Dendritic cell biology and cancer therapy

Abstract

Dendritic cells (DCs) are nature’s best antigen-presenting cells. They possess attributes that allow them to effectively fulfill the requirements for priming/activating T cells and mediating tumor-specific immune responses. In this review, emphasis is placed on those aspects of DC biology that best illustrate their usefulness in immunotherapy of cancer. Culture, maturation, and polarization conditions for human DC are discussed, as are strategies for antigen-loading of DCs and for their delivery to patients with cancer. A concise recommendation for monitoring of DC-based vaccination trails is provided.

Introduction

Dendritic cells (DCs) are potent antigen-presenting cells (APCs) that play a key role in antitumor host responses. Much has been recently written about DCs, and several excellent reviews are available describing their biologic and immunologic properties [3, 5, 8, 21]. This minireview will briefly refer to those DC characteristics that have been instrumental in advancing the use of these cells for human therapy. It will largely focus on practical aspects of generating DCs for immunotherapy, as well as on the available evidence for clinical utility of DCs as components of antitumor vaccines. Many DC-based vaccination trials are in progress at this time, and a recently held meeting (the 4th International Expert Meeting on Clinical Dendritic Cell Immunotherapy, Amsterdam, 3–16 June 2003) has brought together investigators from all over the world to consider various aspects of the adoptive transfer of DCs to treat human diseases. Two issues appear to be critical for the development of “optimal” DC-based immunization strategies for patients with cancer. One deals with the ability of therapeutically used DCs to process and present to T cells a broad range of MHC class I– and class II–restricted epitopes in order to induce effective antigen-specific T-cell immunity. The other focuses on the in vivo expansion and maintenance of T_h1/T_c1-type responses (both CD8⁺ and CD4⁺ mediated), which appear to be necessary for tumor regression and establishment of long-term immunologic memory. In this minireview, we will summarize a collective experience currently available regarding DC generation for therapy of cancer and suggest new avenues of exploration aimed at the production of DCs, which are best able to induce and sustain antitumor activities in patients with cancer.

Biology of dendritic cells

The central role of DCs in the immune system has focused much attention on their biology, including their origin, differentiation, maturation, and functional characteristics. Perhaps the most important lesson emerging from these studies is that DCs are a heterogeneous group of cells. They contain multiple cell subsets that display differences in the cell surface phenotype as well as functions and localize to distinct anatomical sites. A brief survey of DC properties presented below is not intended to be comprehensive but rather to provide a framework for considering DCs as components of therapeutic strategies broadly used today for therapy of cancer and other diseases.

Origins of DCs

DCs originate from CD34⁺ bone marrow stem cells. Human DC precursors are found in the bone marrow as well as peripheral blood and, in a more mature form, in lymphoid and nonlymphoid tissues. Three different subtypes of DC have been defined: Langerhans cells, interstitial DCs, and plasmacytoid DCs. Human skin contains the first two types of DCs, which are generally referred to as myeloid DCs. The plasmacytoid cells are derived from the lymphoid lineage and are found in the T-cell areas of lymphoid organs, the thymus, and in the peripheral circulation [21].

Major DC subsets

DCs developing from a myeloid or lymphoid lineage are both phenotypically and functionally distinct. In addition, within the myeloid DC lineage system, different DC subsets have been identified, based on their functional properties. Several in vitro studies have shown that human CD14 (monocyte)-derived DCs (CD11c⁺) prime T cells to preferentially activate T_h1-type responses, and IL-12 is required for this process [12, 26]. CD14-derived DCs, but not CD1a-derived DCs, can also activate naïve B cells to secrete IgM in the presence of CD40L and IL-2 [7]. In contrast, human CD123⁺ plasmacytoid DCs generally induce T_h2-type responses. These intrinsic functional differences in DC subsets derived from the myeloid vs lymphoid lineage are underscored by their tissue localization. The myeloid pathway of differentiation gives rise to DCs that home to peripheral tissues, where they take up and process exogenous antigens and then migrate to the secondary lymphoid tissues to present these antigens to naïve T cells. On the other hand, thymic DCs are involved in the presentation of self-antigens to developing thymocytes and thus initiate the subsequent deletion of autoreactive T cells. DCs in both differentiation pathways remain highly susceptible to inflammatory cytokines, and environmental “polarization” can shape their behavior. This ability of DCs to adopt relatively stable polarized phenotypes characterized by preferential induction of T_h1 (cellular)- vs T_h2 (humoral)-type responses raises the possibility of using polarized DC subsets to induce “tailored” immune responses driven by a cocktail of selected cytokines.

DC phenotype

DCs express characteristic cellular markers, including substantially higher levels of antigen-presentating molecules (e.g., CD1a, MHC-I, and MHC-II) than any other cell in the body. In addition, DCs have a high surface density of accessory/costimulatory molecules, including leukocyte functional antigen (LFA)-3/CD58, B7-1/CD80, B7-2/CD86, CD40, and intracellular adhesion molecules (ICAMs; ICAM-1/CD54 and ICAM-3/CD50), which facilitate both the interaction with, and stimulation of, lymphocytes [35]. DCs also express a repertoire of receptors for efficient receptor-mediated antigen uptake, including FcγRII (CD32), FcγRI (CD64), FcΣRI, and the C3bi complement receptors (CD11b), which increase the efficiency of immune complex endocytosis. C-type lectin receptors, such as the macrophage mannose receptor and DEC-205, which bind bacterial carbohydrates, are also found on the surface of DCs [15, 30]. This repertoire of surface costimulatory molecules and receptors endow DCs with the capability to readily respond to changes in their milieu and to internalize a broad variety of infectious agents and other antigenic molecules.

DC maturation

DCs residing in tissues are in an “immature” state (iDCs) and are unable to stimulate naïve T cells. In vivo, DC maturation occurring in response to the microenvironmental signals allows DCs to switch their functional phenotype from antigen-processing cells when immature to T-cell stimulatory cells when mature. A similar maturation process can be induced in vitro using various exogenous stimuli, such as LPS, dsRNA, apoptotic cells, immune complexes, CpG DNA, proinflammatory cytokines (TNF-α) and prostaglandins (PGE₂) [16, 35]. The process of maturation is accompanied by a series of morphologic and functional alterations. Thus, the expression pattern of many surface markers and factors secreted by DCs changes, so that matured DCs (mDCs) acquire a phenotype distinct from that of iDCs. Matured DCs increase their level of membrane HLA-DR (MHC class II) and have significantly higher expression of accessory/costimulatory molecules (CD40, CD54, CD50, CD58, CD80, and CD86) than iDCs. They acquire expression of CD83 and CCR7 molecules. During maturation, DCs lose their ability to efficiently take up and process soluble antigens and, instead, acquire numerous processes (veils, dendrites), allowing mDCs to increase their motility and migrate toward the regional lymph nodes (LN). Matured DCs have a distinct cytokine and growth factor profile and are characterized by the increased ability to produce IL-12 in response to environmental stimuli, a cytokine necessary for the development of T_h1-type T-cell responses [31]. Also, the ability to produce angiogenic factors (bFGF, VEGF), which is a characteristic of iDCs, tends to decrease in mDCs (Odoux et al., submitted). However, the window for cytokine/growth factor production is narrow in DCs and tends to shift in the course of their maturation. Thus ex vivo manipulations to achieve optimal DC polarization are both time-dependent and regulated by the content of stimulatory cytokines. In vivo, it is almost certain that DCs must reach LNs and contact cognate antigen-specific T cells during the relatively short period of cytokine production. Therefore, tissue migratory properties are important features of in vivo maturing DCs.

Functions of DCs

The most widely emphasized functions of DCs are antigen uptake and processing. DCs are considered to be the nature’s best APCs. Immature DCs have several features that allow them to capture antigens. They can take up antigens via receptor- or nonreceptor-mediated mechanisms. Upon being internalized, tumor antigens are processed, split into peptides in the cytosol or endocytic vesicles in DC, which are then reexpressed on the cell surface in association with MHC molecules. To accomplish this complex series of events, DCs are equipped with a sophisticated molecular array of cell components representing the antigen-processing machinery (APM). The multiple APM components in human DCs can be identified and their expression quantified using mAbs, such as mAbs developed in Dr Soldano Ferrone’s laboratory (personal communication). The APM is essential for the uptake and processing by DCs of tumor-derived antigens, so that tumor-derived epitopes can be cross-presented to T cells.

Unlike other APCs, such as B cells, which only activate memory T cells, DCs are unique in their ability to stimulate naïve T cells. Antigens that are bound to MHC-I molecules and expressed on the surface of DCs, can be perceived and recognized by T cells equipped with a “relevant” T-cell receptor (TCR). This is how tumor antigen-specific cytotoxic T lymphocytes (CTLs) are generated. Antigens expressed on the DC cell surface and linked to MHC class II molecules interact with CD4 helper T cells (T_h). Both CD8⁺ CTL and CD4⁺ T_h cells are essential for the development of coordinated antitumor responses which lead to tumor regression [5, 23].

DC activation signals

DCs are sensitive to a wide range of stimuli that serve not only to promote immune responses via the release of chemokines and proinflammatory mediators, but also to trigger DC migration toward regional LNs to generate antigen-specific (adaptative) immunity. The pathogens, such as influenza virus, Escherichia coli, Candida albicans, or mycobacteria, represent one group of DC-activating signals [11, 14]. Lipopolysaccharide associated with bacterial cell walls activates DCs via CD14, acting in concert with Toll-like receptors (TLR) 2 and 4 [40].

A second group of DC-activating signals originates from distressed or dying cells. As a consequence of infection or tissue damage, cells release factors that can activate DCs present in the environment. Furthermore, DCs that encounter apoptotic tumor cells undergo maturation in vitro: this maturation involves autocrine/paracrine secretion of IL-1β and TNF-α and is thought to be mediated by αvβ5 integrin and CD36 [1].

One of the primary events that occur after DC activation is the increased production and release of chemokines, such as RANTES, monocyte chemotactic protein (MCP-1), macrophage inflammatory protein (MIP)-1α, and MIP-1β. Activated DCs are also a major source of many cytokines, namely, interferon α (IFN-α), IL-1, IL-6, IL-7, IL-12, and IL-15, and also produce macrophage inflammatory protein (MIP1γ), all of which are important in the elicitation of a primary immune responses [9, 22, 27]. Also, newer evidence indicates that the pattern of cytokine secretion of DCs derived from plastic-adherent monocytes and cultured in GM-CSF and IL-4 can be directed toward the T_h1 (IL-12) or T_h2 (IL-10) cytokine secretory pathway. As indicated above, the production of IL-12 is critical for the promotion of an effective cellular immune response by activated T lymphocytes, which upon interaction with DCs enter the T_h1 differentiation pathway. IL-12 production appears to be inhibited by various tumor-derived substances, including nitric oxide (NO), prostaglandin E₂ (PGE₂), IL-10, IFN-α, the p40 homodimer of IL-12, and transforming growth factor β (TGF-β), which is regarded predominantly as an immunosuppressive cytokine [27, 36]. DCs activated by LPS and CD40L also produce substantial quantities of angiogenic factors (Odoux et al., submitted).

DC migration

One result of DC maturation and activation is increase in their motility and their propensity to migrate toward lymphoid organs. DC migration is an essential feature for DCs which traffic from blood to tissues. While iDCs capture antigens, upon maturation they migrate from tissues to draining lymphoid organs, where, converted to mDCs, they prime naïve T cells. Chemokines play an important role in controlling DC traffic. Immature DCs respond to many CC and CXC chemokines (MIP-1α, MIP-1β, MIP-3α, MIP-5, MCP-3, MCP-4, RANTES, TECK, and SDF-1), which are inducible upon inflammatory stimuli, and for which each immature DC population displays a unique spectrum of chemokine receptors [8]. In contrast to iDCs, mDCs lose their responsiveness to most of these inflammatory chemokines, but acquire responsiveness to ECL/MIP-3β and SLC/6Ckine, as a consequence of the gradual up-regulation of the receptor CCR7 on DC cell surfaces during activation [8]. The ligands for this receptor, 6Ckine and MIP-3β, are produced primarily in the T-cell rich parafollicular areas of LNs [20]. DC migration mediated through chemokine gradients is of special interest in the scientific community. A better understanding of the regulation of DC trafficking might offer new opportunities for therapies able to differentially stimulate or suppress antitumor immune responses.

The properties of DCs discussed above qualify DCs as an excellent cell type for adoptive immunotherapy. Their biologic activities, including the ability to facilitate their own migration by producing angiogenic factors (Odoux et al., submitted), allow DCs to deliver tumor-specific epitopes to T cells found at various sites in the body and to orchestrate the generation of tumor-specific immunity.

DC-based immunotherapy of cancer

The concept of DC transfer

The goal of immunotherapies is to induce or up-regulate T cell–mediated tumor-specific immune responses. As discussed above, DCs have attributes that make them particularly attractive candidates for achieving this goal. Adoptive cell transfers with DCs have become a part of many therapeutic vaccination strategies in cancer as well as chronic infectious diseases in recent years. There are many good reasons for considering DC-based therapy in these diseases. The inability of the host’s immune response to cope with the tumor or viral antigen load either in advanced or recurrent/metastatic disease, represents an opportunity to introduce a cell type capable of restoring effective immune responses specific for the tumor or an infectious agent. The major objective of active specific immunotherapy, including antitumor vaccines, is to generate tumor-specific CTLs as well as helper T cells capable of targeting the tumor. To achieve this goal through an increase of the immunogenic potential of tumor-associated antigens (TAAs) or epitopes derived from these antigens and to provide adequate help in the form of epitope-specific CD4⁺ T cells, it appears necessary to employ DCs.

Various strategies have been used for eliciting robust antitumor T-cell responses in vitro and in vivo, including up-regulation of MHC molecules and increasing expression of cytokine genes or costimulatory signals on human tumors. The use of autologous ex vivo activated DCs pulsed with tumor antigens in expectation of effective delivery of these antigens to T cells has been particularly promising [4]. The rationale for this form of therapy is that tumors are neither good inducers of T-cell responses nor good targets for these responses. They often have defects or abnormalities in antigen-processing machinery [33], express little or no MHC or costimulatory molecules [24], and, through a broad range of ingenious escape strategies (reviewed in [39]), create the immunosuppressive microenvironment in which immune cells cannot function normally (hyporesponsiveness) or are eliminated by apoptosis [38]. It seems reasonable, therefore, that to effectively deal with these tumor-orchestrated suppressive mechanisms and to achieve robust and sustained antitumor T-cell responses, specifically targeted and particularly effective therapeutic approaches are necessary. The most potent T-cell stimulatory APCs are unquestionably DCs, which express a complete coterie of MHC as well as costimulatory molecules and secrete cytokines and factors necessary for driving T_h1-type cell responses [5]. Thus, the rationale for using DCs as a component of antitumor vaccines is compelling: DCs are not only capable of internalizing and processing tumor epitopes, of migrating to regional LNs and optimally presenting these epitopes to T cells, but also of secreting cytokines promoting amplification of T-cell responses as well as the development of immunologic memory. In effect, DCs are such potent APCs because of the adjuvant-like effects they mediate. Another compelling reason for adoptive transfer of cultured DCs as vaccines is provided by the recent discovery that DCs are often functionally compromised in patients with cancer [13]. Ex vivo differentiated autologous DCs obtained either from peripheral blood monocytes or bone marrow stores of patients with cancer are fully functional, however, and thus can be used in vaccines.

DC-based therapy in tumor-bearing animals

The ability of DCs pulsed in vitro with tumor-derived antigens and then transferred to tumor-bearing animals to induce tumor regression was initially evaluated in murine models. To date, DC-fed apoptotic tumor cells, fused with tumor cells or pulsed with lysates made from tumor cells, selected synthetic peptides, or tumor cell-derived RNA, have all been used to successfully immunize tumor-bearing animals against parental tumors. In animals with established tumors, therapy with DCs has led to regression of palpable tumors and the development of long-lived, protective antitumor immunity [10]. In animal models of cancer prevention, DC immunization was shown to induce long-lasting and protective antitumor immunity [25]. The collective experience has been that regardless of the presence or stage of tumor, the route of DC administration or the cell dose, delivery of DCs to tumor-bearing animals is safe, accompanied by minimal toxicity and not associated with development of autoimmune sequelae or an autoimmune disease. In vivo studies of DC migration with labeled DCs injected at different tissue sites indicated that DCs have the ability to reach the regional LNs and engage those antigen-specific T cells, which are poised to mediate immune responses [6]. In addition, the presence of matured DCs at tissue sites has been associated with the phenomenon of epitope-spreading, whereby the inflammatory milieu created by DCs is instrumental in activating T cells with a variety of specificities unrelated to epitopes being presented by the DCs. Overall, however, animal experiments in various tumor models have not provided a clear-cut set of guidelines for DC administration, and thus, routes of delivery, as well as DC numbers, dose, and potency, which are optimal for therapy, remain to be defined.

Culture of DCs for therapy

DCs can be easily obtained in large numbers from either peripheral blood or from bone marrow progenitors. In humans, immature DCs are most commonly generated by culture of peripheral blood monocytes in the presence of IL-4 and GM-CSF, according to a method described by Sallusto and Lanzavecchia [29]. Monocytes are relatively easy to scale up for clinical use: they can be purified from peripheral blood by a variety of methods, including immunoselection, counterflow elutriation, and adherence onto plastic. These are all methods effective in monocyte enrichment, but it is essential that endotoxin-free materials be used, because even a transient exposure of monocytes to endotoxin impairs subsequent IL-12 production, which is a key factor in DC functions [19]. A variety of culture methods for therapeutic-grade DCs have been introduced. The overall objective is to devise a GMP-compatible system for clinical applications that is simple, economical, and reliable in providing functionally active DCs for therapy. The approaches may vary from manually handled plastic flasks or bags to commercially available automated, closed systems such as an Aastrom Replicell System or Miltenyi’s Clinimax. While all these methods offer a possibility for generating large numbers of human DCs from leukapheresis products (e.g., starting with about 2×10⁹ PBMCs, from 5×10⁷ to 5×10⁸ DCs can be obtained), the quality of the final product, in addition to its being sterile and free of endotoxin and mycoplasma contamination, is important. The presence of lymphocytes in a final batch of DCs is common but not desirable, and release criteria established by any laboratory should specify the purity as well as potency of the final DC product.

Introduction of tumor antigens

Immature DCs from cultures of peripheral blood–derived monocytes in the presence of IL-4 and GM-CSF express MHC class I and class II antigens, CD80, CD86, and CD11c, and are CD14 negative. They are able to phagocytose and process complex antigens, secrete IL-12p70 as well as angiogenic cytokines, and are highly responsive to environmental stimuli. At this stage of DC maturation, feeding or loading with complex antigens such as tumor cells, tumor cell lysates, antigen-antibody complexes or other protein mixtures generally takes place. This practice is based on evidence that iDCs are poised to take up antigens for intracytoplasmic processing and do so more efficiently than mDCs. In contrast, pulsing with peptides is best done with mDCs. The rationale is that complex antigens need to be processed before being presented, while peptides can be fitted into the groove of the MHC molecules liberally displayed on the surface of mDCs in a matter of minutes for presentation to T cells. The process of loading or pulsing of DCs ex vivo has been widely debated, because of the concern that antigens might compete for internalization or presentation by DCs. However, DCs appear to be able to handle a wide variety of antigens as well as cells, and are especially efficient when presented with Ag-Ab complexes.

While most clinical trials to date have relied on the use of multiple peptides to pulse DCs, in situations where the tumor antigens are unknown (as is the case with a majority of human solid tumors), autologous tumor cells or their products can be fed to DCs to serve as a source of tumor antigens [28]. Such delivery of a broad repertoire of both MHC class I– and class II–restricted epitopes offers a possibility for polyvalent immunization and synergistic CD4⁺ and CD8⁺ T-cell responses. While multiepitope DC-based vaccines have an advantage of combining preselected, well-defined tumor-specific peptides, they also have disadvantages: MHC-restricted and largely limited to melanoma, where immunogenic cytotoxic as well as helper peptides have been defined for several tumor-associated proteins, they are not broadly applicable. For most other tumor types, DC-based vaccines have to rely on the use of live tumor cells, apoptotic tumor cells, tumor cell products, or DC-tumor fusion products [2]. All of the above can be successfully fed to DCs, but the strict requirement for autologous tumor presents a practical difficulty of obtaining and processing the patient’s tumor at the time of surgery and banking it for use in the vaccine.

Engineering of DC maturation

Following loading with antigens, DCs undergo the process of maturation in the presence of a cytokine cocktail. This is a critical step in the generation of therapeutic DCs. Considerable evidence attests to the fact that the “cytokine profile” of DCs changes both qualitatively and quantitatively during DC maturation, and that it determines the nature of the immune response generated. In culture, the cytokine profile of DCs largely depends on the cytokine content of a cocktail provided to the cells. As the cocktail can be altered by mixing various cytokines, it is possible to direct DC maturation or to “polarize” DCs toward a desired functional phenotype [17]. For example, DCs polarized in the presence of IFN-γ and poly-I:C (the so-called type-1 polarized DC or DC1) have been shown to mature into highly effective antigen-presenting cells even in serum-free cultures. The polarized DC1 induced 4–50-fold higher levels of melanoma peptide-specific CTLs after a single in vitro sensitization (IVS) of T cells obtained from patients with melanoma. Further, the DC1 strongly enhanced T_h1-type responses (i.e., IFN-γ as opposed to IL-5 production in ELISpot assays) of antimelanoma CD4⁺ T cells. In addition, it seems that repolarization of T_h2-type responses in T-cell clones is possible in the presence of autologous DC1 and IL-12 [18]. These data suggest that when compared with the current “gold standard” (i.e., DCs matured in the presence of IL-1β, TNF-α, IL-6, and PGE₂) [16], DC1 are more potent and express a more stable functional phenotype. These attributes of DC1 make them a preferable type of APC for therapy.

The potential of DCs used for therapy to induce T_h1 vs T_h2 immune responses is likely to determine their utility. In order to induce tumor regression and promote long-term disease-free status, DCs have to be able to drive the in vivo expansion and maintenance of T_h1/T_c1-type T cells, both CD8⁺ and CD4⁺. Therefore, procedures used for DC polarization in vitro may be a critical step in the generation of therapeutically effective DCs.

Although no specific release criteria for DC potency based on the functional characteristics of cultured, antigen-pulsed, and matured DCs are broadly accepted today, a need exists for a better definition of such cells. The generally acceptable perception that therapeutic DCs should be viable (>90%) and able to produce cytokines after ex vivo stimulation is clearly not sufficient. At present, the functional quality of DCs administered to patients in different medical centers is likely to differ widely, depending on conditions of their culture and maturation. Therefore, disparate therapeutic results are not altogether unexpected.

Phenotypic and functional evaluations of DC products

Before matured, antigen-loaded or peptide-pulsed, and sterile DCs can be released for therapy, their phenotypic characteristics are determined by multicolor flow cytometry and their functions evaluated. Up-regulation of certain surface markers on matured DCs is expected, including expression of CD40, CD80, CD83, CD86, CD11c, CD25, and CCR7, a chemokine receptor which marks cells able to migrate to LNs [20]. Currently, the cytokine profile of DCs prepared for therapy can only be gleaned via the assessment of their ability to produce IL-12. In the future, it may be possible to determine it more precisely using multiplex technologies. DCs require two signals (e.g., CD40L and IFN-γ) for IL-12 production, and mature DCs can make substantial levels of IL-12p70 (e.g., >1,000 pg/ml/24 h) following stimulation with these agents for 24 h [34]. DCs can also be tested in the MLR-type assay for their ability to stimulate T cells, although this functional characteristic of DCs is highly variable. An ELISpot assay with an antigen-specific T-cell line as responder cells is perhaps the best available way of evaluating in vitro the ability of DCs to present the relevant antigen. Unfortunately, it is often not possible to perform this assay in the clinical setting, as antigen-specific T-cell lines are not commonly available.

The question of DC viability and potency is particularly relevant when a DC-based vaccine has to be divided into lots and frozen for administration at later time points. Comparisons performed in various laboratories between fresh and frozen/thawed DCs report different results, with some seeing a loss of viability as well as function and others reporting little or no changes in DC functions, if the cells are cryopreserved in a rate-controlled device. As multiple courses of DC-based vaccines are often delivered to patients with cancer, the functional activity of thawed cells is an important consideration that needs to be addressed more fully in the future.

DC administration to patients with cancer

The ability of DCs to traffic and to localize in appropriate regions of lymphatic tissues is critical for the success of DC-based vaccines. The route of DC administration as well as their maturation state could affect tissue localization of these cells. Similarly, the number of DCs and their potency are likely to influence in vivo interactions of DCs with other cells. Therefore, the dose and route of DC administration have been intensely debated. DCs are typically administered either intradermally, intravenously, or, in special circumstances, intraperitoneally. Clearly, one of the critical issues for induction of effective antigen-specific T_h1-type immunity in patients with cancer depends on defining a strategy for DC delivery that facilitates antigen presentation in vivo. The presumption that ex vivo pulsed DCs are stable and retain their cytokine-driven functions for prolonged periods of time may not be correct. The concern exists, therefore, that the delivery of “exhausted” or “overly matured” DCs could induce tolerance or favor T_h2-type responses, which would not be to the advantage of patients with cancer. Also, studies of DC trafficking in experimental animals, using labeled DCs have determined that only a very small percentage (0.1–2%) of DCs injected intradermally ever reach the tissue-draining lymph nodes [6]. DCs injected intravenously are rapidly sequestered by lung macrophages. Hence, most clinical protocols require very high numbers of DCs for vaccination. This costly and logistically difficult requirement may not be necessary, as the possibility of injecting DCs directly into lymphatics or intranodally allows for the majority of delivered DCs to reach the LNs. Cannulation of lymphatics or ultrasound-guided delivery are broadly used for administration of therapeutic agents. These routes of cell delivery offer the advantage of requiring fewer DCs and achieving a rapid cell localization to the T-cell areas of lymph nodes. The newer approaches to DC delivery seem to have embraced the idea of smaller doses of highly potent DCs, which retain their functions during migration in vivo, rapidly localize to lymph nodes, and effectively interact with CD8⁺ and CD4⁺ T cells. It is expected, although not yet proven, that this type of DC-based vaccination will produce dramatically improved therapeutic results.

In vivo antitumor activity of transferred DCs

Once the DC-based vaccine is delivered to the patient, its immunologic and therapeutic effects become of paramount interest. These effects can be broadly ascribed to at least two different types of mechanisms mediated by transferred DCs. One type of interaction may involve delivery of DCs directly to the tumor site or the migration of DCs from the vaccine-draining LN to the tumor. In the tumor, activated DCs pick up available antigenic epitopes and produce cytokines, which create a proinflammatory microenvironment. The DCs present a spectrum of tumor-derived epitopes to T cells at the tumor site or in the tumor-draining LN. This leads to epitope spreading and amplification of immune responses, which are not necessarily directed at antigens used for vaccination, but nevertheless target the tumor. If a cytokine, such as IL-12 or IL-2, is delivered together with the DCs, it may further enhance the epitope-spreading phenomenon and contribute to antitumor reactivity. It has been suggested that epitope spreading may be the major mechanism through which DCs amplify antitumor effects of multipeptide vaccines. Delivery of polarized DCs directly to the accessible tumor lesions may be an effective therapeutic strategy, based on the rationale that creating an inflammatory environment favors host antitumor mechanisms.

The second, and more widely accepted, view of DC in vivo interactions with the potential to induce tumor demise, incorporates systemic or intradermal transfer of ex vivo activated antigen-loaded, autologous DCs. Antitumor effects of DC-based therapies have been ascribed to the ability of DCs to induce both T-cell mediated helper and cytolytic responses specific for epitopes presented by transferred DCs. Endogenous DCs in patients with cancer may be compromised because of tumor-induced derangements in the DCs’ differentiation or their survival [32]. It thus makes sense to try and amplify antitumor responses by transfer of ex vivo activated, antigen-primed DCs. However, to succeed, this therapeutic strategy has to be able to engage immune cells in an orchestrated performance, which requires not only antigen presentation but primarily coordinated activation of T_h1 and T_h2 responses as well as survival of the activated effector and helper T cells in the tumor microenvironment. To this end, an extensive search for tumor-specific “helper” epitopes, which could replace nonspecific epitopes such as PADRE, KLH, or tetanus peptides in tumor vaccines, has been ongoing. It is yet to be determined how effective the newer vaccines, incorporating these helper epitopes, will be in cancer patients. Preclinical data indicate that T_h1-type peptide-specific CD4⁺ T cells (which secrete IFN-γ) can be generated from precursors in the peripheral blood of patients with NED, while patients with active disease predominantly make T_h2-type CD4⁺ T cells (which secrete IL-5) [37]. In the presence of peptide-pulsed DCs, it has been possible to re-direct the cytokine response from T_h2- to T_h1-type ex vivo, using peripheral blood cells obtained from cancer patients with active disease [18]. The implication of these findings is that the immune system of cancer patients with active disease is biased toward T_h2-type help, and that DCs might be able to promote the switch from the T_h2- to T_h1-type responses [37]. If confirmed, this finding provides a strong rationale for therapeutic transfers of peptide-pulsed DCs in patients with cancer.

Monitoring of DC-based therapies

As immune mechanisms are thought to be largely responsible for DC-mediated antitumor effects, monitoring designed to confirm these effects after DC transfers should target functions of immune cells. Today, it is fortunately possible to focus on tumor-specific immune responses. The endpoints could be DTH to tumor antigens and the frequency of circulating tumor-specific T cells. Single cell assays, ELISPOT, CFC, and tetramer analysis, all offer an opportunity for quantification of CD8⁺ and CD4⁺ T cells generated in response to specific peptides, which may or may not be components of the vaccine, as well as to autologous tumor. In addition, the functional status (i.e., responses to specific stimuli by cytokine production or by cytotoxic granule release) of these T cells can be determined. These are powerful monitoring tools, which have the potential of helping us dissect qualitatively and quantitatively the nature of T-cell responses targeting the tumor. In the context of serial monitoring of clinical trials, these methods combined with molecular analyses are likely to uncover the sought-after correlations between immune and clinical responses. However, care should be taken to perform the monitoring under conditions that guarantee a reliable assay performance, preferably in a reference laboratory operated as a good laboratory practice (GLP) facility.

Conclusions

DC-based therapy in patients with cancer is now largely in phase II trials. With the safety of DC transfers established, the challenge of the ongoing clinical studies will be to determine effective therapeutic doses and to obtain evidence for clinical efficacy of this form of immunotherapy. Opportunities will be available in the context of these clinical trials to acquire a better understanding of how DCs mediate antitumor effects. A number of questions that have to be addressed concern: e.g., optimization of culture conditions for DCs, especially if “tailored” subsets of polarized DCs are to be produced; definition of cytokine/chemokine profiles that characterize different DC subsets; establishment of conditions for DC polarization or repolarization toward clinically beneficial type-1 T-cell responses; and, finally, finding the means to sustain DC functions in the hostile tumor microenvironment. Studies are necessary to be able to understand which subsets of DCs exert immunogenic vs tolerogenic effects in vivo. The plasticity of DCs and the potential for differential regulation of their state of maturation have to be carefully handled to assure that cancer patients receive adoptive transfers of immunogenic DCs, engineered to promote T_h1- and T_c1-type tumor-specific responses. These and other issues are components of future translational research aimed at the understanding of the biology of DC subsets, their mechanism of action, and their utility for immunotherapy not only of cancer but also of other diseases.