Dendritic cells represent a minor proportion of the hematopoietic compartment present in the postnatal thymus under steady-state conditions. Thymic DCs are able to present antigen in a tolerogenic fashion9 to those CD4+CD8+ double positive (DP) thymocytes that have survived positive selection. Depending on the overall TCR signal quantity and/or quality delivered during this interaction, potentially self-reactive thymocytes are either deleted by apoptosis (negative selection), or undergo commitment towards CD4+ CD25+ Foxp3+ natural regulatory T cells (nTregs), both mechanisms critically contributing to the establishment of central tolerance.10–12 While the role that thymic DCs play in negative selection has been firmly established, their contribution to the generation of nTreg cells remains controversial, especially because thymic epithelial cells (TECs) can function as inducers of nTregs.13–16 Adding complexity to this scheme, several groups have shown that thymus-resident DCs include both pDCs and cDCs subtypes.17–20 In the last years, nTregs have emerged as key players in the induction of immunological tolerance and, therefore, in the protection against fatal autoimmunity throughout life.21 As a consequence, a great interest has emerged in understanding how each intrathymic DC subtype could contribute to the generation of nTregs, especially in humans.

Seminal studies by Liu and coworkers provided direct evidence that human thymic cDCs display tolerogenic potential and generate nTregs, at least under allogeneic conditions in vitro.22 A latter work revealed that only one of the two subsets of cDCs found in mouse thymus supports nTregs development in vivo, and suggested that thymic pDCs display poor tolerogenic function.15 In contrast, we have recently shown that pDCs resident in the human thymus support the differentiation of autologous nTregs from DP thymocytes as efficiently as cDCs, through a mechanism that may involve the CD40-CD40L pathway.23 Moreover, pDCs and cDCs were shown to play a non-redundant functional role, as they selectively induced one of the two nTreg subsets, i.e. IL-10- or TGF-β-producing nTregs, previously identified in the human thymus.24 Therefore, current information points to thymic DCs as key regulators of central tolerance. Whether functional specialization of thymic pDCs and cDCs does exist owing to their distinct antigen presenting abilities25 is an interesting and still open question.

The important tolerogenic role of thymic DCs and their expected therapeutic potential has led to a great interest in defining the developmental origin of these cells. Pioneering work by Ardavin and collaborators showed that early thymic progenitors (ETPs) from the murine thymus displayed cDCs potential in vivo.26 Thereafter, our group extended this finding to human ETPs, which were shown to display non-T cell potentials, including cDCs, macrophage and NK cell potentials in vitro.27–29 As a whole, these data support the notion that thymic DCs are generated in situ from ETPs that display multipotent potential and commit to the T-cell lineage upon thymus entry. More recently, independent studies have shown that human ETPs can also generate pDCs both in vitro30 and in vivo,31 supporting their intrathymic origin. How non-T cells, and specifically DC, developmental pathways branch off from the main T-cell stream has been largely discussed. However, the recent identification in the human postnatal thymus of myeloid-restricted progenitors with pDC/cDC/macrophage potentials (unpublished results) supports a myeloid origin for thymic DCs.

In conclusion, current views support the notion that at least part of thymic pDCs and cDCs are derived in situ from myeloid-primed progenitors. How specific developmental fates are imposed within the thymic microenvironments at specific niches is still an open question that requires further study.

The involvement of DCs in central and peripheral tolerance was soon considered as one of the functions played by DCs. It was earlier shown that, under steady-state conditions, iDCs may capture apoptotic bodies derived from natural cell turnover and, after migration to the draining LNs, DCs present self-antigens and silenciate autoreactive T cells.32 It is also well documented that the role of thymic DCs in tolerance induction against self-antigens and numerous studies have revealed mechanisms by which subsets of DCs induce or maintain self-tolerance. Tolerogenic DCs are immature, maturation-resistant or alternatively activated DCs that express low levels surface MHC molecules, have a low ratio of costimulatory to inhibitory signals, and an impaired ability to synthesize Th1- or Th17-cell-driving cytokines (such as IL-12p70, IL-23 or TNF-α) compared with fully mature cells. Furthermore, the expression of inhibitory receptors as well as the secretion of anti-inflammatory cytokines (IL-10 and TGF-β) contributes to tolerance induction. In general, the molecular mechanisms by which DCs exhibit a tolerogenic function include: antigen presentation with inappropriated costimulation (induction of anergy), presentation of very low levels of antigen in the absence of other stimuli, production of cytokines such as IL-10, transforming growth factor (TGF-α), tumor necrosis factor (TNF)-α, granulocyte colony-stimulating factor (G-CSF) or retinoic acid (RA), expression of inhibitory receptors such as ILT3, expression of some molecules that induce T cell death (deletion) such as indoleamine 2,3-dioxygenase (IDO), and also express membrane receptors that may instruct T-cell deletion, such as the interaction through the Fas-ligand, or the programmed death ligands (PDL)-1 and -2. Several evidences support that DCs, through the above mentioned mechanisms, contribute to peripheral tolerance by limiting the expansion of autoreactive T cells and by activating Tregs.33,34

Remarkably the role of DCs in maintaining tolerance is fully independent of their maturation state, since immature DCs, semi-mature DCs and also mature DCs have shown to expand antigen-specific Treg cells both in vitro and in vivo.35–38 Hence, authors recommend defining DCs as ‘immunogenic DCs’ or ‘tolerogenic DCs’ according to their functionality.

Some of the mechanisms indicated above have been exploited for the generation of tolerogenic DCs for therapeutic purposes. Of course, the clinical potential of TolDCs in autoimmunity and transplantation relates in part in the capacity of DCs to induce and expand regulatory T lymphocytes (Tregs) that suppress unwanted immunity. A number of murine studies suggest the possibility that potent antigen-specific regulatory T cells can be readily induced in vivo by targeting antigens to DC specific receptors under non-activating conditions. In certain conditions, mature DCs expand CD4+CD25+ Tregs with retained suppressive activity and are capable of inhibiting diabetes development in NOD mice. Tolerogenic properties of immature DCs have already been documented in healthy human volunteers, providing proof of principle that antigen-specific T cell tolerance can be achieved using this approach in humans.39In vitro, several agents, including dexamethasone, vitamin D3, retinoic acid, or IL-10, can be used to render DCs resistant to maturation when DCs are derived from monocytes (Fig. 1). However, so far, only one recent clinical study has taken advantage of their specific tolerogenic properties by utilizing CD40, CD80 and CD86 antisense transfected DCs to treat diabetic patients.40 The application of these tolerogenic DCs was safe and well tolerated. It is not difficult to speculate that tolerogenic DCs will be evaluated in other autoimmune or chronic inflammatory diseases like multiple sclerosis, arthritis or prevent organ transplant rejection. In vitro experiments have already shown promising results in some of these situations. Moreover, GMP grade TolDCs may be already obtained in GMP facilities, thus further supporting their potential use in a therapeutic scenario.

A number of viral pathogens can infect cells of the immune system – particularly DCs and macrophages – reducing or even impairing the host's capacity for immune defense. DCs are highly responsive to stimuli such as microbial infection, inflammation, and tissue damage, and their maturation is further driven by interaction with T cells.41 The manner in which different viruses interact with DC function depends both on the virus and on the DC subset involved.42,43 In contrast, the outcome is less dependent on the capacity of the virus to replicate in the DC. A crucial issue is how antigens are presented to T cells by DCs.

The identification of DCs as key regulators of T and B cell adaptive immunity prompted their introduction as adjuvants in vaccination strategies aimed to induce anti-tumor/virus antigen-specific T cell responses. In general, these strategies consist in the isolation of DCs from patients to be then loaded with tumor/viral antigens followed by their in vitro maturation using different stimuli previous to their transfer into the patient. A more direct strategy involves the “in vivo loading” or “targeting” of antigens to DCs through their surface receptors. Ralph Steinman et al. identified and exploited the use of CD205, a C-type lectin receptor (CLR) which is highly expressed in DCs, for the in vivo antigen targeting. CD205 recycles through late endosomal and lysosomal compartments and mediates antigen presentation very efficiently.71 They demonstrated that proteins could be targeted to DEC to greatly enhance antigen presentation to both CD4+ and CD8+ T cells in vivo. They assessed the potential of antigen targeting to dendritic cells to improve immunity, by the incorporation of ovalbumin protein (OVA) into a monoclonal antibody to the DEC-205 receptor. Targeting of anti-DEC-205:OVA to DCs in the steady state induced 4–7 cycles of T cell division, but the T cells were then deleted and the mice became specifically unresponsive to rechallenge with OVA. In contrast, simultaneous delivery of a DC maturation stimulus via CD40, together with anti-DEC-205:OVA, induced strong immunity with the activation of cytolytic CD8+ T cells producing large amounts of interleukin 2 and interferon gamma. They found that DEC-205 provides an efficient receptor-based mechanism for DCs to process proteins for MHC class I presentation in vivo, leading to tolerance in the steady state and immunity after DC maturation.72 Steinman et al. demonstrated that a DEC-targeted vaccine is able to induce a strong and protective CD4+ T cell response, which will improve vaccine efficacy as a stand-alone approach or with other modalities.73 To extend their finding to humans, they coupled the HIV p24 gag protein to anti-DEC antibodies and assessed cross-presentation for recognition by human CD8+ T cells. They found that DCs and DEC-205 can cross-present several different peptides from a single protein, supporting the testing of anti-DEC-205 fusion mAb as a protein-based vaccine for HIV.74,75 When administered together with synthetic double-stranded RNA, polyriboinosinic:polyribocytidylic acid (poly IC) or its analogue poly ICLC as adjuvant, HIV gag-p24 within anti-DEC-205 mAb was found to be highly immunogenic in mice, rhesus macaques, and in ongoing research, healthy human volunteers (reviewed in76). Steinman et al. made a great effort to develop a safe T-cell-based protein vaccine to target antigens to DCs and exploit their pivotal role in initiating adaptive immunity. Many laboratories have extended this approach using a variety of DC receptors. Thus, mannose receptors, DC-SIGN, LOX1, Dectin-1, Fc receptors, CD11c, CD209, CD40 (reviewed in Tacken et al.,77 CD36,78 CD91,79 or different ligands for toll-like receptors such as CpG,80 flagellin81 or the extra domain A from fibronectin82 among others are being exploited to target Ag to DCs and induce humoral and cellular immune responses. Thus, DC-targeted protein vaccines are a potential new vaccine platform, either alone or in combination with other adjuvants, to induce integrated immune responses against microbial or cancer antigens, with improved ease of manufacturing and clinical use.

The properties of DCs as professional APC have led to their use as vaccines in different therapeutic settings. Although most of these applications have been related to the induction of anti-tumor immunity,83 chronic viral infections may also benefit of the advantages of DCs as immunogens. It has been described that among other mechanisms, some viruses may subvert anti-viral immunity by targeting functional properties of DCs,84–87 a feature common to some viruses causing chronic infections. This can be mediated either by a direct effect upon viral infection of DCs or by indirect mechanisms, such as interaction of viral products with DC receptors. Therefore, these individuals are often characterized by their poor T cell responses, mainly those of the Th1 subtype, which display the most potent anti-viral effects.88 The lack of a correct DC function during the natural course of viral infection has suggested that those strategies based on the administration of fully competent DCs may be able to activate T cells with proper anti-viral activity. Thus, initial preclinical studies carried out in animal models showed that DC vaccination could induce potent anti-viral immunity.89 Indeed, it has been shown that in some cases DC vaccination is the most appropriate strategy to break tolerance against viral antigens.90 Regarding human studies, although DCs obtained directly from blood of patients with chronic infections caused by viruses like HIV, HBV or HCV may present some phenotypic and/or functional impairment, protocols have been developed to differentiate functional monocyte-derived from these individuals. In vitro studies using these cells have shown that they can activate T cell responses against viral antigens, suggesting that this approach may be useful in these individuals.91,92 According to it, several clinical trials using DCs have been carried out. HIV infection is the field with the highest number of antiviral therapeutic vaccination clinical trials based on DCs (reviewed in93). They are characterized by their heterogeneity, both on trial design as well as on their DC manufacturing process. Although most of them use monocyte-derived DCs differentiated with GM-CSF and IL-4 (Fig. 1), they include several types of antigens (virus, recombinant proteins, peptides), as well as different loading strategies and maturation stimuli. These clinical trials have shown that DC vaccination is safe in these individuals, showing only minor local side-effects in some patients, in agreement with results obtained in cancer patients. Concerning immunogenicity, DC vaccination induced T cells of the Th1 profile able to recognize viral antigens. However, only in some cases a virological effect was reported, with some one-log decreases of viral load. Other fields using DC vaccination are chronic infections caused by HBV, HCV, HPV and EBV, although few examples exist in these cases.94–97 As for HIV, these protocols have shown to be safe and no significant toxicities were reported. In a similar manner, anti-viral T cell responses were induced or boosted after vaccination. Finally, poor clinical results have been reported, detectable only in a minority of vaccinated patients. The low number of clinical trials as well as their different design and experimental procedures precludes from obtaining solid conclusions regarding their antiviral effect. Nevertheless, current results are promising and further trials are warranted to ascertain the efficacy of DC vaccines as anti-viral therapies.

It is now clear that an extraordinarily complex network of DC subsets contributes in different ways to the induction of immune responses and the maintenance of tolerance. A major challenge now will be to determine the overall contributions of different DC subsets to these processes, to understand the changes in the transcriptional network that are induced by inflammatory states to ensure the mobilization and activation of each DCs precursor and also to decipher DC interactions with other cells in the immune system.

DCs have long been considered ideal candidates for targeted vaccine approaches, but this promise remains largely unfulfilled. However, DC research has reached a point where clinical application has become a real possibility. So far, the rapid pace of immunology research and the fact that a vast number of human diseases have been proven to exhibit an immune or inflammatory component, suggests that the applications envisaged at the moment will likely be expanded in the future in a way that we cannot yet anticipate.

The authors declare no conflict of interest.