Introduction

Acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) are common hematopoietic diseases characterized by uncontrolled clonal malignant cell proliferation with leukemic blasts replacing the cells that perform normal physiological hematopoiesis and associated symptoms of anemia, bleeding, and infections [1,2,3]. High-dose induction chemotherapies, followed by allogeneic hematopoietic stem cell transplantation (HSCT), are the only curative options in selected patients [1, 4, 5]. With advances in combined treatment options, the majority of patients can achieve remission after induction chemotherapy; however, only a minority of patients enjoy durable responses. Drug resistance and relapse remain the major challenges, and the 5-year overall survival (OS) rate of AML patients has stagnated at less than 30% [1, 6, 7].

It is well established that the curative effect of allogeneic HSCT derives from allogeneic lymphocyte-mediated graft versus leukemia (GVL) reactivity [8, 9]. However, the complexity, cost and high rates of HSCT related morbidity and mortality have limited the clinical application of HSCT in all patients. Alternative curative approaches to harness antileukemia immunity have been under active investigation in recent decades. Dendritic cell (DC) vaccination as immunotherapy in patients with AML was initiated a few years after DCs were discovered by Ralph Steinman in 1973 [10] but did not see much progress for a long time. Recently, DC vaccination in AML and MDS has received renewed attention with new technologies applied in the vaccine development and patient selections [11, 12]. Different types of DC vaccination strategies using a variety of antigen sources to promote antitumor immunity have been explored. Recently, DC vaccines combined with conventional chemotherapy, systemic monoclonal antibodies or immune checkpoint-targeting strategies, such as those targeting programmed death 1/programmed death-ligand 1 (PD-1/PD-L1), have been studied. Although the development of DC vaccination in patients with AML and MDS has progressed somewhat, there is still a long way to go before therapeutic DC vaccines can be translated from the research laboratory to the bedside. In this review, we summarize the recent advances in DC vaccines immunotherapy for AML/MDS as well as other nonleukemia malignancies.

DC vaccine cell types and clinical data

DCs are major antigen presenting cells (APCs) that process and present antigens via major histocompatibility complex I and II (MHC I and II) molecules to the innate and adaptive immune systems [13] and play a key role in the interface and crosstalk of the innate and adaptive immune systems [14]. DCs activate NK cells to control pathogens through the innate immune system and activate the adaptive immune system to realize immune memory [15]. Furthermore, DCs form immunological synapses with T cells, resulting in potent T-cell activation against the presented antigens [16]. DCs can enable CD4+ T cells to activate B and CD8+ T cells, mediate immune memory, and activate Tregs to exert important immunosuppressive functions [33, 34]. Additionally, a special flow cytometric gating strategy has been developed. Using patient-specific blast-staining antibodies together with some unique DC-staining antibodies, some unique antigens that are not expressed on leukemic blasts before DCleu generation can be detected [34, 35]. After blast/DC populations are cultured in DC-generating media, the blast/DC population can be further divided into different subpopulations, such as leukemia-derived DCs, nonleukemia-derived DCs and nonconverted blasts [34]. It has been demonstrated that only mature DCleu can activate immune reactive cells. These mature DCleu express chemokine receptor 7 (CCR7), which is crucial for the migratory capacity of DCleu [12, 36]. Furthermore, mature DCleu also expresses CD83 and secrete IL-12 [37].

Both myeloid leukemia cells and DCs are derived from myeloid progenitor cells, but they have completely different characteristics and T cell stimulation functions. The majority of leukemia cells do not express CD80, only express low levels of CD86 [38, 39] and have poor T cell stimulatory capacity or even induce T cell anergy [40, 41]. Myeloid leukemia cells can be induced to differentiate into DCleu [42] by using various stimulants, such as GM-CSF plus IL-4 with either TNF-α or CD40 ligand (CD40L) [91].

In WT1 mRNA-electroporated DC vaccination clinical trials in AML patients, the OS rates and the WT1-specific CD8+ T cell response were improved significantly [92]. In 30 AML patients with a very high risk of relapse, 13 patients demonstrated an obvious antileukemic response. Nine patients achieved molecular remission and 5 of them sustained this remission after a median follow-up of 109.4 months [92]. Furthermore, patients with different tumors including leukemia, brain tumors, prostate cancer, renal cell carcinoma, pancreatic cancer, as well as HIV infection have been treated with ex vivo-generated mRNA-transfected DCs [93]. In AML and chronic myeloid leukemia (CML) patients, both autologous and allogeneic DCs have been administered as cellular therapy [94]. However, the immunogenicity of leukemia cell vaccines can be limited by many factors [42], and new strategies are urgently needed to induce a potent antileukemic immune response.

DC vaccines in nonleukemia malignancies

Melanoma

Melanoma is another malignant tumor in addition to AML and MDS in which DC vaccines have been widely studied. In the past few decades, great progress has been made in the clinical application of DC vaccines loaded with personalized neoantigens, which have been proven to be safe, immunogenic and feasible treatment strategies in patients with melanoma. With the increasing in-depth understanding of DC biology, the next generation of highly efficient cancer vaccines may provide a new immunotherapy strategy for patients with melanoma [95]. Mo-DC vaccines loaded with tumor lysates can affect the tumor microenvironment (TME) and promote the transformation of a “cold” tumor into a “hot” tumor by inducing the activation and infiltration of CD8+ T lymphocytes and the upregulation of PD-L1 expression in patients with metastatic melanoma [96]. The combination of a DC vaccine with immune checkpoint inhibitors (ICIs) has been shown to be effective in treatment of melanoma patients. Even after recurrence in patients who received adjuvant DC vaccination, treatment with first- or second-line PD-1 inhibitor monotherapy resulted in a response rate of 52% [97]. A clinical study demonstrated complete and long-lasting clinical responses in patients with immune checkpoint inhibitor-resistant, metastatic melanoma treated with adoptive T cell transfer combined with DC vaccination, with clinical responses induced by tumor-infiltrating lymphocyte (TIL) therapy combined with DC vaccination seen in 4 out of 4 treated metastatic melanoma patients who previously failed ICI therapy [98]. Another study showed that although more patients showed a clinical response to TIL + DC therapy, the combination of TILs and DCs showed no difference in the persistence of MART-1 TILs compared with TIL therapy alone [99]. A study revealed that blockade of inducible costimulatory molecule ligand (ICOSL) on DCs reduced priming of antigen-specific CD8+ and CD4+ T cells from naïve donors in vitro and that dysregulated NF-κB-dependent ICOSL expression in human DC vaccines impaired T cell responses in patients with melanoma, which supports the implementation of targeted strategies to augment these pathways for improved immunotherapeutic outcomes in patients with cancer [100].

A phase II clinical trial of DC vaccines for patients with stage III/IV melanoma demonstrated improvement in patient survival over the course of a year. However, patients in the DC vaccine treatment group had a higher rate of early local regional relapse than those in the control group, and 80% of patients reported swelling and erythema at the site of intradermal DC-injection [101]. Another study showed that DC vaccines in combination with cisplatin in stage III and IV melanoma patients did not improve clinical outcomes compared to DC vaccination monotherapy [102].

More clinical trials are currently underway. As of the end of November 2021, there were 19 ongoing clinical trials with DC vaccines in patients with melanoma (Table 3). A DC vaccine with natural myeloid DCs loaded with synthetic peptides is currently in a clinical trial for the treatment of stage IIIB and IIIC melanoma patients (NCT02993315). In addition, a personalized vaccine including autologous DCs exposed to autologous whole tumor cell lysate in combination with the chemotherapy drug cyclophosphamide has been explored to treat advanced solid tumor patients with high tumor mutation burden in a phase I clinical trial (NCT03671720).

Glioma/glioblastoma

As the most frequent and aggressive malignant primary brain tumor, glioblastoma multiforme (GBM) has a highly fatal prognosis and disease recurrence is universal. There is no effective therapy for recurrent disease, and the median survival after relapse is 6.2 months [103]. Animal studies on glioblastoma have demonstrated that DC vaccines can reduce tumor growth, prolong survival, and induce tumor-specific IFN-γ and cytotoxic T-lymphocyte (CTL) responses associated with T cell infiltration of tumors [103]. Numerous clinical trial studies have been initiated in GBM patients and have confirmed the feasibility and safety. Many of these studies reported induction of an antitumoral immune response and indicated improved survival after DC vaccine [104,105,106,107,108]. In a large phase III clinical trial of an autologous DC vaccine in newly diagnosed glioblastoma, the median OS was 23.1 months in patients who underwent surgery and DC vaccination, and vaccination-related grade 3 or 4 AEs were observed in only 2.1% of patients [104]. In another study of vaccination with DCs loaded with TAAs and/or mRNA of neoantigens in combination with low-dose cyclophosphamide in glioma patients, vaccination with DCs loaded with TAAs and the mRNA of neoantigens increased the life expectancy of patients. The median OS was 19 months and no grade 3 or higher AEs were observed [105]. A phase II clinical trial of alpha-type-1 polarized DC-based vaccination in newly diagnosed high-grade glioma revealed a significant survival-prolonging effect in DC-treated glioma patients. Ten of 15 evaluable patients showed positive CTL responses. After 6 years of observation, five patients were still alive, and two of these patients were relapse-free [106]. However, two publications of meta-analyses of randomized controlled studies on the efficacy of DC vaccines for newly diagnosed glioblastoma suggested that dendritic cell vaccines provide no obvious benefits for newly diagnosed glioblastoma [107, 108]. As of the end of November 2021, there were 15 ongoing clinical trials regarding DC vaccines in patients with glioma/glioblastoma (Table 3).

Lung cancers

Lung cancer is a common malignant tumor that threatens human life and is associated with high morbidity and mortality rates. Calreticulin (CALR) is a characteristic antigen involved in immunogenic cell death in non-small-cell lung cancer (NSCLC). A recent study showed that the CALR-TLR4 complex inhibits NSCLC progression by regulating the migration and maturation of DCs, providing a theoretical basis and ideas for immunotherapy of NSCLC [109]. In a pilot clinical trial study with a personalized neoantigen pulsed DC vaccine for advanced lung cancer (NCT02956551), the objective effectiveness rate was 25%, the disease control rate was 75%, the median progression-free survival (PFS) was 5.5 months, and the median OS was 7.9 months [110]. A randomized-controlled phase II trial of salvage chemotherapy after immunization with a TP53-transfected DC-based vaccine (Ad.p53-DC) in patients with recurrent small-cell lung cancer (SCLC) revealed that the vaccine was safe, with mostly grade 1/2 toxicities and some grade 3 toxicities. The rate of positive immune responses were between 20 and 43.3% in different experimental arms. Although the vaccine failed to improve ORRs to second-line chemotherapy, its safety profile and therapeutic immune potential remain [111]. Small-scale manufacturing of neoantigen-encoding messenger RNA for early-phase clinical trials in lung cancer patients (NCT04078269) has been successfully applied for the clinical evaluation of MIDRIXNEO, a personalized mRNA-loaded dendritic cell vaccine targeting tumor neoantigens [112]. As of the end of November 2021, there were 8 ongoing clinical trials with DC vaccines in patients with lung cancers (Table 3).

Furthermore, a preclinical study on the effect of a DC vaccine loaded with tumor cell lysate (TCL-DCV) on the percentage of CD166+ cancer stem cells (CSCs) in the lungs of mice exposed to benzo(a)pyrene (BP) revealed that TCL-DCV reversed the tumorigenic effect of BP in the lungs. Compared to cisplatin, TCL-DCV significantly decreased the percentage of CD166+ CSCs, suggesting its potential as a cure for lung cancer [113]. A new PD-1-blocking nanobody (PD-1 Nb20) in combination with tumor-specific DC/tumor cell-fusions augments the broad spectrum of antitumor activity of CD8+ T cells, providing an alternative and promising immunotherapeutic strategy for tumor patients who have T cell-dysfunctional or no sensitivity to anti-PD-1 therapy [114].

Prostate cancer

Patients with high-risk prostate cancer can experience relapse and develop noncurative disease. Vaccines targeting TAAs or tumor-specific antigens have been applied in clinical trials as prostate cancer treatment. Different types of vaccines including DC-based (e.g., sipuleucel-T), peptide/protein-based, or gene-(DNA/RNA) based vaccines have been applied as adjuvant therapy in patients with prostate cancers [115]. Despite the initial success with sipuleucel-T, further DC vaccines have failed to progress. Emerging antigen loading and presentation technologies, such as nanoparticles, antibody-antigen conjugates and virus codelivery systems have been used to improve efficacy [116]. In a phase I trial of an antigen-targeted autologous DC-based vaccine with in vivo activation of inducible CD40 for advanced prostate cancer, immune upregulation and antitumor activity were observed, as were prostate-specific antigen decreases, objective tumor regression and robust efficacy of posttrial therapy [117]. Recently, a long-term first-in-human phase I/II study of an adjuvant DC vaccine in patients with high-risk prostate cancer after radical prostatectomy showed promising results. Among 12 patients with grade 5 prostate cancers, five achieved remission after 84 months, and all mounted immune responses [118]. As of the end of November 2021, there were 5 ongoing clinical trials with DC vaccines in patients with prostate cancers (Table 3).

Lymphoma

DC vaccine immunotherapy has been used in patients with lymphoma for a long time. The 15-year follow-up of relapsed indolent non-Hodgkin lymphoma patients vaccinated with tumor-loaded DCs demonstrated the absence of toxicity and benefit of active immunization. The results showed that the 5-year and 10-year PFS rates were 55.6% and 33.3%, respectively; 10-year OS rate was 83.3%. Female patients experienced a better PFS and 22% of patients experienced a long-lasting complete response. Different genes including KIT, ATG12, TNFRSF10C, PBK, ITGA2, GATA3, CLU, NCAM1, SYT17 and LTK were differentially expressed in responding tumor patients [119]. The induction of an immune response after allogeneic WT1 DC vaccination and donor lymphocyte infusion (DLI) in patients with hematologic malignancies and posttransplantation relapse demonstrated that vaccines could be successfully produced from samples from all donors. DC vaccination and DLI are well tolerated, and DC vaccines can be used to sensitize the repopulated allogeneic donor immune system to WT1 [120].

In a phase I clinical study in patients with follicular lymphoma (FL), Mo-DCs generated in the presence of IFN-α and GM-CSF (IFN-DC) in combination with low doses of rituximab were administered intranodally. The results indicated that IFN-DCs can synergize with rituximab leading to increased cytotoxicity and T cell tumor infiltration. The overall response rate was 50% and the PET-negative complete response rate was 37%. No grade 3 or higher AEs were observed [121]. As of the end of November 2021, there were 4 ongoing clinical trials with DC vaccines in patients with lymphoma, including intranodally or intratumorally administered vaccines, combined with immunotherapy, radiation or cryosurgery (Table 3).

Methods for vaccine delivery

A variety of methods to deliver DC-based vaccines to patients have been used, such as intravenous [48, 122, 123], intradermal [123,124,125] and less frequently intranodal [48, 126] and intratumoral routes [127,128,129,130], as well as in vivo DC induction [12] (Table 4). Currently, there is no consensus as to which route of dendritic cell administration is the best for effectively sensitizing T cells. While antigen-loaded DCs can prime T cell immunity regardless of the route, the quality of responses and induction of antigen-specific antibodies may be different depending upon the route of administration [128]. Intravenous administration of antigen-pulsed DCs and subcutaneous administration of immature DCs have been demonstrated to be effective methods for generating sensitized T cells [131, 132]. In a mouse model, local carbon-ion radiotherapy combined with IV DCs augmented the immunogenicity of tumor cells and the maturation of DCs to stimulate antitumor immunity to decrease lung metastases [133]. In a phase I study using a 3 + 3 dose escalation design, the immunogenicity and efficacy of an intravenous DC-targeted liposomal vaccine in twelve patients with metastatic cutaneous melanoma showed that the DC vaccine was well tolerated and did not induce clinically significant toxicity. Partial response and stable disease were observed in one and two patients, respectively [134].

Table 4 Routes of dendritic cell vaccine administration
Full size table

Theoretically, intranodal administration of DCs may be the best route since DC migration is not required. In mouse models, compared to subcutaneous or intravenous immunization, intranodal injection of peptide-pulsed DCs induced significantly greater expansion of antigen-specific T lymphocytes in the spleen and a stronger antigen-specific Th1-type response. Thus, intranodal administration was an effective and feasible method for DC vaccination [94]. It has been reported that intranodal vaccination with semimature DCs primed strong, long-lasting CD4 T cell responses with a Th1-type cytokine profile in advanced melanoma patients [135]. All 5 metastatic melanoma patients in a tumor peptide-based DC vaccination trial developed strong and long-lasting delayed-type hypersensitivity reactivity correlated with the induction of CD4 T cell proliferation in vitro. In vitro stimulation results showed a significant increase in interferon-γ and IL-2 but not IL-4, IL-5, or IL-10 secretion by bulk T cells [135]. Similarly, intranodal administration of adenovirus encoding chimeric CD154 for CML [136], a tolerogenic DC-based vaccine for multiple sclerosis [137] and neoantigen peptide-loaded DC vaccines for ovarian cancer [138] has been reported. The results of a first-in-human phase I trial of intranodal direct injection of adenovirus expressing a chimeric CD154 molecule in fifteen patients with chronic lymphocytic leukemia (CLL) have been reported. The results showed that preliminary clinical responses, including reductions in leukemia cell counts, lymphadenopathy, and splenomegaly were observed. Six patients did not require additional therapy for more than 6 months, and three achieved a partial remission. These results provide rationale for phase II studies in patients with CLL, lymphomas, and CD40-expressing solid tumors [136]. A harmonized study protocol for two phase I clinical trials comparing intradermal and intranodal cell administration has been established, and clinical trials are underway [137]. Another trial demonstrated the clinical and immunological effects of neoantigen peptide-loaded DC-based immunotherapy in a patient with recurrent and chemoresistant ovarian cancer: following four rounds of vaccination with this therapy, CA-125 levels were remarkably decreased, tumor cells in ascites were decreased, and tumor-related symptoms such as respiratory discomfort improved without any adverse reactions [138]. However, studies on intranodal cell administration in AML/MDS patients are lacking.

Intratumoral administration of vaccines has been successfully used in solid tumors, such as breast, ovarian, lung, colorectal and renal cell carcinomas and melanoma. Intratumorally administration of TAAs in combination with immunostimulatory agents was able to activate tumor-infiltrating DCs and induce strong immune responses, resulting in tumor shrinkage or remission [139]. Based on this concept and these research results, TVEC has been approved by the US FDA for clinical use in patients with advanced malignant melanoma. While intratumoral administration of immunostimulatory agents and noncoding RNAs in solid tumors is a plausible method [140] and may remodel tumor metabolism and the immune microenvironment [147].

Another strategy to enhance DC vaccine efficacy is to combine DCs with immunomodulators targeting regulatory immune cells to overcome the immunosuppressive mechanisms of leukemia cells [42]. Previous studies have demonstrated that whole leukemia cell vaccines are suppressed by various immunosuppressive mechanisms of leukemia, including B7/cytotoxic T lymphocyte-associated protein 4 (CTLA4) and PD-1/PD-L1, Tregs, and myeloid-derived suppressor cells (MDSCs) [42, 148]. Therefore, blockade of the CTLA4 and PD-1 pathways could be used in combination with whole leukemia cell vaccines [42]. In a recent study, it was observed in a mouse model that T cell exhaustion in Langerhans cell histiocytosis was overcome with PD-1 blockade and targeted MAPK inhibition. The combination of a MAPK inhibitor and anti-PD-1 treatment significantly decreased both CD8+ T cells and myeloid Langerhans cells in a synergistic fashion. These results indicate that combined MAPK and checkpoint inhibition is a potential therapeutic strategy [149]. Therapeutic antibodies blocking the PD-1 pathway have been widely used in solid tumors [150]. DC/PD-1 immunotherapy combinations are currently under preclinical and clinical investigation in recurrent advanced brain tumors, advanced and relapsed NSCLC, MM and advanced renal cell carcinoma [151]. Additionally, various other combination therapies that exploit alternative immune targets and other therapeutic modalities have been explored for cancer treatment [152,153,154], especially with the development of a new generation of immune checkpoint inhibitors and other inhibitory targets [155,156,157,158,159,160,161]. The combination of a DC/AML cell fusion vaccine and checkpoint blocking therapy provides unique synergy to induce durable activation of leukemia specific immunity, protect against lethal tumor challenges, and selectively amplify tumor-reactive clones [162]. Different combinations designed to activate the endogenous T cell response through checkpoint blockade appear suitable and are being increasingly tested [163].

In some nonhematological malignancies, DC vaccines have shown an increase in the survival rate of patients in the late stage of tumor disease and have had a significant impact on the destruction of distant metastases. The combination of DC vaccines with another immunotherapy or traditional anticancer methods can be used in treating patients in the early stage of disease or preventing recurrence and metastasis [164].

Combining cancer vaccines with immunomodulatory drugs is currently regarded as a highly promising approach for boosting tumor-specific T cell immunity and eradicating residual malignant cells. Recently, a study of the FL mouse model with a new vaccine including IFN-DCs loaded with apoptotic lymphoma cells demonstrated that lenalidomide improves the therapeutic effect of an interferon-α-dendritic cell-based lymphoma vaccine, which may lead to evaluation of the combination in clinical applications [165].

Challenges of DC vaccines

Although much progresses has been made in the field of DC vaccines, there are several challenges to the wider application of leukemic DC vaccines. In previous trials, failure to generate sufficient numbers of qualified AML-DCs was the most common reason for excluding patients from the study [46]. The high cost of the stimulants required to differentiate leukemic DCs was another challenge [42].

A critical lesson learned, however, is the insufficient therapeutic efficacy of vaccination using genetically modified GM-CSF-secreting leukemia DCs, which is probably due to the immunosuppressive effects of GM-CSF [166]. The lack of immunogenicity of the whole leukemia cell vaccine may be due to the immunosuppressive effect of phosphatidylserine (PS) exposure to inactivated immune responsive T cells [167]. However, studies have shown that inhibition of the PS recognition process increases the immunogenicity of irradiated lymphoma cells, suggesting that endogenous adjuvants combined with dying tumor cells can be used to target tumor immune rejection [168].

Immunosuppressive factors from malignant cells can impede the functions of both DCs and T cells and hinder the vaccine-generated protective immune response. Defects in hematopoietic progenitor cells and an abnormal bone marrow niche render hematopoiesis seriously ineffective in leukemia patients. These factors bring additional challenges and accentuate systemic immunosuppression and DC malfunction [66]. Overall, the objective response rate (ORR) of DC vaccine was reported to be approximately 15% [122]; its therapeutic efficacy has to be improved.

Different mechanisms of weak immunogenicity of DCs have been reported, including failure to induce leukemia-specific CTLs [49], failure to activate NK cells or γδ T cells [65, 169, 170], failure to overcome the immunosuppressive action of Tregs and MDSCs [171], and undesired immune effects from Tregs and MDSCs [172]. Therefore, novel strategies, including DC vaccines combined with HMAs [146], NK cell infusion or immune checkpoint blockade therapy in relapsed/refractory AML and high-risk MDS patients need to be further validated.

Conclusions and future directions

Many difficulties remain and prevent the widespread application of DC vaccine immunotherapy. These limitations include weak cellular immune responses, high costs and time-consuming processes [13]. Despite the limited clinical efficacy, DC vaccinations still constitute a promising new strategy for AML and MDS treatment, as well as the treatment of other malignancies [13, 66, 164] with the validated safety and feasibility [62]. In addition to DC vaccination combined with systemic monoclonal antibody and immune checkpoint blockades, incorporation of PD-1/PD-L signaling into DCs can enhance DC mediated activation of T/NK cells and prevent Tregs stimulation [150, 173]. With the increasing clinical efficacy and application of DC-based vaccine therapy in patients with solid tumors [174], there is increasing interest in combining DC vaccines with conventional therapy, such as HMAs in the frontline treatment of elderly AML patients [108]. A recent study demonstrated that transgelin-2 is an essential protein for cancer and immunity and can act as a double-edged sword for cancer therapy. Engineering and clinical application of this protein may lead to a new era of DC-based cancer immunotherapy [175]. We believe that use of a carefully designed personalized DC vaccine in combination with other appropriate treatment strategies may constitute a valuable option for future treatment of patients with AML/MDS and other malignancies.