Distinct CD16a features on human NK cells observed by flow cytometry correlate with increased ADCC

Abstract

Natural killer (NK) cells destroy tissue that have been opsonized with antibodies. Strategies to generate or identify cells with increased potency are expected to enhance NK cell-based immunotherapies. We previously generated NK cells with increased antibody-dependent cell mediated cytotoxicity (ADCC) following treatment with kifunensine, an inhibitor targeting mannosidases early in the N-glycan processing pathway. Kifunensine treatment also increased the antibody-binding affinity of Fc γ receptor IIIa/CD16a. Here we demonstrate that inhibiting NK cell N-glycan processing increased ADCC. We reduced N-glycan processing with the CRIPSR-CAS9 knockdown of MGAT1, another early-stage N-glycan processing enzyme, and showed that these cells likewise increased antibody binding affinity and ADCC. These experiments led to the observation that NK cells with diminished N-glycan processing capability also revealed a clear phenotype in flow cytometry experiments using the B73.1 and 3G8 antibodies binding two distinct CD16a epitopes. We evaluated this "affinity profiling" approach using primary NK cells and identified a distinct shift and differentiated populations by flow cytometry that correlated with increased ADCC.

Keywords: N-glycosylation; ADCC; CD16a; Fc γ receptor IIIa; Natural killer cell.

Figure 1

Polyclonal IgG reduced the ADCC (by TR-F) of kifunensine-treated YTS-CD16a cells. (A) YTS-CD16a cells show a slight trend in ADCC decrease only upon IgG blocking at high concentration (2 mg/mL), (B) YTS cells show no ADCC or IgG blocking due to lack of CD16a receptor, (C) Kifunensine-treated (Kif., 20 µM) YTS-CD16a cells show significant decrease in ADCC with IgG blocking at high concentration (2 mg/mL), (D) Kifunensine-treated (Kif., 100 µM) YTS-CD16a cells show significant decrease in ADCC with IgG blocking at both low and high concentrations (0.5 and 2 mg/mL). For all panels, data points represent the mean for three independent experiments collected on three different days at 20:1 E:T ratio, each with three replicates, ± SD by one-way ANOVA. ns, not significant; RTX, rituximab (20 µg/mL); **p < 0.01; ****p < 0.0001.

Figure 2

Minimally-processed N-glycans are found on NK cells with greater ADCC. Relative intensity of the ten most abundant N-glycans identified from YTS-CD16a cells using NSI-MS/MS. (A) Wild-type cells, (B) the MGAT1 knockdown clone E1, (C) cells treated with kifunensine (Kif., 20 µM), (D) cells treated with kifunensine (Kif., 100 µM). Pie charts represent the distribution of all N-glycan types identified. Cartoons indicate one possible glycan configuration consistent with the composition, isobaric species were not distinguished. Monosaccharide residues in the cartoons are defined by the CFG convention. The code for the glycan composition on the x-axes reports the numbers of the following residues as identified by MS (HexNAc-hexose-deoxyhexose-Neu5Ac).

Figure 3

IgG blocks ADCC (by TR-F) of primary NK cells. (A-C) IgG blocking (2 mg/mL) of ADCC following treatment with kifunensine (Kif., 20 µM) for primary NK cells isolated from three donors at 20:1 E:T ratio. Means are indicated with a horizontal black line ± SD. RTX, rituximab (20 µg/mL). For all panels, one-way ANOVA; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure 4

MGAT1 knockdown affects YTS-CD16a NK cell effector function and IgG binding sensitivity to CD16a receptor. (A) Schematic diagram of the dual CRISPR-Cas9 plasmid system and strategy used in this study to generate MGAT1 knockdown clones with the YTS-CD16a cell line, (B) Representative flow cytometry plots of the multiple rounds of bulk-sorting used in this study to increase GFP + YTS-CD16a cell population, (C) Western blot and densitometry of the MGAT1 knockdown clone E1, (D) Representative flow cytometry histograms showing decreased L-PHA lectin staining of MGAT1 knockdown clone E1, (E) ADCC (by TR-F) of the MGAT1 knockdown YTS-CD16a clone E1 and WT cells. Data shown include three independent experiments collected on different days, each with three replicates, by two-way ANOVA, (F) The MGAT1 knockdown clone E1 is likewise sensitive to blocking IgG (2 mg/mL) in ADCC (by TR-F) by two-way ANOVA. Data shown include three independent experiments collected on different days, each with three replicates. RTX, rituximab (20 µg/mL). For panels E and F, ***p < 0.001; ****p < 0.0001.

Figure 5

IgG sensitivity and blocking can differentiate distinct populations in YTS-CD6a cells based on CD16a epitope staining. (A) Representative flow cytometry plots of untreated and kifunensine (Kif.)-treated (20 and 100 µM) YTS-CD16a cells distinctly showing that the 3G8 binding of CD16a on YTS-CD16a cells (x axis) and B73.1 binding of CD16a (y-axis) is affected by IgG blocking and no significant difference is seen comparing both treatment concentrations. Dashed lines indicate the mean fluorescence intensities in each dimension. (B) Representative flow cytometry plots of untreated and kifunensine (Kif.)-treated (20 µM) YTS-CD16a cells gaging sensitivity in differentiating distinct populations based on N-glycan composition at varying percentages of WT and Kif-treated populations.

Figure 6

IgG affinity profiling of primary human NK cells. (A) Flow cytometry plots of primary NK cell donors showing decreased 3G8 staining for kifunensine (20 µM) cells after incubation with 10% human serum, (B) a plot of the mean fluorescence intensities (MFIs) from (A) showing increased B73.1 staining and decreased 3G8 staining associated with high affinity IgG binding following kifunensine treatment, (C) representative flow cytometry plots of primary NK cell donors of healthy adults and children showing considerable heterogeneity, with high affinity phenotype (44%) observed in adult population and less percentage in children population (20%), (D) MFIs for the B73.1+ and 3G8+ population plotted against a heavy black y = x line with NK cells exhibiting greater antibody-binding affinity expected above this line.

Figure 7

ADCC (by flow cytometry) following treatment with kifunensine for primary NK cells. (A,D,G) Lysis, (B,E,H) CD107a, (C,F,I) hIFNγ production of NK cells from donors NK138, NK139 and NK140 treated with kifunensine (“Kif.”; 20 µM) at 20:1T:E ratio. Means are indicated with a horizontal black line ± SD. RTX, rituximab (20 µg/mL). For all panels, one-way ANOVA, *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns not significant, n.d. not detected.