Dendritic cells (DC) have been viewed traditionally as ‘foes’ that instigate the rejection process. More recently, they have gained considerable attention as antigen (Ag)-presenting cells (APC) that are also capable of induction/maintenance of immunologic tolerance (1–3). This contemporary view of DC, reflects new insights into how these cells regulate T cell responses. Herein we review accumulating evidence that manipulation of DC can suppress Ag-specific immune responses and build a case that this may provide a novel approach to therapy of allograft rejection. Observations in rodents, nonhuman primates and humans that support or are consistent with tolerogenic activity of DC are summarized in Table 1 (3–17).
Table 1. Transplant-related observations in rodents, humans and nonhuman primates, consistent with tolerogenic properties of DC
Donor DC (splenic) can prolong organ allograft survival (mice) (4)
Persistence of donor DC in long-term rodent or human recipients of organ allografts, including stable patients off immunosuppression (5, 6)
Propagation of donor DC from tolerant but not acutely rejecting organ allograft recipients (mice) (7)
Infusion of immature donor DC prolongs fully allogeneic graft survival in nonimmunosuppressed recipients (mice) (8, 9)
Donor DC are essential for DST-induced transplant tolerance (rats) (10)
Recipient DC pulsed with donor MHC class I peptide promote organ transplant tolerance (rats) (11)
Genetically modified (FasL-transduced) donor DC prolong organ allograft survival (mice) without adjunctive immunosuppression (12)
Immature or mature donor DC combined with anti-CD154 mAb induce heart transplant tolerance (mice) (13, 14)
Transplant tolerance is associated with in situ inhibition of DC maturation in host lymphoid tissue (rhesus monkey) (15)
Diversely acting anti-inflammatory and immunosuppressive drugs inhibit DC maturation (3)
In vitro pulsing of allogeneic T cells with immature DC induces T regulatory cells (human) (16)
Ag-pulsed autologous DC induce Ag-specific tolerance in normal human volunteers (17)
DC: Crucial Regulators of Immunity
DC are rare, uniquely well-equipped Ag-processing and presenting leukocytes resident within virtually all tissues. They play pivotal roles in innate and adaptive immunity. DC constantly sample the local microenvironment by uptake of self and exogenous Ag. In response to infection/inflammation, or following organ transplantation, DC migrate from peripheral sites to T cell areas of secondary lymphoid tissue. This translocation is associated with their maturation, whereby surface expression of major histocompatibility complex (MHC) and costimulatory molecules (e.g. CD40,80,86) necessary for T cell activation is up-regulated. DC produce the potent T helper (Th) cell-driving cytokine (IL-12) and are the only APC that can activate naïve T cells. Apart from their role as the principal instigators of immune reactivity, DC appear to be important in central and peripheral T cell tolerance. Thus, thymic DC are well recognized to delete potentially self-reactive T cells, whereas immature DC freshly isolated or propagated from nonlymphoid tissues, such as the liver, respiratory tract or intestine, exhibit T cell tolerogenicity. Harnessing and optimizing the potential tolerogenicity of DC may lead to improved outcomes in cell and organ transplantation, including the alleviation of chronic rejection. Hematopoietic growth factors that potently and selectively mobilize DC may assist in reaching this goal.
Mouse and Human DC Compared
The origin of DC is a controversial issue in immunology. As leukocytes, DC are unique in the sense that they can be produced from either common lymphoid or common myeloid progenitors. Classic ‘myeloid’ DC, identified initially in murine secondary lymphoid tissue, can be isolated from virtually all tissues or propagated from blood or bone marrow with GM-CSF and IL-4. In mice, as in humans, several DC subsets have been described, including Langerhans cell-derived DC, CD8α– (myeloid) and CD8α+ DC, B cell-like DC and plasmacytoid DC (18–20). All subsets express CD11c, MHC class II and CD40, but exhibit differences in several other parameters. Initial findings suggested that CD8α+ DC (as opposed to classic CD8α– myeloid DC) were comparatively weak stimulators of CD4+ and CD8+ T cells, and could promote apoptotic death in alloreactive T cells via surface CD95 ligand (Fas ligand; FasL). This lead to speculation that these DC were involved in peripheral tolerance. Both CD8α– and CD8α+DC, however, mature into potent immunostimulatory APC. CD8α+ CD11b– DC secrete IL-12, prime naive CD4+ T cells to secrete Th1 cytokines, and promote anti-viral immunity by inducing cytotoxic T lymphocytes (CTL) via secretion of interferon-γ (IFN-γ) (21). CD8α– CD11b+ DC, on the other hand, are strong stimulators of Th2 cell responses and secrete neither IL-12 nor IFN-γ. The murine CD8α+ DC subset has been shown to cross prime T cells (i.e. to take up, process and present exogenous Ag on class I MHC) and, at very low relative concentrations, to inhibit the immunostimulatory capacity of CD8α– DC for delayed-type hypersensitivity responses (22). Whether or not this same cell type is capable of cross-tolerance has yet to be determined.
Recent phenotypic and functional analyses show considerable differences between murine and human DC. At least five types of human DC have been described: myeloid or ‘monocytoid’ DC [CD11c+, CD1a+ and CD123– (IL-3 receptor-α)], ‘plasmacytoid’ DC that are CD11c–, CD1a– and CD123+, Langerhans cell-derived DC (CD11c+, CD1a+ and CD123–), B cell-like DC, and follicular DC-nonleukocytes that retain immune complexes in B cell follicles and that are believed to play an important role in immunologic memory. All express MHC class II and CD40. Both human monocytoid and plasmacytoid DC (sometimes referred to as DC1 and DC2, respectively) can polarize to either Th1 or Th2 responses. In humans, plasmacytoid DC are the principal IFN-producing cells of the immune system. Moreover, IL-12 production is a feature of human DC1 as compared with murine CD8α+ DC. Current understanding of the differentiation, classification and function of murine and human DC subsets has been reviewed comprehensively (19, 23, 24). Although it has been speculated that specialized ‘regulatory’ DC play a role in peripheral tolerance, present evidence is more consistent with the role of DC in tolerance being dependent on their stage of development or maturation (25).
DC and Peripheral Tolerance
Through their capacity to efficiently ingest exogenous Ag and to present processed Ag to autologous T cells, DC are key instigators of the body's defense system. However, they are not constantly called upon to instigate defense against invading pathogens. In the normal steady state, DC constitute a relatively constant proportion of cells within afferent lymph – trafficking from peripheral sites to draining lymphoid tissue. Important new light has been shed recently on the purpose of this cellular migration, by recognition that immature DC are highly efficient at the uptake of apoptotic bodies in healthy tissues. Although DC may phagocytose necrotic as well as apoptotic bodies, only material derived from the former stimulates T cell immunity (26). This suggests a mechanism by which tolerance to self Ag (as acquired by DC in the form of apoptotic bodies generated by constant cell turnover) may be achieved, and a function for immature tissue-derived DC at those times when they are not required for active body defense. Evidence in support of this concept was provided by Huang et al. who identified apoptotic bodies of intestinal epithelial cell origin within rat DC trafficking to draining mesenteric lymph nodes (27). This putative role for DC in peripheral tolerance may be critical in the maintenance of self-tolerance to Ags that cannot be presented by thymic DC within the neonatal period (such as mature ovarian or mature breast tissue Ags). Using DC loaded with antigenic peptide via the multilectin receptor DEC (CD) 205, Hawiger et al. have shown that, under steady-state conditions, such peptide-loaded DC do not induce Th polarization or either sustained T cell expansion or activation in vivo. Rather they induce Ag-specific peripheral tolerance (T cell deletion/unresponsiveness) (28). How, precisely, Ags acquired within the periphery and transported to lymphoid tissues by DC are rendered nonimmunogenic is not clear. A possible role for murine CD8α+ DC resident within draining lymph nodes in T cell deletion has been suggested (29). Given this information, the ability to target Ags to DC and control their function has significant implications for the development of therapies for allo- and autoimmunity.
Targeting the Role of DC in Direct and Indirect Allorecognition
Rejection of organ allografts is a process associated, traditionally, with the migration of interstitial donor ‘passenger’ leukocytes to recipient lymphoid tissue. At the time of transplantation, various leukocytes (including B lymphocytes, monocytes, DC) and other cells, such as endothelial cells, make up the APC constituency of the graft. Migration of donor-derived DC into host lymphoid tissue allows direct presentation of highly immunogenic, donor-derived MHC Ags to recipient naïve T cells. Mature DC are the most potent APC, capable of directly activating naïve, alloAg-specific cytotoxic CD8+ T cells via CD40 signaling, without CD4+ T cell help (18). With respect to the indirect pathway, by which exogenous Ag is cross-presented by DC, resulting in generation of MHC class I peptide complexes, there is new evidence that in vitro, mature DC can tolerize CD8+ T cells in the absence of CD4 Th cells, or a stimulus for CD40 (30).
While donor DC are clearly important in direct allorecognition, recipient DC also play a significant role in graft rejection via the indirect pathway. Indirect allorecognition occurs when APC of host origin present donor peptides to host T cells in the context of recipient MHC molecules. The relative contribution of direct and indirect allorecognition to murine skin graft rejection has been examined recently (31). 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 the manipulation of donor-derived DC to prevent acute graft rejection and to promote tolerance induction. Indirect allorecognition on the other hand, is thought to be of greater significance in the pathogenesis of chronic rejection (32). Thus, manipulation of host-derived DC by pulsing with donor MHC class I peptide to promote their tolerogenicity in organ and pancreatic islet cell transplantation, as described in rats (33), may be an equally important approach to therapy of this process.
Mechanisms by Which DC May Promote Transplant Tolerance
T cell deletion
The critical role that DC may play in determining the balance between transplant tolerance and rejection is indicated by murine liver transplant studies in which the number of potential allostimulatory donor DC within the graft is dramatically increased by treatment of donors with the potent endogenous hematopoietic growth factor, fms-like tyrosine kinase 3 ligand (Flt3L). When Flt3L is administered to donor mice, liver allografts that would otherwise be accepted without immunosuppression, are rejected acutely. Studies of spontaneously accepted murine liver allografts in the B10 (H2b) to C3H (H2k) strain combination suggest that early apoptosis of alloreactive T cells (day 2–4 post-transplant) is associated with successful liver engraftment and induction of donor-specific tolerance. When acute liver rejection is induced by donor treatment with Flt3L (34), or by recipient treatment with IL-12 or IL-2 (35), T cell apoptosis within the liver graft is dramatically reduced. This suggests that an active process of donor APC-induced activation-induced cell death (AICD) may facilitate the induction of liver transplant tolerance. It also raises the possibility that potentiation of this process might promote allograft survival in various types of organ transplantation. Blockade of costimulatory molecule expression on allogeneic DC promotes their capacity to induce AICD in alloactivated T cells (36). Moreover, the enhanced tolerogenic potential of immature donor DC when combined with anti-CD40L (CD154) mAb in heart graft recipients is associated with increased apoptotic death of graft-infiltrating cells (37). The importance of AICD in the establishment of peripheral T cell tolerance to alloAgs has been demonstrated recently in liver and pancreatic islet cell transplantation (35, 38), and also by administration of donor leukocytes in liver transplant models (39).
Induction of T regulatory cells/T cell anergy
A mechanism whereby immature donor DC (that deliver signal 1 in the absence of signal 2) could modulate responses to alloAgs within recipient draining lymphoid tissue is by induction of T regulatory (T reg) cells. A subset of CD4+ T reg cells (Tr1 cells) was identified after cloning of alloAg-activated T cells exposed to IL-10 (40). These cells secrete IL-10 and transforming growth factor-β (TGF-β), but low levels of IL-2, and no IL-4. Under noninflammatory (steady state) conditions, Ag transported to the draining lymphoid tissue via immature DC may be presented to T cells in the absence of costimulatory molecule expression (signal 2) and Tr1-like cells may be generated (see Figure 1). Adding credence to this hypothesis are recent observations in normal human subjects given a single intravenous (i.v.) injection of immature, autologous, monocyte-derived DC pulsed with influenza matrix protein. This resulted in the generation of matrix protein-specific IL-10-producing T cells (17). Furthermore, repeated stimulation of naïve human cord blood T cells with immature allogeneic DC can induce IL-10-producing CD4+ T reg cells (16). The implication for transplantation of these and related animal studies is that immature DC are capable of inducing Ag-specific T cell hyporesponsiveness in vivo, by what may involve an active regulatory mechanism. Conversely, these observations carry an important caveat for DC therapy of cancer, in that presentation of tumor Ag by immature DC may (unintentionally) inhibit an effective anti-tumor response (41). A murine B lineage-associated DC derived from liver-resident progenitors has been shown recently to induce Tr1-like cells in vitro, and to prolong organ allograft survival (42). Using CD4+ T cell lines, human plasmacytoid DC precursors (pDC2) enriched from peripheral blood have been shown to induce T cell anergy in vitro (43), prompting the authors to suggest that these DC are involved in maintenance of peripheral T cell tolerance and have potential for suppression of allograft rejection.
Using DC to skew the immune response from Th1 towards Th2 cell predominance is another theoretical approach to harness DC for anti-rejection therapy. In a nontransplant setting, prevention of type-1 autoimmune (insulin-dependent) diabetes by GM-CSF/IL-4 generated autologous DC in nonobese diabetic (NOD) mice was associated with immune deviation (44). Application of this approach, using Th2-inducing donor-derived DC in humans (45), might be successful in the modification of bone marrow (BM) or organ allograft rejection (see Figure 1).
DC and Intrathymic T Cell Tolerance
Utilizing surrogate thymic DC to promote alloAg-specific T cell hyporesponsiveness is another novel and promising strategy. This approach has been tested in a pancreatic islet cell transplantation model. Injection of host BM-derived DC pulsed with a donor MHC class I peptide into thymi of streptozotocin-induced diabetic Wistar–Furth rats, in combination with anti-lymphocyte serum therapy, resulted in permanent (> 200d) islet allograft survival (46). As the allopeptide used in this study was presented by host DC, the success of this approach underscores the importance of the indirect pathway of allorecognition in acquired thymic tolerance. The underlying mechanism is not yet known, but may be linked epitope suppression. Recently, similar prolongation of heart allograft survival was achieved when BM-derived host DC pulsed with immuno-dominant allopeptide were delivered intravenously, thereby circumventing the limitations imposed by intrathymic administration (33).
DC Immaturity In Situ and Allograft Tolerance
Insight into the role of the maturational status of DC in organ transplant tolerance was provided recently by studies in a nonhuman primate renal transplant model. In these experiments, rhesus macaque monkeys were treated with a combination of anti-CD3 immunotoxin (IT) mAb and a 15-d course of deoxyspergualin (DSG), commencing 4h pre-transplant (15). This combination, but not IT alone, was associated with the development of long-term (> 3years) 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 the gene transcription regulatory protein nuclear factor (NF)κB within secondary lymphoid tissue DC. It also blocked their phenotypic maturation (CD83 and CD86 expression), maintaining these cells in an immature, potentially tolerogenic state. This effect on DC in situ was transient, suggesting that the timing of DC immaturity may be critical in the induction of long-term transplant tolerance.
Further credence has been given to the concept that immature donor DC can promote allograft survival and even induce transplant tolerance (8, 9). Thus, BM-derived DC generated from B10 mice with low concentration GM-CSF (20 U/mL) for 8d (GM-CSFlowDC), retained an immature phenotype and were resistant to maturational stimuli, such as LPS, TNF-α and CD40 ligation. Furthermore, when 5 × 105 of these GM-CSFlow donor DC were administered i.v. to fully allogeneic CBA recipients, 7d pre-transplant, they showed impressive ability to prolong cardiac allograft survival indefinitely (> 100d) (47). These studies confirmed earlier findings (9) that the therapeutic effect of immature donor myeloid DC was strictly dependent on the temporal relationship between their administration and that of the graft. Thus, the powerful therapeutic effect was lost when the DC were administered on day-3, day-14, or day-28 relative to grafting. Further studies are required to elucidate the factors that restrict the in vivo efficacy of ‘tolerogenic’ DC. On the other hand, in addition to ATG, anti-T cell mAbs and costimulation blocking agents, several novel approaches have been shown to enhance the potential of DC to prolong allograft survival.
Novel Approaches to Promoting DC ‘Tolerogenicity'
Both anti-inflammatory (e.g. salicylates) and immunosuppressive drugs (e.g. corticosteroids) that inhibit DC maturation in vitro have been shown to mediate this action by suppressing the nuclear translocation of NF-κB (3, 48) that is critical for DC maturation. An alternate means to promote an immature state is to specifically target the NF-κB cell activation pathway by antisense oligonucleotides. Short oligodeoxynucleotides (ODN) with consensus binding sequences to NF-κB inhibit DC allostimulatory capacity by blocking NF-κB translocation. This, in turn, inhibits cell surface costimulatory molecule expression. Murine donor BM-derived DC treated in vitro with NF-κB ODN for up to 36h and administered to fully allogeneic recipients as a single i.v. dose, 7d before organ transplantation, significantly prolong heart graft survival (49).
We and others have genetically modified DC to promote allogeneic T cell hyporesponsiveness in vitro and in vivo using a variety of viral and nonviral vectors and gene products [IL-10 (viral and mammalian), CTLA4Ig, FasL, TGF-β] (50–53). Interestingly, mixtures (1:1) of adenoviral (Ad) IL-10 and Ad TGF-β -transduced donor DC administered by portal venous infusion, 36h before renal transplantation, significantly prolong graft survival (54). Also, prolonged murine cardiac allograft survival has been reported by transduction (and repeated dosing) of donor-derived DC over-expressing FasL that promotes the death of alloactivated T cells (12). Furthermore, ‘killer’ DC – DC hybrids created by fusing donor – and recipient-derived murine DC and expressing FasL and MHC class I and II of both donor and recipient origin (55) may have potential for the prevention of alloimmune responses in organ transplantation.
DC Subsets and the T Cell Response
The capacity of distinct DC populations to differentially affect T cell activation provides a potential means to modify organ transplant outcome. A summary of the reported effects of murine DC subsets on allograft survival in the absence of any anti-rejection treatment is shown in Table 2 (4, 8, 9, 42, 47, 56). In mice and rats, only limited work has been reported on cells other than myeloid DC. When highly purified murine CD8α+ donor strain DC were administered i.v. to allogeneic recipients, 7d prior to vascularized cardiac allografting, mean graft survival time was prolonged significantly from 11d (controls) to 26–29d in the absence of immunosuppressive therapy (56). Interestingly, this therapeutic effect was not dependent (as it was with CD8α– DC) on the state of donor CD8α+ DC maturation.
Table 2. Evidence that distinct donor DC subsets can prolong mouse allograft survival in the absence of immunosuppressive/myelosuppressive therapy
In humans, G-CSF-mobilized pDC2 that selectively induce Th2 cells, and that can be purified following leukopheresis of live donors (45), offer potential for the induction of tolerance in recipients of hematopoietic stem cell or organ allografts. It has been speculated that the presence of such DC progenitors within G-CSF-mobilized donor BM plays a significant role in the prevention of graft-vs.-host disease in recipients of these grafts (45). Whether an anti-rejection effect might be seen with infusions of purified ‘tolerogenic’ donor DC in organ transplantation is currently unknown, but worthy of examination. A cautionary observation, newly reported in mice, is that when compared with normal donor BM, infusion of unfractionated growth factor (Flt3L)-mobilized donor BM (containing markedly enhanced numbers of in vivo immunostimulatory donor DC) in combination with anti-CD154 mAb leads to inferior graft survival (57).
Chimerism and DC
Induction of hematopoietic chimerism has long been an integral part of experimental tolerance-induction strategies (58). However, the long-term persistence of small numbers of donor-derived hematopoietic cells (microchimerism) in successful, conventionally immunosuppressed graft recipients first described by Starzl et al. in 1992, lead to the hypothesis that microchimerism was essential for the induction of transplant tolerance. This concept, and the two-way paradigm of evolving mutual immunologic unresponsiveness (59, 60), has provided the basis for extensive clinical investigation of the augmentation of natural microchimerism (using donor BM infusion at the time of organ transplant) in an effort to improve graft outcome. Donor BM contains a significant population of immature DC and their precursors, that may be essential for the desired immunomodulatory/tolerogenic effect of donor BM. Unfractionated human BM contains a mixture of DC subsets and their precursors. It is therefore possible that, as suggested by recent mouse studies (56, 61), potent in vivo antagonistic effects of these subsets may frustrate the desired clinical effect. Effects of unmodified donor-derived BM cells on organ transplantation in conventionally immunosuppressed, noncytodepleted patients have been reported. A trend has been noted toward donor-specific hypo- or unresponsiveness in human graft recipients given donor BM (62, 63) and preliminary reports have suggested a reduction in rejection incidence favoring the BM augmentation treatment arm in cardiac and lung transplantation (64). It remains to be seen whether in these patients, significant long-term differences in graft survival, including incidence and severity of chronic rejection, or dependency on immunosuppressive drugs, will emerge between those treated with adjuvant BM and those given standard therapy. As discussed recently (65) modifications in timing and dosage of immunosuppression may allow the value of donor hematopoietic cells to become apparent. At the same time, the possible testing of human DC therapy must wait further appropriate pre-clinical evaluation.
Issues that Remain to be Addressed
Although DC-based strategies have shown considerable promise in experimental organ transplantation, including reports of indefinite graft survival in rodent models, there are as yet no reports of evaluation of DC therapy in a clinically relevant large animal model. Direct assessment of DC tolerogenicity, particularly in renal-allografted nonhuman primates is therefore a priority. The contribution of indirect allorecognition, as mediated by recipient DC in allograft rejection alone will necessitate adjuvant immunosuppression to inhibit recipient DC function. Evidence from rhesus monkey renal transplant studies suggests that DC maturation arrest might only be required for a short period (15). Furthermore, it is inconceivable that DC-based therapy could reach a trial in human transplantation without immunosuppressive cover. As suggested by murine studies (13), co-administration of costimulation blocking agents (such as anti-CD154 mAb) with immature donor DC can potentiate the beneficial effect of these cells, coupled with targeting of recipient APC. Remarkably, the impact of anti-CD154 mAb in conjunction with donor DC in achieving indefinite organ allograft survival may not be dependent on sustained immaturity of the donor cells (14, 37).
Genetic engineering of donor or recipient DC to promote allogeneic T cell hyporesponsiveness/deletion has shown promise both in vitro and in vivo, although many challenges remain. As yet, there is no compelling evidence from either small or large animal transplantation studies, that genetic modification of donor DC with a single gene is capable, of itself, of inducing donor-specific tolerance. By contrast, a single injection of immunosuppressive cytokine (IL-4)-transduced DC has been shown to inhibit established experimental autoimmune disease (collagen-induced arthritis) in rodents (66, 67). Questions of the type of DC, the optimal route for DC delivery, choice of vector, dosage of DC required to inhibit T cell immune reactivity, timing of DC therapy and optimal combination therapy remain to be resolved.
The differentiation pathways that generate DC in vivo remain largely unknown. Thus, further understanding of the physiologic development of DC, both in vitro and in vivo, will aid in future assessment of their therapeutic manipulation/potential. Ongoing research into the role that DC play in the induction of tolerance to self or exogenous Ag should continue to provide insights/directions for DC-based allospecific therapy. If means can be found to mimic the induction and maintenance of peripheral tolerance in the steady state and perhaps also acquired thymic tolerance, using DC-based strategies, then biologic therapy of graft rejection without continued dependence on immunosuppressive drugs may be a feasible goal. Amongst these exciting possibilities are testing of the immunomodulatory properties of DC that have captured donor-derived apoptotic bodies (29) and strategies for induction of T reg cells (16, 17, 42) as novel means for inducing DC-mediated T cell hyporesponsiveness. Recent evidence also suggests that in vitro conditioned pig DC (IL-10 or dexamethasone-treated immature DC) are capable of inducing xenospecific (anti-pig) T cell tolerance in vivo (68).
Experimental evidence at both the in vitro and in vivo levels, including nonhuman primate studies, suggests that DC can be manipulated to promote alloAg-specific T cell hyporesponsiveness. The challenge is now to apply the basic knowledge that has been accrued to the development of novel DC-based therapy, which may facilitate induction/maintenance of clinical transplant tolerance.
Don and Lorraine Jacquot Travelling Fellowship, Royal Australasian College of Physicians, CJ Martin Fellowship, National Health and Medical Research Council of Australia (PTC), National Institutes of Health grants DK49745, AI 41011 and the Roche Organ Transplantation Research Foundation (ROTRF #13068349) (AWT).