It is not the aim of this article to extensively review the natural history of TB. However, it is very important to provide an overview of this disease in order to understand the battlefield on which infection takes place. In the case of M. tuberculosis (Mtb), this is especially important, as it targets one of the weakest links in the human immune system, namely the alveolar macrophage (AM).

Contrary to the normal, somewhat anthropocentric-type view, the AM is not some sort of police officer constantly on the look-out for pathogens and ready to deliver its antigens to trigger a fast and strong immune response. Alveoli are designed for gas exchange and have a very delicate structure, which is why we should perhaps consider the AM to be a highly specialized cleaner that must keep the space assigned to it clean (Figure 1). Following this anthropocentric approach, we can imagine this cleaner in a 25 m² room, the door of which opens every 6 s to allow the entry of external air along with large quantities of dust and pathogens.

Figure 1 illustrates this by showing the lung as a large city, such as Barcelona (Catalonia), filled with 250 million 25 m² rooms. This illustrates roughly the low range of the alveoli number in adults (Ochs et al., 2004) and, consequently, the same number of AMs, the most abundant antigen-presenting cell in the alveolar spaces (Guth et al., 2009), as it is logical to infer that each alveolus is occupied by a single AM (Suarez et al., 2012; Peake and Pinkerton, 2015). In this scenario, of healthy lung parenchyma, the presence, and role of dendritic cells (DCs) in the alveoli is very limited, if any, as they are present in very low numbers in the parenchyma as a whole (around 2%) and thus essentially absent in the healthy alveolar space. Moreover, even if present, AMs tend to suppress their activity (Landsman and Jung, 2007; Linton and Thoman, 2014). Consequently, there is a very particular scenario in which the conductor (the AM) is very reluctant to trigger any inflammatory response. This results in a two important delays in immunological surveillance, namely in communication level with the lymph node for the antigenic presentation; and in the attraction of specific lymphocytes that are able to activate the infected AMs.

The alveoli have another important characteristics, namely that they are mobile as they must be inflated and deflated in order to allow the entry of air (Macklem, 2005). This is possible because of the presence of a surfactant, produced by type II pneumocytes, which provides the surface tension required to avoid the collapse of the alveolus, which has a diameter of around 300 μm. At the same time alveoli must be sealed to ensure a lack of contact with the plasma transuded from the capillary network as this would constantly vary the surface tension (Suarez et al., 2012). This means that no antibodies can reach the alveoli, even when there is a minimal space between the capillary and the alveoli in order to allow gas transfer (Figure 2). Fortunately, the surfactant also has a micobactericidal activity (Arcos et al., 2011) and can contribute to the defense, along with some activity of AMs themselves, as a result of the phagocytosis of previously cleaned particles (that can also be detrimental) or a protective genetic background (Abel et al., 2014).

Overall, this means that in the context of a single AM infection, specific lymphocytes are not attracted, and that specific antibodies cannot reach the alveoli, unless a strong inflammatory response is triggered to disturb this equilibrium. This is the perfect scenario for Mtb infection and therefore makes the design of a prophylactic vaccine rather difficult, as has been noted recently (Lalvani et al., 2013; Kaufmann et al., 2015).

Figure 3 shows the interaction between Mtb and the immune response in the scenario described above. This interaction defines a cycle in which a fine balance is established between the host and Mtb. The bacilli initially grow in the AM. This growth is very slow (every 24 h) and may therefore delay the induction of danger signals from the AM (Kimura et al., 2014). Depending on the strain, Mtb has a greater or lesser ability to avoid apoptosis of the infected AM, a process that could arrest the inflammatory response and limit bacillary growth (Lee et al., 2006). Indeed, a necrotic AM favors supplemental growth of the bacilli, thus increasing the bacillary load for the new incoming or for the neighboring AMs. An efficacious inflammatory response is triggered once sufficient AMs have been infected, thus allowing the entry of PMNs and monocytes from the capillary that start to build a granulomatous lesion. As a result of this bypass, the seal of the alveoli is broken and the bacilli can reach the interstitial space and be drained by the lymphatic vessels toward the lymph nodes (Chackerian et al., 2002). Once it reaches the lymph nodes, Mtb infects the local macrophages, thereby initiating lymph node infection; and the dendritic cells (DCs), which start the process of epitope presentation and specific Th1-lymphocyte proliferation (Vilaplana et al., 2014).

Specific lymphocytes are triggered to a broad spectrum of epitopes, mainly those secreted by replicating Mtb (R-Mtb; Andersen, 1997; Figure 4), and are attracted to the granuloma to activate the infected AMs. Meanwhile, DCs are produced in the granuloma to provide feedback to the lymph node on the need for more lymphocytes at the infected site (Harding et al., 2015).

The AM activation by Th1 lymphocytes causes Mtb destruction, but also generates NR-Mtb, which are mainly immersed in the intragranulomatous necrosis generated by the bacilli and the inflammatory response. These cellular debris together with the NR-Mtb are phagocytosed by old activated AMs that then become foamy macrophages (FMs; Cáceres et al., 2009; Russell et al., 2009), which are finally drained toward the bronchial escalator. FMs can be destroyed in this drainage, and the NR-Mtb have the chance to form part of the aerosols induced in the bronchioles, thereby again infecting the parenchyma. However, the majority are drained toward the upper portions of the tree, the pharynx, and are then swallowed into the gastrointestinal tract (Cardona, 2009). This is a very efficient way to control Mtb infection, with the immune response only being focused against R-Mtb; whereas NR-Mtb, which are more resistant to the stress (Dutta and Karakousis, 2014), are simply drained away. As the alveolar spaces are smaller in the murine lung, FM get trapped and are easy to see at the periphery of the granuloma (Cardona et al., 2003; Cáceres et al., 2009). This dynamic paints a scenario in which both R-Mtb and NR-Mtb can be found in the same lung, as described before (Muñoz-Elías et al., 2005). This is possible because of the presence of Mtb resuscitation-promoting factors that allow NR-Mtb to revert to R-Mtb (Mukamolova et al., 1998).

Coming back to Figure 4, upon reinfection the immune response is increasingly directed toward the antigens secreted by R-Mtb, a process that helps the survival of NR-Mtb, as they remain hidden from immune surveillance (Andersen, 1997).

Interestingly, a majority of Mtb strains are able to trigger a Th1 response and to provide a protective immune response, as good as that induced by BCG (Mollenkopf et al., 2004); although some exceptions have been reported (Manca et al., 2016). It is expected that the reinfection process is a common feature in high incidence countries. This means that there is a constant boost of the immune response against R-Mtb, which explains why the usefulness of prophylactic vaccines, such as the Bacillus Calmette Guerin–BCG-, in these countries is mainly reduced to providing an advantage to the host when faced with its very first Mtb challenge, i.e., in neonates. Specifically, it reduces the onset of the immune response by about 5 days (Jung et al., 2005; Cardona and Vilaplana, 2014), but does not prevent the infection itself.

It is interesting to note that the lesion caused is usually based on the presence of AMs, some rafts of PMNs, some intragranulomatous necrosis, and is surrounded by lymphocytes (the “doughnut” appearance). This is because granulomas, where the majority of AMs are activated, tend to inactivate the entry of new lymphocytes. This phenomenon has been well-described in mice and guinea pigs and appears to arise as a consequence of different mechanisms, including the competition to obtain arginine, direct suppression, or apoptosis, amongst others (Zhang and McMurray, 1998; Ríos-Barrera et al., 2006; Jeevan et al., 2007).

The lung parenchyma in large mammals is structured into 2i lobes that build a network of interlobular septae which connects the whole parenchyma with the visceral pleura and transmits the mechanical force induced by the diaphragm to expand it (Webb, 2004; Figure 2). This network allows the alveolar walls to be as thin as possible to facilitate gas exchange (Peake and Pinkerton, 2015), as they do not have to transmit the strength of the whole lung expansion. This fact is very relevant in big mammals as they have a larger lung parenchyma to expand (Peake and Pinkerton, 2015). These interlobular septae are very sensitive to any minimal mechanical change, such as the minimal lesions induced in LTBI (with a diameter of around 0.5 mm), and are able to encapsulate and isolate them from the healthy parenchyma within about 10 days (Gil et al., 2010). This mechanism is very important as it helps to stop the drainage of NR-Mtb and finally kills them as a result of a poly-stress mechanism based on hypoxia, high pH, starvation, and finally by the hyperosmosis caused by the accumulation of calcium (Gil et al., 2010).

Looking at the consequences of LTBI, it is clear that it causes a negligible health impact. Problems arise, however, when the lesions are not well-controlled and become bigger, thereby progressively hampering patient health as a result of destruction of the parenchyma. The main challenge is to understand the mechanism that allows a tiny lesion of less than 1 mm, in the case of LTBI, to become a lesion with a diameter of around 10 mm (which is the minimum required for detection in a chest X-ray; Andreu et al., 2004).

Considering the dynamics of Mtb, it is hard to believe in the concept that a bacillus from an old lesion resuscitates and is able to trigger a massive infiltration that leads to liquefaction of the tissue and extracellular growth of the bacilli, which is the current dogma (Grosset, 2003). In fact, there is considerable experimental evidence that R-Mtb must constantly attempt to resettle in the parenchyma to induce new lesions otherwise its fate is to be drained as illustrated in Figure 3 (Cardona and Ivanyi, 2011).

Recently, an explanation has been provided by the study of a murine mode of TB in the mouse strain C3HeB/FeJ, which develops human-like lesions (Marzo et al., 2014). This model supports the idea that the tiny lesions of LTBI have only a narrow window of opportunity to generate large lesions, otherwise they are quickly encapsulated and become unable to do so. Essentially, this opportunity comes as a result of two factors (Figure 5). The first of these is infiltration by PMNs, which generate Neutrophilic Extracellular Traps (NETs) to stop Mtb replication (Ramos-Kichik et al., 2009). However, this induces the opposite effect as NETs becomes a platform that allows extracellular growth of Mtb. This is very interesting because extracellular growth was usually linked to the presence of liquefaction (Grosset, 2003) and was not thought to occur before. As such, the presence of NETs helps rapid enlargement of the lesion. In fact, the relation between granulocyte accumulation and TB progression was described before in TB-highly-susceptible (Lyadova et al., 2010) and in several KO immunosuppressed mice (Gil et al., 2006). The crucial role of this extracellular bacillary population at the beginning of the transition from infection to active TB is that opens up the possibility for an antibody-based therapeutic vaccine that would help the clearance of this population and stop progression toward active TB. Indeed, passive antibody therapy has provided excellent results as regards controlling the evolution of lesions in an experimental model with SCID mice, which have an impaired B- and T-cell response and build the response against Mtb with NK cells and PMNs (Guirado et al., 2006).

The second factor is the coalescence of different yet closely related lesions, including new ones, as a consequence of local drainage (Prats et al., 2016). Both these factors mean that 10 mm lesions can be induced in about 10 days (Marzo et al., 2014; Vilaplana and Cardona, 2014).

Curiously, this phenomenon follows the same physical processes as soap bubble formation, thus allowing an “in silico” model, known as the “bubble model” (Prats et al., 2016), to be developed to help to understand the induction of active TB.

In humans, TB lesions tend to accumulate in the upper lobe (Dock, 1954). The most favored hypothesis relates this tropism with the higher oxygen pressure in this region that might favor Mtb growth. This is related to the gravitational force that causes the lung to hang from the upper side of the costal thorax, thereby favoring a higher intraalveolar pressure that reduces capillary circulation and gas exchange and resulting in a higher partial higher oxygen pressure.

The mechanical force required to inflate the lungs is generated by the diaphragm at the base of the lung, which moves the parenchyma toward the abdominal cavity (Figure 6), thus resulting in a low breathing amplitude in the upper lobes that also favors hyperoxia (Gurney, 1991; Guo et al., 2011). Furthermore, it also causes a lack of local particle drainage in the upper lobes and, therefore, an increase in the bacillary load faced by the AM (Figure 6). This favors a response similar to that triggered against extracellular bacilli, firstly stimulating PMN infiltration (Gil et al., 2006; Gan et al., 2008) and then an immune response dominated by the Th17 response, which also favors infiltration with PMNs (Korn et al., 2009; Torrado and Cooper, 2010). An “in silico” model has recently been built to illustrate this process (Cardona and Prats, 2016).

Two main lesions can be detected in pulmonary TB in humans: small, proliferative lesions that harbor a low bacillary load and mainly comprise epithelioid cells and are well-structured and fibrous, and exudative lesions, in which PMN infiltration is the major component and pus (liquefaction) may also be present (Figure 7; Prats et al., 2016). The evolution of TB results in a kind of patchwork scenario, comprising proliferative and exudative lesions, that is dependent on local endobronchial dissemination (Medlar, 1955). Other factors, such as a certain genetic tendency of AMs to trigger a stronger inflammatory response against Mtb or to induce more or less apoptosis, may also play a role, although which particular factor might have an influence has not been elucidated to date (Abel et al., 2014).

There is also extrapulmonary TB, which can be linked to immunosuppression in the case of meningeal or miliary forms, even though Mtb is able to infect any tissue. Extrapulmonary TB requires the entry of Mtb in the bloodstream (Krishnan et al., 2010). Although, some authors have claimed this to be due to the ability of Mtb to enter into epithelial cells and blood vessels, classically this has been linked to local disruption of the aberrant capillary net built around Mtb granulomas (Hunder, 1996); with larger lesions leading to a higher risk of dissemination (Medlar, 1955).

In summary, infection with Mtb clearly follows the “damage theory” paradigm (Casadevall and Pirofski, 1999; Cardona, 2010) in which both pathogen type and type of host' response triggered against the pathogen are key to induction of the disease.

Other promising candidates are still in the preclinical phase. Several DNA vaccines codifying mainly for the hsp65, Ag85A, and B antigens have been tested. Some of these have demonstrated their value in murine models, with or without chemotherapy and with a wide range of protection, as reviewed by Lowrie (2006). Recently, a group of therapeutic vaccines comprising different combinations of fusion proteins has provided very promising results. Thus, the “3-iron”vaccine candidate, which includes several iron-acquisition proteins (Shanley et al., 2014), has demonstrated a reduction in pathology and increased survival in the guinea pig model without the help of chemotherapy, although it was unable to reduce the bacillary load in the lung. Interestingly, this candidate showed this protection against highly virulent MDR strains. Similarly, ID93/GLA-SE, a candidate designed by the Reed group (Coler et al., 2013), is a combination of M. tuberculosis proteins associated with virulence (Rv2608, Rv3619, and Rv3620) or latency (Rv1813). ID93/GLA-SE has demonstrated its therapeutic value in combination with rifampicin and isoniazid in SWR/J mice and cynomolgus monkeys. This candidate elicits a robust Th1 response and is able to reduce the pulmonary bacillary load and increase the survival of the animals.

PC was one of the founders of Archivel Farma, the company that developed RUTI, and was its CSO until 2013. He was also the inventor of RUTI. PC is one of the founders and the CEO of Manremyc, the company that is developing the use of M. manresensis. He was also the inventor of this food supplement.