Background
Despite the continuous evolution of better management, new drug development, and additional diagnostic tools available in public health, tuberculosis (TB) still remains as a serious global problem. There were a total of 8.6 million new cases and 1.3 million deaths in 2012; TB is a leading killer of people living with HIV [
1]. It is currently estimated that one third of the world population is infected with
Mycobacterium tuberculosis and that 5–10 % of those infected will develop the disease during their life-time [
2].
The current TB vaccine is
M. bovis bacille Calmette-Guérin (BCG), and it has been employed for nearly a century to prevent the disease. However, while BCG is the most common and broadly used vaccine worldwide, its efficacy remains quite controversial: In endemic sites, the vaccine has evoked unpredictable outcomes [
3]. Epidemiological studies point to BCG as affording more complete protection against the meningeal and miliary clinical forms of TB in children than against the pulmonary clinical form in adults (Reviewed by [
4]). It is generally perceived that the BCG vaccine is not fully effective because of the absence of antigens commonly shared with
M. tuberculosis, but also that the rapid removal of BGC via systemic immunity, primarily targeting atypical environmental mycobacteria, reduces the vaccine’s effectiveness over time. Therefore, more studies are required to better understand how BCG confers protection in humans.
Conventional studies have focused solely on the role of cytokines released by monocytes infected with BCG as a starting point for inducing immunity, such as the pro-inflammatory interleukin 1 beta (IL-1β) and tumor necrosis factor alpha (TNF-α) (Reviewed by [
5,
6]). Thus, one of the issues raised concerns the additional soluble compounds that are potentially secreted by host cells during BCG infection. Also, there may be a later increase in those factors for the initiation of apoptosis, necrosis, necroptosis and pyroptosis. The action of inflammatory caspases, such as caspase-1, has shown promising results for the activation of IL-1β through an inflammasome mechanism during microbial pathogens interactions [
7].
Our prior study supports the hypothesis that BCG induces distinct cell-death patterns during the maturation of the immune system [
8]. In the study, we observed increasing apoptosis in BCG-stimulated monocytes from healthy, vaccinated adults, associated with a release of IL-1β and TNF-α, but not with metalloproteinase-9. Conversely, higher monocyte necrosis, but not apoptosis, was observed following the infection of umbilical vein cells from naïve neonates. This pattern was paralleled by different pro-inflammatory cytokine levels when compared to adults. In addition, necrosis has been defined as a mechanism used by virulent bacteria to exit macrophages, evade the host defenses, and disseminate; apoptosis has been associated with diminished pathogen viability [
9,
10].
To address remaining critical issues beyond these premises, we have constructed a model of conditioned cell culture medium, i.e., culture supernatants obtained from host cells infected with BCG for 48 h, for the study of the factors secreted by those cells that potentially are responsible for monocyte cell-death during the context of infection. Our hypothesis has a basis in compelling evidence from one recent study where eicosanoid lipid intermediaries regulated the cell death program of human macrophages infected with
M. tuberculosis [
10]. Thus, a cross-sectional population study of samples from Brazil was the means to uncover critical aspects of the in vitro degree of apoptosis and necrosis induced by BCG Moreau in monocytes from individuals, whether sensitized (adults) or non-sensitized (neonates). Using this method, there is input for a better understanding of the protective factors afforded by BCG against TB that would help to identify the processes by which this protection is achieved, thus opening up a horizon for its future improved clinical applicability.
Discussion
Prophylactic medicine via vaccination is still the most cost effective means to reduce the spread of infectious diseases, and it also minimizes the burden on the health system. Because TB has imposed centuries-long threat to humans, the majority of the population in the tropics has been immunized with BCG in the infancy, but real protection against disease during adulthood has not yet been achieved. There is a long road ahead to better understand the entire clinical utility of BCG Moreau vaccine, and currently this task is our primary goal (Reviewed by [
4]).
Here, consistent data have shown an enhancement of the in vitro apoptosis rate in monocytes from both adult and neonate individuals only after repeated long-term incubation with CM from autologous, infected BCG Moreau cultures that have been regularly used in a previous model. Hence, after a total of 48 h of infection, BCG stimulated a sufficient output of soluble factors in the CM, in both PBMC and CBMC cultures, to drive additional 120 h of autologous monocyte cell-death. This result, together with higher in vitro necrosis rate in the UV group, is generally in line with our previous data suggesting the distinct cell-death patterns induced by BCG [
8]. That was associated both with higher NO released from the BCG-primed group, but not from the naïve individuals, and also by a distinct pattern of additional mediators and cytokines that found a close parallel to its respective cell-death outcome.
Reactive nitrogen intermediates, such as NO, have been reported to possess anti-mycobacterial activity, and it has long been established [
13], as well as currently understood, that NO is crucial in the host response against intracellular pathogens [
14]. Macrophages can directly inhibit the proliferation of mycobacteria by producing NO. In addition to the microbicidal properties, NO is also important as signaling molecules and therefore promotes macrophage activation and prominent apoptosis. This facilitates antigen presentation and helps to connect the innate to the adaptive immune responses [
13,
14]. Hypothetically, the more apoptosis is induced by a vaccine against TB in host cells, the more effective is the response against
M. tuberculosis. However, there is insufficient evidence to determine whether NO is involved in the defense against the TB bacillus in humans. Here, the higher NO levels released after BCG infection of mononuclears in the HD group, but not in the UV group, might denote indirectly a memory phenotype that remains active for years after priming, thus creating effective, long-lasting protection in those individuals. The balance of cytokines that efficiently upregulate NO production, such as IFN-γ, TNF-α and IL-1β, may play a critical role in the defense mechanism against TB [
19].
As above stated, IFN-γ synergizes with TNF-α for NO production. Thus, our two previous in vitro studies corroborate with the present one through lower NO levels detected in neonate individuals: Virtually no IFN-γ production [
20] or TNF-α levels [
8] were found in those assessments. Therefore, studying human immune response using the BCG Moreau vaccine and also including the NO production may give us a better understanding of the TB pathogenesis as well.
Resistance or susceptibility to a given infection is a complex phenotype with a multifactorial trait, with several critical candidates playing an important role. As an example, arachidonic acid metabolites converted to the eicosanoids PGE
2 and LTB
4 are inflammatory mediators which are likely to be involved in apoptosis and necrosis episodes, respectively. Originally, PGE
2 was described as the major prostanoid in the lung with an important antifibrotic role. On one hand, PGE
2 mediates vasodilatation, increases vascular permeability, enhances pain perception by bradykinin and histamine, alters connective tissue metabolism, and enhances osteoclastic bone resorption. On the other hand, LTB
4 causes the accumulation of inflammatory cells in the inflamed sites and degranulation of polymorphonuclear leukocytes [
21]. The preferential synthesis of bioactive eicosanoids, mainly prostanoids, is a critical branch point for the innate antimycobacterial response of the infected macrophages. Distinct virulence determinants can manipulate macrophage inflammation toward both extremes, both inducing anti-inflammatory mediator production and also promoting pro-inflammatory TNF [
10]. More importantly, PGE
2 was recently defined as a pro-apoptotic host lipid mediator which protects against necrosis [
9,
10]. In contrast, suppression of PGE
2 synthesis by a pronecrotic mediator results in mitochondrial damage and inhibition of plasma membrane repair mechanisms, ultimately leading to the induction of necrosis. Thus, the balance between PGE
2 and LTB
4, presumably another pronecrotic lipid intermediary, may actually determine whether BCG-infected macrophages undergo apoptosis or necrosis, and this balance determines the outcome of infection. Accordingly, comparable to previous studies, we observed a significant pattern showing that PGE
2 and LTB
4 were highly correlated to the corresponding cell-death patterns; i.e. BCG-infected mononuclears induced PGE
2 and apoptosis in both groups but induced LTB
4 and necrosis in the UV group only. Finally, the absence of IL-1 decreases PGE
2 in a model of
M. tuberculosis-infected deficient mice, and the addition-of IL-1 or PGE
2 reduces the bacteremia, whereas the TNF-induced LTB
4 rises in wild-type animals [
18]. Additionally, blocking of type I IFN signaling resulted in increased IL-1 and PGE
2. Importantly, protection directly correlated with higher PGE
2 and lower type I IFN and IL1Ra production. Likewise, PGE
2 also inhibited type I IFN production in
M. tuberculosis-infected human macrophages. These findings reveal that suppression of type I IFNs and their pro-bacterial activity is a main mechanism of IL-1- and PGE
2-mediated host resistance against
M. tuberculosis.
Similar to LTB
4, IFN-β seems to be implicated earlier in necrosis [
16] and more recently in necroptosis [
17]. In fact, one of the mechanisms underlying enhanced susceptibility to bacterial infections is the stimulation of IFN-β production, as stated above, and the latter study showed a
Salmonella enterica-induced macrophage death via necroptosis in a type I IFN-dependent manner. That novel cell-death modality is usually triggered by CD95 and TNF family of death receptors, and this programmed cell necrosis is distinct from apoptosis [
22]. In another setting, pDCs were shown to be a major cell population producing IFN-β through toll-like receptor (TLR) signaling [
23]. Generally, pDCs produce most IFN-β within 48 h after pathogen infection and then differentiate into mature DCs with enhanced antigen-presenting capability, after which they become refractory upon secondary stimulation and lose their IFN-β production ability. Our current data of highest IFN-β levels after that analogous period of BCG-infected mononuclears in the UV group again shows this cytokine to be well correlated with the necroptosis pattern shown by Robinson and colleagues [
17]. However, in our system the exact underlying mechanism (necrosis vs. necroptosis) has not yet been fully elucidated.
TGFβ1 might act as an anti-inflammatory arm prone to damper Th1 immune response. Actually, in line with the current data, our very recent study also failed to find any Th3 cell phenotype induced by the BCG vaccine in the sensitized adult cohort: Comparable TGFβ1 levels were found regardless of the stimulus employed [
24]. That study originally established that the BCG Moreau vaccine induced a significantly elevated IFN-γ/IL-10 ratio only in the HD group, confirming a polarized Th1 pattern in vitro.
Based on those previous findings, our current primary results suggest that apoptosis was again induced in monocytes, probably via TNF and/or IL-1β in adults, but not in neonates [
8]. Of note, IL-1β drives PGE
2 synthesis in infected human macrophages [
18]. However, we could not rule out a role for other candidates, such as ATP, TWEAK, TRAMP, TRAIL, and CD95, since there is a long list of potential candidates for apoptosis induction in our system, particularly for the new data regarding naïve individuals [
25].
Taking together, one could conclude that the apoptosis pattern found in resting monocytes from both immunized and naïve groups correlates well with high pro-inflammatory cytokine levels induced in BCG-infected mononuclears. However, necrosis data supports previous findings in neonates only, but additional studies are warranted to rule out other mechanisms such as pyroptosis or necroptosis. These findings support the hypothesis that BCG Moreau strain induces a milieu of soluble factors, which further stimulate distinct cell-death patterns regarding the maturation of the immune system. This pattern may set the stage for a subsequent anti-mycobacterial immune response, which could have profound effects during vaccination.
A possible unifying hypothesis on the central role of cell-death modality in TB pathogenesis is that lesser apoptosis levels in BCG-infected monocytes of neonates could predispose those individuals to the development of inflammatory disease due to lack of protective and specific immune response. This is because only the phagocytosis of apoptotic cells leads to the resolution of inflammation [
26]. Conversely, the progress to necrosis could predispose those individuals to the development of disease, since virulent
M. tuberculosis inhibits apoptosis and triggers necrosis of host macrophages, which delays the initiation of adaptive immunity [
9,
10]. Thus, BCG immunization is the only pathway for increasing the immune response against
M. tuberculosis. On the other hand, BCG-vaccinees show a balance between regulatory and effective immune responses, and a breakdown in this equilibrium would be crucial for TB development.
Despite some tantalizing clues and frequent overstatements, the underlying immunological mechanisms by which the BCG Moreau vaccine evokes protective immune response against M. tuberculosis continues to be poorly understood. The central puzzle that remains to be deciphered is: How exactly does the BCG immunization confer partial protection in the vaccinees?
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
JL, MS and TP participated in the design of the study, performed the experiments, and analysed data and statistics. CP, LP and SM carried out several experiments for cell viability and nitric oxide detection, and collected the patient samples. JL and MS provided helpful discussions and edited the manuscript. LRCB provided clinical expertise and helpful discussions. PRZA conceived the study, participated in the design and coordination of the study, analysed data and statistics, wrote the grant application, and wrote the manuscript. All the experiments were performed at the Instituto Oswaldo Cruz (Fiocruz) laboratory. All authors read and approved the final manuscript.