Innate Control of Adaptive Immunity: Beyond the Three-Signal Paradigm

T cells and B cells are equipped with a diverse repertoire of receptors that are capable of recognizing a vast array of Ags. This diversity allows for protection against constantly evolving pathogens but also gives rise to substantial self-reactivity. Avoiding self-reactivity while maintaining clonal diversity is an intriguing evolutionary design problem. A seminal leap in our understanding of the activation of the adaptive immune system was due to the late Charles Janeway, Jr. He proposed that innate cells should be equipped with germline-encoded pattern recognition receptors (PRRs) to recognize “nonself” conserved microbial components, also referred to as pathogen-associated molecular patterns (PAMPs) (1). Activated innate immune cells would then convey the information about the nature and origin of the Ag to the adaptive immune cells to mount an appropriate adaptive immune response. This idea formed the basis for the now-established three-signal paradigm of innate control of adaptive immunity (2, 3) (Fig. 1A). The first signal provided by innate cells is the presentation of the antigenic peptide, which is necessary for activation of the TCR and clonal expansion of Ag-specific T cells. Because the peptide can be of self or nonself origin, Ag presentation alone is unable to provide any qualitative information about the source of the Ag (4). The second signal is provided via costimulatory molecules that are upregulated on APCs only when the Ag is associated with a PAMP (5, 6). Thus, costimulation is necessary for self versus nonself distinction.

The third signal consists of innate cytokines that are produced as a result of PRR activation (3). The cytokine milieu helps T cells differentiate into protective T cell subsets required for host immunity against a given pathogen (7, 8). The three-signal model described a fundamental link between the innate and adaptive immune systems and defined the requirements for inducing a measurable T cell response, hereafter referred to as “productive immunity.”

The requirement of the concurrent presence of all three signals ensures diverse, but selective, T cell activation. A basic understanding of these signals has helped us design immunogenic Ags that are able to induce productive immunity. However, translating immunogenicity to protective immunity continues to be a challenge (9). So far, most successful vaccines rely on generating B cell responses that result in neutralizing Abs against a given pathogen (10); there has been limited success in defining and inducing protective pathogen-specific T cell immunity. This is largely due to the fact that the three-signal model mentioned above is a vast oversimplification of innate control of adaptive immunity. In addition to the three broad information routes, innate cells provide subtle information about the pathogen to the adaptive immune system, which facilitates protective T cell responses. Experimental settings often limit important host–pathogen-associated variables by using purified ligands, model Ags, unnatural routes of infections, and purified cell types. Although a reductionist approach is essential to gain mechanistic insights into a pathway, such isolated challenges are never presented to the host during a natural infection. Microorganisms carry ligands for multiple PRRs and activate various cells types during the course of an infection. History of prior infection or presence of coinfection can further complicate the process of innate activation (11, 12). Also, it has become increasingly clear that the route of infection and the priming microenvironment have an enormous effect on T cell differentiation (13). Depending on the class of pathogen, innate cells also produce the IL-1 family of cytokines (14) and/or type I IFNs (15) that further modulate T cell function. Even after T cell differentiation, the presence of innate cytokines in the surrounding environment maintains the T cells in a fairly plastic state, suggesting a persistent interaction between the innate and adaptive immune systems (16). In summary, innate control of adaptive immunity is a complex process of information transfer that extends beyond the three-signal paradigm (Fig. 1B). Although fully elucidating such complexities will continue to be a challenge, in this article we discuss examples of context-dependent innate activation that regulate and fine-tune T cell immunity to a far greater extent than previously appreciated.

The discovery of various classes of PRRs and identification of their microbial ligands has led to a detailed understanding of innate immune recognition (15, 17, 18). PRRs are strategically located in subcellular compartments based on the nature of their ligands that activate unique signal-transduction programs necessary for host defense (19). A detailed description of specific PRRs and their signal-transduction pathways can be found in other reviews (18–20). In this article, we focus on how activation of individual or multiple PRRs can affect Ag presentation and, in turn, adaptive immunity.

Although each class of PRR induces distinct innate responses, the activation of adaptive immunity depends on their ability to induce dendritic cell (DC) maturation (5, 6). Following PRR activation, DCs upregulate MHC and costimulatory molecules and migrate to the draining lymph nodes to interact with naive T cells that are specific to the microbial Ag (21). DCs process and present extracellular and endosomal peptides on MHC class II molecules to activate CD4 T cells, whereas peptides derived from cytosolic proteins are presented on MHC class I to activate CD8 T cells (22). Some specialized DCs also have the unique capability to take up and present exogenous peptides onto MHC class I through the process of cross-presentation (23, 24). This enables DCs to activate CD8 T cells against tumors or viruses that did not infect the DCs directly. One of the first steps by which the innate immune receptors affect adaptive immunity is by influencing the handling of endocytic and phagocytic cargo. Activation of PRRs, especially TLRs, was shown to accelerate endocytosis and phagocytosis, thus affecting Ag processing and presentation (19, 25, 26). However, recent studies shed light on how activation of certain TLRs can also dictate cargo handling following uptake (27–29). Although activation of surface TLR4 transiently enhances cross-presentation (30), phagosome-intrinsic TLR signaling provides self versus nonself identity of the cargo and promotes MHC class II presentation and CD4 T cell activation (31, 32). It was recently shown that activation of TLR4 and TLR2 leads to upregulation of transcription factor EB, which drives rapid degradation of phagosomal content and prevents its cross-presentation on MHC class I molecules, thus enhancing antigenic peptide loading on MHC class II molecules (33). This suggests that a predominant CD4 T cell response is generated in response to extracellular bacterial pathogens. Conversely, TLR9 stimulation does not induce transcription factor EB expression, which protects cargo from degradation and makes it available for cross-presentation through the cytosolic route or by capturing the cargo into recycling endosomes (33). Because TLR9 is likely activated by viral DNA following phagocytosis of an infected cell, this process ensures the generation of effective CD8 T cell responses to eliminate viral infections. Activation of TLR3, TLR7, and TLR9 by nucleic acids also was shown to enhance cross-presentation in an IFN-α/β receptor (IFNAR)-dependent manner (34). Type I IFN–treated human DCs showed delayed degradation of internalized Ag that enhanced their capacity to cross-present Ags and promote CD8 T cell activation. These studies highlight the ability of the innate immune system to convey qualitative information about the cargo to the adaptive immune system by dictating how the cargo is handled and presented. Thus, a cell’s decision to present peptides on either MHC class I or MHC class II molecules is not merely a reflection of the origin of the Ag but also a consequence of the complex interplay between the PRRs that are engaged. Although advantageous for pathogen-specific immunity, this mechanism also exposes the host to manipulations by the microbes. For instance, certain viruses have the ability to engage plasma membrane TLRs (35, 36) that can potentially suppress antiviral CD8 T cell activation. Alternatively, bacteria can mask their TLR-activating PAMPs (37), thereby enhancing cross-presentation and the generation of CD8 T cells that can cause immunopathology and tissue destruction, facilitating bacterial dissemination.

In addition to regulating Ag presentation, PRR activation leads to induction of costimulatory molecules that are necessary for T cell activation (5, 6). One of the most well-studied costimulatory interactions is that of CD80/86 with CD28 expressed on T cells. CD28 signaling lowers the threshold for TCR activation and leads to transcription of proliferative genes and IL-2 production (38). In addition, there are various TNFR superfamily members that serve as costimulatory molecules. Several TNFR family members, such as OX40, 4-1BB, and CD27, have the ability to promote effector functions of specific T cell subsets (39, 40). In addition, differential Notch ligand expression by DCs has been implicated in differential priming of T cells (41), and Notch signaling is reported to be important for activation and effector function of CD4 T cells (42). It remains to be examined whether specific PRRs control adaptive immunity by influencing expression of a particular set of costimulatory molecules on DCs. It is also unclear how T cells integrate diverse costimulatory signals to mount a protective response. A substantial amount of work has been done in understanding the role of specific costimulatory molecules in T cell functions; therefore, we refer readers to those reviews (40, 43–45) for information on how various costimulatory molecules affect adaptive immunity.

Although individual PRRs have significant immune-modulatory capabilities, protective host responses are often dependent on the ability of the innate immune system to simultaneously engage multiple PRRs. TLR3 and TLR4 synergize with TLR7, TLR8, and TLR9 to enhance the production of IL-12 and IL-23, as well as the Th1-driving capacity of DCs (46). Cooperation between PRRs was shown to result in unique transcriptional changes that control the quality, quantity, and kinetics of chemokine and cytokine production (47). One very well-known example of such PRR synergy is inflammasome activation by virulent pathogens. Sequential (48) or simultaneous (49) activation of TLRs and NLRs leads to the production of bioactive IL-1β and IL-18, which influence T cell immunity, as discussed later. Similarly, activation of dectin-1 was shown to be important for TLR-induced reactive oxygen species production (50) that can lead to inflammasome activation (51, 52) and IL-1 production, thereby promoting antifungal Th17 responses. Moreover, the consequence of PRR cross-talk is also determined by the innate cell type. Zymosan-mediated coactivation of dectin-1 and TLR2 enhances IL-12 and TNF-α production in bone marrow–derived macrophages (50), whereas, in bone marrow–derived DCs, dectin-1 suppressed TLR2-dependent IL-12 (53). Dectin-mediated modulation of IL-12, IL-6, and IL-23 production can promote Th17 differentiation at the expense of Th1 differentiation. These examples underscore the importance of coincident activation of PRRs and suggest that, although individual PRRs can operate in isolation, their cooperation is important to generate an optimal adaptive immune response. In addition to synergistic cross-talk between PRRs, there are examples of cross-inhibition. Negative cross-talk, which might have evolved to prevent immunopathology, can sometimes result in exacerbation of infection. While isolated activation of endosomal TLRs promotes CD8 T cell priming, simultaneous engagement of surface and endosomal TLRs inhibits CD8 T cell expansion (54). The dominant effect of surface TLRs might be a consequence of preferential recruitment of myeloid DCs that are incompetent in inducing CD8 T cell priming and leads to CD4-biased responses (54). This is particularly relevant in the case of viral–bacterial coinfections and provides a potential mechanism for how bacterial infections suppress antiviral immunity. Similarly, peptidoglycan-mediated activation of NLRs in DCs in vitro inhibited Ag cross-presentation induced by TLR4 activation (55). Another instance of a contextual outcome of PRR cross-talk is the engagement of the Mincle receptor, a C-type lectin that can act as an inhibitory or activating receptor, based on the nature of the pathogen. Activation of Mincle by Leishmania suppresses DC maturation and IL-12 production, likely by negatively regulating inflammatory signaling through other PRRs (56). Hence, deletion of Mincle receptor resulted in improved Th1 differentiation. However, Mycobacterium induces IL-6 production by macrophages in a Mincle-dependent manner, which results in reduced Th17 responses in Mincle-deficient mice (57). These examples illustrate the importance of studying PRR cross-talk using PAMPs associated with the pathogen in their native form, to reveal their true impact on host immunity. Collectively, these studies highlight the extensive intra- and intercellular PRR cross-talk that takes place during infections and its consequential effects on the quality of T cell responses.

In addition to Ag presentation and costimulation, the third signal that shapes the quality of T cell responses is the production of innate cytokines. NF-κB–dependent cytokines, such as IL-12, IL-6, and TNF-α (20), drive acute inflammation and dictate T cell differentiation. Two other major classes of innate cytokines, IL-1 family and type I IFNs, often cooperate with priming cytokines in the generation of protective T cell immunity. Both immune and nonimmune cells can produce the IL-1 family of cytokines (IL-1α, IL-1β, IL-18, and IL-33) as an outcome of recognition of virulence factors, cellular stress, or cell death. In contrast, type I IFNs are produced following activation of nucleic acid sensors in the endosomes (TLR3, TLR7/8, and TLR9) or in the cytosol (RLRs and cGAS) (15). The IL-1 family of cytokines and type I IFNs drive systemic immune responses by inducing an acute-phase response and a global antiviral state, respectively.

Unlike IL-6, IL-12, or IL-4, IL-1 family members and type I IFN cytokines are largely produced as a result of cytosolic invasion by virulent pathogens or extensive tissue damage, with certain exceptions, such as cytosolic sensing of LPS (58, 59) or mRNA from pathogenic and nonpathogenic bacteria (60–64). Hence, they are secreted in a much more rapid fashion than priming cytokines. IL-1a and IL-33 can be rapidly released upon the death of necrotic cells (65, 66), whereas IL-1 and IL-18 are present in a premade pro-form in certain cell types and can be acutely secreted via rapid inflammasome activation (49). In contrast, type I IFNs rely on IRF3- and IRF7-mediated signaling. Plasmacytoid DCs can rapidly secrete type I IFNs because they exhibit higher baseline expression of IRF3 and IRF7 (67). The nature of the cellular source of IL-1 family members and type I IFNs also differs from cells that produce priming cytokines to modulate T cell activity. Recent reports show that nonimmune cells can also provide IL-1–related cytokines to promote effector function of tissue-resident T cells (68, 69). Also, in addition to plasmacytoid DCs, every nucleated cell in the body has the capacity to produce type I IFNs in response to viral infections. However, T cell–priming cytokines are understandably produced by professional APCs, such as DCs, and inflammatory cells, such as monocyte-derived DCs and macrophages. Therefore, it is likely that, whereas the priming cytokines dictate T cell differentiation in the secondary lymphoid organs, the IL-1 and type I IFN family of cytokines operate in the periphery to imprint further differentiation for tissue-specific effector function. Below, we discuss how each of these families has a distinct, as well as context-specific, role in dictating the magnitude and quality of T cell responses.

Distinct differentiation states of CD4 T cells are traditionally defined by the lineage-specific transcription factors and their effector cytokines. Unique T cell subsets are largely shaped by the innate cytokines presented during priming. In brief, IL-12 promotes Th1 differentiation by driving T-bet expression (70). IL-4 drives Th2 differentiation in both a paracrine and autocrine manner (71–74). The combination of IL-6 and TGF-β promotes Th17 differentiation (75). IL-6 and TGF-β can synergize with IL-23 and IL-1 to further stabilize the Th17 program (76). In the absence of an infection, TGF-β produced by migratory DCs induces regulatory T cells (Tregs) that are critical for maintaining peripheral tolerance (77). We refer our readers to other reviews for detailed information on the role of individual priming cytokines (7). We will limit our discussion to the context-dependent roles of innate cytokines in T cell differentiation. T cell subsets are often studied using in vitro systems through the use of defined priming cytokine cocktails that result in highly polarized T cell populations. However, during a real infection, activated DCs generate a complex milieu of innate cytokines, leading to the priming of a heterogeneous CD4 T cell population. Therefore, in vitro–polarized T cells might not represent physiological differentiation processes that give rise to a diverse effector population in vivo. Recent studies using whole-pathogen lysates and live microbes have begun to unravel the complexities of cytokine requirements for T cell differentiation. Priming naive human CD4 T cells using whole-pathogen lysates revealed qualitatively distinct patterns of cytokine production between Th17 populations induced by Staphylococcus aureus and Candida albicans (78). S. aureus–induced Th17 cells secrete IL-17, as well as IL-10, independent of IL-1β stimulation, whereas C. albicans–specific Th17 cells produce IL-17 and IFN-γ, requiring IL-1β in addition to IL-6 and IL-23. Furthermore, a mouse model of Citrobacter rodentium infection showed that IL-22–producing T cells, dependent on IL-23 for differentiation, were required for protection (79). These cells expressed RORγt, the master transcriptional regulator of the Th17 lineage, but failed to produce IL-17. In the context of experimental autoimmune encephalomyelitis, IL-23 was alternatively shown to induce “pathogenic Th17” cells, which produce IFN-γ and GM-CSF in addition to IL-17 (80). In agreement with these studies, transcriptomic analysis showed variable gene-expression profiles within the Th17 population (81). Collectively, these findings suggest that, depending on the immune challenge, naive T cells get exposed to a unique innate cytokine milieu that drives functionally peculiar T cell subpopulations, even though they may belong to the same umbrella lineage. Studying functional T cell subsets based on the innate cytokine stimulation that drove their differentiation, rather than their narrow effector cytokine profile, will likely reveal more valuable information regarding the generation of protective T cell responses against specific pathogens.

Priming cytokines, like IL-6, IL-12, and IL-4, often synergize with the IL-1 family of cytokines for generation of protective immunity. Although the exact role of each IL-1 family member in regulating T cell responses is not clear, there is substantial evidence that IL-1β, IL-18, and IL-33 contribute to Th17, Th1, and Th2 CD4 T cell responses, respectively (82). A series of studies verified the importance of IL-1 in Th17 cell biology (83–86). T cell–intrinsic IL-1R signaling was shown to promote pathogenicity of Th17 cells by synergizing with IL-6 and IL-23 (87). IL-18 has been implicated in Th1 and CD8 T cell function. The combination of IL-12 and IL-18 induces IFN-γ production by memory CD8 T cells and Th1 cells, independent of TCR activation (88, 89). IL-18 can also enhance T-bet expression in human CD4 T cells, thus promoting Th1 function (90). IL-33 is produced as an alarmin when cells undergo inflammatory cell death (66). The adaptive function of IL-33R signaling was first discovered in the context of type 2 immunity, because this receptor was prominently expressed on Th2 cells (91). Recently, it was found that colonic (92), adipose tissue–resident (69), and muscle (68) Tregs also express high levels of IL-33R (ST2). IL-33R signaling promotes proliferation and maintenance of induced Tregs, a novel mechanism of IL-33–dependent suppression of inflammation (69, 92). IL-33R–deficient mice also show higher viral burden during lymphocytic choriomeningitis virus (LCMV) infection, which was recently attributed to Th1-intrinsic (93) and CD8 T cell–intrinsic (94) IL-33R signaling. These studies emphasize the multifaceted nature of the IL-1 family of cytokines and show that they can act as mediators of acute inflammation during host invasion, as well as drivers of homeostatic adaptive immunity in the steady-state.

IL-1 and its relatives are produced under strict regulation, likely as a result of their highly inflammatory nature. Inflammasome activation is a major mechanism for the production of bioactive IL-1β. It is important to note that a major consequence of inflammasome activation is pyroptotic cell death, which can potentially compromise Ag presentation and the generation of T cell immunity. Over the years, many inflammasome-independent pathways involving caspase-8 and neutrophil-derived serine proteases have also been reported for IL-1 maturational cleavage (95, 96). Although there is lack of evidence, it is tantalizing to hypothesize that, although the inflammasome-dependent IL-1β drives systemic inflammation, inflammasome-independent mechanisms of IL-1 production have evolved for much subtler functions, such as influencing T cell biology during priming, as well as reactivation.

Type I IFNs are proinflammatory cytokines that are produced in response to viral infections. In the absence of TCR signaling, type I IFNs trigger a proapoptotic program, potentially to prevent bystander T cell activation (97). However, direct activation of IFNAR can act as a costimulatory signal when engaged in sync with TCR (97). In fact, activation of IFNAR signaling on CD8 T cells was shown to bypass the requirement of CD4 T cell help for CD8 immunity during LCMV infection (98). IFN-α synergizes with IL-12 to provide signal 3 during activation of naive, as well as Ag-experienced, CD8 T cells for their clonal expansion and effector function (99–101). IFN-α/β can act together with IL-12 to enhance Th1 and CD8 T cell immunity (102, 103). IFN-α enhances IL-12 and IL-18 responsiveness by increasing IL-18RA expression on T cells (104). While promoting Th1 bias in human T cells, type I IFNs also actively suppress Th2 cell function (105–107). Interestingly, the nature of the pathogen appears to dictate the requirement of type I IFNs for adaptive immunity. Lack of T cell–intrinsic IFNAR signaling resulted in compromised CD8 T cell expansion and memory formation following LCMV infection, whereas vaccinia virus–specific, vesicular stomatitis virus–specific, and Listeria monocytogenes–specific CD8 T cell expansion was independent of type I IFN–mediated signaling (108). This differential requirement extends to CD4 T cells as well, where type I IFNs were critical for CD4 T cell expansion following LCMV, but not L. monocytogenes, infection (109). Because type I IFNs are emerging as an attractive choice for vaccine adjuvants, these studies suggest that we need to be conscious of their differential pathogen-specific impact on T cell immunity for successful vaccine design (110).

In summary, PRR activation results in the production of a functionally distinct set of innate cytokines that are crucial for T cell differentiation and function. Priming cytokines trigger transcriptional programs for lineage differentiation, and the IL-1 family of cytokines and/or the type I IFN family cooperate with priming cytokines to dictate the final outcome of effector differentiation. Interestingly, there seems to be spatial and temporal distinctions in how various innate cytokines contribute to T cell immunity. Moreover, pathogens often dictate the set of innate cytokines required for protective T cell immunity (109, 111). Therefore, it is crucial to be able to induce specific combinations of innate cytokines with appropriate kinetics and tissue distribution to harness the innate immune system for the generation of long-term pathogen-specific T cell immunity.

Different anatomical tissues are exposed to distinct microbial and environmental challenges. To prevent immunopathology, not every infection experiences the full potential of the immune system. Instead, the intensity of the T cell response is likely to be modulated by the local, as well as the systemic, immune status of the host. Infection at the barrier tissues, such as the skin or gut, presents a moderate risk to the host. Therefore, the immune response is often contained in the tissue and, in some cases, even leads to tolerance. In contrast, systemic infection (bacteremia or viremia) is treated as an immediate threat and triggers a systemic inflammatory response. Previous infections in the host, even after resolution, can further alter the immune response to new infections. Therefore, the anatomical location of the infection and host immune status are critical determinants of protective immunity.

Most of our understanding of the immune system is derived from studies done in secondary lymphoid organs, such as the spleen or lymph nodes. However, it has become increasingly clear that secondary lymphoid organs do not faithfully represent immune responses in the peripheral tissues. Tissue-specific immunity can be largely attributed to resident DC populations that are unique to the tissue and have a differential ability to produce innate cytokines. For example, intestinal DCs at steady-state are primarily exposed to commensal microbiota. Therefore, the gut is enriched with retinoic acid–producing CD11b⁻CD103⁺ cells, which specialize in Treg induction (77). During enteric infection, IRF4-dependent CD11b⁺CD103⁺ DCs populate the gut and drive Th17 cell differentiation by providing TGF-β, IL-6, and IL-23 (112–114). However, when the mucosal barrier is breached, leading to a systemic infection, BATF3-dependent CD8⁺ DCs, which predominantly reside in the spleen, are activated and produce IL-12 to induce a strong Th1 response (115). Jenkins and colleagues (116) elegantly demonstrated the impact of the route of infection on T cell responses. When injected i.v., L. monocytogenes induced a Th1 response, whereas intranasal infection with the same strain resulted in Th17 differentiation (116). This may be due, in part, to the IL-6– and TGF-β–producing DC population in the nasal-associated lymphoid tissues. Interestingly, the requirement for Th17 differentiation also varies based on the priming microenvironment (84). IL-6 was required for Th17 priming in the gut, potentially to overcome the suppressive effects of TGF-β (84). However, Th17 priming in the spleen, which does not have TGF-β–producing DCs, was independent of T cell–intrinsic IL-6R signaling (84).

Similar to the gut, the dermis and epidermis of the skin are also populated with specialized DC populations. Langerhans cells in the mouse epidermis promote Th17 differentiation following fungal infection in the skin (117). Unlike the epidermis, the dermis is lined by CD103⁺ DCs that are responsible for cross-presentation to CD8 T cells, as well as generation of Ag-specific Th1 cells (117–120). Therefore, the precise location of the pathogen, within a tissue, can also influence the nature of the resulting T cell response, despite the fact that the T cells are eventually primed in the lymph node draining the infected area. Moreover, a recently identified dermal DC population marked by the expression of CD301b was reported to be required for Th2 responses after allergen challenge or worm infections (121). A detailed functional and transcriptional profiling of the CD301b⁺ population might prove to be useful in predicting the assortment of innate cytokines that they produce upon activation to identify innate signals that drive Th2 differentiation. It also remains to be examined whether these DCs have dedicated receptors or other sensory mechanisms that alert them to the presence of an allergen or helminth infection.

Recent studies identified a tissue-resident memory T cell population that serves as the front-line defense against reinfections (122–125). Specific DC populations have now been identified that preferentially generate tissue-resident T cells (126, 127). These findings present us with an exciting opportunity to induce tissue-specific immunity by individually targeting relevant DC subsets.

Infection models in mice often ignore the impact of prior infections. However, in reality, no one is absolutely naive to microbial challenge; concurrent or past infections can affect the host response to unrelated pathogens. History of infection or ongoing pathogenic insult can heavily skew the cytokine milieu for T cell differentiation, jeopardizing immunity to the secondary pathogenic insult. In response to helminth infection, DCs promote a strong Th2 response that indirectly inhibits Th1 or CD8 T cell function (128). Helminth infection also drives a regulatory program by promoting the differentiation of alternatively activated macrophages and enhancing production of TGF-β and IL-10 by DCs (129). A regulatory program aids wound healing that helps the worm to navigate through the host, yet it also drastically inhibits production of proinflammatory cytokines, like IL-12 and TNF-α, that are crucial for Th1-mediated immunity. This could be a contributing factor for the widespread prevalence of helminth coinfection with Mycobacterium and HIV in which Th1 cells are essential for protection (130, 131). Another prevalent coinfection is the viral bacterial coinfection. How viral infection enhances susceptibility to bacterial infection is still an open question, but many potential mechanisms have been suggested over the years. Viral infections appear to alter metabolic profiles in immune and nonimmune cells that might not be conducive for antibacterial responses (132, 133). There is also a significant reduction in tolerance toward tissue damage during coinfection that can further contribute to morbidity and mortality (134).

Overall, ongoing and past infections can significantly dictate the quality and strength of innate immune activation. One of the major challenges in developing successful vaccines is the heterogeneity of immune responsiveness toward the vaccine. Genetic diversity of the population contributes to variability; however, it is worthwhile to characterize an individual’s immune status as a prognostic tool for predicting the efficacy of immunization.

We have discussed the contextual nature of the three traditional signals that are critical for innate control of adaptive immunity. However, it is becoming clear from several recent studies that the composition of the microbiota and the quality of the metabolites produced by the commensal microbiota can have considerable influence on myeloid and T cell behavior, especially in the tissues (135). In the absence of infection, innate immune cells, such as DCs, provide tolerogenic signals to T cells to avoid aberrant T cell activation against innocuous self-antigens. This is primarily achieved by providing self-peptide–MHC interactions with self-reactive TCRs in the absence of costimulation and inflammatory cytokines. However, commensal microorganisms pose a unique challenge because they possess nonself molecules (PRR ligands), much like virulent pathogens, which can induce all three innate signals required for T cell activation and differentiation. Harmful immune activation of the adaptive immune system against commensals seems to be contained by the host by actively sensing microbial-derived metabolites. Commensal microbiota produce abundant quantities of short-chain fatty acids (SCFAs), such as butyrate, acetate, and propionate, as end products of the fermentation process. In particular, butyrate produced by a certain class of microbiota appears to downregulate LPS-induced proinflammatory cytokine production by lamina propria macrophages (136). Similar findings were made in human monocyte-derived DCs, in which butyrate and propionate suppress inflammatory responses by reducing the secretion of IL-6, IL-12, and chemokines important for leukocyte trafficking (137). This could have a profound effect on suppressing T cell activation induced by aberrant recognition of PAMPs on commensal microorganisms that breach the intestinal barrier. Interestingly, the anti-inflammatory effects of butyrate are independent of TLR signaling and are mediated by inhibition of histone deacetylases (136). In addition to suppressing proinflammatory responses by activated myeloid cells, SCFAs can facilitate M2 polarization of macrophages by directly activating STAT6 signaling (138). Whether SCFAs can influence innate control of T cell responses has not been directly tested; however, butyrate-producing Clostridia was shown to attenuate graft-versus-host disease (139), a disease with a strong T cell component. Testing the precise role of microbial metabolites in the innate control of adaptive immunity is challenging, because butyrate and propionate can also act directly on T cells to promote colonic Treg generation and function by enhancing FOXP3 transcription (140–142). Moreover, SCFAs can also affect intestinal homeostasis by acting on nonimmune cells. The presence of butyrate in colonic crypts inhibits stem cell proliferation and delays intestinal wound repair (143). Therefore, deletion of G protein–coupled receptors that sense SCFAs specifically in innate immune cells, such as DCs or macrophages, is required to test the effects of metabolites on the innate control of adaptive immunity. In addition to SCFA, AHR-binding tryptophan metabolites (144) and the vitamin A lipid metabolite, retinoic acid, can control intestinal homeostasis by fine-tuning the Treg/Th17 balance (145). Although the exact role of various metabolites and their mechanism of action are still under investigation, it is clear that metabolites can modulate adaptive immunity independently of the innate signals discussed earlier in this review, extending the three-signal paradigm.

Qualitative information about the commensal microbiota can also be relayed to DCs via epithelial cells. Specific members of the microbiota, segmented filamentous bacteria, tightly adhere to the intestinal epithelium and stimulate Th17 differentiation (146). Other adhesive bacteria, such as C. rodentium and Escherichia coli O157, were also shown to induce Th17 responses in an adhesion-dependent manner (147). Adhesion of bacteria results in production of serum amyloid A by epithelial cells that can act on DCs to promote Th17 differentiation (148). It is important to note that the presence of segmented filamentous bacteria could also promote protective Th17 responses against C. rodentium, suggesting that adhesive bacteria are not exposed to APCs but perhaps modulate the DC–T cell interaction milieu (146). How and what about the adhesion of bacteria to the intestinal epithelium is sensed by the immune system are not clearly understood and are exciting venues for future research.

In this review, we attempted to expand our linear understanding of innate control of adaptive immunity to highlight the complexities of innate signals that dictate the outcome of adaptive immune responses (Fig. 1). Although our current knowledge base allows us to generate specific T cell lineages, there is still a lack of understanding of how to precisely induce tailored T cell responses that are protective against a given pathogen. Inducing protective T cell immunity includes, but is not limited to, generating the appropriate composition of T cell lineages, imprinting tissue-homing information, and providing cues for memory formation (recirculating and tissue resident). We have highlighted the merits of studying the Ag in association with its native PAMPs and using natural routes of infection. Moreover, it is becoming clear that architectural cell types, such as epithelial and endothelial cells, have an integral role in modulating host immunity (149, 150). As discussed above, attachment of certain commensal bacteria to the epithelial surface is crucial for preferentially inducing Th17 differentiation (147). Additionally, dietary habits play an important role in shaping our steady-state immune status. Such nonimmune cues are often disregarded in experimental settings because there are no in vitro models that faithfully represent the interplay between the epithelia and the immune system. Organoid culture systems that can accommodate immune and nonimmune cells might prove to be beneficial for future studies. Nevertheless, it is virtually impossible to study every immune response in all its complexity. An evolutionary perspective can provide important clues into the blueprint of innate control of adaptive immunity. There are several instances in which a fundamental theme underlies the complexity of immune responses that we see today. For example, MyD88 serves as a signaling adaptor for TLRs and the IL-1 family of receptors; therefore, our understanding of the TLR signaling pathway in innate cells can help us to delineate the role of the IL-1 family of receptors in T cell biology. Similarly, as various DC subsets are being identified, functional patterns are also becoming apparent. Migratory DCs that are critical for Ag delivery to the secondary lymphoid organs are present in the gut, as well as the skin, and are characterized by CD103 expression. Likewise, DNGR1⁺ DCs were shown to induce tissue-resident memory T cells during liver and skin infection (126, 127). These findings point toward a possibility of convergent evolution that took place under diverse tissue-specific immunological pressures and gave rise to highly specialized, but functionally similar, DC subsets. Identifying such underlying principles is going to be as important as collecting contextual information to reach the ultimate goal of successfully reverse-engineering the process of innate control of adaptive immunity and rationally generating protective T cell responses.

We thank all of the members of the Pasare Lab for helpful discussions and critical reading of the manuscript.

The authors have no financial conflicts of interest.