The Role of Mast Cells in Tuberculosis: Orchestrating Innate Immune Crosstalk?

Tuberculosis causes more annual deaths globally than any other infectious disease. However, progress in developing novel vaccines, diagnostics, and therapies has been hampered by an incomplete understanding of the immune response to Mycobacterium tuberculosis (Mtb). While the role of many immune cells has been extensively explored, mast cells (MCs) have been relatively ignored. MCs are tissue resident cells involved in defense against bacterial infections playing an important role mediating immune cell crosstalk. This review discusses specific interactions between MCs and Mtb, their contribution to both immunity and disease pathogenesis, and explores their role in orchestrating other immune cells against infections.

Tuberculosis causes more annual deaths globally than any other infectious disease. However, progress in developing novel vaccines, diagnostics, and therapies has been hampered by an incomplete understanding of the immune response to Mycobacterium tuberculosis (Mtb). While the role of many immune cells has been extensively explored, mast cells (MCs) have been relatively ignored. MCs are tissue resident cells involved in defense against bacterial infections playing an important role mediating immune cell crosstalk. This review discusses specific interactions between MCs and Mtb, their contribution to both immunity and disease pathogenesis, and explores their role in orchestrating other immune cells against infections.
Keywords: Mycobacterium tuberculosis, mast cells, degranulation, extracellular traps, granuloma, infection inTRODUCTiOn Tuberculosis (TB) is the world's major infectious disease killer, accounting for 1.4 million deaths in 2015 (1). Progress in developing vaccines, diagnostics, and therapies has been hampered by an incomplete understanding of the immune response to the causative pathogen, Mycobacterium tuberculosis (Mtb).
Following entry of Mtb-containing droplets into the airways, bacilli are initially phagocytosed by alveolar macrophages (AMφ), providing a comfortable niche in which Mtb can reside, replicate, and evade immune cell detection (2). Mycobacterial pathogen-associated molecular patterns engage pattern recognition receptors (PRRs) to trigger signaling pathways, resulting in the release of various chemokines and cytokines and the recruitment and activation of immune cells (3). This process results in the internalization of mycobacteria by dendritic cells (DCs), which migrate to lymph nodes where they polarize naïve T cells to antigen-specific Th1 effector cells in an IL-12-dependent manner (4,5). IFN-γ produced by Th1-polarized T cells activates mycobactericidal mechanisms in AMφ (6). Various immune cells are sequentially recruited to the sites of infection; neutrophils in the earliest stages as well as T cells, NK cells, and fibroblasts. These surround the infected AMφ to form a mycobacterial granuloma (2, 7), which acts as a "physical barrier" limiting bacillary dissemination. However, chronic granulomas also promote Mtb's intracellular survival and impair elimination, resulting clinically, in latent TB disease (2). One-third of the world's population is latently infected with Mtb and between 1 and 10% will develop progressive TB disease following "reactivation" of Frontiers in Immunology | www.frontiersin.org October 2017 | Volume 8 | Article 1290 infection in later life (8). The mechanism of TB reactivation is unclear as yet, however, it is suggested that a failure of granuloma maintenance may be the cause (9). Thus, the granuloma is the result of a non-efficient immune control that will eventually progress to a chronic infection, rather than mycobacterial clearance. Mast cells (MCs) are tissue resident cells strategically located in mucosal tissues (10) and are among the first cells to come in contact with pathogens (11). MCs contribute to bacterial immunity through multiple mechanisms such as bacterial recognition, activation, recruitment of immune cells to the site of infection, release of inflammatory mediators, and direct bacterial killing by extracellular traps (ETs). However, their main role may be orchestrating other immune cells against infections (11)(12)(13)(14).
Since little is known about the MC contribution to TB pathogenesis, this review summarizes the MC strategies used in bacterial defense as well as potential and reported interactions occurring between Mtb and MCs (Figure 1).

MC OnTOGenY AnD FUnCTiOnS
Mast cells originate from pluripotent CD34 + and CD117 + hematopoietic stem cells and migrate as progenitors to various tissues where they mature influenced by the local microenvironment (15,16). MCs are mainly located in the skin (~12,000/mm 3 ) and mucosa including the lungs (~500-4,000/mm 3 ) (10).

MC invOLveMenT in eARLY MYCOBACTeRiAL inFeCTiOn
Although interactions between MCs and Mtb have been reported (Figure 1), the role of MCs in TB pathogenesis remains unclear. MC involvement in mycobacterial immunity was first observed in guinea pigs using electron microscopy. Shortly, after intratracheal infection with Mtb, a significant increase in MCs was detected in guinea pig lungs (19). A later study demonstrated that the number of MCs in mice lungs increases by ~23% after 15 days of Mtb exposure (20).

Bacillary Recognition
Rat peritoneal MCs (rPMCs) recognize Mtb via CD48 that is a glycosyl phosphoinositol-anchored cell surface protein ( Figure 1A). Incubation with increasing concentrations of anti-CD48 antibodies together with Mtb correlate with a proportional decrease in histamine release (21). Previous work has shown that CD48 recognizes FimH protein expressed by fimbriated bacteria (such as Escherichia coli and Staphylococcus aureus) resulting in MC degranulation (12,22,23), raising the question of how precisely CD48 recognizes Mtb, which is not known to be fimbriated (24).
TLRs are a key receptor family implicated in pathogen recognition (22). Carlos and colleagues found MC TLR2 to be relevant in Mtb recognition ( Figure 1A) (9). The transfer of TLR2 +/+ MCs into TLR −/− Mtb-infected mice showed an increase of cytokine release and cell recruitment, suggesting MC TLR2 as key receptor upon mycobacterial challenge (9). MCs also express TLR4 that serves as a mannose receptor in complex with soluble CD14. In the absence of CD14, high concentrations of TLR4 ligands are required for MC-TLR4 activation (25)(26)(27). Presently, it is unclear if TLR4 (or other TLRs) are involved in Mtb recognition and if the TLR4-CD14 complex is necessary to trigger MC functions. Besides the TLRs and CD48, the CR3, CR4, C3aR, and C5aR complement receptors and the FcεRI, FcγRI, FcγRII, and FcγRIII receptors mediate MC responses to other bacteria, but it is as yet unknown if they are involved in MC-Mtb interactions (12,22,28).

Bacillary Binding and Uptake
Muñoz and colleagues suggested that Mtb is internalized by MCs via lipid rafts ( Figure 1A). Cholesterol depletion of the rat basophilic leukemia cells (RBL-2H3) reduced internalization of mycobacteria. Interestingly, once internalized, mycobacteria survived intracellularly for 4 days after which most of the infected MCs (70%) had undergone lysis. These findings implicate MCs as a reservoir for Mtb (29). However, still unclear are the requirements and dynamics of Mtb internalization by MCs since few studies have demonstrated it experimentally and all were performed in animal models (21,29). It is therefore important that the Mtb internalization is demonstrated in human MCs, together with the underlying mechanisms of internalization (aside from lipid rafts) and the intracellular compartments in which Mtb may reside. Published data describing MC interactions with other bacteria may guide the design of these investigations. For example, Salmonella typhi is taken up by complement receptors while E. coli is attached and internalized by FimH to the MC surface (30). Furthermore, preincubation of human MCs with IFN-γ increases the membrane attachment of S. aureus (31). Recently, the MC phagosome was demonstrated to interact with NOD-like receptors thus modulating cytokine production (32), and indicating that MC phagocytosis acts as an inducer of other MC activities.

Bacterial-induced Mediator Release
Few studies have characterized the Mtb-induced MC production of soluble mediators of inflammation. Muñoz et al. reported that after stimulation with Mtb, rPMCs released de novo synthesized TNF-α and IL-6, followed by secretion of histamine and β-hexosaminidase. In addition, specific Mtb antigens [MPT-63 and early secretory antigenic target 6 (ESAT-6)] have also been shown to induce rPMCs to release TNF-α, IL-6, histamine, and β-hexosaminidase (Figure 1) (21).

Cytokines and Chemokines
Cytokines and chemokines released by MCs contribute to protective immunity in the context of bacterial infections. For example,  it has been shown that MC depletion reduces TNF-α (important in mycobacterial granuloma maintenance) concentrations in bronchoalveolar lavage (BAL) of Bordetella pertussis-infected mice (33). Furthermore, following infection with Streptococcus pneumoniae, increased TNF-α concentrations in the BAL and MC numbers in the lung correlated with protection (34). Similarly, MC-derived IL-6 produced during Klebsiella pneumoniae challenge was observed to improve mouse survival by promoting neutrophil recruitment and intra-neutrophil killing (35), the importance of which in mycobacterial immunity is being increasingly recognized (36). MCs also produce a wide variety of soluble mediators potentially relevant to mycobacterial immunity, including IL-13, IL-12, IL-6, IL-4, TNF-α, CCL5, CXCL2, CCL7, and CCL2, following infection with Streptococcus equi (37). Since the cocktail of soluble mediators produced by MCs appears dependent on the specific pathogen, a comprehensive proteomic assessment of mediators produced by MCs in response to Mtb is needed.

Antimicrobial Peptides (AMPs)
Antimicrobial peptides kill pathogens by forming pores in cytoplasmic membranes; defensins and cathelicidins are the most studied (38). Early in infection, cathelicidins promote phagocytosis, upregulate the expression of costimulatory molecules in DCs and stimulate Th1 cytokine production, while later in the course of the disease they inhibit the production of pro-inflammatory molecules (39). Although little is known regarding the AMP repertoire that MCs may release, MC supernatants reduce bacterial burden (40,41). Cathelicidin LL-37 is expressed in human dermal skin MCs while the respective murine homolog cathelicidinrelated AMP (CRAMP) is produced by bone marrow-derived MCs (BMMCs). Upregulation of CRAMP expression by LPS reduces group A streptococcus extracellular titers (42), while MC-derived LL-37 promotes clearance of Enterococcus faecalis (43). Finally, β-hexosaminidase, which is released by MCs after degranulation, was observed to exhibit antimicrobial activities upon intracellular Listeria monocytogenes infection (39) and mMCP-6, a mouse tryptase, was essential for K. pneumoniae clearance in mMCP-6 −/− mice (44). Thus, antimycobacterial molecules are likely to be secreted by MCs upon Mtb exposure ( Figure 1B).

MC Degranulation and Histamine Release
Mast cell degranulation in Mtb-infected mice is associated with a decrease in leukocytes, neutrophils, mononuclear cells, IL-1β, TNF-α, MIP-2, IL-12, IFN-γ, and MCP-1 (20). Histamine is released during MC degranulation. Carlos and colleagues used histamine-deficient C57BL/6 mice to investigate the role of histamine in Mtb infection, which is detectable in high concentrations 28 days after Mtb infection. Mice lacking histamine showed decreased neutrophils numbers, as well as TNF-α and IL-6 levels in lung tissue, while IL-12 and IFN-γ concentrations were increased. Furthermore, the histamine-deficient lungs showed lymphocytic infiltration with an increase in the number of CD4 + T cells that correlated with reduced bacterial growth (45). Taken together, these findings suggest that MC degranulation may have a complex role in modulating the inflammatory response to Mtb.
It would be important to investigate the redundancy of these pathways in Mtb infection using in vivo models, as well as determine whether MCs are the source of histamine in this context, which may indicate novel therapeutic avenues.

MC eXTRACeLLULAR TRAPS
The formation of ETs, named ETosis, is a type of cell death characterized by release of DNA (46). ETosis differs from apoptosis and necrosis since it lacks DNA fragmentation, disruption of the nuclear envelope, absence of phosphatidylserine in the outer membrane and caspase-independent activation (47). The formation of MCETs upon cell stimulation relies on ROS production by MCs that in turn promotes nuclear envelope disruption and release of DNA together with granular components with antimicrobial properties. The DNA backbone in combination with histones, proteases, and AMPs (39,48) forms physical traps that catch and expose pathogens to high concentrations of antimicrobial molecules ( Figure 1B) (48,49). A human mast cell line (HMC-1), showed ET formation upon L. monocytogenes infection. This Gram-positive bacterium was shown to promote disruption of the nuclear envelope followed by an increase in ROS production. Interestingly, the presence of β-hexosaminidase in the traps was observed to have an antimicrobial activity to intracellular L. monocytogenes (39). E. faecalis was also found to induce MCET after 3 h of incubation with BMMCs. However, the level of MCET observed was lower compared with the one promoted by other bacteria (43). This was possibly due to the low multiplicity of infection (MOI) used in this study (MOI 1:1). Streptococcus pyogenes induced MCET after infection (MOI 25:1) in HMC-1 cells exhibiting cathelicidin LL-37, histones, and tryptase in the traps and in murine BMMCs displaying tryptase and histones (50). This suggests that high bacterial burden promotes MCET activation. Interestingly, besides bacterial stimulation, IL-12 and IL-1β were found to induce MCET containing IL-17 after the stimulation of MCs from skin explants of patients with psoriasis (51).

Mycobacterium-MCeT
Mycobacterium tuberculosis induces neutrophil ETs (NETs). However, Ramos-Kichik et al. have reported that although mycobacteria induce the formation of NETs, which include elastase and histones, the AMPs contained in the NETs are unable to kill mycobacteria (52). Three hour incubation of human neutrophils with the virulent Mtb and the less virulent Mycobacterium canettii showed that both mycobacteria were entrapped in NETs; however, neither Mtb nor M. canettii were killed. In fact, mycobacteria were not eliminated even at low bacterial concentrations (MOI 0.1:1), nor did NETs restrict ongoing mycobacterial replication.
Virulent Mtb secretes ESAT-6 and CFP-10 (10-kDa culture filtrate protein) through the ESX-1 secretory system. Both factors are important for the pathogenic intracellular pore-forming activities and phagosomal subversion observed in the early phase of TB (53). Interestingly, ESAT-6 has been shown to induce extracellular NETs by Ca + influx (54). In addition, Mtb can induce ETs in human macrophages (Mφ) via the ESX-1 system, which is enhanced by IFN-γ (55). By contrast, in highly infected Mφ, it has been reported that after IFN-γ initiates necrosis without NETosis (56). This information suggests that virulent mycobacteria may actively promote NET formation to achieve their own ends of persistence, raising the hypothesis that MCETs may also be involved ( Figure 1B). Although mycobacteria-MCETs have not been demonstrated, MCs produce a large repertoire of immunomodulatory mediators that are known to be contained in traps. Therefore, it is important that future studies investigate the mycobacteria induction of MCETs and the inclusion of antimycobacterial mediators.

MC iMMUne CROSSTALK AnD THe MYCOBACTeRiAL GRAnULOMA
Studies involving a broad array of bacterial pathogens have demonstrated the important role of MCs in promoting recruitment, maturation, and bactericidal activity of Mφ, DCs, and neutrophils ( Table 1). However, the potential roles of MCs in modulating the delicate orchestration of immune crosstalk in mycobacterial immunity have not yet been described ( Figure 1C). MCs could easily coordinate granuloma formation and maintenance. In support of this notion, Taweevisit and Poumsuk reported a correlation between MC numbers and granuloma formation (57). Briefly, 45 lymph nodes from patients with TB lymphadenitis were analyzed to determine the frequency of MCs in the granulomatous region. The authors observed that the number of MCs positively correlated with the number of granulomas. A similar correlation was found between multinucleated giant cells and MCs in the lymph nodes (57). Similar studies have been performed using skin biopsies of patients suffering with leprosy (Mycobacterium leprae) (58). Lepromatous leprosy (disseminated disease with high bacillary load) was associated with the lowest dermal MC density compared with paucibacillary and localized tuberculous leprosy. This suggests that MC functions may have a role in driving a differential susceptibility to these polar forms of leprosy, an as yet poorly explained clinical phenomenon (59). Furthermore, the higher MC numbers located around granulomas in the tuberculous group were considered to be indirect evidence of the role of MCs in activating the immune response to M. leprae infection. Interestingly, numerous MCs were found in the highly fibrotic dermal area and in the epineurial layer of lepromatous leprosy lesions, suggesting that MCs could be involved in the induction of fibrosis, including fibrotic leprosy neuritis (60).

MC-Derived Soluble Mediators and Mycobacterial Granuloma Maintenance
Mast cell-derived LL-37 and CRAMP are bactericidal for Mtb (67). Ramos-Espinosa and colleagues reported that the administration of adenovirus encoding the human cathelicidin LL-37 (AdLL37) and TNF-α (AdTNFα) had a protective role in inducing granuloma maintenance, and thus TB disease reactivation (68). These observations suggest that MCs and MC mediators in particular are involved in granuloma maintenance (Figure 1C). Similarly, von Stebut and colleagues reported that upon encounter with pathogens MCs release pre-stored TNF-α that induces neutrophil recruitment to the site of granulomatous inflammation. This was followed by the release of neutrophil-derived MIP-1α/β and MIP-2 chemokines both responsible for Mφ recruitment. Lack of this immediate pre-stored TNF-α release delayed Mφ recruitment and granuloma formation (69). The study by Carlos et al. discussed earlier also described that during Mtb infection, TLR2 engagement induces cytokine release in the lung; as observed in a reconstitution murine model after 60 days of infection (9). The transfer of TLR2 +/+ MCs into TLR −/− Mtb-infected mice showed diminished lung bacterial growth and an increase of TNF, IL-6, IL-1β concentrations, and neutrophil and mononuclear cell recruitment resulting in the restoration of granuloma formation (9). Therefore, MCs may be involved not only in the early but also in the late phase of infection. Furthermore, MC-derived IL-6 and TNF-α in this phase of infection may contribute to granuloma maintenance (9). The precise contribution of MCs in mycobacterial granuloma maintenance remains an important open question, and in vivo Mtb infection models combined with MC reconstitution experiments may yield critical insights into this area.

COnCLUSiOn
Tuberculosis is a highly contagious infectious disease caused by Mtb which infects billions, and kills millions of people worldwide. Although many efforts have been made to reduce TB mortality, the infection remains one the most important threats to human health. MCs are key lung resident immune sentinels that contribute to antibacterial immunity and are likely to play a key role in TB pathogenesis. A potential important and unique function of MCs is the crosstalk with other immune cells to orchestrate multiple effector functions, which may contribute to granuloma formation and maintenance. We have highlighted the potential roles MCs may play during TB that once addressed could inform the design of novel therapeutic strategies.

AUTHOR COnTRiBUTiOnS
KG-R and SB-P conceived, wrote, designed, and coordinated the manuscript. AG, MA-R, and RH-P provided helpful discussions and edited the manuscript. All the authors read and approved the final manuscript.