Implication of Interleukin-12/15/18 and Ruxolitinib in the Phenotype, Proliferation, and Polyfunctionality of Human Cytokine-Preactivated Natural Killer Cells

Abstract

A brief in vitro stimulation of natural killer (NK) cells with interleukin (IL)-12, IL-15, and IL-18 endow them a memory-like behavior, characterized by higher effector responses when they are restimulated after a resting period of time. These preactivated NK cells, also known as cytokine-induced memory-like (CIML) NK cells, have several properties that make them a promising tool in cancer immunotherapy. In the present study, we have described the effect that different combinations of IL-12, IL-15, and IL-18 have on the generation of human CIML NK cells. Our data points to a major contribution of IL-15 to CIML NK cell-mediated cytotoxicity against target cells. However, the synergistic effect of the three cytokines grant them the best polyfunctional profile, that is, cells that simultaneously degranulate (CD107a) and produce multiple cytokines and chemokines such as interferon γ, tumor necrosis factor α, and C-C motif chemokine ligand 3. We have also analyzed the involvement of each cytokine and their combinations in the expression of homing receptors CXCR4 and CD62L, as well as the expression of CD25 and IL-2-induced proliferation. Furthermore, we have tested the effects of the Jak1/2 inhibitor ruxolitinib in the generation of CIML NK cells. We found that ruxolitinib-treated CIML NK cells expressed lower levels of CD25 than non-treated CIML NK cells, but exhibited similar proliferation in response to IL-2. In addition, we have also found that ruxolitinib-treated NK cells displayed reduced effector functions after the preactivation, which can be recovered after a 4 days expansion phase in the presence of low doses of IL-2. Altogether, our results describe the impact that each cytokine and the Jak1/2 pathway have in the phenotype, IL-2-induced proliferation, and effector functions of human CIML NK cells.

Keywords: cytokine preactivation; cytokine production; degranulation; memory-like natural killer cells; natural killer cells; polyfunctionality; ruxolitinib.

Figure 1

Graphic representation showing the culture conditions of natural killer (NK) cells. During the preactivation phase (16 –18 h), cells were stimulated with different combinations of IL-12, IL-15, and IL-18 in the presence and absence of ruxolitinib. During the expansion phase, NK cells were cultured with IL-2 for 4 days. At the end of the preactivation and expansion phases, the phenotype and effector functions of NK cells were tested.

Figure 2

Cytotoxicity of cytokine preactivated natural killer (NK) cells at the end of the preactivation phase. (A) Bar graph showing the specific lysis of K562 target cells by control non-preactivated and cytokine preactivated NK cells at 10:1 E:T ratio (n = 4). The bar graph shows the mean with SEM. Differences between cytokine-induced memory-like (CIML) NK cells and the rest of conditions were established with Friedman test. (B) Dot plot graphs showing the specific lysis of K562 target cells by control non-preactivated and CIML NK cells exposed or not to 0.1 µM ruxolitinib at 10:1 (left) and 5:1 (right) E:T ratios (n = 4). **p < 0.01.

Figure 3

Polyfunctionality of cytokine preactivated natural killer (NK) cells after the preactivation phase. (A) Pie charts representing the percentages of control non-preactivated and cytokine preactivated NK cells expressing CD107a, interferon (IFN)γ, tumor necrosis factor (TNF)α, and/or C-C motif chemokine ligand (CCL)3 (n = 4). (B) Pie charts representing the percentages of control non-preactivated and cytokine-induced memory-like (CIML) NK cells, and CIML NK cells preactivated in the presence of 0.1 and 1 µM ruxolitinib, expressing CD107a, IFNγ, TNFα, and/or CCL3 (n = 4). Differences between pie charts were established with non-parametric permutation test. The p-values are in the boxes below the pie charts. Significant differences are in the red cells.

Figure 4

Phenotype of cytokine preactivated natural killer (NK) cells after the preactivation phase. (A) Representative experiment (top) and bar graphs (bottom) showing the expression of CD25, CD62L, and CXCR4 on control non-preactivated and cytokine preactivated NK cells (n = 8). (B) A representative experiment (top) and bar graphs (bottom) showing the expression of CD25, CD62L and CXCR4 on control non-preactivated and cytokine-induced memory-like (CIML) NK cells, and CIML NK cells preactivated in the presence of 0.1 and 1 µM ruxolitinib (n = 8). The bar graphs show the mean with SEM. Differences between CIML NK cells and the rest of conditions were established with RM one-way ANOVA with the Greenhouse-Geisser correction (for normally distributed data) or Friedman test (for non-normal distributed data). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 5

IL-2-induced proliferation of cytokine preactivated natural killer (NK) cells at the end of the expansion phase of 4 days. (A) Histograms of a representative experiment (left) and bar graphs showing the division index (DI) (right) of control non-preactivated and cytokine preactivated NK cells (n = 6). (B) A representative experiment (left) and bar graphs showing the DI (right) of control non-preactivated and cytokine-induced memory-like (CIML) NK cells, and CIML NK cells preactivated in the presence 0.1 and 1 µM ruxolitinib (n = 5). (C) A representative experiment (left) and bar graphs showing the DI (right) of control non-preactivated and CIML NK cells, preactivated in the presence and absence of 1 µM ruxolitinib, exposed to 2 or 20 U/mL IL-2 during the expansion phase (n = 5). The bar graphs show the mean with SEM. Differences between CIML NK cells and the rest of conditions were established with Friedman test. *p < 0.05, **p < 0.01.

Figure 6

Polyfunctionality of cytokine-induced memory-like (CIML) natural killer (NK) cells after the expansion phase. Pie charts represent the percentages of control non-preactivated and CIML NK cells, preactivated in the presence and absence of 1 µM ruxolitinib, expressing CD107a, IFNγ, and/or TNFα. Cells were exposed to 2 U/mL (lower six pie charts) or 20 U/mL (upper six pie charts) of IL-2 during the expansion phase of 4 days, and then washed and incubated with or without K562 cells for 6 h (n = 6). Differences between pie charts were established with non-parametric permutation test. The p values are in the box below the pie charts. Significant differences are in the red cells.

Figure 7

CD62L and CXCR4 expression of cytokine-induced memory-like (CIML) natural killer (NK) cells after the expansion phase. Bar graphs showing the percentage of CD62L+ NK cells and the expression (MFI, median fluorescence intensity) of CXCR4 in control non-preactivated and CIML NK cells, preactivated in the presence and absence of 1 µM ruxolitinib, exposed to 2 or 20 U/mL IL-2 during the expansion phase (n = 6). Bar graphs show the mean with SEM. Statistical differences were established with Wilcoxon matched-pairs signed rank test. *p < 0.05.