Dendritic cells, T cell tolerance and therapy of adverse immune reactions

Abstract

Dendritic cells (DC) are uniquely able to either induce immune responses or to maintain the state of self tolerance. Recent evidence has shown that the ability of DC to induce tolerance in the steady state is critical to the prevention of the autoimmune response. Likewise, DC have been shown to induce several type of regulatory T cells including Th2, Tr1, Ts and NKT cells, depending on the maturation state of the DC and the local microenvironment. DC have been shown to have therapeutic value in models of allograft rejection and autoimmunity, although no success has been reported in allergy. Several strategies, including the use of specific DC subsets, genetic modification of DC and the use of DC at various maturation stages for the treatment of allograft rejection and autoimmune disease are discussed. The challenge for the future use of DC therapy in human disease is to identify the appropriate DC for the proposed therapy; a task made more daunting by the extreme plasticity of DC that has recently been demonstrated. However, the progress achieved to date suggests that these are not insurmountable obstacles and that DC may become a useful therapeutic tool in transplantation and autoimmune disease.

INTRODUCTION

Dendritic cells (DC) are now regarded as critical inducers and regulators of adaptive immune responses [1–3]. This contemporary view of DC, once regarded solely as highly effective instigators of T cell-mediated immunity, reflects insights gained over the past several years into how these highly specialized antigen (Ag)-presenting cells (APC) can regulate T cell activation and function, both in vitro and in vivo. Their tolerogenic properties, first recognized within the thymus [4,5], are now well-documented in relation to peripheral immune responses in various model systems. In the healthy steady state, the principal role of DC in the periphery may be to maintain tolerance to self Ag, encountered as apoptotic cells – the result of normal tissue turnover, and that they convey to T cell areas of secondary lymphoid tissue [6]. Evidence has also accumulated, both from animal and human experiments, that DC, particularly those at an immature stage, can suppress peripheral T cell responses to foreign Ag, and induce Ag-specific tolerance [7,8]. Thus, in addition to being viewed in the traditional sense, as natural adjuvants and as potential targets for inhibition of adverse immune reactions, the regulatory functions of DC have generated considerable interest in the potential of these cells for therapy of such diverse immune responses as allograft rejection, autoimmune inflammation and allergic hypersensitivity.

Several DC subsets with distinct phenotypic and functional characteristics have been identified in mice and humans. These have been the subject of recent comprehensive reviews [9,10]. Here we review the properties of murine and human DC that have been associated with T cell tolerance in vitro and in vivo, insights into DC tolerogenicity that have been gained from disease models, and the current status of various strategies using DC for therapy of immune-mediated inflammatory disorders.

DENDRITIC CELLS: CRITICAL REGULATORS OF IMMUNITY

Dendritic leukocytes are rare (<1% of circulating blood mononuclear cells) heterogeneous, uniquely well-equipped bone marrow (BM)-derived Ag-presenting and processing cells present in blood and lymph and resident within virtually all tissues. DC in the periphery constantly sample the local microenvironment by uptake of self and exogenous Ag via macropinocytosis/endocytosis. Until recently, it was thought that there was a comparatively low level of migration of DC to draining secondary lymphoid tissue via afferent lymphatics, in the normal healthy steady state. Recent experiments on the kinetics of DC migration have revealed however, that, depending on their location, DC traffic rapidly to and through secondary lymphoid organs [11,12]. Using BrdU labelling, Kamath et al. [11] showed that DC in the spleen or mesenteric lymph node had average lifespans of between 2 and 9 days. Interestingly, the cutaneous lymph node exhibited a slower accumulation of BrdU-labelled DC, but this was related to a longer lifespan in the skin since, once the DC reached the lymph node, they had a short lifespan of 2–3 days. These data are compatible with earlier findings demonstrating a rapid turnover of DC from the airway [13] and intestinal epithelium [14]. Conveyance and presentation of epithelial cell-derived Ag captured as apoptotic cells/bodies, – the result of normal tissue turnover, may promote/maintain tolerance to self Ag, and the recent demonstration of a gastric self Ag acquired and presented by DC in the gastric lymph node supports this hypothesis [15]. Induction of CD8⁺ T cell tolerance in mice is likely to involve the cross presentation of Ag, predominantly by CD8α⁺ DC [16,17]. The DC that, in the steady state, convey self Ags are immature and a recent study has demonstrated that these cells are capable of inducing T cell deletion in the periphery [18]. In these experiments, Ag was targeted to splenic DC by coupling to an anti-CD205 monoclonal antibody (mAb). Ag-specific T cells expanded for the first three days and then disappeared. More importantly, mice treated in this way were tolerant to a subsequent challenge with Ag in the presence of adjuvant [18]. Other mechanisms for the induction of tolerance by DC in the steady state include the targeting of Ag via the inhibitory FcγRIIb [19], and the presence of inhibitory receptors, such as immunoglobulin (Ig)-like transcript (ILT)3 and ILT4 [20].

In response to infection and microbial products (activation of DC via Toll-like receptors (TLR)/inflammation), or following organ transplantation, the rate and magnitude of DC migration from the periphery to T cell areas of lymph nodes or spleen is greatly enhanced. This translocation is associated with DC maturation, whereby Ag uptake/processing ability is down-regulated and surface expression of chemokine receptors (e.g. CCR7) that facilitate their transendothelial homing to secondary lymphoid tissue is enhanced. In addition, expression of major histocompatibility complex (MHC) and costimulatory molecules (CM) essential for T cell activation is increased. Depending on the nature of the infection, DC produce IL-12p70 – a potent T helper-1 (Th1) cell driving cytokine critical for the development of delayed-type hypersensitivity (DTH) responses such as those that underlie pathological processes in allo- and autoimmunity.

MECHANISMS BY WHICH DENDRITIC CELLS REGULATE T CELL REACTIVITY

DC can either initiate T cell activation and proliferation or promote peripheral tolerance through the deletion of autoreactive T cells depending on their state of maturation. DC whose capacity to activate T cells is impaired/modified, either as a result of incomplete maturation, or the influence of specific inhibitory cytokines (e.g. IL-10 or transforming growth factor β; TGFβ), can induce reversible Ag-specific T cell hyporesponsiveness (anergy) or apoptosis in vitro, and suppress immune reactivity. Inhibitory effects on T cells are not confined to immature DC. Recently, Munn et al.[21] described a human monocyte-derived DC subset (CD123⁺[IL-3Rα⁺] CCR6⁺) expressing the tryptophan-catabolizing enzyme indoleamine 2,3 dioxygenase (IDO) that could inhibit T cell proliferation in vitro. The IDO-mediated suppressor activity was present in both immature as well as mature CD123⁺ DC. Although these cells have been identified in vivo, their role has yet to be confirmed. Specific subsets of mouse splenic DC are highly effective in mediating IDO-dependent apoptosis of T cells in vitro[22]. Of note, it appears that ‘reverse’ signalling resulting from B7 ligation by CTLA4 on T cells may contribute to tolerogenic effects via interference with tryptophan catabolism [23]. Kawahata et al.[24] have shown that, using DC derived from lymphoid tissue of transgenic mice, DC expressing a nuclear autoAg lead to a persistent anergic state of CD4⁺ autoreactive T cells. The results suggest that peripheral tolerance to nuclear autoAg may be due to continuous presentation of the self-peptide by DC, and that a low level of peptide expression may also be involved in the induction of T cell hyporesponsiveness.

The rediscovery of regulatory/suppressor cells [25] has revealed that DC play an important role in the induction and maintenance of several regulatory T cell populations [20,26,27,28,29,30]. Thus, an IL-10-secreting, nonproliferating human CD4⁺ T cell population with regulatory properties can be induced in vitro by repeated stimulation of naïve T cells with allogeneic immature DC [31]. In mice, a novel, liver-derived, B cell-like DC (CD205⁺B220⁺CD19⁻) can induce allogeneic T cells with a cytokine profile resembling Tr1 cells [32]. Moreover, immature Ag-pulsed autologous myeloid DC induce Ag-specific, IL-10-producing CD8⁺ T cells in humans [33]. Using CD4⁺ Ag-specific T cell lines, human plasmacytoid DC (pDC) enriched from blood induce reversible T cell anergy in vitro. The anergic T cells produce IFNγ and IL-10 upon stimulation. Notably, these pDC can be mobilized selectively into human blood by granulocyte colony-stimulating factor (G-CSF) administration and following activation induce T cells that produce predominantly IL-10 and IL-4 [34]. It has also been shown recently that DC found either in the bronchial or intestinal mucosa induce regulatory T cell populations, and this seems to be an intrinsic property of DC in these specific locations [35–37]. In the gut, DC preferentially induce Th2/Th3 cells that secrete IL-4, IL-10 and TGFβ[35,36] which play an important role in maintaining tolerance to oral Ags. In the respiratory tract, DC produce large amounts of IL-10 following encounter with Ag and induce the production of IL-10-producing Tr1 cells [38]. Thus, depending on their location, DC have the unique capacity to stimulate specific regulatory T cells that protect those sites from potential autoimmunity.

Since many pathogenic immune responses, including allograft rejection and autoimmunity, involve Th1 responses, there has been a great deal of interest in analysing the ability of DC to deviate the immune response away from Th1 and towards Th2. Indeed, DC populations that can preferentially induce either Th1 or Th2 cells have been identified in both mouse and human [39,40] and this has been correlated to the amount of IL-12p70 that is produced by these cells. While it was thought initially that different DC lineages/subsets would stimulate either Th1 or Th2, it is now clear that DC of all lineages can be induced to stimulate either Th1 or Th2, depending on the nature of the stimulating Ag and the local environment (reviewed in [10]). Thus, Ag-specific suppression of cell-mediated immunity, achieved by intravenous (i.v.) administration of murine Ag-pulsed Langerhans cells or splenic DC, has been attributed to selective activation of Th2 cells [41]. DC grown in the presence of prostaglandin [PG]E₂ and unable to secrete IL-12p70, promote the development of Th2 cells [42]. IL-10 skews the Th1/Th2 balance to Th2 cells by inhibiting CM expression and suppressing IL-12p70 synthesis by DC [43]. Conceivably, the capacity of DC subsets to skew selectively toward different types of Th cell response in vivo may reflect differential sensitivity of Th subsets to DC-mediated apoptosis.

Interestingly, anergizing capacity can be acquired by DC exposed to regulatory T cells. A distinct population of human T reg (T suppressor; Ts) cells, characterized by their CD8⁺ CD28⁻ phenotype, suppresses Ag- specific CD4⁺ Th cell responses by acting on the priming APC that present peptide-MHC class I complexes to which the Ts cells have been previously primed. The suppression of DC function is due to inhibition of the gene transcription regulatory protein nuclear factor (NF)-κB activation and transcription of CM in the DC. Exposure of immature DC to CD8⁺ CD28⁻ Ts cells increases expression of the genes encoding Ig-like inhibitory receptors ILT3 and ILT4 that render the DC capable of anergizing CD4⁺ Th cells [20].

ORGAN TRANSPLANTATION: THE ROLE OF DENDRITIC CELLS IN DIRECT AND INDIRECT ALLORECOGNITION

Rejection of organ allografts has been associated traditionally with the migration of interstitial donor ‘passenger’ leucocytes to T cell areas of recipient lymphoid tissue. This trafficking of donor-derived DC allows direct presentation of highly immunogenic, donor-derived MHC Ags to recipient naïve T cells. Whilst donor DC are clearly important in direct allorecognition, host DC also play a significant role in graft rejection via the indirect pathway of allorecognition. This occurs when host APC present donor peptides to host T cells in the context of recipient MHC molecules. The relative contributions of direct and indirect allorecognition to murine skin graft rejection have been examined [44]. During acute rejection, <10% of T cells recognized allopeptides presented indirectly. By contrast, the remaining 90% of responding T cells responded to directly presented donor MHC peptides. This predominant role of the direct pathway provides a rational basis for manipulation of donor-derived DC (that are readily accessible in live-organ donation) to prevent acute graft rejection and to promote tolerance induction. Since indirect allorecognition is thought to be of greater significance in the pathogenesis of chronic rejection [45], manipulation of host-derived DC by pulsing with donor MHC class I peptide to promote their tolerogenicity [46,47] may be an equally important approach to therapy of this process that leads commonly to graft destruction. Approaches that have been adopted for treatment of allograft rejection based on utilization/targeting of DC are summarized in Table 1.

Table 1.

Approaches to therapy of allograft rejection by using/targeting donor or host DC

Source of DC	Approach	Reference
Donor DC
	Unmodified (in vitro-generated or in vivo-mobilized)	[48–50]
	Immature DC	[34,120]
	Specific DC subset (? plasmacytoid)
	Manipulated
	Immunological (e.g. IL-10; anti-IL-12 mAb; CTLA4Ig)	[121–123]
	Pharmacological (e.g. aspirin; dexamethasone; rapamycin; deoxyspergualin)	[51,54, 124–126]
	Genetically engineered (e.g. viral vectors encoding IL-10; TGFβ1; FasL; CTLA4Ig)	[104–106, 113,127]
	Antisense (e.g. NF-κB decoy oligodeoxyribonucleotides)	[53,114]
Recipient DC
	Intrathymic injection
	Promote central tolerance through administration of DC pulsed with donor MHC class I peptide	[46]
	Pharmacological manipulation	[54]
	Genetic engineering
	Transgenic expression of donor MHC I	[115]
	Uptake of apoptotic cells
	May promote induction of T reg cells	[128–130]

DENDRITIC CELLS THERAPY OF ALLOGRAFT REJECTION

Intrathymic injection of alloAg-pulsed DC

Injection of host BM-derived DC pulsed with a donor MHC class I peptide into the thymi of streptozotocin-induced diabetic Wistar-Furth rats, in combination with anti-lymphocyte serum therapy, results in permanent (>200 day) pancreatic islet allograft survival [46]. Recently, similar prolongation of heart allograft survival has been achieved using host BM-derived DC pulsed with immunodominant allopeptide delivered intravenously, thereby circumventing the limitations imposed by intrathymic DC administration [47].

Tolerogenic properties of immature dendritic cells in organ transplantation

Over the past several years, credence has been given to the concept that immature donor DC can promote organ or pancreatic islet allograft survival [48,49], and induce donor-specific tolerance [50] (Table 2). Thus BM-derived DC generated from C57BL10 (B10) mice in low concentration GM-CSF for 8 days, retain an immature phenotype and are resistant to maturational stimuli such as bacterial lipopolysaccharide (LPS), TNF-α or CD40 ligation. When 5 × 10⁵ of these immature donor DC are administered i.v. to fully allogeneic CBA recipients, 7 days before transplant, they demonstrate an impressive ability to prolong cardiac allograft survival indefinitely (>100 days) in the absence of immunosuppressive therapy [50]. Both anti-inflammatory (e.g. salicylates) [51] and immunosuppressive drugs (e.g. corticosteroids) [52] that inhibit DC maturation in vitro have been shown to suppress nuclear translocation of NF-κB that is critical for DC maturation. Another means to promote and maintain the immature state is to specifically target the NF-κB cell activation pathway using antisense oligonucleotides. Short oligodeoxynucleotides (ODN) with consensus binding sequences to NF-κB inhibit DC allostimulatory capacity by blocking NF-κB translocation [53]. This in turn, inhibits cell surface CM expression. BM-derived DC from B10 mice treated with NF-κB ODN for up to 36 h and administered to C3H recipients as a single dose (2 × 10⁶) 7 days before organ transplantation, significantly prolong B10 vascularized heart graft survival [53].

Table 2.

Therapeutic effect of distinct DC subsets alone in the prolongation of transplant survival and prevention of autoimmune disease

DC subset	Condition	DC source	Therapeutic effect	Reference
Transplantation
Immature donor myeloid (CD8α−) DC	Cardiac allograft	In vitro-generated	Prolonged graft survival	[49]
Immature donor myeloid DC (GMlowDC)	Cardiac allograft	In vitro-generated	Indefinite (>100 days) graft survival	[50]
CD205+B220+CD19 donor liver-derived DC	Cardiac allograft	In vitro-generated	Prolonged graft survival	[32]
Immature or mature ‘lymphoid – related’ CD8α+ donor DC†	Cardiac allograft	Freshly isolated	Prolonged graft survival	[120]
Autoimmune disease
Myeloid (CD8α−) DC	Autoimune diabetes (NOD* mouse)	Freshly isolated from pancreatic lymph node	Prevention of diabetes	[77,78]
Mature myeloid DC	Autoimmune diabetes (NOD mouse)	In vitro generated	Prevention of diabetes, induction of Th2 response	[79,82]
‘Semi-mature’ DC	EAE† (C57BL/6 mice)	In vitro generated-matured  with TNF-α	Prevention of EAE, induction of IL-10-producing T cells	[26]
‘EAE’-DC	EAE (Lewis rats)	In vitro generated from rats with EAE	Prevention of EAE and resistance to subsequent challenge	[88]
Splenic DC	EAMG‡ (Lewis rats)	Freshly isolated, exposed to TGFβ	Prevention of EAMG, reduced levels of specific antibody	[92]

^*Non-obese diabetic;

^†experimental allergic encephalomyelitis;

^‡experimental autoimmune myasthenia gravis.

A critical role of the maturational status of DC in allograft tolerance has been implicated in a nonhuman primate renal transplant model. Rhesus macaque monkeys were treated with a combination of anti-CD3 immunotoxin (IT) mAb and a 15-day course of deoxyspergualin (DSG), commencing 4 h before transplant [54]. This combination, but not IT alone, was associated with the development of long-term (>3 years) kidney allograft survival, without evidence of chronic graft nephropathy. Split skin grafts from third party donors were rejected promptly, indicating donor specificity of the IT-DSG therapy. Treatment of graft recipients with DSG inhibited nuclear translocation of NF-κB within secondary lymphoid tissue DC and their phenotypic maturation (up-regulation of CD86 and CD83 expression), maintaining these cells in an immature, potentially tolerogenic state. This effect on DC in situ was transient however, suggesting that the timing of DC immaturity may be critical in the induction of long-term tolerance. Interestingly, the DSG analogue LF15-0195 can induce donor DC to expand potent donor-specific regulatory CD4⁺ CD25^+/−T reg cells in rat MHC-mismatched heart allograft tolerance [55].

AUTOIMMUNITY: THE ROLE OF DENDRITIC CELLS

In view of the critical role that DC play in the maintenance of central and peripheral tolerance it is perhaps not surprising that abnormalities in DC function have been implicated in several autoimmune diseases [56]. In type-1 insulin-dependent diabetes, several groups have reported abnormalities in DC phenotype and function in both the mouse and human, which may result in the skewing of the response towards pathogenic Th1 cells [57–61]. In addition, non-obese diabetic (NOD) mice, the murine model for type-1 diabetes, exhibit defects in the number of regulatory T cells, including natural killer (NK)T cells, CD4⁺ CD25⁺ T cells and Th2 cells [62–64], – defects which may, in the future, be attributed to abnormalities of DC function. In systemic lupus erythematosus (SLE), recent reports have revealed abnormalities in the levels of interferon (IFN)-α production and in plasmacytoid DC that produce this important cytokine [65,66]. In murine SLE models, increased numbers of myeloid DC have been noted [67,68] and in one case, this was correlated with increases in the DC-mobilizing haematopoietic factor fms-like tyrosine 3 kinase ligand (Flt3L) in the BM [67]. Increased numbers of DC have also been observed in the affected joints of rheumatoid arthritis (RA) patients [69,70]. More recently, it was shown that RA synovial fluid contains DC precursors and myeloid DC growth factors, such as GM-CSF, TNF, stem cell factor, and IL-13, which suggests that the RA joint could potentiate the differentiation of DC that could then further activate the autoimmune response [71].

Studies of DC in multiple sclerosis (MS) patients have revealed higher levels of IFNγ, TNFα and IL-6-secreting DC than in healthy subjects [72] whereas elevated expression of CCR5 (a chemokine receptor for CCL5 by circulating monocytoid DC may contribute to the recruitment of these cells to cerebrospinal fluid in MS [73]. Moreover, secretion of extracellular matrix-degrading metalloproteinases and their inhibitors by blood monocyte-derived DC is imbalanced in MS [74]. DC have also been implicated in Sjögrens syndrome [75] and thyroiditis [76].

DENDRITIC CELLS AS THERAPEUTIC AGENTS IN AUTOIMMUNE DISEASE

Autoimmune (type-1) diabetes

Ten years ago, Clare-Salzler et al.[77] demonstrated that local sc injection of DC isolated from draining lymph nodes of the pancreas (but not spleen DC) could protect NOD mice from disease development, and suggested that these DC could suppress autoimmunity by induction of regulatory cells. These observations have been confirmed and extended in more recent studies, in which an immunoregulatory role of CD1α-restricted invariant NKT cells has been attributed to the recruitment of tolerogenic myeloid DC to the pancreatic lymph nodes and inhibition of autoimmune inflammation [78]. Feili-Hariri et al.[79] have demonstrated that mature BM-derived DC are more effective in preventing diabetes in NOD mice than immature DC. These results have been confirmed in several other studies [80,81]. We have further demonstrated that the DC-mediated protection is a result of the induction of a regulatory Th2 response [82], and that this was correlated with the DC expressing high levels of CM and producing low levels of IL-12p70 [83]. This approach may have corrected the intrinsic regulatory defect in NOD mice that exhibit a strong Th1 bias in their generalized immune responsiveness [84] and may be applicable in humans with autoimmune diabetes.

Other therapeutic strategies have been shown to exert their effect through action on DC. The selective expression of IL-4 by β islet cells was found to act directly on local DC and alter the relative expression levels of CD80 and CD86, thereby inhibiting the differentiation of diabetogenic cytotoxic T lymphocytes (CTL) [85]. Recently, CD40L blockade in the rat insulin promoter-lymphochoriomeningitis virus model of autoimmune diabetes was found to completely prevent disease and the protection could be transferred by a cell population that had features of both DC and NK cells [86]. Therapeutic effects of DC that have been reported in models of type-1 diabetes and other autoimmune disorders are summarized in Table 2.

Experimental allergic encephalomyelitis (EAE) and multiple sclerosis (MS)

EAE is an animal model closely related to multiple sclerosis (MS). In a series of studies, Xiao et al.[87] have shown that tolerance can be induced against acute EAE in Lewis rats by BM-derived DC pulsed with encephalitogenic myelin basic protein (MBP) peptide 68–86 and injected s.c. into healthy rats before immunization with MBP 68–86 plus complete Freund's adjuvant (CFA). This therapeutic effect is associated with immature DC expressing high levels of IL-10 and low levels of IL-12, increased IFN-γ expression, nitric oxide (NO) production, gradually reduced proliferation and apoptosis of CD4⁺ T cells and/or up-regulation of TGFβ-expressing cells in T cell areas of lymph nodes induced by DC-derived NO [88,89]. The same group has reported that DC-derived NO, that promotes apoptosis in autoreactive T cells, is involved in (intranasal) IL-4-induced suppression of EAE in the same Lewis rat model [90]. In mice, repetitive injection of wild-type DC matured with TNFα (‘semimature’ DC) and pulsed with autoantigenic peptide can induce peptide-specific, IL-10-producing CD4⁺ T cells in vivo and prevent EAE [26]. Efficient suppression of EAE has also been demonstrated in mice using a minor splenic DC subset (CD8α⁻ CD4⁺) capable of binding a disease-ameliorating Ig-chimeric molecule carrying the myelin-oligodendrocyte glycoprotein 35–55 peptide [91]. These findings have been taken as evidence that DC-based immunotherapy could be beneficial at least as a complement to conventional therapy of MS.

Experimental autoimmune myasthenia gravis (EAMG)

DC exposed in vitro to TGFβ1 (but not untreated DC) and administered s.c. (single injection only) to Lewis rats (2×10⁶/rat) on day 5 after immunization with acetylcholine receptor (AChR) in CFA mediate protection against EAMG, and inhibit production of anti-AChR IgG Abs [92]. This finding indicates that DC-based therapy offers potential for inhibition of Ab-mediated autoimmune diseases.

DC as inducers of autoimmunity

While DC have been used effectively in several models of autoimmune disease as a means to prevent or treat disease, it is important to realize that under some circumstances, DC can induce autoimmunity. This has been shown in models of EAE [93,94], diabetes [95] and autoimmune thyroiditis [96]. The finding is perhaps not surprising in view of the efficiency of these cells as inducers of the immune response. These studies highlight the requirement for a thorough understanding of DC biology and function prior to exploiting these cells as therapeutic agents. In addition, it is important to comprehend the autoimmune response, such that specific DC populations may be used to achieve the desired therapeutic effect.

DENDRITIC CELLS IN ALLERGIC ASTHMA

There is recent evidence that mature pulmonary DC in bronchial lymph nodes of mice exposed to respiratory allergen induce the development of IL-10-producing T reg cells. This process is dependent on T cell costimulation via the inducible costimulator (ICOS)-ICOSL pathway [38]. These T reg cells, production of which is IL-10-dependent, block the development of airway hypersensitivity. Recent studies in patients with allergic asthma have revealed that DC are recruited rapidly to the bronchial mucosa following exposure to allergen [97]. Similar findings have been observed in animal models [98] and recently, a long-lived Ag-bearing DC population was observed in a model of allergic asthma in the mouse [99]. A newly described IL-7-like cytokine, thymic stromal lymphopoietin (TSLP), has been shown to act on DC to induce Th2-stimulating function. This factor may have an important role in allergic hypersensitivity since it is highly expressed in the skin of patients with atopic dermatitis [100]. To date, there have been no reports of the ability of DC subsets to prevent or treat asthma, rather they appear to contribute to the disease and could be future targets of therapeutic modalities [101,102]. Thus, future therapies of allergic asthma may aim to recover the normal tolerogenic nature of airway DC, possibly using gene therapy approaches, as described below.

THERAPEUTIC POTENTIAL OF GENETICALLY ENGINEERED DENDRITIC CELLS

An attractive conceptual approach to enhancement/stabilization of the tolerogenic potential of DC is their genetic modification to express ‘immunosuppressive’ molecules that can either (i) inhibit or block cell surface CM expression (e.g. IL-10, TGFβ, or CTL Ag 4 [CTLA4]Ig) and skew the Ag-specific T cell response towards Th2 predominance) or (ii) promote the deletion (apoptosis) of Ag-specific T cell clones (e.g. FasL [CD95L] or TNF-related apoptosis-inducing ligand [TRAIL]) [103]. In principle, ectopic expression of these molecule(s) by DC trafficking to the precise microenvironment in which Ag presentation and Ag-specific T cell responses are initiated, minimizes systemic delivery of the immunosuppressive gene product, and diminishes potential undesired side-effects.

Progress in transplant models

These findings have prompted early evaluation of genetically engineered DC in experimental (murine) organ transplantation (Table 3). The best results following single gene transfer to donor DC have been obtained by Min et al.[104] who found that multiple intraperitoneal (i.p.) injections of DC transduced using lipofection to express human FasL markedly extended murine vascularized heart allograft survival. Electroporation of cDNA encoding CTLA4Ig into a mouse DC line renders the cells capable of prolonging pancreatic islet allograft survival [105]. Coates et al.[106] showed that NOD-SCID mice engrafted with human skin and reconstituted with allogeneic human PBMC mixed with adenovirus (Ad)IL-10-transduced DC autologous to the skin donor, exhibited reduced evidence of graft rejection. To date, there have been no reports of donor-specific tolerance being achieved across MHC barriers using genetically modified donor-derived DC alone. In one report concerning IL-4-transduced donor DC, pretreatment of heart graft recipients with these cells exacerbated rejection [107]. Studies in large animals have been limited to preliminary work on AdIL-10-transduced DC in sheep [108]. A recent report indicates that rhesus monkey DC genetically modified to over-express TGFβ1 by targeting Ad to DC CD40 inhibit CD4⁺ and CD8⁺ T cell responses in an alloAg-specific manner [109].

Table 3.

Evidence that genetically engineered myeloid DC can prolong allograft survival or inhibit autoimmune disease

Origin of DC	Gene transduced (method)	Regimen	Graft/disease	Reference
Transplantation
Donor	FasL (lipofection)	Six postoperative i.p. injections of 5·106 DC	Heart	[104]
Donor	CTLA4Ig* (electroporation)	Twenty-five × 106 DC i.v. on day 0 and 6	Pancreatic islets	[105]
Donor	TGFβ and IL-10† (adenovirus)	Portal venous injection of 5·106 DC 36 h  before transplant	Kidney	[113]
Donor	CTLA4-Ig‡ (adenovirus)	Two × 106 DC i.v. 7 days before transplant	Heart	[114]
Donor	IL-10 (adenovirus)	i.p. injection of 106 DC autologous to the skin  donor + allogeneic human PBMC	Human skin in  NOD-SCID mice	[106]
Donor	IL-4 (retrovirus)	Two × 106 DC i.v. 7 days before transplant	Heart§	[107]
Donor	TGFβ (retrovirus)	Two × 106 DC i.v. 7 days before transplant	Heart	[127]
Recipient	Donor MHC class I  (adenovirus)	One month before transplant + anti-CD4 mAb	Heart	[115]
Autoimmune disease
Autologous	IL-4 (adenovirus)	One × 106 DC i.v. to mice with established disease	CIA¶	[116]
Autologous	IL-4 (retrovirus)	Three × 105 DC s.c., i.v. or i.p. to mice 15 days  after collagen immunization	CIA	[117]
Autologous	FasL (adenovirus)	One × 106 DC i.v. to mice with established disease	CIA	[118]
Autologous	IL-4 (adenovirus)	Two × 5 × 105 DC i.v.to NOD mice at 10–11  weeks of age	Autoimmune diabetes	[119]

^*DC cell line

^†1:1 mixture of DC transduced with either Ad IL-10 or Ad TGFβ

^‡adenoviral transduction of DC treated with NFκ^B antisense oligodeoxynucleotides

^§exacerbation of rejection

^¶CIA, collagen-induced arthritis.

The limited efficacy of genetically engineered DC observed in transplantation models to date can be ascribed to a number of non-exclusive factors that constitute important issues for further investigation. These include (i) unsustained immaturity of the potentially tolerogenic DC; (ii) inappropriate/inadequate numbers of injected DC; (iii) suboptimal route/frequency of cell injection; (iv) administration of a suboptimal/inadequate DC subset; (v) absence of persistent (long-term) transgene expression; (vi) transduction of only a minor population of administered DC (especially with retrovirus); (vii) adverse (immunostimulatory) effects of (Ad) vectors, and viii) failure of manipulation of donor DC to inhibit the indirect pathway of allorecognition. Interestingly, genetic modification of mouse embryonic stem-cell line-derived DC indicates therapeutic potential for this approach [110].

Innovative strategies

Retroviral transduction of mouse BM-derived myeloid DC to produce viral (v)IL-10 substantially impairs their allostimulatory activity [111], but the transduction strategy is comparatively inefficient. Generation of a ‘sorting’ retroviral vector, encoding both vIL-10 and enhanced green fluorescent protein, permits selection (by flow cytometry) of positive transfectants with even further reduced T cell stimulatory activity (potentially tolerogenic, vIL-10-secreting DC) [112]. Combined use of DC populations transduced to overexpress different immunosuppressive genes offers an alternative approach. Thus, significantly prolonged renal allograft survival can be achieved in mice with a mixture (1:1): of donor-derived myeloid DC transduced with either AdTGFβ or AdIL-10 via the portal venous route, 36 h before transplantation. This effect correlates with inhibition of CTL induction and with enhancement of Th2 responses [113].

The capacity of Ad virus to induce DC maturation limits the therapeutic efficacy of recombinant Ad-transduced DC. An important recent finding is that NFκB-specific ‘decoy’ ODN stably inhibit DC maturation [53]. Moreover, they also markedly suppress DC maturation induced by various stimuli, including rAd vectors. Significantly, immunosuppressive transgene (CTLA4Ig) expression is not inhibited. Of even greater significance, is the finding that a single, pretransplant injection of NFκB ODN-treated, rAd CTLAIg-transduced donor myeloid DC can markedly prolong vascularized heart allograft survival. Forty percent of the animals exhibited long-term (>100 day) graft survival, and donor-specific skin graft tolerance [114].

As discussed above, cross-presentation of alloAg by recipient DC may play an important role in allotolerance. Billing et al. [115] have reported that administration of transgenic CBK (H2^k + K^b as a transgene) DC to CBA (H2^k) recipient mice, 27 days before transplant in conjunction with two doses of anti-CD4 mAb on days −28 and −27, induces long-term survival in 75% of fully allogeneic B10 (H2^b) heart graft recipients. This group has further demonstrated that infusion of immature, recipient-derived DC transduced with an Ad vector encoding donor-type MHC class I gene (H2K^b) one month before transplant, together with anti-CD4 mAb, prolongs the survival of fully allogeneic cardiac grafts [115]. These findings suggest that genetically modified autologous DC may prove useful for delivery of donor alloAg in the induction of transplant tolerance.

Efficacy of genetically engineered dendritic cells in autoimmune disease models

The potential of genetically engineered DC for therapy of chronic systemic autoimmune disease (Table 3) has recently been demonstrated convincingly in a murine model of collagen-induced arthritis (CIA) [116,117]. Thus, a single systemic (i.v.) injection of 10⁶ Ad IL-4-transduced BM-derived DC to mice with established CIA reduced the severity of the disease, and disease was completely ameliorated within 1 week in at least 50% of the animals. By contrast, mice injected with AdIL-10 DC showed continuous disease progression. In a separate study, a single injection of retrovirally transduced IL-4 DC also reduced the incidence and severity of CIA and suppressed established Th1 responses and associated humoral responses. These effects were achieved despite only transient persistence of the injected DC in the spleen. IL-4-transduced T cells or fibroblasts failed to alter the course of the disease, indicating that the therapeutic effect was restricted to DC. A recent study demonstrated that DC engineered to express FasL were also effective in the prevention of CIA [118].

We have also shown recently that DC engineered to express IL-4 can prevent autoimmune diabetes in NOD mice with already advanced insulitis. Between 8 and 10 weeks of age, insulitis progresses from a mild peri-insulitis to aggressive intraislet infiltration. At 10–11 weeks of age, unmodified DC are no longer effective, but those that express IL-4 prevent diabetes in a significant number of animals [119]. This protection is long-lasting and is associated with an increase in Th2 cytokines in the pancreas.

CONCLUSIONS

Over the last several years, it has become apparent that DC can function both as initiators of the immune response and as inducers of central and peripheral tolerance. The factors that determine how a given DC functions depends on its state of maturation and the local microenvironment. The use of DC in the treatment of allograft rejection is aimed at using the tolerogenic capacity of the cells to prevent the initiation of an immune response, either by deletion, the induction of anergy, or the stimulation of regulatory T cells. Thus, the goal would be to exploit the ability of DC to maintain tolerance in the steady state to inhibit the response to the ‘novel’ foreign allo-Ag. Since however, most autoimmune and allergic diseases are associated with an ongoing immune response, it is not sufficient to rely on the steady state tolerogenic function of DC. Rather it is important to utilize the ability of specific DC populations to modulate the immune response, thereby converting it from a pathogenic response to a protective one. This can be achieved via the induction of regulatory T cells, such as Th2, Tr1, suppressor cells and NKT cells. The most effective therapy will be determined by the nature of the pathogenic immune response. The challenge is to identify conditions that allow the generation and use of DC with defined characteristics, a task made more daunting by the highly plastic nature of DC themselves. The use of genetic modification of DC may be fruitful means to achieve this goal.