Glycoconjugate vaccine using a genetically modified O antigen induces protective antibodies to Francisella tularensis

Abstract

Francisella tularensis is the causative agent of tularemia, a category A bioterrorism agent. The lipopolysaccharide (LPS) O antigen (OAg) of F. tularensis has been considered for use in a glycoconjugate vaccine, but conjugate vaccines tested so far have failed to confer protection necessary against aerosolized pulmonary bacterial challenge. When F. tularensis OAg was purified under standard conditions, the antigen had a small molecular size [25 kDa, low molecular weight (LMW)]. Using milder extraction conditions, we found the native OAg had a larger molecular size [80 kDa, high molecular weight (HMW)], and in a mouse model of tularemia, a glycoconjugate vaccine made with the HMW polysaccharide coupled to tetanus toxoid (HMW-TT) conferred better protection against intranasal challenge than a conjugate made with the LMW polysaccharide (LMW-TT). To further investigate the role of OAg size in protection, we created an F. tularensis live vaccine strain (LVS) mutant with a significantly increased OAg size [220 kDa, very high molecular weight (VHMW)] by expressing in F. tularensis a heterologous chain-length regulator gene (wzz) from the related species Francisella novicida Immunization with VHMW-TT provided markedly increased protection over that obtained with TT glycoconjugates made using smaller OAgs. We found that protective antibodies recognize a length-dependent epitope better expressed on HMW and VHMW antigens, which bind with higher affinity to the organism.

Keywords: Francisella; O antigen; antibody affinity; glycoconjugate vaccine; intracellular pathogen.

Fig. 1.

F. tularensis OAg is susceptible to acid hydrolysis, but its structural identity is conserved. (A) Silver staining and immunoblot analysis of LPS from F. tularensis (Ft) LVS using an OAg-specific mAb (mAb2034). RU, repeating unit. (B) SEC profiles of HMW and LMW OAgs extracted from F. tularensis LVS. A profile run on a Superose 6 10/300 GL column at 0.5 mL⋅min⁻¹ with 1× PBS (pH 7.4) is shown. The average molecular weight was calculated with a dextran calibration curve. mAU, milli-absorbance unit. (C) ¹H NMR spectrum of HMW and LMW OAgs extracted from F. tularensis LVS.

Fig. 2.

HMW OAg produces a glycoconjugate vaccine conferring protection against intranasal challenge with F. tularensis LVS. (A) Immunization schedule. BALB/c mice received three immunizing doses (2 wk apart) and were challenged intranasally after a 4-wk rest period with F. tularensis LVS (50-fold the intranasal LD₅₀). Animals were monitored daily for 25 d after challenge. (B) Anti-OAg IgG ELISA units to the F. tularensis (Ft) LPS coating agent measured at day 49. Dots represent individual animals. Horizontal bars represent mean ± SD values. i.n., intranasally. (C) Survival of mice challenged intranasally (IN) with F. tularensis after immunization with HMW-TT, LMW-TT, or F. tularensis ∆kdhAB or administration of adjuvant alone. There were six mice per group. **P ≤ 0.01.

Fig. 3.

Heterologous expression of Wzz2 from F. novicida to produce OAg of larger molecular size in F. tularensis. (A) Silver staining analysis comparing LPS from F. tularensis (Ft) LVS with LPS from F. novicida U112. (B) Generation of an Ft LVS mutant with an increased OAg size by heterologous expression of chain-length regulator gene wzz2 from the related subspecies F. novicida. (C) SEC profile of VHMW in comparison to HMW and LMW OAgs. This profile was run on a Superose 6 10/300 GL column at 0.5 mL⋅min⁻¹ with 1× PBS (pH 7.4). The average molecular weight was calculated with a dextran calibration curve.

Fig. 4.

VHMW OAg produces a glycoconjugate vaccine conferring protection against intranasal challenge with a high dose of F. tularensis LVS. (A) Anti-OAg IgG ELISA units to the F. tularensis (Ft) LPS coating agent measured at day 49. Dots represent individual animals. Horizontal bars represent mean ± SD values. i.n., intranasal. (B) Survival of mice intranasally (IN) challenged with Ft (100-fold the IN LD₅₀) after immunization with VHMW-TT, HMW-TT, LMW-TT, or Ft ∆kdhAB or administration of adjuvant alone. There were eight mice per group. **P ≤ 0.01. ns, not significant.

Fig. 5.

Size of the OAg affects relative affinity to LPS-specific antibodies. (A) Anti-OAg IgG ELISA units to the F. tularensis ∆wzz1/wzz2^Fn LPS coating agent measured at day 49. Dots represent individual animals. Horizontal bars represent mean ± SD values. (B) F. tularensis (Ft) OAg competitive ELISA. Recognition of Ft OAg-specific polyclonal antibody by UV-killed Ft-coated ELISA plates in the presence of LMW (orange), HMW (green), or VHMW (red) OAg as a soluble competitor is shown. The inhibition percentage is calculated in relation to the ELISA signal (A₄₀₅) with no competition. Data points represent competition percentage values at indicated inhibitor concentrations. Each data point is the mean of duplicate determinations from a representative experiment. Horizontal bars represent mean ± SD values. *P ≤ 0.05.