Wolbachia lipoprotein stimulates innate and adaptive immunity through Toll-like receptors 2 and 6 to induce disease manifestations of filariasis

Abstract

Wolbachia endosymbiotic bacteria have been implicated in the inflammatory pathogenesis of filariasis. Inflammation induced by Brugia malayi female worm extract (BMFE) is dependent on Toll-like receptors 2 and 6 (TLR2/6) with only a partial requirement for TLR1. Removal of Wolbachia, lipids, or proteins eliminates all inflammatory activity. Wolbachia bacteria contain the lipoprotein biosynthesis genes Ltg and LspA but not Lnt, suggesting Wolbachia proteins cannot be triacylated, accounting for recognition by TLR2/6. Lipoprotein databases revealed 3-11 potential lipoproteins from Wolbachia. Peptidoglycan-associated lipoprotein (PAL) and Type IV secretion system-VirB6 were consistently predicted, and B. malayi Wolbachia PAL (wBmPAL) was selected for functional characterization. Diacylated 20-mer peptides of wBmPAL (Diacyl Wolbachia lipopeptide (Diacyl WoLP)) showed a near identical TLR2/6 and TLR2/1 usage compared with BMFE and bound directly to TLR2. Diacyl WoLP induced systemic tumor necrosis factor-alpha and neutrophil-mediated keratitis in mice. Diacyl WoLP activated monocytes induce up-regulation of gp38 on human lymphatic endothelial cells and induced dendritic cell maturation and activation. Dendritic cells primed with BMFE generated a non-polarized Th1/Th2 CD4+ T cell profile, whereas priming with Wolbachia depleted extracts (following tetracycline treatment; BMFEtet) polarized to a Th2 profile that could be reversed by reconstitution with Diacyl WoLP. BMFE generated IgG1 and IgG2c antibody responses, whereas BMFEtet or inoculation of TLR2 or MyD88-/- mice produced defective IgG2c responses. Thus, in addition to innate inflammatory activation, Wolbachia lipoproteins drive interferon-gamma-dependent CD4+ T cell polarization and antibody switching.

FIGURE 1.

The inflammatory stimuli of BMFE are lipoproteins that primarily signal via TLR2/6. A, HEK-TLR2 cells were transfected with plasmids encoding small interfering (psi) RNA specific for TLR1 or TLR6 before being stimulated with BMFE or control stimuli (doses stated are in micrograms/ml). Accumulations of IL-8 secreted by HEK-psiTLR1 or -psiTLR6 triplicate cultures 20 h post-stimulation are plotted as mean (±1 S.E.) percentages of corresponding HEK-TLR2 responses (mean ± 1 S.E. max IL-8 concentrations as follows: TNFα = 7859 ± 98 pg/ml, PAM₃CSK = 9807 ± 175 pg/ml, FSL-1 = 2001 ± 345 pg/ml, and BMFE = 9495 ± 137 pg/ml). Significant differences compared with HEK-TLR2 responses are indicated: ***, p < 0.001; **, p < 0.01. B, peritoneal macrophages from WT, TLR1^−/−, TLR6^−/− were stimulated with BMFE in triplicate (doses stated are micrograms/ml), and production of TNFα after 20 h is plotted as mean ± 1 S.E. percentages of WT response to 400 μg/ml BMFE (mean ± 1S.E. max TNFα concentration = 188.4 ± 10.36 pg/ml). Significant differences compared with WT are indicated ***, p < 0.001; **, p < 0.01. C, triplicate HEK-TLR2 cultures were stimulated with BMFE or control stimuli (doses stated are micrograms/ml) before or following Cleanascite^TM or BindPro^TM treatment. Data plotted are mean IL-8 ± 1S.E. All data are representative of three independent experiments.

FIGURE 2.

A, anti-rwBmPAL antibodies bind to a Wolbachia product present within BMFE. Anti-rwBmPAL polyclonal antibody immunoblot of three batches of BMFE (a–c) and a corresponding preparation of Wolbachia-depleted BMFEtet, separated by one-dimensional SDS-PAGE. Molecular weight markers are stated in kilodaltons. B, anti-rwBmPAL staining of Wolbachia-infected C6/36 cells (an A. albopictus mosquito cell line, top right panel) and Wolbachia-free C6/36 cells (top left panel). Antibody reactivity was detected with goat anti-rabbit FITC conjugate and counterstained with Evans Blue. Human onchocercoma sections (lower left panel) and B. malayi adult females (lower right panel) were stained using the affinity-purified anti-rwBmPAL antibody. Antibody reactivity was detected using alkaline phosphatase with a permanent red chromogenic substrate and counterstained with hematoxylin.

FIGURE 3.

Synthetic diacyl-lipopeptide analogue of wBmPAL (Diacyl WoLP) replicates BMFE-TLR2/6-specific activation of inflammation. A, HEK-TLR2 cells were transfected with plasmids encoding small interfering (psi) RNA specific for TLR1 or TLR6 before being stimulated with Diacyl WoLP or Triacyl WoLP (doses stated are in micrograms/ml). Accumulations of IL-8 secreted by HEK-psiTLR1 or -psiTLR6 triplicate cultures 20 h post-stimulation are plotted as mean ± 1 S.E. percentages of corresponding HEK-TLR2 responses (mean ± 1S.E. max IL-8 concentrations are as follows: PAM₃CSK = 9807 ± 175 pg/ml, FSL-1 = 2001 ± 345 pg/ml, Diacyl WoLP = 8195 ± 199 pg/ml, Triacyl WoLP 571 ± 27 pg/ml, and BMFE = 9495 ± 137 pg/ml). Significant differences compared with HEK-TLR2 responses are indicated: ***, p < 0.001; **, p < 0.01; and *, p < 0.05. B, peritoneal macrophages from WT, TLR1^−/−, and TLR6^−/− were stimulated with Diacyl WoLP or Triacyl WoLP and control TLR1/6 ligands PAM₃CSK₄ and FSL-1 (doses stated are in nanograms/ml) in triplicate, and production of TNFα after 20 h is plotted as mean ± 1S.E. All data are representative of three independent experiments.

FIGURE 4.

TLR2 enhances binding of Diacyl WoLP. A, confocal microscope images of a HEK293 cell (upper panels) or HEK-TLR2 stable transfectant (lower panels) cultured on coverslips with 5 μg/ml Diacyl WoLP:AF488. Scale bar = 10 μm. B, flow cytometric analysis of HEK293 (black histogram) or HEK-TLR2 (red histogram) transfectants labeled with 5 μg/ml Diacyl WoLP:AF488. PBS processed for AF488 dye labeling was used as a negative control to determine background levels of fluorescent labeling (gray, filled histogram = PBS:AF488-treated HEK293; dashed black histogram = PBS:AF488-treated HEK-TLR2). C, quantification of differences in frequency of cells labeled (% positive) and amount of labeling (median fluorescent intensity (MFI)) of Diacyl WoLP:AF488 in HEK versus HEK-TLR2 transfectants determined by flow cytometry. Bars are mean MFI or % positive ±1S.E. from triplicate labeling reactions each containing 100,000 cells. Significant differences are indicated ***, p < 0.001. All data are representative of two independent experiments.

FIGURE 5.

Diacyl WoLP induces filarial disease manifestations. A, Diacyl WoLP induces neutrophil infiltration following intra-corneal injection into WT mice in a dose-dependent manner (left-hand panel). Bars are mean ± 1S.E. neutrophil numbers from groups of three animals. Doses given are in nanograms/ml. Neutrophil accumulation is impaired in TLR2^−/− or TLR6^−/− but not TLR1^−/− mice following a 1000 ng/ml intra-corneal injection (right-hand panels). Bars are mean ± 1S.E. neutrophil numbers from groups of three animals. B, corneal haze is induced by Diacyl WoLP in a dose-dependent fashion in WT mice (right-hand panels). Bars are mean ± 1S.E. corneal haze from groups of three animals. Doses given are in nanograms/ml. Corneal haze is significantly reduced in TLR2^−/− and TLR6^−/− mice but not TLR1^−/− mice following a 1000 ng/ml dose. Significance differences between Diacyl WoLP inoculation and negative control or between WT and specific KO are indicated: ***, p < 0.001; **, p < 0.01; and *, p < 0.05. C, Diacyl WoLP induces a systemic TNFα response in plasma 6 h following a 50-μg intraperitoneal injection in WT but not TLR2^−/− mice. Bars indicate median TNFα levels from groups of five animals. Significant difference is indicated: **, p < 0.01 (Mann-Whitney non-parametric test). D, Wolbachia- and Diacyl WoLP-dependent gp38 up-regulation on lymphatic endothelial cells. THP-1 cells (human monocytic cell line) were stimulated with 200 μg/ml BMFE or BMFEtet or 10 μg/ml WoLP for 24 h. Monocyte supernatants were harvested and added to HMVECdly at a 1:3 dilution. Recombinant human TNFα (100 ng/ml) and IL-1β (10 ng/ml) were added to HMVECdly as a positive control. Surface expression of gp38 was determined by flow cytometry after 16 h. Differences in MFI of anti-gp38 labeling in HMVECdly exposed to TNFα and IL-1β or supernatants from treated THP-1 cells (treatments given are in micrograms/ml) were compared with unstimulated THP-1 supernatant exposed HMVECdly. Bars are mean ± 1S.E. MFI from triplicate labeling reactions. Significant differences between treatment and medium control are indicated: ***, p < 0.001; *, p < 0.05. All data are representative of two independent experiments. E, histograms represent fluorescent intensity of anti-gp38-stained HMVECdly exposed to unstimulated THP-1 supernatant (dashed line) or stimulated THP-1 supernatants/positive control (red lines). The isotype control is plotted as a black solid line.

FIGURE 6.

Diacyl WoLP provides necessary signals for DC maturation and activation. A, increases in DC CD40, CD80, CD86, and MHCII following 20-h exposure to 0.1 μg/ml LPS, 0.1 μg/ml Diacyl WoLP, 200 μg/ml BMFE, or 200 μg/ml BMFEtet (all red line histograms), 0.01 μg/ml Diacyl WoLP (red dashed-line histograms), compared with untreated DC (black histograms). Gray shaded histograms = untreated DC isotype control, dashed-line histograms = exposed DC isotype control. B, CD40, CD80, CD86, and MHCII surface expression levels in unexposed or DC exposed to LPS, Diacyl WoLP, BMFE, or BMFEtet. Bars are mean ± 1S.E. median fluorescent intensities of triplicate labeling reactions. Doses stated are in micrograms/ml. C, production of DC TNFα, IL-12/IL-23p40, IL-12p70, and IL-23p40/p19 following 20-h exposure to LPS, BMFE, or BMFEtet. Doses stated are in picograms/ml. Bars are mean ± 1S.E. cytokine production from triplicate cultures. D, production of DC IL-12p70 and IL-23p40/p19 following 20-h exposure to various doses (stated in micrograms/ml) of Diacyl WoLP or LPS. Bars are mean ± 1S.E. cytokine production from triplicate cultures. Significant differences are indicated: ***, p < 0.001; **, p < 0.01; and *, p < 0.05. All data are representative of three independent experiments.

FIGURE 7.

DC maturation and activation by Wolbachia and Diacyl WoLP requires MyD88, TLR2, and TLR6 but not TLR4. A, increase in CD80 or CD86 surface molecules following 20-h exposure to LPS, FSL-1, Diacyl WoLP, or BMFE in DC derived from WT, MyD88^−/−, TLR2^−/−, TLR4^−/−,or TLR6^−/− mice. Doses stated are in micrograms/ml. Bars represent mean fold increase in MFI ± S.E. compared with unstimulated cells of triplicate labeling reactions. B, stimulation of TNFα, IL-12/IL-23p40, and IL-12p70 by LPS, FSL-1, Diacyl WoLP, or BMFE (doses stated are in micrograms/ml) from DC derived from WT, MyD88^−/−, TLR2^−/−, TLR4^−/−, or TLR6^−/− mice. Bars are mean ± 1S.E. cytokine production from triplicate cultures. C, Diacyl WoLP mediates an expansion of mature CD11c⁺ DC in vivo. Increases in CD86 surface expression on CD11c⁺ splenocytes 6 h following intraperitoneal inoculation with 50 μg of Diacyl WoLP were compared with sham inoculated WT mice. Numbers are percentages of splenocytes in the upper left and right quadrants. D, significant differences in CD80 and CD86 MFI on CD11c⁺ splenocytes derived from WT mice were compared with TLR2^−/−-deficient mice 6 h following inoculation with 50 μg of Diacyl WoLP intraperitoneally. Bars are mean MFI from groups of three mice. E, significant differences in levels of IL-12/IL-23 p40 measured in spleen extracts from WT mice compared with TLR2^−/−-deficient mice 6 h following inoculation with 50 μg of Diacyl WoLP intraperitoneally. Bars are mean cytokine levels from groups of three mice. Significant reductions compared with WT are indicated: ***, p < 0.001; **, p < 0.01; and *, p < 0.05. All data are representative of two independent experiments.

FIGURE 8.

Diacyl WoLP modulates antigen-specific CD4⁺ T-cell activation and differentiation. A, DO11.10 OVA transgenic CD4⁺ T-cell proliferation following 72-h co-culture of OVA peptide-loaded DC primed with BMFE, BMFEtet, BMFEtet+Diacyl WoLP, Diacyl WoLP, LPS, or untreated (medium only). Doses of stimuli stated are micrograms/ml. Bars are mean ± 1S.E. tritiated thymidine incorporation (counts per minute). Significant differences to untreated DC are indicated: ***, p < 0.001; **, p < 0.01; and *, p < 0.05. B, ratios of IL-4:IFNγ from DO11.10 OVA transgenic CD4⁺ T cells following 72-h co-culture of OVA peptide-loaded DC primed with BMFE, BMFEtet, BMFEtet+Diacyl WoLP, Diacyl WoLP, LPS, or untreated (medium only). Doses of stimuli stated are in micrograms/ml. Bars are the mean ratio ±1S.E. of IL-4:IFNγ secreted in supernatant from triplicate co-cultures 24 h after polyclonal (phorbol myristate acetate/ionomycin) stimulation. Significant different ratios compared with untreated DC are indicated: **, p < 0.001; *, p < 0.05. All data are representative of two independent experiments.

FIGURE 9.

Anti-BMFE IgG2c production is dependent on MyD88 and TLR2 but not TLR4. A, time course of BMFE-specific IgG1 or IgG2c antibody production following inoculation with 50 μg of BMFE intraperitoneally at days 0 and 7 in MyD88^−/−, TLR2^−/−, TLR4^−/−, or WT mice. Data plotted are mean antibody levels from groups of four mice. B, day 25 anti-IgG1 or -IgG2c levels. Bars represent mean levels of antibody. Significant differences compared with WT mice are indicated: **, p < 0.01. C, day 25 anti-IgG1 or -IgG2c levels from WT mice inoculated with 50 μg of BMFE or BMFEtet intraperitoneally at days 0 and 7. Horizontal bars represent mean levels of antibody. Significant differences are indicated: ***, p < 0.001. Data are representative of two independent experiments.