Following BCG intradermal inoculation, resident epidermal macrophages interact with BCG via several pattern-recognition receptors (PRRs), including complement receptor 3 (CR3) (18) and toll-like receptors 2 and 4 (TLR2/4) (19). C-type lectin family receptors such as the mannose receptor (MR) and the macrophage inducible Ca²⁺-dependent lectin (MINCLE) receptor are expressed on macrophages (20–22), but direct interaction between them and BCG, and its subsequent outcome, has not yet been described. Because the peripheral lipid portion of the cell wall is very similar between BCG and M.tb (23), it is predicted that their ability to infect tissue macrophages will be similar. However, BCG’s first contact occurs with resident epidermal macrophages, whereas M.tb contact, in the majority of cases, occurs with resident alveolar macrophages (AMs). Differences in the mechanisms of Ag recognition, Ag uptake, Ag processing, and Ag presentation between these two types of resident tissue macrophages remains unclear and may contribute to the reasons behind why BCG is not fully protective. Thus, resolving this discrepancy will be important to identify if epidermal vaccination is sufficient to protect against lung disease. In fact, the effect of serum opsonization is often overlooked during BCG vaccination (24–26). This process is thought to be crucial in initiating immune responses to the BCG vaccine. As an example, the host opsonin factor H, a regulatory protein of the complement system that downregulates the alternative complement cascade, can bind to the BCG cell surface (27) and partially inhibit its uptake by epidermal macrophages. Macrophages infected by factor H opsonized BCG can secrete elevated amounts of pro-inflammatory cytokines, potentially driving an acute response (high IL-6/TNF) (27). Thus, factor H opsonization of BCG could be detrimental to the primary goal of BCG vaccination (to generate a strong T cell memory response). This is probably because opsonized BCG gets killed too quickly, reducing the amount of time Ag is available for presentation, and thus negatively impacting T cell proliferation and recruitment to the site of infection. Like factor H, there are other serum opsonins in the epidermal tissue and thus, further studies are necessary to assess their effects in the generation of immunity to BCG.

From the mycobacterial perspective, the TB field assumes that BCG cell wall components will interact in a similar fashion with macrophage receptors as M.tb cell wall components do (28). However, differences exist, such as in the case of the mannose-capped lipoarabinomannan (ManLAM) of M.tb vs. BCG. The degree and pattern of mannose capping and fatty acid content in M.tb ManLAM vs. BCG ManLAM differs (28) and thus, it may dictate how this molecule is differentially recognized by macrophage receptors such as the MR, TLR2 (dimerized with TLR1 or TLR6), or TLR4, generating different immune responses that subsequently may affect Ag presentation. In fact, BCG macrophage stimulation via TLR2 or TLR4 drives differences in pro-inflammatory responses, T cell proliferation, and IFNγ secretion in vitro and results in differences in bacterial burden in the lung in vivo (19). Indeed, prolonged Ag stimulation of TLR2 in macrophages downregulates the expression of major histocompatibility complex (MHC) class II and affects MHC I Ag cross processing, thus, reducing Ag presentation to T cells (29). It is still not clear to which degree suppression of MHC I and II via TLR2 will negatively impact the protective immune response generated by BCG vaccination and/or against M.tb infection. In this regard, studies have shown the potential of introducing an adjuvant during BCG vaccination targeting TLR-7/9 signaling, which restores and/or increases the expression of MHC II in macrophages, thus enhancing their ability to present Ag (30).

From the host perspective, differences in tissue-resident macrophages (epidermal vs. lung) expressing different PRR types and levels (31, 32), as well as human polymorphisms in these receptors, could influence immune responses to BCG and subsequently its effectiveness. Studies performed using BCG and M.tb indicate that the nature of their cell wall components engaging a specific macrophage PRR determines the host immune response generated. For example in macrophages, engagement of ManLAM or higher-order phosphatidyl-myo-inositol mannosides (PIMs) to the MR drives an anti-inflammatory response; however, engagement of trehalose 6,6′-dimycolate (TDM) to MINCLE or engagement of lower-order PIMs to CR3 and/or TLRs results in pro-inflammation (28). Thus, the balance between pro- and anti-inflammatory lipids present on the BCG and M.tb cell wall is a factor to consider in the protective ability of BCG. Reducing pro-inflammation at the site of vaccination may be beneficial as macrophages will be exposed to the Ag for a longer period of time, and thus provide additional time for the generation of T cell immunological memory. If we consider that epidermal macrophages and AMs in the lung have and/or use different PRRs for BCG recognition that leads toward site-distinct mechanisms of Ag presentation, then a safe delivery of BCG into the lungs could be a way to optimize BCG efficacy.

Differences in intracellular trafficking and Ag processing could also explain the failure of BCG to confer long-term immunity. In fact, BCG, as well as M.tb, can block phagosome maturation (33–35) by engaging specific receptors, i.e., the MR (36), and although MHC Ag presentation and activation of the adaptive immune system is unaffected, it may be less than optimal (37, 38). In this context, exogenous induction of phagosome–lysosome (P–L) fusion (39), autophagy (12), and/or increasing phagosome leakage of Ag across the phagosome membrane (40), could enhance the Ag-presentation process and thus, BCG efficacy.

An important step forward in the development of a new, effective TB vaccine, focused on optimizing Ag presentation, came from the development of the rBCG ΔureC∷hly vaccine (41). This vaccine, when administered percutaneously, has superior efficacy than BCG by further reducing the bacterial burden in the lung of M.tb-infected mice. Recent studies also indicate that vaccination with BCG ΔureC∷hly increases macrophage apoptotic vesicle formation, thereby inducing more robust CD4 and CD8 T cell responses (42), as well as increasing the gene expression of “Absent In Melanoma 2.” This increases the autophagic pathway and inflammasome activation, which in turn improves control of M.tb infection (43). Thus, although P–L fusion is an important mechanism, modulation of apoptosis and autophagy can also offer novel avenues for vaccines against TB. Mechanistic studies such as these, focusing on the underlying factors behind enhanced immunity offer valuable insight into the requirements for efficient mycobacterial clearance.

Extending on the above findings, ex vivo M.tb-infected lung macrophages from BCG-vaccinated guinea pigs are shown to secrete larger amounts of IFNγ, TNF, and IL-12p40, highlighting their appropriate anti-mycobacterial response (44). Other supporting ex vivo studies indicate that AMs from BCG-vaccinated guinea pigs challenged with M.tb express significantly less IL-10 and more IL-12p40, compared to unvaccinated controls (45). This same study reported an elevated expression of MHC II on peritoneal macrophages from BCG-vaccinated guinea pigs (45). Thus, based on our current understanding of the requirements for mycobacterial immunity; immunity induced by the BCG vaccine results in responses to M.tb at the early phase of infection that directly impact macrophage function.

Altogether, there is a significantly large body of knowledge on the interactions between BCG and macrophages and how these responses can influence protective immune responses against M.tb. Receptor interaction between the macrophages and bacteria are critical in initiating the response, where interactions with different macrophage receptors can differentially modulate trafficking pathways and processing and presentation of Ag. However, additional studies determining the exact role of epidermal macrophages vs. AMs following BCG vaccination are needed before macrophage functions can be exploited to further improve vaccine development strategies. The innate immune response to BCG are highlighted in Figure 1.

Dendritic cells are classified as one of the most potent APCs (46) and classically described as the modulators of cross talk between innate and adaptive immunity (47). Phenotypic and functional differences exist between DCs and macrophages; however, the primary difference is that DCs, in contrast to macrophages, migrate from tissues via the lymphatic system and enter the draining lymph nodes, where they present Ags to naïve T cells (48) leading to the induction of effector T cell responses (49, 50). As DCs migrate, they mature by upregulating MHC II and CD80, CD86, CD40, and CD54 costimulatory molecules, all involved in the activation of adaptive immune cells (48, 51). Several phagocytic receptors on DCs recognize BCG, including CR3 (CD11b/CD18), CR4 (CD11c/CD18), dendritic cell-specific ICAM-3 grabbing non-integrin (DC-SIGN, CD209), and DEC-205 (CD205) (52–54). Of all these receptors, DC-SIGN is possibly one of the most important as neutralization of DC-SIGN with antibodies (Abs) inhibits the interaction of BCG with DCs by 80% (51). Signaling receptors such as TLR2 and TLR4 are also shown to be involved in DC activation and maturation by their interaction with BCG cell wall components (i.e., mAGP complex) (51, 55, 56).

Following BCG vaccination epidermal DCs initiate adaptive immune responses in vivo by trafficking from the site of inoculation to draining lymph nodes where they present Ag to adaptive immune cells. In this context, migratory DCs display an EpCAM^lowCD11b^high phenotype and are capable of priming CD4⁺ T cells via interleukin-1 receptor and myeloid differentiation primary response gene 88 signaling pathways (57). However, most of our understanding of DCs and BCG immunology is from studies conducted in vitro. Stimulation of DCs with BCG in vitro induces homotypic aggregation, upregulation of surface Ag presenting molecules, downregulation of endocytic activity, and release of TNF (58), implying DCs initially respond to BCG. However, it is difficult to discern whether these responses are adequate for optimal immunity. BCG can be engineered to enhance DC activation by the addition of M.tb genes encoding specific mycobacterial proteins into BCG (59), suggesting that DC responsiveness to BCG may not be optimal. Though immunity can be improved by augmenting DC activation, we still lack a fundamental understanding of the exact interactions or processes that result in optimal immunological protection. From the perspective of the host, mice lacking specific genes (e.g., IFNγ, IL-12, and TNF) have been valuable in deciphering some of the requirements for effective immunity to BCG and M.tb; however, they offer little insight as to the concentration and spatial organization required for their optimal efficacy. Transgenic mouse models allowing manipulation of specific gene products or conditional knockout mice allowing for spatial regulation could be useful in elucidating the necessary elements for optimal BCG efficacy in terms of space and time.

Stimulation of human DCs with BCG can increase surface expression of MHC-II, CD40, CD44, CD54, CD80, and CD86, all markers involved in DC activation, maturation, migration, and Ag presentation to T cells (60). DCs infected with BCG also secrete TNF, IL-1β, IL-6, IL-4, and IL-10, but not IL-12 (61). Strikingly, IL-4 is secreted at large concentrations indicating that its presence could shift immunity from a T_h1 polarized response toward a T_h2 response, potentially affecting BCG efficacy (62). However, recent studies disproved this hypothesis by showing that BCG-infected DCs cocultured with T cells are capable of inducing T cell proliferation and IFNγ secretion in vitro, the defined T_h1 cytokine critical for the control of M.tb (63). However, IFNγ is shown to correlate poorly with protection against mycobacteria infection and disease (64, 65). Thus, more refined markers such as abundance of memory lymphocytes in the lung and multifunctional T cell diversity have arisen as better indicators of functional immunity (discussed below) (17, 66). Furthermore, reports indicate that M.tb hinders the efficiency and effectiveness of Ag presentation by macrophages and DCs (67, 68), whereas BCG is much more efficient at stimulating CD4⁺ T lymphocytes. Thus, although BCG may effectively stimulate the process of Ag presentation, the capacity of BCG-infected DCs to present Ag may be at a saturation point, and any further stimulation of the adaptive immune system may not be achievable (69). In this context, attempts such as BCG-expressing Fms-like tyrosine kinase 3 ligand [Flt3L, a hematopoietic growth factor that stimulates DC proliferation, and whose deletion reduces T cell responses by 50% (70)] did not show further enhancement of BCG efficacy against M.tb challenge. This finding is independent of early expansion of DCs and increased stimulation of BCG-reactive IFNγ-secreting T cells (71). As BCG may be less efficient in inducing DC maturation than M.tb, understanding the differences between them could lead to greater approaches in improving the interaction between BCG and DCs and thus, host immunity (72).

Several groups have exploited DC maturation to improve BCG immunity, with studies centered in modulating TNF and IL-10 dominating the field. TNF neutralization inhibits DC maturation post-BCG inoculation suggesting an important role for TNF in immunity generated by BCG (60). As a result, an adequate concentration of TNF is required for optimal DC Ag-presenting efficacy. BCG-infected DCs also secrete large quantities of IL-10, but not IL-12 (61). IL-12 is, however, highly expressed in IL-10^−/− mice following inoculation with BCG indicating the existence of an important balance between IL-10 and IL-12, which could activate/accelerate maturation of DCs (35) and serve as a potential host-directed therapy. These findings may explain why DCs from IL-10^−/− mice are more efficient at activating T cells in the draining lymph nodes (DLNs) (73). Thus, IL-10 may not only downregulate the migration of infected DCs to the DLN but also regulate IL-12 production and subsequent DC capacity to mature, diminishing T cell activation and proliferation, repressing the adaptive immune response generated by BCG (73, 74), and possibly affecting the development of immunological memory. Hence, downregulation of IL-10 and upregulation of IL-12 could to be important for generating an optimal immune response to BCG.

Apart from being powerful mediators of the immune response to BCG by initiating innate responses, DCs also translate information to the adaptive branch of the immune system and initiate the first steps that subsequently give rise to immunological memory. With this in mind, it is logical to explore mechanisms that can enhance DC–BCG interactions with intentions of amplifying BCG efficacy. Although the DC–BCG interaction appears to be highly efficient, it remains unclear why they fail in the lung. One could speculate that too much focus has been placed on DC responses to BCG in vitro and thus, it has misconstrued our notion of DC responses.

Circulating human neutrophils comprise approximate 60% of the blood cells, have a short half-life of 6–10 h, and are one of the first cells to respond to foreign molecules (75). Neutrophils secrete large amounts of chemokines and cytokines, priming long-lived phagocytes (76). Interaction of neutrophils with BCG increases their expression of adhesion markers CD11b and CD18, FcγRs II and III, and stimulates their upregulation of cytokines (e.g., IL-1α, IL-1β, and TGFβ) and chemokines (e.g., IL-8, CCL2, and CCL3) (77). How changes in neutrophil phenotype upon contact with BCG can influence the generation of protective immunity against M.tb is still unclear.

Neutrophils are capable of shuttling live BCG via the lymphatic system into DLNs and into the vicinity of DCs and T cells (78). The cross talk between BCG-infected neutrophils and DCs in this location delivers maturation signals to immature DCs and also assists DCs in their presentation of BCG Ags to prime CD4⁺ and CD8⁺ T cells (78–80). The delay in apoptosis observed in BCG-infected neutrophils supports the concept that BCG may use them as a vehicle to disseminate from the site of inoculation to DLNs (81). However, whether the host benefits from the neutrophil–BCG interaction remains unanswered. Additionally, the longer BCG remains within neutrophils increases the probability the bacteria will be killed by neutrophil intracellular mechanisms, possibly enhancing pathways involved in immune activation via Ag presentation (82). Studies conducted in the C3HeB/FeJ mice, which develop necrotic and hypoxic tubercle granulomas, found that BCG vaccination in these mice was associated with long-lasting immunity and reduced bacterial burden. The association was attributed to a reduction in the numbers of neutrophils (83). Thus, neutrophils may behave as a double-edged sword; on one hand, they seem to facilitate Ag presentation by shuttling live bacilli to the vicinity of DCs, but on the other hand, they may be responsible for preventing the development of long-lasting immunity. The balance between BCG’s intracellular survival and/or digestion, processing, and presentation by the neutrophil may impact DC maturation, the subsequent adaptive immune response (limiting the clonal expansion of Ag-specific effector lymphocytes), and the long-term establishment of a state of protective immunity.

Conversely, neutrophils can also direct activation of T cells in lymph nodes, and thus participate in the generation of adaptive immune responses (80, 84). A recent study indicates that neutrophils expressing CCR7 migrate to the lymph nodes in response to CCR7 ligands, CCL19 and CCL21. This migration seems enhanced during in vivo injection of complete Freund’s adjuvant (containing inactivated M.tb cell wall) in wild-type mice, but not CCR7^−/− mice (85). To further illustrate a role for neutrophils, studies using mice inoculated with BCG via the intranasal route show neutrophils entering the lungs in two waves. The first wave arrives between 1 and 3 days post-inoculation and could kill M.tb. The second wave of neutrophils enters the lung 3 weeks post-inoculation together with IFNγ- and IL-17A-producing T cells. This second wave of neutrophils is not associated with the ability to kill M.tb, and their movement into the lungs is dependent on the expression of IL-17RA (86). Subsequent studies show that M.tb-infected neutrophils are a prominent population in the lungs early during infection (87), and that M.tb-infected neutrophils promote adaptive immune responses to M.tb infection. Recent studies using BCG also demonstrate that neutrophils regulate inflammation via the secretion of IL-10, impairing the control of M.tb growth during chronic infection (88); thus, further establishing the dual role for neutrophils during infection, being active in not only controlling M.tb infection but also regulating the inflammatory response generated during infection. Since neutrophils enter tissues much sooner relative to other host cells, they could be a critical mediator in clearance or persistence of mycobacteria. However, how the regulatory functions of neutrophils influence adaptive memory responses generated by BCG vaccination remains unanswered.

Altogether, neutrophils can have a wide range of effects in the context of BCG vaccination. On one hand, they may help facilitate adaptive immune responses by aiding in Ag presentation and shuttling live bacteria into the vicinity of professional APCs, but on the other hand, their presence and propensity to induce strong inflammatory responses may be detrimental to the tissue in which they reside, propagating the disease, and potentially inhibiting the development of long-lasting immunity. Indeed, it remains unclear whether neutrophil shuttling of BCG is beneficial or detrimental as they may ultimately be used as a vehicle from which to disseminate to other organs. Further studies as to the role of neutrophils and mechanism in which they may be involved post-BCG vaccination could shed some light on the complex behavior of neutrophils.

T cell responses arise in parallel through engagement of the T receptor with foreign Ag presented by Ag-presenting cells (89). With few exceptions, all vaccines stimulate the proliferation of CD4⁺ helper T cells (T_h) and CD8⁺ cytotoxic T cells (T_c) (14). In the context of BCG immunity, helper T cells primarily differentiate into two distinct classes of effector cells during an immune response to vaccination: Th1 cells identified by the production of IFNγ, and Th17 cells identified by the production of IL-17A, though IL-4 producing T_h2 cells can also be generated (90). Other cell subsets such as T regulatory cells (T_reg) and CD1-restricted T cells also arise following BCG vaccination, albeit to a lesser extent. BCG also induces cytotoxic T cells, whose main function is to lyse infected cells through osmotic disruption (91). It is clear that both CD4⁺, and to a lesser extent CD8⁺, T cells are critical for protection against M.tb infection; experimentally highlighted using mouse knockout models (92–95) supporting a dominant role for T cells as the main effector cells following immunization with BCG. The adaptive immune responses to BCG are highlighted in Figure 2.

The generation of long-lasting immunity to most pathogens by current effective vaccines relies on long-lived humoral immune responses that mediate protection via Abs (14). The BCG vaccine is a strong inducer of humoral immunity; however, due to the intracellular nature of M.tb infection, the potential of Abs as significant contributors to protective immunity has been largely disregarded. However, new evidence has emerged identifying a potential role for Abs and B cells in the immune response generated by BCG vaccination (147, 148). Researchers have begun to unravel a more significant role for B cells than solely Ab production in the context of mycobacterial infections (149, 150).