Innate immune recognition of flagellin limits systemic persistence of Brucella

Abstract

Brucella are facultative intracellular bacteria that cause chronic infections by limiting innate immune recognition. It is currently unknown whether Brucella FliC flagellin, the monomeric subunit of flagellar filament, is sensed by the host during infection. Here, we used two mutants of Brucella melitensis, either lacking or overexpressing flagellin, to show that FliC hinders bacterial replication in vivo. The use of cells and mice genetically deficient for different components of inflammasomes suggested that FliC was a target of the cytosolic innate immune receptor NLRC4 in vivo but not in macrophages in vitro where the response to FliC was nevertheless dependent on the cytosolic adaptor ASC, therefore suggesting a new pathway of cytosolic flagellin sensing. However, our work also suggested that the lack of TLR5 activity of Brucella flagellin and the regulation of its synthesis and/or delivery into host cells are both part of the stealthy strategy of Brucella towards the innate immune system. Nevertheless, as a flagellin-deficient mutant of B. melitensis wasfound to cause histologically demonstrable injuries in the spleen of infected mice, we suggested that recognition of FliC plays a role in the immunological stand-off between Brucella and its host, which is characterized by a persistent infection with limited inflammatory pathology.

Fig. 1. Flagellin-deficient B. melitensis mutants infect macrophages in vitro with the same kinetics as wt bacteria but show enhanced persistence in mice.

(A) Western blot analysis of the production of flagellin (FliC, upper panel) by B. melitensis strains harvested at the early log phase and the log phase of growth in 2YT rich medium. Anti-Omp89 detection was used as a loading control (lower panel). Data are representative of two independent experiments. ΔfliC pfliC is the complemented strain. (B) Intracellular replication of B. melitensis 16M wt and ΔfliC strains in RAW264.7 murine macrophages. Error bars represent the standard deviation of triplicates in one representative experiment out of three. (C-D) Infection kinetics in the spleens of wt BALB/c mice (n=5) inoculated intraperitoneally (i.p.) with 4 × 10⁴ CFUs of B. melitensis 16M wt, ΔfliC, complemented ΔfliC pfliC, ΔflbT, or ΔfliF strains. (E) Infection kinetics in the spleens of wt C57BL/6 mice (n=5) inoculated i.p. with 4 × 10⁴ CFUs of B. melitensis 16M wt or ΔfliC strains. Data represent the mean CFUs per organ and error bars represent standard deviation. Results have been analyzed by ANOVA I after testing the homogeneity of variance (Bartlett). ** and *** denote highly significant (p<0.01 and p<0.001 respectively) differences in relation to wt infection. These results are representative of at least two independent experiments.

Fig. 2. Enhanced persistence of B. melitensis ΔfliC in mice is associated with increased pathology

(A) Kinetics of splenomegaly in wt female BALB/c mice (n=5) injected i.p. with 4 × 10⁴ CFUs of wt or ΔfliC strains of B. melitensis 16M. Data represent the mean spleen weight and error bars represent standard deviation. Results have been analyzed by ANOVA I after testing the homogeneity of variance (Bartlett). *** denotes highly significant (p<0.001) differences in relation to wt infection. (B) Splenic pathology caused by a 28 day-infection was determined using the histopathology scoring system as described in the Material and methods. Data were analysed using a Mann Whitney test, and the mean histopathology scores were significantly different (P=0.009) (C) Representative photomicrographs (x10) of histopathology of spleens from BALB/c mice uninfected or infected for 28 days with B. melitensis wt or ΔfliC strain. WP, white pulp; T, thrombosis; black arrows, granuloma; white arrowhead, neutrophil infiltration. These results are representative of at least two independent experiments.

Fig. 3. Constitutive production of flagellin does not impair replication of B. melitensis 16M in macrophages in vitro, but attenuates its virulence in vivo.

(A) Western blot analysis of flagellin (FliC, upper panel) production in wt and BruFliC^ON strains during early exponential and stationary phases of growth in 2YT rich medium. Detection of Omp89 was used as a loading control. (B) Intracellular replication of wt and BruFliC^ON strains in RAW264.7 murine macrophages. Error bars represent the standard deviation of triplicates in one representative experiment out of two. (C) Infection kinetics in the spleens of wt BALB/c mice (n=5) inoculated i.p. with 4 × 10⁴ CFUs of wt or BruFliC^ON strain. Data represent the mean CFUs per organ and error bars represent standard deviation. Results have been analyzed by ANOVA I after testing the homogeneity of variance (Bartlett). ** and *** denote highly significant (p<0.01 and p<0.001 respectively) differences in relation to wt infection. These results are representative of at least two independent experiments.

Fig. 4. Brucella flagellin lacks TLR5 agonist activity

(A-C) FLAG-tagged flagellins from S. enterica serotype Typhimurium (StFliC) or Brucella abortus (BruFliC) were expressed in an S. Typhimurium fliCfljB mutant, and culture supernatants containing recombinant flagellins were used to treat cells. (A) Western blot showing production of bacterium-associated flagellins from S. Typhimurium wt (lane 1), S. Typhimurium fliCfljBmutant (lane 2), fliCfljB mutant expressing StFliC-FLAG (lane 3) or fliCfljB mutant expressing BruFliC-FLAG (lane 4). Flagellins were detected both in the pellets (left panel) and in the concentrated supernatants (right panel) of S. Typhimurium strains. 30ng of concentrated supernatant proteins from S. Typhimurium strains expressing recombinant flagellins were used to treat HEK293/hTLR5 cells for 4 or 24h (B) and T84 cells for 8h (C). IL-8 in cell supernatants was measured by ELISA. (D) Activation of p38 and ERK MAPK in T-84 cells by purified recombinant flagellins from Brucella (GST-BruFliC) and S. Typhimurium (GST-StFliC) was measured by Western blot analysis with anti- p38, anti-phosphorylated (P-)p38, anti- ERK, and anti-P-ERK. Detection of tubulin was used as a loading control. Purified flagellins treated with proteinase K (PK) were used as a control. All data shown are from an individual experiment that was repeated at least twice with similar results.

Fig. 5. B. abortus flagellin induces IL-1β in an NLRC4-independent manner

(A) Primary bone marrow-derived macrophages from C57BL/6 mice were primed with LPS and inoculated with B. melitensis 16M wt or the ΔfliC mutant and IL-1β was measured in the culture supernatants by ELISA. Results are shown as the mean ± standard deviation of data from an individual experiment that was repeated 4 times with similar results. (B) Immortalized, LPS-primed C57BL/6 or Nlrc4^-/- bone marrow-derived macrophages were inoculated with B. melitensis 16M wt or the BruFliC^ONstrain. IL-1β in the supernatant was measured at 6h after inoculation. Data shown are combined from three independent experiments with triplicate samples, and represent the mean ± standard deviation of all data.

Fig. 6. Introduction of recombinant Brucella flagellin into the host cell cytosol results in ASC-dependent, but NLRC4-independent secretion of IL-1β

Graded amounts of GST-BruFliC and GST-StFliC fusion proteins were delivered to the cytosol of LPS-primed primary bone marrow-derived macrophages from C57BL/6 (A), Nlrc4^-/- (B) or Asc^-/- (C) mice, using the cationic lipid DOTAP. Treated macrophages were incubated for 3h before measurement of IL-1β in the supernatants by ELISA. Results are expressed as the mean of triplicate samples, with error bars representing the range of the data from one of two independent experiments with the same outcome.

Fig. 7. NLRC4 inflammasome is implicated in the control of B. melitensis infection in vivo

Wild type, Nlrc4^-/- (A) and Casp1^-/- (B) C57BL/6 mice (n=5) were injected i.p. with 4 × 10⁴ CFUs of B. melitensis wt, BruFliC^ON or ΔfliC strain, as indicated in the figure. Mice were sacrificed 21 days post-infection and CFUs per spleen were determined. These results are representative of at least two independent experiments. Data have been analysed by ANOVA I after testing the homogeneity of variance (Bartlett). * ,** and *** denote significant (p<0.05, p<0.01 and p<0.001 respectively) differences in relation to C57BL/6 wt infection by wt bacteria. # and ## denote significant (p<0.05 and p<0.01 respectively) differences in relation to knock-out mice infection by wt bacteria.

Fig. 8. The distribution of Bru-positive cells is different in the spleen of mice infected by the ΔfliC mutant, compared to wt infection

Localization of Bru⁺ cells (green) and CD11b⁺ cells (red) in the spleen of BALB/c mice non-infected or infected with B. melitensis wt or the ΔfliC strain. The graph represents the relative number of clusters of Bru⁺ cells. Errors bars are the standard deviation calculated on countings of four mice from two independent experiments.