|Home | About | Journals | Submit | Contact Us | Français|
Recognition of intracellular bacteria by macrophages leads to secretion of type I Interferons. However, the role of type I IFN during bacterial infection is still poorly understood. Francisella tularensis, the causative agent of tularemia, is a pathogenic bacterium that replicates in the cytosol of macrophages leading to secretion of type I IFN. Here, we investigated the role of type I IFN in a mouse model of tularemia. Mice deficient for type I IFN receptor (IFNAR1−/−) are more resistant to intradermal infection with F. tularensis subspecies novicida (F. novicida). Increased resistance to infection was associated with a specific increase in IL-17A/F and a corresponding expansion of an IL-17A+ γδ T cell population, indicating that type I IFN negatively regulate the number of IL-17A+ γδ T cells during infection. Furthermore, IL-17A-deficient mice contained fewer neutrophils compared to WT mice upon infection, indicating that IL-17A contributes to neutrophil expansion during F. novicida infection. Accordingly, an increase in IL-17A in IFNAR1−/− mice correlated with an increase in splenic neutrophil numbers. Similar results were obtained in a mouse model of pneumonic tularemia using the highly virulent Francisella tularensis subspecies tularensis SchuS4 strain and in a mouse model of systemic Listeria monocytogenes infection. Our results indicate that the type I IFN-mediated negative regulation of IL-17A+ γδ T cell expansion is conserved during bacterial infections. We propose that this newly described activity of type I IFN signaling might participate in the resistance of the IFNAR1−/− mice to infection with F. novicida and other intracellular bacteria.
Interferons (1) are potent cytokines induced and secreted during infection. Type I Interferons are comprised of several IFN-α proteins, a single IFN-β, as well as atypical IFN (2). Type I IFN are the major antiviral cytokines and their roles have been extensively studied during various viral infections. Type I IFN are also secreted in response to the recognition of “classical” LPS by the cell surface receptor TLR4 (3, 4) and in response to the detection of bacteria in the lysosomes by TLR7 (5). In addition, type I IFN have recently been shown to be secreted by macrophages and other infected cells, independently of the TLRs, in response to numerous bacteria such as Listeria monocytogenes (L. monocytogenes) (6), Francisella tularensis (F. tularensis) (7), Legionella pneumophila (8), Mycobacterium tuberculosis (M. tuberculosis) (9) and group B Streptococcus (10). Despite the recent attention on the signaling pathway leading to the secretion of type I IFN during bacterial infection (5, 10–13), the role of type I IFN signaling in this context is still largely unknown.
Mice deficient for the type I IFN receptor (IFNAR1−/−) have lower bacterial levels than wild-type (WT) mice upon infection with the intracellular bacteria L. monocytogenes (14–16), Chlamydia murinarum (C. murinarum) (17, 18) and M. tuberculosis (9). Resistance of the IFNAR1−/− mice to C. murinarum and L. monocytogenes infections have been associated with a decrease in apoptosis of macrophages (18) and T cells, respectively (14). A current model (19) that may explain the increased resistance of IFNAR1−/− mice during L. monocytogenes infection is that type I IFN sensitize lymphocytes to listeriolysin O-mediated apoptosis. Recognition of apoptotic cells by macrophages leads to an IL-10-dependent, anti-inflammatory response, which creates an environment favorable for bacterial replication. However, the type I IFN receptor is ubiquitously expressed and type I IFN have pleiotropic effects, suggesting that type I IFN might have additional roles during bacterial infections. Finally, since IFNAR1−/− mice are more resistant than WT mice to infection with most intracellular bacteria tested so far (9, 14–18), the contribution of specific virulence factors such as listeriolysin O to this resistance phenotype is unclear. Therefore, we decided to investigate the role of type I IFN signaling during infections with F. tularensis. We then extended our studies to infections with L. monocytogenes, a model facultative intracellular Gram-positive bacterial pathogen that has been used extensively to study host innate and adaptive immune responses during bacterial infection.
F. tularensis is a facultative intracellular Gram-negative bacterium that causes tularemia in humans. Infections occur naturally through the skin resulting in ulceroglandular tularemia. Alternatively, exposure to aerosolized F. tularensis can lead to pneumonic tularemia. The severity of infection is dependent on the F. tularensis subspecies (20). F. tularensis subspecies tularensis (referred to as F. tularensis) is a highly virulent subspecies with as few as one organism causing lethal pneumonic disease. In contrast, F. tularensis subspecies novicida (F. novicida) is avirulent in immunocompetent humans but is highly virulent in mice. Intradermal injection of F. novicida has thus been used as a mouse model of tularemia. The ability of Francisella tularensis species to cause disease is linked to its ability to replicate in the cytosol of host cells, particularly macrophages. Francisella species are Gram-negative bacteria with a non-classical LPS that is not recognized by TLR4 (21). However, macrophages infected with F. novicida or the F. tularensis subspecies holartica live vaccine strain (LVS) secrete large amounts of type I IFN in vitro, suggesting that a cytosolic surveillance pathway is able to detect the presence of Francisella species leading to type I IFN secretion (7, 13). Similarly, in mice infected with F. novicida, IFN-β transcript is highly up-regulated in the spleen (7). We therefore decided to study the effect of type I IFN signaling in vivo during infection with F. novicida. IFNAR1−/− mice were more resistant than WT mice to infection, indicating that type I IFN signaling is detrimental for the host during F. novicida infection. Using an unbiased approach, we demonstrated that type I IFN signaling in vivo negatively regulates the size of an IL-17A+ γδ T cell population. In agreement with the known role of IL-17A in neutrophil recruitment/expansion, IL-17A-deficient mice infected with F. novicida contained fewer neutrophils. Consistent with this finding, the increase in IL-17A observed in IFNAR1−/− mice correlated with an increase in neutrophil number. Furthermore, we also demonstrated the type I IFN-dependent negative regulation of IL-17A/F using a pneumonic model of tularemia with the highly virulent F. tularensis SchuS4, as well as an intravenous infection model with L. monocytogenes. Overall our results highlight a novel conserved role for type I IFN signaling during bacterial infections, which could partially explain the resistance of IFNAR1−/− mice to infection with intracellular bacteria.
F. novicida strain Utah 112 (U112) were grown in tryptic soy broth (TSB) supplemented with 0.1% cysteine or on Mueller Hinton (MH) agar plates supplemented with 2.5% bovine serum, 0.2% cysteine, 1% glucose, 0.025 % ferric pyrophosphate. F. tularensis Schu S4, originally isolated from a human case of tularemia, was obtained from the U.S. Army Medical Research Institute for Infectious Diseases (USAMRIID) (Frederick, MD). F. tularensis Schu S4 strain was cultured on modified MH chocolate agar plates or in MH broth (Difco Laboratories, Lawrence, KA) supplemented with ferric pyrophosphate and Iso-Vitalex (BD Biosciences, San Jose, CA). Active mid-log phase bacteria were harvested and stored in liquid nitrogen; one ml aliquots were thawed periodically for use. For mouse infections, L. monocytogenes strain 10403s were grown in brain heart infusion (BHI) media overnight at 30°C. Cultures were then backdiluted in BHI and grown to mid-logarithmic phase (0.45<OD600<0.6) at 37°C shaking for 1–2 hours.
Six- to 12-week-old B6.129S2-Ifnar1tm1Agt (IFNAR1−/−) mice on the C57BL/6J bred at the Stanford University or the University of California, Berkeley animal facilities were used for F. novicida and L. monocytogenes infection, respectively. IL-17A-deficient mice on the C57BL/6J background have been described before (22). WT mice were acquired from the Jackson Laboratory. Mice were kept under specific pathogen-free conditions in filter-top cages and provided with sterile water and food ad libitum. Experimental studies were in accordance with Institutional Animal Care and Use Committee guidelines. For intradermal infection with F. novicida, 105 cfu from an overnight culture grown in TSB cysteine in 50 μl endotoxin-free PBS (Gibco) were injected i.d., except for the survival experiments, in which 5×105 cfu were injected i.d.. For L. monocytogenes infection, 1.4×104 cfu in 200 μl endotoxin-free PBS were inoculated by tail vein injection. All Schu S4 challenge experiments were performed in a Centers for Disease Control and Prevention-approved Animal Biosafety Level 3 facility at Albany Medical College. Four to six week old BALB/cJ and IFNAR1−/− mice were anesthetized by i.p. injection of 100 μl of xylazine (20 mg/ml) and ketamine (1 mg/ml). 25 cfu of F. tularensis Schu S4 in 50 μl of PBS were delivered by the i.n. route. At various times PI, the mice were euthanized; blood was collected by intracardiac puncture. Organs were weighed and processed as follows: typically, 1/3 of the spleen was ground in PBS for cfu determination, 1/3 of the spleen was ground in TRIzol (Invitrogen) for RNA extraction and 1/3 of the spleen was processed for flow cytometry. For SchuS4 infected mice, organs were ground in 1–2 ml PBS; IL-17A, IFN-γ and TNF-α protein levels were quantified by ELISA.
7 weeks-old recipient mice (B6.SJL-Ptprca Pepcb/BoyJ) (The Jackson Laboratories) which harbor the CD45.1 (Ptprca) marker were irradiated twice with 500 rad four hours apart and reconstituted with 2.106 bone-marrow cells from WT C57BL/6J mice, from IFNAR1−/− mice or with a mixture of 1 106 bone-marrow cells from WT B6.SJL-Ptprca Pepcb /BoyJ mice and 1 106 bone-marrow cells from IFNAR1−/− mice to create chimeric mice. 8 weeks post-transplantation, as analyzed by Ptprcb Prprca repartition, γδ T cells in the spleen were 83% from the donor and 17% from the recipient (range 78–90%). In the chimera situation, on average 53% of γδ T cells were WT and 47% IFN-α/βR−/− (range 49%–60% WT and 40%–51% IFNAR1−/−). 8 weeks post transplant, mice were injected i.d. with 1 105 F. novicida cfu and analyzed at 48h PI.
Single-cell suspensions from spleen were prepared and stained with the following antibodies: anti mouse neutrophils (7/4) from Serotec, TCR β chain (clone H57) from Biolegend, Gr-1 (RB6-8C5), F4/80 (BM8), CD3ε (145-2C11), TCRγδ (GL-3), CD8α (53–6.7), CD27 (LG.7F9), RORγt (AFKJS-9) from eBioscience, CD11b (M1/70), CD19 (1D3), CD4 (L3T4), IL-17A (TC11-18H10), CD45.1 (clone A20), CD45.2 (clone 104) from BD Biosciences and LIVE/DEAD® fixable dead cell stain kit (Invitrogen). For intracellular staining, 107 splenocytes per ml were incubated for 5h at 37C in RPMI-1640 with Glutamax, supplemented with non-essential amino-acids, sodium pyruvate, β-Mercapto-ethanol 50μM, gentamicin 10 μg/ml and monensin (eBioscience) before staining. After fixation and permeabilization with Perm/Wash buffer (BD bioscience), intracellular staining was performed using anti-IL-17A, anti-RORγt or rat IgG1, κ and rat IgG2a isotype controls. Cells were analyzed using an analytical LSR II flow cytometer (Becton Dickinson) at the Stanford FACS facility. Raw data were processed using FlowJo software (Tree Star, Inc.). Aggregates and dead cells were excluded based on FSC-H versus FSC-A plot and Live/Dead stain. Among the resulting live cell singlets, B cells were determined as CD3ε−, CD19+; T cells as CD19−, CD3ε+; γδ T cells as CD19−, CD3ε+, TCR αβ−, TCR γδ+; neutrophils as CD3ε−, CD19−, 7/4very high, Gr-1very high, CD11b+ and F4/80−; monocytes as CD3ε−, CD19−, 7/4high, Gr-1intermediate, CD11b+; macrophages as CD3ε−, CD19−, 7/4−, Gr-1intermediate, CD11b+ and F4/80+.
For cell culture, 106 cells were lysed in 1 ml TRIzol. For in vivo experiments, ~30 mg of organ were ground in 1mL TRIzol. RNA was extracted using chloroform, followed by 70% ethanol-mediated precipitation. RNA was then isolated using the RNeasy Mini kit (QIAGEN). Quantitative real-time RT-PCR was performed on a real-time detection system (7300 real time PCR system; Applied Biosystems) using rTth enzyme (Applied Biosystems) and SYBR green. Primers are described in supplemental table S1. Gene-specific transcript levels were normalized to the amount of β-actin or GAPDH mRNA and expressed as fold increase over the normalized level of the corresponding transcript in WT uninfected sample.
Bone-marrow derived macrophages (BMMs) were prepared from mice femurs, cultured and infected at the indicated multiplicity of infection (MOI) as previously described (23). Splenocytes were cultured in RPMI-1640 with Glutamax, supplemented with non-essential amino-acids, sodium pyruvate, β-Mercapto-ethanol 50μM, gentamicin 10 μg/ml at a cell density of 5×106 per ml. To test the role of rIFN-β on WT cells, splenocytes were isolated from 2 WT mice at 24h post i.d. injection of 105 F. novicida cfu. Total splenocytes (106 cells per well) or γδ T cells (105 cells per well) obtained by immuno-magnetic enrichement using TCRγδ+T Cell Isolation Kit (MACS-Miltenyi Biotec) were restimulated on anti-CD3ε coated plates (purified NA/LE hamster anti-mouse CD3ε antibody (BD Bioscience) coated at 10 μg/ml overnight in PBS) in the presence of various concentrations of rIFN-β (R&D systems). For the co-culture experiments, 5×104 BMM were infected for 30 min with F. novicida at a MOI of 10:1, and washed 3 times before addition of medium containing gentamicin at 10 μg/ml. 1.5 h later, 106 splenocytes from uninfected mice were added to the cells. When MACS-purified cells were used, splenocytes from WT or IFNAR1−/− mice infected for 24h were separated in 3 populations: CD19+ or Gr1+ cells, γδ-depleted T cells or γδ-enriched T cells using TCRγδ+T Cell Isolation Kit (MACS-Miltenyi Biotec). 105 MACS-purified cells were then added to 5×104 BMM as described above. In all cases, supernatant was collected at 20–24h PI and assayed for IL-17A by Elisa (R&D systems).
Blood was collected at 48h PI by intracardiac puncture. Serum concentrations for 23 analytes were determined at the Animal Resources and Laboratory Services Core facility of the PSW-RCE at UC Davis using the multiplex kit Bio-Plex by Bio-Rad (Hercules, CA). The kit was used per the manufacturers' instructions as previously described (24). Analytes serum concentrations for each mouse was divided by the average of the concentration of the corresponding cytokine in the group of WT uninfected mice, leading to expression of the analyte level as fold increase over WT uninfected level. Mice were ordered within each group based on increasing bacterial burden. A heat map was built using the Spotfire software with the help of Andrew Hoston.
Statistical analysis was performed using Prism 5 software (GraphPad). Mann-Whitney analysis was performed; two-tailed p values are presented. For the survival experiment, log rank (Mantel-Cox) test was performed.
We have previously shown that IFN-β is up-regulated in the skin and the spleen of infected mice 48h after intradermal (i.d.) injection of F. novicida (7). To better define the kinetics of IFN-β induction in the spleen following i.d. infection, we tracked IFN-β transcript induction by qRT-PCR throughout the course of infection (Fig. 1A). IFN-β induction was statistically significant at 24h post-infection (PI), reaching a 1000-fold induction over uninfected spleen at 48h PI (Fig. 1A). To assess the role of type I IFN in vivo, WT, IFNAR1+/−, and IFNAR1−/− mice were infected i.d. with 5×105 F. novicida cfu and their relative susceptibility was evaluated in a survival experiment (Fig. 1B). The median time to death of infected WT and IFNAR1+/− mice was 4.25 days and 84% of the mice within this group were dead by day 5. In contrast, the median time to death of IFNAR1−/− mice was greater than 10 days and 55% of the mice survived the infection. This experiment demonstrated that type I IFN is detrimental for the survival of the host during F. novicida infection. To investigate if the survival difference was due to a difference in bacterial burden, we quantified the bacterial colonization in the spleen of WT and IFNAR1−/− mice at various times PI. The bacterial burdens in the spleen of WT and IFNAR1−/− mice were similar until 48h PI (Fig. 1C). However, at 72h PI, the bacterial loads in WT spleen were 60-fold higher than in IFNAR1−/− spleen. A higher bacterial colonization was also observed in the liver of WT mice compared to the liver of IFNAR1−/− mice (data not shown). These results indicate that a type I IFN-mediated response that occurs between 48 and 72h PI is key to enhancing bacterial replication, ultimately leading to the death of WT mice. I.d. injection of a sublethal dose of F. novicida (103 cfu; supplemental Fig. S1) also showed a detrimental role for type I IFN signaling in controlling the bacterial burden indicating that this effect is independent of the infectious dose. To gain a better understanding of the nature of this response, we determined, in an unbiased way, the levels of a large panel of cytokines and chemokines in the serum of IFNAR1−/− and WT mice by a multiplex microbead immunoassay. We focused on the 48h PI time point because the WT and IFNAR1−/− mice contained the same levels of bacteria, thus, allowing us to investigate the early role of type I IFN signaling during infection without differences in the bacterial burden as a confounding factor. The heatmap corresponding to the fold increase in the serum levels of the 23 analytes over the corresponding levels in WT uninfected mice is shown in Fig. 2A, while the raw data are presented in supplemental information (Table S2). As expected, we did not detect any significant differences in serum cytokines levels between WT and IFNAR1−/− uninfected mice. Overall, this large scale analysis showed that the cytokine/chemokine pattern is very similar between F. novicida-infected IFNAR1−/− and WT mice (Fig. 2A). Particularly, we did not observe statistical differences in the levels of typical Th1 or Th2 cytokines (e.g. IL-2, IFN-γ, IL-4, IL-5, IL-13) between WT and IFNAR1−/− mice. At 48h PI, the concentrations of only 4 of the 23 cytokines/chemokines tested (IL-1α, IL12p40, IL12p70, IL-17A) were statistically different between WT and IFNAR1−/− mice. The serum levels of each of these differentially-regulated cytokines in each mouse are shown in Fig. 2B. A similar negative regulation of IL-12p70 by type I IFN has been observed during L. monocytogenes infection (15, 25) and in other experimental systems (26) indicating that at least some effects of type I IFN signaling are conserved among a large set of stimuli. The serum levels of IL-1α and IL-17A were significantly increased in infected IFNAR1−/− mice, while no significant increase was observed in infected WT mice (Fig. 2B). IL-17A is central to the recruitment and the expansion of critical innate immune cells (27). To our knowledge, there is no report of type I IFN regulating IL-17A levels during infections. Therefore, we decided to further investigate this finding.
To examine if the higher level of IL-17A in the serum of infected IFNAR1−/− mice correlated with an increase of the corresponding transcript in the spleen, we performed qRT-PCR on spleen extracts from WT or IFNAR1−/− mice (Fig 3A). We did not see any statistical difference in the basal level of IL-17A between uninfected WT and IFNAR1−/− animals. In agreement with the data for IL-17A protein, IL-17A transcript levels increased upon infection and this increase occurred with faster kinetics and greater magnitude in infected IFNAR1−/− mice compared to infected WT mice (Fig. 3A). Indeed, a significant increase in IL-17A transcripts in the spleen of IFNAR1−/− mice could be detected as soon as 24h PI while no difference was seen in IL-17A transcript levels in WT mice before 48h of infection. Furthermore, at 48h PI, IL-17A transcripts were induced 2000-fold in IFNAR1−/− mice compared to 300-fold in WT mice.
The IL-17 cytokine family comprises six isoforms. IL-17F is the closest homologue to IL-17A and has partially concordant expression with IL-17A (28). Type I IFN signaling during i.d. infection with F. novicida negatively regulated IL-17F transcript levels (Fig. 3B) similar to what we described for IL-17A. This regulation was specific since the levels of IFN-γ transcripts were the same in WT and IFNAR1−/− mice (Fig. 3C). IL-17A (Fig. 3D) and IL-17F (Fig. 3E) splenic transcript levels in WT mice correlated with the bacterial burden at 24h (data not shown) and 48h PI. We wanted to make sure that on a single mouse basis, the lower IL-17 mRNA levels observed in WT mice were due to specific negative regulation of IL-17 transcription and not due to differences in bacterial load. The levels of IL-17A and IL-17F transcripts as well as bacterial burden were determined in the spleen of each infected animal. The splenic levels of IL-17A (Fig. 3D) and IL-17F (Fig. 3E) transcripts were higher in IFNAR1−/− mice compared to WT mice, irrespective of bacterial load. This analysis on a single mouse basis clearly established that type I IFN negatively regulated IL-17A and IL-17F transcripts levels independently of the bacterial burden. To confirm that this difference in IL-17A transcript impacted IL-17A protein secretion by splenocytes, we determined IL-17A protein concentration by ELISA in the supernatants of splenocytes cultured ex-vivo for 24h. In agreement with the results at the mRNA level, splenocytes from IFNAR1−/− mice infected for 48h released IL-17A (149+/− 46 pg/ml) while splenocytes from WT infected mice or from uninfected mice did not release a detectable amount of IL-17A (Fig. 3F). Overall, our results demonstrate that type I IFN negatively regulate the transcript levels of IL-17A and IL-17F in the spleen, IL-17A protein levels in the serum as well as IL-17A production by splenocytes ex vivo.
To identify the cells that produce IL-17A in IFNAR1−/− mice during infection with F. novicida, we adapted a technique previously described by De Pascalis et al to identify IFN-γ-producing cells in vivo (29). Briefly, single cell suspensions were made from the spleens of infected mice, incubated ex vivo for 5h at 37°C with monensin, which blocks the secretion of cytokines, and analyzed for intracellular IL-17A expression by multiparameter flow cytometry. During this ex vivo period, restimulation occurs from cells infected in vivo and is thus more physiologically relevant than the standard 24h ex vivo restimulation using anti-CD3ε. Since we have previously shown that IL-17A levels were much higher in IFNAR1−/− mice than in WT mice infected with F. novicida, we performed this ex vivo analysis on splenocytes from IFNAR1−/− mice infected for 48h. We did not detect significant numbers of Gr-1+ cells, CD19+ cells or CD4+ cells (Th17 cells) that produced IL-17A (Supplemental Fig. S2 and Fig. 4A). The absence of Th17 cells is consistent with the early time of infection (before the onset of adaptive immunity) and with the study by Woolard et al. showing the appearance of Th17 cells at 10 days post-infection with LVS (30). In contrast, 20–50% of the γδ T cells isolated from IFNAR1−/− infected spleen at 48h PI were positive for IL-17A (Fig. 4A) indicating that γδ T cells represent the majority of the IL-17A-producing cells in IFNAR1−/− F. novicida-infected spleens. In agreement with what Ribot et al (31) have observed in WT mice, the IFNAR1−/− IL-17A+ γδ T cells were CD27− (Fig. 4B). We then performed intracellular IL-17A staining on splenocytes from WT and IFNAR1−/− infected and uninfected mice focusing on the γδ T cell subset. We did not detect any IL-17A+ γδ T cells in WT mice infected for 24h or in uninfected mice, although there was a small increase (~ 0.5 %) in IL-17A+ γδ T cells in WT mice at 48h PI (Fig. 4C and D). In contrast, we saw a large increase in the percentage of γδ T cells producing IL-17A (9% of the total γδ T cells) in IFNAR1−/− mice infected for 24h (Fig. 4D). Furthermore, at 48h PI, 50 times more IL-17A+ γδ T cells were detected in the spleen of IFNAR1−/− mice compared to WT mice (Fig. 4C and D). The same type I IFN-mediated control of IL-17A+ γδ T cells was also observed in the liver after a 4h ex vivo restimulation with Phorbol 12-myristate 13-acetate (PMA) and a calcium ionophore (32) (supplemental information Fig. S3). We also detected more IL-17A+ γδ T cells in the spleen of infected IFNAR1−/− mice than in the spleen of infected WT mice following I.V. injection of brefeldin A and in the absence of any ex vivo restimulation (33) (Supplemental information Fig. S4). Although this technique lead to the identification of a lower number of IL-17A+ cells when compared to the ex-vivo restimulation, it indicated to us that this latter step was not responsible for the difference observed between WT and IFNAR1−/− mice. All together, these results indicate that type I IFN signaling constrains the number of IL-17A+ γδ T cells in the spleen and in the liver. Moreover, the expression of IL-17A in this particular subset was much higher in IFNAR1−/− mice (mean fluorescence intensity (MFI) =2346) than in WT mice (MFI= 981) suggesting that type I IFN signaling negatively regulates both the number of IL-17A-expressing cells and the IL-17A expression level in those cells. Type I IFN-mediated regulation of IL-17A in γδ T cells was also statistically significant upon i.d. injection of 103 F. novicida indicating that this phenotype is robust over a large range of inocula (supplemental information Fig. S1).
All the above results on type I IFN signaling were obtained by comparing WT and IFNAR1−/− mice. To demonstrate directly a role for type I IFN in IL-17A regulation, splenocytes and γδ T cells from infected WT mice were treated ex vivo with rIFN-β. Importantly, addition of rIFN-β led to a dose-dependent decrease of the IL-17A secretion obtained by restimulation with anti-CD3ε antibody (Fig. 4E).
Type I IFN have been reported to sensitize lymphocytes to apoptosis during L. monocytogenes infection (25) and more TUNEL-positive cells are observed in the spleen of F. novicida-infected WT mice than in the spleen of F. novicida-infected IFNAR1−/− mice (7). Thus, a possible explanation for the lower number of IL-17-producing γδ T cells in WT mice is that these cells are dying in an IFNAR1-dependent manner. Although we observed a decrease in the number of γδ T cells during F. novicida infection, the number of splenic γδ T cells in infected mice was not statistically different between WT and IFNAR1−/− mice (Fig. 4F). Therefore, our data suggest that the difference in the number of IL-17A+ γδ T cells between WT and IFNAR1−/− mice during F. novicida infection is not due to a difference in cell death.
Type I IFN receptor is expressed ubiquitously and while the previous experiments clearly established that type I IFN signaling results in decreased IL-17A production by γδ T cells during infection, it was unknown if this decrease resulted from a direct signaling in γδ T cells or from infected phagocytes. To address this issue, we set up an in vitro assay in which we co-cultured macrophages and splenocytes from either WT or IFNAR1−/− mice. Briefly, 5×104 bone marrow-derived macrophages (BMM) from either WT or IFNAR1−/− mice were infected for 2h before addition of 106 splenocytes isolated from either uninfected WT or IFNAR1−/− mice. This in vitro assay was performed in the presence of gentamicin in the medium to ensure that only the BMM were infected. At 24h PI, the IL-17A concentration in the supernatant was determined by ELISA. Importantly, using this in vitro system, we were able to recapitulate the IL-17A difference seen in vivo (Fig. 5A). Indeed, WT splenocytes incubated with infected WT BMM released much less IL-17A over a 24h period than IFNAR1−/− splenocytes incubated with infected IFNAR1−/− BMM (Fig. 5A). We could not detect any IL-17A release in the absence of splenocytes or when macrophages were left uninfected (Fig. 5A and not shown). To address the role of type I IFN signaling in infected macrophages, we then added splenocytes from uninfected WT mice to either WT or IFNAR1−/− BMM, infected with F. novicida. WT splenocytes released significantly less IL-17A (3-fold) when co-incubated with infected WT BMM compared to when they were co-incubated with infected IFNAR1−/− BMM (Fig. 5B). Although this experiment demonstrated that type I IFN signaling in macrophages plays a role in the negative regulation of IL-17A, there was a far greater reduction in secreted IL-17A levels when both splenocytes and BMM from WT mice were used. Indeed, WT splenocytes stimulated with infected WT BMM secreted ~20-fold less IL-17A compared to IFNAR1−/− splenocytes stimulated with infected IFNAR1−/− BMM (Fig. 5A). This suggested to us that type I IFN signaling in uninfected splenocytes was mostly responsible for the inhibition of IL-17A release. Consistent with this hypothesis, IFNAR1−/− splenocytes co-incubated with infected WT BMM secreted 18 times more IL-17A than WT splenocytes co-incubated with infected WT BMM (Fig. 5C). Collectively, our in vitro data suggest that type I IFN signaling in uninfected splenocytes is largely responsible for the inhibition of IL-17A secretion.
Since the γδ T cells were the main cells secreting IL-17A (Fig. 4), we decided to further investigate if type I IFN signaling specifically in γδ T cells was responsible for the IL-17A inhibition. Splenocytes from WT or IFNAR1−/− mice infected for 24h with F. novicida were separated into 3 populations (Gr1+ or CD19+ cells; Gr1−, CD19−, γδ− cells and Gr1−, CD19−, γδ+ T cells) by magnetic sorting using TCR γδ+ T cell isolation kit (Miltenyi Biotech). 105 γδ+ T cells were added ex vivo to 5χ104 IFNAR1−/− BMM that had been infected with F. novicida at a MOI of 10:1 for 1h. IL17A release was determined at 24h PI. γδ-enriched T cells from WT mice produced basal to undetectable levels of IL-17A (≤ 15 pg/ml; Fig. 5D). In contrast, γδ-enriched T cells from IFNAR1−/− mice secreted 260 pg/ml of IL-17A (Fig. 5D). Taken together, our in vitro data indicate that γδ T cells are the main cell type producing IL-17A during F. novicida infection and suggest that type I IFN signaling in γδ T cells plays a major role in the inhibition of IL-17A production during infection. However, a preliminary in vivo experiment using chimeric mice presenting a mixture of WT and IFNAR1−/− bone-marrow derived cells suggests that in vivo type I IFN signaling in WT cells has a dominant effect to inhibit IL-17A production by γδ T cells independently of the presence or the absence of the type I IFN receptor on this particular subset (Fig. 6). This preliminary result indicates that our in vitro system may not be fully representative of the in vivo situation, in which the relative contribution of type I IFN signaling in all the different cell types on the IL-17A inhibition remains to be determined.
Recently, Guo et al found that IL-27, a potent inhibitor of IL-17 production by Th17 cells (34, 35), is secreted in an IFNAR1-dependent manner by macrophages (36). To investigate if IL-27 is regulated by type I IFN in F. novicida-infected macrophages, we performed qRT-PCR for this cytokine. IL-27 was strongly up-regulated in WT BMM upon infection (more than 30 fold at 12h PI) (Fig. 7A). In contrast, no induction was seen in IFNAR1−/− BMM indicating that in F. novicida-infected macrophages, IL-27 is induced in an IFNAR1-dependent manner. This result suggests that the inhibitory effect of type I IFN signaling in macrophages on IL-17A production by γδ T cells might partially rely on controlling the secretion of this inhibitory cytokine. Both IL-27 and type I IFN signal through Stat1 in T cells. Stat1 is thought to regulate both directly and indirectly the level of the transcription factor retinoic orphan receptor γt (RORγt) (37), a T cell-specific transcription factor critical for IL-17A/F expression. Accordingly, RORγt transcripts level was 2-fold higher in IFNAR1−/− mice at 48h PI compared to WT infected mice (Fig. 7B). In contrast, the mRNA level of another T cell-specific transcription factor not involved in IL-17A/F regulation, Forkhead Transcription Factor (FoxP3), was not affected by type I IFN signaling during infection (Fig. 7C). Furthermore, intracellular detection of RORγt showed a large population of RORγt+ cells in CD27− γδ T cells from infected IFNAR1−/− mice but neither from infected WT mice nor from uninfected mice (Fig. 7D, 7E). Collectively, our data indicate that during i.d. F. novicida infection, type I IFN negatively regulates IL-17A production by CD27− γδ T cells. This negative regulation is due to signaling in macrophages possibly through the positive regulation of IL-27 (Fig. 7) as well as by a possible direct signaling in γδ T cells (Fig. 5). Ultimately, type I IFN signaling results in down-regulation of RORγt, a critical transcription factor controlling IL-17 production.
IL-17A and F are potent cytokines for the mobilization/recruitment and activation of granulocytes (38). We thus wondered if the high level of IL-17 during infection in IFNAR1−/− mice was impacting the cell composition of the spleen at 48h PI (Fig. 8A). Indeed, the proportion of neutrophils in infected IFNAR1−/− mice was significantly higher (2-fold) than the numbers in infected WT mice as early as 24h PI compared to uninfected mice, which correlates with the elevated IL-17A transcript levels in IFNAR1−/− mice compared to WT mice at 24h (Fig. 3). By 48h PI, there was an even larger difference in the number of neutrophils in the spleens from infected IFNAR1−/− mice compared to WT mice (Fig. 8B and Fig. S1F). For example, spleens from infected IFNAR1−/− mice contained 1.5×106 neutrophils while spleens from infected WT mice contained 0.6×106 neutrophils (Fig. 8A, B). The same trend was observed for CD11b+ macrophages (1.3×105 in IFNAR1−/− infected mice compared to 5×104 macrophages in WT infected). These results established that type I IFN signaling negatively regulates granulocyte influx/replication/cell survival in the spleen upon infection, which correlates with the negative regulation of IL-17A. Thus, we tested the contribution of IL-17A to the increase in splenic neutrophil number upon i.d. infection using IL-17A-deficient mice. At 48h PI, we saw a significant reduction in the number of neutrophils in the spleens of IL-17A-deficient mice compared to WT mice (Fig. 8C), suggesting that IL-17A up-regulation contributes to the increase in the number of splenic neutrophils during infection with F. novicida. However, since uninfected IL-17A-deficient mice contained fewer splenic neutrophils (Fig. 8C), we cannot exclude that the difference observed at 48h PI is due to this initial difference. Taken together, our results demonstrate that type I IFN signaling leads to a decrease in IL-17A production by γδ T cells. This decrease in IL-17A in WT mice compared to IFNAR1−/− mice correlates with a decrease in splenic neutrophil numbers, similar to what is seen in IL-17A-deficient mice.
Type I IFN is secreted in mice infected with numerous intracellular bacterial pathogens that have been inoculated via various routes of infection (7, 9, 15, 17, 18, 25). Moreover, the route of LVS inoculation (e.g. intranasal-i.n. or i.d.) influences Th17 numbers in the lungs of infected mice (30). Thus, we decided to investigate if the same type I IFN-mediated regulation of IL-17A could be observed during infections with other bacteria and/or using different routes of inoculation. To test the route of inoculation, we first inoculated WT or IFNAR1−/− mice with 50 F. novicida cfu. Similarly to what was observed during i.d. inoculation, type I IFN signaling was detrimental both for the survival of the mice (Fig. 9A) and the control of the bacterial burden (Fig. 9B) following i.n. inoculation. The resistance of the IFNAR1−/− mice correlated with an increase in IL-17A transcript in the spleen (Fig. 9C), a large increase in the number of IL-17A+ γδ T cells (Fig. 9D, 9E) and an expansion of the neutrophil population in the spleen (Fig. 9F).
Francisella tularensis subspecies tularensis is highly pathogenic in humans and has been classified as a class A bioterrorism agent by the Center for Disease Control and Prevention due to its very low lethal dose upon i.n. inoculation (39). To test whether type I IFN were induced and whether type I IFN signaling also negatively regulates IL-17A secretion during intranasal infection of mice with F. tularensis strain SchuS4, WT or IFNAR1−/− BALB/cJ mice were inoculated with 25 CFU i.n.. Spleens (Fig. 10A), lungs (Fig. 10B) and livers (not shown) from WT and IFNAR1−/− mice contained similar bacterial burdens at 48h and 96h PI. Bacterial counts in the spleen were low to undetectable at 48h PI. At 96 h PI, IFN-β Transcript levels in infected WT spleen (Fig. 10C) and lung (Fig. 10D) were 50- and 1000-fold higher than the levels in uninfected mice, respectively. To determine whether type I IFN signaling resulted in negative regulation of IL-17A in BALB/cJ mice similar to what we had seen in F. novicida-infected C57BL/6J mice, IL-17A transcript levels were measured in the lung, spleen and liver of mice infected for 96h PI. Indeed, the level of IL-17A mRNA in the spleen of infected IFNAR1−/− mice was 34-fold higher than the level in infected WT mice (Fig. 10E). We saw a similar increase in the level of IL-17A protein in the spleen of IFNAR1−/− mice at 96h (Fig. 10G), while the levels of IFN-γ (Fig. 10H) and TNF-α (Fig. 10I) proteins were similar in the spleen of WT and IFNAR1−/− mice. In contrast, IL-17A mRNA and IL-17A protein in IFNAR1−/− lungs (Fig. 10F and Fig. 10J) and livers (data not shown) were not significantly different than the levels in WT mice. Further experiments are required to demonstrate that this decrease in splenic IL-17A production is due to decrease IL-17A secretion by γδ T cells. As specified above, the experiments performed with the SchuS4 strain were obtained in a different mouse background, BALB/cJ, than the F. novicida experiments, C57BL/6J. Our data demonstrate that Type I IFN are regulators of IL-17A production in the spleen regardless of the mouse background, F. tularensis subspecies or the route of inoculation.
F. novicida and F. tularensis are very similar at the genetic level (20). To exclude any possibility, that the type I IFN-mediated IL-17A regulation was due to the presence of F. tularensis–specific factors, we investigated the potential role of type I IFN signaling in IL-17A secretion during L. monocytogenes infection. 1.4×104 cfu were injected IV and at 48h PI, animals were euthanized and the spleens were processed for bacterial colonization, qRT-PCR and flow cytometry analysis. As previously described (15, 16, 25), IFNAR1−/− mice contained 10-fold lower levels of bacteria in the spleen compared to WT mice at 48h PI (Fig. 11A). Despite this difference in the bacterial burden, spleens from infected IFNAR1−/− mice contained ~6-fold higher levels of IL-17A transcript compared to WT infected mice (Fig. 11B; 1300- and 230-fold increase over uninfected, respectively). In contrast, the levels of IFN-γ Transcript in WT infected mice were significantly higher than in infected IFNAR1−/− mice (Fig. 11C). This difference likely reflects the difference in bacterial burden between the two mouse strains (Fig. 11A). Similar to mice infected with F. novicida, and in agreement with the results observed in the liver by Hamada S. et al (32), γδ T cells produced the majority of the IL-17A in the spleens of mice infected with L. monocytogenes (supplemental information Fig. S5). Indeed, 17% of the total γδ T cells in IFNAR1−/− mice infected with L. monocytogenes produced IL-17A (Fig. 11D, 11E). In contrast, less than 2% of γδ T cells in WT infected mice produced IL-17A (Fig. 11D, 11E). The difference between the IFNAR1−/− and WT mice might be partially due to cell death of IL-17A+ γδ T cells in WT mice, since there were fewer γδ T cells in WT mice than in IFNAR1−/− mice (Fig. 11F). Finally, the higher levels of IL-17A in infected IFNAR1−/− mice correlated with higher numbers of splenic neutrophils in infected IFNAR1−/− mice compared to infected WT mice (Fig. 11G). These results are similar to the type I IFN-dependent phenotypes observed during F. novicida infection and demonstrate that during infection with intracellular bacteria, type I IFN negatively regulate the number of IL-17A+ γδ T cells and of splenic neutrophils.
Type I IFN have recently emerged as a major class of cytokines secreted by phagocytes in response to infection with bacteria (5–8, 11). However, the role of type I IFN during bacterial infection remains unclear due in part to the pleiotropic effects of these cytokines and the diversity of target cells. Here, we investigated the effect of type I IFN signaling during i.d. infection with F. novicida and extended our studies to include infections with F. tularensis and L. monocytogenes using multiple routes of infection. We identified IL-17A/F as specific cytokines that are negatively regulated by type I IFN signaling during infections with intracellular bacteria. To our knowledge, this is the first report of type I IFN signaling regulating IL-17A/F during infection.
There are six members in the IL-17 family: IL-17A, IL-17B, IL-17C, IL-17D, IL-17E and IL-17F (40). Many studies have focused on IL-17A and IL-17F and these 2 family members are the most closely related (28). Multiple cell types including CD4+ αβ T cells, γδ T cells, NK cells, and neutrophils have been shown to produce IL-17A (41). IL-17A plays a central role in autoimmune inflammation and in innate immunity to bacterial pathogens (41). We show here that the increased resistance of IFNAR1−/− mice to infections with F. novicida, F. tularensis and L. monocytogenes was associated with a specific increase in IL-17 A/F and a corresponding expansion of an IL-17A+ γδ T cell population. γδ T cells are the major cell subset producing IL-17 in vivo in response to several infections including M. bovis BCG (42), L. monocytogenes (32), and Salmonella enterica (43). One possible explanation for the increased susceptibility of WT mice compared to IFNAR1−/− mice to infection is due to a type I IFN-dependent cell death of lymphocytes that is triggered by the L. monocytogenes-specific virulence factor Listeriolysin O (44). However, we did not observe a difference in γδ T cell number between WT and IFNAR1−/− mice infected with F. novicida, suggesting that the difference in IL-17 is not due to differential apoptosis of T cells in WT mice compared to IFNAR1−/− mice.
The role of IL-17A in the innate immune response to controlling bacterial infections has been associated with its capacity to induce antimicrobial peptides and to its potent role to increase neutrophil number and microbicidal activity (27, 45–47). Indeed the increased levels of IL-17A during infections of LFA1−/− (48) and IFNAR1−/− (this study) mice with L. monocytogenes correlates with increased survival rates. This suggests that an increase in IL-17A, a cytokine important to control L. monocytogenes infection in the liver (32), could at least partially, mediate the resistance of LFA1−/− and IFNAR1−/− mice. While neutrophils are important for protecting the host against L. monocytogenes infection (49), their role during Francisella infection is still controversial (50). Our results indicate that an early increase in splenic neutrophil numbers is observed in IFNAR1−/− mice and that this correlates with an increased resistance to F. novicida infection. Further investigation is required to examine if this increased resistance is due to the increase in neutrophil numbers, the increased number of IL-17A+ γδ T cells (which probably secrete other cytokines beside IL-17A such as IL-17F, IL-21 or IL-22), or the increase in IL-17A. Nevertheless, the regulation of IL-17 production is important for an optimal innate immune response and protection against bacterial pathogens. Thus, we next focused on the mechanisms of the type I IFN-dependent regulation of IL-17 during infection.
The regulation of IL-17 production has primarily been studied in the Th17 lineage of CD4+ T cells. Most of these studies have focused on the involvement of a complex cytokine network including IL-6, IL-23p19, TGF-β, IL-1, IL-10, IL-27, IFN-β and IFN-γ in the regulation of Th17 differentiation (36, 51). IL-10, a negative regulator of IL-17 production (52), levels increase during L. monocytogenes infection in a type I IFN-dependent manner (14). Thus, IL-10 could play a role in the type I IFN-mediated negative regulation of IL-17A/F that we observed during bacterial infections. However, IL-10 protein and transcript levels were similar in WT and IFNAR1−/− mice at 48h PI with F. novicida (supplemental information Fig. S6A, B). Furthermore, infection of IL-10-deficient mice with F. novicida resulted in a minor, but statistically insignificant, up-regulation of IL-17A transcript (supplemental information Fig. S6C). Taken together, our data indicate that IL-10 does not play a major role in the early regulation of IL-17 during F. novicida infection.
We used an in vitro reconstitution system to better understand the cell types and signaling pathways involved in the type I IFN-dependent regulation of IL-17 during bacterial infection. We demonstrated that in vitro, type I IFN-dependent signaling in macrophages and in T cells contribute to the control of IL-17A levels (Fig. 5). IL-27 is a cytokine that inhibits the development of Th17 cells (34, 35) and type I IFN signaling regulates Th17 development in vivo during multiple sclerosis through regulation of IL-27 (36). We show here that macrophages infected with F. novicida released increased levels of IL-27 in a Type I IFN-dependent manner (Fig. 7A). Therefore, the production of IL-27 by macrophages likely contributes to the type I IFN-dependent constraint of IL-17 production. However, we also investigated the contribution of type I IFN signaling in T cells to the constraint of IL-17 production. We found in vitro that some effect of type I IFN signaling on IL-17 release comes from direct signaling in γδ T cells although this remains to be established in vivo (Fig. 5, ,6).6). The transcription factor RORγt controls development of Th17 CD4+ αβ T cells and of IL-17 producing γδ T cells in the intestinal lamina propria (53). Type I IFN signaling is associated with a specific decrease in RORγt transcript levels (Fig. 7B). Taken together, our data indicate that Type I IFN signaling results in negative regulation of IL-17A/F production through a modulation of the key transcription regulator RORγt in CD27− γδ T cells.
Collectively, our results shed new light on the role of type I IFN signaling during intracellular bacterial infection. We believe the regulation of the IL-17A+ γδ T cell population is partially responsible for the resistance phenotype observed in IFNAR1−/− mice infected with intracellular bacteria. Type I IFN have a major anti-inflammatory role during auto-immune diseases (36). Similarly, we propose that, by negatively regulating the number of IL-17A+ γδ T cells and of neutrophils in the spleen, type I IFN levels influence the balance between a host-beneficial response (anti-bacterial inflammation) and a host deleterious response (excessive inflammation associated with tissue damage). A better understanding of how this balance is maintained may lead to treatments that optimize the inflammatory response to eliminate pathogens and still protect organs from tissue injury.
We are grateful to Paul Luciw and the Animal Resources and Laboratory Services Core of the PSW-RCE at UC Davis for the multiplex analysis, to Andrew Hoston and Gary Nolan for the help with Spotfire software, to Jeffery Margolis for critical reading of the manuscript, to Patrik Viatour and Dennis Levenson-Gower for sharing their expertise on liver cells, and to Hanza Mathew for technical assistance. We thank Jeff Cox, Anita Sil, Joe DeRisi, Russell Vance, Greg Barton, Dan Portnoy and others members of the PO1 group for stimulating discussions.
This work was supported by grants AI 063302 and AI065359 from the NIH-NIAID to DMM, NIH grant PO1 AI 56320 to DWM and an American Lung Association Senior Research Training Fellowship to GSK. The multiplex analysis was possible thanks to the RCE grant AI065359 to Paul Luciw and the Animal Resources and Laboratory Services Core of the PSW-RCE at UC Davis. T.H. was supported by a fellowship from the European Molecular Biology Organization.
Publisher's Disclaimer: “This is an author-produced version of a manuscript accepted for publication in The Journal of Immunology (The JI). The American Association of Immunologists, Inc. (AAI), publisher of The JI, holds the copyright to this manuscript. This version of the manuscript has not yet been copyedited or subjected to editorial proofreading by The JI; hence, it may differ from the final version published in The JI (online and in print). AAI (The JI) is not liable for errors or omissions in this author-produced version of the manuscript or in any version derived from it by the U.S. National Institutes of Health or any other third party. The final, citable version of record can be found at www.jimmunol.org.”