Dendritic Cell Control of Immunity

Author: Octav Cristea

Introduction

The purpose of this paper is to review evidence for the ability of dendritic cells (DC) to
induce T cell activation or T cell tolerance with emphasize on cancer. DC are the most
potent of antigen presenting cells (APC), having the ability to activate naïve and memory
T cells about 100-fold higher than macrophages (1, 2). These functions have prompted
investigation of mechanisms by which DC promote immune responses, and how to use
these cells therapeutically. Conversely, DC have been shown to have a role in induction
of tolerance and clonal deletion (3). Classically, myeloid DC have been thought to
possess T cell stimulatory activity while lymphoid DC as being T cell suppressive and
responsible for negative selection in the thymus. While DC can be obtained by
differentiating CD34+ cells in culture (4), in vivo it is more likely that DC arise from
CD14+ monocytes differentiating into the myeloid subset (5), CD34 cells differentiating
directly into a circulating DC (5a) or a thymic precursor differentiating into a lymphoid
subset (6).

The concept of myeloid and lymphoid DC subsets is based on differential
functional characteristics, anatomical location, and surface marker expression between
the two types (7). Recent data suggests the division between these two subtypes may not
be clear cut (reviewed in 8). Evidence exists that myeloid derived DC which are CD8-
can be induced to express CD8 during maturation (9) and that a myeloid precursor can
give rise to both myeloid and lymphoid DC (10). One possibility is that DC lineages may
not be different entities but cells at different stages of maturation. However, in order to
have meaningful discussions, the myeloid and lymphoid lineages will be treated as
separate subsets based on surface marker expression. In the murine system, lymphoid
DC will be identified as DEC 205+, CD11b-, CD8+, while myeloid DC will be
characterized by the phenotype DEC 205-, CD11b+ (7). In the human system, the
identification scheme proposed by Faratian et al will be used, in which lymphoid DC are
CD4+, CD11c-, CD123+ (IL-3Rα), and myeloid DC are CD1a+, CD11c+, CD83+ (11).
Throughout the chapter, the term “immature” and “mature” DC will be used, this refers to DC of the myeloid lineage, since little is known about immature lymphoid DC.

T Cell Activation vs Tolerance

T cell activation is critical for immune responses to ensue. Such activation is usually
detected by the ability of the T cell to proliferate (CD4 T cell), induce cytotoxicity (CD8
T cell) or to secrete cytokines (both CD4 and CD8). T cell activation requires at least
two signals, one through the TCR and another through a costimulatory molecule such as
CD80 or CD86 on the APC. Activation through TCR alone will drive the cell into a state
of unresponsiveness termed anergy (12). Molecular mechanisms underlying the anergic
state are beyond the purpose of this essay but are discussed in the following references
(13, 14, 15). Anergic T cells can be taken out of their unresponsive state by
administration of the cytokine interleukin-2 (16). This observation provides rational for
the utilization of immune augmentative therapies in diseases, such as cancer, where
therapeutic gain may occur by activation of anergized T cells. Conversely, the ability of

T cells to exit a state of anergy provides opportunity for autoimmunity to arise by
breaking down of self tolerance through stimulation of autoreactive T cells.
Anergy of T cells may also be induced by crosslinking of inhibitory receptors on T cells.
One such receptor is CTLA-4 which binds with high affinity to CD80 and CD86 on the
APC. Recently, it was demonstrated that T cell inhibition through CTLA-4 can occur
through two mechanisms: intracellular signaling of the CTLA-4 molecule, and CTLA-4
blocking activation of CD28 on the T cell by virtue of competition for CD80 and CD86
(17).

While anergy implies T cell unresponsiveness, tolerance to an antigen may also come
about by death of the antigen-specific T cells. In an attempt to induce tolerance, Min et
al administered Fas ligand transfected donor DC into recipient mice receiving cardiac
allografts. Mice receiving transfected donor DC displayed prolonged graft survival only
when the DC were of donor origin (18). A naturally occurring example of tolerance
induced by T cell death is seen in pregnancy. Using H-Y-specific TCR transgenic mice,
Jiang et al demonstrated that maternal T cells specific for fetal antigens decrease in a
specific manner during pregnancy. This decrease was partially explained by the presence
of Fas ligand on the maternal cells in the placenta (19).

In order to provide a framework for understanding DC-T cell interactions several of the
main molecules on T cells responding to dendritic cells are listed in Table I.

Table I. T Cell Molecules Associated With Activation and Tolerance
Molecule Inducible (I)/Constitutive © Function Reference
CD28 C Costimulatory (12)
ICOS I Costimulatory (20)
CTLA-4 I Inhibitory (21)
CD30 I Inhibitory (22)
PD-1 I Inhibitory (23)
Fas I Death of T cell (24)
DR-4 C Death of T cell (25)

Activation of T cells by DC

The myeloid DC is exquisitely suited to act as the immune system’s watchtower against
invading pathogens. This type of DC has the capability to endocytose antigen in it’s
immature state, become partially activated by certain “danger signals” and home into the
lymph node where it becomes fully activated and presents antigens to T cells (26). DC
activitory signals could be heat shock proteins, bacterially derived products, debris from
dead cells and various molecular entities associated with tissue damage. The ability of
the DC to potently activate T cells is dependent on several key features the mature DC
possesses:

a) Mature DC highly express MHC I and MHC II. These molecules contain peptide
antigens derived from the intracellular (cytosolic) and extracellular (endosomal)
compartments. The combination of MHC I-peptide, and MHC II-peptide can provide a
primary signal for activation of CD4 and CD8 T cells. The need for high levels of MHC
during antigen presentation is exemplified in experiments where expression of MHC II is
suppressed on APC by exogenous administration of CLIP, this results in weaker recall
response and ability to ameliorate autoimmune disease (27). In addition, DC have the
unique ability to acquire extracellular antigens, process them through the cytosolic
pathway, and present them on MHC I in order to activate CD8 T cell (28). This ability is
important in priming cytotoxic T cell responses to antigens which are present inside cells
that may not have the ability to activate the CD8 T cell by themselves. One example is
tumor antigens, which are in many cases intracellular. Most tumor cells have
deficiencies in TAP, low expression of MHC I, and absence of costimulatory molecules
such as CD80 or CD86 (29). This would be the tumor cell it’s self a poor stimulator of
CD8 T cells. Apoptotic bodies from tumor cells can be picked up by DC, processed by
the DC in the cytosolic pathway, and the DC can activate the tumor-specific CD8 cell
(30,31). Once the CD8 cells are activated by the DC, they can start to recognize the
tumor and induce an appropriate response.

b) Mature DC express costimulatory molecules. It is established that DC are the most
potent expressers of the costimulatory molecules CD80 and CD86 (2). However, a more
novel finding is that the costimulatory molecule CD86 is found in close apposition with
MHC I and MHC II inside the MIIC and CIIV compartments of immature and mature
DC, respectively (32). In the same study, confocal microscopy was utilized to examine
surface distribution of CD86 in comparison to MHC I and MHC II. All three molecules
seemed to be aggregated in raft-like structures. This was in contrast to the control surface
molecule, CD18, which was evenly distributed across the membrane. From this data it is
tempting to speculate that DC aggregate immunologically important molecules into
microdomains which may serve as counterparts for rafts and/or supermolecular activation
complexes (SMACs) on T cells (33).

Immature DC have been shown to express a costimulatory molecule, LIGHT, which is
the ligand for the herpesvirus entry mediator and lymphotoxin ß receptor found on T
cells. Blockade of LIGHT on DC during allogenic mixed lymphocyte reaction (MLR),
suppresses proliferation of responding T cells. Furthermore, exogenously administered
LIGHT can costimulated anti-CD3 induced proliferation in naïve T cells (34). The idea
of immature DC costimulating T cells is contradictory to findings, which will be
described below, that suggest immature DC are tolerogenic and inhibit MLR.
LIGHT is a member of the TNF-α superfamily. Several other costimulatory molecules
of this superfamily are found on DC. 4.1BB ligand (4.1BBL), is expressed on activated
DC and it’s receptor, 4.1BB, is found on activated T cells. Gramaglia et al showed that
in pigeon cytochrome c-pulsed fibroblasts, transfection with 4.1BBL was as effective at
stimulating proliferation of T cells as transfection with CD80 (35). The importance of
costimulation through 4.1BB in vivo was demonstrated in two models of highly
aggressive cancers where administration of crosslinking anti-body to 4.1 BB was able to
induce effective immune recognition and destruction of established murine tumors (36).
OX-40 ligand (OX-40L) is a member of the TNF-α superfamily which is expressed on
DC and can activate T cells to proliferate and secrete cytokines by binding to OX-40
(37). Crosslinking of CD3 can induce proliferation of T cells from a CD28 knockout
mouse when OX-40L is added, thus indicating costimulation can occur in absence of
CD80/CD86 (38).

Suppression of T cells by Lymphoid DC

Lymphoid DC are important for negative selection of autoreactive T cells in the thymus.
This as been demonstrated using an transgenic system in which MHC class II I-E
molecules were specifically target to DC by placing the gene under control of the CD11c
promoter. This approach demonstrated that I-E expression on thymic DC was sufficient
to negatively select I-E reactive CD4+ T cells (3). The ability of thymic DC to induce
tolerance to antigen has been demonstrated in an animal model of autoimmunity. Khoury
et al protected rats from onset of experimental allergic encephalomyelitis (EAE) by
intravenous injection of thymic DC pulsed with peptides from the autoantigen myelin
basic protein (MBP) (39).

Lymphoid DC can also be found in peripheral organs such as the spleen and liver based
on expression of DEC 205+, CD11b-, CD8+, in the mouse or CD4+, CD11c-, CD123+ (IL-3Rα) in the human. Hepatic allografts in mice have a comparatively longer survival than other organs. This has been attributed in part to the high proportion of resident lymphoid DC (40, 41). Liver derived lymphoid DC, in contrast to bone marrow derived myeloid DC induce production of the tolerance promoting cytokines IL-4 and IL-10 in T cells (42). Unfortunately, a successful strategy for expansion of these DC is lacking. An
attempt at such an expansion was made by systemic administration of the DC expanding
factor, Flt-3 ligand, to donor mice. Instead of promoting tolerance the grafts were
rejected at an accelerated rate (43).

Three mechanisms of lymphoid DC-induced tolerance:

a) Expression of Fas ligand (FasL). Activated T cells express the Fas receptor which
allows them to undergo activation-induced cell death. The importance of this receptor is
seen in a mouse strain, lpr, which develops autoimmunity and hyperproliferation of T
cells (44). Splenic lymphoid DC have been demonstrated to kill activated CD4 T cells
through expression of FasL (45). The relevance of this finding to DC biology remains to
be confirmed in other systems. However, as described above, the transfection of myeloid
DC with FasL can antigen-specifically prolong allograft acceptance (18). Therefore
further studies in this area are needed.

b) Suppression of IL-2 production. Kronin et al demonstrated that splenic lymphoid
DC are poor stimulators of MLR for CD8 cells. This was not attributed to FasL
expression on the DC but by the ability of the DC to block IL-2 production from the T
cell. No mechanism was proposed (46).

c) Expression of inhibitory receptors. It has been demonstrated for many years that
portal vein infusion of donor-specific lymphocytes prolongs renal allograft survival. A
molecule identified on lymphoid dendritic cells, OX-2, is speculated to be the cause for
this tolerogenic effect (47, 48). Administration of blocking antibodies to OX-2 inhibits
the ability of donor-specific lymphocytes to prolonged allograft survival (49). In the
same study it was demonstrated that infusion of donor lymphocytes into the portal vein
was associated with increased levels of IL-10 mRNA in the graft, but this was abrogated
by administration of the anti-OX-2 antibody. The OX-2 knockout mouse is susceptible to
collagen induced arthritis and EAE, thus hinting that this molecule may be important for
maintaining of self tolerance (50). In a personal communication, Grant McFadden stated
myxoma poxvirus contains an OX-2 homologue which is essential for virulence. Could
lymphoid dendritic cells possess such potent immune suppressor molecules that even
viruses would want to steal them?

Suppression of T cells by Myeloid DC

An example for the immunologically importance of immature DC is in cancer. Studies
by Rita Young have shown that several murine tumors produce systemic amounts of GMCSF (51, 52) and that administration of blocking antibody to GM-CSF could abrogate tumor-induced immune suppression (53). The mechanism of GM-CSF-induced immune suppression was production of IL-10, TGF-β, and nitric oxide from the immature DC (53a). GM-CSF given to bone marrow cells at low quantities in vitro, in the absence of IL-4, has been demonstrated to cause expansion of immature myeloid DC which promote allograft tolerance (54). These immature DC express low levels of CD80 and CD86, which may in part account for their ability to induce tolerance. Production of GM-CSF has been recently demonstrated in a battery of human tumor cell lines and this has been postulated by Bronte et al to be responsible for suppression of CD8 cytotoxic T cells seen in cancer bearing mice and patients (55).

Immature myeloid DC have been found in circulation of cancer patients, the number of
DC positively correlating with disease stage (56). These DC inhibit T cell proliferation
in an antigen-nonspecific manner, through mechanisms which remain to be elucidated.
Presence of these DCs is dependent on tumor size, since after surgical resection of tumors
their numbers drop (57). Immature circulating DC are also dependent on the
angiogenesis promoting protein vascular endothelial growth factor (VEGF). When
antibody to VEGF is administered clinically, the number of circulating immature DC
drops. Administration of anti-VEGF antibody to murine models of cancer increases the
ability of DC to activate T cell responses and induce tumor clearance (58).

Another system in which DC are tolerance promoting is the ultraviolet B (UVB) light
treated skin Langerhans cells (which are a type of myeloid DC that resides in the dermis).
These cells are capable of inactivating T cells or skewing their response to a tolerogenic
Th2 profile (reviewed in 59). Irradiation of murine Langerhans cells in vitro led the their
inability to present keyhole limpet hemocyanin (KLH) to Ag-specific, H2d restricted Th1
clones, but not to Th2 clones with the same specificity (60). Immunization of mice with
antigen at the site of UVB irradiation leads to induction of antigen-specific tolerance
(61). A postulated mechanism was that UV-treated APC stimulate activation of T
regulatory cells which kill DC through a Fas-dependent manner. UVB-treated DC
secrete an IL-12 antagonistic protein, the IL-12 p40 homodimer which can block the
ability of splenocytes to secrete IFN-γ after treatment with naturally occurring IL-12 (62).
In another system, DC which can not secrete agonistic IL-12 (generated in the presence
of prostaglandin E2 (PGE-2)) prime naïve human T cells to secrete the Th2 cytokines IL-
4 and IL-5, but can not prime secretion of Th1 cytokines such as IL-2, IFN-γ (63).
Chemically, the mechanism by which UVB radiation programs DC to promote
tolerogenic responses is postulated to involve the transformation of urocanic acid from its
nonimmunomodulatory trans-isoform to its immunosuppressive cis-isoform in the skin
strateum corneum (64, 65). Administration of antibody specific to the cis-isoform of
uraconic acid blocked the ability of UVB to suppress the MLR stimulating ability of
Langerhans cells derived from C3H mice (66).

Another method by which DC can inhibit activation of T cells and/or induce tolerance is
through metabolism of the amino acid tryptophan by the enzyme indoleamine 2,3-
dioxygenase (IDO). Since T cell proliferation and activation is highly dependent on
tryptophan, depletion of this amino acid can result in inhibit of T cell function. Munn et
al demonstrated that administration of an IDO-specific inhibitor, 1-methyl-tryptophan, to
allogenically pregnant mice resulted in rapid onset of immunologically mediated
abortion, thus suggesting expression of IDO may be required for maintaining tolerance to
the fetal allograft (67). IDO expression has been found on IFN-γ or T cell activated
human myeloid DC and administration of IDO-inhibitors to these activated DC can
augment their ability to stimulate T cell proliferation (68). Furthermore, abrogation of
the T cell suppressive effects of murine lymphoid DC by artificial ligation of CD40,
results in downregulation of their ability to degrade tryptophan (69).
From this point of the essay and onward, the term “DC” will refer to myeloid DC, since
the experiments described are recent and have not been performed in the lympoid lineage.

Indirect Activation of T cells by DC

The DC can condition the local immunological environment by virtue of stimulating
other cell types in order to facility activation of T cells. One example of such
conditioning is the ability of DC to activate NK cells. In an early study, it was
demonstrated monocytes can expand IL-2-induced NK cell numbers and augment
cytotoxicity (70). Since monocytes are precursors of myeloid DC, Fernandez et al
investigated whether culturing of monocyte-derived DC can stimulate NK cytotoxicity in
vitro. NK cells were endowed with potent tumor-cytolytic activity and IFN-γ synthesis
after coculture with DC in a contact-dependent manner. Furthermore, when both
monocyte-derived DC and Flt-3L-expanded DC were able to induce NK-dependent
antitumor effects in vivo (71). DC activation of NK cells derived from CD34+ cord
blood progenitors was demonstrated to occur via secretion of IL-12 and IL-18 by DC
(72). The possibility of NK activation by DC is particularly interesting in light of recent
observations that human and murine NK cells can be activated by cells expressing CD80
and/or CD86, in a CD28 independent manner (73,74). NK cells can increase the number
of activated T cells by secretion of immunomodulators such as IFN-γ (review in 75).
Another mechanism by which DC may indirectly augment activation of T cells is through
activation of NKT cells. These cells are potent producers of the cytokines IL-4 and IFN-
γ in response to activation of their conserved TCR, Vα14 Jα281 (murine) or Vα24Vβ11
(human) (76). Priming and expansion of NKT cells in humans and mice is most potently
induced when their conserved ligand, α-galactosylceramide, is presented to them in the
context of DC (77,78). In the NKT-DC interaction, IL-12 production by the DC is
required for NKT cell activation. Administration of α-galactosylceramide in mice is
associated with acquisition of antigen-specific cytotoxic activity by CD8 T cells, as well
as induction of activation marker CD69 on CD4 T cells. The importance of NKT cells is
demonstrated in NKT deficient mice administered α-galactosylceramide, in which T cell
activation does not occur (79).

Indirect Tolerance Induction by DC

DC activation of NKT cells can result in the secretion of tolerogenic cytokines such as
IL-4 and IL-10. Using the model of anterior chamber associated immune deviated
(ACAID), Faunce et al demonstrated a DC-like population is found in the spleen after
intraocular immunization. These cells secrete a chemokine, MIP-2, which is responsible
for the chemoattraction of NKT cells which induce systemic T cell tolerance to the
immunized antigen (80). This tolerance can be abrogated by administration of antibodies
to IL-10 (81).

As mentioned above, DC can aquire tolerogenic properties after exposure to UVB. The
possibility of tolerogenic DC (ie DC which possess low levels of costimulatory
molecules) activating NKT cells to secrete Th2 cytokines is implied from in vivo studies
showing activation of NKT cells in absence of costimulation leads to secretion of IL-4
and suppressed IFNγ (82). Moodycliffe et al demonstrated the transfer of UVB-induced
antigen-specific immune suppression by NKT cells. These cells act as regulatory cells
suppressing activation of CD4 and CD8 T cells (83).

In the non-obese diabetic mouse (NOD) and diabetic patients there is a defect in DC
differentiation (84, 85). It will be interesting to determine whether this is in part
responsible for the reduced NKT cell activity in NOD mice (86, 87).

Therapeutic Applications of DC-Induced T cell Activation

The ability of DC to act as “natural adjuvants” (88), has stimulated an intense interesting
their utilization for inducing immune responses against poorly immunogenic antigens,
such as those found on tumors and HIV. The first DC clinical trial was aimed at inducing
immunity toward the idiotypic antigen (ID) found on B cell lymphomas. ID is the
specific antibody recombination which is found on all of the B lymphoma cells but is
patient-specific. Immunization of 4 patients with antigen-pulsed (ID and a control
antigen, KLH) monocyte-derived DC induced cell mediated but not antibody response to
ID in all patients. All patients produced antibody and cell mediated response to KLH. Of
the four patients, at 21 month follow-up, one had no evidence of disease, one had
complete remission, and another had partial remission. Administration of DC was well
tolerated (89).

Although clinical trials with Ag-pulsed DC are in progress for melanoma, prostate cancer
and renal cell carcinoma, results to date have been less than optimal with average
responses occurring in 20-30% of patients (90). This may be due to T cell abnormalities
in cancer patients (91), the production of DC inhibitory factors by the tumor (92), or the
ability of the tumor to induce direct killing of DC (93). When murine DC transfected
with the anti-apoptotic gene Bcl-xl were injected into mice bearing established prostate
cancer, immunologically mediated regression of tumors was observed (94).

A clinical trial using HIV DC pulsed with gag and pol immunogenic peptides was
performed in six HIV patients. Although a cellular and antibody responses to the
peptides were present, no reduction in viral load was seen (95). Expansion of DC using
Flt-3L prior and to, and concominant with HIV-peptide pulse DC vaccine has yielded
stronger antiviral immunity in mice than injection without expansion (96). Clinical trials
using this approach have not been performed.

A caveat in all dendritic cell immunotherapy studies is the origin of the DC. Most
clinical studies utilize monocyte-derived DC grown in the presence of GM-CSF and IL-4
due to the ease of collecting peripheral blood CD14+ cell. DC generated in this manner
have recently been questioned as an ideal immune stimulator due to their defective
migration and NK cell activation abilities, thus generation of DC from other sources may
lead to more profound clinical result (97).

Therapeutic Applications of DC-Induced Tolerance

Clinical trials using DC to induce tolerance have not been conducted. Animal studies
suggest that this approach may be useful. One method of increasing the ability of DC to
induce tolerance is abrogation of their T cell costimulatory capacity. CTLA-4 binds with
high affinity to CD80 and CD86, blocking their interaction with the immunostimulatory
molecule, CD28, on the T cell (17). By treating donor-derived DC with the CTLA-4-Ig,
it is possible to enhance skin graft survival in a rat model of allotransplantation, as well
as to antigen-specifically inhibit T cell proliferation to the alloantigens (98). Using a
similar approach, transfection of CTLA-4-Ig into donor DC can induce acceptance of
BALB/c islet allografts into C57/B6 recipients (99). In these studies the donor-derived
DC were lacking co-stimulatory molecules, however, intragraft DC were not transfected.
The concept of dominant tolerance, perhaps by treated DC inducing T cell anergy must
be addressed in the future. While co-stimulatory molecule expression is needed for
activation of T effector cells it is also required for generation of T regulatory cells.
Blocking the CD80/CD86 interaction with CD28 by administration of CTLA-4-Ig in
NOD mice accelerates disease and leads to a fall in circulating CD4+ CD25+ T regulatory cells (100). Therefore, co-stimulation blockade is an immunomodulatory techniquewhich needs to be more fully understood before clinical application.

Oral tolerance is witnessed by systemic alteration to antigen-specific immune response
after ingestion of the antigen, in part by activation of TGF-β-secreting Th3 cells (101).
Although induction of oral tolerance against self antigens possesses some efficacy in
animal models of autoimmunity, in the clinic this approach has yielded limited results.
An explanation for this may be the high amount of antigen which needs to be ingested to
cause systemic immunomodulation (102). DC are believed to be, at least in part,
responsible for initiating the immunological changes which occur during feeding of
antigen (103). With the idea of augmenting the effect of oral tolerance by in vivo
expansion of DC, Viney et al administered Flt-3L to mice before feeding of the antigen.
This approach reduced the amount of fed antigen needed for induction of oral tolerance
(104). Besides the obvious clinical relevance of these studies, it will be interesting to
determine the characteristics of gut-derived DC which enable them to induce tolerance.
These studies are needed because in some circumstances, feeding of autoantigen can
trigger autoimmunity (105).

Transfection of DC with immunosuppressive molecules may also render them
tolerogenic. IL-10 has been associated with suppression of delayed type hypersensitivity
and induction of tolerance in several systems (106). Donor-derived myeloid DC
transfected with the IL-10 gene allows prolonged survival of renal allografts. The
tolerogenic effects of these DC can be further augmented by transfection of the DC with
TGF-β (107). In contrast, administration of donor-derived immature DC transfected with
the IL-10 gene cause accelerated rejection of cardiac allografts (108). Therefore, the
utility of IL-10-transfected DC as tolerance-inducing cells needs further investigation.

Conclusion

DC have the potential to activate and tolerize T cells depending on lineage (lymphoid or
myeloid), stage of differentiation, expression of surface antigens, and prior manipulation.
Although the study of DC is relatively new, understanding of the biology of these cells
has already provided opportunities to induce and suppress immunological responses at an
antigen-specific level. The clinical utility of DC appears promising in areas such as
cancer immunotherapy, while tolerance-induction therapy is awaiting entrance into the
clinical arena.

---

---

Social & Bookmarking:           

---

1 Comment | Add Comment

TOP

Send us your feedback?