Preclinical development of an in vivo BCG challenge model for testing candidate TB vaccine efficacy

Abstract

There is an urgent need for an immunological correlate of protection against tuberculosis (TB) with which to evaluate candidate TB vaccines in clinical trials. Development of a human challenge model of Mycobacterium tuberculosis (M.tb) could facilitate the detection of such correlate(s). Here we propose a novel in vivo Bacille Calmette-Guérin (BCG) challenge model using BCG immunization as a surrogate for M.tb infection. Culture and quantitative PCR methods have been developed to quantify BCG in the skin, using the mouse ear as a surrogate for human skin. Candidate TB vaccines have been evaluated for their ability to protect against a BCG skin challenge, using this model, and the results indicate that protection against a BCG skin challenge is predictive of BCG vaccine efficacy against aerosol M.tb challenge. Translation of these findings to a human BCG challenge model could enable more rapid assessment and down selection of candidate TB vaccines and ultimately the identification of an immune correlate of protection.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Figure 1. Timescale showing BCG persistence in the ears and LNs of id-injected mice up to 12 weeks post immunization.

Log BCG CFU in the ears (estimated by culture) are shown for (a) the high dose group (7000 CFU id); and (b) the low dose group (60 CFU id). Log BCG CFU in the auricular LNs are shown in (c) the high dose group (7000 CFU id); and (d) the mid dose group (60 CFU id). Datasets include individual data points for each mouse; the bars represent the median per group in (a) and (b), and a line connects the means for each group in (c) and (d). * indicates P<0.05. Both ears and LNs were homogenized and plated onto 7H11 Middlebrook agar. Log BCG CFU in the ears estimated by culture (e) and BCG genome copies/mHPRT copies estimated by PCR (f) are shown in a timescale for the high dose group, up to 12 weeks post BCG immunization. Quantitative PCR was performed with BCG-specific primers. Individual data points are shown for each mouse with a line connecting the means for each group.

Figure 2. Splenic IFN-γ responses to PPD up to 12 weeks post BCG immunization.

In (a) high dose group (7000 CFU id); (b) mid dose group (60 CFU id) and (c) low dose group (1 CFU id). Datasets include individual data points for each mouse; the bars represent the median value per group. Results expressed as SFC/million splenocytes.

Figure 3. Effect of single dose vaccines (subunits MVA85A and Ad85A, and BCG) on an id BCG challenge.

(a) BALB/c mice were immunized id with 1×106 pfu MVA85A or 2×109 vp Ad85A. Control mice (Naïve) received no immunization. Four weeks later all mice were challenged id with 1×105 CFU BCG, contralaterally to the site of vaccination. Ears and LNs were harvested 4 weeks after BCG challenge and processed for CFU quantification. (*P<0.05, n = 10 except naïves, n = 5). (b) Corresponding intracellular cytokine staining (ICS) of the local draining LNs. Red bars represent the proportion of IFN-γ-secreting CD4+ T cells in response to 85A, blue bars represent the same for CD8+ T cells (M, n = 4; Naïve, n = 3; Ad, n = 4. *P<0.05). (c) Effect of BCG vaccine compared to subunit MVA85A on an id BCG challenge. BALB/c mice were immunized id with either 1×106 pfu MVA85A or 2.2×104 cfu BCG. “Naïve” and antibiotic-treated (I+R) mice received no immunization. Four weeks later all mice were challenged with 6×103 CFU BCG, except the BCG control group who received no challenge. In the I+R group, challenge was followed by 4 weeks treatment with isoniazid and rifampicin. Ears and LNs were harvested 4 weeks after BCG challenge and processed for CFU quantification. Log10 BCG CFU individual data points for each mouse are shown. Bars represent the median per group. (c) Ears (**P<0.01, “non-significant, ND”, n = 10); (d) LNs (**P<0.01, “non-significant, ND”, n = 10).

Figure 4. Effect of BCG on a 4-week id and in BCG challenge.

BALB/c mice were immunized id with 2.5×103 CFU BCG. Naïve and antibiotic-treated (I+R) mice received no immunization. Four weeks later all mice were challenged either id or in with 4×103 CFU BCG, except the BCG control group who received no challenge. Immediately post-challenge the I+R group was treated for 4 weeks with isoniazid and rifampicin. Ears (a), LNs (b) and lungs (c) were harvested 4 weeks after challenge (spleen cfu data not shown). (d) shows the effect of in BCG on in challenge in a separate experiment. Here, BALB/c mice were immunized in with 1×103 CFU BCG. Naïve mice received no immunization. 4 weeks later in-immunized and naïve mice were challenged with 4×104 CFU BCG in. The “BCG in control” group received no challenge. Lungs were harvested 4 weeks after BCG challenge in all groups. CFU from plating of fresh tissues are shown. Log10 BCG CFU individual data points for each mouse are shown. Bars represent the median per group. (a) Ears (**P<0.01, n = 10); (b) LNs (**P<0.01, n = 10); (c) Lungs (**P<0.01, n = 10); (d) Lungs P<0.01, n = 8 naïve; n = 6 BCG in; n = 8 BCG in control).

Figure 5. Effect of BCG immunization on a 16-week id BCG challenge.

Timeline is shown in (a). BALB/c mice were immunized id with 1×104 CFU BCG. Naïve mice received no immunization. 16 weeks later all mice were challenged with 103 CFU BCG, except the (BCG+(I+R)(no challenge)) group. This group and the BCG+(I+R) group received 6 weeks of isoniazid and rifampicin (starting 8 weeks post initial BCG immunization). In the BCG+(I+R) group, there was a 2 week wash-out period between cessation of antibiotics and subsequent BCG challenge. (b) shows the splenic IFN-γ ELISpot responses to PPD and TB10.3, in naïve, BCG-vaccinated, and BCG+(I+R)(no challenge) animals, 4 weeks post BCG challenge. Whiskers represent minimum to maximum values, boxes the interquartile range, and the bars the median values for each group (*P<0.05, **P<0.01, “non-significant”, ND. n = 10, except for BCG+(I+R); n = 2). Ears and LNs were harvested 4 weeks after BCG challenge and processed for CFU quantification. Log10 BCG CFU of challenge are shown, for groups 1 (Naïve), 2 (BCG) and 3 (BCG+(I+R)). For group 4 (BCG+(I+R) no challenge), the ear and LN CFU correspond to the CFU remaining from the priming BCG immunization (zero in all animals). Individual data points for each mouse are shown. Bars represent the median per group. (c) Ears (*P<0.05,**P<0.01, n = 10); (d) LNs (**P<0.01, n = 10).

Figure 6. Effect of BCG, BCG-MVA85A (B–M), and BCG-Ad85A (B-Ad) prime-boost regimes on an id BCG challenge.

Timeline shown in (a): BALB/c mice were immunized with 104 CFU BCG id and then boosted after 13 weeks with either 1×106 pfu MVA85A id or 2×109 vp Ad85A id. All animals were challenged 4 weeks later with 6×103 CFU of BCG id. Organs were harvested 4 weeks after challenge. Log10 BCG CFU of challenge are shown in (c) Ears (***P<0.001, **P<0.01, n = 8); and (d) LNs (***P<0.001, **P<0.01, n = 8, “NS” = no significant difference). **/*** indicate significance of immunization regimes over naïve mice. (b) Pre-challenge immunogenicity as measured in the blood by ELISpot. IFN-γ responses were assessed after stimulation with PPD, a H-2d CD4+ T cell epitope, and a H-2d CD8+ T cell epitope present in the M.tb antigen 85A, on blood samples taken one day pre- BCG challenge. IFN-γ responses within all three vaccinated groups (BCG (B), BCG-MVA85A (B-M), and BCG-Ad85A, (B-Ad), n = 10) are shown. Whiskers represent minimum to maximum values, boxes the interquartile range, and bars the median values for each group. Correlations between LN cfu and pre-challenge blood ELISpot responses to (e) PPD; (f) CD4+ epitope; and (g) CD8+ epitope. Spearman correlation analysis (with individual data-points for all 30 mice in the three vaccination groups) is shown.