A new paradigm emerging from studies
of lymphocyte regulation proposes that
cells of the immune system use a form of
bidirectional communication, commonly
referred to as reverse signalling, that allows
pairs of ‘co-receptors’ on adjacent cells to
engage in a crosstalk by reciprocally acting
as ligands and receptors. (The term reverse
signalling is used here conventionally to
indicate a two-way communication between
cells or cell types, whereby information is
actually flowing in both directions, but one
direction (‘forward’) is of greater or longer-
standing importance.) Reverse signalling,
thanks to primary ligands having evolved
into ancillary receptors, enables an immedi-
ate feedback to the ligand-bearing cell in
response to the forward signal, and it typi-
cally involves tumour-necrosis factor (TNF)
Reverse signalling also applies to the
triad of co-receptors consisting of cyto-
toxic T-lymphocyte antigen 4 (CTLA4),
CD28 and B7 molecules (CD80 and
CD86), and this enables bidirectional
and univocal conditioning of T cells and
dendritic cells (DCs)2–4. As a result, mouse
and human DCs5 respond to CTLA4
engagement of their surface B7 molecules
with activation of the immunoregula-
tory pathway of tryptophan catabolism6.
This pathway, initiated by the enzyme
indoleamine 2,3-dioxygenase (IDO;
encoded by INDO), is controlled by inter-
ferons (IFNs)7, through the involvement of
poorly characterized transcription factors
of the nuclear factor-κB (NF-κB) family2,8,
and it is often associated with the expres-
sion of the anti-inflammatory cytokine
Renewed interest in regulatory T cells
has focused on the CD4+CD25+ regulatory T
(TReg)-cell population. It has become increas-
ingly clear that these cells not only exist as
natural and adaptive subsets that contribute
to the maintenance of self tolerance, but
that they also have a potential for treating
allergic and chronic inflammatory diseases.
However, the origin, recognition properties
and molecular basis for the suppressive
activity of human and mouse TReg cells are
controversial, as is their relationship to other
populations of regulatory cells. Whereas
some of the inhibitory effects appear to be
mediated by the production of immuno-
suppressive cytokines — including IL-10
— other mechanisms, which mostly operate
in the control of autoimmune and allergic
conditions, involve direct interactions of
regulatory T cells with responding T cells or
Reverse signalling in DCs leading to
the activation of IDO expression following
T-cell contact features as one of the contact-
dependent effector mechanisms of natural
regulatory T cells that express surface
CTLA4 (REFS 15–17), thereby reconciling a
long-established role of IDO in mammalian
pregnancy18 with the more recent role of
IDO in TReg-cell function19. There is, how-
ever, increasing recognition of a broader
and truly immunoregulatory role for IDO in
physiopathology, far beyond its function
in pregnancy or as a grossly immuno-
suppressive mechanism20. The role of IDO
has shifted in importance from that of a
metabolic regulator of tryptophan avail-
ability in local tissue microenvironments, to
one that is central to immune homeostasis
and the plasticity of the immune system
(FIG. 1), with implications for many aspects
of immunopathogenesis21–25. This Opinion
article focuses on the nature and mecha-
nisms of the mutual relationship between
IDO and TReg cells.
Non-canonical NF-κB signalling
The NF-κB family comprises seven struc-
turally related transcription factors that have
a central role in the cellular stress response
and in inflammation by controlling a
network of gene expression26. Although the
NF-κB subunits are ubiquitously expressed,
their actions are regulated in a cell-type-
and stimulus-specific manner, allowing for
a diverse range of effects. Recent molecular
dissection of NF-κB activation has shown
that NF-κB can be induced by the so-called
‘canonical’ (classical) and ‘non-canonical’
(alternative) signalling pathways, leading
to distinct patterns in the individual NF-κB
subunits that are activated and the down-
stream genetic responses that are induced.
IDO and regulatory T cells: a role for
reverse signalling and non-canonical
Paolo Puccetti and Ursula Grohmann
Abstract | The immunoregulatory enzyme indoleamine 2,3-dioxygenase (IDO)
suppresses T-cell responses and promotes immune tolerance in mammalian
pregnancy, tumour resistance, chronic infection, autoimmunity and allergic
inflammation. ‘Reverse signalling’ and ‘non-canonical activation’ of the
transcription factor nuclear factor-κB (NF-κB) characterize the peculiar events
that occur in dendritic cells when T-cell-engaged ligands work as signalling
receptors and culminate in the induction of IDO expression by dendritic cells
in an inhibitor of NF-κB (IκB) kinase-α (IKKα)-dependent manner. In this Opinion
article, we propose that IDO acts as a bridge between dendritic cells and CD4+
regulatory T cells, and that regulatory T cells use reverse signalling and
non-canonical NF-κB activation for effector function and self-propagation. This
mechanism may also underlie the protective function of glucocorticoids in
NATuRe RevIeWs | immunology?
vOLume 7 | OCTOBeR 2007 | 817
© 2007 Nature Publishing Group
Nature Reviews | Immunology
Naive T cell
Natural killer cell
Regulatory T cell
IDO+ pDC or
CD8α+ CD19+ DC
The canonical pathway involves activation
of the inhibitor of NF-κB (IκB) kinase-β
(IKKβ), which leads to phosphorylation-
induced proteolysis of the inhibitor IκBα
and consequent nuclear translocation of
the ReL-A (also known as p65; a subunit of
NF-κB) transcriptional activator in the form
of p50 (also known as NF-κB1)–ReL-A
dimers. In the non-canonical pathway, acti-
vation of IKKα by NF-κB-inducing kinase
(NIK) results in the processing of p100 to
p52 and consequent formation of p52 (also
known as NF-κB2)–ReL-B dimers, which
translocate into the nucleus and activate
Although much attention has been
focused on the pro-inflammatory signal-
ling of NF-κB, recent data indicate that
IKKα and IKKβ could have opposing
roles. Whereas IKKβ mediates NF-κB
activation in response to pro-inflamma-
tory stimuli, IKKα accelerates both the
turnover of ReL-A and its removal from
pro-inflammatory gene promoters28. As a
result, IKKβ is indispensable in the canonical
pathway, whereas IKKα is pivotal in the
non-canonical activation that leads to
resolution of the early inflammatory proc-
ess and to the onset of tolerance to self29 or
adaptive immunity to foreign antigens27,28.
Thus, the cross-regulation between canoni-
cal and non-canonical signalling pathways
is crucial in promoting an optimally pro-
tective response that is balanced between
inflammation and tolerance28.
We have recently found that non-
canonical NF-κB activation is necessary
for the induction of IDO expression in
response to reverse signalling, a require-
ment that might be in part due to the
presence of a putative binding site for
p52–ReL-B dimers in the INDO pro-
moter30. The mouse Indo promoter contains
a putative partial binding site (GGGAGA)
at position –3,566 that is recognized by the
non-canonical NF-κB dimer, p52–ReL-B31,
and this site is conserved in the human
gene (position –2,100). Recently, another
enzyme with IDO-like activity, IDO-like
protein 1 (INDOL1), has been described
in both mice and humans. The mouse and
human genes encoding INDOL1 and IDO
have a similar genomic structure and are
situated adjacent to each other on chromo-
some 8 of mice and humans32. Although
IDO and INDOL1 have similar enzymatic
activities, their expression patterns in tis-
sues are different, yet both mouse Indol1
(positions –3,180; –2,640; –2,024) and
human INDOL1 (positions –3,357; –993)
have multiple GGGAGA sequences and
therefore may also be regulated by
Overall, these unanticipated findings
could have significant implications in
immune regulation, including a potential
role for non-canonical NF-κB signalling in
the development of TReg-cell responses.
The bipolar nature of DCs and IFNs
Immune receptors on plasmacytoid DCs
(pDCs) are potent activators of innate
immunity. Through their ability to produce
type I IFNs, pDCs drive protective antiviral
inflammation and promote the function
of bystander myeloid DCs, B cells, T cells
and natural killer cells33. These responses,
however, have also been implicated in the
induction and exacerbation of the inflamma-
tory process associated with autoimmunity
and allergy34,35. Nevertheless, a protective
function of type I or type II IFNs has also
been demonstrated in several experimental
models of autoimmunity and allergy6,36. The
protective functions of pDCs, and of the
associated type I IFN response, appear to be
an intrinsic ability of the immune system to
co-activate cytostatic mechanisms, induce
the death of pathogenic T cells and polarize
T cells towards a TReg-cell phenotype11,12.
Tryptophan catabolism may, in principle,
fulfil all the requirements to mediate these
functions, including an arrest in T-cell pro-
liferation20, induction of T helper 2 (TH2)-
cell apoptosis37, reversible impairment of
T-cell activity through downregulation of
expression of the T-cell receptor ζ-chain38
and the generation of IL-10-producing
regulatory T cells38.
Figure 1 |?A?model?of?crosstalk?between?dendritic?cells?and?T?cells?via?reverse?signalling.?In local
tissue microenvironments, the activity of indoleamine 2,3-dioxygenase (IDO)-expressing plasmacytoid
dendritic cells (pDCs)3,40 and/or CD8α+CD19+ DCs43,87 driven by type I or type II interferons (IFNs) will
result in the sustained IDO enzymatic production of tryptophan catabolites, collectively known as
kynurenines (Kyn). In turn, Kyn could recruit other cell types to the regulatory response, including pDCs
in which the function of IDO, but not of the Kyn pathway enzymes downstream of IDO, is inhibited
post-translationally64. (Under conditions of post-translational blockade of IDO, the IFNγ-inducible
enzymes of the Kyn pathway can be recruited to a tolerogenic response if cells take up and further
metabolize external Kyn that are downstream of the initial, IDO-dependent degradation product of
tryptophan.) The combined effects of tryptophan starvation, caused by IDO+ pDCs, and the high Kyn
production, resulting from the actions of IDO+ and IDO– pDCs, is expected to have various effects on
target T cells38,88 and other cell types89,90, in part involving the stress-response kinase GCN2 (general
control non-derepressible 2), which is a sensor of amino-acid deficiency38,42. Regulatory T cells could
have a crucial role in establishing an IFNγ-rich environment that activates IDO– and IDO+ pDCs, either
by reverse signalling to pDCs or by direct production of the cytokine91.
818 | OCTOBeR 2007 | vOLume 7
© 2007 Nature Publishing Group
Further regarding the relationship
between type I IFNs and non-canonical
NF-κB signalling in pDCs, it is interesting
to note that although it is possible that
type I IFNs contribute to non-canonical
NF-κB activation by an inhibitory ligand39,
non-canonical NF-κB activation is also
observed in pDCs from mice lacking type I
IFN receptors (in other words, induction of
the non-canonical pathway by the inhibitory
ligand is not entirely dependent on IFNs)30.
Reciprocally, transcriptional regulation of
type I IFN genes in pDCs is controlled by
IFN-regulatory factor 3 (IRF3) and/or IRF7
(REF. 34), and the Irf3 promoter contains a
non-canonical NF-κB binding site recog-
nized by p52–ReL-B dimers31. so it is pos-
sible that non-canonical NF-κB contributes
to type I IFN production in response to an
inhibitory ligand. Finally, owing perhaps to
a required function of sTAT1 (signal trans-
ducer and activator of transcription 1), con-
comitant IFN (type I or type II) and NF-κB
signalling may be necessary for an inhibitory
ligand to condition pDCs to express high
levels of functional IDO30,40.
In conclusion, true to their dual nature,
increasing evidence links type I IFNs with
resistance to specific forms of immuno-
pathogenesis, including those associated
with infection10, by way of a potential bridge
between pDCs and TReg cells41. Here, we raise
the possibility that IDO represents the func-
tional bridge between pDCs and TReg cells38,42
and is, at the same time, a main participant
in maintaining the tenuous balance between
those opposing actions of IFNs in immune
protection and pathology43,44.
TLRs, NF-κB and IDO
Toll-like receptors (TLRs) trigger the
induction of type I IFN production, thereby
providing a crucial mechanism of antiviral
defence. TLRs are evolutionarily conserved
receptors that recognize similarly con-
served pathogen-associated molecular
patterns (PAmPs) present on various micro-
organisms. The role of TLRs as arbitrators of
the discrimination between self and non-self
implies that they have a central role in innate
immunity, as well as in the initiation of
adaptive immunity35,45,46. so far, 13 mouse
TLRs have been described and most, if not
all, of these can trigger signals to activate
NF-κB26. TLRs have varied tissue distribu-
tion and recognize many different PAmPs,
including lipopolysaccharide (LPs), double-
stranded RNA (dsRNA), non-methylated
CpG-containing DNA and flagellin. The
intracellular domain of TLRs has a high
degree of homology with that of the IL-1
receptor, and this shared Toll/IL-1 receptor
(TIR) domain mediates interactions with
downstream signalling adaptors that lead
to the activation of three key transcription
factors — NF-κB, activator protein 1 (AP1)
Of interest, signalling through TLR4, the
receptor for bacterial LPs, neither activates
non-canonical NF-κB signalling per se26 nor
induces IDO expression47,48. By contrast,
signalling through TLR9 activates IKKα
(REF. 49) and induces IDO expression37,50,51.
In a mouse model of systemic lupus erythe-
matosus, the absence of TLR9 results in
exacerbation of the autoimmune disease52.
In addition, IKKα is required for the
development of self tolerance53, and NIK
(an integral component of the non-canonical
pathway) may be necessary for the genera-
tion of autoimmune-preventive TReg cells54.
These data suggest that TLRs, NF-κB and
IDO are intimately linked in the preven-
tion of immunopathogenesis that involves
Reverse and non-canonical signalling
In TReg-cell generation. IDO has a role in the
peripheral generation of regulatory T cells,
under physiological38 or pathological condi-
tions55. Recently, a mechanism has been
identified that intrinsically links maturing
pDCs to the generation of IL-10-producing
regulatory T cells, through TLR-dependent
and TLR-independent pathways, and that
may provide a means to prevent excessive
inflammation during infection14. On the
one hand, a potential role for TLRs, NF-κB
and IDO in TReg-cell generation is consist-
ent with the observations reported above.
On the other hand, the TLR-independent
mechanism appears to be mediated by
CD40 signalling, and CD40 signalling
activates NF-κB to induce IDO expression
under environmental conditions that tip the
balance in favour of the non-canonical path-
way56–59. We have recently found that reverse
signalling through glucocorticoid-inducible
TNF receptor-related protein (GITR) ligand
(GITRL) also activates non-canonical NF-κB
signalling and IDO expression in pDCs30.
Overall, these data indicate that differ-
ent ligands acting on pDCs60 — including
TLR9 ligands49,61,62, CTLA4 (REFS 2,3),
CD200 (REF. 63), 4-1BB ligand22, CD40
ligand12,14,59 and GITR30 — all contribute to
IDO-mediated regulatory T-cell generation
through pathways that converge on non-
canonical NF-κB signalling. It has likewise
been suggested that TReg cells use reverse
signalling and consequent IDO production
by pDCs to expand their own population in
the periphery30,38,64. Accordingly, a model of
TReg-cell generation by IDO+ pDCs can be
envisioned in which the combined actions
of CTLA4+GITR+ TReg cells on the one hand,
and NF-κB signalling in the pDCs on the
other hand, are pivotal in maintaining a
regulatory environment (FIG. 2).
In the gut and airways. Traditionally
recognized for its role in infection7, preg-
nancy18,20, transplantation2, autoimmunity65
and neoplasia21,66,67, the IDO mechanism
has revealed an unexpected potential in
the control of inflammation, allergy and
allergic airway inflammation, which are
all conditions in which pDCs could have a
protective function. The first indication that
IDO is expressed in the normal colon and
is upregulated as a protective mechanism
during inflammation came from an elegant
study showing that inhibition of IDO activ-
ity during experimental colitis resulted in
increased mortality and an augmentation
of the normal inflammatory response68.
Not surprisingly, therefore, the administra-
tion of soluble GITR is highly protective in
this experimental setting69, consistent with
the ability of GITR to induce IDO expression
in target cells30. As predicted by the ‘hygiene
hypothesis’ — that is, a reduction in micro-
bial burden at a young age may predispose
individuals to allergy70 and autoimmunity71
— epidemiological and experimental data
now suggest that certain microorganisms
induce a state of protective tolerance10.
Recognition of commensal bacteria by
TLRs is crucial in maintaining intestinal
homeostasis72. moreover, the commensal
flora could have anti-inflammatory effects
through the inhibition of canonical NF-κB
signalling73,74. A key role for DCs in probi-
otic functionality correlates with reduced
colonic expression of pro-inflammatory
genes and increased expression of IFNγ and
IDO by DCs exposed to probiotic bacteria75.
so TLR-mediated induction of IDO expres-
sion and inhibition of the canonical NF-κB
signalling work together to maintain intestinal
homeostasis in experimental settings.
However, the model that most clearly
shows the protective effects of reverse
signalling and IKKα-dependent induction
of IDO expression in mucosal epithelia is
that of experimental asthma. Allergic asthma
is characterized by chronic inflammation
associated with airway remodelling. This
process results in subepithelial fibrosis, an
increase in smooth muscle mass and an
increase in the number of mucous glands.
Chronically allergic mice develop sustained
eosinophilic airway inflammation and airway
NATuRe RevIeWs | immunology?
vOLume 7 | OCTOBeR 2007 | 819
© 2007 Nature Publishing Group
Nature Reviews | Immunology
Type I IFNs
hyperresponsiveness to inhaled antigen76.
In two experimental models of allergic
asthma, soluble CTLA4 (REF. 77) and a TLR9
ligand37 can independently inhibit airway
eosinophilia and hyperresponsiveness by
regulating TH-cell subsets. In a third model,
TReg cells78, pDCs61 and IDO expression10
all contribute to protection from allergic
asthma. In this model, we found that effec-
tive anti-inflammatory treatment inhibited
TH2-cell responses and allergy, and induced
the expression of forkhead box P3 (FOXP3;
a TReg-lineage specification factor) in CD4+
T cells through mechanisms dependent on
IKKα-induced IDO30. These data suggest
that modulation of IDO expression by com-
ponents of non-canonical NF-κB signalling
— in balance with canonical signalling coun-
terparts — is essential for the maintenance of
TLR-driven immune homeostasis in the air-
ways (FIG. 3a,b). Clinical trials of TLR9-based
immunotherapy are presently ongoing and
suggest that TLR9 ligands may significantly
improve the treatment of allergic diseases79.
In glucocorticoid function. Although gluco-
corticoids have been widely used since the late
1940s, the molecular mechanisms responsible
for their anti-inflammatory activity are still
under investigation. since the discovery of
NF-κB in 1986, and the cloning of the genes
encoding the NF-κB components and IκB
proteins, several molecular studies have
demonstrated that these widely used drugs,
known for their varied therapeutic activities,
inhibit NF-κB activity, usually among other
biological effects. Glucocorticoids act by
binding to the glucocorticoid receptor that,
upon activation, translocates to the nucleus
and either stimulates or inhibits the expres-
sion of genes encoding anti-inflammatory
proteins or pro-inflammatory transcription
factors. It is widely believed that the anti-
inflammatory properties of glucocorticoids,
as well as non-steroidal anti-inflammatory
drugs, are in part related to their inhibi-
tion of NF-κB80. A recent study in patients
with asthma suggested that glucocorticoid
treatment is not only immunosuppressive
and anti-inflammatory, but that it also
promotes or initiates the differentiation of
naive T cells towards a TReg-cell phenotype in
a FOXP3-dependent manner81.
using naive mice, we found that gluco-
corticoid treatment in vivo increased
the amount of GITR expressed by CD4+
T cells and of GITRL expressed by pDCs30.
Glucocorticoid treatment also conferred
immunoregulatory properties on pDCs that
were dependent on GITR expression by the
host and required functional IDO in vivo.
Figure 2 |?Regulatory?T‑cell?generation?via?reverse?and?non‑canonical?signalling?to?pDCs.? ?
In?plasmacytoid dendritic cells (pDCs), several different signals emanating from different receptors
may tip the balance of canonical and non-canonical pathways of nuclear factor-κB (NF-κB) activation
in favour of inhibitor of NF-κB (IκB) kinase-α (IKKα)-dependent signalling, leading to p52–ReL-B-
driven transcription of Indo and Irf3, which encode IDO (indoleamine 2,3-dioxygenase) and IRF3
(interferon-regulatory factor 3), respectively. The resulting production of IDO, boosted by the auto-
crine type I interferons (that is, IFNα and IFNβ), generates a regulatory environment through the
combined effects of the integrated stress response and immunoregulatory tryptophan catabolites.
Reverse and CD40 signalling via co-receptor systems — that is, the pairs of cytotoxic T-lymphocyte
antigen 4 (CTLA4) and CD80; glucocorticoid-induced tumour-necrosis factor receptor (GITR) and
GITR ligand (GITRL); and CD40 ligand (CD40L) and CD40 — may sustain IKKα-dependent induction
of IDO expression. signalling through specific Toll-like receptors (TLRs) is expected to reinforce or
mimic these events, either by directly affecting the balance of canonical and non-canonical pathways
of NF-κB activation (as in the case of TLR7 and TLR9, which signal through the intracellular adaptor
protein myeloid differentiation primary-response gene 88 (MyD88) in association with IRF7) or by
activating IRF3 (as is the case for TLR3 and TLR4, which act through TRIF (Toll/interleukin-1 receptor
(TIR)-domain-containing adaptor protein inducing IFNβ). Type I IFNs may, in turn, activate
non-canonical NF-κB signalling.
820 | OCTOBeR 2007 | vOLume 7
© 2007 Nature Publishing Group
Nature Reviews | Immunology
Regulatory T cell
Therapeutic regulatory T-cell generation
Regulatory T cell
so it seems likely that through synergistic
effects on T-cell expression of GITR and
pDC expression of GITRL, glucocorticoid
acts on the GITR–GITRL co-receptor system
to induce IDO expression via non-canonical
NF-κB signalling. In addition, in the model
of allergic asthma, we found that glucocor-
ticoids inhibited TH2-cell responses and
allergy, and induced FOXP3 expression in
CD4+ T cells by mechanisms dependent
on tryptophan catabolism30 (FIG. 3c). This
supports the view that glucocorticoids, pDC
expression of GITRL and TReg-cell activ-
ity are linked by a positive feedback loop,
whereby CTLA4+GITR+ TReg cells would
expand their own population by inducing the
production of IDO by pDCs. more gener-
ally, these data suggest that glucocorticoids
function by taking advantage of reverse
signalling through GITRL to activate the
non-canonical NF-κB pathway, oppose the
canonical NF-κB pathway and induce IDO
expression. This mechanism would explain
the apparently paradoxical observation in
humans of an enhanced IDO-dependent
antimicrobial effect by glucocorticoids82,
which are otherwise immunosuppressive.
Perhaps more importantly, these data might
explain the unexpected finding of increased
IL-10-dependent TReg-cell activity, induced
by IDO, in asthmatic patients treated with
Although our appreciation of both the
complexity and potential for therapeutic
intervention of the IDO mechanism has
expanded enormously in recent years, key
unanswered questions still remain. These
relate to the nature of the adverse effect of
soluble CTLA4 on TReg-cell survival after
short-term administration of the recom-
binant protein83; the mechanisms through
which DAP12-associated receptors48,84 (such
as specific isoforms of the CD200 receptor)
and the combined effects of IL-6 and sOCs3
(suppressor of cytokine signalling 3)4,85
restrain IDO activity; and how the integrated
stress response38,42,43 and various — natural
or synthetic — tryptophan catabolites24,64
contribute to IDO-dependent regulatory
responses in vivo. Despite these limitations,
it is possible to introduce a conceptually
new model that effectively incorporates
non-canonical NF-κB signalling in tolerance
mechanisms for IDO and glucocorticoids in
a TReg-cell-dominated scenario. The resulting
Figure 3 |?non‑canonical?nF‑κB‑mediated?induction?of?iDo?expres‑
airways.?a | Harmless aeroantigens are prevented from initiating air-
way inflammation by the integrity and antimicrobial defence of the
epithelium, in an environment in which Toll-like receptor 9 (TLR9)-
driven expression of IDO by lung plasmacytoid dendritic cells (pDCs)4,85
and other cell types92,93 will inhibit the expansion and activation of
T helper 2 (TH2) cells30. Although some degree of activation of the
canonical nuclear factor-κB (NF-κB) signalling pathway (epitomized
here by inhibitor of NF-κB kinase-β (IKKβ)) probably contributes
towards maintaining the integrity of the epithelial-cell barrier, non-
canonical NF-κB signalling (that is, activated through IKKα) could
contribute IDO-dependent regulatory effects to achieve an overall
local immune homeostasis. b | Under pathological conditions, allergic
inflammation may develop, resulting in a breach in the epithelial-cell
barrier and the production of pro-inflammatory cytokines such as
interleukin-6 (IL-6; not shown). Canonical NF-κB signalling favours IL-6-
dependent suppression of IDO94 and TH2-cell expansion95. However,
regulatory T (TReg) cells — expressing surface cytotoxic T-lymphocyte
antigen 4 (CTLA4) and glucocorticoid-induced tumour-necrosis factor
receptor (GITR) — could accumulate and further expand their own
population through reverse signalling in pDCs, after engagement of
CD80 and GITR ligand (GITRL), respectively. c?| Used therapeutically,
glucocorticoids and TLR9 ligands or modulators could greatly help to
restore local homeostasis, by directly inhibiting the canonical NF-κB
pathway (as is the case for glucocorticoids) or by promoting GITRL- and
TLR9-dependent activation of the non-canonical NF-κB pathway and
TReg-cell generation. PAMP, pathogen-associated molecular pattern.
NATuRe RevIeWs | immunology?
vOLume 7 | OCTOBeR 2007 | 821
© 2007 Nature Publishing Group
paradigm is that reverse and non-canonical
signalling is an essential component of T-cell
regulatory function, whether in mucosal
homeostasis of the gut and airways, or in the
action of glucocorticoids as physiological
mediators or therapeutic agents. Ligands79
and modulators61,84,86 of TLR9 signalling, sol-
uble forms of co-stimulation antagonists15,69,
and inhibitors of canonical NF-κB signal-
ling80,81 may follow tolerogenic pathways
that converge on TReg-cell generation. All
of this could ultimately offer considerable
promise in facilitating our understanding of
the general mechanisms of tolerance.
Paolo Puccetti and Ursula Grohmann are
at the Department of Experimental Medicine,
Section of Pharmacology, University of Perugia,
Perugia 06126, Italy.
Correspondence to P.P.
Published online 3 September 2007
Eissner, G., Kolch, W. & Scheurich, P. Ligands working
as receptors: reverse signaling by members of the
TNF superfamily enhance the plasticity of the immune
system. Cytokine Growth Factor Rev. 15, 353–366
Grohmann, U. et al. CTLA-4–Ig regulates tryptophan
catabolism in vivo. Nature Immunol. 3, 1097–1101
Mellor, A. L. et al. Cutting edge: induced indoleamine
2,3 dioxygenase expression in dendritic cell subsets
suppresses T cell clonal expansion. J. Immunol.
171, 1652–1655 (2003).
Orabona, C. et al. CD28 induces immunostimulatory
signals in dendritic cells via CD80 and CD86.
Nature Immunol. 5, 1134–1142 (2004).
Munn, D. H., Sharma, M. D. & Mellor, A. L.
Ligation of B7-1/B7-2 by human CD4+ T cells triggers
indoleamine 2,3-dioxygenase activity in dendritic cells.
J. Immunol. 172, 4100–4110 (2004).
Grohmann, U., Fallarino, F. & Puccetti, P. Tolerance,
DCs and tryptophan: much ado about IDO. Trends
Immunol. 24, 242–248 (2003).
Taylor, M. W. & Feng, G. S. Relationship between
interferon-γ, indoleamine 2,3-dioxygenase, and
tryptophan catabolism. FASEB J. 5, 2516–2522
Du, M. X., Sotero-Esteva, W. D. & Taylor, M. W.
Analysis of transcription factors regulating induction
of indoleamine 2,3-dioxygenase by IFN-γ. J. Interferon
Cytokine Res. 20, 133–142 (2000).
Munn, D. H. et al. Potential regulatory function
of human dendritic cells expressing indoleamine
2,3-dioxygenase. Science 297, 1867–1870 (2002).
10. Romani, L. & Puccetti, P. Protective tolerance to
fungi: the role of IL-10 and tryptophan catabolism.
Trends Microbiol. 14, 183–189 (2006).
11. Gilliet, M. & Liu, Y. J. Generation of human CD8
T regulatory cells by CD40 ligand-activated
plasmacytoid dendritic cells. J. Exp. Med.
195, 695–704 (2002).
12. Moseman, E. A. et al. Human plasmacytoid dendritic
cells activated by CpG oligodeoxynucleotides induce
the generation of CD4+CD25+ regulatory T cells.
J. Immunol. 173, 4433–4442 (2004).
13. Bluestone, J. A. & Tang, Q. How do CD4+CD25+
regulatory T cells control autoimmunity? Curr. Opin.
Immunol. 17, 638–642 (2005).
14. Ito, T. et al. Plasmacytoid dendritic cells prime
IL-10-producing T regulatory cells by inducible
costimulator ligand. J. Exp. Med. 204, 105–115
15. Finger, E. B. & Bluestone, J. A. When ligand becomes
receptor—tolerance via B7 signaling on DCs.
Nature Immunol. 3, 1056–1057 (2002).
16. Fallarino, F. et al. Modulation of tryptophan
catabolism by regulatory T cells. Nature Immunol.
4, 1206–1212 (2003).
17. Mellor, A. L. & Munn, D. H. IDO expression by
dendritic cells: tolerance and tryptophan catabolism.
Nature Rev. Immunol. 4, 762–774 (2004).
18. Munn, D. H. et al. Prevention of allogeneic fetal
rejection by tryptophan catabolism. Science
281, 1191–1193 (1998).
19. Aluvihare, V. R., Kallikourdis, M. & Betz, A. G.
Regulatory T cells mediate maternal tolerance
to the fetus. Nature Immunol. 5, 266–271 (2004).
20. Mellor, A. L. & Munn, D. H. Tryptophan catabolism
and T-cell tolerance: immunosuppression by
starvation? Immunol. Today 20, 469–473 (1999).
21. Uyttenhove, C. et al. Evidence for a tumoral
immune resistance mechanism based on tryptophan
degradation by indoleamine 2,3-dioxygenase.
Nature Med. 9, 1269–1274 (2003).
22. Seo, S. K. et al. 4-1BB-mediated immunotherapy of
rheumatoid arthritis. Nature Med. 10, 1088–1094
23. Muller, A. J., DuHadaway, J. B., Donover, P. S.,
Sutanto-Ward, E. & Prendergast, G. C. Inhibition of
indoleamine 2,3-dioxygenase, an immunoregulatory
target of the cancer suppression gene Bin1,
potentiates cancer chemotherapy. Nature Med.
11, 312–319 (2005).
24. Platten, M. et al. Treatment of autoimmune
neuroinflammation with a synthetic tryptophan
metabolite. Science 310, 850–855 (2005).
25. Boasso, A. et al. HIV inhibits CD4+ T-cell proliferation
by inducing indoleamine 2,3-dioxygenase in
plasmacytoid dendritic cells. Blood 109, 3351–3359
26. Hayden, M. S. & Ghosh, S. Signaling to NF-κB.
Genes Dev. 18, 2195–2224 (2004).
27. Bonizzi, G. & Karin, M. The two NF-κB activation
pathways and their role in innate and adaptive
immunity. Trends Immunol. 25, 280–288 (2004).
28. Lawrence, T., Bebien, M., Liu, G. Y., Nizet, V. &
Karin, M. IKKα limits macrophage NF-κB activation
and contributes to the resolution of inflammation.
Nature 434, 1138–1143 (2005).
29. Kinoshita, D. et al. Essential role of IκB kinase α in
thymic organogenesis required for the establishment
of self-tolerance. J. Immunol. 176, 3995–4002
30. Grohmann, U. et al. Reverse signaling through
GITR ligand enables dexamethasone to activate
IDO in allergy. Nature Med. 13, 579–586 (2007).
31. Bonizzi, G. et al. Activation of IKKα target genes
depends on recognition of specific κB binding sites by
RelB:p52 dimers. EMBO J. 23, 4202–4210 (2004).
32. Ball, H. J. et al. Characterization of an indoleamine
2,3-dioxygenase-like protein found in humans and
mice. Gene 396, 203–213 (2007).
33. Asselin-Paturel, C. et al. Mouse type I IFN-producing
cells are immature APCs with plasmacytoid
morphology. Nature Immunol. 2, 1144–1150
34. Colonna, M., Trinchieri, G. & Liu, Y. J. Plasmacytoid
dendritic cells in immunity. Nature Immunol.
5, 1219–1226 (2004).
35. Trinchieri, G. & Sher, A. Cooperation of Toll-like
receptor signals in innate immune defence.
Nature Rev. Immunol. 7, 179–190 (2007).
36. Wood, K. J. & Sawitzki, B. Interferon γ: a crucial
role in the function of induced regulatory T cells
in vivo. Trends Immunol. 27, 183–187 (2006).
37. Hayashi, T. et al. Inhibition of experimental asthma
by indoleamine 2,3-dioxygenase. J. Clin. Invest.
114, 270–279 (2004).
38. Fallarino, F. et al. The combined effects of
tryptophan starvation and tryptophan catabolites
down-regulate T cell receptor ζ-chain and induce a
regulatory phenotype in naive T cells. J. Immunol.
176, 6752–6761 (2006).
39. Yang, C. H., Murti, A. & Pfeffer, L. M. Interferon
induces NF-κB-inducing kinase/tumor necrosis factor
receptor-associated factor-dependent NF-κB
activation to promote cell survival. J. Biol. Chem.
280, 31530–31536 (2005).
40. Fallarino, F. et al. Ligand and cytokine dependence
of the immunosuppressive pathway of tryptophan
catabolism in plasmacytoid dendritic cells.
Int. Immunol. 17, 1429–1438 (2005).
41. Tang, Q. & Bluestone, J. A. Plasmacytoid DCs
and Treg cells: casual acquaintance or monogamous
relationship? Nature Immunol. 7, 551–553 (2006).
42. Munn, D. H. et al. GCN2 kinase in T cells mediates
proliferative arrest and anergy induction in
response to indoleamine 2,3-dioxygenase.
Immunity 22, 633–642 (2005).
43. Manlapat, A. K., Kahler, D. J., Chandler, P. R.,
Munn, D. H. & Mellor, A. L. Cell-autonomous
control of interferon type I expression by indoleamine
2,3-dioxygenase in regulatory CD19+ dendritic cells.
Eur. J. Immunol. 37, 1064–1071 (2007).
44. Puccetti, P. On watching the watchers: IDO and
type I/II IFN. Eur. J. Immunol. 37, 876–879 (2007).
45. Medzhitov, R. Toll-like receptors and innate immunity.
Nature Rev. Immunol. 1, 135–145 (2001).
46. Hoebe, K. et al. Genetic analysis of innate immunity.
Adv. Immunol. 91, 175–226 (2006).
47. Braun, D., Longman, R. S. & Albert, M. L.
A two-step induction of indoleamine 2,3 dioxygenase
(IDO) activity during dendritic-cell maturation.
Blood 106, 2375–2381 (2005).
48. Orabona, C. et al. Towards the identification of a
tolerogenic signature in IDO-competent dendritic
cells. Blood 107, 2846–2854 (2006).
49. Hoshino, K. et al. IκB kinase-α is critical for
interferon-α production induced by Toll-like
receptors 7 and 9. Nature 440, 949–953 (2006).
50. Mellor, A. L. et al. Cutting edge: CpG oligonucleotides
induce splenic CD19+ dendritic cells to acquire potent
indoleamine 2,3-dioxygenase-dependent T cell
regulatory functions via IFN type 1 signaling.
J. Immunol. 175, 5601–5605 (2005).
51. Wingender, G. et al. Systemic application of CpG-rich
DNA suppresses adaptive T cell immunity via
induction of IDO. Eur. J. Immunol. 36, 12–20 (2006).
52. Christensen, S. R. et al. Toll-like receptor 7 and TLR9
dictate autoantibody specificity and have opposing
inflammatory and regulatory roles in a murine model
of lupus. Immunity 25, 417–428 (2006).
53. Lomada, D., Liu, B., Coghlan, L., Hu, Y. & Richie, E. R.
Thymus medulla formation and central tolerance
are restored in IKKα–/– mice that express an IKKα
transgene in keratin 5+ thymic epithelial cells.
J. Immunol. 178, 829–837 (2007).
54. Lu, L. F., Gondek, D. C., Scott, Z. A. & Noelle, R. J.
NFκB-inducing kinase deficiency results in the
development of a subset of regulatory T cells,
which shows a hyperproliferative activity upon
glucocorticoid-induced TNF receptor family-related
gene stimulation. J. Immunol. 175, 1651–1657
55. Curti, A. et al. Modulation of tryptophan catabolism
by human leukemic cells results in the conversion
of CD25– into CD25+ T regulatory cells. Blood
109, 2871–2877 (2007).
56. Hwu, P. et al. Indoleamine 2,3-dioxygenase
production by human dendritic cells results
in the inhibition of T cell proliferation. J. Immunol.
164, 3596–3599 (2000).
57. Grohmann, U. et al. Functional plasticity of dendritic
cell subsets as mediated by CD40 versus B7
activation. J. Immunol. 171, 2581–2587 (2003).
58. Vacca, C. et al. CD40 ligation prevents onset of
tolerogenic properties in human dendritic cells treated
with CTLA-4-Ig. Microbes Infect. 7, 1040–1048
59. Tas, S. W. et al. Non-canonical NF-κB signaling
in dendritic cells is required for indoleamine 2,3-
dioxygenase (IDO) induction and immune regulation.
Blood 4 May 2007 (doi:10.1182/blood-2006-11-
60. Ochando, J. C. et al. Alloantigen-presenting
plasmacytoid dendritic cells mediate tolerance to
vascularized grafts. Nature Immunol. 7, 652–662
61. Romani, L. et al. Thymosin α1 activates dendritic cell
tryptophan catabolism and establishes a regulatory
environment for balance of inflammation and
tolerance. Blood 108, 2265–2274 (2006).
62. Fallarino, F. & Puccetti, P. Toll-like receptor 9-mediated
induction of the immunosuppressive pathway of
tryptophan catabolism. Eur. J. Immunol. 36, 8–11
63. Fallarino, F. et al. Murine plasmacytoid dendritic cells
initiate the immunosuppressive pathway of tryptophan
catabolism in response to CD200 receptor
engagement. J. Immunol. 173, 3748–3754 (2004).
64. Belladonna, M. L. et al. Kynurenine pathway enzymes
in dendritic cells initiate tolerogenesis in the absence
of functional IDO. J. Immunol. 177, 130–137 (2006).
65. Grohmann, U. et al. A defect in tryptophan catabolism
impairs tolerance in nonobese diabetic mice.
J. Exp. Med. 198, 153–160 (2003).
66. Munn, D. H. et al. Expression of indoleamine
2,3-dioxygenase by plasmacytoid dendritic cells
in tumor-draining lymph nodes. J. Clin. Invest.
114, 280–290 (2004).
822 | OCTOBeR 2007 | vOLume 7
© 2007 Nature Publishing Group
67. Munn, D. H. Indoleamine 2,3-dioxygenase, Download full-text
tumor-induced tolerance and counter-regulation.
Curr. Opin. Immunol. 18, 220–225 (2006).
68. Gurtner, G. J., Newberry, R. D., Schloemann, S. R.,
McDonald, K. G. & Stenson, W. F. Inhibition of
indoleamine 2,3-dioxygenase augments
trinitrobenzene sulfonic acid colitis in mice.
Gastroenterology 125, 1762–1773 (2003).
69. Santucci, L. et al. GITR modulates innate and adaptive
mucosal immunity during the development of
experimental colitis in mice. Gut 56, 52–60
70. Wills-Karp, M., Santeliz, J. & Karp, C. L. The germless
theory of allergic disease: revisiting the hygiene
hypothesis. Nature Rev. Immunol. 1, 69–75 (2001).
71. Bach, J. F. The effect of infections on susceptibility to
autoimmune and allergic diseases. N. Engl. J. Med.
347, 911–920 (2002).
72. Rakoff-Nahoum, S., Paglino, J., Eslami-Varzaneh, F.,
Edberg, S. & Medzhitov, R. Recognition of commensal
microflora by Toll-like receptors is required for
intestinal homeostasis. Cell 118, 229–241 (2004).
73. Neish, A. S. et al. Prokaryotic regulation of epithelial
responses by inhibition of IκB-α ubiquitination.
Science 289, 1560–1563 (2000).
74. Kelly, D. et al. Commensal anaerobic gut bacteria
attenuate inflammation by regulating nuclear-
cytoplasmic shuttling of PPAR-γ and RelA. Nature
Immunol. 5, 104–112 (2004).
75. Foligne, B. et al. A key role of dendritic cells in
probiotic functionality. PLoS ONE 2, e313–e324
76. Elias, J. A., Zhu, Z., Chupp, G. & Homer, R. J.
Airway remodeling in asthma. J. Clin. Invest.
104, 1001–1006 (1999).
77. Padrid, P. A. et al. CTLA4Ig inhibits airway
eosinophilia and hyperresponsiveness by regulating
the development of Th1/Th2 subsets in a murine
model of asthma. Am. J. Respir. Cell Mol. Biol.
18, 453–462 (1998).
78. Montagnoli, C. et al. Immunity and tolerance to
Aspergillus involve functionally distinct regulatory
T cells and tryptophan catabolism. J. Immunol.
176, 1712–1723 (2006).
79. Hayashi, T. & Raz, E. TLR9-based immunotherapy for
allergic disease. Am. J. Med. 119, 897.e1–897.e6
80. De Bosscher, K., Vanden Berghe, W. & Haegeman, G.
The interplay between the glucocorticoid receptor
and nuclear factor-κB or activator protein-1:
molecular mechanisms for gene repression.
Endocr. Rev. 24, 488–522 (2003).
81. Karagiannidis, C. et al. Glucocorticoids upregulate
FOXP3 expression and regulatory T cells in asthma.
J. Allergy Clin. Immunol. 114, 1425–1433 (2004).
82. Turck, J., Oberdorfer, C., Vogel, T., Mackenzie, C. R. &
Daubener, W. Enhancement of antimicrobial effects
by glucocorticoids. Med. Microbiol. Immunol.
194, 47–53 (2005).
83. Tang, Q. et al. Cutting edge: CD28 controls peripheral
homeostasis of CD4+CD25+ regulatory T cells.
J. Immunol. 171, 3348–3352 (2003).
84. Orabona, C. et al. Enhanced tryptophan catabolism
in the absence of the molecular adapter DAP12.
Eur. J. Immunol. 35, 3111–3118 (2005).
85. Orabona, C. et al. Cutting edge: silencing suppressor
of cytokine signaling 3 expression in dendritic cells
turns CD28-Ig from immune adjuvant to suppressant.
J. Immunol. 174, 6582–6586 (2005).
86. Blasius, A. L. & Colonna, M. Sampling and signaling
in plasmacytoid dendritic cells: the potential roles
of Siglec-H. Trends Immunol. 27, 255–260 (2006).
87. Fallarino, F. et al. Functional expression of
indoleamine 2,3-dioxygenase by murine CD8α+
dendritic cells. Int. Immunol. 14, 65–68 (2002).
88. Fallarino, F. et al. T cell apoptosis by tryptophan
catabolism. Cell. Death Differ. 9, 1069–1077 (2002).
89. Terness, P. et al. Inhibition of allogeneic T cell
proliferation by indoleamine 2,3-dioxygenase-
expressing dendritic cells: mediation of suppression by
tryptophan metabolites. J. Exp. Med. 196, 447–457
90. Frumento, G. et al. Tryptophan-derived catabolites
are responsible for inhibition of T and natural killer cell
proliferation induced by indoleamine 2,3-dioxygenase.
J. Exp. Med. 196, 459–468 (2002).
91. Sawitzki, B. et al. IFN-γ production by alloantigen-
reactive regulatory T cells is important for their
regulatory function in vivo. J. Exp. Med.
201, 1925–1935 (2005).
92. Zegarra-Moran, O. et al. Double mechanism for apical
tryptophan depletion in polarized human bronchial
epithelium. J. Immunol. 173, 542–549 (2004).
93. Beutelspacher, S. C. et al. Expression of indoleamine
2,3-dioxygenase (IDO) by endothelial cells: implications
for the control of alloresponses. Am. J. Transplant.
6, 1320–1330 (2006).
94. Grohmann, U. et al. IL-6 inhibits the tolerogenic
function of CD8α+ dendritic cells expressing indoleamine
2,3-dioxygenase. J. Immunol. 167, 708–714 (2001).
95. Zaph, C. et al. Epithelial-cell-intrinsic IKK-β
expression regulates intestinal immune homeostasis.
Nature 446, 552–556 (2007).
We thank G. Andrielli for help with the original art work.
Support for the work in our laboratory came in part from grants
from the Juvenile Diabetes Research Foundation (U.G. and P.P.)
and the Italian Association for Cancer Research (P.P.).
Competing interests statement
The authors declare no competing financial interests.
CD28 | CD80 | CD86 | CTLA4 | GITR | GITRL | IDO | INDOL1
Reflections on the clonal-selection
Melvin Cohn, N. Av Mitchison, William E. Paul, Arthur M. Silverstein,
David W. Talmage and Martin Weigert
Abstract | How do we account for the immune system’s ability to produce
antibodies in response to new antigens? It has been 50 years since F. Macfarlane
Burnet published his answer to this question: the clonal-selection theory of
antibody diversity. The idea that specificity for diverse antigens exists before these
antigens are encountered was a radical notion at the time, but one that became
widely accepted. In this article, Nature Reviews Immunology asks six key scientists
for their thoughts and opinions on the clonal-selection theory, from its first
proposal to their views of it today.
What was revolutionary about the
clonal-selection theory as a solution
to the specificity-of-antibody problem?
melvin?Cohn. The clonal-selection theory
(CsT), as it was viewed by F macfarlane
Burnet1, meant nothing more than cellular
selection (not clonal selection) even as
late as 1961 (REF. 2). In any case, it made no
contribution to the debate of the instructionist
theory versus the selectionist theory. In fact, it
was the experiments of Luria and Delbrück3,
Newcombe4 and Joshua and esther
Lederberg5 that were the bases for disproving
Burnet himself rejected Niels K. Jerne’s
selectionist theory6 based on the implausibil-
ity of a self-replicating antibody molecule,
and instead suggested that the B-cell recep-
tor (BCR) was located on cells that replicate.
I, myself, was unaware of Burnet’s paper1
until 1959 when it was referenced in Burnet’s
book7. I had independently concluded, as
had David W. Talmage8, that antibodies
had to act as receptors on cells, and by 1959
my colleagues and I were in the middle of
an experiment that was derived from the
demise of instructionism — namely, to
determine the number of antibodies that
a single cell could make. The moment one
considers that an antigen receptor is located
on cells, its genetics and the pathway of its
expression come into play. At one extreme,
instructionism required that one cell pro-
duce all antibodies. At the other extreme,
selectionism required that one cell produce
one antibody. Today, I prefer to refer to
selectionism as somatic evolution and cite
Burnet’s CsT as but one of the initiating
n.?Av?mitchison. Biology in the 1950s was
coming to terms with ‘DNA and all that’. We
saw the problem of antibody synthesis in
terms of read-out from constant DNA (that
is, the germline theory, which states that the
information that is required for the produc-
tion of all necessary antibodies is provided
by the genome), although Linus Pauling’s
NATuRe RevIeWs | immunology?
vOLume 7 | OCTOBeR 2007 | 823
© 2007 Nature Publishing Group