Reducing the activity and secretion of microbial antioxidants enhances the immunogenicity of BCG

Abstract

Background: In early clinical studies, the live tuberculosis vaccine Mycobacterium bovis BCG exhibited 80% protective efficacy against pulmonary tuberculosis (TB). Although BCG still exhibits reliable protection against TB meningitis and miliary TB in early childhood it has become less reliable in protecting against pulmonary TB. During decades of in vitro cultivation BCG not only lost some genes due to deletions of regions of the chromosome but also underwent gene duplication and other mutations resulting in increased antioxidant production.

Methodology/principal findings: To determine whether microbial antioxidants influence vaccine immunogenicity, we eliminated duplicated alleles encoding the oxidative stress sigma factor SigH in BCG Tice and reduced the activity and secretion of iron co-factored superoxide dismutase. We then used assays of gene expression and flow cytometry with intracellular cytokine staining to compare BCG-specific immune responses in mice after vaccination with BCG Tice or the modified BCG vaccine. Compared to BCG, the modified vaccine induced greater IL-12p40, RANTES, and IL-21 mRNA in the spleens of mice at three days post-immunization, more cytokine-producing CD8+ lymphocytes at the peak of the primary immune response, and more IL-2-producing CD4+ lymphocytes during the memory phase. The modified vaccine also induced stronger secondary CD4+ lymphocyte responses and greater clearance of challenge bacilli.

Conclusions/significance: We conclude that antioxidants produced by BCG suppress host immune responses. These findings challenge the hypothesis that the failure of extensively cultivated BCG vaccines to prevent pulmonary tuberculosis is due to over-attenuation and suggest instead a new model in which BCG evolved to produce more immunity-suppressing antioxidants. By targeting these antioxidants it may be possible to restore BCG's ability to protect against pulmonary TB.

Figure 1. Construction of modified BCG vaccines.

(A) Southern hybridization demonstrating inactivation of secA2 and sigH. DraIII-digested DNA from each strain was probed for secA2, sigH and the hygR cassette used to inactivate sigH. Lane 1, BCG Tice; lane 2, BCGΔsecA2; lane 3, BCGΔsigH; lane 4, BCGΔsecA2ΔsigH (DDBCG); lane 5, a colony of BCGΔsecA2 with only one copy of sigH inactivated. secA2 contains two internal DraIII sites and DraIII-digested DNA was predicted to yield fragments of 307 bp, 785 bp, and 10,544 bp. sigH was predicted to be on a 2895 bp fragment. The hygR cassette contains an internal DraIII site and was predicted to yield 2938 bp and 1612 bp fragments after inactivation of sigH. (B) Differences in the localization and activity of SodA in BCG, DDBCG, and 3dBCG. A representative experiment shows the SOD units per ml of concentrated supernatant or lysate. (C) Immunoblot showing comparable amounts of SodA in lysates of DDBCG (lane 1) and 3dBCG (lane 2), despite the marked difference in SOD activity as shown in (b).

Figure 2. Cytokine-producing CD4+ T-cells in the spleens of vaccinated mice after intravenous challenge.

(A) Representative plots show IFN-γ+, TNF-α+ and IFN-γ+TNF-α double-positive cytokine production by splenocytes of mice vaccinated subcutaneously with BCG or 3dBCG at 28 days after intravenous challenge with 2×10⁷ CFU of BCG. Splenocytes were re-stimulated on IFN-γ-treated uninfected bone-marrow derived macrophages and BCG-infected macrophages. Percent values represent frequency of cytokine+ CD4+ T cells for each stimulation condition. (B) Number of IFN-γ+, TNF-α+, IL-2+ and dual cytokine+ CD4+ T cells in spleens 26 to 30 days after intravenous challenge with BCG. Each data point represents the number of BCG-specific cytokine+ T cells in one mouse spleen, calculated from percent values of cytokine+ CD4+ T cells, as in (a), and total number of splenocytes isolated from each mouse. The bars represent median values from 7 to 9 mice in each vaccinated group. Media/- mice are age-matched controls that were vaccinated with media but not challenged, C = challenged. The variances of the median values of the three challenged groups were compared by the Kruskal-Wallis non-parametric test: IFN-γ, P = 0.0031; TNF-α, P = 0.0006; IL-2, P = 0.0165; IFN-γ+TNF-α, P = 0.0007; and IFN-γ+IL-2, P = .0131. The Mann-Whitney P values for two group comparisons are displayed in the figure.

Figure 3. Spleen mycobacterial titers post-challenge and correlation between bacilli counts and T cell responses.

(A) BCG colony-forming units (CFU) persisting in the spleens of mice on day 26–30 after intravenous challenge with 2×10⁷ CFU of BCG. Data are from six mice per group from two experiments. Individual and median values are displayed. The variance of the medians of the three groups was significant, P = 0.0186, Kruskal-Wallis test. Mann-Whitney P values for two-group comparisons are shown in the figure. (B) Plot of median Log₁₀ CFU in spleens of vaccinated mice versus the median CD4+ IFN-γ responses for the three vaccination arms. Coefficient of correlation, R² was calculated using median values from each vaccination group.

Figure 4. Kinetics of the primary immune response and peak CD8+ T cell responses.

(A) Numbers of IFN-γ+ CD8+ T-cells in spleens after re-stimulation with TB10.3/10.4_20-28 peptide (GL9) at 8, 9, 12, 17 and 30 days after intravenous inoculation of media, BCG or 3dBCG. The symbols and error bars represent median and range of responses of 3 to 9 mice per time point. (B) Numbers of IFN-γ+ CD4+ T cells in spleens after re-stimulation with TB10.4_74-88 peptide (ST15) on days after vaccination. (C) Numbers of IFN-γ+, and IFN-γ+TNF-α+, and IFN-γ+IL-2+ CD8+ T cells in spleens after re-stimulation with TB10.3/10.4_20-28 peptide - GL9 at the peak of the primary CD8+ T cell response (day 9 post-inoculation). Each symbol represents the number of cytokine-producing CD8+ T cells from one mouse. The bars represent median values for each group and Mann-Whitney P values for two-group comparisons are displayed in the figure. The variances of the medians of the three groups were significant: IFN-γ, P = 0.0009; IFN-γ+TNF-α, P = 0.0013; and IFN-γ+IL-2, P = 0.0415, Kruskal-Wallis test.

Figure 5. Memory CD4+ T cell responses after subcutaneous vaccination with BCG or 3dBCG.

Kinetics of (A) IFN-γ+ and (B) IL-2+ CD4+ T cell response in spleens of subcutaneously vaccinated mice in response to TB10.4_74-88 peptide - ST15 stimulation. The symbols represent median and range of cytokine responses of 3–7 mice per time point. (C) Numbers of IFN-γ+ and IL-2+ CD4+ T cells in spleens and the IL-2+/IFN-γ+ ratio of CD4+ T cells in each spleen after re-stimulation with TB10.4_74-88 peptide - ST15 at the memory phase of the primary CD4+ T cell response (day 30 post-vaccination). Each data point represents the number of cytokine+ CD4+ T cells or the ratio from one mouse, with 7 mice per group. The bars represent median values for each group and Mann-Whitney P values for two-group comparisons are displayed in the figure. The variances of the medians of the three groups were significant: IFN-γ, P = 0.0012; and IL-2, P = 0.0005, Kruskal-Wallis test.

Figure 6. Expression of genes encoding Th polarizing cytokines.

Mice were inoculated by the retro-orbital route and spleens harvested 72 hours later. Each data point represents the level of gene expression by individual mice in each group as determined by RT-PCR and normalized to the mean value of the PBS-vaccinated group. The bars represent median values for each group and Mann-Whitney P values for two-group comparisons are displayed in the figure. Statistical comparisons of the variances of the medians of the three groups were: IL-12b, P = 0.0006; IFN-γ, P = 0.0033; IL-4, P = 0.0702; IL-21, P = 0.0011; IL-6, P = 0.0033; and RANTES, P = 0.0112, Kruskal-Wallis test.

Figure 7. In vitro growth and in vivo persistance of BCG and 3dBCG.

(A) Plot of the absorbance of cultures of BCG and 3dBCG during ten days of cultivation. A representative experiment shows A₆₀₀ values every two days after Middlebrook 7H9 media containing Tween 80 in roller bottles was seeded with BCG or 3dBCG. The initial inoculum was adjusted to an A₆₀₀ value of 0.2. (B) Spleen CFU at four weeks after intravenous inoculation of Balb/C mice with BCG or 3dBCG. Inocula were prepared based on absorbance values and then titered to determine the number of viable bacilli administered to each group of three mice – 6.6×10⁶ CFU for BCG; 7.7×10⁶ CFU for 3dBCG. The numbers represent the weights of the spleens in grams.