Toll-Like Receptor 8 Is a Major Sensor of Group B Streptococcus But Not Escherichia coli in Human Primary Monocytes and Macrophages

Abstract

TLR8 is the major endosomal sensor of degraded RNA in human monocytes and macrophages. It has been implicated in the sensing of viruses and more recently also bacteria. We previously identified a TLR8-IFN regulatory factor 5 (IRF5) signaling pathway that mediates IFNβ and interleukin-12 (IL-12) induction by Staphylococcus aureus and is antagonized by TLR2. The relative importance of TLR8 for the sensing of various bacterial species is however still unclear. We here compared the role of TLR8 and IRF5 for the sensing of Group B Streptococcus (GBS), S. aureus, and Escherichia coli in human primary monocytes and monocyte-derived macrophages (MDM). GBS induced stronger IFNβ and TNF production as well as IRF5 nuclear translocation compared to S. aureus grown to the stationary phase, while S. aureus in exponential growth appeared similarly potent to GBS. Cytokine induction in primary human monocytes by GBS was not dependent on hemolysins, and induction of IFNβ and IL-12 as well as IRF5 activation were reduced with TLR2 ligand costimulation. Heat inactivation of GBS reduced IRF5 and NF-kB translocation, while only the viable E. coli activated IRF5. The attenuated stimulation correlated with loss of bacterial RNA integrity. The E. coli-induced IRF5 translocation was not inhibited by TLR2 costimulation, suggesting that IRF5 was activated via a TLR8-independent mechanism. Gene silencing of MDM using siRNA revealed that GBS-induced IFNβ, IL-12-p35, and TNF production was dependent on TLR8 and IRF5. In contrast, cytokine induction by E. coli was TLR8 independent but still partly dependent on IRF5. We conclude that TLR8-IRF5 signaling is more important for the sensing of GBS than for stationary grown S. aureus in human primary monocytes and MDM, likely due to reduced resistance of GBS to phagosomal degradation and to a lower production of TLR2 activating lipoproteins. TLR8 does not sense viable E. coli, while IRF5 still contributes to E. coli-induced cytokine production, possibly via a cytosolic nucleic acid sensing mechanism.

Keywords: human TLR8; infection; inflammation; pattern recognition receptors; primary human phagocytes.

Figure 1

Group B Streptococcus (GBS) induce IFNβ and TNF via a TLR2-independent mechanism and is a more potent trigger than Staphylococcus aureus from stationary growth. Human primary monocytes (Mo) and monocyte-derived macrophages (MDM) were infected with viable GBS wt, lipoprotein diacylglyceryl transferase deficient GBS (Δlgt) and S. aureus 113-strain. Cytokine induction was determined by quantitative real-time PCR (qPCR) 3 h postinfection in monocytes (A–C) [multiplicity of infection (MOI): 0.4–4.0 for S. aureus, and 0.2–2.0 for GBS], or 4 h postinfection in MDM (D) (MOI 4.0–40 for S. aureus and 2.0–20 for GBS). Statistical significance was tested relative to S. aureus. n = 4.

Figure 2

Group B Streptococcus (GBS)-induced production of IFNβ in monocytes is not dependent on hemolysin and is antagonized by TLR2 ligand. (A) Infection with viable GBS wt, GBS Δlgt, and GBS hemolysin deficient strains (ΔcylEΔcfb), which is deficient for both the β-hemolysin CylE gene (cylE) and the cohemolysin Christie Atkins Munc-Petersen (CAMP) factor gene (cfb), and the CylE deficient (ΔcylE) strain was done for 3 h (MOI 2.0). (A) IFNβ and TNF levels in the supernatant determined by ELISA (n = 4). (B) Cytokine transcript induction determined by quantitative real-time PCR (qPCR) (n = 6). (C) Infection of monocytes for 18 h with viable GBS wt (MOI 0.02-0.20-2.00) with or without TLR2 ligand FSL-1 (100 ng/ml) costimulation. Levels of IFNβ in the supernatants were determined with ELISA (n = 4), while IL-12-p70 levels were determined by bioplex (n = 3). In non-infected samples the levels were below the limit of detection (1 pg/ml for IFNβ and 20 pg/ml for IL-12-p70).

Figure 3

Group B Streptococcus (GBS) is a more potent activator of interferon regulatory factor 5 (IRF5) nuclear accumulation in monocytes than stationary grown Staphylococcus aureus and Escherichia coli. Monocytes were infected with viable S. aureus Cowan strain (MOI 0.20–0.64–2.00), GBS wt (MOI 0.10–0.32–1.00), or E. coli (MOI 0.20–0.64–2.00) for 2 h. Fixation and immunofluorescence (IF) double staining of IRF5 and NF-kB (p65/RelA) was subsequently performed, and quantification was done by high-content screening (Scan^R, 20×). (A) Percentage of nuclei positively stained for IRF5, and (B) p65. (C) Representative images of IRF5 and p65 staining in cell with no stimuli (NS), and GBS and E. coli infected monocyte cultures (scale bar 50 µm). Significance was tested relative to NS. n = 4 (n = 2 for the medium bacterial dose).

Figure 4

IFN regulatory factor 5 (IRF5) is activated by TLR8 ligands and cytosolic dsDNA sensing, IRF3 is activated by TLR4 signaling and cytosolic dsDNA and dsRNA sensors, while only TLR8 is antagonized by TLR2 ligand. (A) Monocytes were stimulated with CL75 (1 μg/ml), LPS (K12, 100 ng/ml), pdA:dT (dsDNA, 1 μg/ml), or poly(I:C) (dsRNA, 1 μg/ml). The nucleic acid ligands were transfected using L2K. Fixation was done at 30, 60, and 120 min poststimulation and IF staining of IRF5 and IRF3 was performed. (B) Monocytes were stimulated with the TLR8 agonist polyuridylic acid (pU) in complex with poly-l-arginine (pLA) or with pdA:dT/L2K. The ligands (1 μg/ml) were added alone or together with FLS-1 (100 ng/ml), and cells were fixed and stained after 120 min of incubation. Quantification of nuclear accumulation was done by high-content screening (Scan^R, 20×). Significance was tested relative to no stimuli (NS). n = 4.

Figure 5

Activation of IFN regulatory factor 5 (IRF5) by TLR8 ligand and Group B Streptococcus (GBS) is antagonized by TLR2 costimulation, while activation of IRF5 by viable Escherichia coli is not. Monocytes were stimulated with CL75 (1 μg/ml), viable or heat-inactivated (HI) GBS (MOI 1.0) and E. coli (MOI 2.0), and with or without FSL-1 costimulation (100 ng/ml) for 2 h. The monocytes were subsequently fixed and double IF stained IRF5 and p65/RelA. Nuclear accumulation was quantified by high-content screening (Scan^R, 20×). NS, no stimuli. nd, not done. n = 4.

Figure 6

TLR8 accumulates on Group B Streptococcus (GBS) phagosomes in macrophages and monocytes. (A) THP-1 TLR8-KO cells were transduced with TLR8 mNeonGreen construct (TLR8mNG) and differentiated with PMA. Macrophages were infected with viable GBS (MOI 1.0) prestained with Alexa 647. Gentamycin was added 1 h postinfection and cells were fixed with paraformaldehyde after 20 h. DNA was stained with Hoechst 33342. (B) Primary monocytes were infected with viable GBS (MOI 10), and Gentamycin was added 1 h postinfection. Cells were fixed with paraformaldehyde after 3 h and TLR8 was IF stained with a mAb. DNA was stained with Hoechst 33342 in PBS/saponin to visualize GBS phagosomes (arrowhead) and monocyte nuclei. Images were acquired using Olympus Scan^R (60×). Scale bars are 10 µm.

Figure 7

Group B Streptococcus (GBS) trigger cytokine production in part via TLR8-IRF5 signaling, while Escherichia coli induces cytokines independently of TLR8, yet partly dependent on IRF5. Monocyte-derived macrophages (MDM) were transfected with siRNA against TLR7, TLR8, IRF5, and STING, as indicated. Following successful silencing the cells were stimulated with pU/pLA or infected with viable GBS wt (MOI 1.0) or E. coli (MOI 2.0). Cytokine induction relative to non-stimulated cells after 4 h of infection was determined by quantitative real-time PCR (qPCR). Significance of gene silencing was tested in relation to non-targeting control (NTC). n = 5.