There was initial scepticism when studies
by Sakaguchi et al. in 1995 hinted at a
re-emergence of the suppressor-T-cell
concept1, which fell out of favour in
the 1980s as, at the time, there were no
consistent ways to study the phenotype
or function of these cells2. However, these
cells have once again taken centre stage
in immunology research. They are now
identified as being CD4 positive, with high
levels of cell-surface expression of CD25;
they are known as TReg cells (CD4+CD25+
regulatory T cells)3,4 and have been shown
to have a regulatory function in various
disease states in mice3,4. TReg cells are also
found in humans and, as in mice, they only
constitute a small proportion (<10%) of
the CD4+ T-cell pool3,4. TReg cells can also
be identified by their expression of the
forkhead winged-helix transcription factor
FOXP3 (forkhead box P3)5,6.
The expression of FOXP3 has been
proposed to be the crucial switch factor in
the induction of the TReg-cell population.
The immune dysregulation that is found in
Foxp3-knockout mice, which develop a fatal
autoimmune disease, and the association
of defective FOXP3 function with IPEX
and enteropathy, X-linked) syndrome in
humans substantiates a role for this molecule
in immune regulation5,6. Patients with IPEX
develop autoimmune diseases, such as
type 1 diabetes and thyroditis, inflammatory
bowel disease and severe allergies7. These
observations indicate the exciting possibility
of therapeutic intervention, directed towards
the TReg cells themselves or at key target
molecules, such as FOXP3, to modulate
disease activity that arises from immune
In addition to this naturally occurring
population of TReg cells, which develop in
the thymus1, other regulatory T-cell popula-
tions are generated in the periphery. These
are often termed adaptive TReg cells and
they are remarkably similar to the naturally
occurring TReg-cell populations9. It is impor-
tant to understand the biology of regulatory
T cells and how they are generated in vivo
before any attempts are made to manipulate
them therapeutically. This note of caution is
particularly relevant given the severe prob-
lems that arose from the use of super agonistic
CD28-specific antibodies in a recent
clinical trial, in which the reagent, although
tolerated well in animals, had unexpected
effects in humans10. In this Opinion article,
we address the question of how FOXP3+
TReg cells are maintained throughout life and
question to what extent studies in mice
can be extrapolated to humans.
Heterogeneity of TReg cells
Many regulatory T-cell populations have been
described (TABLE 1). In this article, we focus
on FOXP3+ TReg cells, which are the best-
defined population at present. In addition
to CD25 and FOXP3, several other markers,
such as CTLA4 (cytotoxic T-lymphocyte
antigen 4), GITR (glucocorticoid-induced
protein), HLA-DR, OX40 and LAG3
(lymphocyte-activation gene 3) have been
associated with this natural TReg-cell popula-
tion3,4. There are some important differences
between mouse and human T cells, as human
CD4+CD25– effector T cells can upregulate
FOXP3 expression transiently, following
activation11–13, whereas mouse effector T cells
do not14. So, in conditions of immune stimu-
lation, FOXP3 might not be a good marker
for human TReg cells, as it can be expressed by
both regulatory and activated effector T-cell
At present, the best way to identify both
mouse and human TReg cells, in situations
where there is no overt immune activation
in vivo, is by the co-expression of high levels
of CD25 and FOXP3 (REF. 15). In addi-
tion, low-level expression of CD127 (the
interleukin-7 receptor α-chain) can be used
as a further discriminating marker for this
population16,17. However, it is important to
acknowledge the limitations of CD25 by itself
as a marker for regulatory populations, as it is
also upregulated by activated effector T cells,
and some regulatory CD4+ T cells that are
FOXP3+CD127low express little or no CD25
(REFS 16,17). Although most FOXP3+ natu-
rally occurring TReg cells in adults express
markers of primed T cells, such as CD45RO,
a small proportion of these cells express the
marker CD45RA, and this CD45RA popula-
tion decreases with age18–20. However, both
populations have identical mechanisms of
Lifelong maintenance of TReg cells
TReg cells can develop in the thymus by the
high-affinity interaction of the T-cell recep-
tor (TCR) with self-peptides, as reviewed
extensively elsewhere21. However, whereas
The dynamic co-evolution of
memory and regulatory CD4+ T cells
in the periphery
Arne N. Akbar, Milica Vukmanovic-Stejic, Leonie S. Taams and Derek C. Macallan
Abstract | Whereas memory T cells are required to maintain immunity, regulatory
T cells have to keep the immune system in check to prevent excessive inflammation
and/or autoimmunity. Both cell types must be present during the lifetime of the
organism. However, it is not clear whether both subsets are regulated in tandem or
independently of each other, especially because thymic involution severely restricts
the production of T-cell populations during ageing. In this Opinion article, we
discuss recent evidence in both mice and humans that supports the hypothesis that
some CD4+CD25+FOXP3+ regulatory T cells can differentiate from rapidly
proliferating memory T cells in the periphery.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 7 | MARCH 2007 | 231
© 2007 Nature Publishing Group
Naive T cell
Memory T-cell pool
Memory-derived TReg cell
the output of an infant’s thymus is in the
order of 109 CD4+ T cells per day, by 20
years of age it is reduced to 15% of this value
and continues to decrease thereafter22. Old
individuals (>70 years of age) have decreased
numbers of CD45RA+ TReg cells18,20, as well
as decreased numbers of conventional
naive T cells, which supports the assump-
tion that thymic involution equally affects
the generation of nascent regulatory and
non-regulatory T cells. However, compared
to young individuals (<35 years old), old
individuals have the same or increased
numbers of CD45RO+ TReg cells with the
same function as those from young indi-
viduals23,24, therefore, other mechanisms in
addition to thymic output must contribute to
the lifelong maintenance of these cells.
There are three potential mechanisms to
explain how TReg cells are maintained in vivo.
First, there might be naturally occurring
TReg cells of thymic origin that are resistant
to death. However, we and others have
shown that most TReg cells in adults express
CD45RO and these cells are highly suscep-
tible to apoptosis25–27. Second, TReg cells in
adults might originate from thymic-derived
TReg cells that are continuously replenished
by proliferation in the periphery (FIG. 1a).
Indeed, mouse TReg cells have been shown
to proliferate extensively while maintain-
ing their regulatory function in vivo28–30,
and such proliferation also occurs for both
human and mouse TReg cells in vitro31–34. This
might explain why freshly isolated TReg cells
from both humans and mice express a highly
differentiated phenotype (CD45RBlow),
indicating that they have proliferated exten-
sively26,35–38. Third, TReg cells might also be
generated from activated effector or memory
CD4+CD25– T cells in the periphery, and
these cells are known as adaptive or induced
TReg cells39–41 (FIG. 1b). The phenotype and
function of adaptive TReg cells is remarkably
similar to naturally occurring TReg cells
The small subset of the human naturally
occurring TReg-cell population that expresses
CD45RA has robust proliferative activity
and might represent a self-regenerating
TReg-cell population18–20. However, CD45RA+
TReg cells upregulate CD45RO expression after
activation and become highly susceptible
to apoptosis18–20, and are therefore likely to
be lost rapidly after proliferation. Although
it is possible that these CD45RA+ TReg cells
may re-populate the CD45RO+ TReg-cell
pool throughout life, the observations that
CD45RA+ TReg cells decrease considerably
with age, together with their susceptibility
to death, suggests that it is unlikely18–20. The
occurrence of thymic involution indicates
that TReg cells in old humans have to be gener-
ated from either proliferative natural TReg cells
or TReg cells induced from effector memory
T cells, in both cases proliferation is central
to their lifelong maintenance. This raises a
problem because there is a limit to the extent
to which cells can proliferate continuously
in vivo that is set by the extent of telomere
erosion42,43. Therefore, the two crucial issues
Table 1 | Types of regulatory T cell: origin, phenotype and function
IL-10, TGFβ, CTLA4
Expansion of natural
and in some cases
TGFβ and/or IL-10
expansion of non-
regulatory CD4+ T cells
TH3 cellsFOXP3+/– (not well
CD8+ T cells
Cell contact, ILT3
*For simplicity, we have provided a consensus of the most widely accepted data on the characteristics of regulatory
T cells. ‡ Most of the natural regulatory CD4+CD25+ T (TReg) cells express CD45RO, which is characteristic of a memory
phenotype. However, a small proportion of circulating TReg cells express CD45RA and are sometimes referred to as
naive natural TReg cells. These cells are abundant in cord blood, but decrease with age. In terms of phenotype and
function, they are very similar to the CD45RO+ TReg cells. CTLA4; cytotoxic T-lymphocyte antigen 4; FOXP3, forkhead
box P3; GITR, glucocorticoid-induced tumour-necrosis-factor-receptor-related protein; IL, interleukin;
ILT, immunoglobulin-like transcript; TGFβ, transforming growth factor-β; TH3, T helper 3; TR1, T regulatory 1.
Figure 1 | Possible mechanisms for the origin and maintenance of human CD4+CD25+ regula-
tory T cells in vivo. Depicted are two possible models of CD4+CD25+ regulatory T (TReg)-cell mainte-
nance. a | In the first model, naturally occurring TReg cells are generated in the thymus as a separate
lineage. However, as thymic output decreases, TReg cells need to be maintained by continuous prolif-
eration in the periphery. Telomere erosion and lack of telomerase activity would limit their capacity
for self-regeneration. b | In the second model, TReg cells are continuously generated from responding,
effector memory T-cell populations in the periphery.
232 | MARCH 2007 | VOLUME 7
© 2007 Nature Publishing Group
Naive T-cell pool
Memory T-cell pool
Regulatory T-cell pool
New cells: <0.5%
∼8% per day
Division of cells within
the naive pool
Division of cells within
the memory pool
new TReg cells
that have to be addressed to assess the resid-
ual replicative capacity of TReg cells are: first,
the rate of turnover of these cells, and second,
whether they have sufficient telomeric reserve
to sustain the rate of division throughout life.
Turnover rate of human TReg cells in vivo
The development of the deuterated glucose
method for labelling proliferating cells
in vivo has enabled the turnover rate of
different subsets of human T cells to be
determined (FIG. 2; TABLE 2). However, there
are few data available on the turnover rate
of TReg cells. In one study in mice, using a
combination of BrdU (bromodeoxyuridine)
and CFSE (5,6-carboxyfluorescein diacetate
succinimidyl ester) labelling, it was shown
that a subset of TReg cells that was gener-
ated in vitro has a rapid turnover rate after
adoptive transfer in vivo44. In a second study,
in patients with acute myeloid leukaemia,
there was evidence for a relatively high
rate of TReg-cell proliferation in vivo45.
More recently, a study of TReg-cell turnover
in healthy individuals without any overt
immune activation, showed that CD45RO+
TReg cells had a rapid doubling time of 8 days,
compared with the CD4+CD25–CD45RO+
memory T cells (doubling time of 28 days)
or CD4+CD25–CD45RA+ naive T cells
(doubling time of 199 days)24. In this study,
CD45RO+ TReg cells expressed FOXP3
(REF. 24) and were CD127low (A. McQuaid,
M.V.-S. and A.N.A., unpublished observa-
tions). The rapid proliferation was matched
by the rapid death rate of CD45RO+ T cells.
Furthermore, the kinetics for prolifera-
tion and death of these regulatory T cells
in young and old individuals were very
similar24. This suggests that human TReg cells
comprise a dynamic population, which is
generated continuously from either pre-
existing CD45RO+ TReg cells28,30,44 or from
CD4+CD25– memory T cells that proliferate
rapidly then enter the TReg-cell pool during
the labelling period (FIG. 1).
Can TReg cells proliferate indefinitely?
Every time a cell divides, between 50–100
base pairs of telomeric DNA are lost from
the ends of chromosomes42,43. The induction
of expression of the enzyme telomerase can
retard telomere erosion and it is possible
that some T-cell populations might be main-
tained indefinitely because this enzyme is
constitutively active42,43. The telomere length
in human T cells is 4–12 kilobases (kb)
and growth arrest or replicative senescence
occurs when telomere length falls below
4 kb42,43. The telomere length of human
CD45RO+ TReg cells was found to be short
compared with either CD4+CD25–CD45RO+
memory T cells or the total CD4+ T-cell pool
in the same subjects24,46. In addition, the
TReg cells from both young and old subjects
were unable to upregulate telomerase after
activation24. It is therefore unlikely that the
TReg-cell pool in old humans is derived from
self-regenerating populations of naturally
occurring TReg cells that continuously pro-
liferate throughout life24. Furthermore, the
propensity of replicative senescence together
with their susceptibility to apoptosis26,27
Table 2 | Relative in vivo proliferation rates of CD4+ T-cell subtypes
Persistence of chromosomal
Cell type studied
Main phenotypic definition
CD45RO+ versus CD45RA+
Memory T cells > naive T cells
Mouse spleen and
lymph-node T cells
CD44+ and CD45RB+
Memory T cells > naive T cells96
Deuterium-labelled glucoseCD45RO+ or CD45RA+ and
CCR7+ versus CCR7– (CD45RO+
and CD45RA+ populations)
CD25, CD62L, CD44
Memory T cells > naive T cells97,98
Deuterium-labelled water Human PBMCsEffector memory T cells > naive
Effector memory T cells > central
memory T cells > naive T cells
‘Activated’ TReg cells (defined as
CD62LlowCD44hi) > ‘resting’ TReg cells
CD25hi ≈ CD25mid ≈ CD25low
Deuterium-labelled glucoseHuman PBMCs
Survival after adoptive transfer
using CFSE and BrdU labelling
Mouse spleen and
lymph-node T cells
CD25hi, CD25mid and CD25low
(CD4+ T-cell populations)
CD25hi versus CD25– (CD45RO+
population) versus CD45RA+
Deuterium-labelled glucoseHuman PBMCsTReg cells > memory T cells > naive
BrdU, bromodeoxyuridine; CFSE, carboxy-fluorescein diacetate succinimidyl ester; PBMCs, peripheral blood mononuclear cells.
Figure 2 | A model of regulatory T-cell homeostasis derived from deuterium-labelled glucose
measurements of in vivo proliferation. The size of the circle schematically represents the size of
the T-cell population; the size of the dark wedge represents the proportion of new cells per unit of time
within that population. Most circulating CD4+ T cells are slowly dividing or non-dividing CD45RA+
naive T cells. Cells within the smaller memory T-cell pool divide more often, whereas TReg cells
(CD4+CD25+ regulatory T cells) comprise the smallest population (about one-thirtieth of circulating
CD4+ T cells) with the fastest rate of turnover24. Cell death is not shown for the sake of clarity but must
balance the appearance of new cells within each pool. This data is compatible with the available
reports in mice44 and humans45 showing that TReg cells have a high rate of proliferation.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 7 | MARCH 2007 | 233
© 2007 Nature Publishing Group
would explain why these cells do not accu-
mulate in vivo despite their rapid prolifera-
tion. This raises the question of whether
TReg cells are derived from proliferating
memory CD4+CD25–CD45RO+ T cells. It is
important to emphasize that as mice have a
considerably shorter lifespan than humans,
but have up to tenfold longer telomeres25,
the thymus-derived TReg-cell population
might not be restricted to the same extent by
continuous proliferation. Self-regenerating
naturally occurring TReg cells might therefore
constitute a greater proportion of the TReg-cell
pool in mice compared with humans.
Are memory T cells and TReg cells related?
Studies in mice have shown that TReg cells
can be generated from CD4+CD25– effec-
tor T cells under certain conditions40,41,47,48
(TABLE 3). To address the relationship
between human CD45RO+ TReg cells and
CD4+CD25–CD45RO+ memory T cells, sev-
eral researchers have compared the TCR Vβ
repertoires in the two subsets to determine
to what extent both populations express
receptors that belong to the same TCR
family. These studies showed that these two
populations were related46,49. These studies
were extended to investigate the clonal rela-
tionship between TReg-cell and memory T-cell
populations within an individual TCR Vβ
family from the same individual24. The key
observation was that there was an average
of 80% clonal homology between CD45RO+
TReg-cell and CD4+CD25–CD45RO+ memory
T-cell populations, and this related identity
has been confirmed by other research-
ers using DNA spectratyping49 or CDR3
(complementarity-determining region 3)
sequence analysis50. Furthermore, antigen-
related expansions of CD4+CD25– effector
T cells within certain Vβ families were
associated with parallel increases of FOXP3+
TReg cells with the same TCR Vβ usage24. As it
is unlikely that these human regulatory T cells
are self-regenerating, at least a proportion of
them must be derived from rapidly dividing
memory CD4+ T cells. However, an alterna-
tive possibility that must be considered is
that the shared clonality in CD4+CD25– and
CD4+CD25+ T-cell subpopulations might
actually represent shared sequences between
CD25– TReg-cell-like and CD25+ TReg-cell
populations. This might be resolved in the
future by the use of MHC class II tetramers
to show that the same antigen-specific CD4+
T cells are equally represented in regulatory
and non-regulatory pools.
From memory T cell to TReg cell. It has been
shown in both mice and humans that trans-
forming growth factor-β (TGFβ)48,51,52 and
interferon-γ (IFNγ)53 can induce CD4+CD25–
FOXP3– effector T cells to become FOXP3+
TReg cells (TABLE 3). As these cytokines signal
through different pathways, it will be impor-
tant to determine their point of convergence.
Other regulatory populations, such as those
that secrete interleukin-10 (IL-10; T regula-
tory 1 (TR1) cells)54 or TGFβ (T helper 3
(TH3) cells)55,56, can also be generated from
naive and/or memory CD4+CD25– T cells
under specific conditions, but their relation-
ship to naturally occurring or adaptive
FOXP3+ TReg cells is unclear. One important
caveat that applies to humans is that not all
CD4+CD25–FOXP3– T cells that are induced
to express FOXP3 also acquire regulatory
function12. Although this has led to the sug-
gestion that FOXP3 is not an optimal marker
for human TReg cells, we think this is too
narrow a conclusion. The fact that FOXP3
is upregulated by CD4+CD25– T cells after
activation is consistent with our hypothesis
that activated T cells can differentiate into
TReg cells under appropriate conditions. The
crucial question to be addressed is what
signals induce the stable upregulation of
FOXP3 and committed regulatory function
in CD4+CD25– T cells.
Apart from cytokine induction, it has
been shown that the induction of anergy in
activated CD4+ T cells by providing a TCR
signal in the absence of co-stimulation,
or by specific peptide being presented by
activated T cells to each other in humans or
Table 3 | Studies showing extra-thymic generation of CD4+CD25+ regulatory T cells from CD4+CD25– non-regulatory T cells
T-cell sourceTReg-cell-inducing stimulus Induction of
Mouse in vitro studies
CD4+CD25– T cells
CD4+CD25– T cells Yes 53
Mouse in vivo studies
CD4+CD25– T cells
CD4+CD25– T cells (OVA-specific
TCR-transgenic Rag–/– mice)
CD4+CD25– T cells (HA-specific
TCR-transgenic Rag–/– mice)
CD4+ (OVA-specific TCR-transgenic
Human in vitro studies
CD4+CD25– T cells
CD4+CD25– T cells
Homeostatic proliferation and/or conversion
Intravenous or oral administration of OVA
Prolonged exposure to a low dose of antigenYes 39,41
Prolonged exposure to peripheral antigenYes 40
CD3 and/or CD28 or antigen-specific stimulation
Thrombospondin and/or CD47 interaction
IL-4 and/or IL-13
T-cell-to-T-cell antigen presentation; immobilized
CD4+CD25–CD45RA+ T cells Yes48,52,103
CD4+CD25–CD45RA+ T cells
CD4+CD25–CD45RA+ T cells
CD4+ effector T-cell clones
A.N.A. et al., unpublished
FOXP3, forkhead box P3; HA, haemagglutinin; IFNγ, interferon-γ; ND, not determined; OVA, ovalbumin; Rag, recombination-activating gene; TCR, T-cell receptor; TGFβ,
transforming growth factor-β; TReg cell, CD4+CD25+ regulatory T cell.
234 | MARCH 2007 | VOLUME 7
© 2007 Nature Publishing Group
CD4+ T cells
Memory T cells
TReg cellsTReg cells
Antigen removedForeign antigen
rats, induces the development of regulatory
T cells in vitro57–59. However, it is not clear
whether these populations are identical to
the FOXP3+ TReg cells that are found in vivo25.
It has also been shown that the interaction
between thrombospondin and CD47 on the
surface of human CD4+CD25–FOXP3– naive
or memory T cells induced the development
of a FOXP3+ TReg-cell population60. These
data collectively highlight that the CD4+
T-cell repertoire in both mice and humans is
considerably more flexible than previously
appreciated and that exogenous mediators
and specific cell-surface triggers might
be able to regulate the differentiation of
responsive T-cell populations into regulatory
T-cell populations. The context and location
of antigenic encounter, for example in the
presence of IFNγ, IL-10 or TGFβ, could
influence which type of regulatory T cell
(FOXP3+ TReg cell, TR1 cell or TH3 cell) is
induced. In addition, as FOXP3+ TReg cells
have a highly differentiated phenotype26,35, it
is likely that the cues that induce this stable
transition affect only memory CD4+ T cells
that have reached a particular stage of matu-
rity. This is a characteristic that has been
observed for other functional lineage-related
transcription factors, such as T-bet, which is
involved in inducing a TH1-type functional
profile in CD4+ T cells61.
Co-evolution of memory T cells and TReg cells.
Based on the earlier discussion, we postulate
different pathways for the maintenance
of TReg cells in mice and humans. In mice,
FOXP3+ TReg cells could originate either
from naturally occurring, thymus derived
TReg cells that are self-regenerating or from
activated and proliferating CD4+CD25–
FOXP3– memory T cells (FIG. 1). However,
in humans, especially older individuals, we
postulate that FOXP3+ TReg cells are generated
continuously by activated and proliferating
CD4+CD25–FOXP3– T cells. The rapid gen-
eration of these cells is counterbalanced by
rapid loss, thereby maintaining homeostasis,
which allows a relatively constant number of
TReg cells to be maintained during ageing23,24.
This hypothesis is supported by the observa-
tion that TReg cells develop in parallel with
activated CD4+ T cells within the memory
T-cell pool24,50. The inescapable conclusion
from this is that most human FOXP3+
TReg cells are transient cells that are generated,
lost and replaced continuously in vivo.
This proposes a simple model that links
the level of regulatory T-cell generation
to the extent of immune stimulation that
occurs (FIG. 3). A robust antigenic challenge
that results in the extensive proliferation of
specific CD4+ T cells will give rise to many
short-lived TReg cells of the same specificity,
whereas a smaller immune response would
induce less TReg-cell generation. When the
immune response subsides, owing to removal
of the antigen, the regulatory cells disappear
rapidly as they are susceptible to apoptosis26.
This prevents the accumulation of regulatory
cells that could paralyse specific immunity
and therefore allows subsequent responsive-
ness to a previously encountered antigen.
Conversely, in the case of self-antigens and
dietary antigens, there will be prolonged
antigenic encounter, leading to chronic
stimulation of antigen-specific CD4+ T cells.
This in turn will lead to the continuous gen-
eration of TReg cells. This would prevent the
development of autoimmunity, and might be
a crucial mechanism at the interface between
antigen-induced immunity and tolerance.
Implications for therapy
According to this model of TReg-cell genera-
tion, strategies to deplete TReg cells in vivo
to boost antitumour immune responses62,63
might only have limited benefit, as these
cells may be continuously replenished
from the memory pool64. In fact, in a
recent clinical trial, using a recombinant
IL-2–diphtheria-toxin conjugate to deplete
TReg cells and thereby boost the antitumour
response, TReg cells were restored to 75% of
the original number within two months63. In
addition, infusion of these cells into humans
to reduce inflammation or autoimmunity
will only confer short-term benefit as these
cells are susceptible to apoptosis and might
be cleared rapidly in vivo. In this context,
factors such as IL-2 that have been shown to
be essential for the generation of TReg cells,
might actually function by counteracting
TReg cell susceptibility to cell death65–67.
However, the proximity of these cells to rep-
licative senescence because of short telom-
eres might be a useful characteristic from the
point of view of safety, as their capacity for
continuous proliferation would be limited in
situations of adoptive transfer.
If TReg cells differentiate from memory
or activated CD4+ T cells, then changes
in FOXP3+ TReg cells that have been found in
human diseases, such as multiple sclerosis68,
Figure 3 | A model that links the duration and intensity of immune stimulation to regulatory
T-cell generation. a | Antigenic challenge resulting in extensive proliferation of specific CD4+
T cells will give rise to many short-lived TReg cells (CD4+CD25+ regulatory T cells) of the same
specificity. The number of TReg cells generated will be proportional to the size of the immune
response. Once the antigen is removed, the immune response subsides and the regulatory cells
disappear rapidly, as they are susceptible to apoptosis. This prevents the over-accumulation of
regulatory T cells that could paralyse subsequent responses to the same antigen. b | Prolonged
antigenic encounter, as in the case of self-antigens and dietary antigens, will lead to chronic
stimulation of antigen-specific CD4+ T cells. This in turn will lead to the continuous generation of
TReg cells that might prevent the development of autoimmunity.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 7 | MARCH 2007 | 235
© 2007 Nature Publishing Group
rheumatoid arthritis69–71, type 1 diabetes72 and
atopic eczema73, could be linked to defective
differentiation from effector or memory
CD4+ T cells rather than to their dysfunction
per se. In addition, if TReg cells are generated
continuously throughout life, then persistent
infection with viruses, such as cytomegalo-
virus (CMV), that induce the large-scale
differentiation of human memory T cells74–76
would be expected to also induce a dispropor-
tionate representation of CMV-specific regu-
latory T cells in the total TReg-cell pool, which
indeed has been shown24. Such an imbalance
might contribute to the CMV-related
immune defect found in old individuals77.
Finally, the concept that TReg cells might
develop from CD4+CD25– memory or acti-
vated CD4+ T cells by antigenic stimulation
might explain some historical data. Low-zone
tolerance, which is achieved by the repeated
administration of low doses of antigen to
an animal, is one of the fundamental tenets
of immunology but the mechanism for this
remains unclear78. More recent data in mice,
in which continuous infusion of antigen
induces TReg cells, provide a modern counter-
part of these studies, suggesting that low-zone
tolerance might occur as a result of regulatory
T-cell induction from the responding popula-
tion39–41,79. The possibility of clinically appli-
cable induction of regulatory T cells that can
restore or maintain self-tolerance was clearly
shown by studies of autoimmune diabetes in
mice. In this study, treatment of non-obese
diabetic mice with CD3-specific antibodies
induced TGFβ-dependent regulatory T cells
from CD25– T-cell populations80. Treatment
with CD3-specific antibody was also effective
in dampening disease progression in patients,
presumably by a similar mechanism81. Peptide
immunotherapy in humans has been used
successfully to treat allergies by inducing
responsive cells to secrete IL-10, and this is
analogous to the studies of low-zone toler-
ance82. The IL-10-producing CD4+ TR1-cell
population that is induced by peptide
immunotherapy overlaps with the TReg-cell
population83–85. These studies indicate that the
way forward in the use of human TReg cells for
therapy might be to identify and manipulate
mechanisms that regulate their induction
from responsive populations in vivo.
The regulatory T-cell field has diversified
considerably and increased in complexity
in the last decade, as different researchers
use different systems and focus exclusively
on particular regulatory T-cell populations.
Searching for points of convergence between
regulatory T-cell types and experimental
systems used is essential to allow simplifica-
tion of the field in the future. The model
proposed here provides a simple explanation
of how the activation of memory T cells is
closely linked to the induction of regula-
tory activity. The current limitation to the
model proposed here is that the relationship
between different types of human regula-
tory T cell is unclear and it remains to be
determined whether they arise from a com-
mon precursor. Second, although the shared
clonotype between TReg cells and memory
T cells provides compelling evidence that
both populations are related, naturally
occurring TReg cells are considered to be
specific for self-antigens, whereas adaptive
TReg cells have the same non-self-antigen
specificity as the effector memory T-cell
pool. The possible implication for auto-
immunity of the natural attrition of the natu-
rally occurring TReg-cell population owing to
thymic involution, in parallel with the accu-
mulation of the adaptive TReg-cell population
during ageing, requires further study. The
current challenge is to identify mechanisms
that induce the stable development of regula-
tory CD4+ T cells from activated populations
to pinpoint therapeutic targets for the in vivo
manipulation of these cells in humans.
Arne N. Akbar and Milica Vukmanovic-Stejic are at the
Department of Immunology and Molecular Pathology,
Division of Infection and Immunity, Windeyer Institute
of Medical Sciences, University College London,
London W1T 4JF, UK.
Leonie S. Taams is at the Department of
Immunobiology, King’s College London, Guy’s Hospital,
London SE1 9RT, UK.
Derek C. Macallan is at the Centre for Infection,
Division of Cellular and Molecular Medicine,
St George’s University of London, London
SW17 0RE, UK.
Correspondence to A.N.A.
Sakaguchi, S. et al. Immunologic self-tolerance
maintained by activated T cells expressing IL-2 receptor
α-chains (CD25). Breakdown of a single mechanism of
self-tolerance causes various autoimmune diseases.
J. Immunol. 155, 1151–1164 (1995).
Tada, T. & Takemori, T. Selective roles of thymus
derived lymphocytes in the antibody response. I.
Differential suppressive effect of carrier primed T cells
on hapten specific IgM and IgG antibody responses.
J. Exp. Med. 140, 239–252 (1974).
Sakaguchi, S. Regulatory T cells: key controllers of
immunologic self-tolerance. Cell 101, 455–458
Shevach, E. M. CD4+CD25+ suppressor T cells:
more questions than answers. Nature Rev. Immunol.
2, 389–400 (2002).
Hori, S., Nomura, T. & Sakaguchi, S. Control of
regulatory T cell development by the transcription
factor Foxp3. Science 299, 1057–1061 (2003).
Fontenot, J. D., Gavin, M. A. & Rudensky, A. Y.
Foxp3 programs the development and function
of CD4+CD25+ regulatory T cells. Nature Immunol.
4, 330–336 (2003).
Wildin, R. S., Smyk-Pearson, S. & Filipovich, A. H.
Clinical and molecular features of the
enteropathy, X linked (IPEX) syndrome. J. Med. Genet.
39, 537–545 (2002).
Bluestone, J. A. Regulatory T-cell therapy: is it ready
for the clinic? Nature Rev. Immunol. 5, 343–349
O’Neill, E. J. et al. Natural and induced regulatory
T cells. Ann. NY Acad. Sci. 1029, 180–192 (2004).
10. Sheridan, C. TeGenero fiasco prompts regulatory
rethink. Nature Biotech. 24, 475–476 (2006).
11. Allan, S. E. et al. The role of 2 FOXP3 isoforms in the
generation of human CD4+ Tregs. J. Clin. Invest. 115,
12. Gavin, M. A. et al. Single-cell analysis of normal and
FOXP3-mutant human T cells: FOXP3 expression
without regulatory T cell development. Proc. Natl
Acad. Sci. USA 103, 6659–6664 (2006).
13. Walker, M. et al. Induction of FoxP3 and acquisition of
T regulatory activity by stimulated human CD4+CD25–
T cells. J. Clin. Invest. 112, 1437–1443 (2003).
14. Yagi, H. et al. Crucial role of FOXP3 in the
development and function of human CD25+CD4+
regulatory T cells. Int. Immunol. 16, 1643–1656
15. Roncador, G. et al. Analysis of FOXP3 protein
expression in human CD4+CD25+ regulatory T cells at
the single-cell level. Eur. J. Immunol. 35, 1681–1691
16. Liu, W. et al. CD127 expression inversely correlates
with FoxP3 and suppressive function of human CD4+
T reg cells. J. Exp. Med. 203, 1701–1711 (2006).
17. Seddiki, N. et al. Expression of interleukin (IL)-2 and
IL-7 receptors discriminates between human
regulatory and activated T cells. J. Exp. Med. 203,
18. Seddiki, N. et al. Persistence of naive CD45RA+
regulatory T cells in adult life. Blood 107, 2830–2838
19. Fritzsching, B. et al. Naive regulatory T cells: a novel
subpopulation defined by resistance toward CD95L-
mediated cell death. Blood 108, 3371–3378 (2006).
20. Valmori, D. et al. A peripheral circulating
compartment of natural naive CD4 Tregs. J. Clin.
Invest. 115, 1953–1962 (2005).
21. Picca, C. C. & Caton, A. J. The role of self-peptides in
the development of CD4+ CD25+ regulatory T cells.
Curr. Opin. Immunol. 17, 131–136 (2005).
22. Mackall, C. L. & Gress, R. E. Thymic aging and T-cell
regeneration. Immunol. Rev. 160, 91–102 (1997).
23. Gregg, R. et al. The number of human peripheral
blood CD4+ CD25high regulatory T cells increases with
age. Clin. Exp. Immunol. 140, 540–546 (2005).
24. Vukmanovic-Stejic, M. et al. Human CD4+ CD25hi
Foxp3+ regulatory T cells are derived by rapid
turnover of memory populations in vivo. J. Clin. Invest.
116, 2423–2433 (2006).
25. Akbar, A. N., Taams, L. S., Salmon, M. & Vukmanovic-
Stejic, M. The peripheral generation of CD4+ CD25+
regulatory T cells. Immunology 109, 319–325
26. Taams, L. S. et al. Human anergic/suppressive
CD4+CD25+T cells: a highly differentiated and
apoptosis-prone population. Eur. J. Immunol. 31,
27. Fritzsching, B. et al. Cutting edge: in contrast to
effector T cells, CD4+CD25+FoxP3+ regulatory T cells
are highly susceptible to CD95 ligand- but not to TCR-
mediated cell death. J. Immunol. 175, 32–36 (2005).
28. Walker, L. S. et al. Antigen-dependent proliferation of
CD4+ CD25+ regulatory T cells in vivo. J. Exp. Med.
198, 249–258 (2003).
29. Cozzo, C., Larkin, J., III & Caton, A. J. Cutting edge:
self-peptides drive the peripheral expansion of
CD4+CD25+ regulatory T cells. J. Immunol. 171,
30. Klein, L., Khazaie, K. & von Boehmer, H. In vivo
dynamics of antigen-specific regulatory T cells not
predicted from behavior in vitro. Proc. Natl Acad.
Sci. USA 100, 8886–8891 (2003).
31. Earle, K. E. et al. In vitro expanded human
CD4+CD25+ regulatory T cells suppress effector T cell
proliferation. Clin. Immunol. 115, 3–9 (2005).
32. Hoffmann, P. et al. Large-scale in vitro expansion of
polyclonal human CD4+CD25high regulatory T cells.
Blood 104, 895–903 (2004).
33. Godfrey, W. R. et al. In vitro-expanded human
CD4+CD25+ T-regulatory cells can markedly inhibit
allogeneic dendritic cell-stimulated MLR cultures.
Blood 104, 453–461 (2004).
34. Yamazaki, S. et al. Direct expansion of functional
CD25+ CD4+ regulatory T cells by antigen-processing
dendritic cells. J. Exp. Med. 198, 235–247 (2003).
236 | MARCH 2007 | VOLUME 7
© 2007 Nature Publishing Group
35. Itoh, M. et al. Thymus and autoimmunity: production Download full-text
of CD25+CD4+ naturally anergic and suppressive
T cells as a key function of the thymus in maintaining
immunologic self-tolerance. J. Immunol. 162,
36. Baecher-Allan, C., Brown, J. A., Freeman, G. J. &
Hafler, D. A. CD4+CD25high regulatory cells in human
peripheral blood. J. Immunol. 167, 1245–1253
37. Jonuleit, H. et al. Identification and functional
characterization of human CD4+CD25+T cells with
regulatory properties isolated from peripheral blood.
J. Exp. Med. 193, 1285–1294 (2001).
38. Dieckmann, D. et al. Ex vivo isolation and
characterization of CD4+CD25+ T cells with regulatory
properties from human blood. J. Exp. Med. 193,
39. Apostolou, I. & von Boehmer, H. In vivo instruction of
suppressor commitment in naive T cells. J. Exp. Med.
199, 1401–1408 (2004).
40. Knoechel, B. et al. Sequential development of
interleukin 2-dependent effector and regulatory
T cells in response to endogenous systemic antigen.
J. Exp. Med. 202, 1375–1386 (2005).
41. Kretschmer, K. et al. Inducing and expanding
regulatory T cell populations by foreign antigen.
Nature Immunol. 6, 1219–1227 (2005).
42. Akbar, A. N., Beverley, P. C. & Salmon, M. Will
telomere erosion lead to a loss of T-cell memory?
Nature Rev. Immunol. 4, 737–743 (2004).
43. Hodes, R. J., Hathcock, K. S. & Weng, N. P. Telomeres
in T and B cells. Nature Rev. Immunol. 2, 699–706
44. Fisson, S. et al. Continuous activation of autoreactive
CD4+ CD25+ regulatory T cells in the steady state.
J. Exp. Med. 198, 737–746 (2003).
45. Wang, X. et al. Increased population of CD4+CD25high,
regulatory T cells with their higher apoptotic and
proliferating status in peripheral blood of acute
myeloid leukemia patients. Eur. J. Haematol. 75,
46. Taams, L. S. et al. Antigen-specific T cell suppression
by human CD4+CD25+ regulatory T cells. Eur. J.
Immunol. 32, 1621–1630 (2002).
47. Curotto de Lafaille, M. A., Lino, A. C., Kutchukhidze, N.
& Lafaille, J. J. CD25- T cells generate CD25+Foxp3+
regulatory T cells by peripheral expansion. J. Immunol.
173, 7259–7268 (2004).
48. Chen, W. et al. Conversion of peripheral CD4+CD25–
naive T cells to CD4+CD25+ regulatory T cells by
TGF-β induction of transcription factor Foxp3.
J. Exp. Med. 198, 1875–1886 (2003).
49. Kasow, K. A. et al. Human CD4+CD25+ regulatory
T cells share equally complex and comparable
repertoires with CD4+. J. Immunol. 172, 6123–6128
50. Hayashi, Y. et al. Antigen-specific T cell repertoire
modification of CD4+CD25+ regulatory T cells.
J. Immunol. 172, 5240–5248 (2004).
51. Fantini, M. C. et al. Cutting edge: TGF-β induces a
regulatory phenotype in CD4+CD25– T cells through
Foxp3 induction and down-regulation of Smad7.
J. Immunol. 172, 5149–5153 (2004).
52. Zheng, S. G. et al. Generation ex vivo of TGF-β-
producing regulatory T cells from CD4+CD25–
precursors. J. Immunol. 169, 4183–4189 (2002).
53. Wang, Z. et al. Role of IFN-γ in induction of Foxp3 and
conversion of CD4+ CD25– T cells to CD4+ Tregs.
J. Clin. Invest. 116, 2434–2441 (2006).
54. Groux, H. et al. A CD4+ T-cell subset inhibits antigen-
specific T-cell responses and prevents colitis. Nature
389, 737–742 (1997).
55. Miller, A. et al. Suppressor T cells generated by oral
tolerization to myelin basic protein suppress both
in vitro and in vivo immune responses by the release
of transforming growth factor β after antigen-specific
triggering. Proc. Natl Acad. Sci. USA 89, 421–425
56. Chen, Y. et al. Regulatory T cell clones induced
by oral tolerance: suppression of autoimmune
encephalomyelitis. Science 265, 1237–1240
57. Chai, J. G. et al. Anergic T cells act as suppressor cells
in vitro and in vivo. Eur. J. Immunol. 29, 686–692
58. Taams, L. S. et al. Anergic T cells actively suppress
T cell responses via the antigen-presenting cell.
Eur. J. Immunol. 28, 2902–2912 (1998).
59. Lombardi, G., Sidhu, S., Batchelor, R. & Lechler, R.
Anergic T cells as suppressor cells in vitro. Science
264, 1587–1589 (1994).
60. Grimbert, P. et al. Thrombospondin/CD47 interaction:
a pathway to generate regulatory T cells from human
CD4+CD25– T cells in response to inflammation.
J. Immunol. 177, 3534–3541 (2006).
61. Grogan, J. L. & Locksley, R. M. T helper cell
differentiation: on again, off again. Curr. Opin.
Immunol. 14, 366–372 (2002).
62. Shimizu, J., Yamazaki, S. & Sakaguchi, S. Induction
of tumor immunity by removing CD25+CD4+ T cells:
a common basis between tumor immunity and
autoimmunity. J. Immunol. 163, 5211–5218 (1999).
63. Dannull, J. et al. Enhancement of vaccine-mediated
antitumor immunity in cancer patients after depletion
of regulatory T cells. J. Clin. Invest. 115, 3623–3633
64. Ghiringhelli, F. et al. Tumor cells convert immature
myeloid dendritic cells into TGF-β-secreting cells
inducing CD4+CD25+ regulatory T cell proliferation.
J. Exp. Med. 202, 919–929 (2005).
65. Zorn, E. et al. IL-2 regulates FOXP3 expression in
human CD4+CD25+ regulatory T cells through a STAT-
dependent mechanism and induces the expansion of
these cells in vivo. Blood 108, 1571–1579 (2006).
66. Ahmadzadeh, M. & Rosenberg, S. A. IL-2
administration increases CD4+CD25hi Foxp3+
regulatory T cells in cancer patients. Blood 107,
67. Zhang, H. et al. Lymphopenia and interleukin-2
therapy alter homeostasis of CD4+CD25+ regulatory
T cells. Nature Med. 11, 1238–1243 (2005).
68. Viglietta, V., Baecher-Allan, C., Weiner, H. L. &
Hafler, D. A. Loss of functional suppression by
CD4+CD25+ regulatory T cells in patients with
multiple sclerosis. J. Exp. Med. 199, 971–979
69. Ehrenstein, M. R. et al. Compromised function of
regulatory T cells in rheumatoid arthritis and reversal
by anti-TNFα therapy. J. Exp. Med. 200, 277–285
70. Valencia, X. et al. TNF downmodulates the function of
human CD4+CD25hi T-regulatory cells. Blood 108,
71. van Amelsfort, J. M. et al. CD4+CD25+ regulatory
T cells in rheumatoid arthritis: differences in the
presence, phenotype, and function between peripheral
blood and synovial fluid. Arthritis Rheum. 50,
72. Lindley, S. et al. Defective suppressor function in
CD4+CD25+ T-cells from patients with type 1
diabetes. Diabetes 54, 92–99 (2005).
73. Verhagen, J. et al. Absence of T-regulatory cell
expression and function in atopic dermatitis skin.
J. Allergy Clin. Immunol. 117, 176–183 (2006).
74. Fletcher, J. M. et al. Cytomegalovirus-specific CD4+
T cells in healthy carriers are continuously driven to
replicative exhaustion. J. Immunol. 175, 8218–8225
75. Khan, N. et al. Cytomegalovirus seropositivity drives
the CD8 T cell repertoire toward greater clonality in
healthy elderly individuals. J. Immunol. 169,
76. Appay, V. et al. Memory CD8+ T cells vary in
differentiation phenotype in different persistent virus
infections. Nature Med. 8, 379–385 (2002).
77. Akbar, A. N. & Fletcher, J. M. Memory T cell
homeostasis and senescence during aging. Curr. Opin.
Immunol. 17, 480–485 (2005).
78. Mitchison, N. A. Induction of immunological paralysis
in two zones of dosage. Proc. R. Soc. Lond. B Biol. Sci.
161, 275–292 (1964).
79. Thorstenson, K. M. & Khoruts, A. Generation of
anergic and potentially immunoregulatory CD25+CD4
T cells in vivo after induction of peripheral tolerance
with intravenous or oral antigen. J. Immunol. 167,
80. Belghith, M. et al. TGF-β-dependent mechanisms
mediate restoration of self-tolerance induced by
antibodies to CD3 in overt autoimmune diabetes.
Nature Med. 9, 1202–1208 (2003).
81. Chatenoud, L. CD3-specific antibody-induced active
tolerance: from bench to bedside. Nature Rev.
Immunol. 3, 123–132 (2003).
82. Akdis, C. A. et al. Role of interleukin 10 in specific
immunotherapy. J. Clin. Invest. 102, 98–106 (1998).
83. Francis, J. N., Till, S. J. & Durham, S. R. Induction of
IL-10+CD4+CD25+ T cells by grass pollen
immunotherapy. J. Allergy Clin. Immunol. 111,
84. Smith, T. R. et al. Cat allergen peptide immunotherapy
reduces CD4+ T cell responses to cat allergen but
does not alter suppression by CD4+ CD25+ T cells:
a double-blind placebo-controlled study. Allergy 59,
85. Gardner, L. M. et al. Induction of T ‘regulatory’ cells by
standardized house dust mite immunotherapy: an
increase in CD4+ CD25+ interleukin-10+ T cells
expressing peripheral tissue trafficking markers.
Clin. Exp. Allergy 34, 1209–1219 (2004).
86. Read, S., Malmstrom, V. & Powrie, F. Cytotoxic
T lymphocyte-associated antigen 4 plays an essential
role in the function of CD25+CD4+ regulatory cells
that control intestinal inflammation. J. Exp. Med.
192, 295–302 (2000).
87. Zhang, X., Izikson, L., Liu, L. & Weiner, H. L.
Activation of CD25+CD4+ regulatory T cells by oral
antigen administration. J. Immunol. 167, 4245–4253
88. Shimizu, J. et al. Stimulation of CD25+CD4+
regulatory T cells through GITR breaks immunological
self-tolerance. Nature Immunol. 3, 135–142 (2002).
89. Stephens, L. A., Mottet, C., Mason, D. & Powrie, F.
Human CD4+CD25+ thymocytes and peripheral
T cells have immune suppressive activity in vitro.
Eur. J. Immunol. 31, 1247–1254 (2001).
90. Walker, M. R. et al. De novo generation of antigen-
specific CD4+CD25+ regulatory T cells from human
CD4+CD25– cells. Proc. Natl Acad. Sci. USA 102,
91. Skapenko, A., Kalden, J. R., Lipsky, P. E. &
Schulze-Koops, H. The IL-4 receptor α-chain-binding
cytokines, IL-4 and IL-13, induce forkhead box
P3-expressing CD25+CD4+ regulatory T cells from
CD25–CD4+ precursors. J. Immunol. 175, 6107–6116
92. Levings, M. K. et al. IFN-α and IL-10 induce the
differentiation of human type 1 T regulatory cells.
J. Immunol. 166, 5530–5539 (2001).
93. Barrat, F. J. et al. In vitro generation of interleukin 10-
producing regulatory CD4+ T cells is induced by
immunosuppressive drugs and inhibited by T helper
type 1 (TH1)- and TH2-inducing cytokines. J. Exp. Med.
195, 603–616 (2002).
94. Chang, C. C. et al. Tolerization of dendritic cells by T S
cells: the crucial role of inhibitory receptors ILT3 and
ILT4. Nature Immunol. 3, 237–243 (2002).
95. Michie, C. A., McLean, A., Alcock, C. & Beverley, P. C.
Lifespan of human lymphocyte subsets defined by
CD45 isoforms. Nature 360, 264–265 (1992).
96. Tough, D. F. & Sprent, J. Turnover of naive- and
memory-phenotype T cells. J. Exp. Med. 179,
97. McCune, J. M. et al. Factors influencing T-cell turnover
in HIV-1-seropositive patients. J. Clin. Invest. 105,
98. Macallan, D. C. et al. Measurement and modeling of
human T cell kinetics. Eur. J. Immunol. 33,
99. Hellerstein, M. K. et al. Subpopulations of long-lived
and short-lived T cells in advanced HIV-1 infection.
J. Clin. Invest. 112, 956–966 (2003).
100. Macallan, D. C. et al. Rapid turnover of effector-
memory CD4+ T cells in healthy humans. J. Exp. Med.
200, 255–260 (2004).
101. Liang, S. et al. Conversion of CD4+CD25– cells into
CD4+CD25+ regulatory T cells in vivo requires B7
costimulation, but not the thymus. J. Exp. Med. 201,
102. Morgan, M. E. et al. Expression of FOXP3 mRNA is
not confined to CD4+CD25+ T regulatory cells in
humans. Hum. Immunol. 66, 13–20 (2005).
103. Zheng, S. G. et al. Natural and induced CD4+CD25+
cells educate CD4+CD25– cells to develop suppressive
activity: the role of IL-2, TGF-β, and IL-10. J. Immunol.
172, 5213–5221 (2004).
We thank M. Soares for help preparing the original figures.
This work was financed by the Biotechnology and Biological
Sciences Research Council, the British Skin Foundation,
Dermatrust and the Sir Jules Thorne Charitable Trust and
Research into Ageing.
Competing interests statement
The authors declare no competing financial interests.
The following terms in this article are linked online to:
Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.
CD4 | CD25 | CD45 | CD127 | FOXP3
Access to this links box is available online.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 7 | MARCH 2007 | 237
© 2007 Nature Publishing Group