2.1 DC dysfunction influenced by cytokines

Cancer cells, myeloid-derived suppressor cells, tumor-associated macrophages (TAMs), regulatory T cell (T_reg) and other cells produce various interleukins (ILs) (such as IL-6, IL-10, and IL-13), indoleamine 2,3-dioxygenase (IDO), growth factors (vascular endothelial growth factor [VEGF], granulocyte-macrophage colony-stimulating factor [GM-CSF]), and other factors (ganglioside and prostaglandin E2 [PGE2]) in the TME may influence or even damage the function of DCs (Figure 1, Pathway D) [22-24].

In some immunological reactions, such as inflammation stimulated by lipopolysaccharide, T_reg cells secrete a large amount of the cytokine IL-10 in the TME [25], which not only reduces the expression of B cell lymphoma 2 (Bcl-2) in DCs but also inhibits the differentiation of monocytes into DCs [26]. In addition to IL-10, the abnormal expression of IL-6 by tumor-infiltrating DCs affects the differentiation of T helper cells to type 2 T helper cells, which impairs immune responses against pathogens [27]. Tumor-derived versican is a chondroitin sulfate proteoglycan that binds to toll-like receptor 2 to enhance the autocrine function of IL-6 and IL-10, and the upregulation of their respective cell surface receptors can lead to DC dysfunction, such as IL-10-producing conventional dendritic cells (cDC) [28-31]. Moreover, massive abnormally secreted bioactive substances in the TME, such as GM-CSF, gangliosides, and interferon-γ (IFN-γ), dysregulated DC maturation, inhibited the antigen presentation ability of DCs, and subsequently suppressed the immune response [32-34]. VEGF, another growth factor, inhibited DC differentiation and maturation [35], and treatment with anti-VEGF antibodies alleviated DC dysfunction in the TME [22, 36]. In addition, PGE2, an inflammatory mediator that accumulates in the TME, inhibited the antigen presentation function of DCs and attenuated the activation of natural killer (NK) cells, cytotoxic T lymphocytes (CTL), and other immune cells [37, 38]. Cytokines, as crucial regulators of the immune system, also play critical roles in the function of DCs, especially negative regulatory cytokines that induce DC dysfunction.

2.2 DC dysfunction mediated by hypoxic environment

As tumors demonstrate rapid growth, they require an increased amount of nutrients and oxygen; however, the blood vessels within the tumor failed to transport sufficient nutrients to the inner region of the tumor, leading to the development of a hypoxic environment and the production of numerous metabolic byproducts within the tumor cells [39-41]. Immune cell function was limited by several factors in the TME, such as oxygen, pH, glycogen, and metabolic byproducts (Figure 1, Pathway C) [14, 24, 39, 40, 42, 43].

Hypoxia-inducible factor 1 (HIF-1), a heterodimeric DNA-binding transcription factor, negatively regulates DC function under hypoxic conditions [39, 44]. In the hypoxic environment, chemokine receptors (CC-chemokine receptor 2/3/5, CX3C chemokine receptor 1, C5a receptor gene 1, and formyl peptide receptor 3) were up-regulated to polarize immature DCs (iDCs) into a migrating phenotype [45], which was mediated by the hypoxia-inducible factor-α (HIF-α) to enhance the migration ability of hypoxic iDCs via phosphoinositide 3-kinase/ protein kinase B signaling [46-48]. Hypoxic DCs down-regulated the expression of RhoA/Ezrin-Radixin-Moiesin and lectin receptor cluster of differentiation 206 (CD206), which impaired the ability of DCs to capture antigens [46]. Moreover, the secretion of VEGF and IL-10 mediated by HIF-αinhibited DC function [48-50].

Moreover, the hypoxic TME continuously activated the endoplasmic reticulum (ER) stress factor X-box binding protein l due to mutations in oncogenes and tumor metabolism [45, 51]. This induced the ER stress response of the unfolded protein response, and ultimately damaged the antigen processing and presentation functions of DCs [52]. Under normal conditions, low levels of reactive oxygen species (ROS) mediated DNA oxidation enhanced the immune recognition of DCs [53]. However, the secretion of nicotinamide adenine dinucleotide phosphate oxidase in DCs induced by high ROS level contributed to proton depletion in the phagosome and eventually increased the intracellular pH, which inhibited antigen presentation by DCs [54, 55]. Therefore, further studies on the regulatory mechanisms of ROS and hypoxia in the TME are warranted to strategize effective therapies [45].

2.3 Immune tolerance of the DCs mediated by TEXs

TEXs are extracellular membrane vesicles secreted by tumors [56] that contain various soluble components such as micro ribonucleic acids, proteins, enzymes, lipids, and cytokines [56, 57]. Hypoxia increased exosome secretion from epithelial ovarian cancer (EOC) cells. These exosomes were engulfed by unpolarized macrophages, and the TAM phenotype was adjusted to promote the progression of EOC [58]. Moreover, TEXs secreted during ferroptosis of cancer cells inhibited the maturation of DCs and impaired antigen cross-presentation [59]. In addition, fatty acid-carrying tumor-derived exosomes induced dysfunctional DCs to facilitate immune evasion, which was reversed by PPARα inhibition [18]. Based on the mechanisms of TEXs, cancer-derived TEXs as antigens in DC vaccines may exert more potential effects on DC maturation and cross-presentation of MHC molecules than those of conventional tumor-associated antigen (TAA) lysates [60]. In summary, although TEXs exerted a negative effect on DCs, they were potential targets used not only as drug delivery materials but also as anti-tumor DC vaccines [61].

2.4 Suppression of the immune system by other metabolites in the TME

As previously described, the hypoxic environment in the TME promoted the production of large amount of lactic acid in the tumor cells [62], which induced a low pH in the TME. Low pH further attenuated antigen presentation by DCs and inhibited their maturation [62]. Moreover, hypoxia led to glycogen accumulation in the tumor cells and altered the metabolism of tumor-infiltrating lymphocytes and other immune cells, which eventually promoted the growth of the tumor cells [63]. Aberrant cyclooxygenase activity associated with hypoxia in tumors induced the production of prostaglandin E2, which mediated cDC1 dysfunction through the cDC loss of interferon regulatory factor 8 [64]. In addition, tumor-derived α-fetoprotein downregulated the transcriptional expression of the enzymes, resulting in lipid disorders and inhibiting DC maturation and development [65]. Abnormal lipid metabolism in the TME resulted in ER stress and functional damage to DCs [52]. Adenosine produced in the TME also affected the immune function of DCs [66]. Abnormal regulation of N6-methyadenosine (m6A) RNA levels in the TME partly impaired antigen presentation by DCs via the YT521-B homology (YTH) domain-containing family protein 1 (YTHDF1) signaling pathway [67]. Therefore, the metabolites produced in the TME mediated DC dysfunction and promoted tumor growth.

2.5 DC dysfunction mediated by co-inhibitory molecules

A few molecules displayed co-inhibitory functions in DCs (Figure 1, Pathway D). Cytotoxic T-lymphocyte antigen-4 (CTLA-4) expressed on tumors blocked the co-stimulatory signal of DCs to T cells by binding to CD28, which competed with CD86 expressed on DCs [14, 68]. T cell immunoglobulin and mucin domain 3 expressed on tumor-infiltrating DCs inhibited the tumor nucleic acid monitoring system to prevent DCs from cleaning abnormal nucleic acids and impeded DC-mediated immunogenic cell death [69, 70]. CD73, an immunoinhibitory protein that plays a key role in tumor growth and metastasis, suppressed the anti-tumor immunity of DCs [71]. Blocking CD73 enhanced the recruitment of cDC1 to tumor sites after radiotherapy and improved anti-tumor efficiency [72]. Therefore, analyzing these co-inhibitory molecules in DC dysfunction may contribute to the discovery of novel immunotherapies targeting DCs.

These findings indicate that DC dysfunction is induced by several external factors, such as cytokines, hypoxia, TEXs, tumor metabolites, and co-inhibitory molecules. Therefore, focusing on these factors may facilitate the development of targeted therapies to overcome DC dysfunction. Besides the external factors, intracellular signal transduction plays an important role in DC dysfunction, which needs to be considered.

3.1 Wnt/β-catenin signaling and relevant therapies

In abnormal cDC1, the dysregulated Wnt/β-catenin signaling pathway has been identified as a key factor promoting tumor progression [73]. The over-activated Wnt signal in the TME elevated β-catenin levels, which in turn reduced the downstream gene transcriptional level of chemokines (or ligands), which subsequently prevented intra-tumoral migration of CD103⁺ DCs and impaired the infiltration of T cells into the TME [74, 75]. This was also confirmed by a study on anti-programmed cell death 1 (anti-PD-1) immunotherapy, in which elevated β-catenin levels impeded the recruitment of CD103⁺ DCs and triggered the tolerance of PD-1 therapy in mice with hepatocellular carcinoma [76]. These studies suggested that DC dysfunction was mediated by Wnt/β-catenin signaling [77, 78]. Targeting Wnt/β-catenin signaling may be a promising approach to overcome DC dysfunction and improve the efficacy of current cancer immunotherapies [73, 79], such as therapies through PORCN (porcupine O-acyltransferase) inhibitors, Dickkopf-1 (DKK1) antibodies, and cAMP response element-binding protein-binding protein/β-catenin (CBP/β-catenin) inhibitors.

3.2 YTHDF1 drives the tumor evasion from DC surveillance

YTHDF1 regulated the translation of cancer-associated genes via m6A methylation and impaired the antigen presentation ability of DCs, thereby affecting the cytotoxic CD8⁺ T cells’ function [96]. Furthermore, YTHDF1 promoted the synthesis of lysosomal cathepsin in DCs and prevented tumor antigens from being processed and loaded onto MHC I molecules by DCs [97]. When treated with PD-1 inhibitors in YTHDF1 knockout mice, the therapeutic efficacy of the PD-L1 blockade was enhanced [97]. This underscores the potential synergistic effects of combining PD-L1 blockade with YTHDF1 inhibitors. Therefore, YTHDF1 plays a crucial role in the cross-presentation function of DCs, and therapies targeting YTHDF1 may contribute to the activity of DCs against cancer. However, because YTHDF1 is also expressed in normal cells, further verification is needed to avoid toxicity and autoimmunity caused by off-target effects [97, 98].

Despite the substantial role of Wnt/β-catenin and YTHDF1 signaling pathways in DC dysfunction, therapeutic strategies targeting these pathways have not shown notable progress. Considerable efforts are needed to explore the mechanism of DC dysfunction in basic research and develop novel and more efficient therapies that leverage these two signaling pathways.

4.1 DC-based adjuvant

The early development of DC therapy was most associated with some immune adjuvants inducing inflammation factors (Figure 2A), such as aluminum salts, oils, cytokines (IFN-γ, IL-12, and GM-CSF), and synthetic compounds such as polyinosinic-polycytidylic acid (poly [I:C]) [100, 101]. The administration of these adjuvants boosted the body's immune reaction and exerted a specific anti-tumor effect [14, 101]. Among them, TLR agonists are highly promising adjuvants for vaccines against complex and life-threatening diseases such as malaria, acquired immunodeficiency syndrome, and cancer. More than 50 clinical trials have been conducted using TLR agonists such as TLR3, TLR4, TLR 7, TLR 8, and TLR 9 agonists (Supplementary Table S1). Most were in phase I and II clinical trials (94.91%) (Figure 3A). These agonists elicited a “danger” signal to induce an effective immune response, which may offer durable protection [102]. The results are summarized in Table 4. Most studies reported that TLR agonists were safe, well tolerated in patients and induced an immune response. The TLR-3 receptor agonist (poly [I:C]), the most commonly used anti-tumor adjuvant, enhanced the function of CD8⁺ and CD4⁺ T cells [103]. The combination treatment of poly (I:C) and chimeric antigen receptor T (CAR-T) cell therapy enhanced the activity of CAR-T and increased the secretion of IFN-γ [104]. TLR-3 receptor agonists prolonged the survival of glioblastoma patients [105]. Another synthetic-specific TLR-3 adjuvant, ARNAX, has attracted considerable attention. The TLR-3 adjuvant activated only the TLR-3 signal, which reduced systemic inflammatory side effects without activating melanoma differentiation-associated protein 5 (MDA5) [106]. In addition, ARNAX mitigated immunosuppression in the TME. The combination of anti-PD-1/PD-L1 antibodies with ARNAX further improved the therapeutic effects of monoclonal antibodies [107]. Despite this progress, the therapeutic effect of adjuvant treatment on cancer cells remains limited and sometimes triggered autoimmunity [101, 108]. Therefore, developing safe and more effective adjuvants is highly desirable for anti-tumor immunotherapy.

4.2 DC-based cancer vaccine

As shown in Figure 2B, the design principle of DC vaccines involves isolating monocytes from the human body. Monocytes are induced to become precursor DCs [121]. The precursor DCs were then incubated with TAA or tumor cell lysate and cytokines (such as GM-CSF, and IL-12) to differentiate into antigen-specific DCs, which were subsequently injected into the body to enhance the antigen presentation function [122-125]. In the past decades, Sipuleucel-T (Provenge®) was the only prostate cancer vaccine based on autologous DCs approved by the Food and Drug Administration in 2010 [126]. Sipuleucel-T isolates precursor DCs from the human body. These precursor cells were then stimulated to differentiate into mature DCs by GM-CSF and prostatic acid phosphatase. Ultimately, mature DCs were injected back into the patients to elicit an anti-tumor immune response [127]. However, Sipuleucel-T demonstrated limited clinical benefits, with only a 4.1-month improvement in median survival and no improvement in progression-free survival [128]. Low proportion of stimulated mature DCs and the complexity of administration and dosage were the potential disadvantages of first-generation DC vaccines [129, 130]. These problems have led to low clinical response rates to DC vaccines [130], warranting further exploration of diverse approaches to improve the efficacy of available vaccines. To date, many DC vaccines have been evaluated in clinical trials (Supplementary Table S2), and most vaccines are in phase I and II trials (93.94%) (Figure 3B). In these clinical trials, different approaches, such as exploring DC subsets and neoantigens, combining mRNA and DC vaccines, and combination therapy with ICBs, have been explored to identify the most promising candidates. Based on the results, these clinical trials have been summarized in Table 5. Most trials confirmed the safety and tolerability of DC-based vaccines and demonstrated clinical benefits for different clinical indications. Novel strategies for developing DC vaccines, including the development of new DC subsets, neo-antigen identification, mRNA vaccines, and immunomodulatory molecules, have been discussed in the subsequent sections.

Suitable DC subsets with better antigen presentation capabilities should be identified by cell surface markers [122, 146]. A promising subset of DCs, the basic leucine zipper ATF-like transcription factor 3-dependent X-C chemokine receptor 1⁺ (XCR1⁺) CD8⁺ DCs, exhibited potent stimulating effects in CD4⁺ T cells, CD8⁺ T cells, and NK cells [147-149]. The chemokine receptor XCR1 was specifically expressed on DCs, and its ligand XCL1 promoted cross-presentation [149]. Moreover, the technique used to separate XCR1⁺ DCs was much more straight forward than that for other subsets [149, 150]. These results confirmed that XCR1⁺ DCs may be a suitable subset for DC therapy.

Another strategy for improving the efficacy of DC vaccines was neo-antigen identification to overcome refractory and relapsed tumors [151]. An antigen loaded with DC vaccines often showed tolerance to immunity in refractory patients. Therefore, DC vaccines loaded with new TAA may inhibit tumor escape. Bioinformatics and genomic technological advances have made it possible to anatomize the immune response to personalized neo-antigens encoded by tumor-specific mutations. Ideal neo-antigens are tumor-specific peptides that are absent in normal human tissues [152]. In a study focusing on the genome analysis of tumors in melanoma patients, researchers obtained an individualized neo-antigen, which was loaded into DCs as a vaccine. This neo-antigen-loaded DC vaccine elicited vital T cell responses [153] and showed that neo-antigen identification may be beneficial for improving the anti-tumor activity of DC vaccines.

The mRNA vaccine, a new type of cancer vaccine, stimulated innate immune responses and provided antigens by cell transfection in vivo [154]. In a recent clinical study, an mRNA vaccine (mRNA-4650) comprised mRNA backbones encoding up to 20 different antigens, including autologous tumor antigens and neo-antigens [155]. Although the mRNA vaccine demonstrated a less significant anti-tumor effect, it increased the population of CD8⁺ and CD4⁺ neo-antigen-specific T cells. Another research team obtained relevant mutant neo-antigens from a mouse lung cancer cell line, LL2, and cultured them with DCs in vitro, followed by injection into mice [156], which resulted in continuous activation of CD8⁺ T cells and a large amount of IFN-γ production. In addition, gliomas with low mutations exhibited specific T-cell responses against the tumor using neo-antigens [157]. Based on these reports, neo-antigen DNA or RNA could be delivered as a vaccine to induce a positive immune response [103, 158].

In addition to tumor antigens, a group of DC vaccines loaded with immunomodulatory molecules enhanced the immune regulation [159]. DCs electroplated with TriMix mRNA encoded three immune-modulating molecules: TLR-4, CD40L, and CD70 (TriMixDC-MEL) [133]. According to other reports, melanoma-associated antigen was fused with human leukocyte antigen class II targeting DC-lysosomal membrane proteins to constitute the whole TriMixDC-MEL vaccine [159, 160]. TriMixDCs combined with the anti-CTLA-4 monoclonal antibody ipilimumab demonstrated vital T cell-specific activation ability in multiple experiments [159]. Compared with ipilimumab monotherapy, this combination therapy was beneficial for obtaining long-term clinical responses in melanoma patients [133]. Given the fewer side effects and durable clinical responses, TriMixDC in combination with ipilimumab may be more beneficial [133, 161]. Another engineered DC, SmartDC-tyrosinase-related protein 2 (TRP2), was constructed using a lentivirus expressing IL-4, TRP2 (melanoma antigen), and GM-CSF [162]. SmartDC-TRP2 was effectively transferred from the injection site to the local draining lymph nodes in a mouse model, where it persisted for a few weeks to induce anti-melanoma responses and T cell expansion [162]. These studies suggest that DCs equipped with immunomodulatory molecules exhibited enhanced efficacy, which provided a novel strategy for DC vaccines.

The advent of genetic engineering technology has opened a new era of personalized therapy, such as neoantigen DC vaccines; however, the following problems still need to be resolved: 1) the manufacturing cycle is lengthy and expensive; 2) the extraction of neo-antigens is complex, and the purification process is inefficient; 3) The efficacy of mono-therapy remains limited in the TME [158, 163].

4.3 New DC vaccine delivery system

In addition to antigen-loaded DC vaccines, effective delivery systems have been the focus of research (Figure 2C) [14, 130]. Mannose (MN)-labeled poly lactic-co-glycolic acid (PLGA) nanoparticles (MN-PLGA-NPs) were synthesized to encapsulate the TAA [164]. These nanoparticles target the MN receptor to initiate antigen presentation by DCs through the pattern recognition receptor [164]. Similarly, another team reported that TAA encapsulated in PLGA nanoparticles enhanced the antigen presentation ability of DCs [165]. Tateshita et al. [166] reported that an mRNA vaccine comprising lipoplex and vitamin E scaffolds exhibited greater cytokine release and enhanced CTL activity than that of an mRNA vaccine without lipoplex delivery. A multi-functional compound with lipids was also synthesized to improve stability by avoiding the catalytic hydrolysis of mRNA, which resulted in insufficient stimulation of DC cells [167, 168]. This compound activates TLR7 (or TLR8)/RLR (RIG-I and MDA5) and condenses with mRNA into lipid nanoparticles to promote cellular endocytosis and reduce mRNA degradation [167]. In addition, this study also demonstrated that nanoparticles loaded with OVA mRNA rapidly induced potent T cell responses, increased IFN-γ levels, and inhibited the growth of secondary inoculated tumor cells by producing durable immune effects. The delivery system is a critical component of DC vaccine therapy and can considerably impact the effectiveness of the treatment.

4.4 Combination therapy

Combining DC vaccines with ICBs improved the response rate in patients and prolonged their survival (Figure 2D) [122, 169].

Although PD-1/PDL-1 antibodies demonstrated restoration of immune cell killing ability, more than 50% of patients experienced poor efficacy [170]. PD-L1 was highly expressed in both peripheral and infiltrated DCs in lung cancer patients. Blocking PD-L1 in DCs enhanced T cell activation and proliferation, thereby prolonging patient survival [171]. Therefore, immunoregulation by DCs is critical in therapies targeting PD-L1, which was also proved by the fact that anti-PD-L1 antibodies did not control tumor growth and the CD8⁺ T cell population in the mice did not increase in the cDC1 function-deficient mice [172]. Dammeijer et al. [173] established two animal models of PD-L1 blockade targeting either tumor-draining lymph nodes or the entire body. They demonstrated that blocking PD-L1 in cDC2 cells induced effective tumor immunity. Furthermore, DCs induced tumor-specific T cell responses to ICBs via the stimulation of interferon genes activation [174]. Given that ICB therapy depends on DCs, the efficacy of combination therapy involving DC vaccines and ICBs was superior to that of single treatment modality [13, 133].

In addition to targeting immunosuppressive molecules (PD-1/PD-L1), therapies that leverage immuno-costimulatory molecules, such as CD40, enhance DC targeting strategies. CD40, which is expressed on B lymphocytes, DCs, and monocytes, is a crucial regulator of cellular and humoral immunity [175]. Its ligand, CD40L, is primarily expressed in activated T lymphocytes and platelets [176]. Activation of the CD40 signal on DCs triggered several immunological responses: upregulation of the expression of co-stimulated receptors and MHC molecules, enhancement of antigen presentation, production of pro-inflammatory cytokines (such as IL-12), and promotion of T cell activation [177]. Therapies targeting CD40 and FMS-related tyrosine kinase 3 ligand increased cDC infiltration and restored immune surveillance in pancreatic cancer [178]. Various agonists targeting CD40 have also been developed. Their safety profiles and enhanced immune cell function have been demonstrated in clinical trials [176]. In combination therapy using CD40 mAb and gemcitabine for the treatment of pancreatic ductal adenocarcinoma, 11 of 21 patients showed a favorable clinical response [179]. Salomon et al. [180] designed a new bispecific antibody (BsAb) that preferentially targeted cDC1 cells and activated CD40 signaling to further improve safety and efficacy. Compared to CD40 mAbs, BsAb mediated more robust T cell activation and anti-tumor activity [180]. In addition, a CD40 agonist-conjugated TAA and TLR5 binding domain directly generated a triple functional molecule to target DCs [181]. The activation signal of CD40 significantly improved the activity of DCs in the TME and provided a more durable anti-tumor response in combination therapy, which may be beneficial for patients who are not responsive to conventional immunotherapy [175, 176]. Another BsAb targeting DCs and T cells on PD-L1 and CD3 respectively, achieved durable anti-tumor activity through T-cell rejuvenation [182], which highlights the restoration of T cell function by DCs.

Both traditional combination therapies and novel approaches have demonstrated promising results. Fan et al. [183] designed novel antibody-engineered tDC-Exo-expressing anti-CD3 and anti-EGFR antibodies to mimic CAR-T therapy. This approach promoted the binding of T cells to cancer cells and achieved better efficiency when combined with an anti-PD-L1 antibody. T-cell-engaging BsAbs (T-BsAbs) such as teclistamab, blinatumomab, and mosunetuzumab, demonstrated improved anti-tumor activity than those of their parent mAbs. However, continuous exposure to T-BsAbs may induce T-cells exhaustion [184]. Fortunately, DCs counteracted this exhaustion by maintaining and guiding the differentiation of the precursors of the exhausted T cells [185]. This insight provides a new strategic approach for cancer therapy: combining T-BsAbs such as anti-CD3/tissue factor BsAb, anti-CD3/Lewis Y BsAb, and anti-CD3/EGFRvIII BsAb from our laboratories with therapies targeting DCs [186-188]. Another BsAb targeting PD-L1 and lymphocyte activation gene 3 demonstrated potent anti-tumor activity by promoting the activation of DCs and conjugation of T and tumor cells [189]. This was investigated in a first-in-human trial (NCT05101109). To summarize, therapies involving DCs achieve synergistic effects resulting in “1 + 1> 2” activity when combined with CAR-T cells [190, 191], ICBs [192], NK cells [193], cytokines [194] or other immunotherapies.