Intravenous Bacille Calmette-Guérin vaccination protects simian immunodeficiency virus-infected macaques from tuberculosis

Abstract

Tuberculosis, caused by Mycobacterium tuberculosis (Mtb), is the most common cause of death in people living with human immunodeficiency virus (HIV). Intra-dermal Bacille Calmette-Guérin (BCG) delivery is the only licensed vaccine against tuberculosis; however, it offers little protection from pulmonary tuberculosis in adults and is contraindicated in people living with HIV. Intravenous BCG confers protection against Mtb infection in rhesus macaques; we hypothesized that it might prevent tuberculosis in simian immunodeficiency virus (SIV)-infected macaques, a model for HIV infection. Here intravenous BCG-elicited robust airway T cell influx and elevated plasma and airway antibody titres in both SIV-infected and naive animals. Following Mtb challenge, all 7 vaccinated SIV-naive and 9 out of 12 vaccinated SIV-infected animals were protected, without any culturable bacteria detected from tissues. Peripheral blood mononuclear cell responses post-challenge indicated early clearance of Mtb in vaccinated animals, regardless of SIV infection. These data support that intravenous BCG is immunogenic and efficacious in SIV-infected animals.

Fig. 1. Plasma viraemia of SIV+ unvaccinated and vaccinated animals.

Plasma viral copy equivalents were determined by qRT–PCR. Each point indicates an individual animal (SIV/Unvax, n = 4; SIV/IV BCG, n = 12). a, Plasma viral load (in copy equivalents (CEQ) ml⁻¹) of SIV+ unvaccinated (top, red up-pointing triangles) and SIV+ vaccinated (bottom, gold down-pointing triangles) animals over the course of the study. Horizontal dashed line represents the limit of detection. b, Plasma viral load of each group before (pre), 2 weeks and 12 weeks, relative to BCG vaccination. Time-matched plasma from SIV+ unvaccinated animals served as a control. Repeated measure one-way ANOVAs were performed on each group (SIV/Unvax, P = 0.5759; SIV/IV BCG, P < 0.0001). Multiple comparisons relative to ‘pre-BCG’ were performed using Dunnett’s multiple comparisons tests. Adjusted P values for multiple comparisons are shown. Source data

Fig. 2. Leukocyte and T cell subsets and T cell response in BAL after BCG vaccination.

a, Number of leukocytes per BAL collection. b, Number of T cell subset cells per BAL collection. c, Number of cytokine-responsive (IFNγ, TNF and IL-2) CD4⁺ and CD8⁺ T cells in BAL after 14 h stimulation with PPD. d, Number of IL-17A-responsive CD4⁺ and CD8⁺ T cells in BAL after 14 h stimulation with PPD. Data are mean and s.d. of SIV-naive (light blue) and SIV+ (gold) vaccinated animals (a–d). SIV-naive vaccinated animals: pre-BCG (n = 5), 4 weeks post-BCG (n = 6) and 12 weeks post-BCG (n = 6). SIV+ vaccinated animals: pre-BCG (n = 10), 4 weeks post-BCG (n = 11) and 12 weeks post-BCG (n = 8). Individual animal data are shown in Supplementary Data 5. Mixed effects models with subject as a random variable were used to assess mean differences among timepoints and vaccine groups. Timepoints were compared with ‘pre-BCG’ control using Dunnett’s multiple comparison tests. No significant differences between vaccination groups were found. Significant P values (P < 0.05) across time are shown above each graph. Fixed effect test results and Dunnett’s multiple comparisons are included in Supplementary Table 2a. All statistical tests were two-sided. Source data

Fig. 3. Frequencies of leukocyte and T cell subsets in PBMC after vaccination.

a, Frequencies of leukocyte subsets relative to BCG. mDCs, myeloid dendritic cells; pDCs, plasmacytoid dendritic cells. b, Frequencies of T cells and T cell subsets (MAIT, Vγ9⁺, CD4⁺ and CD8⁺ T cells) relative to BCG. Data are mean and s.d. of SIV-naive (IV BCG, light blue) and SIV+ (SIV/IV BCG, gold) vaccinated animals. SIV-naive vaccinated animals: pre-BCG (n = 7), 4 weeks post-BCG (n = 7) and 12 weeks post-BCG (n = 7). SIV+ vaccinated animals: pre-BCG (n = 12), 4 weeks post-BCG (n = 12) and 12 weeks post-BCG (n = 11). Individual animal data are shown in Supplementary Data 6. Mixed effects models with subject as a random variable were used to assess mean differences among timepoints and vaccine groups. Timepoints were compared to ‘pre-BCG’ control using Dunnett’s multiple comparison tests. Significant P values (P < 0.05) across time are shown above each graph. Significant differences determined by Mann–Whitney tests between vaccination groups at given timepoints are indicated by brackets and significant P values (not adjusted for multiple comparison) indicated: *P < 0.05 and ****P < 0.0001. Fixed effect test results, Dunnett’s multiple comparisons and Mann–Whitney test results are reported in Supplementary Table 2b. All statistical tests were two-sided. Source data

Fig. 4. T cell response in PBMC after BCG vaccination and Mtb challenge.

a,b, Frequency of cytokine-responsive CD4⁺ (a) and CD8⁺ (b) T cells in PBMC after 14 h stimulation with H37Rv whole cell lysate relative to BCG vaccination. Mixed effects models with subject as a random variable were used to assess mean differences among timepoints and vaccine groups. Timepoints were compared with ‘pre-BCG’ control using Dunnett’s multiple comparison tests. Significant P values (P < 0.05) across time are shown above each graph. Fixed effect test results and Dunnett’s multiple comparisons are reported in Supplementary Table 2d. c,d, Frequency of cytokine-responsive CD4⁺ (c) and CD8⁺ (d) T cells in PBMC after 14 h stimulation with ESAT-6/CFP-10. Data shown in weeks relative to Mtb challenge. Linear mixed effects models were used to determine significant mean differences between time and animal group. Šidák’s multiple comparisons test P values are reported in Supplementary Table 2d. Significant P values (P < 0.05) across time for each treatment group are shown above each graph. In a–d, data are mean and s.d. of Unvax (dark blue), IV BCG (light blue), SIV/Unvax (red) and SIV/IV BCG (gold) animals. Unvaccinated animals (Unvax): pre-Mtb (n = 8) and 12 weeks post-Mtb (n = 8). SIV-naive vaccinated animals (IV BCG): pre-BCG (n = 7), 4 weeks post-BCG (n = 7), 12 weeks post-BCG/pre-Mtb (n = 7) and 12 weeks post-Mtb (n = 7). SIV+ unvaccinated (SIV/Unvax): pre-Mtb (n = 4) and 12 weeks post-Mtb (n = 4). SIV+ vaccinated animals (SIV/IV BCG): pre-BCG (n = 12), 4 weeks post-BCG (n = 12), 12 weeks post-BCG/pre-Mtb (n = 11) and 12 weeks post-Mtb (n = 12). Individual animal data are shown in Supplementary Data 7. e, IFNγ production in PBMC before and after Mtb challenge (12 weeks post-Mtb) by ELISpot. SFU, spot-forming units. Individual symbols and lines indicate individual animals. Unvax, pre-Mtb (n = 8) and post-Mtb (n = 8); IV BCG, pre-BCG (n = 7) and post-Mtb (n = 7); SIV/Unvax with historical controls, pre-Mtb (n = 8) and post-Mtb (n = 8); SIV/IV BCG, pre-BCG (n = 10) and post-Mtb (n = 9). Paired t-tests were used to determine significance within each group. P values are shown. All statistical tests were two-sided. Source data

Fig. 5. Mycobacterial-specific antibodies in plasma and BALF following BCG vaccination.

a–f, Antibodies specific for H37Rv whole cell lysate were assessed by ELISA for both vaccinated groups: plasma IgG (a), BALF IgG (b), plasma IgA (c), BALF IgA (d), plasma IgM (e) and BALF IgM (f). AUC was determined by calculating the area under the dilution series curve using GraphPad Prism. Individual animals indicated by symbols for IV BCG (light-blue square; n = 7) and SIV/IV BCG (gold, down-pointing triangle; n = 12). Wilcoxon paired signed rank tests were performed to determine significance. P values are shown. Source data

Fig. 6. Protection against Mtb challenge in BCG-vaccinated groups.

a, Total FDG activity (lung inflammation) relative to Mtb challenge, measured by PET/CT imaging. Lines indicate individual animals. b, Number of lung granulomas relative to Mtb challenge. Animals with granuloma numbers >100 are indicated as too numerous to count (tntc). At 4 and 8 weeks, granulomas were counted by CT, whereas at 12 weeks, granulomas were counted by gross pathology. c, Three-dimensional renderings of PET/CT images of individual animals taken at necropsy. d–f, Lung inflammation (d), overall TB pathology (e) and total Mtb burden (thoracic CFU) (f) at necropsy. Each point indicates an individual animal and horizontal bars indicate group medians (Unvax, n = 8; IV BCG, n = 7; SIV/Unvax, n = 11; SIV/IV BCG, n = 12). Kruskal–Wallis tests were performed with Dunn’s multiple comparisons between SIV-naive, vaccinated and unvaccinated groups (dark-blue circles and light-blue squares, respectively), and SIV+, vaccinated and unvaccinated groups (red, up-pointing triangle and gold, down-pointing triangle, respectively). P values are shown. All statistical tests were two-sided. Source data

Extended Data Fig. 1. Study outline and clinical parameters.

a) Timeline of NHP groups including SIV infection, BCG vaccination, HRE therapy, and Mtb challenge. Sampling schedule of blood draws, BAL, and PET-CT image is also included. Created with BioRender.com. b) BCG CFU/mL of blood cultures collected 2 wks after vaccination. Horizontal bars indicate group medians and symbols indicate individual animals (Unvax, n = 4; IV BCG, n = 7; SIV/Unvax, n = 4; SIV/IV BCG, n = 12). c) ESR (mm/hr) of individual animals in each group relative to Mtb challenge. d) Individual animal weights (kg) of each group relative to Mtb challenge. C & D) Symbols indicate individual animals (Unvax, n = 8; IV BCG, n = 7; SIV/Unvax, n = 4; SIV/IV BCG, n = 12). Source data

Extended Data Fig. 2. Clinical parameters of 192-18 and 82-18.

a) Weight of 192-18 over time relative to Mtb challenge. b) ESR of 192-18 over time. c) Weight of 82-18 over time relative to Mtb challenge. d) ESR of 82-18 over time. e) Plasma viral load over course of study. f) Histopathological signs of lymphoma in 82-18. Photomicrographs were representative of 30 slides examined per animal. Left panel: image of ileocecocolic mass (black arrow). Middle panel (i & ii): 10X (left; scale bar = 100 μm) and 40X (right; scale bar = 10 μm) H&E-stained section of ileocecocolic mass indicating large non-cleaved lymphoid population. Right panel (iii): H&E-stained section of liver (20X), peri-portal lymphoid infiltration (scale bar = 50 μm). Source data

Extended Data Fig. 3. CD4 + T cell phenotype in PBMC after vaccination.

a) Frequency of CD4 + CCR5 + T cells in PBMC relative to BCG vaccination. b) Frequency of CD4 + Th1/Th17 (Th*) T cells (CXCR3⁺CCR6⁺) in PBMC relative to BCG vaccination. c) Frequency of CD4+ T_fh T cells (CXCR5⁺ICOS⁺PD-1⁺) in PBMC relative to BCG vaccination. d) Frequency of CD4+ T_reg T cells (CD25⁺CD127^lo) in PBMC relative to BCG vaccination. a-d) Lines indicate mean and error bars indicate SD of SIV-naïve (light blue) and SIV+ (gold) vaccinated animals. SIV-naïve vaccinated animals: pre-BCG (n = 7), 4 wks post BCG (n = 7), and 12 wks post BCG (n = 7). SIV+ vaccinated animals: pre-BCG (n = 12), 4 wks post BCG (n = 12), and 12 wks post BCG (n = 11). Individual animal data are shown in Supplementary Data 8. Linear mixed effects models with subject as a random variable were used. Fixed effect tests were used to assess mean differences among time points and vaccine groups. Time points were compared to ‘pre-BCG’ control using Dunnett’s multiple comparison tests. Significant p-values (p < 0.05) across time are shown above each graph. Significant differences determined by Mann-Whitney tests between vaccination groups at given time points are indicated by brackets and significant p-values indicated: **p < 0.01 and ***p < 0.001. Fixed effect test results, Dunnett’s multiple comparisons, and Mann-Whitney test results are reported in Supplementary Table 2C. All statistical tests were two-sided. Source data

Extended Data Fig. 4. Bacterial burden of SIV-naïve and SIV+ unvaccinated animals.

Total thoracic Mtb burden at necropsy. Each point indicates an individual animal, horizontal bars indicate group means, and error bars indicate standard deviation. Unpaired t test (two-sided) was performed between SIV-naïve unvaccinated (Unvax, dark blue circles; n = 8) and SIV+ unvaccinated groups (SIV/Unvax, red, up-pointing triangles; n = 11). P-value is shown. Source data

Extended Data Fig. 5. TB outcome by tissue compartment.

a-e) Lung pathology (a), thoracic lymph node (LN) pathology (b), extrapulmonary pathology (c), lung Mtb burden (d), and thoracic LN Mtb burden (e) at necropsy. Each point indicates an individual animal and horizontal bars indicate group medians (Unvax, n = 8; IV BCG, n = 7; SIV/Unvax, n = 11; SIV/IV BCG, n = 12). Kruskal Wallis tests were performed with Dunn’s multiple comparisons between SIV-naïve, vaccinated and unvaccinated groups (dark blue circles and light blue squares, respectively); and SIV+, vaccinated and unvaccinated groups (red, up-pointing triangle and gold, down-pointing triangle, respectively). P-values are shown. All statistical tests were two-sided. Source data

Extended Data Fig. 6. Plasma viral load, PBMC CD4+ T_fh T cells, and BAL CD4 + T cell levels stratified by protection across SIV+ vaccinated animals.

Individual symbols and lines indicate protected animals (thoracic CFU < 100, n = 9) and unprotected animals (thoracic CFU > 100, n = 3). a) Plasma viremia in protected and unprotected SIV+, vaccinated animals over the course of the study. b) Frequency of CD4+ T_fh T cells (CXCR5⁺ICOS⁺PD-1⁺) in PBMC from protected and unprotected SIV+, vaccinated animals relative to BCG vaccination. c) Number of CD4 + T cells per BAL collection in protected and unprotected SIV+, vaccinated animals relative to BCG vaccination. Source data