Working correlates of protection predict SchuS4-derived-vaccine candidates with improved efficacy against an intracellular bacterium, Francisella tularensis

Abstract

Francisella tularensis, the causative agent of tularemia, is classified as Tier 1 Select Agent with bioterrorism potential. The efficacy of the only available vaccine, LVS, is uncertain and it is not licensed in the U.S. Previously, by using an approach generally applicable to intracellular pathogens, we identified working correlates that predict successful vaccination in rodents. Here, we applied these correlates to evaluate a panel of SchuS4-derived live attenuated vaccines, namely SchuS4-ΔclpB, ΔclpB-ΔfupA, ΔclpB-ΔcapB, and ΔclpB-ΔwbtC. We combined in vitro co-cultures to quantify rodent T-cell functions and multivariate regression analyses to predict relative vaccine strength. The predictions were tested by rat vaccination and challenge studies, which demonstrated a clear relationship between the hierarchy of in vitro measurements and in vivo vaccine protection. Thus, these studies demonstrated the potential power a panel of correlates to screen and predict the efficacy of Francisella vaccine candidates, and in vivo studies in Fischer 344 rats confirmed that SchuS4-ΔclpB and ΔclpB-ΔcapB may be better vaccine candidates than LVS.

Fig. 1. Leukocytes from ΔclpB-vaccinated mice control intramacrophage bacterial growth and effect mediator production in vitro.

BMMΦ from C57BL/6J were infected with LVS and co-cultured with PBLs or splenocytes obtained from naive or vaccinated mice, as indicated. After two or three days of co-culture, BMMΦ were lysed to evaluate the recovery of intracellular bacteria (a). Supernatants from the corresponding co-cultures were collected and analyzed for IFN-γ (b) and NO (c). Values shown are the averages from five independent experiments of similar design. Error bars indicate the standard error of the mean (s.e.m.). Brackets indicate significant differences among vaccine groups, calculated by two-way ANOVA and Tukey’s multiple comparison test (P < 0.05). Symbols indicate values obtained from independent experiments. Data from naive samples were significantly different from the vaccine groups (not shown by brackets, for clarity), with the exception of NO results at 48 h, for which values from naive leukocytes were not significantly different from that of LVS-R.

Fig. 2. Rat leukocytes control intramacrophage bacterial growth and effect mediator production in vitro in a pattern that suggests differential vaccine efficacies.

BMMΦ from Fischer 344 rats were infected with LVS and co-cultured with PBLs (a, c, e) or splenocytes (b, d, f) obtained from naive or vaccinated rats, as indicated. After 2 days of co-culture, BMMΦ were lysed to evaluate the recovery of intracellular bacteria (a, b). Supernatants were collected and analyzed for NO (c, d) and IFN-γ (e, f). Values shown are the average from four independent experiments of similar design. Error bars indicate standard error of the mean (s.e.m.). Brackets indicate significant differences among vaccine groups and calculated by two-way ANOVA and Tukey’s multiple comparison test (P < 0.05). Symbols indicate values obtained from independent experiments. Not shown by brackets are the differences calculated with naive samples against all vaccine groups, which were significant, except vs ΔclpB-ΔfupA (CFU, spleen; NO, PBLs at 48 h; NO, spleen), vs ΔclpB-ΔcapB (NO, spleen at 72 h), and vs ΔclpB-ΔwbtC (NO, spleen at 72 h).

Fig. 3. Titration of rat PBLs in co-cultures highlights a differential in vitro between vaccine groups.

BMMΦ from Fischer rats were infected with LVS and co-cultured with decreasing amounts of PBLs obtained from naive or vaccinated rats, as indicated. After 2 days of co-culture, BMMΦ were lysed to evaluate the recovery of intracellular bacteria (a). Supernatants from the corresponding co-cultures were collected and analyzed for NO (b) and IFN-γ (c). Values shown are the average from samples run in triplicate; circle symbols indicate values for single data points. Error bars indicate the standard error of the mean (s.e.m.) (P < 0.05). Within each PBL dilution, differences were calculated by two-way ANOVA. Matching symbols indicate significant differences in comparison to LVS.

Fig. 4. Relative gene expression of a working panel of correlates of protection in PBLs suggests a hierarchy of vaccine efficacy.

BMMΦ from Fischer rats were infected with LVS and co-cultured with PBLs obtained from naive or vaccinated rats, as indicated. After 2 days of co-culture, PBLs were collected and analyzed for relative gene expression. Semiquantitative analyses of gene expression were performed using a custom array as described in “Methods”. Fold changes were calculated in comparison to naïve cells and data shown are averages of fold change ± standard error of the mean (s.e.m.) calculated from four independent experiments of a similar design. The heat map shown was generated within each gene across vaccine groups; darker colors indicate higher upregulation, while white indicates lower upregulation, within each gene. Matching symbols indicate significant differences in comparison to values observed for LVS-immune cells (P < 0.05).

Fig. 5. Anti-Francisella antibody titers do not show obvious differences across vaccine groups.

Fischer rats (n = 12–24) were vaccinated as indicated. Thirty-nine days after vaccination, serum samples were collected and analyzed for IgG antibodies using plates coated with heat-killed SchuS4. The results from two independent studies are shown; values indicate the median titer for each sample analyzed from one to four times. All vaccines induced antibodies against SchuS4 (P < 0001 vs. PBS), and pair-wise comparisons indicated the only significant difference is between ΔclpB-ΔcapB and ΔclpB-ΔwbtC sera (P = 0.03).

Fig. 6. Survival outcomes after aerosol challenge with SchuS4 of vaccinated rats reflect in vitro activities of leukocytes.

Fischer 344 rats were vaccinated as indicated. Six weeks after vaccination, rats were challenged by aerosol either with a high (b) or low (a) dose of SchuS4. Survival was monitored for 28 days. The experiment was repeated three times with various combinations of the vaccines for a total of 8–20 animals per vaccine-challenge dose group. For each challenge dose, bacterial lung deposition was monitored and was comparable across the three experiments. The figure depicts the time to death and survival of results combined from all three experiments. # and * indicate significant differences between survival outcomes for ΔclpB- and LVS-vaccinated rats, and between survival outcomes for ΔclpB-ΔcapB- and LVS-vaccinated rats, respectively, calculated by Kaplan–Meier and log-rank (Mantel–Cox) analysis.