NK cells activated by Interleukin-4 in cooperation with Interleukin-15 exhibit distinctive characteristics

Significance

Natural killer (NK) cells are well known to serve as effecter cells in Th1-type immune responses, whereas their roles in Th2-type immune responses are largely unknown. In this study, we describe a previously unidentified pathway wherein IL-4, an initiator cytokine of Th2-type immune responses, induces development of highly activated murine NK cells in vivo. We show that IL-4 overexpression remarkably increases tissue-resident macrophages, which contributes to NK-cell proliferation via production of IL-15. The NK cells induced by IL-4 secrete a large amount of IFN-γ in response to IL-12, an initiator cytokine of Th1-type immune responses, and also exhibit high cytotoxicity against tumor cells. These results reveal the novel role of IL-4 in immune responses through the induction of unique NK cells.

Abstract

Natural killer (NK) cells are known to be activated by Th1-type cytokines, such as IL-2, -12, or -18, and they secrete a large amount of IFN-γ that accelerates Th1-type responses. However, the roles of NK cells in Th2-type responses have remained unclear. Because IL-4 acts as an initiator of Th2-type responses, we examined the characteristics of NK cells in mice overexpressing IL-4. In this study, we report that IL-4 overexpression induces distinctive characteristics of NK cells (B220^high/CD11b^low/IL-18Rα^low), which are different from mature conventional NK (cNK) cells (B220^low/CD11b^high/IL-18Rα^high). IL-4 overexpression induces proliferation of tissue-resident macrophages, which contributes to NK cell proliferation via production of IL-15. These IL-4–induced NK cells (IL4-NK cells) produce higher levels of IFN-γ, IL-10, and GM-CSF, and exhibit high cytotoxicity compared with cNK cells. Furthermore, incubation of cNK cells with IL-15 and IL-4 alters their phenotype to that similar to IL4-NK cells. Finally, parasitic infection, which typically causes strong Th2-type responses, induces the development of NK cells with characteristics similar to IL4-NK cells. These IL4-NK–like cells do not develop in IL-4Rα KO mice by parasitic infection. Collectively, these results suggest a novel role of IL-4 in immune responses through the induction of the unique NK cells.

Interferon-γ and IL-4 exhibit opposite regulatory roles in the immune system. IFN-γ promotes differentiation of CD4⁺ T cells to Th1 cells and consequently induces Th1 immune responses, whereas IL-4 promotes differentiation of CD4⁺ T cells to Th2 cells and consequently induces Th2 immune responses (1, 2). Furthermore, IFN-γ inhibits Th2 immune responses, whereas IL-4 inhibits Th1 immune responses (3–5).

Natural killer (NK) cells are the third largest lymphocyte population, major IFN-γ producers, and cytolytic effector cells that play major roles for clearance of tumors and virus-infected cells. In immune responses, NK cells are activated by Th1-type cytokines such as IL-2, -12, or -18 (6). Therefore, NK cells are assumed to contribute to Th1 immune responses (7, 8). Conversely, NK cells express IL-4Rα, and key effector functions of NK cells are regulated by IL-4. Previous studies reported that IFN-γ production from murine NK cells cultured in the presence of IL-4 was not detected (9). IL-4 exposure also suppressed the cytotoxic capacity of NK cells in both in vitro and in vivo models (10, 11). However, in contrast to those findings, IL-4 was suggested to induce IFN-γ production by NK cells. Administration of the IL-4 and anti–IL-4 mAb complex (IL-4C), a long-acting formulation of IL-4 in vivo, induced IFN-γ production by NK cells (12). Interestingly, IL-13, which shares many IL-4 effects, failed to stimulate IFN-γ production and suppressed basal IFN-γ production (12). In addition, Kitajima et al. showed that memory type 2 helper T cells enhanced NK cell cytotoxic activity in an IL-4–dependent manner in vivo (13). Thus, there are many apparently contradictory reports on the effect of IL-4 on NK cells, and the role of IL-4 for NK cells remains unclear.

In this study, we investigated the role of IL-4 for murine NK cells in vivo by expressing IL-4 in the liver by the hydrodynamic tail vein injection method (14–16) and show that IL-4 overexpression induces distinctive characteristics of NK cells in vivo. Compared with mature conventional NK (cNK) cells, those IL-4–induced NK (IL4-NK) cells show differences in the sensitivity to IL-12 and -21 and the capacity of cytokine production. Moreover, IL4-NK cells exhibit a higher cytotoxic capacity against tumor cells compared with cNK cells. Importantly, IL-4 induces these phenotypic changes of NK cells directly and also indirectly. Finally, we show that IL4-NK–like cells are also induced in physiological condition, such as parasitic infection. Thus, these findings uncover the novel role of IL-4 for NK cells in vivo.

Results

Overexpression of IL-4 Induces Distinctive Characteristics of NK Cells.

To examine the role of IL-4 on NK cells in vivo, we used the hydrodynamic tail vein injection method to express IL-4 in the liver. At 5 d after the injection with control vector or pLIVE-IL-4 vector, we confirmed the expression of IL-4 mRNA in hepatocytes and an increase of IL-4 protein in the serum by injection of pLIVE-IL-4 vector (Fig. S1A). We then analyzed the cell surface markers on hematopoietic cells in the liver by flow cytometry and found that CD45⁺NK1.1⁺B220⁺CD3e⁻CD19⁻ cells were significantly increased in the liver from the mice overexpressing IL-4 compared with control mice (Fig. 1A). These IL-4–induced cells expressed NK-cell markers such as NK1.1, CD49b, NKp46, and NKG2D. Interestingly, however, they did not express CD3e (T-cell and NK T-cell marker), CD19 (B-cell marker), TNF-related apoptosis inducing ligand (TRAIL) (ILC1 marker), and IL-7Rα (ILC1 marker) (Fig. 1B). Therefore, we concluded that these IL-4–induced cells are NK cells and referred to them as IL4-NK cells. In addition, we examined various cell surface molecules and found that IL4-NK cells showed an expression pattern distinct from mature cNK cells (CD45⁺NK1.1⁺CD11b⁺CD3e⁻CD19⁻) (Fig. 1 A and B and Table S1). Particularly, the expression levels of B220, CD11b, IL-4Rα, IL-18Rα, and IL-21Rα were significantly different between cNK cells and IL4-NK cells (Fig. 1B and Table S1). Moreover, we also found that IL4-NK cells showed an expression pattern distinct from immature CD11b⁻ NK cells (CD45⁺NK1.1⁺CD11b⁻CD3e⁻CD19⁻) (Fig. S1 B and C). In particular, the expression levels of B220, TRAIL, IL-7Rα, and IL-21Rα on IL4-NK cells were different from those on immature CD11b⁻ NK cells (Fig. S1 B and C). The morphology of IL4-NK cells was quite different from that of cNK cells and immature CD11b⁻ NK cells: IL4-NK cells showed a larger size, lower nuclear/cytoplasm ratio, and increased granules (Fig. 1C and Fig. S1D). Although some immature NK cells exhibited a high expression level of B220 and a low expression level of IL-18Rα, their morphology was quite different from that of IL4-NK cells (Fig. S1D). IL4-NK cells were also found in other tissues, including spleen, mesenteric lymph nodes (MLNs), bone marrow, and peripheral blood (Fig. 1D). Because IL4-NK cells were increased significantly in the mice overexpressing IL-4, we investigated the proliferation potential of IL4-NK cells and found that IL4-NK cells markedly incorporated 5-ethynyl-2′-deoxyuridine (EdU) compared with cNK cells or immature CD11b⁻ NK cells in both the liver and the spleen (Fig. 1E and Fig. S1E). These results suggest that IL-4 overexpression induces distinctive characteristics of NK cells in vivo.

Fig. S1.

Comparison of IL4-NK cells with immature NK cells. (A–D) Control vector or pLIVE-IL-4 vector (5 µg) were injected intravenously into C57BL/6 mice. These mice were analyzed 5 d after the injection. (A) Hepatocytes and serum were harvested from these mice. (Left) Expression of IL-4 mRNA was determined by quantitative RT-PCR (qPCR). (Right) The level of IL-4 in serum was measured by ELISA. Data are shown as means ± SEM of three samples. (B) Hematopoietic cells were isolated from the livers of control mice and stained for CD45, CD3e, CD19, NK1.1, and CD11b and analyzed by flow cytometry. Numbers indicate percentages. The graph shows the percentages of CD11b⁻ NK cells or cNK (CD11b⁺ NK) cells among total NK cells, and bars indicate the means of seven mice. Statistical significance was determined by Student’s t test. (C) Immature CD11b⁻ NK cells from mice injected with control vector and IL4-NK cells from mice injected with pLIVE-IL-4 vector were stained for the indicated cell surface markers. Immature CD11b⁻ NK and IL4-NK cells were gated on CD45⁺NK1.1⁺CD11b⁻CD3e⁻CD19⁻ and CD45⁺NK1.1⁺B220⁺CD3e⁻CD19⁻, respectively. Data shown are representative of three independent experiments. (D) Cytospin preparations of immature CD11b⁻ NK cells and immature B220^highIL18Rα^low NK cells were stained by May–Grünwald–Giemsa. (Scale bars, 20 μm.) (E) At 4 d after the injection with vectors, these mice were intraperitoneally injected with 1 mg of EdU or PBS as a control 2 h before death. Histogram plots show the amount of EdU incorporation in cNK cells, immature CD11b⁻ NK cells, and IL4-NK cells from the liver or spleen. Shaded histograms indicate NK cells from the mice injected with PBS. Numbers indicate percentages. Data shown are representative of three independent experiments. **P < 0.01. N.D., not detected.

Fig. 1.

IL-4 overexpression induced distinctive characteristics of NK cells. (A–D) Control or pLIVE-IL-4 vector (5 µg) were injected intravenously into C57BL/6 mice. These mice were analyzed at 5 d after the injection. (A and B) Hematopoietic cells were isolated from the livers of these mice. (A) Cells were stained for CD45, CD3e, CD19, NK1.1, B220, and CD11b and analyzed by flow cytometry. Numbers indicate percentages. R1 and R2 indicate cNK cells (CD45⁺NK1.1⁺CD11b⁺CD3e⁻CD19⁻) and IL4-NK cells (CD45⁺NK1.1⁺B220⁺CD3e⁻CD19⁻), respectively. (B) cNK cells (R1) and IL4-NK cells (R2) were stained for the indicated cell surface markers. cNK cells and IL4-NK cells were gated on CD45⁺NK1.1⁺CD11b⁺CD3e⁻CD19⁻ and CD45⁺NK1.1⁺B220⁺CD3e⁻CD19⁻, respectively. (C) Cytospin preparations of cNK cells and IL4-NK cells were stained by May–Grünwald–Giemsa. (Scale bars, 20 μm.) (D) Hematopoietic cells were isolated from the liver, spleen, MLN, bone marrow (BM), and peripheral blood (PB) of these mice. The graphs show the percentages of cNK distribution (R1; Left) or IL4-NK distribution (R2; Right) in CD45-positive cells. Data are shown as means ± SEM of three mice. (E) At 4 d after the injection with control or pLIVE-IL-4 vector, these mice were intraperitoneally injected with 1 mg of EdU or PBS as a control 2 h before death. The graph represents the means ± SEM of the percentages of EdU-positive cells. (A and B) Data shown are representative of three independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001; N.D., not detected; N.S., not significant.

Table S1.

Expression levels of surface markers on cNK and IL4-NK cells

Markers	cNK cells	IL4-NK cells
Common NK marker
CD49b**	180.0 ± 6.1	86.1 ± 2.5
Maturation markers
B220*	20.2 ± 2.0	145.3 ± 17.4
CD11b***	189.7 ± 1.9	27.5 ± 2.4
TRAIL	6.0 ± 0.7	5.1 ± 0.6
Activating receptors
NK1.1*	491.2 ± 21.4	720.0 ± 53.1
NKp46*	325.7 ± 21.5	222.3 ± 26.0
NKG2D	35.6 ± 0.8	27.2 ± 4.6
2B4*	209.0 ± 16.4	681.3 ± 59.1
Ly108**	9.3 ± 0.8	24.7 ± 2.4
Inhibitory receptors
KLRG1	161.0 ± 33.0	73.2 ± 16.8
NKG2A***	30.6 ± 0.8	11.1 ± 0.3
Activation marker
Thy1.1**	294.3 ± 89.1	963.5 ± 133.9
Resting marker
Ly6C**	106.5 ± 11.2	28.6 ± 5.4
Cytokine receptors
IL-4Rα*	10.1 ± 0.9	31.0 ± 4.9
IL-7Rα	5.3 ± 1.1	4.6 ± 0.1
IL-18Rα*	26.6 ± 4.0	5.0 ± 0.3
IL-21Rα**	4.5 ± 0.3	17.4 ± 0.9

The cNK cells in the liver were derived from the mice at 5 d after the injection with control vector (5 μg). The IL4-NK cells in the liver were derived from the mice at 5 d after the injection with pLIVE-IL-4 vector (5 μg). The numbers indicated MFI ± SEM of at least three separate experiments. *P < 0.05; **P < 0.01; ***P < 0.001.

IL-4 Overexpression Converts cNK Cells to IL4-NK Cells in Vivo.

To investigate the possibility that cNK cells are converted to IL4-NK cells in the mice overexpressing IL-4, we performed an in vivo transplantation assay. We first injected control vector or pLIVE-IL-4 vector intravenously into nonirradiated CD45.1 congenic mice (Fig. 2A). We then isolated cNK cells (CD49b⁺CD11b⁺CD3e⁻FCεR1α⁻) from the spleens of CD45.2 C57BL/6 mice and adoptively transferred into those CD45.1 mice at 24 h after the injection of vectors (Fig. 2A and Fig. S2A). At 3 d after transplantation, the transferred CD45.2⁺ NK cells were sorted from the recipient liver and the spleen, and their cell surface markers were examined (Fig. 2 B and C and Fig. S2B). The CD45.2⁺ NK cells from the mice overexpressing IL-4 displayed higher expression levels of B220 and IL-21Rα than those from the mice injected with control vector (Fig. 2 B and C). Moreover, the former also displayed lower expression levels of CD11b and IL-18Rα than the latter (Fig. 2 B and C). We also examined the contribution of immature CD11b⁻ NK cells to IL-4 NK cells. We isolated cNK (CD11b⁺ NK) and CD11b⁻ NK cells from the spleens of CD45.2 C57BL/6 mice and labeled cNK cells with carboxyfluorescein diacetate succinimidyl ester (CFSE) (Fig. S2C). We then adoptively transferred both CFSE-labeled cNK cells and nonlabeled CD11b⁻ NK cells into CD45.1 mice at 24 h after the injection of pLIVE-IL-4 vector (Fig. S2C). We examined surface markers on donor cells at 1 or 3 d after the transplantation by flow cytometry (Fig. S2D). Both cNK cells and immature CD11b⁻ NK cells were converted to IL4-NK cells gradually. The percentages of donor cells derived from immature CD11b⁻ NK cells at day 1 were lower than those from cNK cells, whereas the former were higher than the latter at day 3 (Fig. S2E). Next, we investigated whether NK cells required the direct IL-4 signal for the conversion and found that NK cells isolated from IL-4Rα KO mice were not converted to IL-4 NK cells in the recipient mice overexpressing IL-4 (Fig. S3 A and B). Interestingly, the change of fluorescence intensity of CFSE revealed that proliferation of NK cells derived from IL-4Rα KO mice was lower compared with NK cells derived from WT mice. Because the IL-4Rα KO mouse was nonresponsive to both IL-4 and -13, we also investigated the role of IL-13 for NK cells and found that overexpression of IL-13 did not induce IL4-NK cells (Fig. S4). Therefore, these results indicate that IL4-NK cells can be derived from cNK and immature CD11b⁻ NK cells by IL-4 in vivo.

Fig. 2.

Conversion of cNK cells to IL4-NK cells. (A) Schematic diagram is shown. (B and C) Control or pLIVE-IL-4 vector (5 µg) was injected intravenously into CD45.1 congenic mice. At 24 h later, cNK cells (CD49b⁺CD11b⁺CD3e⁻FCεR1α⁻) sorted from the spleens of CD45.2 C57BL/6 mice were transferred into these CD45.1 mice. At 3 d after the transfer, hematopoietic cells were harvested from the livers and spleens of the recipient mice, stained for CD45.2, CD49b, and other cell surface markers, and analyzed by flow cytometry. (B) The histograms show surface expression profiles of donor NK cells. Data shown are representative of three independent experiments. (C) Mean fluorescence intensity (MFI) was measured by flow cytometry and normalized to MFI of donor NK cells from the mice injected with control vector. The graphs represent the means ± SEM of relative expression of three mice. *P < 0.05; **P < 0.01; ***P < 0.001.

Fig. S2.

Immature CD11b⁻ NK cells were converted to IL4-NK cells. (A) Splenocytes from CD45.2 C57BL/6 mice were stained for CD49b, CD3e, FCεR1α, and CD11b and CD49b⁺ cells among them were sorted by autoMACS pro. cNK cells (CD49b⁺CD11b⁺CD3e⁻FCεR1α⁻) or CD11b⁻ NK cells (CD49b⁺CD11b⁻CD3e⁻FCεR1α⁻) were then sorted by Moflo XDP. (B) Control vector or pLIVE-IL-4 vector (5 µg) were injected intravenously into CD45.1 congenic mice. At 24 h later, cNK cells sorted from the spleens of CD45.2 C57BL/6 mice were transferred into these CD45.1 mice. At 3 d after the transfer, hematopoietic cells were harvested from the livers and spleens of the recipient mice, stained for CD45.2 and CD49b, and analyzed by flow cytometry. Numbers indicate percentages, and data shown are representative of three independent experiments. (C) Schematic diagram is shown. (D and E) pLIVE-IL-4 vector (5 µg) was injected intravenously into CD45.1 congenic mice. cNK cells and immature CD11b⁻ NK cells were sorted from the spleens of CD45.2 C57BL/6 mice, and cNK cells were labeled with CFSE. At 24 h after the injection, both donor cNK and immature CD11b⁻ NK cells were transferred into these CD45.1 mice. At 1 or 3 d after the transfer, hematopoietic cells were harvested from the livers of the recipient mice, stained for CD45.2, CD49b, and other cell surface markers, and analyzed by flow cytometry. (D) Numbers indicate percentages. The histogram plots show surface expression profiles of donor NK cells. Data shown are representative of three or four independent experiments. (E) The graphs show the percentages of donor cells derived from immature CD11b⁻ NK or donor cells derived from cNK cells among CD45.2⁺ cells, and bars indicate the means of three or four mice. Statistical significance was determined by Student’s t test. **P < 0.01; ***P < 0.001.

Fig. S3.

IL-4Rα–deficient NK cells were not converted to IL4-NK cells. (A) Schematic diagram is shown. (B and C) pLIVE-IL-4 vector (5 µg) was injected intravenously into BALB/c WT mice. Splenocytes were harvested from WT or IL-4Rα KO mice and labeled with CFSE. At 24 h after the injection, donor NK cells were transferred into these BALB/c mice. At 3 d after the transfer, hematopoietic cells were harvested from the livers of the recipient mice, stained for NKp46 and other cell surface markers, and analyzed by flow cytometry. (B) Numbers indicate percentages. The histogram plots show surface expression profiles of donor NK cells. (C, Left) The histogram plot shows the fluorescence intensity of CFSE of donor NK cells. (C, Right) Bars indicate the means of three or five mice. Statistical significance was determined by Student’s t test. ***P < 0.001.

Fig. S4.

IL-13 overexpression did not induce IL4-NK cells. Control vector or pLIVE-IL-13 vector (20 µg) were injected intravenously into C57BL/6 mice. These mice were analyzed 5 d after the injection. (A) Serum was harvested from these mice. The level of IL-13 in serum was measured by ELISA. Data are shown as means ± SEM of three samples. (B and C) Hematopoietic cells were isolated from the livers of these mice. (B) Cells were stained for CD3e, CD19, NK1.1, and CD11b and analyzed by flow cytometry. Numbers indicate percentages. (C) cNK cells (NK1.1⁺CD11b⁺CD3e⁻CD19⁻) derived from mice injected with control vector or CD11b⁻ NK cells (NK1.1⁺CD11b⁻CD3e⁻CD19⁻) and CD11b⁺ NK cells (NK1.1⁺CD11b⁺CD3e⁻CD19⁻) derived from mice injected with pLIVE-IL-13 vector were stained for the indicated cell surface markers. (B and C) Data shown are representative of three independent experiments.

IL-4–Stimulated Macrophages Contribute to NK-Cell Proliferation via IL-15.

As described above, IL4-NK cells proliferated actively in the mice overexpressing IL-4 (Fig. 1E and Fig. S1E). However, cNK cells were not increased when cultured with IL-4 alone (Fig. S5A), suggesting that additional factors are necessary for the development of IL4-NK cells. Previous studies reported that tissue-resident macrophages were increased by IL-4 in vivo (17, 18). Macrophages and dendritic cells (DCs) have been known to transpresent IL-15 to NK cells, which is essential for the survival of NK cells (19–21). Therefore, we examined macrophages in the liver and found that macrophages were remarkably increased in the mice overexpressing IL-4 compared with control mice by immunohistochemical staining (Fig. 3A). Additionally, we observed that granzyme B-positive cells were in close contact with liver macrophages, and most of the granzyme B-positive cells were identified as NK cells by flow cytometry (Fig. 3A and Fig. S5B). These results suggested that macrophages transpresented IL-15 to NK cells. To evaluate the role of macrophages, we depleted macrophages by administration of clodronate liposomes and then, at 24 h later, we overexpressed IL-4 by using the hydrodynamic tail vein injection method. In these experiments, we found that a high dose of IL-4 induced the severe phenotype in the mice when macrophages were depleted. Thus, to avoid the severe phenotype, we injected 1 μg of pLIVE-IL-4 rather than 5 μg. At 3 d after the injection, we analyzed NK cells in the liver and the spleen by flow cytometry (Fig. 3B). Depletion of macrophages suppressed the induction of IL4-NK cells in both the liver and the spleen by overexpressing IL-4, but did not affect the NK cells in the steady state (Fig. 3B and Fig. S5 C and D). We also tested whether administration of clodronate liposomes induced the liver inflammation (Fig. S5E). Although the level of alanine aminotransferase, a specific indicator of liver inflammation, in serum derived from mice injected with clodronate liposomes was slightly higher than that with control liposomes, it remained at a low level. Next, we investigated whether IL4-NK cells require IL-15 produced by liver macrophages for their survival. We performed coculture of cNK cells derived from WT mice with liver macrophages derived from the mice overexpressing IL-4. In this coculture, the addition of IL-15–blocking antibodies (Abs), anti–IL-15/IL-15Rα Abs, inhibited the survival of NK cells (Fig. 3C). These results indicate that overexpression of IL-4 induces proliferation of macrophages, which contributes to NK cell proliferation via IL-15.

Fig. S5.

Macrophages contribute to NK-cell proliferation in the mice overexpressing IL-4. (A) Sorted cNK cells were cultured with IL-4 (50 ng/mL) for the indicated time periods, and cell viability was assayed. Data are shown as means ± SEM of three samples. Data shown are representative of three independent experiments. (B) Control vector or pLIVE-IL-4 vector (5 µg) were injected intravenously into C57BL/6 mice. Hematopoietic cells were isolated from the livers of these mice at 5 d after the injection. Intracellular granzyme B and surface CD49b and CD3e were stained and analyzed by flow cytometry. Numbers indicate percentages. (C–E) Control liposomes containing PBS or clodronate liposomes were intraperitoneally injected into C57BL/6 mice. (C) At 4 d later, liver and spleen sections from these mice were stained for CD68 (macrophage marker; green) by immunohistochemistry. Nuclei (blue) were counterstained with Hoechst 33342. (Scale bars, 100 μm.) (D) The graphs show the total NK cell (NK1.1⁺CD3e⁻CD19⁻) numbers or B220^highIL-18Rα^low NK cell (NK1.1⁺ CD3e⁻CD19⁻B220^highIL-18Rα^low) numbers in the liver or spleen. (E) The graph shows the level of ALT in serum. Serum was harvested at 4 d after the administration of liposomes. Bars indicate the means of four or five mice. (C–E) Data shown are representative of two independent experiments. Statistical significance was determined by Student’s t test. **P < 0.01; N.S., not significant.

Fig. 3.

Macrophages induced by IL-4 promoted NK-cell proliferation. (A) Control or pLIVE-IL-4 vector (5 µg) was injected intravenously into C57BL/6 mice. At 5 d later, liver sections from these mice were stained for CD68 (macrophage marker; green) and granzyme B (NK marker; red) by immunohistochemistry. Nuclei (blue) were counterstained with Hoechst 33342. (Scale bars, 50 μm.) (B) Control liposomes containing PBS or clodronate liposomes were intraperitoneally injected into C57BL/6 mice. At 24 h later, pLIVE-IL-4 vector (1 µg) was injected intravenously into these mice. At 3 d after the injection of pLIVE-IL-4 vector, hematopoietic cells were harvested from the livers (Left) and spleens (Right) and stained for CD45, CD3e, CD19, and NK1.1. Numbers indicate percentages, and data shown are representative of four independent experiments. The graphs show the percentages of NK cells and absolute NK cell numbers from four mice, and bars indicate the means of four mice. Statistical significance was determined by Student’s t test. (C) cNK cells derived from WT mice were cocultured with liver macrophages derived from the mice overexpressing IL-4 with isotype control or anti–IL-15/IL-15Rα Abs. Absolute NK cell numbers were determined in triplicate cultures. Data are shown as means ± SEM of triplicate. Data shown are representative of three independent experiments. *P < 0.05; **P < 0.01.

Different Phenotypes Between cNK and IL4-NK Cells.

NK-cell subsets with a distinct expression pattern of surface markers display differences in cytokine production and cytotoxicity (16, 22–24). Because cNK cells and IL4-NK cells showed distinct expression patterns of surface markers (Fig. 1B), we next examined cytokine production and cytotoxicity of these NK cells. IFN-γ is a major cytokine secreted by NK cells, and IL4-NK cells were found to secrete a higher level of IFN-γ than cNK cells when stimulated with IL-12, -21, and anti-NK1.1 Abs (Fig. 4A). Because previous studies reported that NK cells also produce IL-10 and GM-CSF (25–27), we examined the production of IL-10 and GM-CSF and found that IL4-NK cells secreted both IL-10 and GM-CSF at much higher levels than cNK cells (Fig. 4 B and C). These data suggest that each NK cell shows different sensitivity to IL-12 and -21 and that the capacity of cytokine production is different. In addition to cytokine production, cytotoxicity is another major function of NK cells in the immune system (28). To evaluate the cytotoxic capability of each NK cell, we first investigated the production of granzyme B, which is a cytotoxic granule released from NK cells and kills target cells. Production of granzyme B by IL4-NK cells was much higher than that by cNK cells or immature CD11b⁻ NK cells (Fig. 4D and Fig. S6). Moreover, IL4-NK cells exhibited a higher cytotoxic capacity against YAC-1 cells compared with cNK cells (Fig. 4D). Collectively, these results indicate that the phenotype and functions of IL4-NK cells are different from cNK cells.

Fig. 4.

Cytokine production and cytotoxicity of cNK cells and IL4-NK cells. cNK cells were sorted from the spleens of C57BL/6 mice, and IL4-NK cells were sorted from the livers of C57BL/6 mice injected with pLIVE-IL-4 vector (5 µg) at 5 d after the injection. (A) Sorted NK cells (5 × 10⁴ cells per well) were stimulated with IL-12 (100 ng/mL), IL-21 (100 ng/mL), or anti–NK1.1-Ab in the presence of IL-2 (100 ng/mL). After 24 h of incubation, the level of IFN-γ in the culture supernatants was measured by ELISA. (B) Sorted NK cells (1 × 10⁵ cells per well) were stimulated with IL-12 (100 ng/mL) or IL-21 (100 ng/mL) in the presence of IL-2 (100 ng/mL). After 24 h of incubation, the level of IL-10 in the culture supernatants was measured by ELISA. (C) Sorted NK cells (1 × 10⁵ cells per well) were cultured in the presence of IL-2 (100 ng/mL) or IL-15 (100 ng/mL). After 24 h of incubation, the level of GM-CSF in the culture supernatants was measured by ELISA. (D) Sorted NK cells (5 × 10⁴ cells per well) were cultured in the presence of IL-2 (100 ng/mL). After 24 h of incubation, the level of granzyme B in the culture supernatants was measured by ELISA (Left). CFSE-labeled YAC-1 cells (1 × 10⁴) were cocultured with sorted NK cells for 5 h at various effector-to-target ratios in the presence of IL-2 (100 ng/mL) (Right). The graphs represent the means ± SEM of percentages of CFSE⁺/PI⁺ YAC-1 cells. Data are shown as means ± SEM of three samples. All data shown are representative of three independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001. N.D., not detected; No stim., no stimulation.

Fig. S6.

Representative data from flow-cytometric analysis of the production of intracellular granzyme B. Control vector or pLIVE-IL-4 vector (5 µg) were injected intravenously into C57BL/6 mice. Hematopoietic cells were isolated from the livers of these mice at 5 d after the injection. Immature CD11b⁻ NK and cNK cells from mice injected with control vector and IL4-NK cells from mice injected with pLIVE-IL-4 vector were stained for intracellular granzyme B and surface CD3e, CD19, CD49b, and CD11b and analyzed by flow cytometry.

Development of IL4-NK Cells Requires both IL-4 and -15.

We next examined the direct effect of IL-4 on NK cells in culture. Because it seemed that IL4-NK cells received the IL-15 signal, we added IL-15 to the culture medium of cNK cells. The expression level of IL-18Rα on NK cells cultured for 4 d with IL-15 and -4 was lower than that in NK cells cultured with IL-15 alone. However, expression levels of B220, CD11b, IL-4Rα, and -21Rα were nearly the same in both NK cells (Fig. 5A). These results suggest that the difference of surface markers between IL4-NK and cNK cells is caused by both the direct effect of IL-4 and indirect effect of IL-4, namely, IL-15. In addition, NK cells cultured with IL-15 and -4 secreted a higher level of IFN-γ than NK cells cultured with IL-15 in response to IL-12 and -21 (Fig. 5B). Although NK cells cultured with IL-15 and -4 produced more IL-10 than NK cells cultured with IL-15 in response to IL-12, the former did not produce more IL-10 in response to IL-21 (Fig. 5C). Conversely, production of GM-CSF and granzyme B by NK cells cultured with IL-15 and -4 was much higher than that by NK cells cultured with IL-15 (Fig. 5 D and E). Interestingly, phenotypic changes induced by IL-4 were not caused by IL-13 (Fig. S7). Collectively, these data suggest that the phenotype of NK cells cultured with IL-15 and -4 is similar to that of IL4-NK cells and that induction of IL4-NK cells requires both IL-4 and -15.

Fig. 5.

IL-4 changed the phenotype of cNK cells to that similar to IL4-NK cells in vitro. cNK cells were sorted from the spleens of C57BL/6 mice and cultured with or without IL-4 (200 ng/mL) in the presence of IL-15 (100 ng/mL) for 4 d. (A) The histogram plots show surface expression profiles of cNK cells (shaded) from WT mice, NK cells cultured with IL-15 (dotted lines), and NK cells cultured with IL-15 and -4 (solid lines). (B and C) NK cells (2 × 10⁴ cells per well; B) or (5 × 10⁴ cells per well; C) were stimulated with IL-12 (100 ng/mL) or IL-21 (100 ng/mL) in the presence of IL-2 (100 ng/mL). After 24 h of incubation, the levels of IFN-γ and IL-10 in the culture supernatants were measured by ELISA. (D and E) After 4 d of incubation, NK cells (5 × 10⁴ cells per well; D) or (2 × 10⁴ cells per well; E) were cultured with or without IL-4 (200 ng/mL) in the presence of IL-15 (100 ng/mL) for an additional day. The levels of GM-CSF and granzyme B in the culture supernatants were then measured by ELISA. Data are shown as means ± SEM of three samples. All data shown are from a representative of three independent experiments. Statistical significance was determined by Student’s t test. **P < 0.01; ***P < 0.001. N.D., not detected; No stim., no stimulation; N.S., not significant.

Fig. S7.

IL-13 did not change the phenotype of cNK cells to that similar to IL4-NK cells in vitro. (A) cNK cells were sorted from the spleens of C57BL/6 mice and cultured with or without IL-13 (200 ng/mL) in the presence of IL-15 (100 ng/mL) for 4 d. The histogram plots show surface expression profiles of NK cells cultured with IL-15 (dotted lines) and NK cells cultured with IL-15 and -13 (solid lines). (B) NK cells (2 × 10⁴ cells per well) were stimulated with IL-12 (100 ng/mL) in the presence of IL-2 (100 ng/mL). (Left) After 24 h of incubation, the level of IFN-γ in the culture supernatants was measured by ELISA. After 4 d of incubation, NK cells (5 × 10⁴ cells per well) (GM-CSF) or (2 × 10⁴ cells per well) (granzyme B) were cultured with or without IL-13 (200 ng/mL) in the presence IL-15 (100 ng/mL) for an additional day. (Center and Right) The levels of GM-CSF and granzyme B in the culture supernatants were then measured by ELISA. Data are shown as means ± SEM of three samples. All data shown are representative of three independent experiments. Statistical significance was determined by Student’s t test. No stim., no stimulation; N.S., not significant.

IL4-NK–Like Cells Are Induced by Parasitic Infection.

We further investigated whether IL4-NK cells are induced in physiological conditions. IL-4 production is known to be induced by parasitic infection and plays important roles in elimination of parasites. Therefore, we examined NK cells in mice infected with Nippostrongylus brasiliensis (Nb). We first analyzed cell surface markers on NK cells and found that B220^highIL-18Rα^low NK cells, similar to IL4-NK cells, were increased in the MLN 10 d after the infection (Fig. 6 A and B). Moreover, these IL4-NK–like cells were not induced by parasite infection in IL-4Rα^−/− mice (Fig. 6 A and B). IL4-NK–like cells also expressed a lower level of CD11b and higher levels of IL-4Rα and -21Rα than cNK-like cells (CD45⁺NKp46⁺B220^lowCD3e⁻CD19⁻) (Fig. 6C). These results suggest that IL4-NK–like cells are also induced in an IL-4–dependent physiological condition. Previous studies reported that IL-4–expressing cells were increased in the lung from day 5 after the infection of Nb and, consequently, Th2 immune responses were induced (29, 30). Therefore, we next analyzed NK cells in the lung and found that IL4-NK–like cells were also increased at days 9 after infection with Nb, whereas the number of total NK cells was not significantly changed (Fig. 6D). To investigate the role of NK cells in the Nb-infected lung, we depleted NK cells by administration of anti-asialo GM1 Abs (Fig. S8A). We then found that the expression levels of Eotaxin-1/CCL11 and KC/CXCL1 in the lung were significantly increased when NK cells were depleted (Fig. 6E). Eotaxin-1/CCL11 promotes chemotaxis of eosinophils, which produce IL-4 and contribute to the induction of Th2 immune responses. Conversely, KC/CXCL1 promotes chemotaxis of neutrophils, which produce IL-13 and contribute to the induction of Th2 immune responses in the Nb infection model (31). The expression levels of Eotaxin-1/CCL11 and KC/CXCL1 in the lung were not significantly changed in steady state by depletion of NK cells (Fig. S8B). Therefore, these results suggest that NK cells in the lung infected with Nb might keep the balance of Th1/Th2 immune responses.

Fig. 6.

IL4-NK–like cells in parasitic infection. BALB/c WT mice or IL-4Rα^−/− mice were infected by s.c. injection of Nb. (A–C) At 10 d later, hematopoietic cells were harvested from the MLN, stained for CD45, CD3e, CD19, NKp46, B220, and IL-18Rα, and analyzed by flow cytometry. (A) Numbers indicate percentages. Data shown are representative of five or six independent experiments. (B) The graphs show the percentages of B220^highIL-18Rα^low NK cells among total NK cells and absolute B220^highIL-18Rα^low NK cell numbers, and bars indicate the means of five or six mice. (C) Cells were stained for CD11b, IL-4Rα, and -21Rα. cNK-like and IL4-NK–like cells were gated on CD45⁺NKp46⁺B220^lowCD3e⁻CD19⁻ and CD45⁺NKp46⁺B220^highIL-18Rα^lowCD3e⁻CD19⁻, respectively. The histogram plots show surface expression profiles of cNK-like cells (shaded) and IL4-NK–like cells (solid lines). Data shown are representative of four independent experiments. (D) At 9 d after the infection, hematopoietic cells were harvested from the lung and analyzed by flow cytometry. The graphs show the numbers of total NK cells (NKp46⁺CD3e⁻CD19⁻FCεR1α⁻) (Left) and IL4-NK–like cells (NKp46⁺B220^highIL-18Rα^lowCD3e⁻CD19⁻FCεR1α⁻) (Right) per lung weight. Bars indicate the means of three or five mice. UI, uninfection. (E) BALB/c WT mice were intraperitoneally injected with PBS or anti-asialo GM1 Abs at days −1, 0, and 5 after infection with Nb. The lungs were harvested at day 9 after infection. Expression levels of Eotaxin-1/CCL11 mRNA and KC/CXCL1 mRNA in the lung were determined by quantitative RT-PCR. Bars indicate the means of five or six mice. Data shown are representative of two independent experiments. *P < 0.05; **P < 0.01. N.S., not significant.

Fig. S8.

Representative data from flow cytometric analysis and gene expressions in the lung injected with anti-asialo GM1 Abs. (A) BALB/c WT mice were intraperitoneally injected with PBS or anti-asialo GM1 Abs at days −1, 0, and 5 after infection with Nb. The lungs were harvested at day 9 after infection. The graphs show the numbers of total NK cells (Left) and IL4-NK–like cells (Right) per lung weight. Bars indicate the means of four or five mice. (B) BALB/c WT mice were intraperitoneally injected with PBS or anti-asialo GM1 Abs, and the lungs were harvested at day 4 after the injection. Expression levels of Eotaxin-1/CCL11 mRNA (Left) and KC/CXCL1 mRNA (Right) in the lung were determined by quantitative RT-PCR. Bars indicate the means of five or six mice. Data shown are representative of two independent experiments. *P < 0.05; **P < 0.01; N.S., not significant.

Discussion

In the present study, we showed that distinctive characteristics of NK cells (IL4-NK cells) were induced in the mice overexpressing IL-4. The cell surface markers on these IL4-NK cells are significantly different from cNK cells and immature CD11b⁻ NK cells. CD11b is a well-known maturation marker on NK cells and is highly expressed on mature NK cells (22, 32). Thus, it might be assumed that IL4-NK cells are immature NK cells. However, IL4-NK cells highly express CD49b, which is expressed on mature NK cells, and do not express TRAIL and IL-7Rα, which are expressed on immature NK cells (16, 33, 34). In addition, the morphology of IL4-NK cells and their capacity to produce granzyme B were quite different from cNK and immature NK cells. Moreover, mature cNK cells were converted to IL4-NK cells in the mice overexpressing IL-4. These results indicate that IL4-NK cells may not simply represent an immature stage of NK cells. Interestingly, most of the NK cells in the mice overexpressing IL-4 are IL4-NK cells and immature CD11b⁻ NK cells are also converted to IL4-NK cells, suggesting that most of NK-lineage cells could become IL4-NK cells. The percentage of donor cells derived from immature CD11b⁻ NK cells 3 d after transplantation was higher than that from cNK cells in the mice overexpressing IL-4. However, because the number of immature CD11b⁻ NK cells is lower than that of cNK cells in the steady state, both cNK and immature CD11b⁻ NK cells may equally contribute to the development of IL4-NK cells. Of note, the present study demonstrates that these unique NK cells are induced and increased in vivo by IL-4.

Because cNK cells are neither increased nor converted to IL4-NK cells when cultured with IL-4 alone, additional signals are necessary for the development of IL4-NK cells. In this study, we found that liver macrophages were remarkably increased in the mice overexpressing IL-4 and that these IL-4–stimulated macrophages contributed to NK-cell proliferation via IL-15. The present study links NK-cell proliferation and IL-4–stimulated macrophages.

The capacity of cytokine production and cytotoxicity is also different between cNK and IL4-NK cells. Interestingly, IL4-NK cells secrete a much higher level of IFN-γ than cNK cells when stimulated with IL-12, an initiator of Th1 responses (35). IL4-NK cells also secrete IL-10, which inhibits not only Th1 immune response, but also Th2 immune response (36, 37). Collectively, these results suggest that IL4-NK cells might be a balancer or regulatory cells of Th1/Th2 immune responses. Notably, GM-CSF is secreted from IL4-NK cells without stimuli such as IL-12, -21, and the signal from activating receptors.

The phenotype of cNK cells is changed to that of IL4-NK cells in culture with IL-4 and -15. Down-regulation of IL-18Rα is a direct effect of IL-4 on cNK cells, whereas changes of other surface markers may be caused by IL-15. However, the induction of IL4-NK cells in vivo is highly dependent on the IL-4 signal. The change in fluorescence intensity of CFSE revealed that proliferation of CFSE-labeled NK cells isolated from IL-4Rα KO mice was lower than that of NK cells isolated from WT mice. These results suggest that the IL-4 signal in NK cells might affect the IL-15 signal in NK cells by unknown mechanisms. Also IL-4 directly induces the asecretion of GM-CSF from NK cells. Collectively, it is suggested that conversion of cNK to IL4-NK cells requires both direct and indirect effect of IL-4. As described, GM-CSF plus IL-4 promotes differentiation of DC to enhance the production of the bioactive IL-12 heterodimer, an activator of NK cells (38). IL4-NK cells might interact with DCs via GM-CSF. However, further studies are necessary to uncover the interaction between NK cells and DCs.

Finally, to confirm the induction of IL4-NK–like cells in physiological conditions, we used a typical model of Th2 response, Nb infection, and found that IL4-NK–like cells were induced. Moreover, these IL4-NK–like cells were not induced in IL-4Rα^−/− mice. Because IL4-NK cells were not induced by IL-13, the induction of IL4-NK–like cells is IL-4–dependent. It is known that IL-4–expressing cells are increased in the lung from day 5 after the infection of Nb, which results in the induction of Th2 immune responses (29, 30). In the Nb infection model, the roles of NK cells remain largely unknown. Therefore, we investigated the effect of depletion of NK cells on gene expressions in the lung where Th2 immune responses were induced by Nb and found a significant increase of the expression of Eotaxin-1/CCL11 and KC/CXCL1 in the lung by depletion of NK cells, suggesting that NK cells might regulate migration of eosinophils and neutrophils. Eosinophils are well known to contribute to the induction of Th2 immune responses, and neutrophils also contribute to the induction of Th2 immune responses in the Nb infection model (31). Considering these results and the strong capacity of IL4-NK cells to produce cytokines, the roles of IL4-NK–like cells in the Nb infection model might keep the balance of Th1/Th2 immune responses. However, because administration of anti-asialo GM1 Abs depletes total NK cells, selective depletion of IL4-NK–like cells is necessary to address the roles of IL4-NK–like cells in the Nb infection model.

In the present study, we described a novel function of IL-4 in mice. IL-4 overexpression induces distinctive characteristics of NK cells, indicating the importance of the effect of IL-4 on NK cells in vivo. We further revealed that IL-4 affects NK cells directly and also indirectly through macrophages. Although the effect of IL-4 on NK cells is controversial among previous reports, it might be explained by whether the indirect effect is considered in each study. Because IL-4 has multiple functions on various hematopoietic cells, the indirect effect should be considered more carefully. Collectively, this study suggests a novel role of IL-4 in immune responses through the induction of these unique NK cells.

Materials and Methods

WT C57BL/6J and BALB/c mice were purchased from CLEA Japan, Inc. CD45.1 congenic C57BL/6 mice (RBRC 00144) were provided by RIKEN BioResource Center through the National Bio-Resource Project of the Ministry of Education, Culture, Sports, Science and Technology, Japan. IL-4Rα^−/− mice (39) were provided by Frank Brombacher, University of Cape Town, Cape Town, South Africa, and Masato Kubo, Tokyo University of Science/RIKEN Center for Integrative Medical Sciences, Tokyo. All mice were maintained under standard specific pathogen-free conditions and were used between 6 and 12 wk of age. The animal studies were performed according to the guidelines set by the Institutional Animal Care and Use Committee of the University of Tokyo. Materials and methods are described in greater detail in SI Materials and Methods.

SI Materials and Methods

Statistical Analysis.

Unless otherwise stated, data are shown as means ± SEM and were compared by using Welch’s t test. A value of P < 0.05 was taken to indicate statistical significance. Statistical analyses were performed by using GraphPad Prism (GraphPad Software).

Overexpression by Hydrodynamic Tail Vein Injection Method.

Mouse IL-4 and -13 cDNA was amplified by PCR with the following primers: 5′-cacagagctattgatgggtc-3′ and 5′-ctacgagtaatccatttgcatg-3′ for IL-4 and 5′-ctctgggcttcatggcgctc-3′ and 5′-ctcattagaaggggccgtgg-3′ for IL-13, from the mix of bone marrow-cDNA and thymus-cDNA derived from C57BL/6J mice. cDNA fragment of IL-4 and -13 were inserted into the BamHI and XhoI sites of a pLIVE vector (Mirus). pLIVE-IL-4 (5 μg) or pLIVE-IL-13 (20 μg) was diluted with TransIT-EE Hydrodynamic Delivery Solution (Mirus) and injected intravenously into CD45.1 congenic mice. pLIVE-SEAP (secreted alkaline phosphatase) (5 or 20 μg) was used as a control.

Abs and other reagents.

FITC-conjugated anti-B220 and CD45.2; phycoerythrin (PE)-conjugated anti-B220 and CD3e; PE/Cy7-conjugated anti-CD45; biotin-conjugated anti-Thy1.1 and IL-7Rα; purified anti-NK1.1; PE-, allophycocyanin (APC)-, and PE/Cy7-conjugated streptavidin were purchased from BD Pharmingen. FTTC-conjugated CD11b; PE-conjugated anti-NK1.1, FCεR1α, 2B4, KLRG1, NKG2A, IL-4Rα, IL-21Rα, and CD49b; PE/Cy7-conjugated anti-Ly6C; APC/Cy7-conjugated anti-CD11b; biotin-conjugated anti-CD49b, FCεR1α, Trail, NKp46, NKG2D, and F4/80; APC-conjugated CD49b and IL-21Rα; Alexa Fluor 647-conjugated IL-18Rα; BV421-conjugated anti-CD3e and CD19; purified anti-CD68 were purchased from BioLegend. Purified anti-granzyme B and PE-conjugated anti-granzyme B were purchased from eBioscience. For IL-15 neutralization assay, LEAF-purified Rat IgG1κ isotype Ab and Functional Grade purified mouse IL-15/IL-15Rα complex Ab were purchased from BioLegend and eBioscience, respectively. Recombinant murine IL-2 (PROSPEC), IL-4 (Recentec), IL-12 (PeproTech), IL-13 (BioLegend), IL-15 (eBioscience), and IL-21 (BioLegend) were purchased.

Cell isolation and flow cytometry.

A single-cell suspension from the liver was obtained by a modified collagenase perfusion method as described (15). In brief, liver specimens were perfused with a basic perfusion solution containing 0.5 g/L collagenase type IV (Sigma-Aldrich) and 50 mg/L DNase I (Sigma-Aldrich). The digested liver was passed through a 70-μm cell strainer. After centrifugation twice at 35 × g for 1 min, the precipitated cells were used as hepatocytes after Percoll (GE Healthcare) density centrifugation. The supernatant was transferred to a new tube and centrifuged at 35 × g for 2 min repeatedly until no pellet was visible. The supernatant was centrifuged at 450 × g for 5 min, and the precipitated cells (nonparenchymal cells) were resuspended in 33% (vol/vol) Percoll and centrifuged at 500 × g for 15 min at 20 °C for purification of lymphocytes. To obtain single-cell suspension from the spleen or MLNs, the tissue was triturated by a plunger on a 70-μm cell strainer and passed with PBS containing 2% (vol/vol) FBS. To obtain single-cell suspension from the lung, the tissue was minced and stirred with a basic perfusion solution containing 0.5 g/L collagenase type IV and 50 mg/L DNase I for 20 min at 37 °C. The tissue was then passed through a 70-μm cell strainer. Aliquots of cells were RBC-depleted and blocked with anti-FcγR Ab, costained with fluorescence- and/or biotin-conjugated Abs, and then incubated with allophycocyanin-conjugated streptavidin (BD Biosciences) if needed. The stained cells were analyzed by using a FACS Canto II (BD Biosciences) and the FlowJo software (Version 8.8.7; TreeStar), or sorted by Moflo XDP (Beckman-Coulter) or autoMACS pro (Miltenyi Biotec) with antiallophycocyanin microbeads or streptavidin microbeads. Dead cells were excluded by propidium iodide staining. For isolation of liver macrophages, liver specimens were perfused with a basic perfusion solution containing 0.25 g/L collagenase-Yakult (Yakult Pharmaceutical Industry Co.) and 50 mg/L DNase I. Then, nonparenchymal cells were stained with biotin-conjugated anti-F4/80 and APC-conjugated streptavidin. Thereafter, the cells were sorted by autoMACS pro. The percentages of F4/80-positive cells were >90%.

Quantitative RT-PCR.

Total RNA was extracted from hepatocytes or lungs with TRIZOL reagents (Invitrogen) and reverse-transcribed with PrimeScript RT Master Mix (Takara Bio Ink). Quantitative real-time RT-PCR was performed on a LightCycler 96 (Roche Applied Science) by using SYBR premix Ex TaqII reagent (Takara Bio Ink). HPRT was used as an internal control. The sequence of primers was as follows: 5′-agttgtcatcctgctcttctttctc-3′ and 5′-atggcgtcccttctcctgt-3′ for IL-4, 5′-agagctccacagcgcttct-3′ and 5′-gcaggaagttgggatgga-3′ for Eotaxin/CCL11, 5′-gactccagccacactccaac-3′ and 5′-tgacagcgcagctcattg-3′ for KC/CXCL1, and 5′-tgacactggtaaaacaatgc-3′ and 5′-tatccaacacttcgagaggt-3′ for HPRT. The expression levels were normalized to HPRT.

ELISA.

At 5 d after the injection with control vector, pLIVE-IL-4 vector (5 µg), or pLIVE-IL-13 vector (20 µg), sera were harvested from these C57BL/6 mice and analyzed by using an ELISA kit for IL-4 (BD Biosciences) or IL-13 (eBioscience). Sorted NK cells were cultured in the presence of IL-2 (100 ng/mL), IL-15 (100 ng/mL), IL-15 (100 ng/mL) with or without IL-4 (200 ng/mL), or IL-15 (100 ng/mL) with or without IL-13 (200 ng/mL). After 24 h of incubation, cell-free supernatants were analyzed by using an ELISA kit for GM-CSF (BD Biosciences) or granzyme B (R&D Systems). Then, sorted NK cells were stimulated with IL-12 (100 ng/mL), IL-18 (100 ng/mL), IL-21 (100 ng/mL), or anti-NK1.1 Abs in the presence of IL-2 (100 ng/mL). The 96-well plates were coated overnight with 20 μg/mL purified isotype control or anti-NK1.1 Abs and then washed by PBS. After 24 h of incubation, cell-free supernatants were analyzed by using ELISA kits for IFN-γ (BD Biosciences) or IL-10 (BD Biosciences). ELISAs were performed in triplicate.

May–Grünwald–Giemsa staining.

Sorted NK cells ware prepared with the cytospin. These cells were stained with May–Grünwald, followed by 5% Giemsa stain in phosphate buffer (pH 6.4). Images were obtained with a fluorescence microscope (Axio Observer.Z1; Zeiss).

EdU assay.

At 4 d after the injection with control or pLIVE-IL-4 vector (5 µg), these C57BL/6 mice were intraperitoneally injected with 1 mg of EdU. At 2 h later, cells were harvested from the liver and spleen, and cNK, immature CD11b⁻ NK, or IL4-NK cells were gated on CD45⁺NK1.1⁺CD11b⁺CD3e⁻CD19⁻, CD45⁺NK1.1⁺CD11b⁻CD3e⁻CD19⁻, or CD45⁺NK1.1⁺B220⁺CD3e⁻CD19⁻, respectively. Cells were then stained for intracellular incorporation of EdU (Click-iT Plus EdU Alexa Fluor 647 Flow Cytometry Assay Kit: Molecular Probes) and analyzed by FACS Canto II.

Adoptive cell transfer.

Control or pLIVE-IL-4 vector (5 µg) was injected intravenously into CD45.1 congenic or BALB/c WT mice. At 24 h later, sorted NK cells (4 × 10⁵ to 1 × 10⁶ cells) from the spleens of C57/BL6 mice or splenocytes (2 × 10⁷ cells) derived from BALB/c WT mice or IL-4Rα^−/− mice were injected intravenously into these nonirradiated CD45.1 congenic mice or BALB/c WT mice, respectively. In some experiments, donor cells were labeled with CFSE (5 µM) (eBioscience).

Intracellular staining of granzyme B.

Cells were cultured in the presence of IL-2 (100 ng/mL) with Brefeldin A solution (5 μg/mL) (Biolegend) for 5 h. After incubation, cells were stained for surface markers and intracellular granzyme B by using PE-conjugated anti-granzyme B Ab and the BD Cytofix/Cytoperm Fixation/Permeabilization Solution Kit (BD). The stained cells were analyzed by using FACS Canto II and the FlowJo software (v8.8.7).

Immunohistochemistry.

Frozen sections (8 μm) of the liver were prepared by using a HM525 cryostat (Microm International), placed on amino silane-coated glass slides (Matsunami Glass), and fixed with acetone (Wako). After blocking with 5% skim milk/PBS, the samples were incubated with primary Abs and then with fluorescence-conjugated secondary Abs. Nuclei were counterstained with Hoechst 33342 (Sigma). Images were obtained with a fluorescence microscope (Axio Observer.Z1; Zeiss).

Depletion of macrophages or NK cells.

C57BL/6 mice were intraperitoneally injected with 30 μL of control or clodronate liposomes (KATAYAMA Chemical Industries) for depletion of macrophages. BALB/c mice were intraperitoneally injected with 30 μL of anti-asialo GM1 Abs (Wako) for depletion of NK cells.

Measurement of serum alanine aminotransferase.

Serum alanine aminotransferase (ALT) was measured at the Oriental Yeast Company.

Cell culture.

cNK cells (CD49b⁺CD11b⁺CD3e⁻FCεR1α⁻) were isolated from the spleens of C57/BL6 mice using a MoFlo cell sorter. Purified cNK cells were cultured in RPMI 1640 medium containing 10% (vol/vol) FBS, 2-mercaptoethanol (50 µM), Hepes (20 mM), nonessential amino acids, sodium pyruvate, l-Gln, and gentamycin in the presence of recombinant mouse IL-15 (100 ng/mL) with or without IL-4 (200 ng/mL) or IL-15 (100 ng/mL) and with or without IL-13 (200 ng/mL) for 4 d. Cell viability was measured in triplicate cultures with IL-4 (50 ng/mL) by a CellTiter-Glo kit (Promega).

Coculture of NK cells and macrophages.

At 4 d after the injection with pLIVE-IL-4 vector (5 µg), macrophages were sorted from the liver of these C57BL/6 mice, and 1 × 10⁶ cells were cultured on a BioCoat Collagen I plate (Corning). At 4 h later, the supernatant was removed gently, and sorted NK cells (1 × 10⁶ cells) from the spleens of C57/BL6 mice were seeded on these macrophages with isotype control or anti–IL-15/IL-15Rα complex Ab. At 48 h later, cultured cells were harvested by using Cell Dissociation Buffer (Gibco), and their viability was measured by NK1.1⁺/PI⁻ NK cells by using FACS Canto II.

Cytotoxic assay.

YAC-1 cells (RCB1165; provided by RIKEN BioResource Center through the National Bio-Resource Project of the Ministry of Education, Culture, Sports, Science, and Technology, Japan) were labeled with CFSE (5 µM) (eBioscience). CFSE-labeled YAC-1 cells (1 × 10⁴) were cocultured with sorted NK cells for 5 h at various effector-to-target ratios. Cytotoxicity was measured by CFSE⁺/PI⁺ YAC-1 cells using FACS Canto II.

Parasitic infection.

Mice were inoculated s.c. with 750 viable third-stage Nb larvae in 500 μL of PBS. Mice were killed 9 or 10 d after infection, and hematopoietic cells were harvested from the lungs or MLNs, respectively.