Bacterial vesicles block viral replication in macrophages via TLR4-TRIF-axis

Abstract

Gram-negative bacteria naturally secrete nano-sized outer membrane vesicles (OMVs), which are important mediators of communication and pathogenesis. OMV uptake by host cells activates TLR signalling via transported PAMPs. As important resident immune cells, alveolar macrophages are located at the air-tissue interface where they comprise the first line of defence against inhaled microorganisms and particles. To date, little is known about the interplay between alveolar macrophages and OMVs from pathogenic bacteria. The immune response to OMVs and underlying mechanisms are still elusive. Here, we investigated the response of primary human macrophages to bacterial vesicles (Legionella pneumophila, Klebsiella pneumoniae, Escherichia coli, Salmonella enterica, Streptococcus pneumoniae) and observed comparable NF-κB activation across all tested vesicles. In contrast, we describe differential type I IFN signalling with prolonged STAT1 phosphorylation and strong Mx1 induction, blocking influenza A virus replication only for Klebsiella, E.coli and Salmonella OMVs. OMV-induced antiviral effects were less pronounced for endotoxin-free Clear coli OMVs and Polymyxin-treated OMVs. LPS stimulation could not mimic this antiviral status, while TRIF knockout abrogated it. Importantly, supernatant from OMV-treated macrophages induced an antiviral response in alveolar epithelial cells (AEC), suggesting OMV-induced intercellular communication. Finally, results were validated in an ex vivo infection model with primary human lung tissue. In conclusion, Klebsiella, E.coli and Salmonella OMVs induce antiviral immunity in macrophages via TLR4-TRIF-signaling to reduce viral replication in macrophages, AECs and lung tissue. These gram-negative bacteria induce antiviral immunity in the lung through OMVs, with a potential decisive and tremendous impact on bacterial and viral coinfection outcome. Video Abstract.

Keywords: Alveolar epithelial cell; Antiviral innate immunity; Bacterial and viral co-infection; Extracellular vesicles; Macrophage; Outer membrane vesicles; Pneumonia.

Fig. 1

Characterization of bacterial vesicles and response in human macrophages. A Separation of bacterial extracellular vesicles from free proteins via size exclusion chromatography from different bacterial supernatants (Legionella pneumophila (Lp), Klebsiella pneumoniae (Kp), Escherichia coli (Ec), Salmonella enterica serovar Typhimurium (Sal), and Streptococcus pneumoniae (Sp)). Vesicle concentration in each fraction was determined by nano-flow cytometry (nFCM) and proteins were quantified by BCA. B Vesicle size distributions of purified OMVs/MVs were determined by nFCM. C TEM images of purified OMVs/MVs. Scale bar = 50 nm. D–G BDMs were stimulated with OMVs/MVs (1 µg/mL each) from different bacteria or left untreated for control for up to 48 h. D CXCL8 release was determined by ELISA and is depicted in ng/mL. Expression of IL1B (E) and IL12B (F) were determined by qPCR, results are normalized to RPS18 and are depicted relative to untreated control cells. G After 1 h incubation with bacterial vesicles, expression and phosphorylation of IRAK-1, p38 and TBK-1 were determined by Western Blot. Representative results of four biological independent replicates are shown. Bars represent mean values + SEM from three (B) to four (D–F) independent experiments. Statistics: 2-way ANOVA (D-F); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns = not significant; n = 3–4

Fig. 2

Bacterial extracellular vesicles activate type I interferon response in BDMs. BDMs were stimulated with OMVs/MVs (1 µg/mL each) from different bacteria or left untreated as control. A Phosphorylation and expression of IRF-3 was determined after 1 h of OMV/MV incubation by Western Blot. Representative result of three biological independent replicates is shown. Expression of IFNA1 (B), IFNB (C), IFIT1 (E) and IFI44 (F) was measured by qPCR, results are normalized to RPS18 and are depicted relative to untreated control cells. D Phosphorylation of STAT1 was determined by Western Blot after 2 h of OMV/MV stimulation. Representative result of four biological independent replicates is shown. G Phosphorylation of STAT1 and expression of Mx1 were determined by Western Blot after 20 h of bacterial vesicle stimulation. Representative results of four biological independent replicates are depicted. Bars represent mean values + SEM from four independent experiments. Statistics: 2-way ANOVA; *p < 0.05, **p < 0.01, ****p < 0.0001; n = 4

Fig. 3

Mx1-inducing OMVs block influenza A virus replication in THP-1 cells. A Influenza A virus replication in differentiated THP-1 cells. Cells were either pre-treated with OMVs/MVs (1 µg/mL each), or left untreated for control (–). After 20 h pre-treatment, cells were infected with A/WSN/33(H1N1) (MOI 0.001) for 24 and 48 h. IAV replication was determined by qPCR against IAV-NP normalized to 18S. Mean values ± SEM of three to five independent experiments are shown. B Differentiated THP-1 cells were pre-treated with OMVs/MVs (1 µg/mL each) or left untreated for control (–). After 20 h pre-incubation, cells were infected with A/WSN/33(H1N1) (MOI 0.1) for 4 h. After fixation, cells were stained with an α-influenza NP antibody (yellow) and DAPI (blue). A representative result from four biological independent experiments is shown. C Quantification of NP positive (NP⁺) area from (B). Bars represent mean values of four independent experiments + SEM. D THP-1 cells stably overexpressing Mx1 (Mx1oex) and empty vector control (VC) cells were infected with influenza virus A/WSN/33(H1N1) (MOI 0.1) for 4 h. After fixation and immunofluorescence staining with α-influenza NP, the NP⁺ area was quantified and is depicted relative to VC cells. Bars show mean values of four independent experiments + SEM. E Mx1oex and VC cells were infected with A/WSN/33(H1N1) (MOI 0.001) for 6 h. Viral replication was determined by plaque assay and results are depicted as plaque forming units (pfu) per mL. Bars represent mean values + SEM of four independent experiments. Statistics: 2-way ANOVA (A), 1-way ANOVA (C), unpaired t-test (D + E); *p < 0.05, **p < 0.01, ****p < 0.0001; n = 3–5

Fig. 4

Inhibition of JAK-signalling rescues influenza A virus replication in macrophages. A OMVs activate TRIF-IRF-signalling in macrophages, which in turn induces IFN-β expression and release with subsequent IFNAR-signalling. IFNAR signals via JAK/STAT and induces the expression of ISGs, one of which is Mx1. This leads to a block of IAV replication. B–E THP-1 cells were pre-incubated for 1 h with 10 µM JAK inhibitor (JAKi) before addition of OMVs (1 µg/mL; Kp/Sal) for 20 h. B–C CXCL8 (B) and Mx1 (C) expression was determined by qPCR and results are normalized to RPS18 and depicted relative to untreated control. Bars show mean values of three to four independent experiments + SEM. D + E OMV pre-treated cells were additionally infected with IAV (MOI 0.001). D Western blot shows Mx1 protein expression at 0–3 h post infection (p.i.). E Viral replication was determined by plaque assay 24 h after infection. Bars show mean values of four independent experiments + SEM. Statistics: 1-way ANOVA (B + C + E); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; *compared to DMSO control, # as depicted in the graph; ns = not significant; n = 4

Fig. 5

OMVs induce antiviral signalling in macrophages via TRIF. A THP-1 Dual reporter cells induce the expression of secreted embryonic alkaline phosphatase (SEAP) after activation of the NF-κB pathway and the expression and secretion of Lucia luciferase upon activation of the IRF pathway. Additionally, TRIF^−/− cells have a stable knockout of the adapter molecule TRIF. (B-F) THP-1 reporter cells (= Dual; black bars) and TRIF^−/− cells (grey bars) were differentiated and subsequently stimulated with KpOMV (1 µg/mL) for 20 h or left untreated for control. After the indicated time, supernatant, RNA and/or proteins were collected. B Representative Western Blot image of Mx1 protein expression. C + D Lucia reporter activity (C) and SEAP reporter activity (D) was determined in cell culture supernatant. The same supernatant was used to determine the activity of both reporters. E Relative mRNA expression for STAT- and NF-κB-dependent target genes (from top to bottom: IFIT1, IFI44, Mx1, CXCL8) was determined by qPCR and results are normalized to RPS18 and depicted relative to unstimulated Dual control. Fold changes were log2 transformed. Cells were additionally infected with A/WSN/33(H1N1) (MOI 0.1) for 24 h (F) or VSV (MOI 0.1) for 12 h (G). Viral replication was determined by plaque assay. Bars show mean values of four (C–F) to five (G) independent experiments + SEM. Statistics: 2-way ANOVA (C–G); *p < 0.05, **p < 0.01, ****p < 0.0001; * compared to unstimulated Dual control, # as depicted in the graph; ns = not significant; n = 4–5

Fig. 6

OMV-induced Mx1 expression is lost after LPS inhibition. THP-1 cells were incubated for 20 h with OMVs (1 µg/mL; Kp or Clear coli (Cc)) alone or in combination with 20 µg/mL Polymyxin B (PB) or left untreated for control. A Mx1 protein expression was determined by Western Blot. A representative result of three biological independent experiments is shown. Mx1 B and CXCL8 C expression was determined by qPCR. Bars show mean values of four independent experiments + SEM. D THP-1 cells were incubated for 20 h with OMVs (1 µg/mL; Kp or Cc) alone or in combination with 20 µg/mL PB or left untreated for control and then additionally infected with A/WSN/33(H1N1) (MOI 0.001). Virus replication was determined by plaque assay 24 h after infection. Bars show mean values of four independent experiments + SEM. Statistics: 1-way ANOVA B–D; *p < 0.05, **p < 0.01, ****p < 0.0001; *compared to control, #compared to KpOMV; ns = not significant; n = 3–4

Fig. 7

Influenza A virus replication-inhibiting effect of OMVs is transferable to AECs. THP-1 cells were incubated with LpOMV or KpOMV for 20 h or left untreated for control. Supernatant (SN) was sterile filtered and used for pre-stimulation of A549 cells for 20 h alone or in combination with 10 µM JAKi. A + B Mx1 (A) and CXCL8 B expression were determined by qPCR at the time point of infection (0 h p.i.). Bars show mean values of four independent experiments + SEM normalized to untreated control cells. C Pre-treated A549 cells were additionally infected with influenza virus A/Hamburg/5/2009(H1N1pdm) (MOI 0.01) for 24 h. Viral replication was determined by plaque assay. Bars are mean values of four independent experiments + SEM. Statistics: 1-way ANOVA; **p < 0.01, ****p < 0.0001; * compared to control-SN; # compared to KpOMV-SN; n = 4

Fig. 8

KpOMV reduce influenza A virus replication in human precision-cut lung slices. Human PCLS were incubated with 1 µg/mL OMVs (Lp/Kp) for 20 h and then additionally infected with influenza A/California/04/2009(H1N1pdm) for 48 h. UV inactivation of virus served as a control. A Cytotoxicity was determined by quantification of released LDH from PCLS and is depicted in % compared to a total lysis. B Mx1 expression was quantified by qPCR and is presented relative to RPS18 and untreated control PCLS. C CXCL10 release was determined by ELISA and is depicted in ng/mL. D Viral replication was determined by plaque assay and is depicted in pfu/mL. Bars show mean values of three to five biological replicates + SEM. E Proposed model for induction of antiviral immunity of OMVs in the lung. Statistics: 1-way ANOVA (A-C), Friedman-test (D); *p < 0.05, ***p < 0.001, ****p < 0.0001; * compared to IAV infected, but not pre-treated PCLS, # as depicted in the graph; ns = not significant; n = 5