These T cells, expressing the unconventional γδ TCR, were discovered from the accidental identification of the TCRγ chain (8). γδ TCRs and αβ TCRs have qualitatively distinct modes of antigen recognition; γδ TCRs are not restricted to the recognition of peptides bound to MHC molecules (9). Unlike conventional αβ T cells, cytokine stimulation, or bacterial contact, is sufficient for activation γδ T cells, making these cells rapid and potent mediators of inflammation.

They are much less abundant than classical αβ T cells (1–4%) in thymus and lymphoid organs of adult mice, but they are in highest abundance in mucosal sites, being ~20–40% of the intestinal intraepithelial T cells, ~10–20% of total T cells in the reproductive tracks, and ~50–70% of skin dermal T cells (10).

In humans, the population of peripheral blood γδ T cells is increased in response to infections (11). Initial characterization of human γδ T cells suggested that antigens recognized by γδ T cells were small, non-peptide compounds that contained critical phosphate residues (12). The mainstream γδ T cells in human peripheral blood express the TCR Vγ9Vδ2, and they can recognize (E)-4-hydroxy-3-methyl-but-2enyl pyrophosphate (HMBPP), which are usually referred as phosphoantigens, derived from various bacteria (13). Moreover, human Vγ2Vδ2⁺ T cells can expand 2- to 10-fold during infections and recognize primary alkylamines derived from microbes, releasing interleukine-2 (IL-2) (14). Lysates or culture supernatants from many bacteria (including mycobacteria, other Gram-negative and Gram-positive cocci, protozoal parasites, and even plants extracts) stimulate Vγ2Vδ2⁺ T cells (15). Thus, human peripheral blood γδ T cells can respond to specific antigens from bacteria [e.g., Mycobacterium tuberculosis (16) and Listeria monocytogenes (17)]. Non-peptidic mycobacterial ligands in human Vγ9Vδ2⁺ T cells induce massive tumor necrosis factor (TNF) production (18). Moreover, Vγ2Vδ2⁺ T cells respond to non-peptide bacterial antigens predominantly producing Th1 cytokines such as interferon-γ (IFN-γ), although few of them (<5%) also produce IL-4 (15). It has been reported that Helicobacter pylori can directly interact with human peripheral γδ T cells in vitro, upregulating the activation molecule CD69, TNF-α, IFN-γ, and chemokines, such as MIP-1β and RANTES, favoring an inflammatory environment (19).

In mice, γδ T cells expand dramatically after challenge with Mycobacteria, Listeria, and Salmonella spp. (20–22), rapidly producing cytokines. They are able to produce IFN-γ after L. monocytogenes infection and IL-4 in response to Nippostrongylus brasiliensis (23). Moreover, it has also been reported that γδ T cells are the major IL-17 producers during infections (24, 25).

Invariant NKT cells express an invariant TCRα chain and recognize lipid and glycolipid antigens presented by CD1d, a non-polymorphic MHC class I molecule (37). They also express several receptors, such as NK1.1 (in some mouse strains), and members of the Ly49 family that are typical of the NK cell lineage. iNKT cells are most abundant in liver, thymus, spleen, and bone marrow, but are also found in lymph nodes, peripheral blood, adipose tissue, skin, and mucosal surfaces of intestine and lungs.

It has been described that iNKT cells participate in the response to microbial pathogens in mice (38, 39) (Figure 1B). Upon activation, they produce cytokines such as IL-4 and IFN-γ. iNKTs can be activated by TCR stimulation with microbial antigens presented by CD1d (direct activation) or with endogenous antigens and/or cytokines (indirect activation) produced by APCs. The glycosphingolipid (GSL) α-galactosyl ceramide (α GalCer) was the first antigen identified recognized by mouse [Vα14 (40)] and human [Vα24 (41)] iNKT cells. Recently, it has been described a novel GSL antigen for iNKT cells, DB06-1, which induces preferentially IFN-γ in vivo in mice (42).

It has been reported that mouse iNKTs recognize cell wall GSL expressed by Sphingomonas spp., which are Gram-negative bacteria and have abundant GSLs, similar to α GalCer (43). Mice deficient of Vα14 iNKT cells present reduced spirochete clearance and are more susceptible to chronic inflammation following Borrelia burgdorferi infection (44). Furthermore, iNKTs recognize glycolipids expressed by Helicobacter pilori (45) and diacylglycerol-containing glycolipids from Streptococcus pneumoniae and group B Streptococcus (46). After infection of mice with S. pneumoniae, Vα14 iNKT cells produce IFN-γ, through TCR engagement (46). Recently, it has been shown that respiratory infection with Francisella tularensis, Gram-negative facultative intracellular bacteria that cause lethal pulmonary tularemia, activates iNKT cells which produce IFN-γ and propagates a sepsis-like proinflammatory response (47).

On the other hand, it has been demonstrated in both humans and mice that the responses by iNKT to some bacteria, such as Salmonella typhimurium, is due to indirect recognition of endogenous lysosomal GSL expressed by activated DCs combined with TLR activation (43, 48).

Mucosal-associated invariant T cells express a semi-invariant TCRα chain that recognizes small molecules, pterin analogs, and riboflavin metabolites, presented by the non-polymorphic MHC class I-related molecule, MR1 (53). MR1 is highly conserved (90% gene sequence identity between mouse and human), allowing for considerable species cross-reactivity of MAIT cells. Human MAIT cells develop effector capacity before exiting the thymus, in contrast to conventional T cells that remain naïve until antigen stimulation in the periphery (54, 55). Human MAIT cells are found principally in lungs, liver, and blood (56, 57), but they are less abundant in common laboratory mouse strains. Recently, it has been discovered that MAIT cells are more frequent in inbred CAST/EiJ mice than in C57BL/6 (58).

The role of MAIT cells in the control of bacterial infection was observed due to the absence of peripheral MAIT cells in germ-free mice and its expansion after microbial colonization (59). Both human and mouse MAIT cells can recognize bacterial (and fungi) infected cells in an MR1-dependet manner (56, 60). Human MAIT cells produce proinflammatory cytokines, e.g., IFN-γ and TNF in response to infection with Mycobacterium smegmatis, Escherichia coli, Salmonella enterica, or Staphylococcus aureus (56). However, not all microorganisms tested can activate MAIT cells. Several bacteria, such as Enterococcus faecalis, group A Streptococcus, and L. monocytogenes, do not stimulate human and mouse MAIT cells, neither do viruses (56, 60). Indeed, only the microorganisms that can synthetize riboflavin metabolites which bind MR1 are able to activate MAIT cells (i.e., most bacteria and some fungi) (53, 61). In agreement, MAIT cells accumulation in lungs of mice infected with S. typhimurium depends on microbial riboflavin synthesis (62). However, despite the fact the viruses do not produce riboflavin, viral infections (e.g., HIV) can reduce the numbers of peripheral MAIT cells (by mechanisms that are not totally understood), therefore increasing the susceptibility to opportunistic infections of bacteria and fungi (63). Nevertheless, an open question remains to be elucidated is how MAIT cells are able to recognize specific pathogens. Recently, using MR1 tetramers, it has been identified populations of MR1-restricted cells which assist to have different antigen recognition in humans (64) and mice (65).

In addition to the secretion of inflammatory cytokines, human MAIT cells achieve antibacterial immunity destroying infected cells by secretion of cytotoxic granzyme and perforin (66, 67). The role of MAIT cells during an infection in vivo has been demonstrated using MR1-deficient mice. These mice infected with Klebsiella pneumoniae develop higher bacterial burden, hypothermia, and have increased mortality in the first 4 days of infection compared with infected WT mice (68). In other infection model, using Mycobacterium bacillus Calmette–Guérin (BCG), MR-1 deficient mice also show higher bacterial burden in the lung compared to the WT mice (69). In both models, protection by mouse MAIT cells occurs within the first days of the infection, suggesting that they act as innate lymphocytes (Figure 1A).

Finally, it remains unresolved whether MAIT cells expressed TLRs influencing its activation or function upon stimulation.

It is well known that B-lymphocytes, components of the adaptive response and responsible for humoral immunity, are also APCs. They can capture soluble antigens or antigens exposed by macrophages, DCs, and follicular (FO) DCs (70). There are several subtypes of B cell lymphocytes that include B-1 (B-1a and B-1b) and conventional B-2 cells that comprise two populations designated as MZ and FO B cells.

B-1 cells are not considered part of the adaptive immune system, as they do not develop into memory cells. B-1 cells have been identified in both human and mouse, but due to the logistical difficulties in isolating B-1 cells of humans, the vast majority of the studies have been performed in mouse models (71). B-1 resides principally in the peritoneal and pleural cavities but is in a minor fraction in lymph nodes and spleen (72). B-1 cells in the peritoneal cavity express CD11b (Mac-1) and are subdivided based on the expression of CD5.

B-1 cells play a relevant role in innate immunity by their contribution in the first line of defense against bacterial infection. B-1 cells alter their normal migration patterns (73), accumulating rapidly in the omentum, lymph nodes, and spleen, following activation by stimuli, such as IL-10, IL-5 (74), TLR agonists, such as LPSs (73, 75), or even whole bacteria, such as S. pneumoniae (76) or Borrelia hermsii (77). The exit of these cells from peritoneal cavity in response of LPS or bacteria is controlled by myeloid differentiation primary response protein 88 (MyD88), a key adaptor for TLRs signaling, that downregulates the expression of integrins and CD9, thereby promoting cell migration (73). After migration, B-1 cells differentiate and secrete rapidly large amount of IgM and/or IgA (73, 75). They are able to produce antibodies in response to T-cell-independent type 2 antigens (mainly repetitive structures from encapsulated bacteria) along with MZ B cells (76).

Regarding the phagocytic capacity of B-1 cells after bacteria encounter, there are several reports showing that B-1 cells are able to phagocyte S. aureus, E. coli, and polystyrene fluorescent microspheres (78, 79). Upon phagocytosis, B-1 cells kill internalized bacteria via phagolysosomes and present bacterial antigens in MHC-II molecules (78, 79). Salmonella spp. has been shown to be degraded through both proteasomal and lysosomal processing, resulting in MHC-I antigen presentation (80). On the other hand, not all bacteria are killed; some Salmonella can survive within B cells, using therefore B cells as Trojan horses to disseminate through the infected host (81), similarly to what has been observed in myeloid cells (82). It has been also described that B-1 cells undergo differentiation to acquire a mononuclear phagocyte phenotype in vitro (B-1CDP), and they are able to phagocytose Coxiella burnetii and kill them more effectively than peritoneal macrophages and bone marrow-derived macrophages (BMMf) (83). Moreover, it has been demonstrated that B-1 cell differentiation into phagocytes occurs also in vivo (84). These results revealed that mammalian B-1 cells have phagocytic and microbicidal abilities to strengthen the innate nature of these cells (Figure 1B). In agreement with the innate behavior of B1 cells, a recent report shows that during aging, mouse B1 cells express high levels of the costimulatory molecule CD86 and become potent activators of CD8⁺ T cells, a role deserved for specialized populations of APCs (85).

The current consensus is that B-1 cells are phagocytic, whereas the phagocytic abilities of mouse conventional B cells and the mechanisms for bacterial uptake are less clear (79). BCR-blocking antibodies do not alter the internalization of bacteria, indicating that BCR is not involved in bacteria entry. In contrast to these data, it has been demonstrated that mouse liver B cells (both B-1 and B-2 cells) actively phagocytose and kill bacteria, such as E. coli, in a complement-dependent manner (86). Accordingly, it has been shown that splenic mouse B cells can internalize opsonized Brucella abortus (87), but B. abortus can survive inside B cells. In addition, it has also been reported that S. typhimurium are able to infect and survive within both mouse splenic B-1 and B-2 cell subpopulations. Salmonella infection stimulates expression of PD-L1 on mouse B cells, suggesting that PD-1/PD-L1 pathway may be involved in turning off the cytotoxic effector response during persistent infection (80). Furthermore, it has been reported that human primary B cells are able to internalize S. typhimurium (88). This process is BCR mediated and leads to efficient antigen loading into MHC-II, inducing CD4⁺ T cell help to boost Salmonella-specific antibody production. Salmonella-specific B cells that phagocytose Salmonella upon BCR ligation reactivate human memory CD8⁺ T cells via cross-presentation (89). Additionally, it has been demonstrated that both human peripheral blood and mouse splenic B-lymphocytes serve as a niche for intracellular Salmonella promoting systemic spreading of infection (81).

Conventional B-2 cells are divided into two populations designated as MZ and FO B cell. MZ B cells are a special population of mostly non-recirculating B cells enriched primarily in the MZ of the spleen. They are one of the first cells that take contact with blood-borne pathogens, supporting the first line of host defense. Pathogens trapped in the MZ activate MZ B cells, which maturate to plasma cells secreting IgM or to APCs. It has been reported that MZ B cells capture, process, and present antigens to T cells more efficiently than FO B cells both in vitro (90) and in vivo (91). MZ B cells appear, therefore, as excellent APCs, which, together with lymphoid DCs, play essential roles in the initial steps of in vivo T-cell activation. Consequently, they participate in T-cell-dependent (TD) immune response through the capture and import of blood-borne antigens to FO areas of the spleen (91).

Mice depleted of MZ B cells and infected with B. burgdorferi show elevated pathogen burden and reduced levels of B. burgdorferi-specific IgG and IgM, correlated with diminished splenic CD4⁺ T-cell responses (92). Similarly, these mice show an increased susceptibility to S. aureus (93). The clearance of the bacteria L. monocytogenes depends on the interactions between marginal zone macrophages (MZ M) and MZ B cells (94). MZ M bind pathogens and capture antigens through various PRRs, including scavenger receptors and C-type lectin receptors (CLRs) (95), and then, they expose native antigens and establish direct cell–cell contact for the activation of MZ B cells (96) that are required for potent responses (94).

Innate response activator (IRA) B cells have recently been described as a B cell population that protects against microbial sepsis in mice. They are accumulated in the spleen in a mouse model of sepsis and in response to E. coli infection, indicating that IRA B cell expansion is an overall characteristic of the body’s reaction to bacteria (111). IRA B cells are different phenotypically and functionally from other B cell populations. They contain large amounts of intracellular IgM and spontaneously secrete IgM, but not IgA or IgG1. In addition, they are able to secrete granulocyte macrophage colony-stimulating factor (GM-CSF) and IL-3 (Figure 1B). IRA B cells derived from B-1 cell precursors are activated by TLR stimuli, and they protect against septic shock by controlling neutrophil-dependent bacterial clearance (111).

B-1a cells migrate to the lung in response to microbial airway infection, producing IgM. This process is depended on IRA B cells, which controls IgM production via autocrine GM-CSF signaling, conferring a first-line defense against bacteria in the lungs (112).

Recently, it has been described IRA B cells in humans. They reside in tonsils, within FO areas, which are the first route of defense from infection of the upper respiratory tract and are able to phagocyte bacteria, such as S. aureus, at least in vitro (113).

Therefore, IRA B cells seem to play important roles in bacterial clearance, but further work is required to clarify its function in vivo, and it remains to be studied whether IRA B cells directly interact with infecting bacteria and the nature of such interactions.

Innate lymphoid cells are a recently identified member of the lymphoid lineage, which are enriched at epithelial barriers, such as skin, intestine, and lung, where contacts with microorganisms normally occur (114, 115). These innate lymphocytes mediate immune responses against infections and regulate homeostasis and inflammation (116, 117). They neither express TCRs or BCRs nor respond in an antigen-specific manner.

Innate lymphoid cells are divided into three subsets: group 1 ILCs (ILC1s and NK cells), group 2 ILCs (ILC2s), and group 3 ILCs (ILC3s and LTi cells). This nomenclature was unified to classify these emerging cell populations, which had been called by different terms including NK-22 cells, LTi-like cells, natural helper cells, nuocytes, and innate helper cells (118). ILCs are crucial in the protective immunity against bacteria (ILC1s and ILC3s) (119–121), intracellular parasites (ILC1s) (122), fungi (ILC3) (123), and parasitic worms (ILC2s) (124, 125).

In addition to the specialized lymphocyte populations with innate functions described above, we have recently described that conventional CD4⁺ T cells, the paradigm of the adaptive immunity, also play innate-like roles during bacterial infections, contrary to the current view of immunology (7). CD4⁺ T cells of both mouse and human origin are able to internalize different bacteria (pathogenic and non-pathogenic) such as L. monocytogenes, S. aureus, E. coli, and S. enterica from infected DCs, in a process called transinfection. Bacteria play a passive role in this process, driven by T cells (7); therefore, it would be more appropriate to term it transphagocytosis. Transphagocytic (ti) CD4⁺ T cells kill internalized bacteria in a manner reminiscent of innate immune cells and secrete proinflammatory Th-1 cytokines (IFN-γ, TNF-α, and IL-6) in a rapid innate-like response (Figure 1B). Furthermore, tiCD4⁺ T cells protect against bacterial infections in vivo, highly reducing the bacterial load found in liver and spleen 24 and 48 h after infections, contributing to the early innate immune response (7). This route of bacterial capture by T cells could be used for some pathogenic bacteria to spread. In this regard, it has been shown that T cells can serve as reservoir of bacteria in vivo (160–162). Moreover, Shigella flexneri manipulates the migration capacity of infected T cells in a type III secretion system-dependent manner (161–163). Transphagocytosis depends on T cell cytoskeleton, but the molecular mechanisms of how T cells can capture bacteria remain largely unknown. T cells are unable to directly capture bacteria (7); transphagocytosis requires T cell/DC intimate contact, and it is enhanced by antigen recognition by the TCR. On the other hand, T cells are unable to uptake latex beads from DCs, indicating that bacterial PAMPs are also involved in the transphagocytic process and suggest a role of T cell TLRs in this recently discovered process of bacterial uptake by CD4⁺ T cells.