Cutting Edge: Developmental Regulation of IFN-γ Production by Mouse Neutrophil Precursor Cells

Induction of immune responses to microbial pathogens is generally divided into distinct stages, in which the recognition of microbial molecules mediated by innate immune receptors results in maturation of dendritic cells and macrophages, followed by secretion of activating cytokines that regulate the effector phase of host defense (1, 2). This model is particularly well studied in the context of TLR-mediated immunity, in which DC-specific TLR activation results in production of IL-12 that subsequently regulates IFN-γ production by NK and T cells (3). In vivo experiments with the protozoan parasite Toxoplasma gondii formally tested this model and revealed that whereas NK and CD4⁺ T cell IFN-γ production was regulated by TLR11-dependent activation of MyD88 (4–8), there is an unforeseen component of IFN-γ–dependent host response mediated by neutrophils (9). In contrast to NK cells (10, 11), T cells (5, 12, 13), and innate lymphoid cells (14), the ability of neutrophils to produce IFN-γ does not depend on IL-12– or TLR-dependent parasite recognition (9), and therefore how neutrophils produce IFN-γ is incompletely understood. This knowledge is crucial, because although neutrophils have long been viewed as effector cells that mediate their protective effects via the release of lytic granules, recent evidence suggests that neutrophils secrete cytokines that have pleiotropic effects on host defense to microbial infections (15, 16).

In this study, we investigated how neutrophils produce IFN-γ. Our data revealed that neutrophil IFN-γ production is regulated during hematopoietic development of neutrophil precursor cells and that IFN-γ accumulates in primary granules produced in the promyelocyte stage of neutrophil development. The accumulation of IFN-γ in primary granules is independent of microbial infection or the endogenous microbiota. Neutrophil IFN-γ effector mechanisms are achieved by activation-induced degranulation. This represents a broad innate immune mechanism of IFN-γ–mediated host defense required for host protection from intracellular pathogens.

C57BL/6, germ-free C57BL/6, and TLR11^−/− mice have been previously described in experimental toxoplasmosis (5, 6, 17). All experiments were performed with protocols approved by the Institutional Animal Care and Use Committees at the University of Texas Southwestern Medical Center.

C57BL/6 and TLR11^−/− mice were infected i.p. with 20 cysts of T. gondii (ME49 strain) as previously described (6, 9). Cells were not given any stimulation or Golgi inhibitors unless otherwise indicated. For indicated experiments, peritoneal cells were incubated for 5 h at 37°C in media with or without GolgiPlug (1 μg/ml). To block protein translation, peritoneal cells were incubated in media with GolgiPlug and increasing concentrations of cycloheximide, with or without PMA (50 ng/ml) and ionomycin (750 ng/ml). To induce neutrophil degranulation, cells were activated with 10 nM fMLP, 100 ng/ml LPS, 50 ng/ml PMA, and/or 750 ng/ml ionomycin, incubated for 1 h at 37°C, supernatants were collected for IFN-γ ELISA (eBioscience), and cells were stained for flow cytometry.

For bone marrow (BM) and peritoneal cells, single-cell suspensions were stained in PBS plus 1% FBS and 0.5 mM EDTA. Blood samples were directly stained and erythrocytes lysed with ACK lysing buffer. After extracellular staining, samples were fixed and permeabilized for intracellular staining using the Foxp3 staining buffer kit by eBioscience, per the manufacturer’s instructions. The following Abs were used for staining and were from eBioscience unless otherwise indicated: anti-NK1.1, anti-CD3ε, anti–IFN-γ (XMG1.2; BD Biosciences), anti–Ly-6G (BD Biosciences), anti-CD34, anti-CD117, anti-CD16/32, anti–Sca-1, anti-CD19, anti-F4/80, anti-Ter119, anti-CD11b, anti–Gr-1, and isotype IgG1κ. These data were acquired on FACSCalibur and FACSCanto cytometers or on the MoFlo cytometer and analyzed with FlowJo.

RNA from defined immature neutrophils was isolated using the PureLink RNA Mini kit. cDNA synthesis was done using SuperScript III reverse transcriptase. Samples were analyzed on MyiQ real-time PCR system, and data were processed for relative expression to control gene hypoxanthine phosphoribosyltransferase.

Magnetic bead sort-purified Ly-6G⁺ BM cells were fixed in 4% PFA, blocked in PBS with 2% BSA and 0.2% Triton X-100, and stained with Abs of interest. The following Abs were used for microscopy analysis: anti–IFN-γ Alexa Fluor 488 (XMG1.2; eBioscience), anti-myeloperoxidase (ab9535; Abcam), anti-lactoferrin (ab135710; Abcam), anti-gelatinase (ab38898; Abcam), and donkey anti-rabbit IgG Alexa Fluor 568 (A10042; Life Technologies). Images were acquired with Leica TCS SPE or SP5 laser scanning confocal microscopes. Downstream images were examined with ImageJ, Z stacks underwent three-dimensional blind deconvolution with AutoQuant X software, colocalization analysis (Manders coefficient) (18) was performed, and three-dimensional modeling was done with Imaris.

All statistical analysis on bar graphs is done using the standard unpaired t test, and error bars shown are mean ± SD.

T. gondii elicits potent innate and adaptive IFN-γ responses from multiple cell types, including neutrophils, innate lymphoid cells, NK cells, and T cells (19). Although TLR activation and IL-12 production by dendritic cells is required for IFN-γ production by lymphoid cells, how neutrophils mediate TLR- and IL-12–independent IFN-γ production is largely unknown (20). Therefore, we began to address this question by examining the appearance of IFN-γ–producing neutrophils triggered by T. gondii infection.

Neutrophils are present in large numbers in the peritoneal exudate cells isolated from the site of T. gondii infection in wild-type mice and they stain positive for IFN-γ (Fig. 1A). However, the amount of IFN-γ detected in neutrophils by intracellular staining was not increased by in vitro incubation with GolgiPlug, a protein transport inhibitor that leads to intracellular accumulation of IFN-γ seen in NK and T cells (Supplemental Fig. 1A). To boost IFN-γ expression, neutrophils and T cells isolated from T. gondii–infected mice were stimulated with PMA and ionomycin. As expected, stimulated T cells produced large amounts of IFN-γ that was inhibited in a dose-dependent manner by the de novo protein synthesis inhibitor cycloheximide (Supplemental Fig. 1B). In contrast, stimulation of neutrophils failed to increase their IFN-γ positivity, and cycloheximide treatment had no dose-dependent effects on the levels of IFN-γ detected in neutrophils (Supplemental Fig. 1B). Taken together, these data revealed that neutrophils expressed IFN-γ prior to isolation from T. gondii–infected mice.

To address the question of whether neutrophil IFN-γ production is limited to the site of infection in vivo, wild-type mice were infected with T. gondii and several tissues were examined for neutrophil IFN-γ positivity, including the peritoneal cavity, peripheral blood, and BM. Neutrophils stained positive for IFN-γ at a similar level in all locations examined (Fig. 1A, 1B, 1E). T. gondii parasites are largely confined to the site of experimental infection during the first 5 d, and thus these data suggest that an indirect sensing of infection may be responsible for IFN-γ production by neutrophils. However, T. gondii–infected TLR11^−/− mice also have IFN-γ⁺ neutrophils in all tissues examined (Supplemental Fig. 1C). Additionally, the analysis of neutrophils in naive mice revealed IFN-γ⁺ neutrophils in the blood and BM but not in the peritoneal cavity (Fig. 1C–E, Supplemental Fig. 1C). The amount of IFN-γ positivity by neutrophils in the BM and blood was comparable in naive and T. gondii–infected mice (Fig. 1E, 1F). Lack of IFN-γ⁺ neutrophils in the peritoneal cavity of naive mice is not surprising, given that neutrophils are not present in the peritoneal cavity of naive mice. Nevertheless, when thioglycollate was used to elicit neutrophils into the peritoneum of otherwise naive mice, neutrophils stained uniformly positive for IFN-γ (Supplemental Fig. 1D) When combined, these experiments suggested that neutrophils constitutively produce IFN-γ independent of infection.

One possible explanation for the appearance of IFN-γ–producing neutrophils in naive mice is that microbial products derived from intestinal microbiota could be stimulating neutrophils in peripheral tissues to produce IFN-γ, because commensal bacteria are involved in the activation of mature neutrophils (21, 22). To examine this possibility we quantified the presence of IFN-γ⁺ neutrophils in the BM of germ-free C57BL/6 mice. We observed that the colonization status of mice had no effect on the appearance of IFN-γ⁺ neutrophils (Fig. 1G–J). Thus, IFN-γ production by neutrophils is regulated differently from lymphoid cells and does not require the presence of T. gondii or other microbial stimuli. Instead, T. gondii infection triggers neutrophil recruitment to the site of infection, explaining the selective appearance of IFN-γ⁺ neutrophils in the peritoneal cavity of infected but not naive mice (Fig. 1A, 1C).

Analysis of BM neutrophils revealed that in addition to IFN-γ⁺Ly6G^high cells, there is an appearance of Ly-6G⁻ and Ly-6G^low cells that produce IFN-γ in germ-free, naive, and T. gondii–infected mice (Fig. 1). The presence of IFN-γ⁺ in the BM suggests a model of developmental regulation for IFN-γ production by neutrophils. To obtain insight into this hypothesis, we investigated the stages of neutrophil development and their association with IFN-γ production. Flow cytometric analysis of neutrophil precursors in BM excluded T, B, and NK cells, followed by the analysis of c-Kit–, CD34-, and Ly-6G–expressing cells (Fig. 2A). The cells were divided into five populations: myeloblasts (red), promyelocytes (orange), myelocytes (purple), metamyelocytes (green), and mature neutrophils (blue) (Fig. 2). To confirm the successful identification of neutrophil precursors, the immature neutrophil populations were sort purified and analyzed for stage-specific proteins by quantitative RT-PCR, including proteinase 3, a primary granule protein expressed mostly in early promyelocytes, lactoferrin (a secondary granule protein expressed mostly in metamyelocytes), and gelatinase (a tertiary granule protein expressed mostly in late band cells) (Supplemental Fig. 2A) (23). Examination of immature neutrophils for IFN-γ positivity by flow cytometry revealed that neutrophils acquire IFN-γ positivity at the promyelocyte stage, exhibiting a nearly equal split of IFN-γ⁺ and IFN-γ⁻ cells at this stage (Fig. 2B). Subsequent neutrophil stages were uniformly IFN-γ⁺ (Fig. 2B).

We also examined whether earlier hematopoietic cells were capable of producing IFN-γ. Lineage⁻sca-1+c-Kit+ (LSKs), which included the long-term hemopoietic stem cells, short-term hemopoietic stem cells, and multipotent progenitors, did not stain positive for IFN-γ (Supplemental Fig. 2B). Additionally, the immediate precursor to all granulocytes, the GMPs, also did not stain positive for IFN-γ (Supplemental Fig. 2B). These data revealed that IFN-γ expression in the BM is restricted to the neutrophil-specific lineage development (Fig. 2).

Given that promyelocyte stage neutrophils produce IFN-γ (Fig. 2B) and that immature neutrophils begin packaging granules as they transition to the promyelocyte stage (24), we hypothesized that neutrophil IFN-γ could be stored in granules. To examine the localization of IFN-γ, we used immunofluorescence detection of IFN-γ together with known granular proteins. Simultaneous visualization of IFN-γ with myeloperoxidase, lactoferrin, or gelatinase, which localize in primary, secondary, and tertiary granules, respectively, revealed colocalization between IFN-γ and myeloperoxidase (Fig. 3A). Minimal to no colocalization was observed between IFN-γ and lactoferrin or gelatinase (Fig. 3B, 3C). Quantitative colocalization analysis, using Manders coefficient, confirmed that IFN-γ is localized in the same granules that contain myeloperoxidase (Fig. 3). These results formally established that IFN-γ is largely contained within the primary granules of neutrophils.

To further examine the physiological significance of IFN-γ distribution in primary granules, we tested whether induction of neutrophil degranulation would result in IFN-γ release. We first induced neutrophil degranulation with fMLP and LPS and observed that this stimulation resulted in tertiary granule, but little primary granule, degranulation, and minor loss of IFN-γ positivity (Fig. 4A–C). However, the combined treatment of neutrophils with fMLP, LPS, PMA, and ionomycin resulted in primary granule degranulation as seen by the loss of myeloperoxidase staining in the stimulated neutrophils (Fig. 4B). Importantly, primary granule degranulation also resulted in a nearly complete loss of intracellular IFN-γ staining and simultaneous release of IFN-γ into the cell culture supernatant (Fig. 4C, 4D). Taken together, our data revealed a developmental mechanism for IFN-γ expression in immature neutrophils and that the release of IFN-γ is regulated by activation-induced primary granule degranulation.

In the present study, we have identified developmental regulation of IFN-γ expression in neutrophil precursor cells, which represents a novel mechanism for IFN-γ–mediated host defense. Whereas IFN-γ production by innate lymphoid cells, NK cells, and T cells is largely regulated by IL-12–dependent induction of IFN-γ expression postinfection, neutrophil IFN-γ is produced prior to infection in immature neutrophils at the promyelocyte stage. This mechanism prearms neutrophils, ensuring a rapid response to pathogens. Regulated secretion of IFN-γ from neutrophil granules prevents spontaneous effects of IFN-γ in the absence of microbial infections. Whereas IFN-γ–expressing neutrophils are in blood and BM, the effector mechanisms of neutrophil-derived IFN-γ are restricted to the sites of infection because their IFN-γ secretion requires microbial or inflammatory environment–driven degranulation of primary granules. Granular localization of IFN-γ may explain the confusion in the literature regarding the ability of neutrophils to produce IFN-γ, because both isolation-induced degranulation or insufficient granular protein fixation will result in an inability to detect neutrophil IFN-γ. In addition to the priming effects of IFN-γ on other cells types, IFN-γ can prime neutrophils themselves for increased migration, pathogen clearance, Ag presentation abilities, and cytokine production (25), the complexities of which will need to be revisited in light of neutrophil-derived IFN-γ.

The release of IFN-γ by neutrophils represents a bona fide type I innate immune response that can potentially be elicited by a variety of pathogens and host inflammatory responses sufficient to trigger neutrophil primary granule degranulation. In this regard it is of interest that neutrophil IFN-γ expression is developmentally regulated, which is distinct from their cytokine and stimuli-induced production of IL-22 and IL-17, two additional cytokines that were also, until recently, considered to be lymphoid cell–specific effector molecules (26, 27). Both IL-22 and IL-17 production by neutrophils is restricted to mucosal tissues, and in the case of IL-17, the transcription factor retinoic acid–related orphan receptor γt is a master regulator of IL-17 expression by neutrophils (27) and lymphoid cells (28). The localization of IFN-γ within primary granules of neutrophils provides an activation-dependent regulatory mechanism, because stronger inflammatory signals at the site of infection or inflammation are involved in primary granule degranulation. Overall, in this study we unveiled an arm of IFN-γ–mediated innate immunity that is prepared for immediate effector responses and is regulated by neutrophil recruitment and activation.

We thank Stephanie Watowich for critical reading of the manuscript.

The authors have no financial conflicts of interest.