NK cell education after allogeneic transplantation: dissociation between recovery of cytokine-producing and cytotoxic functions

Abstract

Natural killer (NK) cells mediate GVL effects after allogeneic hematopoietic cell transplantation (allo-HCT) by the production of inflammatory cytokines and by direct target lysis. The acquisition of both functions was presumed to be developmentally linked, but this linkage remained unstudied after allo-HCT. We tested the cytokine production and degranulation of reconstituting NK cells after adult unrelated donor or umbilical cord blood grafting. Recipients of T cell-depleted transplants, receiving no immune suppression, showed diminished NK cell degranulation. In contrast, degranulation was normal or increased after T-cell replete transplants given with immune suppression. Strikingly, target cell-induced IFNγ production was markedly diminished in all transplant settings, especially with T cell-depleted or naive T cell-containing umbilical cord blood grafts, suggesting a role for T cells in NK education. Although degranulation was similar in the KIR(+) and KIR(-) populations that coexpressed NKG2A, target cell-induced IFNγ production was limited to the subset of NK cells expressing KIR inhibited by self-ligands. Thus, cytokine production and cytotoxic function do not consistently coexist in NK cells reconstituting after allo-HCT. Exposure to IL-15 rapidly increased target-inducible IFNγ production, indicative of IL-15's potential as a therapeutic tool to enhance NK cell function to protect against infection and relapse after allo-HCT.

Figure 1

Target cell–induced IFNγ production, but not CD107a expression, is decreased after allogeneic transplant. PBMCs from healthy donors and recipients 3 months after T-cell replete, T-cell deplete, and UCB HCT were incubated with media alone, K562 cells, or IL-12/IL-18. CD107a expression and IFNγ production were measured on CD56⁺CD3⁻–gated NK cells. Representative FACS plots are shown. The numbers in each quadrant are the percentages based on total CD56⁺CD3⁻ NK cells.

Figure 2

Target cell-induced IFNγ production is decreased after T-cell replete, T-cell deplete, and UCB transplantation. PBMCs from healthy donors and recipients at 1 (T-cell deplete and UCB only), 3, and 6 months post-HCT were incubated in either media alone (not shown), with K562 cells for 4 hours (top and middle panels), or with IL-12 and IL-18 overnight (bottom panel). After incubation, CD107a expression and IFNγ production were measured on CD56⁺CD3⁻–gated NK cells. Bars represent the mean ± SEM. Donors were compared with recipient samples using the Wilcoxon rank sum test; statistical significance is indicated as *P < .05, **P < .01, and ***P < .0001.

Figure 3

KIR⁺ NK cells produce higher levels of IFNγ after target cell exposure than do KIR⁻ NK cells. (A) CD56⁺CD3⁻–gated NK cells from donors and recipient post-HCT samples at the indicated time points were further gated as KIR⁺ (black bars) or KIR⁻ (white bars) based on staining with a cocktail of antibodies containing CD158a/h, CD158b/j, and CD158e. Expression of CD107a and production of IFNγ were measured after incubation with K562 cells for 4 hours. IFNγ production also was measured after overnight stimulation with IL-12 and IL-18. Bars represent the mean ± SEM. KIR⁺ NK cells were compared with KIR⁻ NK cells at each time point using the Wilcoxon signed rank test; statistical significance is indicated as *P < .05, **P < .01, and ***P < .0001. (B) CD56⁺CD3⁻–gated NK cells from donors and recipient post-HCT samples from the indicated time points were further gated as CD56^bright or CD56^dim. Production of IFNγ was measured after overnight stimulation with IL-12 and IL-18. Bars represent the mean ± SEM. CD56^bright NK cells were compared with CD56^dim NK cells at each time point using the Wilcoxon signed rank test; statistical significance is indicated as *P < .05, **P < .01, and ***P < .0001.

Figure 4

KIR⁺NKG2A⁻ NK cells are decreased and KIR⁻NKG2A⁺ NK cells are increased after HCT. CD56⁺CD3⁻–gated NK cells from donors (black bars) and post-HCT recipients (white bars indicate T-cell replete; light gray bars indicate T-cell deplete, and dark gray bars indicate UCB) at 3 and 6 months were further gated based on staining with a cocktail of KIR antibodies and NKG2A into KIR⁻NKG2A⁻, KIR⁻NKG2A⁺, KIR⁺NKG2A⁺, or KIR⁺NKG2A⁻ populations. Bars represent the mean ± SEM frequency of each NK cell population. Donors were compared with recipient samples using the Wilcoxon rank sum test; statistical significance is indicated as *P < .05, **P < .01, and ***P < .0001.

Figure 5

NKG2A educates reconstituting post-HCT NK cells for degranulation but not for cytokine production. CD56⁺CD3⁻–gated NK cells from donor and post-HCT recipient pairs receiving T-cell replete transplants were further gated as KIR⁻NKG2A⁻ (light gray bars), KIR⁻NKG2A⁺ (white bars), KIR⁺NKG2A⁺ (dark gray bars), or KIR⁺NKG2A⁻ (black bars) based on staining with a cocktail of KIR antibodies and NKG2A. CD107a expression and IFNγ production were measured on all 4 subsets after incubation with K562 cells for 4 hours. IFNγ also was measured after overnight stimulation with IL-12 and IL-18. Bars represent the mean ± SEM. Pairwise comparisons were based on a linear mixed model. Statistical significance is indicated as *P < .05, **P < .01, and ***P < .0001. ns = not significant. NE = not evaluable.

Figure 6

KIR expression positively correlates with IFNγ production. The percentages of CD56⁺CD3⁻–gated NK cells producing IFNγ following incubation with K562 cells were plotted against the percentages of KIR⁺ NK cells or NKG2A⁺KIR⁻ NK cells from both donors and post-HCT recipients of T-cell replete grafts. The estimated regression line is shown with the r value and significance based on the Pearson correlation coefficient. Significance is calculated as P < .05.

Figure 7

Self-KIR educate for cytotoxicity and IFNγ production, and low-dose IL-15 exposure restores the post-HCT defect in IFNγ response. (A) Single KIR⁺NKG2A⁻ NK cells from both donors and post-HCT recipients of T-cell replete grafts were grouped as expressing either self-KIR (black bars) or non-self KIR (white bars) based on the donor or recipient KIR ligand status. Bars represent the mean ± SEM. NK cells expressing self-KIR were compared with NK cells expressing non-self KIR using the paired Student t test. Statistical significance is indicated as *P < .05. (B) PBMCs from recipients 3 months after transplantation with T-cell replete grafts were incubated overnight in media alone (squares) or media containing 1 ng/mL IL-15 (triangles). They were then tested for response induced by K562 exposure; CD107a expression and IFNγ production were measured on CD56⁺CD3⁻–gated NK cells. The paired Student t test was used to compare the function of cells incubated in media versus media containing 1 ng/mL IL-15. Statistical significance is indicated as *P < .05 and **P < .01.