Functional in-vitro evaluation of the non-specific effects of BCG vaccination in a randomised controlled clinical study

Abstract

Bacille Calmette-Guérin (BCG), the only currently licenced tuberculosis vaccine, may exert beneficial non-specific effects (NSE) in reducing infant mortality. We conducted a randomised controlled clinical study in healthy UK adults to evaluate potential NSE using functional in-vitro growth inhibition assays (GIAs) as a surrogate of protection from four bacteria implicated in infant mortality. Volunteers were randomised to receive BCG intradermally (n = 27) or to be unvaccinated (n = 8) and were followed up for 84 days; laboratory staff were blinded until completion of the final visit. Using GIAs based on peripheral blood mononuclear cells, we observed a significant reduction in the growth of the Gram-negative bacteria Escherichia coli and Klebsiella pneumonia following BCG vaccination, but no effect for the Gram-positive bacteria Staphylococcus aureus and Streptococcus agalactiae. There was a modest association between S. aureus nasal carriage and growth of S. aureus in the GIA. Our findings support a causal link between BCG vaccination and improved ability to control growth of heterologous bacteria. Unbiased assays such as GIAs are potentially useful tools for the assessment of non-specific as well as specific effects of TB vaccines. This study was funded by the Bill and Melinda Gates Foundation and registered with ClinicalTrials.gov (NCT02380508, 05/03/2015; completed).

Figure 1

Consort diagram of volunteer recruitment and study schedule. Volunteers (see Table 1 for demographics) were enrolled in two phases and randomised to receive either BCG vaccination or no intervention (a). Volunteers had blood samples collected at screening and days 0 (baseline), 2, 4, 7, 10, 14, 21, 28 and 84. Volunteers from Birmingham followed the same schedule with the exception of follow-up visits on days 4 and 10 which were omitted for logistical reasons. Nasal swabs for determination of S. aureus carriage status were taken at screening and days 0 and 14 (b).

Figure 2

PPD-specific IFN-γ ELISpot responses and control of mycobacterial growth in the direct PBMC MGIA are enhanced following BCG vaccination. Samples were taken from volunteers enrolled into phases 1 and 2 combined. Healthy UK adults were randomised to receive BCG vaccination (n = 27) or to be unvaccinated controls (n = 8). PPD-specific IFN-γ ELISpot responses were measured at baseline and days 7, 14, 21, 28 and 84 in volunteers who received BCG vaccination (a) and those who were unvaccinated controls (b). The direct PBMC MGIA was conducted on cells and serum taken at baseline and days 28 and 84 from volunteers who received BCG vaccination (c) and those who were unvaccinated controls (d). Bars represent the mean values with the standard error of the mean (SEM); dotted lines indicate the baseline mean. For A–B, Wilcoxon tests were performed of each time-point vs. baseline. For C–D, a paired t-test was performed of each time-point vs. baseline. *Indicates a p-value of < 0.05, **indicates a p-values of < 0.005 and ***indicates a p-value of < 0.001. SFC = spot-forming cells.

Figure 3

No effect of BCG vaccination on growth of Gram-positive bacteria in the PBMC GIA. Samples were taken from volunteers enrolled into phases 1 and 2 combined. Healthy UK adults were randomised to receive BCG vaccination (n = 27) or to be unvaccinated controls (n = 8). PBMC bacterial GIAs were conducted on samples taken at baseline and days 2, 7, 14, 28 and 84 following BCG vaccination (closed circles) and at the same time-points in unvaccinated control individuals (open circles). PBMC and autologous serum were co-cultured with S. aureus (a, b) or S. agalactiae (c, d) for 1 h after which time cells were lysed and bacteria quantified by plating on solid blood agar. Bars represent the median values with the interquartile range (IQR); dotted lines indicate the baseline median.

Figure 4

Significant effect of BCG vaccination on growth of Gram-negative bacteria in the PBMC GIA. Samples were taken from volunteers enrolled into phases 1 and 2 combined. Healthy UK adults were randomised to receive BCG vaccination (n = 27) or to be unvaccinated controls (n = 8). PBMC bacterial GIAs were conducted on samples taken at baseline and days 2, 7, 14, 28 and 84 following BCG vaccination (closed circles) and at the same time-points in unvaccinated control individuals (open circles). PBMC and autologous serum were co-cultured with E. coli (a, b) or K. pneumoniae (c, d) for 1 h after which time cells were lysed and bacteria quantified by plating on solid blood agar. Bars represent the median values with the interquartile range (IQR); dotted lines indicate the baseline median. Paired t-tests were performed of each time-point vs. baseline, where *indicates a p-value of < 0.05 and **indicates a p-value of < 0.005.

Figure 5

S. aureus growth in the GIA stratified by S.aureus carriage status. The anterior nares of all volunteers were sampled at screening and days 0 and 14. Individuals were considered ‘persistent’ carriers if two or more consecutive cultures were positive, ‘intermittent’ carriers if one or more non-consecutive cultures were positive or ‘non-carriers’ if all 3 cultures were negative. Baseline S. aureus GIA results stratified by carriage status are shown for whole blood (a) and PBMC (b). Stratified using a reclassification of carriage types into two categories (‘persistent’ carriers and ‘others’) proposed by Van Belkum et al., S. aureus GIA baseline results are shown for whole blood (c) and PBMC (d). Bars represent the median values with the interquartile range (IQR). A Mann–Whitney test was performed where *indicates a p-value of < 0.05 and **indicates a p-value of < 0.005.