Keywords

19.1 Introduction

Mobilization of the immune system for the generation of an effective lymphocyte response against tumor tissue is one of the main goals of immunotherapy. It implies the necessity of a coordinated participation of the innate and adaptive immunity mechanisms in order to both trigger an effective response against tumor cells and preserve the host from autoimmune response. In this aspect, dendritic cells (DCs) perform a fundamental role in linking the innate defenses to the specific responsiveness by lymphocytes.

The very first report on DCs was published in 1868 by Paul Langerhans who found branched skin cells by gold staining (called Langerhans cells), whose “dendritic” extensions of plasmatic membrane resembled nervous cells [1]. A century later Prunieras [2] coined the expression “dendritic cells” for the Langerhans cells and proposed that they can capture antigens and are involved in primary defense against pathogens. However, the key contribution toward the morphological, phenotypical, and functional identification and classification of DCs as a new population of leukocytes was given by Steinman and Cohn, whose seminal reports from 1973 to 1978 are considered the beginning of a new era in this research field [3–7]. There are two main DC populations: the conventional DC, a myeloid-derived cell lineage, and the plasmacytoid DC (pDC), a lymphoid-derived lineage [8]. Although these two populations can be differentiated by morphological and surface markers, each DC type shows a wide phenotypical variation and multifunctional role in the immunosurveillance and regulation of the immune system [9, 10]. Thus, conventional human DC express CD4, CD11c, and CD1a or CD83 and the MHC class I [11, 12]. Maturation/activation of these cells is characterized by the expression of CD80, CD86, CD40, and CCR7 [8]. Differently, lymphoid pDC are featured as CD4+/CD1a−/CD11c−/CD123+ cells [13].

DCs are the main professional antigen-presenting cells (APCs) and perform a continuous surveillance and recognition of the microenvironment of tissues and organs where they are found as immature cells (iDCs). In this condition, they have high capacity for capturing soluble and particulate antigens by endocytosis, phagocytosis, and micropinocytocis [3, 11, 14, 15]. The intakes of opsonized and non-opsonized antigens can be mediated by several surface receptors such as FcγR [11], mannose receptor (MR) [16], DC-SIGN [17], type C lectin receptors (DEC-205) [18], as well as Toll-like receptors [12, 19]. These antigens are then processed into peptides that are subsequently presented to T lymphocytes in the context of the major histocompatibility complex (MHC) [11, 12, 20].

Immature DCs do not have the unique ability for stimulating naïve T cells, since in this state they do not have the co-stimulatory signals required for T-cell activation. Considering that contact between iDC and a specific T cell can drive lymphocytes to cell anergy or induce regulatory cells [21, 22], DC maturation is critical for achieving the balance between effector responsiveness and autotolerance [11].

Proinflammatory signals induce not only the migration of iDC to the secondary lymphoid organs but also their maturation and activation. In contrast to iDC, mature DCs show reduced endocytic and antigen processing ability, while becoming highly efficient presenters of processed antigens for lymphocytes at the T-cell sites of lymphoid organs. Mature DCs express a higher density of CCR7 that drives their chemotactic migration toward the T-cell sites [11, 23].

Maturation is also followed by increased expression of a set of the abovementioned surface markers and by production of several proinflammatory cytokines, such as IL-12, IL-18, TNF-α, IL-23, IL-10, and IFN-α, depending on the stimulating factor [24–26].

Phenotypical and cytokine profile of mature DC contribute to the recruitment, interaction, and activation of lymphocytes for the development of an efficient response against pathogenic microbes, allergens, and allogeneic tissues [27, 28] and were also evidenced in antitumor response [8]. In fact, it was reported that tumor mass-infiltrating DCs are usually suppressed or maintained as iDC in situ. These observations have instigated many authors to try to stimulate infiltrating DCs to play a more effective role against tumor cells [29, 30] or to transfer autologous or allogeneic DCs after in vitro loading with tumor antigens, thus giving rise to several studies on the feasibility of using DC as therapeutic vaccines for active immunization of cancer patients.

Such studies have benefited from the observation that murine DC can be differentiated in vitro from bone marrow precursors. Further investigations were strongly reinforced by the finding that human DC could be differentiated from peripheral blood monocytes through treatment with adequate cytokine cocktails, usually a combination of IL-4 and GM-CSF [8, 31–34].

Being the main professional antigen-presenting cells, DC constitutively express both MHC class I and class II antigens on their surface. This feature is closely associated with their effective antigen-presenting function, whereas strategies for improving the expression of these molecules have been proven to enhance the antitumor response triggered by DC vaccines. In this aspect, it was early observed that increasing the expression of MHC class II molecules on DCs by transfecting them with MHC class II transactivator genes (CIITA). It induces four times more CTL than parental untransfected DC or DC transfected with irrelevant genes [35].

In an early report, even before the flourishing of proposals for DC-based antitumor vaccines (DC vaccine), it was observed that monocyte-derived phagocytic cells could be sensitized by apoptotic bodies obtained by dead tumor cells [36]. Current studies are still using peripheral blood cells to generate human DC and bone marrow cells for murine ones; however, the efficiency of these vaccines appears to be dependent on a number of factors including generation of mature DCs [37–39], sustained production of IL-12 [40–43], and overcoming the suppressive microenvironment provided by regulatory T cells [37, 44–47] and myeloid-derived suppressor cells [48–51]. In fact, there is a variety of approaches to generate DC vaccines and it has been observed that each type of tumor has particular features that can hinder the effectiveness of such preparations.

19.2 Strategies for Developing Clinical Grade DC Vaccines

One of the main issues for generation of clinical grade antitumor DC vaccines is the choice of the technique for DC loading with tumor antigens. They range from the easier antigen preparation of tumor cell lysates by quick freeze-and-thaw cycles to the generation of tumor-DC hybrid cells or their transfection with tumor nucleic acid. However, there is still no definitive agreement on what strategy is the best.

Results with DCs loaded with lysates of tumor cells are controversial since some studies have shown that this approach results in a poor protective role of DCs, whereas other authors have successfully prepared them. Some details can be crucial to the effectiveness of lysate-pulsed DC vaccines. For instance, [52, 53] inhibitory effect of lysate on DC maturation can be reduced when tumor cells are stressed by heating at 42 °C for 25 min prior to the cell lysate preparation. It is hypothesized that the expression of heat shock proteins (HSPs) by tumor cells can avoid the suppressive effect of cell lysate by increasing DC maturation, an observation corroborated by others [54–56]. Induction of HSPs may be a required feature for increasing the immunogenicity of tumor cells by treatment with chemotherapeutic agents. The authors have observed that low nontoxic concentrations of paclitaxel or doxorubicin are able to alter the expression of a number of genes including HSP70, HSP40, and HSP105 mRNA [53].

Aiming to compare different methods for loading DCs with tumor antigens, it was observed that lysate obtained from a homogenate of solid tumor cells exerted a poor effect on the ability of DCs to stimulate antitumor activity [57]. Stressed tumor cells were obtained by freeze-and-thaw cycles or by irradiation at 30 Gy, with the irradiation being more useful than a freeze-and-thaw process. However, the best method for loading DCs in the mentioned study was their fusion with live tumor cells. The authors observed that irradiation of tumor cells at 30 Gy was effective at blocking their proliferative ability and did not affect their usefulness in preparing tumor-DC hybrids. For clinical purposes, loading DCs with tumor-associated proteins or peptides has been preferred in relation to the total tumor lysates. In a phase I study, patients with advanced melanoma were vaccinated with CD34+-derived DC pulsed with melanoma peptides. Some patients showed peptide-specific DTH response, as well as Melan-A- and gp-100-specific CTL in the peripheral blood [45].

One of the limitations of preparing DC vaccines pulsed with tumor lysate is that the available tumor tissue is usally not sufficient for repeated applications for the patient. The use of tumor RNA for encoding tumor antigens was first proposed by Nair and Gilboa’s group [58, 59], and there is substantial evidence that RNA transfection is a superior method for loading antigens onto DC [60–62]. An important point to consider is that tumor RNA can be amplified through molecular biology techniques, so that even a small amount of original RNA can be employed to obtain sufficient material for DC loading. Moreover, both total RNA and selected sequences can be used for DC-pulsing in order to drive the antigen presentation toward a more specific immune response. Finally, RNA shows a safety advantage on DNA, since it cannot be permanently integrated into the host genome.

The strategy of DC transfection with CEA RNA has been used both in murine [63, 64] and human systems [59, 65, 66]. Sakakibara et al. [67] have proposed a method for generating DC vaccines more rapidly by incubating monocytes with GM-CSF and IL-4 for 24 h (fast DC) transfection with tumor mRNA and cultivation with a maturation cocktail for an additional 48 h. The authors observed that mature fast DCs and standard DCs displayed comparable levels of many markers expressed on DCs, including HLA-DR, CD83, CD86, CD208, and CCR7. Both were equally able to elicit specific T-cell response and IFNγ-secreting T cells, leading to the conclusion that mature fast DCs are functional antigen-presenting cells (APCs) capable of inducing primary T-cell responses.

Vaccination with tumor-DC hybridomas using autologous melanoma or renal carcinoma cells and allogeneic DCs is able to change the natural history of the disease, since it may present stabilization [31] or even regression of metastatic lesions with local fibrosis [68]. Whether a patient was unable to fight the tumor development, it is probable that his/her own DCs were unable to efficiently process and present relevant tumor antigens to generate specific CTLs. The fact that most tumor antigen peptides are considered to be self-antigens hampers the generation of an effective CTL response. This point of view has led some authors to suggest the use of allogeneic or semi-allogeneic systems to generate DC vaccines. Fusion of allogeneic DCs with autologous metastatic colon cancer cells was able to activate both CD4+ and CD8+ T cells in just 24 h, in a higher number than controls, while CD8+ cells were significantly able to lyse target cells [69]. It can also solve some practical problems, namely, (a) it is usually possible to generate a limited number of samples of autologous DCs for vaccination, whereas a higher number of DCs could be generated from healthy allogeneic or semi-allogeneic donors; (b) the cellular reactivity triggered by allogeneic or semi-allogeneic DCs for allogeneic MHC antigens could facilitate the elimination of escaped tumor variants, as happens in the recipients of semi-allogeneic bone marrow transplantation; and (c) autologous tumor cells are sometimes scarce, which may be overcome by the use of stable tumor cell lines as the source of allogeneic tumor antigens for pulsing autologous DCs.

Evaluation of the efficiency of syngeneic, allogeneic, and semi-allogeneic DCs has shown that hybrid cells prepared with allogeneic or semi-allogeneic DCs were more effective than syngeneic ones and also worked better as therapeutic vaccines, thus protecting hosts against pulmonary metastasis. Actually, allogeneic and semi-allogeneic DCs more effectively induce CTL activity, as well as NK cytotoxicity, and induce higher levels of IFN-γ, as well as the IFN-γ:IL-10 ratio [70].

The use of exosomes for DC loading has also been proposed by some authors [71–74]. Exosomes are defined as constitutive nanovesicles that can be exocyted by both tumor and DCs displaying a sample of all membrane molecules of original cells [75, 76]. It was observed that vaccination with tumor peptides is more effective when carried on exosomes [72, 77]. Dai et al. [54] revealed that these nanovesicles can be isolated from heat-stressed tumor cells, culturing them for 43 h at 37 °C, followed by incubation for 1 h at 43 °C. After purification by ultracentrifugation on a discontinuous density sucrose cushion, exosomes were used to induce maturation of monocyte-derived DC. DCs loaded with such nanovesicles showed strong upregulation of HLA-DR, CD86, and CD40, as well as the production of IL-12p70 and TNF-α. This technology can also be used for increasing the immunogenicity of tumor cells, since they are able to uptake mature DC exosomes and express themselves, thus activating molecules such as HLA-DR and CD86 [78].

Cross-priming performed by DC is a phenomenon that can enhance the transference of antigenic peptides through HSP, such as gp96 and HSP70 [79–81]. Some HSPs obtained from tumor cells seem to be loaded with tumor antigens and can be internalized by DC through phagocytosis receptors. Such peptides can further be presented in the MHC class I context for inducing CD8+ response and subsequent specific attack toward tumor cells [82–85]. Although the use of HSPs seems to represent a good strategy for enhancing the DC loading with tumor antigens [86–88], the clinical application faces some limitations including the difficulty to construct the HSP-peptide complex and the necessity of a large amount of antigen source for obtaining a sufficient quantity of purified HSPs [89].

19.3 Routes of Administration

Another fundamental aspect of DC-based immunotherapy is the route of choice for administrating ex vivo prepared DCs. Clinical trials have reported various routes of DC administration, aiming to achieve an efficient delivery of cells to the appropriate immune site. Therefore, DCs can be inoculated by intradermal (i.d.), subcutaneous (s.c.), or intranodal (i.n.) routes to deliver loaded cells to regional lymphoid tissues, whereas intravenous (i.v.) methods should be chosen for their systemic distribution. There are also a number of studies showing the feasibility of intratumor (intralesional) inoculation of DC vaccines.

In vivo tracking of s.c.- and i.d.-inoculated DCs in multiple myeloma patients revealed their migration to the regional lymph nodes [90]. In fact, the i.d. route seems to be more efficient than s.c. for cell delivery to lymph nodes of patients with metastatic diseases [91]. Although these routes lack DC migration to the spleen, they appear to be more effective for inducing specific antitumor responses compared to the i.v. method [92, 93]. Tracking studies have also revealed that i.v. inoculation promotes DC distribution to the liver, spleen, lungs, and bone marrow. It was observed that DCs accumulate in the spleen just 3–24 h after inoculation [92]. Since the majority of relapsing diseases result from metastatic tumor cells, it is reasonable to infer that systemic distribution of DCs to the main targets for metastasis (lung, liver, and bone marrow) would be preferred in the protocols developed for preventing them [94–96].

Despite the suppressive microenvironment established at the tumor site, intralesional administration of DC was shown to be feasible, safe, and well tolerated [97–99]. Of course, this choice is limited by the tumor accessibility, while Mirvish et al. [100] suggest that in some cases the combination of different routes should be necessary for achieving successful immunization.

Considering the different designs for tumor antigen delivery, as well as the different administration routes, in the next section we will highlight the clinical experience in relation to selected diseases.

19.4 DC Vaccine for Prostatic Cancer

Prostate cancer is the second most frequent type of neoplasia worldwide, accounting for more than 903,500 new cases each year [101]. Most patients are successfully treated by prostectomy or radiotherapy, but about 30 % of them relapse [102]. In this aspect, immunotherapeutic approaches became an attractive alternative treatment, particularly for patients with the advanced disease, since the conventional treatments are merely directed against the symptoms. In addition, its feature of slow progression facilitates the manipulation of the immune system in order to enhance the recognition of tumor antigens.

The first DC vaccine approved by the U.S. Food and Drug Administration (FDA) for cancer therapy targets prostate cancer [103–105]. This vaccine, called sipuleucel-T (Provenge® – Dendreon, Seattle, WA, USA), was developed for castration-resistant metastasis of PC (for both symptomatic and asymptomatic patients). It is a DC-enriched autologous cell suspension from the own patient prepared by culturing them with a fusion protein called PA2024, which is constituted by the granulocyte-macrophage colony stimulating factor (GM-CSF) and the prostatic acid phosphatase (PAP) widely expressed by tumor cells [105–107]. The analysis of disease progression and overall survival in two phase III studies (D9901 and D9902A) showed that this vaccine was able to increase the overall survival from 4.5 to 6.7 months [104, 105].

A third phase III trial has shown that sipuleucel-T improved patient survival time by 4.1 months, showing a 22 % lower relative risk of death than the placebo group [103]. Another positive result of these trials is that patients have shown variable reduction of PSA levels (prostatic-specific antigen), the main prognostic marker of this disease [104, 108].

The cellular immune response was also improved by treatment with sipuleucel-T, with 73 % of patients presenting an adequate lymphoproliferative response, whereas merely 12 % of the placebo group showed similar responsiveness [103]. In addition, generation of PAP-specific T lymphocytes was significantly higher in vaccinated patients than in those receiving placebo (27.3 % vs. 8.0 %), while minimal and well-tolerated collateral effects were also observed [106, 109].

In another successful approach, prostatectomized patients with biochemical relapse were treated with autologous DCs pulsed with human recombinant PSA (Dendritophage-rPSA) [110, 111]. Nine out of twenty-four patients showed 50 % reduction in PSA levels, whereas 11 others showed less pronounced diminution (6–39 %). In addition, 13 patients showed PSA-specific T-lymphocyte responsiveness. Six of the patients did not present any sign of circulating tumor cells during a 6-month follow-up. These results are favorable since handling patients with biochemical relapse is still a challenge for oncologists, urologists, and radiotherapists, due to the difficulty of ascertaining the correct location of relapsing disease.

Considering the difficulty of obtaining sufficient amounts of tumor antigens, Fong et al. [111] have proposed the use of xenogeneic murine PAP for loading autologous DCs. Six out of twenty-one patients with metastatic prostate cancer showed stabilization of the disease, with no rise of PSA levels nor the development of PSA-specific T cells.

Preparation of tumor-DC hybrid cells was also tested in prostate cancer. Hybridomas prepared with three different cell lines successfully induced an in vitro response in a mixed leukocyte culture by enhancing the IFN-γ production. Results were especially evident when ONYCAP23 and LNCaP were used for fusion (73 % and 67 %, respectively). Interestingly, the ONYCAP23 based hybridoma have induced specific T-cell response to different tumor targets [112].

A phase I/II study using DCs pulsed with allogeneic tumor cell lysate has demonstrated good tolerance and absence of toxic effects. However, although some patients have presented significant in vitro proliferation of specific antitumor lymphocytes, this approach has not achieved relevant clinical results [113].

19.5 DC Vaccine for Melanoma

The first clinical study on DC vaccines in melanoma patients was published by Nestle et al. [114], who analyzed the efficacy of DCs pulsed with HLA-A2-restricted peptides and autologous tumor cell lysates. Two out of six patients presented complete response to vaccination, while four of them developed specific DTH response.

The use of allogeneic tumor cell lysate for loading DCs, assessed in a phase I/II study, found that only 1 out of 15 patients with melanoma treated with autologous iDC pulsed with tumor lysate showed complete remission of metastasis. When the follow-up was discontinued, this patient had maintained an asymptomatic condition for 24 months [115].

More recently, melanoma patients were treated with DCs pulsed with melanoma peptides (HLA-A2+) or tumor lysates (HLA-A2−), in association with IL-12, celecoxib, and metronomic doses of cyclophosphamide (phase II study). This association was well tolerated by patients, and 29 % of patients with metastasis had the disease stabilized for 7–13.7 months. These patients also showed a higher median overall survival than patients with progressive disease (10.5 vs. 6 months). No significant difference of efficacy was observed between DCs loaded with cell lysate and peptides, although no correlation was found between the development of specific immune response and clinical response [116].

The use of autologous tumor RNA for loading autologous DC has promoted increased numbers of IFN-γ-producing CD4+ cells [117]. This result merits attention because the strategy of using RNA aims to stimulate CD8+ response since it implies the generation of tumor peptides at cytoplasm, which would be processed through the cytosolic machinery. Thus, the expected effect on the activation of CD4+ cells can favor the establishment of memory CD8+ cells (Shedlock and Shen 2003; Janssen et al. 2003). In a phase I/II study, Kyte’s group showed that administration of RNA-pulsed DCs was able to induce a specific DTH reaction and in vitro lymphoproliferative responsiveness as well as IFN-γ production [118].

Cell fusion technology was also applied to melanoma and kidney cancer patients, by fusing autologous tumor cells with allogeneic DC obtained from healthy donors [31, 119]. The measurable clinical response from these patients demonstrated that the disease had been stabilized for a median of 6 months, with no relevant collateral effects [31].

19.6 DC Vaccine for Colorectal Cancer

DCs are constituent cells of lamina propria and are involved in every local pathological condition. Mechanical disaggregation and enzymatic digestion of intestine specimens of patients with different types of colon disease – including colorectal cancer, Crohn’s disease, ulcerative colitis, and nonmalignant, noninflammatory conditions – show that DCs correspond to 2 % of cells isolated from lamina propria [120]. As to the ability of these cells to stimulate lymphocyte activity, DC-rich suspension induces mixed lymphocyte response (MLR) by T cells. However, tumor-infiltrating DCs poorly stimulate T lymphocytes in a primary allogeneic culture (MLR) and are not able to induce significant levels of IL-2 or IFN-γ [120].

The C-type lectin DC-SIGN (DC-specific intercellular adhesion molecule-3-grabbing non-integrin) is involved in the recognition of colorectal cancer cells by DCs [121]. Immature DCs within colon tumor tissue expressing DC-SIGN, but not mature DCs, interact with tumor cells by binding to Lewisx and Lewisy carbohydrate of CEA in tumor cells. Interestingly, DC-SIGN do not interact with CEA expressed by normal colon epithelium that shows low levels of Lewis epitopes. Therefore, DCs interact with human colon SW1116 tumor cells that express aberrantly glycosylated Lewis epitopes (Lea/Leb) of CEA and CEA-related cell adhesion molecule 1 (CEACAM1), an interaction that induces the production of immunosuppressive cytokines such as IL-6 and IL-10 [122].

Immunohistochemical analysis of infiltrating cells showed that mature CD83+ DCs are found in almost all primary colon carcinoma samples and in some metastases. Heterogeneous infiltration patterns vary from diffuse cells to clustered DCs that tend to accumulate around vascular structures and the marginal zone of lymphoid aggregates [123]. Data on maturation markers on DCs that infiltrate primary tumors are contradictory. Indeed, some authors observed that around 90 % of CD83+ cells were double-stained by anti-CD40 or anti-CD86 antibodies, indicating their in vivo activation [123], whereas others reported that 64–97 % of cells do not express B-7 molecules [124, 125], even after stimulation with TNF-α, IL-4, and GM-CSF [125]. The density of DCs at the tumor site was higher in patients with a high proportion of activation markers (CD86 and CD40), suggesting that mature DC can actively migrate to or be activated in the tumor microenvironment under exposure to tumor antigens [123].

Immunization of patients with DC vaccine in phase I/II clinical trials showed that the vaccine was effective for 16.7 % of patients in the phase I study and for 23 % of them in the phase II study [59]. Messenger RNA of TAT protein transduction domain and calreticulin increase the immunogenicity of CEA and the effectiveness of mRNA-pulsed human DCs. It is interesting that transfection of DCs with calreticulin mRNA seems to be associated with activation of CD4+ T cells, whereas TAT protein mRNA preferentially stimulates CD8+ cells [126]. Since mRNA represents only up to 5 % of total cell RNA, in vitroamplification of mRNA was shown to be feasible for producing immunogenically active CEA-encoding mRNA [65].

Instead of using mRNA for known specific antigens such as CEA and HER2/neu, DCs transfected with total tumor RNA were able to induce CTL response, while effector cells were able to recognize both the original tumor cell line used for RNA preparation (SW480) and other cell lines, namely, HCT-116 (colon cancer) and A498 (kidney cancer) [127]. Supporting this strategy, a clinical trial using total RNA extracted from metastasis tumor cells for pulsing autologous DCs, followed by inoculation in the patients (four injections, every 4 weeks), showed an ability to induce specific T response to CEA [128].

Analysis of ten clinical samples of colorectal carcinomas showed that 60 % of them overexpressed the antigen EphA2 [129]. Murine DCs pulsed with human EphA2 were observed to induce antitumor response against EphA2-transfected MC38 cells. Results have shown that Eph-DC strongly delayed the tumor growth and induced specific CD8+ cells and CD4+ cells which play a critical role in the antitumor response.

19.7 DC Vaccine for Nervous Tissue Cancer

As reviewed by Montelli et al. (2009), the potential clinical use of DC vaccines against brain tumors has also been investigated by some groups. The first DC vaccination study in patients with malignant glioma was reported in 2001 by [130], showing increased tumor-specific cytotoxicity in four out of seven patients treated with peptide-pulsed DCs. In a phase I clinical trial conducted by Sampson et al. [131], 13 patients with glioblastoma (GBM) and 3 with WHO grade III glioma were i.d. inoculated with autologous DC vaccine. Peripheral blood monocyte-derived DCs were pulsed with peptide from a mutated region of EGFRvIII conjugated with KLH (keyhole limpet hemocyanin). After three doses, immunization resulted in the restoration of immune responsiveness, which was followed only by grade I or II local reaction at the administration site. Treatment resulted in a median survival time of 110.8 weeks, which was higher than usually observed in patients under other types of therapy such as temozolomide (63.3 weeks [132]) and carmustine wafers (59.6 weeks [133]).

Parajuli et al. [134] studied in vitro the ability of different DC-vaccine strategies to induce T-cell response against malignant astrocytomas. Autologous monocyte-derived DCs were pulsed either with autologous tumor lysate, transfection with total tumor mRNA or by fusion of DCs with tumor cells. The authors concluded that all strategies used for pulsing DCs efficiently induced T-cell cytotoxicity, which was further improved by addition of CD40 ligand [135].

Twelve GBM patients followed in a phase I trial were treated with DC vaccines pulsed with peptides eluted from autologous tumor cells. After 3 doses, 50 % of the patients presented increased immunological response against autologous tumor cells and survival time was higher than historical control data [136].

In a very expressive clinical trial, 56 patients with relapsing GBM were treated with at least 3 doses of autologous DCs loaded with autologous tumor lysate, producing a 3-month median progression-free survival and a 9.6-month overall survival. Almost 15 % of patients presented a 2-year overall survival, although some of them have presented relapse during the follow-up [137]. In a phase II study, patients producing increased levels of IFN-γ showed higher overall survival than nonresponders [138].

Polarization of type 1 response can also be achieved by polyinosinic-polycytidylic acid stabilized by lysine and carboxymethylcellulose (poly-ICLC), a type 1 IFN inducer. This product acts on TLR3 [139] to induce the production of IFN-γ, IL-6, TNF-γ, and chemokines including CCL2, CCL5, CCL20, and CXCL10 from astrocyte and microglia [140, 141]. Among the 38 patients with malignant glioma enrolled in the first clinical trial, those inoculated with poly-ICLC showed minimal toxicity associated with the treatment. Sixty-seven percent of the patients exhibited tumor regression or stabilization under radiological evaluation, with a 19-month median survival [142]. Antitumor response was associated with activation of 2′5′-oligoadenylate synthetases, which are antiviral proteins induced by type I IFN [143]. In another study, 30 adult patients with glioblastoma multiforme received poly-ICLC in combination with radiotherapy, thereby demonstrating an advantage in relation to historical studies using radiotherapy alone [144]. Okada’s group is currently analyzing the effect of associating poly-ICLC with DC vaccines generated under INF−α(αDC1), previously shown to be more effective than conventional DCs at inducing an antigen-specific CTL response [145].

19.8 Concluding Remarks

Despite their demonstrated effectiveness and promising results, the clinical use of DC vaccines is promising but not definitive. It can be partially explained by the difficulty of establishing a standard effective source of antigens and because several tumor-associated antigens are shared by normal cells. In addition, the increased Treg cells in advanced cancer, as well as other suppressor cells, can hinder the efficacy of a DC vaccine. In fact, even after activation, the autologous DCs of breast cancer patients induce higher levels of regulatory T cells (Treg) than DCs from healthy donors [146], which determines a low immunogenicity of autologous monocyte-derived DCs usually suppressed or induced to tolerance by Treg cytokines.

Reduction of Treg activity by blocking the regulatory molecule CTLA-4, through a monoclonal antibody, can be a good strategy to overcome this obstacle. The FDA reinforced this possibility through its 2011 approval of anti-CTLA-4 (ipilimumab – Yervoy; Bristol-Myers Squibb) for treatment of metastatic advanced melanoma. Treatment was well tolerated by patients and the combination with autologous DC vaccine or peptide-based vaccination was able to develop a significant antitumor response [147, 148].

In conclusion, despite these limitations, promising results are stimulating the search for the best pathways toward improving tumor immunogenicity, DCs’ antigen-presenting function, responsiveness of effector cells in the tumor microenvironment, as well as overcoming the tolerogenic or suppressive status of the patient’s immune system. Association of different immunotherapeutic approaches or combination of immunotherapy with chemotherapy [53] can open up new avenues for fighting cancer.