Vaccine-induced antibody isotypes are skewed by impaired CD4 T cell and invariant NKT cell effector responses in MyD88-deficient mice

Abstract

The requirement for TLR signaling in the initiation of an Ag-specific Ab response is controversial. In this report we show that a novel OVA-expressing recombinant Salmonella vaccine (Salmonella-OVA) elicits a Th1-biased cell-mediated and serum Ab response upon oral or i.p. immunization of C57BL/6 mice. In MyD88(-/-) mice, Th1-dependent Ab responses are greatly reduced while Th2-dependent Ab isotypes are elevated in response to oral and i.p., but not s.c. footpad, immunization. When the T effector response to oral vaccination is examined we find that activated, adoptively transferred Ag-specific CD4(+) T cells accumulate in the draining lymph nodes, but fail to produce IFN-gamma, in MyD88(-/-) mice. Moreover, CD1d tetramer staining shows that invariant NKT cells are activated in response to oral Salmonella-OVA vaccination in wild-type, but not MyD88(-/-), mice. Treatment with neutralizing Ab to CD1d reduces the OVA-specific Ab response only in MyD88-sufficient wild-type mice, suggesting that both Ag-specific CD4 T cell and invariant NKT cell effector responses to Salmonella-OVA vaccination are MyD88 dependent. Taken together, our data indicate that the type of adaptive immune response generated to this live attenuated vaccine is regulated by both the presence of MyD88-mediated signals and vaccination route, which may have important implications for future vaccine design.

FIGURE 1

A novel Salmonella-OVA vaccine elicits an OVA-specific Ab response in serum and external secretions. A, Parent plasmid pTETnir15 (23) and daughter plasmids pnirOVA and pnirBEM contain an ampicillin (Ap) resistance gene and encode tetanus toxoid fragment C, chicken egg OVA, or no heterologous protein, respectively, under the control of the anaerobically inducible nirB promoter. B, Western blot analysis of bacterial strains carrying indicated plasmid constructs demonstrates that Salmonella carrying pnirOVA (Salmonella-OVA, lanes 3 and 5), but not pnirBEM (Salmonella-BEM, lanes 2 and 4), express OVA protein. Lane 1 contains 25 μg of OVA protein. OVA-specific serum IgG2c (C) and IgA (D) in WT C57BL/6 mice 14 and 28 days after two gastric doses of 2 × 10¹⁰ Salmonella-OVA or Salmonella-BEM measured by ELISA. Twenty-eight days after the second dose, OVA-specific IgA in fecal extracts (E) was measured by ELISA. Data represent three independent experiments (n = 6 mice/ group). Means ± SEM are shown.

FIGURE 2

The absence of MyD88 signaling reduces OVA-specific serum IgG responses to s.c. immunization in the footpad. WT C57BL/6 or MyD88^−/− mice were bled 2 wk after immunization with 150 μg OVA emulsified in CFA and OVA-specific serum IgG1 (A), IgG2c (B), IgG2b (C), and IgG3 (D) measured by ELISA. Pooled data from three independent experiments are shown (n = 10 –11 mice/group). *, p < 0.05. Error bars represent SEM.

FIGURE 3

In the absence of MyD88 signaling, oral vaccination induces serum IgG1 and IgA. C57BL/6 (filled bars) or MyD88^−/− (open bars) mice were given one gastric dose of 3 to 6 × 10¹⁰ Salmonella carrying pnirBEM (Salm-BEM) or pnirOVA (Salm-OVA), and OVA-specific serum IgG2c (A), IgG1 (B), IgG2b (C), and IgA (D) were measured 14, 21, and 28 days postvaccination by ELISA. Pooled data from three independent experiments are shown (n = 9 –13 mice/group). *, p < 0.05. Error bars represent SEM.

FIGURE 4

Intraperitoneal vaccination also induces serum IgG1 in the absence of MyD88 signaling. C57BL/6 (filled bars) or MyD88^−/− (open bars) mice were given one i.p. dose of 2 × 10⁵ Salmonella carrying pnir-BEM (Salm-BEM) or pnirOVA (Salm-OVA), and OVA-specific serum IgG2c (A), IgG1 (B), IgG2b (C), and IgG3 (D) were measured 21, 28, and 35 days postvaccination by ELISA. Pooled data from two independent experiments are shown (n = 3–11 mice/group). *, p < 0.05. Error bars represent SEM. In B, most mice (whether C57BL/6 or MyD88^−/−) did not make an OVA-specific IgG1 response; the few that did were MyD88^−/− mice. Because of this partial response to the vaccine in MyD88^−/− mice, the OVA-specific serum IgG1 responses of MyD88^−/− and C57BL/6 mice vaccinated i.p. with Salmonella-OVA are not statistically significantly different using the unpaired Student t test with unequal variances.

FIGURE 5

Activated Ag-specific CD4⁺ T cells that fail to make IFN-γ accumulate in the draining MLN of MyD88^−/− mice orally vaccinated with Salmonella-OVA. A, Proportion of adoptively transferred CD4⁺Thy1.1⁺ OVA-TCR transgenic OT2 cells in MLN. B, Percentge CD69⁺ among CD4⁺Thy1.1⁺ OT2 cells in MLN. Dots represent individual mice. Lines represent mean percentages for each treatment group. C, Percentage IFN-γ⁺ among CD4⁺Thy1.1⁺ OT2 MLN cells is shown. D, Percentage IL-4⁺ among CD4⁺Thy1.1⁺ OT2 MLN cells is shown. In C and D, 3 days after one gastric dose of ~6 ×10¹⁰ Salmonella, MLN cells were cultured overnight with OVA, pulsed for 4 h with PMA, ionomycin, and GolgiPlug, surface labeled with mAbs against CD4 and Thy1.1, fixed, permeabilized, and intracellularly stained with Abs to IFN-γ or IL-4. Means ± SEM are shown. Salm-BEM indicates fed Salmonella-BEM; Salm-OVA, fed Salmonella-OVA. Pooled data from three independent vaccinations are shown. C57BL/6 Salm-BEM (n = 8), C57BL/6 Salm-OVA (n = 8), MyD88^−/− Salm-BEM (n = 6), MyD88^−/− Salm-OVA (n = 6). *, p < 0.05.

FIGURE 6

IFN-γ production is impaired and IL-4 and IL-17 production enhanced in CD4⁺ and CD4⁻ Thy1.1⁻ cells from the draining MLN of MyD88^−/− mice orally vaccinated with Salmonella-OVA. A, Percentage IFN-γ⁺ among CD4⁺Thy1.1⁻ T cells is shown. B, Percentage IFN-γ⁺ among all host Thy1.1⁻ cells is shown. C, Percentage IL-4⁺ among CD4⁺Thy1.1⁻ T cells is shown. D, Percentage IL-4⁺ among all host Thy1.1⁻ cells is shown. E, Percentage IL-17⁺ among CD4⁺Thy1.1⁻ T cells is shown. F, Percentage IL-17⁺ among all host Thy1.1⁻ cells is shown. In all panels, 3 days after one gastric dose of ~6 × 10¹⁰ Salmonella, MLN cells were cultured overnight with OVA, pulsed for 4 h with PMA, ionomycin, and GolgiPlug, surface labeled with mAbs against CD4 and Thy1.1, fixed, permeabilized, and intracellularly stained with Abs to IFN-γ, IL-4, or IL-17. Means ± SEM are shown (n = 3– 8 mice/group). *, p < 0.05.

FIGURE 7

CD1d-restricted iNKT cell activation in response to oral immunization with Salmonella-OVA is MyD88 dependent. A, Percentage CD3⁺ CD1d tetramer⁺ among liver lymphocytes. B, Absolute number of CD3⁺ CD1d tetramer⁺ cells. In A and B, *, p < 0.05 compared with C57BL/6 naive. C, Bacterial titers in livers of C57BL/6 (●) and MyD88^−/− mice (○) 3 days after one gastric dose of ~6 × 10¹⁰ Salmonella-OVA. Pooled data from two independent vaccinations; dots represent individual mice; lines indicate mean CFU/g tissue. D, Left panel, iNKT cell gate on CD3 intermediate, CD1d tetramer⁺ events in liver. Right panel, Intracellular staining of iNKT cells with rat IgG1 isotype control. E, Representative flow cytometry data showing percentage IL-4⁺ and IFN-γ⁺ among liver iNKT cells from naive and Salmonella-OVA-fed C57BL/6 or MyD88^−/− mice. Numbers in each quadrant represent percentage of liver iNKT cells that are IFN-γ single positive (SP), IL-4 SP, and IFN-γ/IL-4 double positive (DP) 3 days after gastric administration of ~6 × 10¹⁰ Salmonella-OVA and after 4 h of culture with PMA, ionomycin, and GolgiPlug. F, Graphical representation of data in E. Pooled data from four independent vaccinations. Means ± SEM are shown. Naive indicates no treatment; Salm-OVA, fed Salmonella-OVA. C57BL/6 naive (n = 8), C57BL/6 Salm-OVA (n = 10), MyD88^−/− naive (n = 7), MyD88^−/− Salm-OVA (n = 8). *, p < 0.05.

FIGURE 8

Treatment with a neutralizing Ab to CD1d significantly reduces Ag-specific IgG2c responses to oral vaccination in WT mice. C57BL/6 or MyD88^−/− mice were treated or not with 500 μg of CD1.1 blocking Ab i.p. every 2 days starting 1 day before receiving one gastric dose of 4 × 10¹⁰ Salmonella-OVA. OVA-specific serum IgG2c (A), IgG1 (B), IgG2b (C), and IgA (D) were measured 21 days postvaccination by ELISA. Pooled data from two independent experiments are shown (n = 4 –12 mice/group). *, p < 0.05. Error bars represent SEM.