Salmonella Typhi Bactericidal Antibodies Reduce Disease Severity but Do Not Protect against Typhoid Fever in a Controlled Human Infection Model

Abstract

Effective vaccines against Salmonella Typhi, a major cause of febrile illness in tropical regions, can have a significant effect as a disease control measure. Earlier work has shown that immunization with either of two Salmonella Typhi vaccines, licensed Ty21a or candidate M01ZH09, did not provide full immunity in a controlled human infection model. Here, we describe the human humoral immune responses to these oral vaccines and their functional role in protection after challenge with S. Typhi. Serum, obtained from healthy volunteers before and after vaccination with Ty21a or M01ZH09 or placebo and before and after oral challenge with wild-type S. Typhi, was assessed for bactericidal activity. Single-dose vaccination with M01ZH09 induced an increase in serum bactericidal antibodies (p = 0.001) while three doses of Ty21a did not. No association between bactericidal activity and protection against typhoid after challenge was seen in either vaccine arm. Bactericidal activity after vaccination correlated significantly with delayed disease onset (p = 0.013), lower bacterial burden (p = 0.006), and decreased disease severity scores (p = 0.021). Depletion of antibodies directed against lipopolysaccharide significantly reduced bactericidal activity (p = 0.009). We conclude that antibodies induced after ingestion of oral live-attenuated typhoid vaccines or after challenge with wild-type S. Typhi exhibit bactericidal activity. This bactericidal activity is mediated by anti-O:LPS antibodies and significantly reduces clinical symptoms but does not provide sterile immunity. This directs future vaccine studies toward other antigens or mechanisms of protection against typhoid.

Keywords: Salmonella enterica Typhi; bactericidal activity; human challenge model; immune responses; typhoid infection.

Figure 1

Kinetics of bactericidal antibody induction and decay. (A) Serum bactericidal antibody (SBA) titers in each study arm at D28 (pre-vaccination baseline), D0 (day of challenge), and D60 after challenge. (B) SBA titer kinetics after challenge in TD participants (filled diamonds) from each study arm. (C) SBA titers on D28, D0, and D60 after challenge in each study arm, shown for those with no typhoid diagnosis (nTD, empty circles). (D) SBA titers on the day of challenge, in those with typhoid diagnosis (TD) and nTD across all study arms. n.s., not significant, * p < 0.05; ** p < 0.01; *** p < 0.001. Asterisk within vertical lines indicates significance of pair-wise comparison between M01ZH09 and Ty21a vaccine arms for D0. Significance was determined by Wilcoxon signed rank test for paired value comparisons and by Mann–Whitney U-tests between groups.

Figure 2

Bactericidal antibodies do not protect against typhoid infection but high bactericidal antibody activity reduces disease severity and cytokine response in vaccinated participants. (A–F) Correlation of disease parameters and plasma cytokine levels after diagnosis with SBA titers on D0 in participants from the Ty21a and M01ZH09 vaccine arms. Shown are day of diagnosis (A), bacterial load in blood on the day of diagnosis (B), symptom severity score (C), and maximum concentrations of cytokines TNFα, IL6, and IL8 [(D–F), respectively], in plasma samples collected within 48 h from diagnosis.

Figure 3

Bactericidal antibodies are specific to O9:LPS. SBA titer pre-vaccination (D28) and at D28 after challenge of samples undepleted (UD), or depleted of anti-H, anti-O9:LPS or total IgA antibodies. Diamonds indicate placebo, circles indicate Ty21a and squares indicate M01ZH09 samples. Significance was determined by Wilcoxon signed rank tests.

Figure 4

Kinetics of O9:LPS antibody induction. Anti-O9:LPS IgM (A), IgA (B), and IgG (C) antibodies at baseline (D28), on the day of challenge (D0) and at D28, in those with typhoid diagnosis (TD, diamonds), and no typhoid diagnosis (nTD, empty circles) within each study arm. *p < 0.05, **p < 0.01, ***p < 0.001. Significance was determined by Wilcoxon signed rank tests for paired value comparisons.