Variegated Outcomes of T Cell Activation by Dendritic Cells in the Steady State

Dendritic cells (DC) consist of two main populations: conventional (cDC) and plasmacytoid (pDC) (1, 2). cDC are further divided into the type 1 (cDC1) and type 2 (cDC2) subsets based on their development and direct characterization of their expression of specific surface markers (3, 4). Human and murine cDC develop from progenitors in the bone marrow that disseminate through the blood to secondary lymphoid organs such as the spleen and lymph nodes (LNs) to form a network of systemic cDC that constantly interact with T cells (3–5). In addition, some systemic cDC whose progenitors originally seed peripheral nonlymphoid tissues, then migrate to lymphoid organs (6, 7). The systemic cDC are not directly exposed to extrinsic signals, in contrast to the cDC found at anatomical barriers and the associated lymphoid tissues that are exposed to microbiota and other environmental factors. Instead, systemic cDC, the focus of this review, constantly survey local and circulating Ags derived from parenchymal, interstitial, and other nonbarrier tissues. Furthermore, the systemic cDC can present Ags from pathogens during infections as well as those introduced by i.m. vaccinations (8–10) (Fig. 1). It is important to note that migratory cDC that originate at anatomical barriers also drain to certain specialized LNs and, in the case of some respiratory infections, to the spleen (7, 11). Such migratory cDC have key roles in the initiation of effector responses under proinflammatory conditions (11–16).

Initially, cDC were proposed to function in vivo as “nature’s adjuvant” due to their potent abilities to prime T cells that were first identified in vitro (17–19). However, such efficient priming of effector T cells by cDC in vivo requires sensing of additional signals relayed through specific pattern recognition receptors (20–24). These signals result in a proimmunogenic process that enhances specific functions of cDC and is generally referred to as “maturation,” which results in altered expression of costimulatory molecules, MHCs, and cytokines, as well as other changes in cDC that facilitate enhanced effector T cell priming under proinflammatory conditions (22, 25, 26). Overall, the functional partnership between cDC and T cells is the sine qua non to effective adaptive immune responses.

In contrast to proinflammatory conditions, the term “steady state” refers to the absence of a specific inflammatory process (27, 28). The maintenance of the steady state further depends on undisturbed microanatomical architecture supported by stromal cells of the secondary lymphoid tissues that provide a framework for the interactions between cDC and T cells (8, 9, 27, 29, 30). The functions of cDC in the steady state were unclear at first, and an apparent lack of T cell priming in the absence of proinflammatory stimuli originally led to the designation of such cDC as “immature” immune bystanders (18, 31, 32). However, these conclusions were often based on the results of in vitro experiments using bone marrow–derived DC (BMDC). In contrast to BMDC, which only express appreciable amounts of MHC class II (MHC-II) and other costimulatory molecules upon specific stimulation, endogenous cDC in vivo constitutively express at least intermediate amounts of these molecules (6, 21, 33–36). Subsequent experimental results established that in the steady state, endogenous cDC that are present in both the spleen and LNs can very efficiently activate Ag-specific T cells in vivo (37–40). This functional competency may depend on specific mechanisms that operate intrinsically in cDC under homeostatic conditions and may also develop in response to not yet completely identified molecular ligands that are constitutively present in the steady state (26, 32, 38, 41–46). Such emerging concepts of “homeostatic maturation” are also consistent with observations of multiple specific gene expression changes in some cDC under steady state conditions that are comparable in scope to those observed during TLR ligand-mediated maturation (26, 41–44, 46). Although the specific mechanisms governing cDC functions in vivo under steady-state conditions following their initial development from bone marrow precursors are still being uncovered, it is clear that these processes result in phenotypically heterogeneous cDC populations present within lymphoid organs (41, 42, 46).

Instead of being primed for effector differentiation, T cells activated by cDC in the steady state rather acquire various tolerogenic properties, including a conversion into peripheral regulatory T cells (pTregs) (8, 22, 31, 32). Overall, the current model established by the cumulative experimental evidence collected by multiple investigators specifies tolerogenic roles of cDC in the steady state and contrasts these functions with the effector-priming capacity of cDC under proinflammatory conditions (6, 21, 23, 32, 36, 47) (Fig. 2). Nevertheless, such protolerogenic roles of systemic cDC need to be reconciled with the initiation of autoimmune T cell responses and cancer immunosurveillance also observed in the steady state (48–55). The diversity of T cell responses is further accentuated by the recently uncovered induction of specific effector programs in T cells in the steady state, possibly reflecting functional heterogeneity of homeostatically matured cDC (56). Whereas the exact impacts on T cell responses made by different subpopulations of homeostatically matured cDC remain to be uncovered, such various roles of cDC may, at least to some extent, be orchestrated by the specific immunomodulatory molecules that are differentially expressed in distinct subpopulations of cDC (57). Overall, in the steady state, cDC are emerging as versatile sculptors of the functional repertoires of Ag-specific T cells that are directly relevant to multiple homeostatic and pathologic conditions.

In contrast to some in vitro experimental approaches such as those based on BMDC, the functional analysis of endogenous cDC in vivo is more difficult. Such specific challenges stem from the relatively small numbers of endogenous cDC found at the relevant anatomical locations in situ and are further compounded by the short lifespans of cDC and their high sensitivity to external stimulation (6, 27, 32). Some early attempts to bypass these challenges relied on injecting in vitro prepared BMDC back into the animals, but it is unclear to what extent the in vitro derived and manipulated cells correspond to the bona fide endogenous cDC (21, 58). However, these in vitro systems still have utility for translational research, especially in the development of DC-based immunotherapies for cancer (59, 60). Fortunately, the roles of cDC in the steady state could be established directly by multiple other lines of independent experimental evidence. These approaches, which have been specifically designed to observe the impact of endogenous cDC in situ on T cell responses in vivo, have fallen into two main categories: (1) a depletion of the cDC normally present under steady-state conditions and (2) an introduction of defined T cell Ags into such endogenous cDC (32, 61). Although the experimental scope of these approaches differed substantially, the results of these independent studies pointed to the role of cDC in maintaining immune homeostasis and tolerance (8, 32), as discussed below .

The experimental deletion resulting in the absence of cDC in vivo was achieved by specifically expressing diphtheria toxin receptor followed by diphtheria toxin treatment, expressing diphtheria toxin A subunit directly, or by nonspecific chemical depletion (62–66). Such induced absence of cDC resulted in spontaneous autoimmune T cell activation and increased formation of effector T cells (65, 66), defects in induction of tolerance to systemic Ags (64), disturbance of oral tolerance after exposure to specific Ags (62), and increased susceptibility to colitis (63). As indicated above, these results found key roles for cDC in the maintenance of homeostasis, in agreement with other studies that employed transgenic expression of cytosolic ectopic Ags in cDC as well as earlier studies employing the direct delivery of defined Ags to cDC in vivo for measuring the impact of specific antigenic presentation by endogenous cDC in the steady state (37, 67–70). Although the results of the studies using expression of ectopic Ags in the cytosol of all cDC showed the crucial role of cDC in inducing tolerance of self-reactive CD8⁺ T cells in the steady state, such models are limited to defined MHC-I restricted T cell epitopes (68–70). In contrast, the introduction of defined exogenous Ags to cDC in vivo allowed for controlling the availability of specific Ags to both CD4⁺ and CD8⁺ T cells and have served as a versatile tool for studying T cell responses elicited by cDC in settings that mimic the uptake by cDC of apoptotic and other endogenous materials (32, 37, 40, 67, 71–77).

To obtain their specific presentation by cDC, cognate T cell epitopes can be delivered to cDC directly by using modified mAbs specific for molecules expressed on the surfaces of cDC, and such methods have been broadly adopted over the past two decades, as reviewed elsewhere (78, 79). Particularly, the recombinant Abs that include T cell Ags as fusion proteins within their constant regions have proved to have multiple advantages. In contrast to a chemical conjugation of Ags to native Igs, fusion proteins avoid unintentional stoichiometric differences in the amounts of antigenic molecules and also minimize unintended contamination with endotoxin and other undesired molecules, rendering these recombinant Ag-delivery reagents ideal for basic science and clinical applications (32, 78–80). Furthermore, in chimeric Abs, the original constant regions are replaced with the species-specific constant regions. In the case of the recombinant chimeric Abs used in experimental mice, the engineered murine IgG1 also had additional mutations minimizing their nonspecific binding to Fc receptors (37). In addition to full Igs, single-chain fragment V region and single-domain Ab have also been used for Ag delivery under multiple immune conditions (32, 81–88).

The endocytic receptor DEC-205 (DEC205, CD205, LY75) remains as one of the key molecules targeted by recombinant Abs to deliver Ags to murine cDC (37, 67, 78). DEC-205 is expressed at high levels on cDC1 and mediates an efficient endocytic pathway allowing for robust processing and presentation of Ags from proteins binding to this lectin receptor (31, 78, 89, 90). Ags delivered through DEC-205 are presented in the context of both MHC-I and MHC-II without causing other perceivable changes to cDC in the steady state (37, 71, 78). Although DEC-205 also can be expressed in other immune cells, such as germinal center B cells, the expression of DEC-205 in B cells in vivo is considerably increased only after immunization using adjuvants or in vitro after activation with LPS, anti-CD40, and IL-4 (31, 91, 92). Overall, Ag delivery through DEC-205 and other cDC surface proteins has allowed a comprehensive analysis of the outcomes of T cell activation by cognate Ags presented by cDC in the steady state (38–40, 78, 93). Such experimental strategies that combined Ag delivery methods with various genetic models of cDC have additionally broadened our understanding of the roles of specific cDC populations, particularly clarifying the tolerogenic functions of cDC1 in the steady state (32). Furthermore, the specific methods of targeted Ag delivery to cDC are promising avenues for various immunotherapeutic applications (78–81).

Immune tolerance is indispensable for protecting self-tissues from autoimmune attack, and specific tolerogenic mechanisms also help to regulate the immune responses against infectious agents and cancer cells (32, 94, 95). The immune tolerance of T cells depends on thymic mechanisms that include clonal selection as well as the functions of thymically produced regulatory T cells (tTregs) (96). However, such tTregs can be overwhelmed by specific proinflammatory activation in the periphery. This underscores the risks conferred by autoreactive T cells that survive thymic negative selection yet remain responsive to Ags in the periphery because of the cross-reactive TCRs and the differing efficiencies of antigenic presentation in the thymus and in the periphery (32, 97–106).

The peripherally induced mechanisms of tolerance that prevent the unwarranted priming of such autoreactive T cells include T cell anergy, T cell deletion, and a conversion of pTregs (8, 32, 107). The pTregs, which are de novo converted in response to specific self-antigens, bestow a dominant and long-lasting tolerance to peripheral Ags that can ameliorate various autoimmune responses, therefore crucially complementing the immunological tolerance first initiated in the thymus (8, 32, 102). The induction of systemic pTregs is efficiently mediated in the steady state by cDC with tolerogenic functions that belong to the Batf3-dependent cDC1, corresponding to previously defined CD8α⁺DEC-205⁺ cDC present both in LNs and in the spleen (1, 8, 31, 32, 38). These tolerogenic cDC1 are further distinguished by their high expression of B and T lymphocyte associated (BTLA) and constitute a majority of splenic cDC1 that are mostly considered to be “resident” cDC in the steady state (6, 9, 38). BTLA engages herpesvirus entry mediator (HVEM) in CD4⁺ T cells to modulate a CD5-dependent resistance of developing pTregs to effector-differentiating cytokines such as IL-4, IL-6, and IFN-γ (38, 102, 108, 109). Therefore, functions of the BTLA–HVEM–CD5 axis stabilize and promote the process of pTreg conversion that is initiated and further facilitated by other key molecules, such as TGF-β, retinoic acid (RA), IL-10, CD39, and others (42, 47, 57, 109–116). In particular, RA, together with TGF-β, helps to promote induction of Foxp3 expression in pTregs, and β-catenin governs production of RA and TGF-β in some cDC (111, 117). Such cDC-mediated functions of TGF-β depend on the expression of TGF-β–activating integrin α_vβ₈ (118). In addition to these soluble mediators, other pathways dependent on the immunomodulatory properties of PD-L1/PD-1, CD80/CD86/CTLA-4, and B7h/ICOS signaling axes induce Foxp3 expression in developing pTregs as well as facilitate some other mechanisms of tolerance (57, 68, 119–124).

Because pTregs are Ag specific, the availability of peripheral Ags is necessary for the generation of these cells. Apoptotic materials are a particularly abundant source of tissue-derived self-antigens, and cDC1 efficiently mediate the uptake, processing, and presentation of Ags from apoptotic materials (69, 72, 74, 75, 125–127). BTLA^hi cDC1 reside in the spleen and LNs and are ideally positioned to capture apoptotic and other materials that serve as a source of various systemic self-antigens for induction of tolerance (8, 9, 31, 69, 72, 73). Therefore, the de novo pTreg conversion can take place both in the spleen and in LNs (38, 56, 93). Furthermore, the induction of Ag-specific pTregs does not depend on specific receptors expressed in these cDC to mediate Ag uptake (32, 38, 93). However, the efficiency of such pTreg induction is diminished by the presence of high amounts of the specific Ag, especially if such Ags are presented to T cells also by other cDC that lack pTreg-inducing mechanisms (38, 93, 128, 129). Overall, in addition to the functions of the specific immunomodulatory mechanisms discussed above, the efficient conversion of pTregs requires only a moderate initial activation of T cells, also consistent with the diminished pTreg induction upon activation of cDC by specific proinflammatory stimuli (32, 93, 129).

In addition, some cDC2 can induce pTregs, especially in the intestines, expand the existing tTregs, and promote other forms of tolerance, such as T cell anergy and deletion (130–132). However, in contrast to cDC1, cDC2 are generally less efficient at inducing mechanisms of tolerance. By using a diphtheria toxin receptor expression system in the intestines, the deletion of cDC2 did not exacerbate sensitivity to an inflammatory process, but a deletion of cDC1 exacerbated such inflammation (133). Furthermore, various experimental systems, including those based on Ag targeting in vivo, also revealed a superior ability of cDC1 to induce pTregs in comparison with cDC2 (32, 78).

In addition to the inherent tolerogenic functions of some systemic cDC in the steady state as outlined above and referred to as “natural” tolerogenic cDC (8), some microbiota-associated stimuli as well as certain key endogenous metabolites and vitamins can also induce tolerogenic functions in cDC (8, 134, 135). Conversely, a dysregulation of certain metabolic pathways such as adenosine deaminase may lead to autoimmune stimulation by cDC (136). Various types of cDC with induced tolerogenic functions, especially those that migrate to draining LNs, have important roles for maintaining homeostasis at multiple anatomical barriers exposed to the commensal microbiota and other environmental cues (8, 137–143). Overall, induced tolerogenic functions complement the constitutive mechanisms of tolerance mediated by cDC with inherent tolerogenic roles (8).

Migratory cDC transport Ags from parenchymal tissues to the lymphoid organs in the steady state (144). However, systemic soluble Ags can be delivered to lymphoid organs directly (such as to the spleen via the bloodstream) for presentation by resident cDC (8, 9). Some migratory cDC isolated ex vivo can convert Tregs, but it remains unclear if such cDC induce tolerance in the steady state directly in vivo or if these cells ferry Ags from the peripheral tissues and pass them on to other tolerogenic cDC present in the lymphoid organs (8, 12, 42, 44, 128, 140, 145–149). Furthermore, the Ags that trigger pTreg conversion are readily available in vivo to both migratory and resident cDC in the lymphoid organs (32, 38, 40). However, the de novo induction of such Ag-specific pTregs in the steady state is compromised in vivo in the absence of specific tolerogenic functions of BTLA^hi cDC1 (32, 38). In addition, in the case of some tumors, certain cDC acquire regulatory functions within the tumor microenvironment and may migrate to lymphoid tissues to dampen antitumor responses (150–153).

The functions of pTregs initially arising in response to systemically available self-antigens in the steady state confer dominant tolerance that prevents subsequently triggered organ-specific autoimmunity such as in experimental autoimmune encephalomyelitis, a model of multiple sclerosis, as well as other models of autoimmune diseases (32, 102). Such de novo induced tolerance is specifically perturbed in the absence of Hopx (homeodomain-only protein), a transcription cofactor required for the survival of pTregs under proinflammatory conditions (101, 102, 154). Therefore, the presence of Hopx or other factors that control either the initial conversion or the functions of pTregs is required for the ability of Ag-specific pTregs to restrain the subsequently triggered autoimmune process (101, 102, 155).

The induction of peripheral tolerance is not limited only to induction of mechanisms of tolerance in CD4⁺ T cells. By presenting endogenous Ags to both CD4⁺ and CD8⁺ T cells, cross-presenting cDC1 also help to maintain tolerance to self by directly deleting autoreactive CD8⁺ T cells (31, 156). Various recently identified mechanisms that govern the pathways of cross-presentation in cDC1 may therefore also contribute to these functions of cDC1 (157–159). Overall, in addition to being considered as one of the key safeguards against autoimmune responses, the tolerogenic skewing of T cells in the periphery also provides a means for preventing transplant rejection (160).

The general model of immune system function postulates an immunogenic priming mediated by proinflammatory signals (21, 23, 28, 32, 36, 161–163). Therefore, an induction of T cell tolerance in the steady state is critical for the prevention of subsequent autoimmune responses initiated when self-antigens are presented under proinflammatory conditions. However, such a preferential induction of tolerance in the steady state is not consistent with other experimental evidence indicating the induction of effector characteristics in CD4⁺ T cells under otherwise tolerizing conditions (48–50, 164, 165). In addition, in the absence of sufficient immunoregulation, such T cells with effector characteristics may contribute to autoimmune responses (8, 50, 57, 166, 167). Furthermore, the mechanisms of cancer immunosurveillance postulated to constantly remove cancerous cells arising under homeostatic conditions are inconsistent with the predominantly tolerogenic outcomes of Ag-specific T cell activation under homeostatic conditions (51, 168). Similarly, productive antitumor responses can develop in the absence of major perturbations of homeostasis (52–55, 169).

Overall, such an emerging diversity of T cell fates points to an underappreciated complexity of the biological processes mediated by cDC in the steady state. Recent results identified a process resulting in active programming by cDC of naive CD4⁺ T cells with specific epigenetic and transcriptional instructions leading to an acquisition of specific Th effector functions. Therefore, such “pre-effectors” activated in the steady state become poised for subsequent effector differentiation and, upon restimulation under nonskewing conditions in vitro or in vivo, readily express key factors such as IFN-γ that then can trigger expression of T-bet and possibly other effector master regulators (56). Pre-effectors were shown to contribute to initiation of the autoimmune responses and might possibly contribute to cancer immunosurveillance (56, 170).

In contrast to a binary model of effector versus regulatory fate determination, effector programming in the steady state occurs concurrently with the conversion into pTregs of T cells with the same Ag specificity (56). This dichotomous differentiation also fits a broader paradigm of how Th cell fates may diverge, resulting in different functional outcomes (171–173). Overall, the functionally dichotomous effector and regulatory outcomes of CD4⁺ T cell activation in the steady state may increase the range, plasticity, and regulation of the subsequent immune responses.

The proliferative potential and gain of effector functions are also separable events in the differentiation of CD8⁺ T cells that depends on CD4⁺ T cell help (174). In the absence of such CD4⁺ T cell help, the initial proliferation of naive CD8⁺ T cells may lead to their deletional tolerance (175, 176). Therefore, divergent outcomes of CD4⁺ T cell activation in the steady state are also expected to directly affect CD8⁺ T cell responses. Overall, a regulation of CD4⁺ Th cells whose functions in the steady state can be kept in check by the correspondingly arising Ag-specific pTregs may help to better account for the observed complexities of T cell responses initiated in the steady state (56, 177).

In contrast to the well-defined mechanisms employed by cDC1 for inducing pTregs, the specific processes resulting in the dichotomous tolerogenic and effector outcomes in the steady state remain unclear. Both cDC1 and cDC2 can induce effector programming in the steady state, although such outcomes are enhanced by cDC2 that lack specific mechanisms needed for the efficient conversion of pTregs (8, 47, 56). Therefore, it is likely that a pre-effector T cell fate determination is independent of the specific type of T cell–activating cDC, whereas conversion into pTregs is facilitated by specific mechanisms present only in certain cDC1, as discussed above and as shown in (Fig. 3. Furthermore, one might speculate that the BTLA^lo cDC1 excel in inducing pre-effectors in contrast to BTLA^hi cDC1 that are known to efficiently convert pTregs as discussed above. In addition, the induction of pre-effectors may rely on other specific characteristics of some cDC, such as their expression of effector T cell master regulatory genes and specific metabolic reprogramming (157, 178–180). In addition to the differences in processing and presentation of Ags to T cells, as well as the expression of specific immunomodulatory mechanisms within various cDC populations, the specific functions of such cDC subsets and even individual cDC may be determined by their exact anatomical localization within immune organs (9, 32, 39, 57, 181). In this regard, it is important to recognize that the specific anatomical organization of gut-draining LNs plays an important role in balancing the local tolerogenic and immunogenic responses (141). Other recent results also showed the role of the commensal microbiota in licensing the functions of cDC to directly prime some CD8⁺ T cell responses in the steady state (182).

Furthermore, various extrinsic factors may help determine the balance of tolerogenic and immunogenic T cell responses in the steady state. In addition to serving as a source of Ags, apoptotic materials that engage receptors such as CD36 may mitigate effector-inducing properties, whereas necrotic materials derived from injured cells and recognized by Clec9a (DNGR-1) promote proimmunogenic effects (183–187). As already discussed earlier in the text and as recently reviewed (8), other extrinsic factors can modulate functions of specialized cDC present at specific anatomical locations by activating key pathways, including NF-κB, Wnt/β-catenin, and mammalian target of rapamycin. Future research will continue to clarify how such signaling pathways may contribute to orchestrating a balance between proimmunogenic and protolerogenic functions.

The plethora of diverse T cell fates initiated in the steady state underscores the complexities underlying the orchestration of immune responses. An improved understanding of such intricate mechanisms is also likely to open the door to designing more effective immunotherapeutic approaches against different types of autoimmunity and cancer. In this regard, it should be noted that various tools currently used in basic research to probe the crucial outcomes of interactions between cDC and T cells also have a potential translational value. Overall, the emerging understanding of the specific functions in the steady state of cDC and T cells along with their corresponding mechanisms is expanding the current scope of research in immunology.

Figures were created with Biorender.com.