Immune correlates of early clearance of Mycobacterium tuberculosis among tuberculosis household contacts in Indonesia

Abstract

Some individuals, even when heavily exposed to an infectious tuberculosis patient, do not develop a specific T-cell response as measured by interferon-gamma release assay (IGRA). This could be explained by an IFN-γ-independent adaptive immune response, or an effective innate host response clearing Mycobacterium tuberculosis (Mtb) without adaptive immunity. In heavily exposed Indonesian tuberculosis household contacts (n = 1347), a persistently IGRA negative status was associated with presence of a BCG scar, and - especially among those with a BCG scar - with altered innate immune cells dynamics, higher heterologous (Escherichia coli-induced) proinflammatory cytokine production, and higher inflammatory proteins in the IGRA mitogen tube. Neither circulating concentrations of Mtb-specific antibodies nor functional antibody activity associated with IGRA status at baseline or follow-up. In a cohort of adults in a low tuberculosis incidence setting, BCG vaccination induced heterologous innate cytokine production, but only marginally affected Mtb-specific antibody profiles. Our findings suggest that a more efficient host innate immune response, rather than a humoral response, mediates early clearance of Mtb. The protective effect of BCG vaccination against Mtb infection may be linked to innate immune priming, also termed 'trained immunity'.

Fig. 1. The dynamics of innate immune cells in IGRA-converters and persistently IGRA-negative individuals.

A Overview of the flow cytometry dataset. Created in BioRender. Setiabudiawan, T. (2025) https://BioRender.com/w61f180. B Frequencies of circulating innate immune cells (numbers / µL blood) were compared between week 2 and week 14 in IGRA converters (N = 22), and persistently IGRA-negative individuals (N = 48, [FDR <0.1, <0.05, <0.01; *, **, ***; FC = median fold change]). C Persistently IGRA-negatives with BCG scar (N = 38) showed a larger decrease in cell numbers than participants without BCG scar (N = 10) in the innate circulating immune cells from week 2 to week 14.

Fig. 2. Study outline and ex vivo cytokine production.

A Baseline whole blood ex vivo cytokine production, compared between baseline IGRA-negative (N = 328) and IGRA-positive individuals (N = 145), and between IGRA converters (N = 91) and persistently IGRA-negative individuals (N = 237). Created in BioRender. Setiabudiawan, T. (2025) https://BioRender.com/k78q817. B Cytokine production following stimulation with Mtb, BCG, and E.coli, with higher Mtb-induced cytokine production in baseline IGRA-positive individuals, and higher E. coli-induced production in persistently IGRA-negative individuals. C Mtb, BCG, and E.coli-induced IL-6 production (as a representative), stratified for IGRA-status (Mann-Whitney U test after correction for multiple testing). Association between cytokine production and IGRA status at baseline (D) and 14 weeks (E), expressed as odds ratio (OR) using logistic regression adjusting for age, sex, BMI, exposure risk score, blood monocyte count, blood lymphocyte count, and batch. F Relation between baseline ex vivo cytokine E.coli-induced production (in IGRA-negative individuals) and IGRA status at 14 weeks, shown as odds ratios, stratified for BCG vaccination status. All models were corrected for multiple testing (Benjamini-Hochberg). (FDR <0.1, <0.05, <0.01, <0.001; closed circle & *, **, ***, ****).

Fig. 3. Inflammatory proteins in IGRA supernatants.

Inflammatory proteins relative concentrations (NPX unit, log₂ transformed) in IGRA supernatants (nil and mitogen) at baseline were compared between IGRA converters (N = 48) and persistently IGRA-negative individuals (N = 128). A A persistently IGRA-negative status was associated with higher CXCL6 in the baseline IGRA nil tube, and (B) with ADA, MCP-3 (CCL7), TWEAK, IL-17C, and IL-18 in the baseline mitogen tube (logistic regression adjusted for age, sex, BMI, and exposure risk score; FDR <0.1). C Associations between persistently IGRA-negative status and concentrations of all proteins measured in IGRA nil and mitogen tube (Odds ratios [OR], adjusted for age, sex, BMI, and exposure risk score). D In the mitogen tubes, the same proteins, as well as CSF-1, DNER, CD244, and VEGFA, showed a correlation with quantitative TBAg - nil IFNγ IGRA results after correction for multiple testing with a lower FDR cutoff of 0.05.

Fig. 4. Antibody profiles according to IGRA-status at baseline and follow-up.

Antibody profiles were compared between IGRA-positive (N = 100) and IGRA-negative (N = 433) tuberculosis household contacts; and between IGRA converters (N = 51) and persistently IGRA-negative individuals (N = 237), based on strict IGRA criteria (<0.15 IU/mL for negative and > 0.70 IU/mL for positive). Fold differences in antibody levels (shown as the ratio of antibodies corrected for the positive control hemagglutinin [HA]), are shown according to IGRA status at baseline (A red: higher antibody levels in IGRA-positive individuals) and follow-up (D purple: higher antibody levels in IGRA converters). No difference reached statistical significance, thus, numbers not shown in the heatmap (Mann-Whitney U test; FDR <0.1). B Partial least squares discriminant analysis (PLS-DA) using the selected 25 antibodies was used to visualize differences in antibody levels between baseline IGRA-positive (red) and -negative (blue) individuals, and (E) between IGRA converters (purple) and persistently IGRA-negative individuals (blue). C In logistic regression, no antibody was associated with IGRA status at baseline or follow-up (F) after adjustment for sex, age, BMI, and exposure risk score.

Fig. 5. Effect of BCG vaccination on cytokine production and Mtb-specific antibodies.

Heatmap showing fold change (A) and paired boxplot (B) of ex vivo cytokine production in Dutch healthy adults (N = 298) before and 90 days after BCG vaccination. Shown are 24 h stimulation of PBMCs with Mtb and S. aureus, and 7-day stimulation of Mtb for IFN-γ. C shows a fold change of hemagglutinin (HA)-standardized anti-Mtb antibody levels at day 90 compared to the pre-vaccination, with statistically significant fold changes shown in numbers. D Shows changes in IgG3 and IgM antibodies against different Mtb antigens, and raw anti-HA IgM and IgG3 MFI pre-vaccination and 90 days post-vaccination were used for normalization. FDR <0.1, <0.05, <0.01, <0.001; *, **, ***, ****.