Various lipids of neutral, positive (cationic) or negative (anionic) electrical charges can be assembled into liposomes with consequences for how they interact with APCs. Positively charged formulations containing lipids, such as 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP), or 3ß-[N-(N’,N’-dimethylaminoethane)-carbamoyl]cholesterol (DC-Chol) have been associated with DC-activating effects. Here, both the net positive electric charge, supporting a favorable interaction with the negatively charged cell membrane, as well as APC interaction with the lipid head groups appear to be important factors for inducing DC activation (52). Several mouse and in vitro human studies found an upregulating effect on DC maturation markers and on production of pro-inflammatory cytokines by cationic liposomes (53–55). Due to this assumed adjuvant quality, cationic liposomes are popular candidates for the development of new tumor targeting particulate therapies. Moreover, cationic liposomes appear to be toxic to cells, although a disadvantage when targeting DCs, toxicity is a feature which could be exploited for enhanced tumor cell lysis, thereby also enhancing immunogenicity of cancerous cells (56). Intriguingly, DC-Chol and DOTAP liposomes were found superior in stimulating cross-presentation of OVA to OT-I transgenic CD8+ T cells compared to anionic formulations containing Egg L-α-phosphatidylcholine (EggPC) (57). The authors of that study propose an alkalizing effect of the positively charged formulations on BMDC lysosomes, which leaves OVA more intact for cross-presentation (57). If confirmed by further studies, this would yield another argument for the use of cationic liposomes in anti-tumor targeting. Interestingly, several studies in the context of allergy also employ cationic liposomes, since these formulations proved to be superior in preventing mast cell degranulation and lead to a more efficient reduction in airway eosinophilia and OVA-IgE in allergic mouse models compared to neutrally charged formulations or empty antigen (58, 59).

In the tolerogenic field, liposomes containing the negatively charged lipid, phosphatidylserine (PS) gained considerable attention (22, 60, 61). The most prominent theory for a tolerogenic adjuvant effect of PS containing particles puts forth the notion that such particles resemble apoptotic bodies, thereby silencing DC maturation upon their encounter. Utilizing PS liposomes added to mouse DCs in vitro, Shi and colleagues demonstrated that the DCs were resistant to maturation, produced less pro-inflammatory cytokines, acquired the capacity to suppress CD4+ T cell proliferation, as well as to induce PD-1 surface expression on T cells (61). Similar effects were confirmed in a human in vitro study, in the context of Type-I diabetes, where patient DCs pulsed with PS-liposomes retained a tolerogenic profile, and suppressed autologous T cell proliferation (62). Interestingly, in a recent study by Benne and colleagues it was not PS containing liposomes, but liposomes incorporating the anionic lipid 1,2-distearoyl-sn-glycero-3-phosphoglycerol (DSPG) that induced antigen specific Foxp3+ T cells upon in vivo injection in mice (55), the mechanisms of which are yet elusive.

A characteristic that is often altered when changing lipid composition of a formulation is rigidity (22). APCs envelope rigid particles easier than flexible ones leading to more efficient uptake (63–65). In line with these observations Benne et al. confirmed enhanced uptake of more rigid variants of the DSPG containing tolerogenic liposomes mentioned above, where injection of more rigid formulations in mice also correlated with stronger Treg responses (66). Similarly, more solid gel-phase pegylated 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC-PEG) or 1,2-dio- leoyl-sn-glycero-3-phosphocholine (DOPC-PEG) liposomes were taken up better by bone marrow-derived DCs (MDCs) and activated these cells more than their fluid-phase counterparts. Of note, this study used Cholera toxin antigen loaded formulations, and empty formulations did not have an adjuvant effect on BMDCs (67).

Multiple studies emphasize the effect of particle size on uptake mechanisms by APCs, which in turn may influence how liposomal cargo is processed and presented to T cells (22, 68, 69). Similar to how viruses enter cells, particles smaller than 100nm are taken up efficiently by clathrin mediated endocytosis, whereas particles larger than 200nm are phagocytosed or internalized by macropinocytosis (68). This may influence intracellular routing of the liposomal cargo. Cargo escaping the lysosomal route can be more available for cross-presentation on MHC-I, a desirable outcome for DC-mediated cancer therapies but also for tolerizing CD8+ memory T cells employing liposomes with self-antigen (70). Cargo following phagocytosis and the endo-lysosomal route will be preferentially processed towards MHC-II presentation, priming for interaction with CD4+ T cells. Moreover, some studies suggest that particle size has a specific influence on T cell polarization. Nanobeads smaller than 100nm were shown to elicit stronger IFN-γ responses in mice leading to superior T_H1 immunity compared to beads larger than 100nm (71). As already mentioned small particles may be taken up by DCs similarly to viruses mimicking anti-viral immunity (71). Biodegradable polymer particles in a size range of 1-5µm, on the other hand, were shown to adhere to the DC membrane and offload their cargo there, leading to a T_H2 response (72, 73). However, conclusions about liposomes seem to point in the opposite direction: liposomes smaller than 200nm were reported to induce T_H2 immunity whereas larger ones primed towards T_H1 (74, 75). These observed differences may be explained by uptake and intracellular processing differences between solid particles (such as nanobeads and various polymers) and semi-solid or fluid liposomes. In contrast to solid particles which are taken up by active processes, liposomes can also fuse with the cell membrane and offload cargo directly into the cytoplasm (76). This difference in interaction with DCs can lead to different intracellular processing and presentation of cargo to T cells.

In addition to differences in cellular uptake and processing, size of particles influences bioavailability and biodistribution upon in vivo injection. Only particles smaller than 200nm seem to drain to lymph nodes where they can be directly processed by lymph node resident DCs for early T cell activation (22). Larger particles, in contrast, remain at the injection site until phagocytosed by DCs that can migrate to the lymph node, possibly leading to a less vigorous T cell response.

Although less studied, particle shape seems to have an immune modulating effect on APCs as well, where rod-shaped structures (nanorods) were reported as more pro-inflammatory compared to spherical particles (22, 77). However, this feature is less relevant for liposomes as they do not belong to the group of solid nanoparticles.

Despite demonstrated evidence on adjuvant characteristics of certain nanoparticles, a large body of evidence with human cells is lacking. Often several characteristics of a particle are altered at once, making it difficult to discern which characteristic is responsible for observed differences in immunogenicity (22). As visible from the studies discussed in this section, there is also a great need for standardization of different particles or liposomal formulations in order to facilitate a valid comparison between studies. Finally, the therapeutic content loaded in a particle, such as tolerogenic or immune activating adjuvants together with any targeting molecules, may overrule the immune modulatory effect of empty particles (78).

Tumors display a wide variety of (abnormally expressed) tumor-associated antigens (TAA), and TAAs such as MART-1, MUC1, WT1, gp100 and the MAGE-A antigens, have been tested in various vaccines in clinical trials (79–82). Using TAA-loaded nanoparticles in different clinical trials have, so far, resulted in mixed responses (83, 84). Since TAA are expressed on both healthy and transformed cells, it is possible that T cells specific for these antigens are deleted during the negative selection in the thymus, which therefore leads to the observed suboptimal anti-tumor responses in many TAA (85, 86). The potential off-target effects induced by targeting TAA make tumor-specific antigens (TSA) an interesting alternative. These antigens are not expressed on healthy cells, and therefore also have limited tolerance related complications.

TSA include mutated neoantigens, but also antigens from endogenous origin. In particular, because ~12% of human cancers are caused by viruses, the foreign (viral) antigens expressed on the transformed cells are highly immunogenic (87, 88). The VGX-3100 vaccine, targeting the HPV proteins E6 and E7, the main oncoviral antigens of the cancer caused by the virus, is based on a DNA vaccine of the mentioned antigen, and is currently tested in a phase III clinical trial (NCT03185013) (89). Parallel to this non-liposomal vaccine liposomal (archaesomes) delivery systems for DNA encoding the HPV antigens are being tested, and show the induction of a potent anti-tumor immune response in an in vivo cancer model (90). Another class of TSA arise from non-synonymous DNA mutations, and are therefore called neoantigens (91). These antigens are highly immunogenic and therefore highly sought after targets for therapeutic vaccines (92). Since the neoantigens are personal to the patient, they will need to be identified per patient through genomic comparison of tumor and normal tissue (93). Recent personalized clinical trials with vaccines targeting neoantigens feature high immune activation and overall promising results (94, 95). Nanoparticles are already used to deliver the neoantigens (e.g. RNA, DNA or peptides) to DCs in vitro, in vivo and in clinical trials (85, 96, 97). Especially for (m)RNA vaccines, nanoparticles such as liposomes are greatly beneficial since they protect the payload against degradation (98). Accordingly, a recent phase I clinical trial with patient personalized tumor mRNA-loaded nanoparticles showed high tolerability, and the interim results from another recent phase I clinical trial showed encouraging clinical responses (NCT03897881) (99). Exploiting the same pharmacological advantage, the recently successful mRNA based SARS-CoV-2 vaccines have a nanoparticle delivery system (100, 101).

PRRs researched for targeting of DCs in context of cancer therapy include CLRs, the Siglec receptor Siglec-1, and Fc receptors (FcR). CLRs recognize carbohydrate ligands, which makes them important sensors of differently glycosylated PAMPS (102). Some of these receptors are expressed on a broad class of APCs, whereas others are DC or even DC-subset specific. Thus, targeting of these receptors by adding glycan moieties or CLR antibodies to liposomes can be used for reaching APCs with a varying spectrum of cell specificity. DC-SIGN is a CLR expressed on different subsets of APCs, including moDCs, CD14+ dermal DCs, subsets of macrophages and DCs at mucosal sites (103). Targeting liposomes to DC-SIGN via its natural glycan ligand Lewis Y showed increased CD8+ T cell responses in vitro and ex-vivo ( 104). Also, targeting the DC-SIGN receptor with antibodies conjugated to nanostructures, leads to increased immune responses (105). Unfortunately, a clinical trial with a DC-SIGN targeting nanoparticle vaccine Lipovaxin-MM has not resulted in immunogenic anti-tumor responses (106). The vaccine’s antigens (e.g. gp100 and MART-1) were derived from plasma membrane vesicles from a human melanoma cell line, and modified with a liposomal mixture, also containing IFN-γ. A DC-SIGN targeting antibody was also incorporated in the membrane, allowing for targeting to various APCs. While the vaccine was well tolerated, significant immunogenicity of the vaccine was not detected. Similarly to DC-SIGN, the mannose receptor (MR) is also expressed on various APCs, such as macrophages and moDCs. Targeting of MR has already been evaluated in a clinical trial, proving that the administration of the MR-targeting vaccine together with local application of TLR agonists, induces significant humoral and cellular immune responses (107). When mannosylated nanoparticles were used to target the MR, it led to high effector T cell responses and reduced tumor growth in vivo ( 108). Immunization of mice with liposomes made of mannose-mimicking ligands loaded with DNA encoding for MART-1 allowed for efficient transfection of CD11c+ DCs, inducing long lasting melanoma specific prophylactic CTL responses (109). Therapeutic vaccination with these liposomes resulted in delayed tumor growth in mice. In contrast to the previously discussed CLRS, DEC-205 appears more DC-specific with expression demonstrated on cDC1s, cDC2s and moDCs. When DCs were targeted with PLGA nanoparticles coupled to monoclonal anti-DEC205 antibodies, the treatment lead to enhanced internalization, cross-presentation and CD8+ T cell activation (110). Similar results were observed by another group, showing that targeting of nanoparticles to DEC-205, CD40 or CD11c improved priming of cytotoxic CD8+ T cells over untargeted nanoparticles (111). Targeting CLEC9A, a CLR with advantageously restricted expression on cDC1 DCs, also elicited anti-tumor responses in multiple studies, coherent with the potent (cross-)presenting function of these DCs (112, 113). In an additional study, targeting PLGA nanoparticles loaded with the TAA Trp2 and gp100 to CLEC9A expressing DCs via antibodies resulted in strong therapeutic anti-tumor responses in vivo, but also induced in vitro expansion of NKT and CD8+ T cells specific for melanoma in PBMCs from both healthy donors and melanoma patients (114).

Other groups of receptors that have been used for targeting APC in the context of cancer are Siglecs and FcR. While most Siglecs confer tolerogenic responses, the Siglec-1 (or CD169) receptor, expressed on splenic macrophages is of special interest for anti-tumor responses. These cells transferred antigen to cross-presenting cDC1s when targeted with liposomes coated with the Siglec-1 ligand GM3, conferring beneficial anti-tumor CD8+ T cell responses (115). FcR, which bind to the constant domains of antibodies can be targeted by coating liposomes with antibodies. Indeed, IgG coated liposomes bearing the OVA antigen prevented development of OVA-expressing lymphoma, in contrast to the liposomes without IgG coating (116). Also, specifically the Fc fragment of an antibody can be used on the outside of a nanoparticle for targeting purposes, which induced increased cellular and humoral immune responses in mice when a cancer peptide was included in the nanoparticle (117).

Targeting one receptor may in several cases induce either immunity or tolerance, depending on the vaccine formulation and microenvironment. For instance, targeting DC-SIGN can display T_H2 polarizing effects, combined with inhibition of T_H1/T_H17, when targeted with natural ligands (118). Accordingly, depending on the adjuvants used, DEC-205 targeting is used for both tolerogenic and immunogenic purposes (119). Therefore, when targeting DCs via specific receptors, it is not only important to target the appropriate receptor, but also to provide efficient co-stimulatory adjuvants to properly skew the immune response (51).

TLRs are a well-known class of PRRs, and some of these are targeted for induction of antitumor responses (120). Molecules targeting TLR4 are LPS structures derived from bacteria, but as adjuvant the less toxic variant of LPS, monophosphoryl lipid A (MPLA) is used. The incorporation of MPLA in PLGA nanoparticles increases T_H1 and pro-inflammatory responses in comparison to the non-encapsulated administration of MPLA (121). Agonists for TLR7/8 are viral ssRNA and synthetic compounds like R848. Triggering TLR7 for cancer therapy with Imiquimod has already been approved by the FDA, and clinical trials with the TLR7/8 stimulating adjuvant R848 have been conducted (120, 122). The prophylactic and therapeutic effect of OVA mRNA-loaded nanoparticles as vaccine against OVA expressing lymphoma in an in vivo mouse model is increased when R848 is added, and results in increased TAA presentation and antigen specific CD8+ T cell populations (123). The TLR9 stimulating ligand CpG has also been evaluated in clinical trials, with positive outcomes (124). When CpG is incorporated into a nanoparticle, its efficacy in terms of (delayed) tumor growth was superior in comparison to the soluble form of CpG, again highlighting the potential of nanoparticles for antigen and adjuvant delivery (125). TLR3, recognizing dsRNA, is another PRR receptor that is targeted for pro-inflammatory responses (126). The synthetic analogue of TLR3, Poly (I:C), is being evaluated as adjuvant in (pre-) clinical trials, and elicits favorable effects (94). Thereby, encapsulation of the TLR3 adjuvant in a nanoparticle further potentiates the inflammatory immune responses (127). Especially when Poly (I:C) is combined with the TLR9 stimulant CpG in a nanoparticle, it induces protective and therapeutic immune responses in in vivo models (128, 129). Other combinations of TLR ligands in nanoparticles are also used, with established synergistic effects (120, 130). Hence, it is becoming clear that incorporating multiple TLR stimuli in liposomes will be a promising adjuvant strategy for immunogenic vaccines.

An adjuvant that works directly via presentation on DCs is α-galactosylceramide (αGC), a potent immune activating glycolipid which when presented on DCs, initially activates iNKTs (131). In turn, these iNKTs activate other NK-, CD8+-T, B-cells and DCs, via increased cytokine production (132). Since αGC is a glycolipid, it also allows for easy incorporation in- for example- liposomal nanoparticles. Incorporated in liposomes, αGC leads to the increase in CD8+ T cell responses in vitro ( 104). Accordingly, αGC incorporated in liposomes is able to reduce the outgrowth of lung melanoma metastasis in vivo ( 133).

Another way to increase vaccine efficiency is by increasing the numbers, and immunological state, of immune cells in the site of vaccination. The adjuvant (MF59) used in the seasonal influenza vaccine, based on an oil-in-water nano-emulsion, increases the number of APCs and creates an immunogenic microenvironment (134). Enhancement of APC numbers also improves trafficking of antigens to the draining lymph nodes via leukocytes, favoring a stronger immunogenic response. Oil-in-water nano-emulsions are able to encapsulate both antigens and antibodies, for specific targeting to for instance the CLEC9A receptor, thereby becoming self-adjuvating delivery systems for DC vaccination (135).

Repeated exposure to antigen has shown tolerance inducing effects. AIT exploits this very concept with curative success demonstrated for various pollen and venom allergies (3). However, treatment efficacy for other allergies, such as food allergies, could be increased, together with reduction of risk of anaphylaxis by incorporating allergens in nanoparticles. A clinical study with subcutaneous injection of liposomes incorporating house-dust mite extract was already carried out as early as 2002 (136). Unfortunately, this study did not compare treatment with soluble allergen and no clear line of clinical studies followed. Even though the validity of allergic mouse models is somewhat questionable, several recent studies demonstrate therapeutic proof-of concept. In a mouse model of pollen allergy to the weedy plant Chenopodium album, subcutaneous injection of allergen incorporating protamine-DNA liposomes shifted a predominant T_H2 response to the allergen in a T_H1 direction, with decrease of IgE, increase in IgG2a and IFN-γ production specific to allergen (137). Chaisri et al. tested effects of intranasal vaccination with liposomes incorporating Derp1 and Derp2 separately, or in combination (138). Interestingly, even though all formulations reduced T_H2 immune reactions, only the liposomes incorporating single allergens lead to expression of tolerogenic cytokine genes TGF-β, IL-35 and IL-10 in mouse lung cells. In an OVA mouse model of allergy, sublingual treatment with OVA incorporating liposomes preceding allergen challenge was superior to treatment with free OVA (59). Unfortunately, none of the above studies examined airway hyperresponsiveness as a read-out, which could make the conclusions about treatment efficacy stronger.

In contrast to allergies where the antigen is known and clearly defined, autoimmune conditions pose the difficult challenge of unknown causative auto-antigens or foreign antigens that may trigger disease. In rheumatoid arthritis (RA), anti-citrulline antibodies (ACPA) appear in blood before disease onset and can be very specifically linked to RA pathogenesis (139). This knowledge stimulated a quest for putative, citrullinated antigens on cartilaginous surfaces capable of stimulating auto-reactive T cells. Indeed, using a panel of HLA-DRB1*04:01 tetramers, James and colleagues confirmed an increased presence of citrullinated antigen specific T cells in peripheral blood of RA patients compared to healthy subjects (140). Of note, Benham and colleagues performed immunotherapy using ex vivo RA patient-derived moDCs with an NF-κB inhibitor and a mix of citrullinated antigens (141). This (uncontrolled) treatment strategy proved safe and showed some signs of efficacy such as a decrease in effector T cells and improvement in clinical RA scores. As early as 2009, the same research group established that citrullinated antigens loaded in EggPC liposomes can efficiently be used to induce antigen-specific Foxp3+ Tregs in mice, an effect that was specifically mediated by DCs (142).

Several CLRs and Siglecs are under scrutiny for targeting, as an attempt to specify tolerogenic in vivo DC therapies. For an extensive review on this subject we refer the reader to a very recent piece published by our colleagues (21). The CLR DEC-205 is highly expressed on the cross-presenting cDC1 DC subset (as well as on cDC2s and moDCs) and is therefore a widely studied receptor for nanoparticle DC targeting, for example via antibodies (143). Targeting the DEC-205 receptor via antibodies without providing maturation stimuli can lead to specific induction of T-cell anergy as well as increased T cells suppression (119). OVA-loaded PLGA nanoparticles that were conjugated with antibodies to DEC-205 induced IL-10 production in DCs and subsequently T cells, of which the levels were dependent on the amount of antibodies on the nanoparticles (144). However, the T_H1 priming of DCs targeted with these nanoparticles was not impeded.

The F4/80 receptor, expressed on macrophages and DCs, shows promise in inducing tolerance in a nonobese diabetic mouse model. The progression of diabetes in the in vivo model was prevented by vaccinating mice with liposomes coated with anti-F4/80 antibodies and a disease specific short peptide joined with a TLR-2 ligand (145).

Modification of antigens with the carbohydrate ligands (sialic acid) of Siglecs expressed by DCs results in the induction of antigen specific Tregs and alleviation of allergic symptoms in mice (146, 147). Other studies focus on nanoparticles targeting Siglecs to induce tolerance in B-cells or other immune cells, but to date, no studies have been reported using nanoparticles that target to Siglecs on DCs (148, 149).

Similarly to allergies, there is pre-clinical precedent for the use of nanoparticles for developing new treatment modalities against RA and other autoimmune diseases. However, studies with a targeting component towards DCs are not abundant in literature, highlighting a knowledge gap that needs to be addressed.

In order to prevent adverse reactions to allergens or a worsening of autoimmunity it is highly likely that co-delivery of tolerance inducing compounds together with disease relevant antigen will be a necessary element of successful in vivo targeting therapies. Indeed, multiple studies observed immunogenic reactions to nanoparticles delivering antigen only (69, 150–152). This may also apply to AIT utilizing nanocarriers. For example, in a mouse study of cockroach allergy, only allergen encapsulated in liposomes together with a tolerogenic adjuvant induced increased transcription of IL-10, TGF-β and IL-35 as well as IDO1 (153).

The choice of adjuvants with the strongest immune dampening effect may be critical for successful DC-mediated tolerance in vivo. The vitamins D and A have been extensively studied in the tolerogenic context. 1,25alfa-dehydroxycalciferol, or vitamin D3 (VD3) appears to be the most potent immunosuppressant of all forms of vitamin D (154, 155). Both mouse and human studies have demonstrated that VD3 priming of immature and mature DCs results in a tolerogenic phenotype with induction of co-inhibitory receptors, reduced IL-12 production and induced IL-10 secretion (156–160). The tolerogenic effects of VD3 have also been shown in several skin-derived DC subsets where priming of in vitro cultured LCs, CD1a+ DDCs or skin-derived DCs with VD3 resulted in the outgrowth of distinct Treg phenotypes (161, 162). Most importantly, VD3-raised DCs show tolerogenic stability in face of repeated rechallenge with pro-inflammatory stimuli, making VD3 a robust DC-tolerizing candidate in an already inflamed environment (163). In line with that, several recent studies in mice demonstrated that nanoparticles loaded with VD3 and OVA induced tolerogenic DCs with in vitro and in vivo suppressive capacity of OVA-specific T cells (164, 165). Additionally, subcutaneous injection of VD3-loaded particles resulted in effective targeting of PD-L1 high draining lymph node DCs, resulting in amelioration of a RA disease model (165).

In contrast to the well-established tolerogenic role of VD3, there is still considerable debate on whether the active form of Vitamin A, all-trans retinoic acid (atRA) has pro- or anti-inflammatory effects (166). Similarly to murine mucosal CD103+ DCs (167), human moDCs raised with atRA, induce the development of IL-10 producing Tr1 in co-culture (41). Building on the anti-inflammatory potential of atRA, the compound was recently incorporated in PLGA particles together with atherosclerosis autoantigen and improved atherosclerotic lesions in vivo ( 168). In a further recent study, atRA was encapsulated together with another anti-inflammatory adjuvant, triptolide, in galactose-containing nanoparticles (78). In vivo effects in mice consisted of reduced infiltration of these sites by T cells and pro-inflammatory macrophages, together with reduced expression of T_H1-T_H17 polarizing cytokines in inflamed tissue extracts. However, atRA also supports induction of T_H1 and T_H2 responses upon inflammatory stimulation, serving as one explanation of contradictory pro- and anti-inflammatory effects observed (169, 170). Thus, loading of atRA in nanoparticle platforms for tolerogenic purposes may be an interesting option, but with a sidenote of caution.

A central transcription factor downstream of activating signals delivered to DCs is NF-κB, which makes it one of the key targets for immune modulation. Corticosteroids and glucocorticoids are well-known to exert their immunosuppressive effects via NF-κB inhibition (171, 172). Treatment of moDCs with the corticosteroid dexamethasone leads to an immature phenotype with loss of IL-12 secretion and high IL-10 secretion. Similarly to VD3 treated DCs, dexamethasone DCs seem robustly maturation resistant and capable of inducing IL-10 producing Tregs, although these Tregs exert suppression in a non-antigen specific manner (173). To prevent systemic immune suppression, loading of corticosteroids into nanoparticles and targeting the particles only to cells relevant for inflammation is a treatment approach considered in RA and other autoimmune conditions, showing promising treatment efficacy in several mouse studies (174). In addition to these rather a specific inhibitors of NF-κB signaling, also several highly specific inhibitors of (non)canonical signaling have been investigated in human moDCs and demonstrated to potently reduce T cell responses (175). Recently, also VD3 was confirmed to downmodulate NF-κB signaling in human moDCs matured with LPS, providing an elegant bridge between VD3 and NF-κB inhibition as adjuvants (176).

Indoleamine 2,3-dioxygenase (IDO), the enzyme responsible for breaking down the essential amino acid tryptophan into kynurenine can be produced by DCs, which leads to decreased T cell proliferation, induction of Tregs and an anergic T cell phenotype in co-culture experiments (177–179). IDO production in the gut is intricately intertwined with the gut microbiota, and exerts tolerogenic effects on CD103+ DCs as well as dampening tissue T_H1/T_H17 responses (180–182). Several molecules may induce IDO production by DCs, however, some of these, such as type I interferons or IFN-γ are also pro-inflammatory cytokines, hindering their use in tolerance inducing therapies (183). Induction of IDO has been described upon CD40 ligation of moDCs, where the enzyme was induced by non-canonical NF-κB signaling (36). After treating mouse pDCs with TGF-β, IDO was activated as a downstream signaling molecule, leading to inhibitory signaling and the activation of the non-canonical NF-κB pathway, further strengthening IDO expression in a self-feeding loop (184). Further compounds that were demonstrated to induce IDO are soluble CTLA-4, the TLR-agonists LPS and CpG, together with DNA agonists of stimulator of interferon genes (STING) (183). As treatment with TLR-agonists carries pro-inflammatory risks, soluble CTLA-4 and STING agonists seem better suited candidates for DC-treatment. Indeed, nanoparticles constituted of CpG free pDNA were demonstrated to induce IDO via the STING pathway, leading to amelioration in experimental autoimmune encephalomyelitis (183). Moreover, this therapeutic effect could be strengthened when blocking downstream metabolites of IDO additionally to DNA nanoparticle treatment.

In addition to the above compounds discussed in detail, several other materials are tested for their tolerogenic potential, such as various parasite-derived antigens, plant-derived adjuvants or compounds already well-known in the clinic, like the m-TOR inhibitor rapamycin (185). In fact, just like VD3 and dexamethasone, rapamycin was demonstrated to induce robustly tolerogenic, clinical grade DCs (186). Although rapamycin loading into several forms of nanoparticles is actively tested for better delivery to cancer cells or as an inflammation dampening adjuvant, targeting to DCs has not been in the focus of research thus far (187).