NK cell phenotype in patients with HGBL-MYC/BCL2, patients with DLBCL NOS, and HDs. (A-B) Expression analysis of major NK cell subsets (A) and NK cell receptors (B) in HDs (green; n = 15 in A; n = 13 in panel B), patients with HGBL-MYC/BCL2 (red; n = 51 in A; n = 41 in panel B), and patients with DLBCL NOS (blue; n = 35 in A; n = 25 in panel B). (C) In-depth analysis of the computationally identified NK cell populations by FlowSOM expressed higher (red) or lower (blue) in patients with HGBL-MYC/BCL2 vs patients with DLBCL NOS. (D-E) Expression analysis of TIM-3, DNAM-1 (D), and HLA-DR (E) on NK cell subsets in HDs (green; n = 13-15), patients with HGBL-MYC/BCL2 (red; n = 41-51; based on sample availability for the 2 T-cell panels; see also supplemental Figure 1), and patients with DLBCL NOS (blue; n = 25-35) and dimensionality reduction by UMAP of the flow cytometry data of NK cells. Color overlays for DNAM-1 and HLA-DR are depicted. (F) Comparative percentages of kill of K562 and degranulation as measured by CD107a/b surface expression on NK cells in HDs (green; n = 7), patients with HGBL-MYC/BCL2 (red; n = 11), and patients with DLBCL NOS (blue; n = 11) after 4-hour coculture of PBMCs with K562 cell line. Cytotoxicity is calculated relative to the amount of K562 cells without PBMCs. For all box plots, the lower upper hinges correspond to the 25th and 75th percentiles. The middle hinge corresponds to the median. The whiskers extend from the largest to smallest value ± 1.58 ∗ interquartile range. Samples are plotted individually, bone marrow samples (n = 3 HGBL-MYC/BCL2 and n = 1 DLBCL NOS) are indicated as square. Mann-Whitney U test between 2 groups was used for statistical analysis. P < .05 was considered significant.