Antigen-Specific Polyclonal Cytotoxic T Lymphocytes Induced by Fusions of Dendritic Cells and Tumor Cells

The aim of cancer vaccines is induction of tumor-specific cytotoxic T lymphocytes (CTLs) that can reduce the tumor mass. Dendritic cells (DCs) are potent antigen-presenting cells and play a central role in the initiation and regulation of primary immune responses. Thus, DCs-based vaccination represents a potentially powerful strategy for induction of antigen-specific CTLs. Fusions of DCs and whole tumor cells represent an alternative approach to deliver, process, and subsequently present a broad spectrum of antigens, including those known and unidentified, in the context of costimulatory molecules. Once DCs/tumor fusions have been infused back into patient, they migrate to secondary lymphoid organs, where the generation of antigen-specific polyclonal CTL responses occurs. We will discuss perspectives for future development of DCs/tumor fusions for CTL induction.


Introduction
A major area of investigation in cancer vaccines involves the design of dendritic cells-(DCs-) based cancer vaccines. DCs can be distinguished from B lymphocytes and macrophages by their abundant expression of costimulatory molecules and efficient ability to prime both CD4+ helper and CD8+ cytotoxic activities [1]. Exogenous antigens from tumor cells can be taken up by DCs and translocated to the cytoplasm, processed, and presented through endogenous pathway. Both immature and mature DCs are capable of processing and presenting MHC-peptide complexes to T cells. Mature DCs are significantly better at CTL induction due to higher expression of MHC and costimulatory molecules, while presentation of antigens by immature DCs, in the absence of proper costimulation, may lead to tolerance induction [2,3]. After antigens uptake and inflammatory stimulation, immature DCs in peripheral tissues undergo a maturation process characterized by the upregulation of costimulatory molecules. During this process, mature DCs migrate to the regional lymph nodes, where they present antigens to CD4+ and CD8+ T cells through MHC class I and II pathways [1][2][3].
Loading MHC class I and II molecules on the cell surface of DCs with peptides derived from defined tumor-associated antigens (TAAs) is the most commonly applicable strategy for DCs-based cancer vaccines. This strategy has some limitations: (1) a limited number of known tumor peptides available in many HLA contexts whose immunogenicity is uncertain and (2) the relatively rapid turnover of exogenous peptide-MHC complexes that results in comparatively low antigen-presentation by DCs. Although DCs pulsed with antigen-specific peptides have been used in clinical trials for cancer patients, clinical responses have been found in a small number of patients [4,5]. Another strategies have been developed to load DCs with TAAs, including tumor RNA, tumor lysates, and dying tumor cells to induce antigen-specific CTL responses [6][7][8][9][10]. DCs pulsed with 2 Journal of Biomedicine and Biotechnology apoptotic tumor cell fragments or tumor lysates rely on antigen being cross-presented, all of which are usually not efficient [11]. An alternative strategy for inducing efficient CTL responses is the use of fusion cells generated by fusing DCs and tumor cells by polyethylene glycol (PEG) known as a chemical membrane destabilizing agent [12]. In this approach, multiple TAAs, including both known and unidentified, are delivered to DCs, endogenously processed and presented through MHC class I and II pathways in the context of the potent immune-stimulatory machinery of the DCs [13][14][15].

DCs/Tumor Fusions Approach
The chemical agent PEG [12], electroporation [16], and many viruses [17] have been used for the cell fusion strategy. We have used PEG to generate fusions of DCs and tumor cells. In our approach, DCs are usually mixed with tumor cells at a ratio of 10 : 1 in serum-free prewarmed RPMI 1640 medium. After centrifuge, mixed cell pellets are gently resuspended in prewarmed 50% PEG solution (molecular weight = 1,450)/DMSO solution, Sigma-Aldrich, St. Louis, MO; 1 mL per 5 × 10 6 cells) for 3 to 5 minutes at room temperature. Subsequently, the PEG solutions are diluted by slow addition and mixed with 1, 2, 4, 8, and 16 mL of serum-free prewarmed RPMI medium until 50 mL. The cell pellets are resuspended in prewarmed RPMI 1640 supplemented with 10% autologous heat-inactivated serum, GM-CSF (1000 units/mL), and IL-4 (500 units/mL) and cultured in a 5% CO 2 atmosphere at 37 • C for 3 days. The DCs/tumor fusions cannot proliferate but alive until 5 to 7 days after fusion (our unpublished data). Therefore, we have usually cultured fusion cells for 3 days after PEG treatment. After 3 days of culture, DCs/tumor fusion preparations are integrated into a single entity and are loosely adherent to the culture dish. Unfused tumor cells grow firmly attaching to the plates, whereas DCs/tumor fusions are loosely adherent in the culture wells. DCs/tumor fusions can be selected and purified by gentle pipetting, and firmly attached tumor cells are discarded. As this fusion procedure delivers not only the TAAs-epitopes but also the genes encoding the TAAs, DCs/tumors can continue to produce TAAs for several days after fusion [18]. Because fusion efficiency is closely correlated with antitumor immune responses in mice study, DCs/tumor fusions have been harvested on 3 days after fusion process to induce CTL responses in vitro.
There are some methods for enrichment of DCs/tumor fusion cells by FACS cell sorting [19] and the transfection of DCs with Tyr-green fluorescent protein reporter virus [20]. On the other hand, short-term culture of fusion cell preparations can also promote DCs/tumor fusion efficiency and reduce cell aggregates [18]. Therefore, it is not necessary to enrich DCs/tumor fusion cell preparations using special methods. In clinical trials and animal studies, fusion cell preparations, including fused cells, unfused DCs, and unfused tumor cells, have been irradiated before vaccination [12,18,21,22].

Characterization of DCs/Tumor Fusions
Fusions of DCs and whole tumor cells have the essential elements for processing and presenting TAAs to host immune cells and inducing effective antitumor immune responses ( Figure 1). It has been known that there are two different pathways for antigen presentation by DCs. Endogenously synthesized proteins, such as those in viral infections, and certain exogenous antigens are processed and presented through the MHC class I-restricted pathway to CD8+ T cells [1][2][3]. In contrast, exogenous antigens are processed through endogenous pathway, a phenomenon called cross-antigen presentation and displayed in association with MHC class II molecules and recognized by CD4+ T cells [1][2][3]. Importantly, the fusion process facilitates the introduction of tumor antigen to endogenous antigen processing pathway [23]. It has been reported that DCs/tumor fusions have the ability to process and present intracellular proteins derived from tumor cells [24]. Advantage of DCs/tumor fusions strategy over DCs pulsing with tumor lysates is that endogenously synthesized antigens have better access to MHC class I pathway [25]. Indeed, DCs/tumor fusions vaccine is superior to those involving other methods of DCs loading with antigenic proteins, peptides, tumor cell lysates, or irradiated tumor cells [26].
The membranes of DCs/tumor fusions are integrated into a single cell, whereas the nuclei remain separate in the primary hybrid cells [27]. Therefore, the important advantage in DCs/tumor fusions strategy is that modifications of DCs and tumor cells are independently possible while their characters persist after the fusion. This is an important difference between the DCs/tumor fusions and DCs pulsed with whole tumor cells. The fusions deliver not only the TAAs but also the mRNA encoding the TAAs. It has been reported that there are differences in the translation efficiency of mRNA to protein between immature DCs and mature DCs [15]. Stimulation of DCs by Toll-like receptor (TLR) signaling results in increased expression of peptide/MHC class I and II complexes, costimulatory molecules (CD80 and CD86), and IL-12 [28]. Recent studies have shown that cross-presentation is based on the transfer of proteasome substrates that are transcriptionally upregulated by heat-treatment of tumor cells [29,30]. Therefore, we have generated fusions of TLR-stimulated DCs and heat-treated tumor cells to induce efficient CTL induction ( Figure 2). This modified DCs/tumor fusions have been more potent than conventional fusions generated with immature DCs and unheated tumor cells, as demonstrated by (1) upregulation of heat-shock proteins (HSPs), MHC class I and II, TAAs, CD80, CD86, CD83, and IL-12; (2) activation of CD4+ and CD8+ T cells able to produce IFNγ at higher levels; and (3) potent induction of antigenspecific CTL responses [30]. Synergism between TLRstimulated DCs and heat-treated tumor cells may enhance the immunogenicity of DCs/tumor fusions and may provide promising means of inducing therapeutic CD4+ and CD8+ T cells.

How to Induce CTL Responses by DCs/Tumor Fusions
The DCs/tumor fusions possess the properties of both parent cells. Fusions of DCs and tumor cells result in the formation of a heterokaryon that combines DC-derived MHC class I and II, costimulatory molecules, efficient antigen-processing and -presentation machinery, and abundant tumor-derived MHC class I and TAAs (Figure 1) [12]. The DCs/tumor fusions approach facilitates the entry of TAAs that are synthesized de novo in the fusions. As a result, whole TAAs can be processed and presented through both MHC class I and II pathways on the DC compartment in the context of costimulatory molecules [12]. The advantage of DCs/tumor fusions over DCs loading with tumor lysates or peptide is that endogeneously synthesized TAAs have better access to MHC class I and II pathways [25]. Effective TAAs processing and presentation in DCs/tumor fusions are crucial for potent CTL induction. MHC class I and II molecules are important for antigen processing and presentation and subsequent activation of CD8+ and CD4+ T cells, respectively. To investigate the role of MHC class I-and II-restricted antigen presentation and the activation of CD4+ and CD8+ T cells by DCs/tumor fusion cells, we have created four types of DCs/tumor fusions with intact or deficient expressions of MHC class I and/or II molecules by using several kinds of DCs from MHC class I and/or II knockout mice and tumors: (1) wild type fusions (WT-FCs), (2) MHC class I knockout fusions (IKO-FCs), (3) MHC class II knockout fusions (IIKO-FCs), and (4) MHC class I and II knockout fusions (I/IIKO-FCs) [77]. We observed differential impairment of antitumor immunity induced by fusions generated with DCs from MHC class I and/or II knockout mice. Immunization of mice with IKO-FCs resulted in slightly decreased capacity for CTL induction, tumor prevention, and tumor treatment, compared with immunization with WT-FCs. In contrast, immunization with IIKO-FCs abolished IFN-γ production of CD4+ and CD8+ T cells and the induction of CTLs and exhibited severely impaired antitumor immunity. Therefore, MHC class II antigen presentation targeting activation of CD4+ T cells may be indispensable in antitumor immunity. The presentation of TAAs on MHC class II is essential for the activation of CD4+ T cells and the induction of efficient CD8+ CTL responses. It has been demonstrated that predominant effector cells are CD8+ CTLs, as most tumors express only MHC class I molecules. However, there are increasing evidences that CD4+ T cells play a more direct role beyond delivery of assistance in the generation of efficient antitumor immunity [78]. Unlike CD8+ T cells, CD4+ T cells may contribute to antitumor immunity through diverse mechanisms. It has been well documented that CD4+ T cells provide help to CD8+ T cells by activating APCs through CD40-CD40L interaction [79] and/or IL-2 production [80]. In addition to providing help in the priming phase, CD4+ T cells are also essential in the effector phase [81], in which they are required for the maintenance of CD8+ CTLs in vivo and the infiltration of CTLs at the tumor site [82]. Importantly, it has been also reported that adoptive cell transfer of antigenspecific CD4+ T cells can control tumor growth [83].

Stimulation of CD4+ and CD8+ T Cells by DCs/Tumor Fusions
Efficient CTL induction requires the stimulation of both CD4+ and CD8+ T cells [84]. We have previously reported that DCs/tumor fusions migrated to the regional lymph nodes, where they were closely associated with CD4+ and CD8+ T cells. Moreover, DCs/tumor fusions localized to the T cell area in the regional lymph nodes and formed clusters with CD4+ and CD8+ T cells [34]. Simultaneous recognition of cognate peptides presented by MHC class I and II molecules on DCs/tumor fusions is essential in the induction of efficient CTLs [84,85]. DCs/tumor fusions, unlike DCs, do not have to take up exogenous TAAs. Moreover, DCs or Langerhans cells in the patient body can intake the degraded DCs/tumor fusions that have been vaccinated. Therefore, both direct antigen-presentation by DCs/tumor fusions and cross-presentation by DCs in the patients with cancer can participate in CD4+ and CD8+ T cell-activation [14,45,84]. It has been well known that mature DCs are significantly better at CTL induction due to higher expression of MHC and costimulatory molecules, while presentation of TAAs by immature DCs, in the absence of proper costimulation, may lead to tolerance induction [86]. Recent studies indicate that TLRs directly regulate both cancer immunity and tolerance through innate immune responses mediated by regulatory T cells (Tregs), DCs, and other immune cells [87]. Coadministration of TLR ligands with DCs/tumor fusion-based vaccine regulates the function of Tregs and DCs through some mechanisms: (1) stimulation of fusions by TLR signaling results in the increased expres-sion of peptide/MHC class I and II complexes, costimulatory molecules (CD80 and CD86), and cytokines (IL-12) [18], and (2) TLR signaling activation on fusions can render naive T cells refractory to suppression mediated by Tregs [88]. TLR ligands also activate DCs at the tumor site and enhance antigens cross-presentation, migration into regional lymph node, and induction of antigen-specific CTL responses [87,89]. Importantly, TLR ligands prevent the death of activation-induced CTLs by increasing the expression of antiapoptotic mediators (Bcl-xL and c-FLIP), allowing these cells to survive and migrate into the tumor site [90]. Moreover, the apoptosis of DCs can potentially regulate DCs homeostasis and immune responses. It has been reported that Bim, a BH3 protein of the Bcl-2 family, is important for regulating spontaneous cell death of DCs [91]. Therefore, enhanced DCs/tumor fusions survival may play a role in efficient CTL inductions. Another effective adjuvants for enhancing the induction of CD4+ and CD8+ T cells are HSPs, to which the ability of heat-treated tumor cells to enhance immunogenicity has been attributed [92]. Recently it has been shown that cross-priming is based on the transfer of proteasome substrates that are transcriptionally upregulated by heattreatment of tumor cells [29]. This is potentially important in the rational design for cancer vaccines that elicit CD8+ CTL responses. This concept offers additional effects by which heat treatment of tumor cells might enhance antigen-processing and -presentation in MHC class I and II molecules on the surfaces of DCs/tumor fusions. Moreover, extracellular HSPs act as a chaperon and interact with DCs in a receptor-mediated manner, leading to DCs maturation as well as proinflammatory responses, all of which are essential for induction of effective CD4+ and CD8+ CTL responses. Recently, we have shown that heat shock protein 70-peptide complexes (HSP70.PCs) derived from the DCs/tumor fusions possess superior properties compared with HSP70.PCs from tumor cells in animal and human models [93,94]. Namely, HSP70.PCs derived from human DCs/tumor fusions induced T cells that expressed higher levels of IFN-γ and exhibited increased levels of killing of tumor cells, compared with those induced by HSP70.PCs derived from tumor cells. Enhanced immunogenicity of HSP70.PCs from DCs/tumor fusions was associated with improved composition of the vaccine, including increased content of TAAs and their processed intermediates, and the detection of other heat HSPs such as HSP90 and HSP110 [94].

Induction of Antigen-Specific Polyclonal CTL Responses Restricted by MHC Class I and II Molecules
Cancer vaccine approaches that rely on induction of immunity against particular antigens are potentially subject to tumor cell resistance mediated by the downregulation of the single antigen. Therefore, induction of polyclonal antigen-specific CTL responses may have the potential to maximize the protection against various subsets of tumor cells downregulated by certain tumor antigens, which may appear during the course of tumor progression. Fusions of DCs and autologous tumor cells are effective for induction of antigen-specific polyclonal CTL responses. How the fusions assemble and present the MHC class I-restricted multiple peptide complexes is unclear. One possibility is that multiple peptides are complexed with MHC class I molecules in tumor cells and the complexes are simply transferred to the fusion cells. Alternatively, the tumor cells may provide multiple tumor antigens and the fusion process simply facilitates the delivery of tumor antigens to the efficient antigen-processing and -presenting machinery in DCs. Then the tumor antigens are presented in the context of MHC class I and II molecules derived from DCs. We have demonstrated that fusions generated with allogeneic tumor cells and autologous DCs induced polyclonal antigen-specific CD4+ and CD8+ T cells restricted by autologous MHC class I and II molecules [95,96]. These results suggest that the TAAs from allogeneic tumor cells are processed along the endogenous pathway, through the antigen-processing machinery of autologous DCs. Indeed, stimulation of T cells with fusions resulted in MHC class I and II restricted CD4+ and CD8+ T cell proliferation, high levels of IFN-γ production in both CD4+ and CD8+ T cells, and the simultaneous induction of polyclonal antigen-specific CTL responses restricted by multiple HLA molecules [84,[95][96][97][98][99][100][101].

Generation of Tregs by DCs/Tumor Fusions
Immature DCs are one of the mediators of tolerance induction. In peripheral lymphoid organs immature DCs are incapable of eliciting CTL responses. In contrast, if a stimulus for DCs activation is sufficiently coadministered with antigens, mature DCs express high levels of costimulatory molecules, resulting in priming of antigen-specific CTL induction rather than Tregs [86]. There are increasing evidences that tumor-derived soluble factors such as TGF-β and IL-10 promote the induction of tolerance through the generation of CD4+ CD25high Foxp3+ Tregs, which may be linked to compromised immune responses in patients with cancer [102]. In animal models, tumor-derived TGF-β reduces the efficacy of DCs/tumor fusion vaccine via in vivo mechanisms [42]. Moreover, the blockade of tumor-derived TGF-β reduces Treg generation by the DCs/tumor fusions and enhances antitumor immunity [52]. Indeed, the potent immunosuppressive effects by Tregs can explain the failure of many immunotherapeutic approaches to patients with cancer [103]. In our study, fusions of immature DCs generated in the presence of tumor-derived culture supernatants promoted the generation of CD4+ CD25high Foxp3+ Tregs and inhibited CTL induction [88]. Recently, it has also been demonstrated that fusions in human are effective APCs to induce CTL responses but to stimulate inhibitory T cells that limit vaccine efficacy. However, exposure of DCs to TLR agonists, stimulatory cytokines, and anti-CD3/CD28 enhances vaccine efficacy by limiting the Treg responses and promoting expansion of CTLs in vitro [104]. Depletion of Tregs before vaccination may lead to enhanced CTL responses in cancer patients [103]. If the immune-suppressed environment in tumor is sufficiently improved, selective manipulation for the innate immune responses induced by TLR agonists may have more potential to promote DCs maturation and CTL over Treg generation [97]. As many tumors express or induce immunosuppressive cytokines such as IL-10 and TGF-β in tumor environment, patients early in the course of the disease are expected to respond best to clinical responses by fusion cell vaccination. A combination of control of Tregs and concomitant induction of CTLs may be a more effective immunotherapy to reduce recurrence and prolong survival after surgery.

Clinical Trials
Based on these unique features of DCs/tumor fusions, initial Phase I/II clinical trials have been conducted in a variety of tumors (Table 1). Fusion cell vaccination has been reported in patients with melanoma [98,99]. In the report, autologous melanoma cells were fused to allogeneic DCs by electrofusion and assessed them as a vaccine in 17 patients with disseminated melanoma refractory to standard therapy. One patient had a partial response with decrease in size of all evaluable tumor manifestations. In one patient, some of the metastases were regressed despite overall progressive disease, and one patient achieved disease stabilization for 6 months. There were no serious side effects associated with the administration of the vaccine. In our initial clinical trial, 8 patients with malignant glioma were vaccinated with fusions of autologous DCs and autologous tumor cells. In all 6 cases analyzed, the concentration of IFN-γ in the T cell culture supernatant increased after vaccination. However, clinical responses were not observed [101]. Therefore, we conducted a phase I/II clinical study for the safety profile of vaccination with DCs/tumor fusions combined with recombinant human (rh) IL-12 in patients with malignant brain tumor, melanoma, breast, gastric, colorectal, and ovarian cancer [100,105]. Eleven out of 15 patients with malignant brain tumor achieved a stable response and 24 patients had a progressive disease after 8 weeks of the initial treatment. The vaccine was well tolerated and no serious adverse effects were observed. In four patients, magnetic resonance imaging showed a greater than 50% reduction in tumor size. One patient had a mixed response. Therefore, coadministration of the fusions and rh IL-12 can induce more effective antitumor effects than fusions alone in some patients with malignant glioma. In this clinical trial, we vaccinated patients with extremely small amounts of fusion cells and rh IL-12 but resulted in immunological and clinical responses in a subset of patients with malignant brain tumor. The relatively favorable response to malignant brain tumors is of particular interesting since the central nervous system is generally considered to be an immunologically privileged site as a result of the lack of lymphatic drainage and the nature of the blood-brain barrier. However, these studies demonstrated that appropriately activated T cells could cross the blood-brain barrier to access the tumor burden and initiated tumor regression. On the other hand, In the clinical setting of the patients with cancer, a major difficulty for the fusions vaccine is the preparation of sufficient amounts of autologous tumor cells. The specimen of tumor from primary lesion may not provide sufficient numbers of viable tumor cells due to the length of culture time and potential contamination of bacteria and fungus. We have reported that fusions of autologous DCs and allogeneic tumor cell lines induced CTL responses against the autologous tumor cells [95,96]. The basis for using allogeneic tumor cell lines instead of autologous tumor cells is that some antigens are shared by most of tumors.

Future Directions
Although fusion cell vaccination is effective to induce CTL responses, we are searching for optimal strategy to benefit patients significantly. DCs/tumor fusion cell vaccine alone may be insufficient to have a significant contribution to treat advanced cancer patients with compromised immune system. To elicite polyclonal CTL induction in patients with cancer, it may be necessary to combine with chemotherapy, radiotherapy, hormonal therapy, or photodynamic therapy to reduce Treg and enhance CTL responses. Indeed, the combination therapy of fusion cell vaccination and adoptive immunotherapy is a very effective against poorly immunogenic carcinomas in murine studies [41,53,106]. Combination with these approaches has enormous potential to improve the current outcomes from conventional cancer therapy. The next decade will see the first clinical trials testing whether combination of induction of augmented CTLs and depletion of Tregs might be effective in treatment for cancer patients.