Bacille Calmette-Guerin induces NOD2-dependent nonspecific protection from reinfection via epigenetic reprogramming of monocytes

Abstract

Adaptive features of innate immunity, recently described as "trained immunity," have been documented in plants, invertebrate animals, and mice, but not yet in humans. Here we show that bacille Calmette-Guérin (BCG) vaccination in healthy volunteers led not only to a four- to sevenfold increase in the production of IFN-γ, but also to a twofold enhanced release of monocyte-derived cytokines, such as TNF and IL-1β, in response to unrelated bacterial and fungal pathogens. The enhanced function of circulating monocytes persisted for at least 3 mo after vaccination and was accompanied by increased expression of activation markers such as CD11b and Toll-like receptor 4. These training effects were induced through the NOD2 receptor and mediated by increased histone 3 lysine 4 trimethylation. In experimental studies, BCG vaccination induced T- and B-lymphocyte-independent protection of severe combined immunodeficiency SCID mice from disseminated candidiasis (100% survival in BCG-vaccinated mice vs. 30% in control mice). In conclusion, BCG induces trained immunity and nonspecific protection from infections through epigenetic reprogramming of innate immune cells.

Fig. 1.

BCG vaccination increased the nonspecific production of proinflammatory cytokines. (A) Diagram showing the course of the BCG vaccination trial. Blood was collected from 20 naïve (nonexposed) volunteers before and after (2 wk and 3 mo) vaccination with BCG. (B–D) PBMCs isolated from the 20 volunteers before and after (2 wk and 3 mo) vaccination were stimulated in vitro with sonicated M. tuberculosis, heat-killed S. aureus, and C. albicans blastoconidia. Proinflammatory cytokine production (INF-γ [B], TNF-α [C]), and IL-1β [D]) was assessed by ELISA in the supernatants. *P < 0.05, **P < 0.01.

Fig. 2.

BCG alters the phenotype of circulating monocytes in healthy volunteers. (A) Flow cytometry analysis of CD45⁺ cells. (B) Histogram showing the expression level of CD14⁺ (% of total cells) in the cells isolated from 20 volunteers before and after BCG vaccination. (C) Flow cytometry analysis of TLR4 and CD14 in cells isolated from one volunteer before and 3 mo after BCG vaccination. (D and F) Overlays of the surface expression level of TLR4 (D) and CD11b (F) within CD14⁺ cells isolated from one volunteer before and 3 mo after BCG vaccination. (E and G) Average surface expression level of TLR4 (E) and CD11b (G) within the CD14⁺ monocyte population isolated from 20 volunteers before and after BCG vaccination. *P < 0.05, **P < 0.01, ***P < 0.005. (H) il-1β and tnfα mRNA expression after sonicated M. tuberculosis, heat-killed S. aureus and C. albicans hyphae in vitro stimulation of the PBMCs isolated from volunteers before and 3 mo after BCG vaccination (n ≥ 3). (I) ChIP analysis of the enrichment of H3K4me3 at the promoter of tnfα, il6 and tlr4 in human monocytes isolated from three subjects before and 3 mo after BCG vaccination. *P < 0.05, paired t test.

Fig. 3.

BCG vaccination protects mice against lethal C. albicans infection through a T-/B-lymphocyte–independent mechanism. (A) Survival rate of SCID mice infected with live C. albicans (2 × 10⁶ cfu/mouse) injected i.v. The mice were vaccinated i.v. either with PBS (control) or BCG 14 d before inoculation with a lethal C. albicans dose (n ≥ 15 per group, two independent experiments). (B) Fungal burden of kidneys from control- and BCG-vaccinated SCID mice 3 and 14 d after the lethal C. albicans infection (n = 5). (C) TNF-α production of spleen monocytes after restimulation in vitro with LPS from control- and BCG-vaccinated SCID mice 7 d after the lethal C. albicans infection (n = 5). *P < 0.05, **P < 0.01, and ***P < 0.005 vs. control (PBS) animals.

Fig. 4.

BCG primes the production of proinflammatory cytokines. (A) Diagram showing the course of the in vitro preincubation experiment. Cells were preexposed to culture medium or BCG vaccine for 24 h (first stimulation—training). After the first stimulus was washed, the cells were incubated for 7 d in culture medium supplemented with serum. Afterward, a second in vitro stimulation (second stimulation) of cytokine production with various pattern recognition receptor ligands was performed for an additional 24 h. (B) BCG training in vitro using freshly isolated adherent monocytes. (C and D) Inhibition of TLR4 or TLR2 does not affect the training effects induced by BCG. (E and F) BCG training of monocytes is severely affected in cells obtained from NOD2-deficient volunteers (F) but not from dectin-1–deficient volunteers (E). (G) The training effects induced by BCG could be reproduced with MDP but not with Pam3Cys or LPS. (H and I) TNF-α production in BCG-primed monocytes in the absence or presence of Rip2/p38 inhibitor (I), the histone demethylase inhibitor pargyline, or the histone methyltransferase inhibitor MTA (H). (C–I) The ratios of cytokine production in BCG-primed vs. nonprimed monocytes are presented. Data presented in (E) are the mean of two independent experiments (n = 1 + 1). Data presented in (F) were obtained from 12 healthy volunteers (controls) and two different NOD2-deficient individuals (NOD2^−/−). (B–D and G–I) *P < 0.05, **P < 0.01, ***P < 0.005. Data are presented as mean ± SD (n ≥ 6). The Wilcoxon signed rank test was used to detect significant differences.