Lipopolysaccharides from distinct pathogens induce different classes of immune responses in vivo

Abstract

The adaptive immune system has evolved distinct responses against different pathogens, but the mechanism(s) by which a particular response is initiated is poorly understood. In this study, we investigated the type of Ag-specific CD4(+) Th and CD8(+) T cell responses elicited in vivo, in response to soluble OVA, coinjected with LPS from two different pathogens. We used Escherichia coli LPS, which signals through Toll-like receptor 4 (TLR4) and LPS from the oral pathogen Porphyromonas gingivalis, which does not appear to require TLR4 for signaling. Coinjections of E. coli LPS + OVA or P. gingivalis LPS + OVA induced similar clonal expansions of OVA-specific CD4(+) and CD8(+) T cells, but strikingly different cytokine profiles. E. coli LPS induced a Th1-like response with abundant IFN-gamma, but little or no IL-4, IL-13, and IL-5. In contrast, P. gingivalis LPS induced Th and T cell responses characterized by significant levels of IL-13, IL-5, and IL-10, but lower levels of IFN-gamma. Consistent with these results, E. coli LPS induced IL-12(p70) in the CD8alpha(+) dendritic cell (DC) subset, while P. gingivalis LPS did not. Both LPS, however, activated the two DC subsets to up-regulate costimulatory molecules and produce IL-6 and TNF-alpha. Interestingly, these LPS appeared to have differences in their ability to signal through TLR4; proliferation of splenocytes and cytokine secretion by splenocytes or DCs from TLR4-deficient C3H/HeJ mice were greatly impaired in response to E. coli LPS, but not P. gingivalis LPS. Therefore, LPS from different bacteria activate DC subsets to produce different cytokines, and induce distinct types of adaptive immunity in vivo.

FIGURE 1

Experimental design. B6.PL.Thy-1^a (B6.PL) mice, or C57BL/6 mice that were reconstituted with OT-2 cells were injected with either soluble OVA, soluble OVA + E. coli LPS, soluble OVA + P. gingivalis LPS, E. coli LPS alone, or P. gingivalis LPS alone i.p. or in the footpad. Four days later, the spleens or draining lymph nodes were removed for phenotypic and functional analyses, including clonal expansion of OVA-specific CD4⁺ T cells, in vitro proliferation of OVA-specific CD4⁺ T cells, and cytokine production by the OVA-specific CD4⁺ T cells.

FIGURE 2

E. coli LPS and P. gingivalis LPS enhance Ag-specific Th responses in vivo. B6.PL.Thy-1^a (B6.PL) mice reconstituted with OT-2 transgenic T cells were immunized with soluble OVA, OVA + E. coli LPS, or OVA + P. gingivalis LPS, either in the footpad (s.c; B, D, and F), or i.p. (C,E, and G). Four days later, the draining popliteal lymph nodes (footpad injections) or spleens (i.p. route) were removed, and the clonal expansion of OVA-specific CD4⁺ T cells was assessed by flow cytometry, by staining with Thy-1.2 vs CD4. A, Flow cytometry profiles from day 4 of the response in the popliteal lymph nodes, from a representative experiment. B and C, The percentage expansion of OVA-specific CD4⁺ T cells (Thy-1.2⁺, CD4⁺) in the draining lymph nodes (B) and the spleens (C) at day 4. Both E. coli LPS and P. gingivalis LPS significantly enhance the clonal expansion, regardless of the route of injection. Data represent the means of four mice per group from one representative experiment. SDs are indicated, and differences between OVA group and the OVA + LPS groups are highly significant (p < 0.01). D andE, The absolute numbers of Thy-1.2⁺ CD4⁺ cells per popliteal lymph node (D) or per spleen (E) at day 4. Data represent the means of four mice per group from one representative experiment. SDs are indicated, and differences between OVA group and the OVA + LPS groups are highly significant (p< 0.01). F and G, In vitro restimulation of OVA-specific T cells expanded in vivo, by footpad (F), or i.p. (G) injections. Four days after priming, single cell suspensions from the draining popliteal lymph nodes (F) or spleens (G) were restimulated with varying concentrations of OVA for 72 h, and pulsed with ³H for 12 h. Note that injections of E. coli LPS alone or P. gingivalis LPS alone did not result in significant clonal expansion or in vitro proliferation. A–G, Representative of 10 independent experiments.

FIGURE 3

E. coli LPS and P. gingivalis LPS induce distinct types of Ag-specific Th responses in vivo. Culture supernatants from the cultures described in Fig. 2, F and G, were assayed for IL-2, IFN-γ, IL-10, IL-5, and IL-5 with ELISA. Each data point represents mean ± SD of triplicate wells, from a typical experiment. Differences in IL-2 production between the E. coli LPS and the P. gingivalis LPS groups are not significant; IL-5 production in the P. gingivalis LPS group at 500 µg/ml OVA is significant (p< 0.05; detection limit of ELISA assay is 6 pg/ml). Note that injections of E. coli LPS alone or P. gingivalis LPS alone did not result in significant cytokine production. Representative of 10 independent experiments.

FIGURE 4

P. gingivalis LPS induces much higher levels of IL-13 than E. coli LPS. B6.PL.Thy-1^a (B6.PL) mice reconstituted with OT-2 transgenic T cells were immunized with soluble OVA, OVA + E. coli LPS, or OVA + P. gingivalis LPS. Four days after priming, single cell suspensions from the draining popliteal lymph nodes or spleens were restimulated with varying concentrations of OVA for 72 h, and assayed for IL-13 production. Each data point represents mean ± SD of triplicate wells, from a typical experiment. Differences between the E. coli LPS and the P. gingivalis LPS groups are highly significant (p < 0.01). Representative of three independent experiments.

FIGURE 5

E. coli LPS and P. gingivalis LPS enhance Ag-specific CD8⁺ T cell responses in vivo. C57BL/6 mice or B6.PL.Thy-1^a (B6.PL) mice reconstituted with OT-1 transgenic T cells were immunized with soluble OVA, OVA + E. coli LPS, or OVA + P. gingivalis LPS, either in the footpad (s.c; B, D, and F) or i.p. (C,E, and G). Four days later, the draining popliteal lymph nodes (footpad injections) or spleens (i.p. route) were removed, and the clonal expansion of OVA-specific CD8⁺ T cells was assessed by flow cytometry, by staining with CD8 vs Vα2, or CD8 vs Thy-1.2 vs Vα2 (A). B and C, The percentage expansion of OVA-specific CD8⁺ T cells (CD8⁺ Thy-1.2⁺) in the draining lymph nodes (B) and the spleens (C) at day 4. Both E. coli LPS and P. gingivalis LPS significantly enhance the clonal expansion, regardless of the route of injection. Data represent the means of four mice per group from one representative experiment. SDs are indicated, and differences between OVA group and the OVA + LPS groups are highly significant (p < 0.001). D and E, The absolute numbers of OVA-specific CD8⁺ T cells per popliteal lymph node (D) or per spleen (E) at day 4. Data represent the means of four mice per group from one representative experiment. SDs are indicated, and differences between OVA group and the OVA + LPS groups are highly significant (p < 0.001). F and G, In vitro restimulation of OVA-specific CD8⁺ T cells expanded in vivo, by footpad (F), or i.p. (G) injections. Four days after priming, single cell suspensions from the draining popliteal lymph nodes (F) or spleens (G) were restimulated with varying concentrations of OVA for 72 h, and pulsed with ³H for 12 h. Note that injections ofE. coli LPS alone or P. gingivalis LPS alone did not result in significant clonal expansion, or in vitro proliferation. Representative of three independent experiments.

FIGURE 6

E. coli LPS and P. gingivalis LPS induce distinct types of Ag-specific CD8⁺ T cell responses in vivo. Culture supernatants from the cultures described in Fig. 5, E and F, were assayed for IL-2, IFN-γ, IL-10, IL-5, and IL-5 with ELISA. Note that injections of E. coli LPS alone or P. gingivalis LPS alone did not result in significant cytokine production. Each data point represents mean ± SD of triplicate wells, from a typical experiment. Differences in IL-2 production between the E. coli LPS and the P. gingivalis LPS groups are not significant; IL-5 production in the P. gingivalis LPS group at 500 µg/ml OVA is significant (p < 0.05; detection limit of ELISA assay is 6 pg/ ml). Note that injections of E. coli LPS alone or P. gingivalis LPS alone did not result in significant cytokine production. Representative of three independent experiments.

FIGURE 7

BothE. coli LPS and P. gingivalis LPS activate CD8α⁺ and CD8α⁻ DC subsets in vivo. C57BL/6 mice were injected with either PBS (open red histograms) E. coli LPS, or P. gingivalis LPS (open green histograms), either i.v. or i.p., and 6 h later, the expression of CD80, CD86, and CD40 was assessed on gated, splenic CD11c⁺CD8α⁺ and CD11c⁺CD8α⁻ DC subsets by flow cytometry. Isotype controls are filled purple histograms. Data are representative of three experiments.

FIGURE 8

E. coli LPS, but not P. gingivalis LPS, induces IL-12 in CD8α⁺ DCs. Splenic CD11c⁺CD8α⁺ and CD11c⁺CD8α⁻ DCs were isolated from C57BL/6 mice by microbead enrichment, followed by flow cytometry, and stimulated in vitro with 10 µg/ml E. coli LPS or P. gingivalis LPS. Culture supernatants were assayed for IL-12, IL-6, or TNF-α⁻ 24 h later. Representative of 10 independent experiments.

FIGURE 9

P. gingivalis LPS is less dependent on TLR4 signaling than E. coli LPS. Splenocytes from C3H/HeJ mice and C3H/HeN mice were cultured with varying concentrations of E. coli LPS or P. gingivalis LPS for 72 h. The cultures were pulsed with ³H during the last 12 h of culture. Representative of three independent experiments.

FIGURE 10

IL-6 production by splenocytes stimulated with E. coli LPS or P. gingivalis LPS in C3H/HeN and C3H/HeJ mice. Splenocytes from C3H/HeJ mice and C3H/HeN mice were cultured with varying concentrations of E. coli LPS (A and C) or P. gingivalis LPS (B and D) for 12 h (A and B) or 48 h (C and D). IL-6 was measured by cytokine ELISA. Representative of three independent experiments.

FIGURE 11

Cytokine production by splenic DCs from C3H/HeJ and C3H/HeN mice stimulated in vitro withE. coli LPS or P. gingivalis LPS. CD11c⁺ DCs, enriched from the spleens of mice, were cultured with either type of LPS for 24 h, and secretion of IL-12, IL-6, and TNF-α was measured by cytokine ELISA. Representative of three separate experiments.