22 data points. (I) ViSNE clustering analysis of NK cells according to sample origin, showing expression levels of surface markers that were visibly different between 2009-0266 and 2012-0708.