Impact of vector priming on the immunogenicity of recombinant Salmonella vaccines

Abstract

There are conflicting reports concerning the impact of prior vector priming on the immunogenicity of recombinant-Salmonella-based vaccines. A comparison of experimental protocols identified two variables which might account for this inconsistency: the potential of the vector strain to colonize the murine gut-associated lymphoid tissue (GALT) and the nature of the foreign antigen subsequently delivered by the recombinant Salmonella construct. The former was investigated by constructing an aroA mutant of the Salmonella enterica serovar Stanley vector previously used in our laboratory. Although the introduction of an aroA mutation had surprisingly little effect on GALT colonization, it did reduce the strength of antilipopolysaccharide (anti-LPS) antibody responses and the impact of vector priming. Studies were also performed to ascertain the extent to which any observed hyporesponsiveness consequent upon vector priming might be determined by the characteristics of the foreign antigen. S. enterica serovar Stanley was used to deliver either of two Escherichia coli antigens, K88 pilus protein or the LT-B toxin subunit, to vector-primed mice. Both serum immunoglobulin G (IgG) and intestinal IgA responses to K88 were completely abolished, and those to LT-B were significantly reduced, as a consequence of vector priming. When similar experiments were performed with an aroA S. enterica serovar Dublin vector, responses to K88 were significantly reduced but those to LT-B were unaffected by vector priming. Paradoxically, a priming infection with this vector induced stronger anti-LPS antibody responses but was less likely to elicit a state of hyporesponsiveness to subsequently presented foreign antigen. The impact of vector priming thus depends on both the Salmonella strain used and the nature of the foreign antigen, but our present data strengthen concerns that preexisting antivector immunity represents a serious threat to the Salmonella-based vaccine strategy.

FIG. 1.

Peyer's patch colonizing potential of aroA mutants. (A) Mice were fed wt or aroA S. enterica serovar Stanley as follows: 1.1 × 10⁹ CFU of wt S. enterica serovar Stanley on day 0 (□); 1.3 × 10⁹ CFU of aroA S. enterica serovar Stanley on day 0 (▧); 1.5 × 10¹⁰ CFU of aroA S. enterica serovar Stanley on day 0 and 1.2 × 10¹⁰ CFU of aroA S. enterica serovar Stanley on day 4 (▪). (B) Mice were fed aroA S. enterica serovar Dublin, either 1.2 × 10⁹ CFU on day 0 (▩) or 9.4 × 10⁹ CFU on day 0 and 9.5 × 10⁹ CFU on day 4 (▥). Histograms show log₁₀ bacterial burden (GM ± SD; n = 4) in Peyer's patches; lines show limit of detection (20 bacteria).

FIG. 2.

Impact of priming with wt or aroA S. enterica serovar Stanley on subsequent responsiveness to S. enterica serovar Stanley-K88. Mice were orally primed with wt (○, 1.5 × 10⁹ CFU on day −70) or aroA S. enterica serovar Stanley (▪, 1.7 × 10⁹ CFU on day −70; □, 1.9 × 10¹⁰ CFU on day −70 and 9.6 × 10⁹ CFU on day −66) or were held as unprimed controls (•). Ten weeks later, all mice were orally dosed with 1.7 × 10⁹ CFU of S. enterica serovar Stanley-K88. Serum IgG (A and B) and intestinal IgA (C and D) responses to K88 (A and C) and LPS (B and D) were determined by ELISA, sampling alternate subsets of four mice at successive time points. Serum IgG responses are represented as log₁₀ ELISA titers (GM ± SD; n = 4), while gut IgA responses are expressed as log₁₀ ELISA titers (GM ± SD; n = 4) per milligram of IgA. For clarity, SDs are not shown for groups primed with aroA S. enterica serovar Stanley (see the text). Horizontal lines show the limit of detection of serum ELISAs at a titer of 20.

FIG. 3.

Anti-LPS responses induced by oral immunization with wt or aroA S. enterica serovar Stanley or with aroA S. enterica serovar Dublin. Groups of mice were immunized with wt (○, 1.1 × 10⁹ CFU on day 0) or aroA S. enterica serovar Stanley (▪, 1.3 × 10⁹ CFU on day 0; □, 1.6 × 10¹⁰ CFU on day 0 and 1.1 × 10¹⁰ CFU on day 4) or with aroA S. enterica serovar Dublin (♦, 1.1 × 10¹⁰ CFU day 0 and 1.0 × 10¹⁰ CFU on day 4). Serum IgG (A) and FPS IgA (B) responses to LPS are shown as log₁₀ ELISA titers (GM ± SD; n = 5), the latter being standardized per milligram of IgA. For clarity, SDs are not shown for one group. The horizontal line shows the limit of detection of serum ELISAs at a titer of 20.

FIG. 4.

Immunogenicity of S. enterica serovar Stanley-LT-B in vector-primed mice. Mice were orally dosed with either 1.2 × 10⁹ CFU of wt S. enterica serovar Stanley (○) or NaHCO₃ (•) on day −70; 10 weeks later both groups received 1.3 × 10⁹ CFU of S. enterica serovar Stanley-LT-B. Serum IgG (A and B) and intestinal IgA (C and D) responses to LT-B (A and C) and LPS (B and D) were determined by ELISA on serum and FPS samples (alternate subsets of five mice were sampled at successive time points). Serum responses are log₁₀ ELISA titers (GM ± SD; n = 5), while gut IgA responses are log₁₀ ELISA titers per milligram of IgA (GM ± SD; n = 5). Horizontal lines represent the limit of detection of serum ELISAs at a titer of 20. *, P < 0.05; **, P < 0.01 (two-tailed t test).

FIG. 5.

Scatter plots of anti-LT-B responses. Individual serum (A) and gut (B) anti-LT-B responses from vector-primed (○) and control (•) mice (from Fig. 4) are shown as scatter plots. The horizontal line shows the limit of detection of serum ELISAs at a titer of 20.

FIG. 6.

Immune responses to recombinant Salmonella in mice primed with aroA S. enterica serovar Dublin. Mice were orally primed with aroA S. enterica serovar Dublin (○, 1.4 × 10¹⁰ CFU on day −70 and 1.3 × 10¹⁰ CFU on day −66) with other mice set aside as unimmunized controls (•). On days 0 and 4, paired groups of control and primed mice were given boosters of either EL23 (aroA S. enterica serovar Dublin-LT-B; 9.5 × 10⁹ followed by 1.6 × 10¹⁰ CFU [A1 to A4]) or SL1438-K88 (9.5 × 10⁹ followed by 2.0 × 10¹⁰ CFU [B1 to B4]). Graphs A1 and A3 show serum IgG and gut IgA responses to LT-B, whereas graphs B1 and B3 show serum IgG and gut IgA responses to K88. Serum IgG (A2 and B2) and gut IgA (A4 and B4) responses to LPS are also shown. Serum responses are log₁₀ ELISA titers (GM ± SD; n = 5), while gut IgA responses are log₁₀ ELISA titers per milligram of IgA (GM ± SD; n = 5). Horizontal lines show the limit of detection of serum ELISAs at a titer of 20. *, P < 0.05; **, P < 0.01 (two-tailed t test).