Therapeutic Cancer Vaccination with Ex Vivo RNA-Transfected Dendritic Cells—An Update

Over the last two decades, dendritic cell (DC) vaccination has been studied extensively as active immunotherapy in cancer treatment and has been proven safe in all clinical trials both with respect to short and long-term side effects. For antigen-loading of dendritic cells (DCs) one method is to introduce mRNA coding for the desired antigens. To target the whole antigenic repertoire of a tumor, even the total tumor mRNA of a macrodissected biopsy sample can be used. To date, reports have been published on a total of 781 patients suffering from different tumor entities and HIV-infection, who have been treated with DCs loaded with mRNA. The majority of those were melanoma patients, followed by HIV-infected patients, but leukemias, brain tumors, prostate cancer, renal cell carcinomas, pancreatic cancers and several others have also been treated. Next to antigen-loading, mRNA-electroporation allows a purposeful manipulation of the DCs’ phenotype and function to enhance their immunogenicity. In this review, we intend to give a comprehensive summary of what has been published regarding clinical testing of ex vivo generated mRNA-transfected DCs, with respect to safety and risk/benefit evaluations, choice of tumor antigens and RNA-source, and the design of better DCs for vaccination by transfection of mRNA-encoded functional proteins.


Tumors and the Immune System
Therapeutic cancer vaccination is a concept for treating tumor patients by immunizing them against their own tumor. As early as 1891, the application of bacterial substances (which we now know to have served as adjuvants) into tumors was executed by William Coley, who achieved a clinical response rate of 10% in soft tissue sarcoma [1,2]. During the 20th century, chemotherapy and radiation therapy were developed and superseded immunotherapy. The concept of immune surveillance, however, was later resumed and pursued [3,4] (and reviewed by [5,6]). The striking success that was achieved in preventive vaccination against infectious diseases suggested that the immune system could be utilized against malignancies in a similar fashion.

Therapeutic and Preventive Vaccination
There are, however, crucial differences between a preventive vaccine against an infectious disease and a therapeutic vaccination against an existing cancer. Firstly, the malignant cells arise from the body's own healthy cells-therefore, the immune system's primary function to distinguish self from foreign is circumvented. Secondly, the malignancy has developed in the presence of a functioning immune system and has hence adapted to immune surveillance. Thirdly, most tumors influence the immune system in their favor. Due to these circumstances, the induction of an effective anti-tumor immunity

RNA-Transfection of Dendritic Cells
The easiest way to achieve MHC class I-restricted presentation in this fashion is the intracellular expression of the antigens, and a method considered ideal for clinical application is mRNA transfection [76]. Several forms of RNA transfection have been used by different groups over the past decades, e.g.: passive pulsing of mRNA, i.e., simple co-incubation of mRNA with the DCs, lipid-mediated transfection of mRNA, and mRNA electroporation, i.e., transfer of mRNA molecules through the cell membrane by an electrical pulse (reviewed in [77]), with the latter being used by most groups at present. Figure 1 gives a schematic overview. When mRNA electroporation is performed properly, transfection rates above 90% are feasible. This physical method requires no additional chemicals, which is beneficial under GMP conditions. Until now, most clinical trials applying mRNA-transfected DCs have used mRNAs encoding defined non-mutated tumor antigens. However, one can also perform sequencing of the tumor and identify mutated neo-antigens, which are considered more immunogenic, but are hampered by the fact that they seem to be rarely presented and constitute less than 1% of the HLA-ligandome [78][79][80]. To our knowledge, mRNAs encoding such mutated neo-antigens have only been used as a direct vaccine by injection into the lymph nodes [81], and thus far not in a DC vaccine.

RNA-Transfection of Dendritic Cells
The easiest way to achieve MHC class I-restricted presentation in this fashion is the intracellular expression of the antigens, and a method considered ideal for clinical application is mRNA transfection [76]. Several forms of RNA transfection have been used by different groups over the past decades, e.g.: passive pulsing of mRNA, i.e., simple co-incubation of mRNA with the DCs, lipid-mediated transfection of mRNA, and mRNA electroporation, i.e., transfer of mRNA molecules through the cell membrane by an electrical pulse (reviewed in [77]), with the latter being used by most groups at present. Figure 1 gives a schematic overview. When mRNA electroporation is performed properly, transfection rates above 90% are feasible. This physical method requires no additional chemicals, which is beneficial under GMP conditions. Until now, most clinical trials applying mRNA-transfected DCs have used mRNAs encoding defined non-mutated tumor antigens. However, one can also perform sequencing of the tumor and identify mutated neo-antigens, which are considered more immunogenic, but are hampered by the fact that they seem to be rarely presented and constitute less than 1% of the HLA-ligandome [78][79][80]. To our knowledge, mRNAs encoding such mutated neo-antigens have only been used as a direct vaccine by injection into the lymph nodes [81], and thus far not in a DC vaccine. Endocytosed and phagocytosed material is usually presented in MHC class II to CD4 + helper T cells and can be cross-presented in MHC class I only under certain circumstances (not depicted). Proteins from the cytoplasm are, in contrast, primarily presented in MHC class I. mRNA can be introduced into cells by passive pulsing, relying on intrinsic and yet unknown means of uptake. It can be complexed with lipid reagents that mediate entry into the cytoplasm, or it can be transfected by electroporation. By applying a short electric pulse, pores in the cell membrane temporarily open, allowing entry of the RNA molecules. The transferred mRNA is then translated in the cytoplasm, and the encoded antigens are hence presented in a MHC class I context. By encoding signaling and targeting sequences (green) fused to the antigenic protein, this can also be directed towards the endosomal pathway, resulting in efficient MHC class II presentation.
An initial drawback of intracellularly expressed antigens is their limited presentation in MHC class II molecules for recognition by CD4 + helper T cells. While more experience was gathered in the field of DC vaccination, the relevance of T-cell-help in anti-tumor responses emerged. CD4 + helper T cells were shown to facilitate the generation of memory-type CD8 + T cells and, hence, this hurdle was overcome by fusing the mRNA-encoded antigens with targeting sequences, which direct the antigen towards the lysosomal compartment, thus mediating MHC class II-restricted presentation in addition to MHC class I-restricted presentation ( Figure 1). Common targeting sequences were derived from lysosomal-associated membrane proteins (LAMP) like LAMP-1 [57] and DC-LAMP [82]. Endocytosed and phagocytosed material is usually presented in MHC class II to CD4 + helper T cells and can be cross-presented in MHC class I only under certain circumstances (not depicted). Proteins from the cytoplasm are, in contrast, primarily presented in MHC class I. mRNA can be introduced into cells by passive pulsing, relying on intrinsic and yet unknown means of uptake. It can be complexed with lipid reagents that mediate entry into the cytoplasm, or it can be transfected by electroporation. By applying a short electric pulse, pores in the cell membrane temporarily open, allowing entry of the RNA molecules. The transferred mRNA is then translated in the cytoplasm, and the encoded antigens are hence presented in a MHC class I context. By encoding signaling and targeting sequences (green) fused to the antigenic protein, this can also be directed towards the endosomal pathway, resulting in efficient MHC class II presentation.
An initial drawback of intracellularly expressed antigens is their limited presentation in MHC class II molecules for recognition by CD4 + helper T cells. While more experience was gathered in the field of DC vaccination, the relevance of T-cell-help in anti-tumor responses emerged. CD4 + helper T cells were shown to facilitate the generation of memory-type CD8 + T cells and, hence, this hurdle was overcome by fusing the mRNA-encoded antigens with targeting sequences, which direct the antigen towards the lysosomal compartment, thus mediating MHC class II-restricted presentation in addition to MHC class I-restricted presentation ( Figure 1). Common targeting sequences were derived from lysosomal-associated membrane proteins (LAMP) like LAMP-1 [57] and DC-LAMP [82].
An additional merit of mRNA electroporation lies in the possibility of introducing not only antigens, but also functional proteins into the DCs, thus manipulating their phenotype and providing additional activation and maturation signals. The group around Kris Thielemans developed a DC-maturation process completely independent of exogenous cytokines by utilizing a mix of mRNAs encoding CD40L, CD70, and a constitutively active TLR4. This formulation was termed TriMix, and was used in a variety of clinical trials [40][41][42][43]46] (reviewed in [83]). Others used cytokine-matured DCs transfected with CD40L [84] to treat renal cell carcinoma [36] or HIV infection [72,73]. We commonly transfected cytokine-matured DC, but recently utilized a constitutively active form of IKKß to improve T-cell [85] and NK-cell activation [86], which will be tested in a phase I clinical trial soon.

The Total Tumor RNA Approach
The use of one or a few defined tumor antigens still limits the possibly induced anti-tumor immunity, and it has been shown that human anti-melanoma immunity is dominated by T-cell responses directed against somatically and individually mutated antigens [87]. Hence, the idea arose to use the entire tumor transcriptome by extraction, and, if needed, PCR-based amplification of tumor mRNA for subsequent electroporation into the DCs.
This allows an individualized treatment exploiting the complete antigenic repertoire of a given tumor, even if possible defined rejection antigens are yet unknown. In this aspect, total tumor mRNA is similar to the use of dead tumor cells but is not restricted by limitations regarding the size of excised tumors, reproducibility and validation. Exploitation of the total antigenic repertoire is considered critical as it targets not only overexpressed antigens but also the mutated proteins including both passenger as well as oncogenic driver mutations [88,89] and the emerging class of non-mutated neoantigens [90].

From the Bench to the Bedside
The potential clinical benefit of RNA-loaded DC therapy was first demonstrated in mouse tumor models. Mice treated with DCs pulsed with RNA from ovalbumin (OVA)-expressing tumor cells were protected against a challenge with OVA-expressing tumor cells [91]. In the same study, mice in the poorly immunogenic, highly metastatic, B16/F10 (B16) tumor model demonstrated a dramatic reduction in lung metastases in animals treated with DCs pulsed with tumor-derived RNA. Again using the B16 model, a second study showed that treatment with bone marrow-generated DCs, pulsed with either B16 cell extract or B16 total RNA induced specific CTLs against B16 tumor cells [92]. This treatment was able to protect animals from tumor located in the central nervous system (CNS), and led to prolonged survival in mice with tumors placed before initiation of therapy [92].
When this technology was taken from the bench to the bedside (as schematically represented in Figure 2), already the initial clinical trials were promising, demonstrating feasibility and immunogenicity, as well as hints for clinical efficacy. For example, a DC/RNA vaccine was explored in a phase I trial to treat eleven subjects presenting with metastatic RCC [31]. While the primary objectives of this study were safety, feasibility, and immunological assessment, it was noteworthy that tumor-related mortality was unexpectedly low among the 10 evaluable subjects who received the prescribed three administrations. The calculated mean survival following nephrectomy was 19.8 ± 3.1 months, although survival interpretation was confounded by the post-study therapies (predominately cytokine) received by most subjects. There were no adverse drug reactions with the exception of five subjects who experienced grade I injection site reactions consisting of inflammatory skin erythema lasting 48-72 h. A polyclonal tumor-specific T-cell response was detected in six subjects evaluable for immune response following DC treatment. Comparable results were observed in a phase I/II trial of melanoma therapy with autologous tumor mRNA [93,94]. The published clinical trials using tumor-RNA-transfected DCs are summarized in Tables 1-3.    EP with gp100-, tyrosinase-, MAGE-A3-, and -C2-DC-Lamp mRNA Cohort 1: 2 × 10 7 id, 4 × 10 6 iv Cohort 2: 12 × 10 6 id, 12 × 10 6 iv Cohort 3: 4 × 10 6 id, 2 × 10 7 iv Cohort 4: 24 × 10 6 iv 4 vaccinations in biweekly intervals, 5th vaccination with 10 weeks interval 11 pts local reaction grade II 3 pts chills grade II 8 pts flu like symptoms grade II 3 pts fever grade II         (1) Tumor material is isolated by surgery or biopsy. From this material, mRNA can be directly isolated. This RNA is usually amplified via a PCR-based method to gain sufficient amounts of mRNA. The tumor material can also be analyzed by sequencing, immunohistology, or other methods to identify antigens associated with this tumor, including somatic mutations. New bioinformatical methods can be used to identify the most promising T-cell epitopes for each individual patient. mRNA molecules that encode these antigens are transcribed in vitro from plasmid templates. (2) Cells (usually monocytes) that can be differentiated into DCs in vitro are isolated from the patient's blood. Alternatively, DCs, which are present in the blood, can be used. These DCs are matured to become immunogenic. Defined non-mutated antigens were used in most clinical trials with RNA-transfected dendritic cells and have proven immunogenic while autoimmunity was rarely observed, except for some cases of vitiligo in melanoma patients immunized with antigens expressed in both melanoma cells and melanocytes.
In the early studies with RNA-loaded dendritic cells, a simple co-incubation of DCs and RNA as opposed to electroporation was used. Since 2005, however, electroporation has commonly been used as a method to actively introduce mRNA into cytoplasm of the DCs. A total of 47 publications so far have described clinical phase I and II trials using RNA-loaded DCs to treat cancer and virus-infections. Within those, 781 treated patients were described. Most of those suffered from cutaneous melanoma (289) followed by viral infections (85), urogenital (79), and prostate cancer (73). For the complete summary, please see Table 1. In 10 of these trials, cells were loaded with autologous tumor-RNA, one used allogenic tumor RNA, and the other 36 used various defined antigens ( Table 1).
The last publication in a clinical trial with RNA-loaded dendritic cells dates back to 2018. However, several active clinical trials with mRNA-transfected DCs to treat cancer are listed in clinicaltrials.gov. By November 2019, 18 active trials (Table 4)   (1) Tumor material is isolated by surgery or biopsy. From this material, mRNA can be directly isolated. This RNA is usually amplified via a PCR-based method to gain sufficient amounts of mRNA. The tumor material can also be analyzed by sequencing, immunohistology, or other methods to identify antigens associated with this tumor, including somatic mutations. New bioinformatical methods can be used to identify the most promising T-cell epitopes for each individual patient. mRNA molecules that encode these antigens are transcribed in vitro from plasmid templates. (2) Cells (usually monocytes) that can be differentiated into DCs in vitro are isolated from the patient's blood. Alternatively, DCs, which are present in the blood, can be used. These DCs are matured to become immunogenic. (3) DC are transfected with the RNA to express the encoded tumor antigens. The DCs' own processing machinery degrades and presents the included T-cell epitopes in MHC. (4) The DCs are injected into the patient. Intradermal (id) and subcutaneous (sc) injection require migration via the lymphatic vessels towards the draining lymph node. Intravenous (iv) injection necessitates the transfer from the blood stream into lymphatic tissue. The direct injection into lymph nodes (i.n.) is an elegant approach, but is technically very difficult. (5) If the vaccine is successful, the DCs present the tumor-specific epitopes to T cells which are activated and attack the malignant tissue. Usually the DCs are injected repetitively to boost and maintain the responses. (The Motifolio Scientific Illustration Toolkit was used for the generation of this figure).
Defined non-mutated antigens were used in most clinical trials with RNA-transfected dendritic cells and have proven immunogenic while autoimmunity was rarely observed, except for some cases of vitiligo in melanoma patients immunized with antigens expressed in both melanoma cells and melanocytes.
In the early studies with RNA-loaded dendritic cells, a simple co-incubation of DCs and RNA as opposed to electroporation was used. Since 2005, however, electroporation has commonly been used as a method to actively introduce mRNA into cytoplasm of the DCs. A total of 47 publications so far have described clinical phase I and II trials using RNA-loaded DCs to treat cancer and virus-infections. Within those, 781 treated patients were described. Most of those suffered from cutaneous melanoma (289) followed by viral infections (85), urogenital (79), and prostate cancer (73). For the complete summary, please see Table 1. In 10 of these trials, cells were loaded with autologous tumor-RNA, one used allogenic tumor RNA, and the other 36 used various defined antigens ( Table 1).
The last publication in a clinical trial with RNA-loaded dendritic cells dates back to 2018. However, several active clinical trials with mRNA-transfected DCs to treat cancer are listed in clinicaltrials.gov. By November 2019, 18 active trials (Table 4) were listed mainly performed in the US (8), followed by Belgium (3), and Norway (3).

Clinical Efficacy
Within the large number of phase I/II DC trials that have been published, mixed responses (disappearance of some but not necessarily all metastases, also with appearance of new ones) and stabilization of disease were usually reported in a subset of patients. Objective responses, classically defined by disappearance of all tumors (CR) or a reduction of ≥50% (PR) were, however, less frequently observed. Interestingly, however, while overall response was found to be only 3.8% with non-DC-based cancer vaccines, in a much-debated article by Rosenberg et al., tumor regression was seen in 7.1% of patients receiving DC vaccination [95][96][97]. In select DC-vaccination trials, regressions were observed at higher rates, such as in DC-based vaccination for non-Hodgkin lymphoma targeting tumor-specific idiotype immunoglobulin (response rate of 31.6% [13,98]), or in melanoma trials when DCs were loaded with dying autologous tumor cells to vaccinate against the total antigenic repertoire of the individual tumors (20% overall response rate in stage IV melanomas [99,100]). These observations support the use of DC vaccines that target the antigenic repertoire of a given tumor as it can be achieved by loading DCs with total tumor mRNA as a technically more elegant approach which can be validated and thus be performed well under GMP conditions.
According to a review published by Ridgeway in 2003 [101], 78 of 98 analyzed trials included patient outcomes, although none of the clinical studies was designed to demonstrate the efficacy of the DC treatment. There was evidence of clinical response in at least one subject in 48 of the clinical trials, and one or more subjects experienced a complete response (CR) in 16 trials.
A review from Engell-Noerregaard et al. published in 2009 analyzed DC-based vaccination of patients with malignant melanoma [102]. A total of 38 articles were included for analysis, including 626 melanoma patients treated with DC-based vaccines. The objective response rate (CR and PR) was 9% with 20 (3%) complete responses and 37 (6%) partial responses. The clinical response rate (CR, PR, and SD) was 30% with 133 patients (21%) having stable disease. Apart from suggesting a clinical benefit in one third of the patients, the analysis was also interesting because it was found that SD was significantly associated with induction of antigen-specific T cells (p = 0.0003).
Regarding efficacy, it must, however, be emphasized that it has become clear that for active immunotherapies which lead to activation of tumor-specific T cells by either specific active vaccination (Dendreon's first generation DC vaccine Provenge™ [103]) or antigen-unspecific immune activation by taking off the brake from the immune system (anti-CTL-A4 treatment with Ipilimumab™ [104,105]), prolonged overall survival does not necessarily require regressions as defined by classical response criteria. Researchers in the cancer vaccine field were the first to point to this possibility. Indeed, the Provenge™ DC vaccine phase III trials provided the first proof for this concept because time to progression was not significantly prolonged while OS was, so that finally this vaccine got approved by the FDA for the treatment of androgen-independent, metastatic prostate cancer. Treatment with the anti-CTL-A4 antibody Ipilimumab™ in phase II and the subsequently published phase III trial exhibited 4 response patterns associated with survival, with only two of them corresponding to classical regressions [104].
In retrospect, one has to state that it was very optimistic to expect DC vaccines or any other cancer vaccine (such as fashionable neo-antigen vaccines [106]) by themselves to frequently produce significant clinical benefit in the setting of established late stage malignancies like stage IV melanoma, given the increasing evidence that the tumor microenvironment dictates whether tumor-specific T-cell responses will successfully alter the course of the disease [107][108][109] as checkpoint molecules suppress spontaneously arising or vaccine-induced T cells. Nevertheless, the reported clinical responses with mRNA-transfected DCs are at least encouraging. Table 3 summarizes efficacy data from all published clinical trials using RNA-loaded DCs.
In the setting of stage IV melanoma, i.e., significant tumor-load, and also many other tumors, the use of DC vaccines, which are reliably immunogenic, is, therefore, now best explored in combination with other treatments such as anti-CTL-A4 and anti-PD1 treatment. In contrast, it is timely to test in randomized trials immunogenic DC vaccines alone in the setting of minimal tumor load as performed in the adjuvant treatment of resected monosomy 3 uveal melanoma patients (NCT01983748), because in this setting clinical benefit is a more realistic possibility.

Safety of DC Vaccine Therapy
In general, DC-based vaccination is well tolerated, and few severe side effects have been reported. The events most often reported after vaccination with antigen-loaded dendritic cells are local reactions at the dendritic cell injection sites, flu-like symptoms (fever, chills, headache, and myalgia) and fatigue. These immune-related symptoms are meanwhile considered to be reactogenicity to the vaccine, and are valued as a sign of the immunostimulatory effectiveness of the treatment. The local symptoms observed after subcutaneous or intradermal application are usually absent at onset but appear upon repetitive vaccination indicating accumulation of T cells at the injection site/draining lymph node, the systemic symptoms fatigue and increase of temperature probably being an effect of cytokines released.
In about 20% of our patients infused with standard DCs, we have ourselves observed grade 1 to 2 flu-like reactions (including fever up to 39.4 • C) and constitutional symptoms within the first 72 h after infusion. Such side effects resolved upon treatment with paracetamol (usually 1 g administered i.v. followed by oral application) within 8 h. This delayed reaction resembles a mild grade 1 CRS (cytokine release syndrome), which is caused by T-cell activation and mediated by the release of pro-inflammatory cytokines into the plasma such as IL-1, TNFα, and IL-6. Indeed, we detected an increase of these cytokines in the blood of those patients who were vaccinated with standard DC and developed transient fever from 6 h after DC vaccination onwards.
Severe autoimmune side effects/immune-related adverse events (IRAE)-as now often observed as toxic side effect of anti-CTL-A4 [105] and anti-PD1 therapy-have not been a safety issue in the context of DC vaccination, even in patients vaccinated for prolonged periods including patients with tumor regressions. The induction of severe autoimmunity-related side effects is theoretically possible with DC-based immunotherapy. In the case of DC vaccines, so far, the induction of autoantibodies without clinical symptoms has been observed occasionally. Induction of overt autoimmune diseases, with the exception of the occurrence of cosmetically troublesome, yet otherwise harmless vitiligo caused by the spotty destruction of skin melanocytes has, however, not been described.
The absence of autoimmune side effects is, however, not due to the fact that DCs were not sufficiently immunogenic. Cosmetically disturbing vitiligo resulting from destruction of melanocytes in the skin was regularly observed in a small number of patients after vaccination with DC loaded with melanocyte differentiation peptides; yet no other organ damage occurred. As expected, this side effect has also been observed with DCs loaded with mRNA including total mRNA. Out of 31 patients vaccinated with monocyte-derived DC loaded with autologous tumor-RNA, 1 patient developed vitiligo [37]. In one of our phase I trials with stage IV melanoma patients (NCT00126685), 1 out of 8 fully evaluable patients developed vitiligo after vaccination with DC loaded with autologous tumor RNA. Within another phase I/II trial, 9 patients out of 42 developed vitiligo after vaccination with DC, electroporated with MelanA, Mage-A3, and survivin mRNA.
What has been additionally observed in DC trials are laboratory abnormalities including positive anti-nuclear antibody tests [16,19,56,57,[110][111][112][113][114][115][116], positive anti-dsDNA [111], positive anti-thyroid antibody tests [16,113,117,118], and positive rheumatoid factor [56,110,115,116,119]. Apart from four cases of thyroiditis [110], development of autoimmune antibodies was, however, not associated with clinically manifested autoimmune disorders such as systemic lupus erythematosus, rheumatoid arthritis, or dermatomyositis. In a trial using allogeneic DCs loaded with autologous renal cell carcinoma lysate, one patient experienced WHO grade IV thrombocytopenia [120], but it remained unclear whether thrombocytopenia was a side effect of the drug cyclophosphamide used in this trial too, from paraneoplastic origin, or triggered/aggravated by the allogeneic cell therapy.
Side effects observed in 46 publications describing the experience with DCs loaded with either defined RNA or RNA extracted from tumor cells (Table 4) did not significantly differ from those observed in the larger number of trials employing also monocyte-derived DCs but loaded by other methods but introduction of antigen mRNA (such as co-incubation with peptides or dead tumor cells). In the 781 patients treated with mRNA-loaded DCs, SAEs of more than grade II rarely occurred (Table 4). Often the attribution of such SAEs to the administered DCs has not been clear [57,58,69]. In one case, fatigue grade III occurred [38]. One study that must be mentioned here was recently described by Bol et al. [47]. They combined adjuvants from conventional preventive vaccines with DCs which resulted in heavy side effects including grade III flu-like symptoms, local reactions including purulent discharge and liver toxicity. All these symptoms were transient in nature and can be clearly attributed to the use of the adjuvants, because side effects of such extent have never been observed before, and were never observed afterwards.
Overall, the safety profile of DC vaccination including DCs transfected with mRNA is very good-notably as compared to any other treatment regimen for advanced malignancies.

Challenges and Future Perspectives of DC Vaccine Therapy
Although, as described above, DC vaccine therapy has clear merits, it is a very personalized medicinal product, requiring well-educated staff and a GMP-compliant facility for production, limiting its application. Depending on which source of mRNA is used for the transfection of the DCs, the costs and applicability can differ. For example, to gain a completely individualized product using autologous amplified total tumor RNA and autologous DCs, one has to obtain enough tumor RNA to perform the amplification procedure, and the RNA has to be produced for each patient. The same is true for the application in which individually mutated mRNAs are picked for the transfection, which additionally causes high costs for the sequencing of the tumor to find these mutations. On the other side, an off-the-shelf approach can also be chosen by using prepared mRNAs encoding non-mutated antigens often expressed in the tumor, which reduces the costs to some extent. Nonetheless, we strongly believe that the merits of mRNA-DC vaccine therapy overrule the above-mentioned disadvantages.
Adding DC vaccination to any of the standard therapies seems reasonable, since there is sound evidence that these standard regimens will enhance the T-cell induction by vaccination so that an enhanced clinical effect is possible, and a negative impact of the DC vaccine regarding clinical benefit of the standard therapy is unlikely. Importantly, there is no evidence for significant enhancement of undesired side effects as cancer vaccines including DC vaccines have been used without unexpected or clearly enhanced toxicity problems in man together with chemotherapy [121][122][123][124], as well as immune checkpoint blockade (ICB) [43,81,[125][126][127][128]. With respect to combination with checkpoint blockade, anti-PD-1 is a preferred backbone for combinations with small molecules and immune stimulatory agents, but for resistant tumors (like uveal melanoma) double checkpoint blockade is used for triplet therapies as a novel concept. This is evidenced by a search in clinicaltrials.gov, which shows over 120 phase I triplet trials.
It is perhaps unexpected that triplet trials using vaccines as partners for toxic double checkpoint blockade (55% grade 3-4 irAE, 33% DLT) are that advanced. On the other hand, it is logical, as all the vaccine strategies mentioned above have a very low documented toxicity, and thus qualify as preferred combination partners compared to small molecules, cytokines or agonistic checkpoint molecules. This appears particularly true for DC vaccines. The Tri-Mix DC vaccine developed by K. Thielemans' group was also given in combination with anti-CTL-A4, but again no grade 3 or 4 side effects occurred [43], even though Ipilimumab was used at the high dose level of 10 mg/kg, which because of the increased toxicity is avoided nowadays.
The combination of cancer vaccines (including DC vaccines) with chemotherapy has also been explored in the clinic (for a review also of ongoing trials see [129][130][131][132]). While significant synergism seems apparent only in few trials, most of them DC studies [133][134][135][136], a negative impact has never been reported, as the rules for a successful combination-also derived from animal studies-have been taken into account. These principles are: (1) avoid high-dose chemotherapy, (2) avoid combination after prolonged chemotherapy which results in general immunosuppression, and (3) avoid the concomitant administration of vaccines and cytotoxic drugs but rather administer about 1-2 weeks later. This circumvents the inhibition of activated, proliferating vaccine-induced T cells, and can dramatically foster T-cell responses by depletion of unwanted myeloid cells [133]. Gemcitabine has been explored in mice and humans in combination DC and other cancer vaccines with promising results [137][138][139][140][141]. Reassuringly, gemcitabine also synergizes with two other types of immune therapy, namely oncolytic virotherapy and ICB [142][143][144]. Like gemcitabine, fotemustine has already been tested in combination with prolonged vaccination with promising results and no added toxicity [145]. Interestingly, the combination with other immunostimulatory agents, namely IFN alpha + IL-2 and anti-CTL-A4 ICB, was also promising and clinical activity was observed again without evidence for enhanced side effects [146,147].

Conclusions
Undoubtedly, DC-based cancer vaccines are safe and feasible, and RNA-transfection is emerging as an ideal method for antigen-loading and functional manipulation of the applied cells. While other new cancer treatment regimens involve serious side effects, DC vaccination rarely produces adverse events higher than grade II. This allows a combination treatment in patients with a high tumor burden where DC-based monotherapy yielded only limited clinical results. Additionally, DC vaccines should be further extended to the adjuvant setting, to circumvent the massive immunosuppression exerted by a late stage tumor. Exploration of alternative DC origins and maturation protocols and the functional manipulation of the DCs by transfection with mRNA encoding proteins that trigger activation pathways is a consequent perpetuation to increase their immunogenicity. Currently running and future clinical trials explore these new approaches.