Neutrophil depletion impairs natural killer cell maturation, function, and homeostasis

Abstract

Natural killer (NK) cells are bone marrow (BM)-derived granular lymphocytes involved in immune defense against microbial infections and tumors. In an N-ethyl N-nitrosourea (ENU) mutagenesis strategy, we identified a mouse mutant with impaired NK cell reactivity both in vitro and in vivo. Dissection of this phenotype showed that mature neutrophils were required both in the BM and in the periphery for proper NK cell development. In mice lacking neutrophils, NK cells displayed hyperproliferation and poor survival and were blocked at an immature stage associated with hyporesponsiveness. The role of neutrophils as key regulators of NK cell functions was confirmed in patients with severe congenital neutropenia and autoimmune neutropenia. In addition to their direct antimicrobial activity, mature neutrophils are thus endowed with immunoregulatory functions that are conserved across species. These findings reveal novel types of cooperation between cells of the innate immune system and prompt examination of NK cell functional deficiency in patients suffering from neutropenia-associated diseases.

Figure 1.

An extrinsic factor contributes to NK cell hyporeactivity in Genista mice. (A) Circulating NK cells from WT (top) or Genista mice (bottom) were stimulated for 4 h with YAC-1 target cells (right) or medium alone (left). Representative FACS histograms show NK1.1 and CD107a expression gated on NK1.1⁺ CD3⁻ cells. (B) Histograms representing the percentage of CD107a⁺ blood NK cells in WT or Genista mice upon stimulation with YAC-1 tumor cells. Data were pooled from three independent experiments. n = 3. (C) WT and β2mKO splenocytes were stained with two different doses of the fluorescent dye CFSE, mixed at a 1:1 ratio, and transferred into WT or Genista mice. The percentage of specific lysis of MHC-I–deficient cells corresponds to the ratio of the various populations before and 48 h after co-injection in the spleen. The experiment was performed twice. n = 2–3. (D and E) Splenocytes from WT (shaded bars) or Genista (open bars) mice were stimulated for 4 h with YAC-1 tumor targets or medium alone (D), or with anti-NK1.1, anti-NKp46, or isotype control mAb-coated plates (E). NK cell intracellular IFN-γ production (D and E, top) and degranulation (CD107a surface exposure; D and E, bottom) were measured by FACS on NK1.1⁺ CD3⁻ or NKp46⁺ CD3⁻ NK cells. Data were pooled from five independent experiments. n = 3–6. (F) Representative FACS profiles are shown for NK1.1, NKp46 surface expression, or isotype control mAb (IC, dashed) on WT (black) or Genista (gray) NK cells. (G) Splenic NK cells from WT (shaded bars) or Genista (open bars) mice were stimulated for 4 h with a mixture of PMA and ionomycin. Data show the frequencies of IFN-γ⁺ (left) or CD107a⁺ (right) NK cells among the total NK cell population in WT or Genista, or the amount of IFN-γ produced (MFI) per IFN-γ⁺ NK cells from WT or Genista (middle). MFI control values were calculated as the average of the IFN-γ mean fluorescence intensity (MFI) of IFN-γ⁺ NK cells from WT mice. (H) Purified WT spleen NK cells were transferred to WT (shaded bars) or Genista (open bars) hosts, as indicated. 7 d after transfer, the frequencies of IFN-γ–producing cells (left) and CD107a⁺ cells (right) among transferred WT NK cells were analyzed after stimulation with isotype control, anti-NK1.1, or anti-NKp46 mAb-coated plates. Experiments were repeated three times, with n = 2–4 transfer groups. Statistical significance was determined with a Mann-Whitney test. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 2.

Genista mice lack mature neutrophils. (A and B) Representative FACS profiles of blood, spleen, and BM cells from WT or Genista mice stained with anti-CD11b and anti-Ly6G mAb (A) or anti-CD11b and anti-Ly6C mAb (B). Percentage of neutrophils (CD11b^high Ly6G^high; A) and monocytes (CD11b^inter Ly6C^high; B) are indicated. Data are representative of 6–10 independent experiments. n = 3–4. (C) Gating strategy for identification of DC populations in the spleen of WT (top) or Genista (bottom) mice. Live, single, CD11c⁺ Lin⁻ (CD3, CD19, NK1.1) spleen cells were plotted for CD11c and Siglec-H expression, with numbers indicating the percentage of pDCs (CD11c⁺ Siglec-H⁺). CD11c^high DCs were then subdivided in CD11b^high CD8α⁻ (conventional DCs) and CD11b⁻ CD8α⁺ (CD8α DCs) cells. Numbers indicate the percentage of each population. Data represent two experiments. n = 4. (D) Histograms representing the total numbers of neutrophils, monocytes, and DCs subpopulations in the spleen of WT (black) or Genista (white) mice. Live, single, Lin⁻ (CD3, CD19 NK1.1, Ly6G), CD115⁺ CD11b⁺ monocytes were further subdivided into inflammatory (Ly6C⁺) and resident (Ly6C⁻) monocytes. Experiments were performed two to five times. n = 3–4. Statistical significance was determined with a Mann-Whitney test. *, P < 0.05; **, P < 0.01; ***, P < 0.001. (E) Representative FACS profiles of F4/80 and CD11b expression on BMMs from WT or Genista mice obtained after 8 d of in vitro differentiation of BM cells with GM-CSF. (F) CD80, CD69, and CD86 cell surface expression of BMMs from WT (top) or Genista (bottom) mice after overnight incubation with LPS (black histograms) or medium (filled gray histograms). (G) Representative FACS profiles of MHC-II and CD11c expression on BMDCs from WT or Genista mice obtained after 8 d of in vitro differentiation of BM cells with GM-CSF. (H) CD80, CD69, CD86, and MHC-II cell surface expression on BMDCs from WT (top) or Genista (bottom) after overnight incubation with LPS (black histograms) or medium (filled gray histograms). Experiments were performed twice. n = 3.

Figure 3.

Genista NK cells regain full reactivity in a WT environment. Purified Genista spleen NK cells were transferred to WT or Genista (G) hosts, as indicated. 7 d after transfer, the frequency of IFN-γ–producing cells among transferred Genista NK cells (shaded bars) was analyzed after stimulation with isotype control, anti-NK1.1, and anti-NKp46 mAb-coated plates or a mix of PMA and ionomycin. The reactivity of NK cells from WT or Genista hosts was indicated as controls (open bars). Experiments were repeated three times. n = 2–3. Statistical significance was determined with a Mann-Whitney test. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 4.

NK cells and neutrophils are localized in the same areas of the spleen and the lymph nodes and form conjugates. (A) Immunostaining on NKp46^iCre/wtR26R^eYFP/wt spleen (top) or lymph node (bottom) sections at steady state showing neutrophils (anti-MMP9, red), NK cells (anti-GFP, green), B cells (anti-B220, blue), and the lymphatic vessels (Lyve-1, white in the lymph node image). Adjacent overlapping confocal images of spleen and lymph node sections were stitched to generate panoramic images. White arrowheads show contact between neutrophils and NK cells in the spleen. (B) Percentages of conjugates formed between neutrophils and NK cells (NK cells + Neutro), CD3⁺ T cells (T cells + Neutro), or CD3⁻ NK1.1⁻ cells (B cells + Neutro) after 10 min incubation together in vitro at 37°C. Data are representative of three independent experiments. Statistical significance was determined with a Mann-Whitney test. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 5.

Neutrophils are essential for NK cell function at the steady state. (A) Representative FACS profiles of spleen cells from Rat IgG2b- or RB6-8C5–treated mice (top) or, Rat IgG2a-, or 1A8-treated mice (bottom), stained with anti-Ly6C and anti-CD11b mAb. 1A8 is specific for Ly6G, whereas RB6-8C5 recognizes Ly6G and, with a lower affinity, Ly6C (Fleming et al., 1993). As RB6-8C5 recognizes a different epitope on Ly6C than the monoclonal anti-Ly6C (AL-21) antibody (Ribechini et al., 2009), we used CD11b and Ly6C to monitor the percentage of both neutrophils (CD11b^high Ly6C^inter, red circle) and monocytes (CD11b^inter Ly6C^high, black circle). (B) 6 d after treatment, splenocytes from rat IgG2b- and RB6-8C5–treated (top) or, Rat IgG2a- or 1A8-treated mice (bottom), were stimulated by incubation for 4 h on anti-NK1.1, anti-NKp46, or isotype control mAb-coated plates or with a mixture of PMA and ionomycin. NK cell intracellular IFN-γ production was measured by FACS. Experiments were repeated 6–20 times. n = 3–6 mice per group. Statistical significance was determined in a Mann-Whitney test. *, P < 0.05; **, P < 0.01; ***, P < 0.001. (C) Percentages of neutrophils (red circles) and monocytes (black circles) in the blood and the BM of rat IgG2a- or 1A8-treated mice. Data are representative of 6–10 independent experiments. n = 3–4. (D) 6 d after anti-CD4 antibody (GK1.5) or isotype control antibody (rat IgG2b) injection, the presence of CD4⁺ cells among CD3⁺ T cells was monitored in the spleen of rat IgG2b- or GK1.5-treated mice. (E) 6 d after depletion, splenic NK cells were stimulated for 4 h by incubation in plates coated with NK1.1, NKp46, or isotype control mAb, or with a mixture of PMA and ionomycin. The percentages of NK cells expressing intracellular IFN-γ are shown for control isotype-treated mice (shaded bars) and CD4-depleted mice (open bars). Experiments were performed twice. n = 2–3.

Figure 6.

Neutrophils are required for terminal NK cell maturation. (A and B) Representative FACS profiles of splenic NK cells stained with anti-CD27 and anti-CD11b mAb from WT and Genista mice (A) or 1A8, RB6-8C5, or isotype control mAb–treated mice (B). Experiments were performed 6–20 times. n = 3–6. (C) Representative FACS histograms of CD43 expression of splenic NK cells from WT (black line) or Genista (gray line). Tinted histogram shows isotype control staining. Data represent at least 20 independent experiments. (C–E) Percentages of splenic NK cells positive for CD43, for WT and Genista mice (C, right); isotype control-, 1A8-, or RB6-8C5–treated mice (D); or donor WT NK cells 7 d after transfer into WT or Genista hosts (E). Experiments were performed three to six times. n = 3–4. Statistical significance was determined in a Mann-Whitney test. *, P < 0.05; **, P < 0.01; ***, P < 0.001. (F) Representative FACS profiles of splenic NK cells (gated on CD3⁻ NKp46⁺ or NK1.1⁺) from WT mice stained for CD43 and intracellular IFN-γ, 4 h after stimulation by incubation in isotype control, anti-NK1.1, or anti-NKp46 mAb-coated plates. (G) Frequencies of IFN-γ–producing splenic NK cells generated from CD43⁺ or CD43⁻ WT NK cells by stimulation on anti-NK1.1 or anti-NKp46 mAb-coated plates. (H) Amount of IFN-γ secreted per CD43⁺ (shaded bars) or CD43⁻ (open bars) NK cell in response to stimulation by incubation in anti-NK1.1 (top) or anti-NKp46 (bottom) mAb-coated plates. The MFI for IFN-γ of IFN-γ⁺ CD43⁻ NK cells was normalized with respect to that for IFN-γ⁺ CD43⁺ NK cells. Experiments were performed twice. n = 5–8. Statistical significance was determined in a Mann-Whitney test. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 7.

Higher rates of proliferation and poorer survival of NK cells in the absence of neutrophils. (A and B) Representative FACS profiles of splenic NK cells from WT and Genista (A, or isotype control– and 1A8-treated mice (B) stained with anti-NK1.1 and anti-Ki67 mAb. Data a representative of three independent experiments. n = 3–4. (C and D) CFSE-labeled CD45.1⁺ WT spleen NK cells (donor) were transferred into CD45.2⁺ WT or Genista hosts for 7 d. (C) Representative FACS profiles show the percentage of donor WT NK cells (rectangle) after transfer into WT (top left) or Genista (top right) mice. Histograms of CFSE expression gated on donor CD45.1⁺ WT NK cells after transfer into WT (bottom left) or Genista (bottom right) mice. (D) 7 d after transfer into WT (left) or Genista (right) mice, the frequency of undivided (CFSE^high) or divided (CFSE^diluted) donor WT NK cells producing IFN-γ in response to stimulation with isotype control, anti-NK1.1, or anti-NKp46–coated plates was determined. (top) Representative FACS profile upon NK1.1 activation. (bottom) Statistical comparison of NK cell responsiveness in undivided (CFSE^high) and divided (CFSE^diluted) donor NK cells. Experiments were performed three times. n = 2–4. (E) Representative FACS profiles of WT or Genista NK cells (gated on CD3⁻ NKp46⁺ or NK1.1⁺) stained with anti–Annexin V and propidium iodine (Pi) after overnight incubation with medium, 1,000 U/ml IL-2 or 15 ng/ml IL-15. (F) Frequencies of Pi⁺ NK cells from WT (shaded bars) or Genista (open bars) mice after overnight incubation with medium, 1,000 U/ml IL-2 or 15 ng/ml IL-15. Experiments were repeated three times. n = 7–10. Statistical significance was determined in a Mann-Whitney test. *, P < 0.05; **, P < 0.01; ***, P < 0.005.

Figure 8.

Impaired NK cell maturation and function in patients with SCN. Total NK cell (CD3⁻CD56⁺; A) and percentages of CD56^bright NK cells (B) in peripheral blood of healthy controls (black squares), SCN patients not treated with G-CSF (open circles) or SCN patient treated with G-CSF (black diamonds). (C-F) PBMCs from healthy control individuals (black squares), SCN patients not treated with G-CSF (open circles), or SCN patient treated with G-CSF (black diamonds) were stimulated for 4 h with K562 target cells (C–E) or antibody-coated cells (ADCC; D-F) at an effector/target ratio of 5:2 and 50:1, respectively. Frequencies of IFN-γ–producing NK cells (C and D) and NK cell cytotoxic activity (E and F) were determined by FACS. Statistical significance was determined in a Mann-Whitney test. *, P < 0.05; **, P < 0.01; ***, P < 0.005.

Figure 9.

Impaired NK cell maturation and function in patients with AIN. Total NK cell (CD3⁻CD56⁺; A) and percentages of CD56^bright NK cells (B) in peripheral blood of healthy control individuals (black squares) and AIN patients (open circles). (C–F) PBMCs from healthy controls (black squares) or AIN patients (open circles) were stimulated for 4 h with K562 target cells (C–E) or ADCC (D–F) at an effector/target ratio of 5:2 and 50:1, respectively. (C and D) Frequencies of IFN-γ–producing NK cells and (E and F) NK cell cytotoxic activity were measured by FACS. Statistical significance was determined in a Mann-Whitney test. *, P < 0.05; **, P < 0.01; ***, P < 0.005. (G) Representation of the neutrophil count (cells/microliter) in peripheral blood of AIN patient #25 during neutropenia and after remission. Dashed line indicates the neutropenia threshold of 1,500 neutrophils/µl. (H) PBMCs from AIN patient #25 during neutropenia (top) or after remission (bottom) were stimulated for 4 h with K562 target cells, ADCC, or medium. FACS profiles show IFN-γ⁺ NK cells among (CD3⁻ CD56⁺) NK cells. Percentages of IFN-γ⁺ NK cells are indicated.