Autophagy activation by interferon-γ via the p38 mitogen-activated protein kinase signalling pathway is involved in macrophage bactericidal activity

Abstract

Macrophages are involved in many essential immune functions. Their role in cell-autonomous innate immunity is reinforced by interferon-γ (IFN-γ), which is mainly secreted by proliferating type 1 T helper cells and natural killer cells. Previously, we showed that IFN-γ activates autophagy via p38 mitogen-activated protein kinase (p38 MAPK), but the biological importance of this signalling pathway has not been clear. Here, we found that macrophage bactericidal activity increased by 4 hr after IFN-γ stimulation. Inducible nitric oxide synthase (NOS2) is a major downstream effector of the Janus kinase–signal transducer and activator of transcription 1 signalling pathway that contributes to macrophage bactericidal activity via nitric oxide (NO) generation. However, no NO generation was observed after 4 hr of IFN-γ stimulation, and macrophage bactericidal activity at early stages after IFN-γ stimulation was not affected by the NOS inhibitors, NG-methyl-l-arginine acetate salt and diphenyleneiodonium chloride. These results suggest that an NOS2-independent signalling pathway is involved in IFN-γ-mediated bactericidal activity. We also found that this macrophage activity was attenuated by the addition of the p38 MAPK inhibitors, PD 169316, SB 202190, and SB 203580, or by the expression of short hairpin RNA against p38α or the essential factors for autophagy, Atg5 and Atg7. Collectively, our results suggest that the IFN-γ-mediated autophagy via p38 MAPK, without the involvement of NOS2, also contributes to the ability of macrophages to kill intracellular bacteria. These observations provide direct evidence that p38 MAPK-mediated autophagy can support IFN-γ-mediated cell-autonomous innate immunity.

Introduction

Macrophage participates in many essential functions such as tissue homeostasis and clearance of infection.^1,2 The type II interferon (IFN), IFN-γ, a cytokine that is mainly secreted by activated T helper type 1 (Th1) T lymphocytes and natural killer (NK) cells, is a powerful macrophage activator.^3,4 Interferon-γ induces tyrosine phosphorylation of the signal transducer and activator of transcription 1 (STAT1) protein via Janus kinase 1 (JAK1) and JAK2.^5–7 Subsequently, STAT1 dimers bind to IFN-stimulated response elements and induce the transcription of many IFN-stimulated genes.^5,6 In addition to the JAK–STAT1 pathway, expression of another gene is mediated by IFN-γ via STAT1-independent pathways and modulates various signalling cascades, including those involving myeloid differentiating factor 88 (MyD88), p38 mitogen-activated protein kinase (MAPK), phosphatidylinositol 3-kinase (PI3K), and protein kinase C.^7–14 The major immune factors in the signalling pathway downstream of IFN-γ are MHC class I and II; inflammatory and pyrogenic cytokines; chemokines; and antimicrobial proteins, such as inducible nitric oxide synthase (NOS2; also known as iNOS), phagocyte oxidase, and immune guanosine triphosphatases (GTPases).^{5–7,15–18}

Notably, IFN-γ is critical for cell-autonomous innate immunity against intracellular bacteria, such as Listeria, Mycobacteria and Salmonella.^17,19–21 The antimicrobial enzyme NOS2 is largely considered responsible for the bactericidal activity, via nitric oxide (NO) generation, of IFN-γ-activated macrophages against intracellular bacteria.^22,23 However, NOS2-knockout (NOS2^−/−) mice that had been infected with virulent Mycobacterium tuberculosis have been found to survive significantly longer and exhibit some control of lung M. tuberculosis growth when compared with mice lacking IFN-γ or IFN-γ receptor.²⁴ This observation suggested that IFN-γ-dependent, NOS2-independent immunity against intracellular bacteria exists. Recently, it has been shown that, in addition to NOS2, IFN-γ-inducible immune GTPases, including p47 immunity-related GTPases (p47 IRGs) and p65 guanylate-binding proteins (p65 GBPs), regulate autophagy and contribute to the disposal of intracellular pathogens.^{17,18,20,24–28}

Autophagy has emerged as a major immune defence pathway and this cascade can be provoked by host-derived cytokines, IFN-γ, or pattern recognition receptors, including Toll-like receptors and nucleotide-binding oligomerization domain-like receptors.^{25,26,29–35} It has been shown that IFN-γ controls autophagy via several types of IFN-γ-inducible immune GTPases belonging to the IRG family and the GBP family.^{18,25–28,36,37} More recently, we have shown that, in addition to the IFN-inducible GTPase pathway, the p38 MAPK pathway contributes to autophagy activation in the IFN-γ-stimulating cells.³⁸ Interferon-γ is able to activate autophagy through at least two different pathways, the conventional STAT1- and Irgm1-dependent pathway and an alternative p38 MAPK-dependent, STAT1-independent pathway. However, the biological role of IFN-γ-induced autophagy via p38 MAPK remains unclear.

In this study, we demonstrated that macrophage bactericidal activity increased at 4 hr after IFN-γ stimulation in an STAT1- and NOS2-independent manner. Furthermore, this macrophage bactericidal activity that occurred early after IFN-γ stimulation was attenuated by the inhibition of p38 MAPK or autophagic function. These results suggest that the autophagy mediated by p38 MAPK, without the influence of NOS2, also contributes to the ability of macrophages to kill intracellular bacteria. To our knowledge, this study is the first to document that p38 MAPK-mediated autophagy can activate IFN-γ-mediated cell-autonomous innate immunity.

Materials and methods

Reagents

Recombinant mouse IFN-γ was purchased from R&D Systems (Minneapolis, MN) and used at a concentration of 200 U/ml. NG-methyl-l-arginine acetate salt (l-NMMA) and diphenyleneiodonium chloride (DPI) were obtained from Sigma (St Louis, MO) and used at a concentration of 500 or 10 μm. PD 169316 and SB 202190 were obtained from Cayman (Ann Arbor, MI) and used at a concentration of 10 μm. SB 203580 was bought from Calbiochem (Darmstadt, Germany) and used at a concentration of 5 μm.

Mammalian cell culture

RAW 264.7 cells were obtained from the American Type Culture Collection (Manassas, VA) and maintained in RPMI-1640 medium containing 10% fetal bovine serum, 10 mm HEPES and 1 mm sodium pyruvate. The primary bone-marrow derived macrophages (BMMs) were generated from C57BL/6 mice, as reported previously.³⁸ The lentiviral vectors used for expressing short hairpin (sh)RNA against IFN-γR1, STAT1 and Atg7 have been described previously.³⁸ The plasmids for expressing shRNA as a non-target control and for expressing shRNA against Atg5 or p38α were constructed using pLKD.neo³⁸ and the Addgene pLKO.1 protocol (http://www.addgene.org). The RNAi sequences were as follows: for the non-target shRNA control, 5′-CAACAAGATGAAGAGCACCAA-3′; for Atg5, 5′-GCAGAACCATACTATTTGCTT-3′; for p38α, 5′-CCTCTTGTTGAAAGATTCCTT-3′. The ViraPower Lentiviral Expression system (Invitrogen, Carlsbad, CA) was used to co-transfect the viral vector into 293FT (Invitrogen) to produce lentiviruses. The resulting viral supernatant was used for the transfection of RAW 264.7 cells or BMMs, and then stable knockdown (KD) cells were selected with G418 (BD Clontech, Palo Alto, CA).

Bacterial culture

Listeria monocytogenes EGD (serovar 1/2a) was a generous gift from Dr Masao Mitsuyama (Kyoto University Graduate School of Medicine, Kyoto, Japan). It was grown overnight in brain–heart infusion broth (BD Biosciences, Sparks, MD) at 37° and shaken. Listeria monocytogenes cells were washed with RPMI-1640 medium once and used in an infection assay. Salmonella enterica serovar Typhimurium (RIMD1985009) was provided by the Research Institute for Microbial Diseases, Osaka University (Osaka, Japan), and was grown overnight in Luria–Bertani broth (Sigma).

Measurement of bacterial growth in the macrophages

RAW 264.7 cells or BMMs were infected with L. monocytogenes for 1 hr at a multiplicity of infection (MOI) of 5 or with S. typhimurium for 10 min at an MOI of 10, and then the cells were washed three times with PBS. Following this, the cell culture medium was changed to new RPMI-1640 containing 50 μg/ml of gentamicin (Sigma) to exclude the bacteria, which were not taken up by the macrophages. After 0 or 4 hr of Listeria infection, or after 1 or 4 hr of Salmonella infection, the macrophages were lysed with 0·1% Triton X-100, and living bacteria were quantified by the colony-forming unit (CFU) method. The growth rate of bacteria in the cells over a 4-hr period was calculated. For each time-point, counts were obtained from three independent experiments.

Measurement of NO production

Levels of NO were measured as the accumulation of nitrite in the cell culture medium. The nitrite level was determined spectrophotometrically with Griess reagent (Sigma). RAW264.7 cells or BMMs were treated with or without 200 U/ml IFN-γ for 4 or 24 hr. Briefly, 250 μl of cell culture supernatant was mixed with an equal volume of Griess reagent. Following incubation for 10 min, the absorbance at 550 nm was measured, and values were quantified against a standard curve of sodium nitrite.

Western blotting

The macrophages were lysed in cell lysis buffer (50 mm Tris, pH 7·5, 1% Triton X-100 and 150 mm NaCl) plus protease inhibitor mixture (Roche, Mannheim, Germany), and centrifuged at 16 400 g for 20 min. The supernatants were used as cell lysates, and were subjected to SDS–PAGE before transferring to PVDF membranes. Western blotting was carried out with the following antibodies: anti-IFN-γR polyclonal antibody (pAb) and anti-Atg7 pAb (GeneTex, Irvine, CA), anti-STAT1 pAb (GenScript, Piscataway, NJ), anti-GAPDH monoclonal antibody (mAb) (6C5; Santa Cruz Biotechnology, Santa Cruz, CA), anti-phospho-p38, p38, Atg5 pAb (Cell Signaling Technology, Danvers, MA), ant-LC3 mAb (2G6; Nano Tools, Hamburg, Germany), horseradish peroxidase-conjugated anti-mouse IgG antibody and horseradish peroxidase-conjugated anti-rabbit IgG antibody (Jackson ImmunoResearch Laboratories, West Grove, PA). The expression of GAPDH was used as the internal control. The intensity of bands was quantified using ImageJ software (National Institutes of Health, Bethesda, MD).

Statistical analysis

Statistical analyses for differences between group means were analysed using Student's t-test. All metrics are given as mean ± standard deviation values.

Results

Macrophage bactericidal activity increased at 4 hr after IFN-γ stimulation in an IFN-γ R1-dependent, but STAT1-independent manner

Previously, we have demonstrated that autophagy, which plays key roles in the host defence pathway, is activated by IFN-γ stimulation for 4 hr.³⁸ In this study, we analysed the function of macrophages that had been stimulated with IFN-γ for 4 hr, to elucidate the biological significance of IFN-γ-activated autophagy via p38 MAPK. Macrophage bactericidal activity is potentiated by IFN-γ stimulation and this effect is predominantly guided by NO, which is produced from arginine by IFN-inducible NOS2.^22,39 In mouse macrophage-like RAW 264.7 cells, the expression of NOS2 mRNA was observed at 5 hr after IFN-γ (100 U/ml) stimulation, and NO production remained low until 10 hr after IFN-γ treatment.^40,41 Moreover, our result also showed that NO was barely produced at 4 hr after IFN-γ stimulation (Fig. 1a). However, an elevation in bactericidal activity against both intracytosolic (L. monocytogenes) and intravacuolar (S. typhimurium) pathogens was observed in the macrophages that had been treated with IFN-γ for 4 hr (Fig. 1b,c). This kind of IFN-γ-mediated bactericidal activity was inhibited in IFN-γ receptor 1 KD cells, but not in STAT1-KD cells (Fig. 2). These results suggested that bactericidal activity is reinforced at an early stage of IFN-γ stimulation in a STAT1- and NO-independent manner. Unlike NO production, autophagy activation has previously been observed in RAW 264.7 cells and BMMs at 4 hr after IFN-γ stimulation,³⁸ we therefore expected that autophagy may contribute to bactericidal activity at an early stage of IFN-γ stimulation.

Figure 1.

Macrophage bactericidal activity increased on interferon-γ (IFN-γ) stimulation for 4 hr. RAW 264.7 cells or the primary bone-marrow derived macrophages (BMMs) were treated with 200 U/ml IFN-γ for 4 or 24 hr. The nitrite level in the cell culture medium was determined by the Griess method (a). The cells were then infected with Listeria monocytogenes (b) or Salmonella typhimurium (c), and bacterial growth in the macrophages was calculated on the basis of colony-forming units (CFU). The values shown are the mean ± standard deviation values obtained from three independent experiments. *P < 0·01; **P < 0·02.

Figure 2.

Macrophage bactericidal activity increased 4 hr after interferon-γ (IFN-γ) stimulation in an IFN-γR-dependent, signal transducer and activator of transcription 1 (STAT1) -independent manner. (a, b) RAW 264.7 cells, bone-marrow-derived macrophages (BMMs), or stable knockdown cells were treated with 200 U/ml IFN-γ for 4 hr. The cells were then infected with Listeria monocytogenes (a) or Salmonella typhimurium (b), and bacterial growth in the macrophages was calculated on the basis of colony-forming units (CFU). The values shown are the mean ± standard deviation values obtained from three independent experiments. *P < 0·01; **P < 0·1; NS, not significant. (c) The cells were lysed in cell lysis buffer containing protease inhibitors, and the cell lysates were then subjected to SDS–PAGE analysis and Western blotting.

NOS2 is not involved in macrophage bactericidal activity in the early stages after IFN-γ stimulation

Next, we used the NOS inhibitors, l-NMMA and DPI, to rule out the possibility of participation of NOS2 and NO in macrophage bactericidal activity immediately after IFN-γ stimulation. The bactericidal activity at 4 hr after IFN-γ stimulation was not affected by either NOS inhibitor (Fig. 3a,b). The effects of these NOS inhibitors were confirmed by NO production in the macrophages that had been treated with IFN-γ for 24 hr, because no NO production was observed immediately after IFN-γ stimulation (Figs 1a and 3c). Both l-NMMA and DPI were effective at these concentrations (500 μm for l-NMMA and 10 μm for DPI). These results suggested that macrophage bactericidal activity in the early stages following IFN-γ stimulation is independent of NOS2 activity.

Figure 3.

Inducible nitric oxide synthase (NOS2) activity is not required for macrophage activation in the early stages after interferon-γ (IFN-γ) stimulation. RAW 264.7 cells or bone-marrow-derived macrophages (BMMs) were treated with 200 U/ml IFN-γ in the presence or absence of 500 μm of l-NMMA or 10 μm of DPI for 4 hr (a, b) or 24 hr (c). (a, b) The cells were then infected with Listeria monocytogenes (a) or Salmonella typhimurium (b), and bacterial growth in the cells was calculated on the basis of colony-forming units (CFU). (c) The nitrite level in the cell culture medium was determined by the Griess method. The values shown are the mean ± standard deviation values obtained from three independent experiments. *P < 0·01; **P < 0·02; ***P < 0·05; ****P < 0·1; NS, not significant.

The p38 MAPK signalling pathway is employed for IFN-γ-mediated pathogen clearance

Interferon-γ activates various signalling cascades, including MyD88, p38 MAPK, PI3K, and protein kinase C.^8,9,12–14 In previous studies, we have demonstrated that 4 hr after IFN-γ stimulation, autophagy may be activated in macrophages.³⁸ Therefore, we next investigated whether p38 MAPK-mediated autophagy contributes to bactericidal activity (Figs 4 and 5). We found that IFN-γ-mediated autophagy activation and bactericidal activity were inhibited in the presence of p38 MAPK inhibitors (PD 169316, SB 202190 and SB 203580) (Fig. 4) or in the p38α KD cells (Fig. 5), suggesting that macrophage activation at an early stage following IFN-γ stimulation depends on p38 MAPK.

Figure 4.

Macrophage activation in the early stages after interferon-γ (IFN-γ) stimulation is reduced by the p38 mitogen-activated protein kinase (MAPK) inhibitors. RAW 264.7 cells or bone marrow-derived macrophages (BMMs) were treated with 200 U/ml IFN-γ in the absence or presence of p38 MAPK inhibitors (10 μm of PD 169316, 10 μm of SB202190, or 5 μm of SB 203580) for 4 hr. (a, b) The cells were then infected with Listeria monocytogenes (a) or Salmonella typhimurium (b), and bacterial growth in the cells was calculated on the basis of colony-forming units (CFU). The values shown are the mean ± standard deviation values obtained from three independent experiments. *P < 0·01; **P < 0·05; ***P < 0·1. (c) The cells were lysed in cell lysis buffer containing protease inhibitors, and the cell lysates were then subjected to SDS–PAGE analysis and Western blotting. Densitometric LC3-II/GAPDH ratios are shown underneath the blot.

Figure 5.

Macrophage activation in the early stages after interferon-γ (IFN-γ) stimulation was reduced in p38α knockdown cells. (a, b) RAW 264.7 cells, bone-marrow-derived macrophages (BMMs), or stable knockdown cells were treated with 200 U/ml IFN-γ for 4 hr. The cells were then infected with Listeria monocytogenes (a) or Salmonella typhimurium (b), and bacterial growth in the macrophages was calculated on the basis of colony-forming units (CFU). The values shown are the mean ± standard deviation values obtained from three independent experiments. *P < 0·01; **P < 0·02; ***P < 0·05. (c) The cells were lysed in cell lysis buffer containing protease inhibitors, and the cell lysates were then subjected to SDS–PAGE analysis and Western blotting. Densitometric LC3-II/GAPDH ratios are shown underneath the blot.

Autophagy plays an important role in pathogen elimination by IFN-γ-stimulated macrophages

Next, to confirm whether p38 MAPK-mediated autophagy is involved in the function of IFN-γ-activated macrophages, autophagy-deficient Atg5 or Atg7 KD macrophage cells were used (Fig. 6). The antibacterial effect observed at 4 hr after IFN-γ treatment was blocked in autophagy-deficient cells, suggesting that macrophage bactericidal activity in the early stages following IFN-γ stimulation is mediated by autophagy. Our findings collectively suggested that, in IFN-γ-activated macrophages, bactericidal activity is quickly strengthened by p38 MAPK-mediated autophagy, rather than by NO. This action may be advantageous in macrophages before full activation.

Figure 6.

Autophagy plays a critical role in macrophage activation in the early stages after interferon-γ (IFN-γ) stimulation. RAW 264.7 cells, bone-marrow-derived macrophages (BMMs), and stable knockdown cells were treated with 200 U/ml IFN-γ for 4 hr. (a, b) The cells were infected with Listeria monocytogenes (a) or Salmonella typhimurium (b). The bacterial growth in the cells was calculated on the basis of colony-forming units (CFU). The values shown are the mean ± standard deviation values obtained from three independent experiments. *P < 0·01; **P < 0·05; ***P < 0·1; NS, not significant. (c) The cells were lysed in cell lysis buffer containing protease inhibitors, and the cell lysates were then subjected to SDS–PAGE analysis and Western blotting. Densitometric LC3-II/GAPDH ratios are shown underneath the blot.

Discussion

Autophagy has emerged as a major immune defence pathway; it also contributes to inflammation by facilitating an IFN-γ response and signal transduction.^34,42 We have previously shown that the p38 MAPK signalling cascade also activates autophagy in IFN-γ-stimulated macrophages, but the effect of p38 MAPK-mediated autophagy on macrophage function was unknown.³⁸ In this study, we found that macrophage bactericidal activity increased at 4 hr after IFN-γ stimulation in a STAT1- and NOS2-independent, p38 MAPK- and autophagy-dependent manner. Furthermore, this macrophage action was effective against both intracytosolic and intravacuolar pathogens. These findings suggest that p38 MAPK-mediated autophagy can help stimulate IFN-γ-mediated cell-autonomous immunity, with implications for understanding how IFN-γ-induced autophagy is mobilized within macrophages for inflammation and host defence.

Both type I and type II IFNs up-regulate gene transcription via the JAK–STAT, p38 MAPK, and PI3K signalling pathways, but slight differences exist in the regulation enacted by each type of IFN.⁷ Studies using a pharmacological inhibitor of p38 MAPK, over-expression of kinase-inactive p38α mutant, or mouse embryonic fibroblasts from p38α knockout mice, showed that p38 MAPK is required for gene transcription via the IFN-stimulated response element and IFN-γ-activated site in response to type I IFNs, such as IFN-α and IFN-β.^43,44 However, p38 MAPK plays no role in type II IFN or IFN-γ-dependent gene transcription.⁴⁴ In our previous study, phosphorylation of p38 MAPK by IFN-γ stimulation was observed after STAT1 knockdown in RAW 264.7 cells and in BMMs from STAT1-knockout mice.³⁸ Collectively, these data demonstrate that STAT1- and p38 MAPK-dependent pathways operate separately from each other in IFN-γ signalling.

The NOS2 is responsible for bactericidal activity in IFN-γ-stimulated macrophages that combat intracellular bacteria via NO generation.^22,23 In addition to NOS2, Irgm1 – a member of the IFN-γ-inducible p47 IRG family – activates autophagy, and this contributes to the defence mechanism of macrophages acting against M. tuberculosis.^25,26 The IFN-γ-induced macrophage bactericidal activity via NOS2 and Irgm1 depend on STAT1, because this molecule mediates the expression of NOS2 and Irgm1.^24,45

In this study, we showed that macrophage bactericidal activity increased at 4 hr following IFN-γ stimulation. This phenomenon was not attenuated in STAT1-KD cells, suggesting that the macrophage bactericidal activity occurring in the early stages following IFN-γ stimulation is not due to NOS2 or Irgm1 (Fig. 2). In fact, the NOS inhibitors l-NMMA and DPI failed to affect this kind of bactericidal activity (Fig. 3). The synergistic effects that are initiated via NOS2-, Irgm1- and p38 MAPK-dependent signalling cascades are likely to contribute to IFN-γ-mediated host defence.

Disclosure

The authors have no financial conflicts of interest.