IFN-β is a macrophage-derived effector cytokine facilitating the resolution of bacterial inflammation

Abstract

The uptake of apoptotic polymorphonuclear cells (PMN) by macrophages is critical for timely resolution of inflammation. High-burden uptake of apoptotic cells is associated with loss of phagocytosis in resolution phase macrophages. Here, using a transcriptomic analysis of macrophage subsets, we show that non-phagocytic resolution phase macrophages express a distinct IFN-β-related gene signature in mice. We also report elevated levels of IFN-β in peritoneal and broncho-alveolar exudates in mice during the resolution of peritonitis and pneumonia, respectively. Elimination of endogenous IFN-β impairs, whereas treatment with exogenous IFN-β enhances, bacterial clearance, PMN apoptosis, efferocytosis and macrophage reprogramming. STAT3 signalling in response to IFN-β promotes apoptosis of human PMNs. Finally, uptake of apoptotic cells promotes loss of phagocytic capacity in macrophages alongside decreased surface expression of efferocytic receptors in vivo. Collectively, these results identify IFN-β produced by resolution phase macrophages as an effector cytokine in resolving bacterial inflammation.

Fig. 1

Non-phagocytic resolution phase macrophages express IFN-β. a, b Male mice were injected intraperitoneally with zymosan A (1 mg/mouse) followed by an injection of PKH2-PCL at 62 h. After 4 h, the peritoneal cells were recovered and immuno-stained for F4/80 and CD11b. Then, F4/80⁺ macrophages were sorted based on the extent of PKH2-PCL acquisition (PKH2 + /− populations; >98% purity) using the FACSAria III sorter (illustrated in (a). The collected cells were immediately used for RNA extraction (with RNA integrity value above 7.5), and a gene expression microarray analysis was performed using Illumina hiSeq 2500. Differential gene expression analysis and gene ontology (GO) enrichment were performed for genes that were significantly upregulated (b, left panel) or downregulated (b, right panel) in non-phagocytic/satiated (PKH2-PCL^lo) macrophages in comparison to phagocytic (PKH2-PCL^hi) ones. The results indicate the statistical significance of the GO term and the percentage of enrichment is presented. c–e Expression of IFN-β and ISG15 in sorted satiated and phagocytic macrophages. Representative results (c) and mean ± SEM (d, e) for three independent experiments. *P < 0.05 (Student’s t test). g, h Peritoneal exudates were collected from unchallenged mice (0 h) or following peritonitis for 4–96 h. IFN-β content in cell-free fluids was determined by ELISA (f). Results are mean ± SEM from three (24, 72, 96 h) or four (0, 4, 48 h) mice. *P < 0.05, **P < 0.01, ***P < 0.005 (Tukey’s HSD). Alternatively, resolution phase macrophages were recovered 66 h post peritonitis initiation (PPI) and incubated with TGF-β (5 ng/ml), poly (I:C) (4 μg/ml) or apoptotic cells (AC, at a ratio of 1:5) for 24 h. Culture supernatants were then collected and IFN-β content was measured (g). Culture media from apoptotic cells served as control. Results are representative from three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.005 (Tukey’s HSD). Source data are provided as a Source Data file

Fig. 2

Enhanced IFN-β production during the resolution of E.coli pneumonia. Under isoflurane anesthesia, female C57BL/6 mice were injected intratracheally with 5*10⁶ live E. coli. At the indicated times, lungs were collected without lavage and analyzed for E. coli content (a), lung dry-to-wet weight ratio (c) and tissue MPO activity (e). In separate groups of mice bronchoalveolar lavage fluid protein concentration (b), IFN-β levels (d), total leukocyte (f), neutrophil (g) and monocyte/macrophage numbers (h) and the percentage of annexin-V-positive (apoptotic) PMN (i) and the percentage of BAL fluid macrophages containing apoptotic bodies (j) were determined. Results are means ± SEM (n = 6 mice per group). *P < 0.05, **P < 0.01 (Dunn’s multiple contrast hypothesis test). k Neutrophil apoptosis positively correlates with lavage fluid IFN-β levels (Spearman correlation analysis). Source data are provided as a Source Data file

Fig. 3

IFN-β is essential for the resolution of E. coli pneumonia. Female C57BL/6 mice were injected intratracheally with 5*10⁶ live E. coli with anti-IFN-β antibody (1 μg/20 g b.w.) or isotype control (IgG) or saline (“-“). At 24 or 48 h, lungs were removed without lavage and analyzed for E. coli content (a), lung dry-to-wet weight ratio (c) and tissue MPO activity (d). In separate groups of mice bronchoalveolar lavage fluid protein concentration (b), total leukocyte (e), neutrophil (f) and monocyte/macrophage numbers (g), the percentage of annexin-V-positive (apoptotic) PMN (h), and the percentage of macrophages containing apoptotic bodies (i) were determined. Results are means ± SEM (n = 6 mice per group for control and anti-IFN-β, or 4 for IgG). *P < 0.05, **P < 0.01 (Dunn’s multiple contrast hypothesis test). Source data are provided as a Source Data file

Fig. 4

IFN-β promotes apoptosis in mouse and human neutrophils. a Peritoneal PMN were recovered from male Ifnb^+/+ or Ifnb^−/− mice at 4 h PPI and stained immediately with annexin-V and propidium iodide to assess apoptosis and cell viability, respectively with flow cytometry. Alternatively, cells were cultured with or without the pan-caspase inhibitor Q-VD (10 μM) for 24 h and then assessed for apoptosis. Results are mean ± SEM from three independent experiments. *P < 0.05, **P < 0.01 (Tukey’s HSD). b, c Peritoneal PMN were recovered from Ifnb^+/+ mice 24 h PPI and cultured ex vivo with IFN-β (20 ng/ml) and/or Q-VD (10 µM) for 24 h. Apoptosis was evaluated as above. In some experiments, PMN lysates were prepared after 6 h culture and immunoblotted for cleaved (active) caspase-3. Results are representatives for three independent experiments. *P < 0.05 (Tukey’s HSD). d–g Human PMN (5 × 10⁶ cells/ml) were pretreated with human recombinant IFN-β (25–50 ng/ml) for 10 min and then challenged with CpG DNA (1.6 μg/ml) or (h–k) first challenged with CpG DNA (1.6 μg/ml) and then treated with IFN-β (50 ng/ml) at 60 min post-CpG DNA. Cell viability (d, h), annexin-V staining (e, i), mitochondrial transmembrane potential (ΔΨm; CMXRos staining, f, j) and nuclear DNA content (g, k) were analyzed after culturing neutrophils for 24 h with CpG DNA. Results are mean ± SEM of 5 experiments with different blood donors. *P < 0.05, **P < 0.01 (Dunn’s multiple contrast hypothesis test). l, m Human PMN lysates, prepared following 4 h culture with IFN-β (50 ng/ml) with or without CpG DNA, were immunoblotted for Mcl-1 or the ubiquitous transcription factor YY1 as a loading control. Representative blots (l) and densitometry analyses (m) for three independent experiments. *P < 0.05, **P < 0.01 (Dunn’s multiple contrast hypothesis test). Source data are provided as a Source Data file

Fig. 5

IFN-β promotes neutrophil apoptosis through STAT3 activation. a, b Human PMN (5*10⁶ cells/ml) were pretreated with human IFN-β (50–150 ng/ml) for 10 min and then challenged with CpG DNA for 30 min. The cells were then lysed, cytosolic and nuclear fractions were prepared and immunoblotted for phospho-STAT1 (a) or phospho-STAT3 (b). β-actin served as a loading control. Blots are representative for 3 independent experiments. c–f Human PMN were pre-incubated with the STAT1 inhibitor fludarabine (FLU; 25 μM) or the STAT3 inhibitor WP1066 (5 μM) for 30 min before addition of human recombinant IFN-β (50 ng/ml) for 10 min and then challenged with CpG DNA. After culture for additional 24 h, cell viability (c), annexin-V staining (d), mitochondrial transmembrane potential (ΔΨm, CMXRos staining) (e) and nuclear DNA content (f) were analyzed. Results are mean ± SEM of 6 experiments with different blood donors. *P < 0.05, **P < 0.01, ***P < 0.001 (Dunn’s multiple contrast hypothesis test). Source data are provided as a Source Data file

Fig. 6

IFN-β enhances macrophage efferocytosis during the resolution of inflammation. a–c Peritoneal macrophages were recovered from male Ifnb^+/+ or Ifnb^−/− mice at 48 h PPI and stained with Hoechst 33342 and FITC-phalloidin (a). Fluorescent images of select frames were taken using high-resolution microscopy achieved by 3D confocal (z-stack) scanning with a Nikon A1-R confocal fluorescent microscope. The number of apoptotic PMN nuclei in each macrophage were enumerated (see arrowheads for illustration) using the Nikon NIS-Elements microscope imaging software and average engulfment per macrophage (b) or engulfment according to thresholds (c) were calculated. Representative images (a) and means ± SEM (b, c) for six (Ifnb^+/+) and three (Ifnb^−/−) experiments. d–f Peritoneal macrophages were recovered from male Ifnb^+/+ or Ifnb^−/− mice at 48 h PPI and incubated with CypHer-labeled apoptotic Jurkat cells at a ratio of 1:3. After 4 h, unbound cells were washed and macrophages were stained with Hoechst 33342 and FITC-phalloidin (as illustrated in d). Select images of each staining and merged images are shown (e). The number of engulfed apoptotic cells (AC) in each macrophage (M) were counted and average apoptotic cell uptake (f) and the percentage of efferocytosing macrophages (g) were calculated. h The total numbers of F4/80⁺ peritoneal macrophages was detected by flow cytometry and calculated. Representative images (e) and means ± SEM (f–g) for four (Ifnb^+/+) and five (Ifnb^−/−) independent experiments. *P < 0.05, ***P < 0.005 (Student’s t test). Source data are provided as a Source Data file

Fig. 7

IFN-β favors macrophage reprogramming during the resolution of inflammation. a–d Macrophages were recovered from peritoneal exudates of male Ifnb^+/+ or Ifnb^−/− mice at 48–66 h PPI and cultured with LPS (1 μg/ml) for 24 h. Culture supernatants were then collected and levels of IL-10 (a), IL-6 (b), IL-12 (c) and CCL3 (d) were determined by selective ELISAs. Results are means ± SEM from four independent experiments. *P < 0.05, ***P < 0.005 (Tukey’s HSD). e–f Macrophages were recovered from peritoneal exudates of Ifnb^+/+ mice at 48–66 h PPI and cultured with mouse IFN-β or IFN-α (20 ng/ml each) for 48 h. Then, culture supernatants were collected and levels of IL-10 (e) and IL-12 (f) were determined by ELISA. Results are means ± SEM (n = 4). ***P < 0.005 (Tukey’s HSD). g Macrophages were recovered from peritoneal exudates of Ifnb^+/+ mice 48–66 h PPI and incubated with IFN-β (20 ng/ml) for 48 h. The cells were then immunostained for F4/80 and CD11b and the percentage of CD11b^low macrophages was determined by flow cytometry. Results are means ± SEM from three independent experiments. ***P < 0.005 (Tukey’s HSD). h–i Mice undergoing peritonitis were treated with IFN-β (20 ng/mouse, i.p.) or vehicle at 24 h PPI. Peritoneal macrophages were collected at 48 h PPI, lysed and immunoblotted for 12/15-LO, arginase 1, ISG15 and GAPDH. Representative blots (h) and densitometry analysis (means ± SEM) (i) for three independent experiments. *P < 0.05 (Student’s t test). Source data are provided as a Source Data file

Fig. 8

IFN-β treatment accelerates the resolution of E. coli pneumonia. Female C57BL/6 mice were injected intratracheally with 5*10⁶ live E. coli. Six hours later (at the peak of inflammation), they were treated with mouse recombinant IFN-β (50 ng/20 g b.w., intraperitoneally) or vehicle. At 24 or 48 h post-E. coli instillation, lungs were removed without lavage and analyzed for E. coli content (a), lung dry-to-wet weight ratio (c) and tissue MPO activity (d). In separate groups of mice bronchoalveolar lavage fluid protein concentration (b), total leukocyte (e), neutrophil (f) and monocyte/macrophage numbers (g), the percentage of annexin-V-positive (apoptotic) PMN (h), and the percentage of macrophages containing apoptotic bodies (i) were determined. Results are means ± SEM (n = 6 mice per group for 6 and 24 h and 5 mice per group for 48 h). *P < 0.05, **P < 0.01 (Dunn’s multiple contrast hypothesis test). Source data are provided as a Source Data file

Fig. 9

Characteristics of macrophages following loss of phagocytosis. a, b The phagocyte-specific dye PKH2-PCL green was injected I.P. to male WT mice undergoing peritonitis for 20, 44 and 68 h. After 4 h, the peritoneal cells were recovered and immuno-stained for F4/80 and CD11b. PKH2-PCL green acquisition by CD11b^high macrophages was determined by flow cytometry. Results are means ± SEM (n = 8 mice for 24 h, 6 mice for 48 h, and 7 mice for 72 h) showing MFI (a) and the percentage of PKH2⁺ cells (b). *P < 0.05 (Tukey’s HSD). c–g Mice were injected i.p. with 3 × 10⁶ apoptotic Jurkat cells, latex beads, IgG-opsonized latex beads or vehicle together with PKH2-PCL red at 62 h PPI. 4 h later, the peritoneal cells were recovered, immuno-stained for F4/80 and CD11b and F4/80⁺ macrophages were analyzed by flow cytometry for target particle uptake and/or PKH2 engulfment (designated as PKH2-high, -low and -negative populations). Results are representative for three experiments presented as density or histogram plots (c, f) and means ± SEM from 4 independent experiments of % of phagocytic macrophages (d), particles engulfed (e), or PKH2 uptake (g). *P < 0.05, **P < 0.01 (Tukey’s HSD). h, i PKH2-PCL green was injected I.P. to mice at 44 h PPI followed by an injection of PKH2-PCL red at 58  h PPI. Peritoneal cells were recovered at 62 or  66 h PPI and immuno-stained for F4/80 and CD11b and analyzed with flow cytometry (as illustrated in h). PKH2-PCL green vs. red acquisition was determined in F4/80⁺ macrophages and macrophages that displayed increased, sustained or reduced phagocytosis (representative dot-plot is shown in i). j Data are means ± SEM (n = 5 mice for 4 h and 3 mice for 8 h). *P < 0.05, **P < 0.01 (Student’s t test). k, l PKH2-PCL red was injected I.P. to mice at 62 h PPI. 4 h later the peritoneal cells were recovered, immuno-stained for CD36, CD206 or MER, receptors involved in apoptotic cell uptake. Staining for CD45 served as a control. Receptor expression on PKH2-PCL-high (PKH2^hi) or PKH2-PCL-low (PKH2^lo) macrophages was determined by flow cytometry. Results are representative from n = 4 mice presented as dot plots (k) and means ± SEM relative MFI normalized to PKH2^hi macrophage expression (l). *P < 0.05, **P < 0.01, ***P < 0.005 (Student’s t test). Source data are provided as a Source Data file