Salmonella Typhimurium infection inhibits macrophage IFNβ signaling in a TLR4-dependent manner

Abstract

Type I Interferons (IFNs) generally have a protective role during viral infections, but their function during bacterial infections is dependent on the bacterial species. Legionella pneumophila, Shigella sonnei and Mycobacterium tuberculosis can inhibit type I IFN signaling. Here we examined the role of type I IFN, specifically IFNβ, in the context of Salmonella enterica serovar Typhimurium (STm) macrophage infections and the capacity of STm to inhibit type I IFN signaling. We demonstrate that IFNβ has no effect on the intracellular growth of STm in infected bone marrow derived macrophages (BMDMs) derived from C57BL/6 mice. STm infection inhibits IFNβ signaling but not IFNγ signaling in a murine macrophage cell line. We show that this inhibition is independent of the type III and type VI secretion systems expressed by STm and is also independent of bacterial phagocytosis. The inhibition is Toll-like receptor 4 (TLR4)-dependent as the TLR4 ligand, lipopolysaccharide (LPS), alone is sufficient to inhibit IFNβ-mediated signaling. Cells downregulated their surface levels of IFNα/β receptor 1 (IFNAR1) in response to LPS, which may be mediating our observed inhibition. Lastly, we examined this inhibition in the context of TLR4-deficient BMDMs as well as TLR4 RNA interference and we observed a loss of inhibition with LPS stimulation as well as STm infection. In summary, we show that macrophages exposed to STm have reduced IFNβ signaling via crosstalk with TLR4 signaling, which may be mediated by reduced host cell surface IFNAR1, and that IFNβ signaling does not affect cell-autonomous host defense against STm.

Keywords: Salmonella; interferons; macrophages; toll receptors.

Fig 1

STm infection specifically inhibits IFNβ signaling. RAW-Lucia ISG-KO-IRF3 cells were infected with STm at MOI of 10 for 30 min followed by 1-h 100 µg/mL gentamicin chase. Post-infection, cells were stimulated with IFNβ (A) or IFNγ (B) in the presence of 20 µg/mL gentamicin for 20 h and the response of the gene reporter measured by luminescence and plotted as RLUs (A and B, left panels) as well as fold change relative to the uninfected, IFNβ- or IFNγ-treated condition (A and B, right panels). Data points indicate independent biological experiments and are the average from four technical replicates. Statistical analysis for RLUs is multiple unpaired t-tests, and for fold change, is one sample t- and Wilcoxon test where each fold change mean is compared to 1. Significance is as follows: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Fig 2

IFNβ has no effect on restriction of STm in ex vivo BMDM infection. (A) BMDMs from wild-type and Ifnb^-/- mice were infected with STm at MOI of 3 for 30 min followed by 1-h 100 µg/mL gentamicin chase. Next, cells were stimulated with 1 ng/mL IFNβ every 12 h in the presence of 20 µg/mL gentamicin. CFUs were measured at the indicated timepoints. The data shown are mean values and standard deviation of three independent infections (left panel) or normalized to 0 h post-infection data (right panel). (B) LDH cell death assay was performed on cell supernatants obtained from the experiment described in panel A for each given timepoint. Data points indicate independent biological replicates (average from the two technical replicates) and are shown as relative absorbance at 490 nm (left panel) or normalized to 0 h post-infection (right panel). Statistical analysis for raw absorbance values is multiple unpaired t-test with Welch correction where each background strain stimulated with IFNβ is relative to untreated cells for a given timepoint. Statistical analysis for fold change relative to the 0-h timepoint is multiple unpaired t-test with Welch correction.

Fig 3

STm infection results in reduced STAT1 phosphorylation after IFNβ treatment. (A and B) RAW-Lucia ISG-KO-IRF3 cells were infected at MOI of 3 for 30 min followed by 1-h 100 µg/mL gentamicin chase. Next, cells were stimulated with 200 pg/mL IFNβ for the indicated timepoints in the presence of 20 µg/mL gentamicin. Whole cell lysates were collected and immunoblotted for pSTAT1, STAT1, and β-actin (A). Band densities were normalized to either STAT1 or β-actin as indicated. Densitometric ratios are shown relative to UI + IFNβ control (B). The blot shown here is representative of three independent infections, and individual points indicate the average from three independent experiments. (C) Representative images of RAW-Lucia ISG-KO-IRF3 cells infected or not with STm and treated with IFNβ or left untreated (UT). Cells were infected with STm-YFP at MOI of 3 for 30 min followed by 1-h 100 µg/mL gentamicin chase. Following infection and chase, cells were stimulated with 1 ng/mL IFNβ for the indicated times. (D) Cellpose was used to determine relative bacterial burden and MFI within Voronoi area corresponding to the cell nuclei. Points represent the mean of MFI values from four independent experiments. Statistical analysis for B is one sample t- and Wilcoxon test where each fold change mean is compared to 1. Statistical analyses for D are unpaired t-test with Welch correction. Significance is indicated as such: *P ≤ 0.05, **P ≤ 0.01.

Fig 4

STm infection inhibits IFNβ signaling independently of type III and type VI secretion systems as well as bacterial phagocytosis. RAW-Lucia ISG-KO-IRF3 cells were infected with wild-type STm or (A) SPI-1 mutants or (B) SPI-2 mutants to delete the encoded T3SS and (C) SPI-6 mutant to delete the encoded T6SS at MOI of 10 for 30 min followed by 1-h 100 µg/mL gentamicin chase. Following infection, cells were stimulated with 200 pg/mL IFNβ in the presence of 20 µg/mL gentamicin for 20 h. Values are plotted as RLUs as well as fold change relative to an uninfected IFNβ-treated control. Data points indicate independent biological experiments and are the average from four technical replicates. (D) RAW-Lucia ISG-KO-IRF3 cells were pretreated with cytoD for 30 min followed by infection at MOI of 3 for 30 min. CytoD was maintained on the cells at all points during the experiment. Whole cell lysates were collected and immunoblotted for pSTAT1, STAT1, and β-actin. Band densities were normalized to either STAT1 or β-actin as indicated (D, right panels). Densitometric ratios are shown relative to UI + IFNβ control (D). Statistical analysis for RLUs is multiple unpaired t-tests. For fold change and western blot quantification, statistical analyses are one sample t- and Wilcoxon test where each fold change mean is compared to 1. Significance is as follows: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.

Fig 5

LPS stimulation specifically inhibits IFNβ signaling. RAW-Lucia cells were pretreated with 50 ng/mL LPS for 4 h followed by an additional 20 h along with 200 pg/mL IFNβ (A) or 500 pg/mL IFNγ (B). Values are plotted as RLUs as well as fold change relative to their respective −LPS + IFN control. Supernatants from the IFNβ and IFNγ experiments were collected and analyzed for endpoint cell death using LDH release assay (C and D). Values are plotted as absorbance at 490 nm (Abs490) as well as fold change relative to their respective −LPS + IFN control. Individual points indicate independent biological experiments and are the average of four technical replicates. RAW-Lucia ISG-KO-IRF3 cells were pretreated with 50 ng/mL LPS for the indicated timepoints followed by stimulation with either (E) 200 pg/mL IFNβ or (F) 500 pg/mL IFNγ for 30 min. Whole cell lysates were collected and immunoblotted for pSTAT1, STAT1, and β-actin. Band densities were normalized to either STAT1 or β-actin as indicated. Densitometric ratios are shown relative to respective −LPS + IFNβ or −LPS + IFNγ control. Each blot and individual points indicate independent biological experiments. Statistical analysis for each panel is one sample t- and Wilcoxon test where each fold change mean is compared to 1. Statistical significance is as follows: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.

Fig 6

LPS stimulation downregulates cell surface IFNAR1. BMDMs were stimulated with 50 ng/mL LPS and/or 200 pg/mL IFNβ for 30 min each as indicated. Flow cytometry analysis was performed using fluorescent antibodies specific for IFNAR1 (A) and IFNAR2 (B). Histograms are from one experiment representative of three independent experiments. Bar graphs show MFI with each point representing an independent experiment. Statistical analysis is repeated measure one-way ANOVA with Fisher’s least significant difference (LSD) post-comparison test. Significance is as follows: **P ≤ 0.01.

Fig 7

STm infection inhibits IFNβ signaling in a TLR4-dependent manner. BMDMs were infected with STm at MOI of 3 for 30 min followed by 1-h 100 µg/mL gentamicin chase. Cells were then either lysed to determine CFUs post-infection (B) or incubated for an additional 30 min or 60 min followed by 200 pg/mL IFNβ stimulation for 30 min (A). This results in 90 min and 120 min post-infection timepoints. Whole cell lysates were collected and immunoblotted for pSTAT1, STAT1, and β-actin (A and C). Band densities were normalized to either STAT1 or β-actin. The blot is representative of three independent infections with data points (C) indicating independent experiments. To determine successful repression of gene expression, RAW-Lucia ISG-KO-IRF3 cells were transfected with small interfering RNA and allowed to rest for 24 h before lysis followed by immunoblot for TLR4 and β-actin (D). To assess loss of inhibition, RAW-Lucia-KO-IRF3 cells were transfected as previously described here followed by stimulation with 50 ng/mL of LPS (E) or infection with STm at MOI of 3 (F). Post stimulation and infection, cells were stimulated with 200 pg/mL IFNβ for 20 h in the presence of 20 ng/mL gentamicin and the reporter cell response was measured (E and F, left panels). For 7D, blot is representative of two independent transfections. For CFUs, statistical analysis is unpaired t-test with Welch correction. For western blot quantification, statistical analysis is repeated measure one-way ANOVA with Fisher’s LSD post-comparison test. Statistical analysis for RLUs are multiple ratio paired t-tests or one sample t- and Wilcoxon test where each fold change is compared to 1 (E and F). Statistical significance is as follows: **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.