Visual Abstract

NK cells can recognize target cells such as virus-infected and tumor cells through integration of activation and inhibitory receptors. Recognition by NK cells can lead to direct lysis of the target cell and production of the signature cytokine IFN-γ. However, it is unclear whether stimulation through activation receptors alone is sufficient for IFN-γ production. In this study, we show that NK activation receptor engagement requires additional signals for optimal IFN-γ production, which could be provided by IFN-β or IL-12. Stimulation of murine NK cells with soluble Abs directed against NK1.1, Ly49H, Ly49D, or NKp46 required additional stimulation with cytokines, indicating that a range of activation receptors with distinct adaptor molecules require additional stimulation for IFN-γ production. The requirement for multiple signals extends to stimulation with primary m157-transgenic target cells, which triggers the activation receptor Ly49H, suggesting that NK cells do require multiple signals for IFN-γ production in the context of target cell recognition. Using quantitative PCR and RNA flow cytometry, we found that cytokines, not activating ligands, act on NK cells to express Ifng transcripts. Ly49H engagement is required for IFN-γ translational initiation. Results using inhibitors suggest that the proteasome–ubiquitin–IKK–TPL2–MNK1 axis was required during activation receptor engagement. Thus, this study indicates that activation receptor–dependent IFN-γ production is regulated on the transcriptional and translational levels.

This article is featured in In This Issue, p.1681

Natural killer cells recognize and attack target cells, including cancer and pathogen-infected cells, through a combination of activation and inhibitory receptor–ligand interactions. Upon recognition of a target cell through such interactions, NK cells can directly induce lysis of the target, but they also produce the signature cytokine IFN-γ. Activation receptor–dependent IFN-γ production is frequently studied to assess NK cell functionality (1). NK cells can produce IFN-γ in response to cytokines, as well; in particular, IL-12 in combination with IL-18 results in strong IFN-γ production (2). However, whether these pathways intersect is unclear.

Production of IFN-γ by NK cells has been shown to contribute to viral control and tumor rejection. For example, NK cells are the main source of IFN-γ during early stages of murine CMV (MCMV) infection (3). This IFN-γ produced early during infection contributes to MCMV clearance, particularly in the liver (4). A susceptibility locus on mouse chromosome 10 is associated with impaired MCMV control and decreased NK IFN-γ production, whereas IFN-γ produced by T cells is unaffected (5), providing genetic evidence suggesting NK cell–produced IFN-γ is critical for viral control. IFN-γ production during MCMV infection requires IL-12 and depends on STAT4 (3, 6). In addition, IL-18 synergizes with IL-12 to induce IFN-γ during infection (7). Thus, in the context of MCMV infection, the role for cytokines inducing NK cell IFN-γ is well established. NK cell IFN-γ production has been shown to control metastasis formation of B16 melanoma subline (8), implying a role for NK cell IFN-γ in controlling tumors, as well.

It is well established that ligation of activation receptors triggers NK cells to produce IFN-γ, but there is a body of evidence suggesting that stimulation through an activation receptor alone is insufficient for optimal IFN-γ production. Stimulation of mouse NK cells with plate-bound Abs against activation receptors such as NK1.1 or Ly49H triggers IFN-γ production (911). In contrast, stimulation with soluble Abs does not induce IFN-γ, whereas soluble anti-Ly49D has been reported to induce phosphorylation of SLP76 and ERK (12). This indicates that soluble Ab is capable of inducing NK cell activation but not IFN-γ production. Plate-bound anti-NKG2D–dependent NK cell GM-CSF production requires signaling through CD16 (13), suggesting that plate-bound Ab may also trigger Fc receptors. Moreover, Abs against different receptors synergize for human NK cell IFN-γ and TNF-α production when coated on the same beads (14), and a combination of activation receptor ligands and adhesion molecules are required on insect target cells to induce IFN-γ by freshly isolated human NK cells (15). Overexpression of activation ligands on certain cell lines induces IFN-γ by resting mouse NK cells, including overexpression of m157 and NKG2D ligands (5, 16, 17). In addition, NK cells stimulated with MCMV-infected macrophages produce IFN-γ in a Ly49H-dependent manner (17). However, transfer of wild-type (WT) NK cells into a naive host constitutively expressing the Ly49H ligand m157 as a transgene (m157-transgenic [m157-Tg]) did not result in IFN-γ production but rather caused NK cell hyporesponsiveness within 24 hours (18, 19), indicating that additional signals may be required for activation receptor–dependent IFN-γ production.

In this study, we show that activation receptor–mediated IFN-γ production by NK cells indeed requires additional signals, which can be provided by cytokines such as IL-12 and IFN-β. We found that cytokine signaling induces transcription of Ifng mRNA, whereas Ly49H signaling resulted in translation of Ifng mRNA. Furthermore, efficient IFN-γ production required a specific order of these stimuli. Taken together, this study provides a molecular basis for the requirements of NK activation receptor–induced IFN-γ production.

C57BL/6 (and congenic CD45.1 C57BL/6) mice were purchased from Charles River Laboratories. IL-12Rβ2−/− and RAG1−/− mice were purchased from The Jackson Laboratory. DAP12 knock-in mice were kindly provided by E. Vivier (CNRS-INSERM-Universite de la Mediterranee, Marseille, France). DAP10−/− mice were kindly provided by M. Colonna. IFNAR1−/− mice were backcrossed on C57BL/6 background as described previously (20). m157-Tg and Ly49H-deficient B6.BxD8 mice were generated and maintained in-house in accordance with institutional ethical guidelines.

Fluorescent-labeled Abs used were anti-NK1.1 (clone PK136), anti-NKp46 (29A1.4), anti-CD3 (145-2C11), anti-CD19 (eBio1D3), anti-CD45.1 (A20), anti-CD45.2 (104), anti-Eomes (Dan11mag), anti–TNF-α (MP6-XT22), and anti–IFN-γ (XMG1.2), all from Thermo Fisher Scientific. Fluorescent-labeled anti-CD49a (Ha31/8) was purchased from BD Biosciences. Biotinylated anti-Ly49H (3D10) was produced in-house. MG-132 (2094S; Cell Signaling Technology), actinomycin D (AD; A9415; Sigma), cycloheximide (CHX; C7698; Sigma), CGP-57380 (C0993; Sigma), ISRIB (SML0843; Sigma), Pyr-41 (N2915; Sigma), BMS-345541 (B9935; Sigma), Tpl2 kinase inhibitor (sc-204351; Santa Cruz Biotechnology), and PD90859 (P215; Sigma) were used at the indicated concentrations. Where indicated, NK cells were enriched to a purity >90% NK1.1+CD3CD19 cells, verified by flow cytometry for each experiment, from RAG1−/− splenocytes using CD49b microbeads with LS columns (Miltenyi Biotec) or from C57BL/6 mice using the EasySep Mouse NK cell isolation kit (STEMCELL Technologies) according to the manufacturer’s instructions. m157-Tg or WT murine embryonic fibroblasts (MEF) were isolated from day 11.5–13.5 embryos.

Freshly isolated splenocytes and liver single-cell suspensions were prepared as previously described (21). For Ab stimulations, splenocytes were stimulated with 4 μg/ml precoated or soluble Ab against NK1.1 (PK136), Ly49H (3D10), and Ly49D (4E4) that were produced in-house or NKp46 (29A1.4) purchased from BioLegend. IFN-β (800 U/ml; PBL Assay Science) and/or IL-12p70 (25 ng/ml; PeproTech) were added where indicated. For coculture experiments, splenocytes were cocultured with congenic CD45.1/2 or CFSE-labeled m157-Tg or littermate control splenocytes at a 1:1 ratio in the presence of 100–200 U/ml IFN-β or 10–20 ng/ml IL-12p70 where indicated. For experiments with MEF, splenocytes or purified NK cells were cocultured with m157-Tg or C57BL/6 MEF at a 30:1 or a 1:1 ratio, respectively. Monensin (Thermo Fisher Scientific) was added after 1 h, and the culture was stopped 5–8 h thereafter. Samples were subsequently stained with fixable viability dye (Thermo Fisher Scientific), followed by surface staining in 2.4G2 hybridoma supernatant with directly conjugated or biotinylated Abs that were revealed by streptavidin–PE (BD Biosciences). Samples were then fixed and stained intracellularly using the Cytofix/Cytoperm kit (BD Biosciences) or the FOXP3/transcription factor staining buffer set (Thermo Fisher Scientific) for Eomes staining according to the manufacturer’s instructions. Samples were acquired using FACSCanto (BD Biosciences) and analyzed using FlowJo software (Tree Star). Where indicated, representative sample files were combined after acquisition in concatenated files for visualization purposes. Concatenated files were gated for NK cells; sample versus IFN-γ is shown in the generated composite plot. NK cells were defined as ViabilityNK1.1+CD3CD19 or ViabilityNKp46+CD3CD19, and gating for IFN-γ was based on unstimulated controls. For RNA flow experiments, splenocytes were stimulated with MEF, and after Ab staining, they were hybridized with Ifng or control scrambled probe using the PrimeFlow RNA Assay Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions.

RNA was isolated from purified NK cells after the indicated stimulation using TRIzol (Thermo Fisher Scientific). Contaminating DNA was removed using TURBO DNase, and cDNA was synthesized using Superscript III (Thermo Fisher Scientific). Quantification was performed for Ifng (Mm.PT.58.30096391; Integrated DNA Technologies) and Gapdh (Mm99999915_g1; Thermo Fisher Scientific) against plasmid standard curves using TaqMan Universal Master Mix II on a StepOnePlus real-time PCR system (Thermo Fisher Scientific).

Stimulated, purified NK cells were lysed in RIPA lysis and extraction buffer in the presence of protease and phosphatase inhibitors (Thermo Fisher Scientific). Lysates were denatured in Laemmli sample buffer and resolved by SDS-PAGE. Proteins were transferred to PVDF membranes and probed with Abs specific for Erk1/2 (137F5), phospho-Erk1/2 (Thr202/Tyr204; no. 9102), NF-κB p65 (D14E12), phospho–NF-κB p65 (Ser536; 93H1), or β-actin (no. 4967), which were purchased from Cell Signaling Technology.

Statistically significant differences were determined with unpaired, two-tailed Student t tests using Prism (GraphPad). Error bars in figures represent SD and p values from duplicate or triplicate samples as indicated. All experiments shown are representative of two or more independent experiments.

NK cells produce IFN-γ in response to stimulation with immobilized plate-bound Abs against activation receptors, such as anti-NK1.1, anti-Ly49H, anti-Ly49D, and anti-NKp46, as shown in composite plots of concatenated representative samples (Fig. 1). However, stimulation with soluble Ab alone did not result in significant IFN-γ production, suggesting that additional signals are required for IFN-γ production. We investigated a potential role for cytokines in activation receptor–mediated IFN-γ production, which we and others have previously described to be important for NK cell cytolytic functions during viral infections (6, 20, 22). Stimulation of WT NK cells alone with IL-12 and/or IFN-β resulted in low percentages of NK cells producing IFN-γ (Fig. 1). Strikingly, stimulation of NK cells with soluble anti-NK1.1 in combination with IL-12 and IFN-β synergized to induce strong IFN-γ production, reaching levels similar to plate-bound Ab (Fig. 1A, 1C). To investigate whether the requirement for multiple signals was applicable to other activation receptors that signal through adaptor molecules other than those associated with NK1.1 (FcRγ), we analyzed the response to anti-Ly49H (DAP10/12), anti-Ly49D (DAP10/12), and anti-NKp46 (FcRγ/CD3ζ). Soluble anti-Ly49H, anti-Ly49D, and anti-NKp46 also induced optimal IFN-γ production in the presence of IL-12 and IFN-β, suggesting that these receptors have similar requirements. In summary, these results indicate that a wide range of NK activation receptors, signaling through distinct adapter molecules, require synergistic stimulation with cytokines for full IFN-γ production.

FIGURE 1.

Soluble Abs directed against activation receptors are capable of inducing NK cell IFN-γ in the presence of IFN-β or IL-12. Splenocytes from WT mice were stimulated in the presence of soluble or plate-bound Ab and the indicated cytokines. NKp46+ NK cells (A and B) or NK1.1+ NK cells (C and D) were analyzed for IFN-γ production. Combined FACS plots of concatenated representative samples are shown in (A) and (C), and quantification of duplicates is shown in (B) and (D). A representative experiment of three independent experiments is shown. The asterisk (*) symbols indicate statistical comparison with same condition without Ab; the number (#) symbols indicate comparison with soluble Ab only within the same group. ####p < 0.0001, ###/***p < 0.001, ##/**p < 0.01, #/*p < 0.05. ns, not significant.

FIGURE 1.

Soluble Abs directed against activation receptors are capable of inducing NK cell IFN-γ in the presence of IFN-β or IL-12. Splenocytes from WT mice were stimulated in the presence of soluble or plate-bound Ab and the indicated cytokines. NKp46+ NK cells (A and B) or NK1.1+ NK cells (C and D) were analyzed for IFN-γ production. Combined FACS plots of concatenated representative samples are shown in (A) and (C), and quantification of duplicates is shown in (B) and (D). A representative experiment of three independent experiments is shown. The asterisk (*) symbols indicate statistical comparison with same condition without Ab; the number (#) symbols indicate comparison with soluble Ab only within the same group. ####p < 0.0001, ###/***p < 0.001, ##/**p < 0.01, #/*p < 0.05. ns, not significant.

Close modal

To explore whether the synergy between activation receptor and cytokine stimulation extends to innate lymphoid cells, we analyzed liver ILC1 for IFN-γ and TNF-α production in response to anti-NK1.1 under the conditions described above (Supplemental Fig. 1). ILC1 produced high levels of IFN-γ and TNF-α in response to plate-bound anti-NK1.1. In contrast to conventional NK cells, ILC1 produced moderate amounts of IFN-γ in response to IL-12 alone, which was elevated to some extent by the addition of soluble anti-NK1.1. Taken together, ILC1 cytokine production likely requires distinct mechanisms as compared with NK cells.

To investigate the requirements for activation receptor–dependent IFN-γ production in the context of activation receptor engagement by ligand rather than Ab cross-linking, we turned to an in vitro model that we previously published (20). Briefly, we stimulated NK cells among freshly isolated splenocytes with freshly isolated m157-Tg or WT splenocytes in the presence or absence of cytokines. Whereas Ly49H+ NK cells produce IFN-γ in response to MCMV-infected cells upon recognition of m157 (17, 18), in this study we observed that Ly49H+ NK cells produced only low levels of IFN-γ in response to m157-Tg splenocytes alone in vitro (Fig. 2A). These data suggest that, in the context of activation receptor engagement by ligand, additional signals are required, as well. Indeed, coculture of NK cells and m157-Tg splenocytes with IFN-β or IL-12 induced substantial levels of IFN-γ, whereas stimulation of WT NK cells alone with IFN-β or IL-12 did not (Fig. 2A).

FIGURE 2.

Ly49H-dependent IFN-γ production requires cell-intrinsic cytokine and DAP12 but not DAP10 signaling. (A) Splenocytes from CD45.2 WT mice were mixed with CD45.1/2 congenic m157-Tg or WT splenocytes at a 1:1 ratio in the presence of the indicated cytokines. IFN-γ production by the CD45.2+ NK cells is shown. (BE) CD45.2+ splenocytes deficient for the indicated genes were mixed with congenic CD45.1/2+ m157-Tg or control splenocytes that were not deficient for the indicated genes. A representative experiment of 17 (A), 3 (B), and 2 (C) independent experiments performed in duplicate is shown. ***p < 0.001, **p < 0.01, *p < 0.05. ns, not significant.

FIGURE 2.

Ly49H-dependent IFN-γ production requires cell-intrinsic cytokine and DAP12 but not DAP10 signaling. (A) Splenocytes from CD45.2 WT mice were mixed with CD45.1/2 congenic m157-Tg or WT splenocytes at a 1:1 ratio in the presence of the indicated cytokines. IFN-γ production by the CD45.2+ NK cells is shown. (BE) CD45.2+ splenocytes deficient for the indicated genes were mixed with congenic CD45.1/2+ m157-Tg or control splenocytes that were not deficient for the indicated genes. A representative experiment of 17 (A), 3 (B), and 2 (C) independent experiments performed in duplicate is shown. ***p < 0.001, **p < 0.01, *p < 0.05. ns, not significant.

Close modal

To explore whether this cytokine requirement applied to other activation receptors such as NKG2D, we analyzed NK cell IFN-γ production in coculture experiments with SV40-immortalized MEF expressing the NKG2D ligand, Rae1δ (Supplemental Fig. 2A). Similar to m157-Tg stimulation, NK cells produced more IFN-γ upon exposure to Rae1δ-expressing target cells with IFN-β or IL-12. These results strongly suggest that activation receptor engagement by ligand alone is insufficient for optimal IFN-γ production and additional signals such as cytokines are required.

To determine if cytokine stimulation acted directly on NK cells, we used mixed cultures involving NK cells deficient in cytokine receptors. When IFNAR-deficient splenocytes were mixed with IFNAR-sufficient m157-Tg or WT splenocytes, IFNAR-deficient NK cells were unable to respond to IFN-I and m157-Tg stimulation but still produced IFN-γ in response to IL-12 and m157-Tg targets, indicating that IFN-β acts cell-intrinsically on NK cells (Fig. 2B). Conversely, IL-12R–deficient NK cells were unable to respond to IL-12– and IL-12R–sufficient m157-Tg splenocytes but did respond to IFN-β and m157-Tg targets (Fig. 2C), indicating that IL-12 acts directly on NK cells, as well. Furthermore, stimulation of purified NK cells with m157-Tg MEF and cytokines was capable of inducing IFN-γ, providing further evidence that these cytokines act directly on NK cells (Supplemental Fig. 2B).

Interestingly, most IFN-γ was produced by Ly49H+ NK cells with lower levels of Ly49H (Fig. 2A), suggesting that these cells had directly engaged m157, which modulated Ly49H expression, as previously described (20). To confirm the involvement of Ly49H, we analyzed NK cells from B6.BxD8 mice that selectively lacked Ly49H expression and found that m157-Tg target cells were unable to stimulate IFN-γ production even with IFN-I or IL-12 (Fig. 2D). Because Ly49H has been reported to signal through the signaling adapters DAP10 and DAP12 (11, 23), we investigated which of these molecules was required for Ly49H-dependent IFN-γ production by analyzing NK cells deficient in either of these molecules (Fig. 2E). DAP10-deficient NK cells were fully capable of producing IFN-γ, whereas DAP12-deficient NK cells were unable to produce IFN-γ. These data show that signaling through DAP12 but not DAP10 is required for Ly49H-dependent IFN-γ production.

To determine whether cytokine and activation receptor signals were needed simultaneously, we separated cytokine and activation receptor stimulation (Fig. 3A). Upon stimulation of NK cells, first with cytokines for 3 h and then without cytokines for a 1-h rest followed by m157-Tg MEF coculture, we observed IFN-γ production similar to stimulation with both cytokines and m157-Tg MEF throughout the assay (Fig. 3B, 3C). However, when the stimuli were reversed and NK cells were first stimulated with m157-Tg MEF and with cytokines second, we did not observe IFN-γ production by Ly49H+ NK cells (Fig. 3D). These experiments indicate that NK cells do not have to receive cytokine and activation receptor signals simultaneously, but the order by which they receive these signals is important.

FIGURE 3.

Cytokine and activation receptor stimulation is not required simultaneously, but cytokine exposure is needed before activation receptor triggering. (A) Experimental setup. (B) IFN-γ production after simultaneous stimulation with cytokines and m157-Tg MEF. (C) IFN-γ production after pretreatment with cytokines and subsequent stimulation with m157-Tg MEF. (D) IFN-γ production after pretreatment with m157-Tg MEF and subsequent stimulation with cytokines. Samples were analyzed as in Fig. 2. A representative experiment of four independent experiments performed in duplicate is shown. ***p < 0.001, **p < 0.01, *p < 0.05. ns, not significant.

FIGURE 3.

Cytokine and activation receptor stimulation is not required simultaneously, but cytokine exposure is needed before activation receptor triggering. (A) Experimental setup. (B) IFN-γ production after simultaneous stimulation with cytokines and m157-Tg MEF. (C) IFN-γ production after pretreatment with cytokines and subsequent stimulation with m157-Tg MEF. (D) IFN-γ production after pretreatment with m157-Tg MEF and subsequent stimulation with cytokines. Samples were analyzed as in Fig. 2. A representative experiment of four independent experiments performed in duplicate is shown. ***p < 0.001, **p < 0.01, *p < 0.05. ns, not significant.

Close modal

We assessed the contributions of transcription and translation during cytokine and m157-Tg exposure (Fig. 4A, 4B). We blocked with either the transcriptional inhibitor AD or protein synthesis inhibitor CHX during each stage. AD exposure during initial cytokine stimulation blocked IFN-γ production almost completely, whereas Ly49H+ NK cells were still able to produce IFN-γ when AD was added only during subsequent m157-Tg MEF stimulation (Fig. 4B). In contrast, CHX during initial cytokine stimulation did not block IFN-γ production, whereas it completely blocked IFN-γ production during subsequent m157-Tg stimulation. Taken together, these data suggest that cytokine stimulation induces Ifng transcription, whereas m157-Tg stimulation induces translation.

FIGURE 4.

IL-12 and IFN-β induce Ifng mRNA, whereas signaling through Ly49H initiates IFN-γ translation. (A) Experimental setup. (B) Splenocytes were pretreated with IL-12; thereafter, the cells were washed and incubated with m157-Tg MEF in the presence of monensin. During IL-12 (cytokine) or m157-Tg stimulation, transcription inhibitor AD (2.5 μg/ml) or translation inhibitor CHX (5 μg/ml) was added. (C) Splenic NK cells were purified by negative selection. Six hours after stimulation with cytokines and/or m157-Tg MEF, RNA was isolated from NK cells. Ifng transcripts were quantified using TaqMan quantitative PCR and normalized to Gapdh expression. (D) Purified NK cells were stimulated with IL-12 and/or m157-Tg MEF for 6 h and stained for IFN-γ protein and Ifng mRNA using the PrimeFlow RNA Assay. (E) Splenocytes were pretreated with IL-12; thereafter, the cells were washed and stimulated with m157-Tg MEF in the presence of monensin and MNK-1/eIF4 inhibitor CGP-57380 or eIF2 inhibitor ISRIB at the indicated concentrations. A representative experiment of three independent experiments is shown, and each was performed in duplicate (B and E) or triplicate (C). ***p < 0.001, **p < 0.01, *p < 0.05. ns, not significant.

FIGURE 4.

IL-12 and IFN-β induce Ifng mRNA, whereas signaling through Ly49H initiates IFN-γ translation. (A) Experimental setup. (B) Splenocytes were pretreated with IL-12; thereafter, the cells were washed and incubated with m157-Tg MEF in the presence of monensin. During IL-12 (cytokine) or m157-Tg stimulation, transcription inhibitor AD (2.5 μg/ml) or translation inhibitor CHX (5 μg/ml) was added. (C) Splenic NK cells were purified by negative selection. Six hours after stimulation with cytokines and/or m157-Tg MEF, RNA was isolated from NK cells. Ifng transcripts were quantified using TaqMan quantitative PCR and normalized to Gapdh expression. (D) Purified NK cells were stimulated with IL-12 and/or m157-Tg MEF for 6 h and stained for IFN-γ protein and Ifng mRNA using the PrimeFlow RNA Assay. (E) Splenocytes were pretreated with IL-12; thereafter, the cells were washed and stimulated with m157-Tg MEF in the presence of monensin and MNK-1/eIF4 inhibitor CGP-57380 or eIF2 inhibitor ISRIB at the indicated concentrations. A representative experiment of three independent experiments is shown, and each was performed in duplicate (B and E) or triplicate (C). ***p < 0.001, **p < 0.01, *p < 0.05. ns, not significant.

Close modal

To confirm that cytokine stimulation induces Ifng transcription, we analyzed purified NK cells in response to cytokine and m157-Tg MEF stimulation. Stimulation using purified NK cells resulted in similar IFN-γ production compared with the splenocyte cultures (Supplemental Fig. 2B). In response to IFN-β or IL-12, Ifng transcripts were increased ∼100-fold, whereas stimulation with m157-Tg MEF alone did not induce increased Ifng transcript levels in purified NK cells (Fig. 4C). The combination of cytokines and m157-Tg MEF resulted in similar Ifng transcripts as did cytokine alone. We used flow cytometry to visualize Ifng mRNA versus IFN-γ protein following stimulation with both cytokines and m157-Tg on a single-cell level (Fig. 4D). Consistent with our quantitative PCR data, we found that IL-12 alone but not m157-Tg MEF alone stimulated Ifng transcription. The combination of IL-12 and m157-Tg MEF did not result in increased Ifng transcript levels compared with IL-12 alone, but NK cells with high Ifng transcript levels produced more IFN-γ protein. Incubation with IL-12 and IL-18, which resulted in high Ifng transcript and simultaneous IFN-γ protein levels, is shown as a positive control. Thus, cytokine stimulation, but not Ly49H activation, induces Ifng transcription.

To validate the contribution of IFN-γ translation during m157-Tg stimulation, we blocked initiation of translation with inhibitors for two potential initiation pathways, namely, the MNK1-inhibitor CGP57380 for the eIF4 complex and ISRIB for the eIF2 complex. CGP57380 addition during m157-Tg target stimulation prevented IFN-γ production in a dose-dependent manner, whereas ISRIB did not (Fig. 4E). Neither inhibitor affected NK cell viability at the concentrations used (data not shown). These results suggest that Ly49H activation initiates IFN-γ translation and that the eIF4 but not the eIF2 complex is required. Together, the data presented in this study indicate that IFN-β or IL-12 induces Ifng transcription, whereas activation receptor stimulation acts posttranscriptionally to induce IFN-γ protein production.

To determine whether IFN-γ is produced in response to cytokine stimulation but rapidly degraded by the proteasome, we blocked the proteasome (Fig. 5A). Addition of MG-132 during cytokine stimulation did not alter IFN-γ induced upon secondary stimulation with WT MEF, suggesting that IFN-γ is not continuously degraded by the proteasome during initial cytokine stimulation. Unexpectedly, we observed decreased IFN-γ upon MG-132 treatment during subsequent m157-Tg stimulation, indicating that proteasomal degradation was instead required for IFN-γ production at the time of Ly49H activation.

FIGURE 5.

Ly49H signaling requires the proteasome–ubiquitin–IKK–TPL2–ERK axis to induce IFN-γ. (A) As in Fig. 4A and 4B, splenocytes were pretreated with IL-12; thereafter, the cells were washed and incubated with m157-Tg MEF in the presence of monensin. During IL-12 (cytokine) or m157-Tg stimulation, the proteasome inhibitor MG-132 (10 μM) was added. (BD and F) As in Fig. 4E, splenocytes were pretreated with IL-12; thereafter, the cells were washed and stimulated with m157-Tg MEF in the presence of monensin and E1 inhibitor Pyr-41 (B), IKK complex inhibitor BMS-345541 (C), TPL2 inhibitor (D), or ERK inhibitor PD90859 (F) at the indicated concentrations. (E) Purified NK cells were prestimulated with IL-12, after which the NK cells were stimulated with m157-Tg MEF for the indicated time. ERK and phospho-ERK levels were analyzed by Western blot. (G) Splenocytes were stimulated with the indicated plate-bound Ab or IL-12 (12.5 ng/ml) and IL-18 (5 ng/ml) in the presence or absence of 9 μM TPL2 inhibitor. The percentage of IFN-γ–producing NK cells was normalized to the condition without inhibitor. Representative experiments of two (A, D, E, and G) or three (B, C, and F) independent experiments are shown and were performed in duplicate (A–C, F, and G) or triplicate (D). **p < 0.01, *p < 0.05. ns, not significant.

FIGURE 5.

Ly49H signaling requires the proteasome–ubiquitin–IKK–TPL2–ERK axis to induce IFN-γ. (A) As in Fig. 4A and 4B, splenocytes were pretreated with IL-12; thereafter, the cells were washed and incubated with m157-Tg MEF in the presence of monensin. During IL-12 (cytokine) or m157-Tg stimulation, the proteasome inhibitor MG-132 (10 μM) was added. (BD and F) As in Fig. 4E, splenocytes were pretreated with IL-12; thereafter, the cells were washed and stimulated with m157-Tg MEF in the presence of monensin and E1 inhibitor Pyr-41 (B), IKK complex inhibitor BMS-345541 (C), TPL2 inhibitor (D), or ERK inhibitor PD90859 (F) at the indicated concentrations. (E) Purified NK cells were prestimulated with IL-12, after which the NK cells were stimulated with m157-Tg MEF for the indicated time. ERK and phospho-ERK levels were analyzed by Western blot. (G) Splenocytes were stimulated with the indicated plate-bound Ab or IL-12 (12.5 ng/ml) and IL-18 (5 ng/ml) in the presence or absence of 9 μM TPL2 inhibitor. The percentage of IFN-γ–producing NK cells was normalized to the condition without inhibitor. Representative experiments of two (A, D, E, and G) or three (B, C, and F) independent experiments are shown and were performed in duplicate (A–C, F, and G) or triplicate (D). **p < 0.01, *p < 0.05. ns, not significant.

Close modal

To investigate how proteasomal degradation could be required for IFN-γ production, we used a panel of established inhibitors to examine ubiquitin-dependent signaling pathways requiring the proteasome. To determine involvement of the ubiquitin pathway, we blocked the E1 ubiquitin ligase with Pyr-41 and observed a dose-dependent decrease in IFN-γ production, suggesting that ubiquitination is required (Fig. 5B). We analyzed if the IKK complex was involved with BMS-345541, which inhibited IFN-γ production in a dose-dependent manner (Fig. 5C). Stimulation of NK cells with m157-Tg cells resulted in modest p65 phosphorylation at early time points, suggesting that IFN-γ production may be independent of canonical NF-κB (Supplemental Fig. 2C). The IKK complex has also been implicated in posttranscriptional regulation of TNF-α production by macrophages through activation of TPL2 (24, 25). To investigate whether IFN-γ production by NK cells could be similarly regulated, we inhibited TPL2 and detected decreased IFN-γ production in a dose-dependent manner, suggesting that Ly49H signaling through TPL2 is required for IFN-γ production (Fig. 5D). ERK has been reported to act downstream of TPL2 as well as in a number of immune receptor signaling pathways; indeed, we detected NK cell ERK phosphorylation in response to m157-Tg MEF stimulation (Fig. 5E). Additionally, blockade of ERK with PD90859 inhibited IFN-γ production (Fig. 5F). Finally, we analyzed whether this pathway potentially is involved in other means of NK cell stimulation. TPL2 inhibition during stimulation with plate-bound Abs against NK1.1, Ly49H, Ly49D, and NKp46 resulted in a virtually complete reduction in IFN-γ production (Fig. 5G, Supplemental Fig. 2D). Intriguingly, inhibition of TPL2 during stimulation with a combination of IL-12 and IL-18 did not result in a reduction in IFN-γ, even at lower stimulatory cytokine concentrations (Supplemental Fig. 2E), indicating that it is independent of the TPL2 axis and that the TPL2 inhibitor is not toxic at the concentrations used. Taken together, these data suggest that NK cell IFN-γ induced specifically through activation receptors requires the proteasome–ubiquitin–IKK–TPL2–ERK axis.

Whereas we previously described that dual signaling is required for full NK cytolytic function during MCMV infection (20), in this study, we uncovered a distinct role for activation receptor–dependent IFN-γ production that is regulated at the transcriptional and translational levels. The data presented in this study support the following model (Supplemental Fig. 3): stimulation of NK cells with IFN-I and IL-12 induces Ifng transcription. Subsequently, a second signal through an activation receptor initiates translation of Ifng mRNA into protein. In the case of the DAP12-dependent activation receptor Ly49H, signaling requires the ubiquitin–IKK–TPL2–MNK1–eIF4 axis to induce IFN-γ protein. This pathway likely plays a role downstream of other activation receptors, as well, as inhibition of TPL2 blocked IFN-γ production in response to anti-NK1.1, anti-Ly49H, and anti-NKp46. Intriguingly, IFN-γ in response to IL-12 in combination with IL-18 was not dependent on TPL2, suggesting that a distinct pathway is involved, as compared with activation receptor stimulation.

In this study, we found that primary m157-Tg splenocytes and MEF alone were unable to stimulate robust NK cell IFN-γ production. In contrast, cell lines overexpressing m157, such as BAF-m157, are capable of inducing NK cell IFN-γ (5, 16, 17). We observed similar expression levels of m157 in m157-Tg MEF and BAF-m157, indicating that in both situations the ligand levels are similar. The cell lines overexpressing m157 potentially express additional NK ligands or soluble factors that may bypass the requirement for a second signal. To avoid contributions of such factors, we used primary m157-Tg cells that are the same as WT cells except for transgenic expression of m157.

Resting NK cells have increased Ifng transcript levels compared with naive T cells, as reported in Yeti mice that have the Ifng 3′ untranslated region (UTR) replaced by YFP (26). The data presented in this study suggest that these levels are not enough for optimal activation receptor–dependent IFN-γ production. We observed an ∼100-fold increase in Ifng transcript levels upon treatment with IL-12 or IFN-β, which allowed for IFN-γ protein expression upon engagement of Ly49H. Of note, the combination of IL-12 and IL-18 resulted in even higher Ifng transcript levels, presumably partially or completely bypassing the need for activation receptor engagement as shown in this study. NK cells that are previously exposed to MCMV infection also have increased Ifng transcripts compared with naive NK cells, as reported in Yeti mice (27), suggesting that target cell recognition by activation receptors without cytokine exposure may be enough for these NK cells to produce IFN-γ.

We observed increased IFN-γ production upon MG-132 treatment during cytokine stimulation that is followed by m157-Tg stimulation (Fig. 5A). Jak2 and STAT4 are downstream of the IL-12R; both of these molecules have been described to be degraded by the proteasome as part of negative regulation (28, 29). Thus, blocking the proteasome during cytokine stimulation may potentially lead to sustained cytokine signaling and increased Ifng transcription, which may explain the increase in IFN-γ production observed upon MG-132 treatment during cytokine stimulation.

IFN-γ production is known to be regulated at the transcriptional and posttranscriptional levels. Transcription is regulated through a number of transcription factors including T-bet and Eomes as well as microRNAs and long noncoding RNAs. The UTR have been implicated in posttranscriptional control of IFN-γ. The human Ifng 5′ UTR contains a pseudoknot that inhibits eIF2-dependent translation through PKR (30). We found a role for eIF4 and not eIF2 in NK cell IFN-γ production, suggesting that the observed IFN-γ is independent of the pseudoknot. However, deletion of the Ifng 3′ UTR increased IFN-γ–producing NK and T cells treated with IL-12 (31), indicating that the Ifng 3′ UTR may play another important role in posttranscriptional control of IFN-γ production. The ARE-binding protein ZFP36L2 has recently been identified to bind to the Ifng 3′ UTR in T cells, thereby repressing its translation (32). Our data are consistent with a translational repressor that is released upon activation receptor signaling, and perhaps ZFP36L2 may regulate NK IFN-γ production, as well. Yet unclear is how receptor activation signals would relieve ZFP36L2 from its putative repression of IFN-γ translation.

We observed increased IFN-γ production upon CHX treatment during cytokine stimulation followed by m157-Tg stimulation (Fig. 4B); CHX treatment potentially resulted in decreased levels of molecules that negatively regulate IFN-γ production under normal conditions. For example, blocking protein production of ZFP36L2, suppressor of cytokine signaling 1 (SOCS1), and/or downstream regulatory element antagonist modulator (DREAM) (33) may cause or contribute to increased IFN-γ production upon CHX treatment during cytokine stimulation.

TPL2 has been shown to be involved in posttranscriptional control of TNF-α production (24). TPL2 can be activated by IΚB-dependent ubiquitination and degradation of p105 (25). Our data, using established inhibitors blocking the proteasome, E1 ubiquitin ligase, IΚK, and TPL2, suggest that this pathway may also be involved in Ly49H-dependent IFN-γ production. Although inhibitors were used in this study to outline the proteasome–ubiquitin–IKK–TPL2 axis, these pathways have been well established in other systems in which inhibitors have been validated. However, follow-up studies with mice (conditionally) deficient in key players will be required to confirm these findings. TPL2-deficient CD4+ T cells are defective in producing IFN-γ and controlling Toxoplasma gondii (34), suggesting that IFN-γ may be regulated by TPL2 in a similar fashion to TNF-α.

In conclusion, we uncovered a stepwise initiation process of NK cell IFN-γ transcription versus translation by cytokines and activation receptors as well as by proteasome degradation. These pathways are likely to be subject to regulation at multiple points, potentially allowing for specific IFN-γ production during pathogenic infections and tumor surveillance while preventing autoimmunity and immunopathology.

This work was supported by National Institutes of Health Grant R01-AI131680 to W.M.Y., and S.J.P. was supported by the Netherlands Organisation for Scientific Research (Rubicon Grant 825.11.004).

The online version of this article contains supplemental material.

Abbreviations used in this article:

AD

actinomycin D

CHX

cycloheximide

MCMV

murine CMV

MEF

murine embryonic fibroblast

m157-Tg

m157-transgenic

UTR

untranslated region

WT

wild-type.

1
Long
,
E. O.
,
H. S.
Kim
,
D.
Liu
,
M. E.
Peterson
,
S.
Rajagopalan
.
2013
.
Controlling natural killer cell responses: integration of signals for activation and inhibition.
Annu. Rev. Immunol.
31
:
227
258
.
2
Freeman
,
B. E.
,
H. P.
Raué
,
A. B.
Hill
,
M. K.
Slifka
.
2015
.
Cytokine-mediated activation of NK cells during viral infection.
J. Virol.
89
:
7922
7931
.
3
Orange
,
J. S.
,
C. A.
Biron
.
1996
.
Characterization of early IL-12, IFN-alphabeta, and TNF effects on antiviral state and NK cell responses during murine cytomegalovirus infection.
J. Immunol.
156
:
4746
4756
.
4
Loh
,
J.
,
D. T.
Chu
,
A. K.
O’Guin
,
W. M.
Yokoyama
,
H. W.
Virgin
IV
.
2005
.
Natural killer cells utilize both perforin and gamma interferon to regulate murine cytomegalovirus infection in the spleen and liver.
J. Virol.
79
:
661
667
.
5
Fodil
,
N.
,
D.
Langlais
,
P.
Moussa
,
G. A.
Boivin
,
T.
Di Pietrantonio
,
I.
Radovanovic
,
A.
Dumaine
,
M.
Blanchette
,
E.
Schurr
,
P.
Gros
,
S. M.
Vidal
.
2014
.
Specific dysregulation of IFNγ production by natural killer cells confers susceptibility to viral infection.
PLoS Pathog.
10
: e1004511.
6
Nguyen
,
K. B.
,
T. P.
Salazar-Mather
,
M. Y.
Dalod
,
J. B.
Van Deusen
,
X. Q.
Wei
,
F. Y.
Liew
,
M. A.
Caligiuri
,
J. E.
Durbin
,
C. A.
Biron
.
2002
.
Coordinated and distinct roles for IFN-alpha beta, IL-12, and IL-15 regulation of NK cell responses to viral infection.
J. Immunol.
169
:
4279
4287
.
7
Pien
,
G. C.
,
A. R.
Satoskar
,
K.
Takeda
,
S.
Akira
,
C. A.
Biron
.
2000
.
Cutting edge: selective IL-18 requirements for induction of compartmental IFN-gamma responses during viral infection.
J. Immunol.
165
:
4787
4791
.
8
Takeda
,
K.
,
M.
Nakayama
,
M.
Sakaki
,
Y.
Hayakawa
,
M.
Imawari
,
K.
Ogasawara
,
K.
Okumura
,
M. J.
Smyth
.
2011
.
IFN-γ production by lung NK cells is critical for the natural resistance to pulmonary metastasis of B16 melanoma in mice.
J. Leukoc. Biol.
90
:
777
785
.
9
Smith
,
H. R.
,
H. H.
Chuang
,
L. L.
Wang
,
M.
Salcedo
,
J. W.
Heusel
,
W. M.
Yokoyama
.
2000
.
Nonstochastic coexpression of activation receptors on murine natural killer cells.
J. Exp. Med.
191
:
1341
1354
.
10
Kim
,
S.
,
J.
Poursine-Laurent
,
S. M.
Truscott
,
L.
Lybarger
,
Y. J.
Song
,
L.
Yang
,
A. R.
French
,
J. B.
Sunwoo
,
S.
Lemieux
,
T. H.
Hansen
,
W. M.
Yokoyama
.
2005
.
Licensing of natural killer cells by host major histocompatibility complex class I molecules.
Nature
436
:
709
713
.
11
Orr
,
M. T.
,
J. C.
Sun
,
D. G.
Hesslein
,
H.
Arase
,
J. H.
Phillips
,
T.
Takai
,
L. L.
Lanier
.
2009
.
Ly49H signaling through DAP10 is essential for optimal natural killer cell responses to mouse cytomegalovirus infection.
J. Exp. Med.
206
:
807
817
.
12
May
,
R. M.
,
M.
Okumura
,
C. J.
Hsu
,
H.
Bassiri
,
E.
Yang
,
G.
Rak
,
E. M.
Mace
,
N. H.
Philip
,
W.
Zhang
,
T.
Baumgart
, et al
.
2013
.
Murine natural killer immunoreceptors use distinct proximal signaling complexes to direct cell function.
Blood
121
:
3135
3146
.
13
Ho
,
E. L.
,
L. N.
Carayannopoulos
,
J.
Poursine-Laurent
,
J.
Kinder
,
B.
Plougastel
,
H. R.
Smith
,
W. M.
Yokoyama
.
2002
.
Costimulation of multiple NK cell activation receptors by NKG2D.
J. Immunol.
169
:
3667
3675
.
14
Bryceson
,
Y. T.
,
M. E.
March
,
H. G.
Ljunggren
,
E. O.
Long
.
2006
.
Synergy among receptors on resting NK cells for the activation of natural cytotoxicity and cytokine secretion.
Blood
107
:
159
166
.
15
Fauriat
,
C.
,
E. O.
Long
,
H. G.
Ljunggren
,
Y. T.
Bryceson
.
2010
.
Regulation of human NK-cell cytokine and chemokine production by target cell recognition.
Blood
115
:
2167
2176
.
16
Diefenbach
,
A.
,
A. M.
Jamieson
,
S. D.
Liu
,
N.
Shastri
,
D. H.
Raulet
.
2000
.
Ligands for the murine NKG2D receptor: expression by tumor cells and activation of NK cells and macrophages.
Nat. Immunol.
1
:
119
126
.
17
Smith
,
H. R.
,
J. W.
Heusel
,
I. K.
Mehta
,
S.
Kim
,
B. G.
Dorner
,
O. V.
Naidenko
,
K.
Iizuka
,
H.
Furukawa
,
D. L.
Beckman
,
J. T.
Pingel
, et al
.
2002
.
Recognition of a virus-encoded ligand by a natural killer cell activation receptor.
Proc. Natl. Acad. Sci. USA
99
:
8826
8831
.
18
Tripathy
,
S. K.
,
P. A.
Keyel
,
L.
Yang
,
J. T.
Pingel
,
T. P.
Cheng
,
A.
Schneeberger
,
W. M.
Yokoyama
.
2008
.
Continuous engagement of a self-specific activation receptor induces NK cell tolerance.
J. Exp. Med.
205
:
1829
1841
.
19
Bolanos
,
F. D.
,
S. K.
Tripathy
.
2011
.
Activation receptor-induced tolerance of mature NK cells in vivo requires signaling through the receptor and is reversible.
J. Immunol.
186
:
2765
2771
.
20
Parikh
,
B. A.
,
S. J.
Piersma
,
M. A.
Pak-Wittel
,
L.
Yang
,
R. D.
Schreiber
,
W. M.
Yokoyama
.
2015
.
Dual requirement of cytokine and activation receptor triggering for cytotoxic control of murine cytomegalovirus by NK cells.
PLoS Pathog.
11
: e1005323.
21
Sojka
,
D. K.
,
B.
Plougastel-Douglas
,
L.
Yang
,
M. A.
Pak-Wittel
,
M. N.
Artyomov
,
Y.
Ivanova
,
C.
Zhong
,
J. M.
Chase
,
P. B.
Rothman
,
J.
Yu
, et al
.
2014
.
Tissue-resident natural killer (NK) cells are cell lineages distinct from thymic and conventional splenic NK cells.
Elife
3
: e01659.
22
Fehniger
,
T. A.
,
S. F.
Cai
,
X.
Cao
,
A. J.
Bredemeyer
,
R. M.
Presti
,
A. R.
French
,
T. J.
Ley
.
2007
.
Acquisition of murine NK cell cytotoxicity requires the translation of a pre-existing pool of granzyme B and perforin mRNAs.
Immunity
26
:
798
811
.
23
Sjölin
,
H.
,
E.
Tomasello
,
M.
Mousavi-Jazi
,
A.
Bartolazzi
,
K.
Kärre
,
E.
Vivier
,
C.
Cerboni
.
2002
.
Pivotal role of KARAP/DAP12 adaptor molecule in the natural killer cell-mediated resistance to murine cytomegalovirus infection.
J. Exp. Med.
195
:
825
834
.
24
Dumitru
,
C. D.
,
J. D.
Ceci
,
C.
Tsatsanis
,
D.
Kontoyiannis
,
K.
Stamatakis
,
J. H.
Lin
,
C.
Patriotis
,
N. A.
Jenkins
,
N. G.
Copeland
,
G.
Kollias
,
P. N.
Tsichlis
.
2000
.
TNF-alpha induction by LPS is regulated posttranscriptionally via a Tpl2/ERK-dependent pathway.
Cell
103
:
1071
1083
.
25
Beinke
,
S.
,
M. J.
Robinson
,
M.
Hugunin
,
S. C.
Ley
.
2004
.
Lipopolysaccharide activation of the TPL-2/MEK/extracellular signal-regulated kinase mitogen-activated protein kinase cascade is regulated by IkappaB kinase-induced proteolysis of NF-kappaB1 p105.
Mol. Cell. Biol.
24
:
9658
9667
.
26
Stetson
,
D. B.
,
M.
Mohrs
,
R. L.
Reinhardt
,
J. L.
Baron
,
Z. E.
Wang
,
L.
Gapin
,
M.
Kronenberg
,
R. M.
Locksley
.
2003
.
Constitutive cytokine mRNAs mark natural killer (NK) and NK T cells poised for rapid effector function.
J. Exp. Med.
198
:
1069
1076
.
27
Sun
,
J. C.
,
J. N.
Beilke
,
L. L.
Lanier
.
2009
.
Adaptive immune features of natural killer cells. [Published erratum appears in 2009 Nature 457: 1168.]
Nature
457
:
557
561
.
28
Ungureanu
,
D.
,
P.
Saharinen
,
I.
Junttila
,
D. J.
Hilton
,
O.
Silvennoinen
.
2002
.
Regulation of Jak2 through the ubiquitin-proteasome pathway involves phosphorylation of Jak2 on Y1007 and interaction with SOCS-1.
Mol. Cell. Biol.
22
:
3316
3326
.
29
Tanaka
,
T.
,
M. A.
Soriano
,
M. J.
Grusby
.
2005
.
SLIM is a nuclear ubiquitin E3 ligase that negatively regulates STAT signaling.
Immunity
22
:
729
736
.
30
Ben-Asouli
,
Y.
,
Y.
Banai
,
Y.
Pel-Or
,
A.
Shir
,
R.
Kaempfer
.
2002
.
Human interferon-gamma mRNA autoregulates its translation through a pseudoknot that activates the interferon-inducible protein kinase PKR.
Cell
108
:
221
232
.
31
Hodge
,
D. L.
,
C.
Berthet
,
V.
Coppola
,
W.
Kastenmüller
,
M. D.
Buschman
,
P. M.
Schaughency
,
H.
Shirota
,
A. J.
Scarzello
,
J. J.
Subleski
,
M. R.
Anver
, et al
.
2014
.
IFN-gamma AU-rich element removal promotes chronic IFN-gamma expression and autoimmunity in mice.
J. Autoimmun.
53
:
33
45
.
32
Salerno
,
F.
,
S.
Engels
,
M.
van den Biggelaar
,
F. P. J.
van Alphen
,
A.
Guislain
,
W.
Zhao
,
D. L.
Hodge
,
S. E.
Bell
,
J. P.
Medema
,
M.
von Lindern
, et al
.
2018
.
Translational repression of pre-formed cytokine-encoding mRNA prevents chronic activation of memory T cells.
Nat. Immunol.
19
:
828
837
.
33
Fenimore
,
J.
,
H.
A Young
.
2016
.
Regulation of IFN-γ expression.
Adv. Exp. Med. Biol.
941
:
1
19
.
34
Watford
,
W. T.
,
B. D.
Hissong
,
L. R.
Durant
,
H.
Yamane
,
L. M.
Muul
,
Y.
Kanno
,
C. M.
Tato
,
H. L.
Ramos
,
A. E.
Berger
,
L.
Mielke
, et al
.
2008
.
Tpl2 kinase regulates T cell interferon-gamma production and host resistance to Toxoplasma gondii.
J. Exp. Med.
205
:
2803
2812
.

The authors have no financial conflicts of interest.

Supplementary data