REVIEWS

Regulatory T cells, tumour immunity

and immunotherapy
Weiping Zou
Abstract | Tumours express a range of antigens, including self-antigens. Regulatory T cells
are crucial for maintaining T-cell tolerance to self-antigens. Regulatory T cells are thought
to dampen T-cell immunity to tumour-associated antigens and to be the main obstacle
tempering successful immunotherapy and active vaccination. In this Review, I consider the
nature and characteristics of regulatory T cells in the tumour microenvironment and their
potential multiple suppressive mechanisms. Strategies for therapeutic targeting of
regulatory T cells and the effect of regulatory T cells on current immunotherapeutic and
vaccine regimens are discussed.

Regulatory T cells Tolerance is the ‘holy grail’ in the field of immunology serve as a phenotypic marker for CD4+ suppressor
A T-cell population that can and it is generally divided into two broad categories: T cells or CD4+ regulatory T cells8. More recent stud-
functionally suppress an central tolerance and peripheral tolerance. Central ies have shown that the transcription factor forkhead
immune response by tolerance concerns immature lymphocytes as they dif- box P3 (FOXP3) is not only a key intracellular marker
influencing the activity of
another cell type. Several
ferentiate in the primary lymphoid organs. The main but is also a crucial developmental and functional fac-
phenotypically distinct mechanisms of inducing central tolerance are clonal tor for CD4+CD25+ regulatory T cells9–11. Thereafter, the
regulatory T-cell populations deletion and inactivation of self-reactive lymphocytes. notion of regulatory T cells was revived and the field of
exist. The classic regulatory Peripheral tolerance concerns mature lymphocytes regulatory T cells has evolved rapidly. Indeed, over the
T cells are CD4+CD25+
once they have exited primary lymphoid organs and past few years, several phenotypically distinct regulatory
FOXP3+ T cells (TReg cells).
are circulating in the periphery. It is suggested that T-cell populations have been suggested8,12–19 (TABLE 1).
regulatory T cells are responsible for inducing and The classic regulatory T cells are thymus-derived
maintaining peripheral tolerance. Regulatory T cells CD4+CD25+FOXP3+ T cells (TReg cells).
can be defined as a T-cell population that functionally Recent evidence has demonstrated that regulatory-
suppresses an immune response by influencing the T-cell-mediated immunosuppression is one of the cru-
activity of another cell type1. cial tumour immune-evasion mechanisms and the main
Regulatory T cells were initially described by obstacle of successful tumour immunotherapy20–23. By
Gershon et al. in the early 1970s and were called sup- focusing on the tumour microenvironment — the active
pressive T cells2,3. Five years later, Sehon and colleagues battlefront between tumours and the host immune
suggested that these regulatory T cells negatively regu- system — I summarize and review recent data demon-
lated tumour immunity and contributed to tumour strating that tumours actively prevent the induction of
growth in mice4. After another five years, North and tumour-associated antigen (TAA)-specific immunity
colleagues published a series of experiments provid- through induction of regulatory T-cell trafficking, dif-
ing evidence that CD4+CD25+ T cells from tumour- ferentiation and expansion. Mechanisms to revert regu-
bearing mice inhibited tumour rejection, indicating latory T-cell trafficking, differentiation, function and
the existence of tumour-suppressor T cells5–7. These signalling are discussed as novel therapeutic strategies
pioneering studies established the field of regulatory for cancer.
T cells in tumour immunology. Unfortunately, despite
Department of Surgery, the importance of these studies there was extensive Regulatory T cells and mouse tumours
University of Michigan, skepticism in the immunological field about the exist- TReg cells constitute 5–10% of peripheral CD4+ T cells
1500 East Medical Center ence of these cells, and suppressive T cells left the in normal mice and humans. The role of TReg cells in
Drive, Ann Arbor, Michigan
48109-0346, USA.
centre stage of immunology for decades. However, mouse tumour immunopathogenesis has largely been
e-mail: [email protected] in 1995, Sakaguchi and colleagues showed that the defined using reagents that target TReg cells in vivo in
doi:10.1038/nri1806 interleukin-2 (IL-2) receptor α-chain, CD25, could tumour-bearing mice.

NATURE REVIEWS | IMMUNOLOGY VOLUME 6 | APRIL 2006 | 295

© 2006 Nature Publishing Group
REVIEWS

Table 1 | Regulatory T-cell populations

Cell subset Suggested origin Original experimental setting Suggested suppressive References
mechanism
CD4+ T-cell subset
CD4+CD25+FOXP3+ (TReg) cells* Thymus Mouse autoimmune diseases In vitro: cell–cell contact. 8
In vivo: multiple modes of
action
CD4+IL-10+FOXP3– (TR1) cells* Periphery In vitro mouse cell-culture system with IL-10 IL-10 12
+ +
CD4 TGFβ (TH3) cells* Periphery In-vivo-induced oral tolerance in mice TGFβ 13
+
CD8 T-cell subset
CD8+CD25+ T cells Thymus Human thymus TGFβ and CTLA4 14
+ –
CD8 CD28 T cells Periphery Human allogeneic organ transplantation Targeting ILT3 and ILT4 15
on DCs
CD8+CD62L+CD122+ T cells ND Normal neonatal mice ND 16
+ +
CD8 IL-10 T cells Periphery Human ovarian cancer, in vitro culture IL-10 17–19
system
*CD4+ regulatory T cells are conceptually divided into three populations. CD4+CD25+FOXP3+ T cells are thought to be thymus derived, and are termed naturally
occurring regulatory T cells (TReg cells); CD4+IL-10+FOXP3– regulatory T cells can be induced in vitro with various protocols or in vivo in response to exogenous
antigen challenge, and are termed adaptive regulatory T cells, induced regulatory T cells or T-regulatory 1 cells (TR1 cells); CD4+TGFβ+ T cells are induced in the
context of oral tolerance and are termed TH3 cells. Notably, in this classification, the cytokine pattern and suppressive modes are not mutually exclusive and they
often overlap. For example, human and mouse tumour CD4+CD25+ TReg cells express both FOXP3 and IL-10 (as discussed in the text). CTLA4, cytotoxic T-lymphocyte-
associated antigen 4; DC, dendritic cell; FOXP3, forkhead box P3; IFN, interferon; IL-10, interleukin-10; ILT, immunoglobulin-like transcript; ND, not determined;
TGFβ, transforming growth factor-β; TH, T helper.

Treatment with CD25-specific antibody. CD25 is GITR ligation directly reduced suppressor activity
expressed by activated effector T cells and TReg cells. of TReg cells in vitro43. Furthermore, administration of
The role of TReg cells in mouse tumour immunity was GITR-specific antibody protected mice from B16 tumour
initially studied by systemic depletion of CD25+ T cells. challenge41, and it induced tumour regression in mice
Early studies demonstrated that in vivo administration bearing methylcholanthrene (methA)-induced sarcoma
of CD25-specific antibody (PC61) suppressed growth of and N-nitroso-N-methylurethane-induced, undifferen-
several progressively growing tumours24,25 (TABLE 2). tiated colon carcinoma (CT26)34. Theoretically, GITR-
These studies showed a correlation between reduced specific antibody might attenuate the suppressor activity
TReg-cell numbers and reduced tumour volume. In sup- of TReg cells in vivo as reported in vitro43. However, in
port of these findings, depletion of total CD4+ T cells addition to TReg cells, other cell populations express
was found to improve tumour immunity and induce GITR; therefore, it will be interesting to provide a link
effective tumour rejection26–28. The antitumour effects between the treatment with GITR-specific antibody and
of in vivo administration of CD25-specific antibody the potential deficits in numbers, functional competence
were confirmed in several mouse tumour models29–39 or specificity of TReg cells in vivo. For example, it remains
(TABLE 2). Some of these studies tested several combina- to be defined whether purified TReg cells from mice
torial treatments, which demonstrated possible additive treated with GITR-specific antibody show reduced sup-
or synergistic protective effects27,29,32,33,35 (TABLE 2). pressor activity in vitro. Nonetheless, the mechanism of
In addition to studies of CD25+ T-cell depletion, action and potential synergistic effects of GITR-specific
experiments with adoptively transferred human40 and antibody with current immunotherapeutic strategies
mouse41,42 TReg cells provided a direct link between needs careful investigation.
Antigen-presenting cells
(APCs). Cells that uptake, TReg cells and reduced tumour immunity. Tumour-
process and present antigen to specific CD8+ T cells were transferred with either TReg cells Treatment with CTLA4-specific antibody. Mouse TReg cells
other immune cells to initiate or CD4+CD25– T cells into mice bearing B16 melanoma. constitutively express cytotoxic T-lymphocyte-associated
and activate immune In mice that received TReg cells, but not in mice that antigen 4 (CTLA4)44,45. Interest in the role of TReg cells has
responses. Monocytes,
macrophages, dendritic cells
received CD4+CD25– T cells, CD8+ T-cell-mediated been provoked by studies of CTLA4-specific antibody
and B cells are APCs. Dendritic immunity was abolished41,42. These data indicated that treatment in both mouse tumour models and patients
cells are the most potent APCs. TReg cells inhibit mouse TAA-specific immunity. with cancer.
In 1996, Allison and colleagues reported that admin-
Cytotoxic T-lymphocyte-
Treatment with GITR-specific antibody. In addition to istration of CTLA4-specific antibody resulted in mouse
associated antigen 4
(CTLA4). Following ligation by CD25-specific antibody, an agonistic antibody that is tumour rejection, even of pre-established tumours.
B7.1 (CD80) or B7.2 (CD86) specific for the glucocorticoid-induced tumour-necrosis Furthermore, this treatment induced immunity to a
on APCs, CTLA4 signalling in factor (TNF)-receptor related protein (GITR; also known second tumour challenge46. At the time, this study was
activated T cells induces cell- as DTA1) has been used to study the role of TReg cells in not deliberately designed to deplete CTLA4+ TReg cells
cycle arrest, reduces cytokine
production and diminishes
mouse tumour immunity. GITR is expressed on the sur- or to alter CTLA4 signalling in TReg cells46. Instead, the
T-cell responses. TReg cells face of TReg cells, but is also expressed to various degrees data were interpreted to mean that CTLA4, a counter-
constitutively express CTLA4. on CD4+CD25– T cells and antigen-presenting cells (APCs)43. receptor for the B7 family of co-stimulatory molecules,

296 | APRIL 2006 | VOLUME 6 www.nature.com/reviews/immunol

© 2006 Nature Publishing Group
 REVIEWS

Table 2 | In vivo depletion of CD25+ cells and tumour models

Tumour type* Treatment Observation References
Mouse melanoma (B16) CD25-specific antibody Reduced tumour growth and increased 25,30
survival
CD25- and B7-H1-specific antibodies Improved tumour immunity and reduced W.Z. et al., unpublished
tumour growth observations
CD25-specific antibody and IFNα Induced tumour regression 29
Mouse melanoma (B16-F10) CD25-specific antibody and DCs Improved DC-mediated immunity and tumour 29,33
regression
CD25-specific antibody and IL-12 Improved tumour immunity against tumour 32
challenge
Mouse melanoma (B16-BL6) CD25- and CTLA4-specific antibodies Improved tumour immunity and mouse survival 27
Mouse spontaneous leukaemia CD25-specific antibody Reduced tumour growth and/or induced 24
(ASL1) tumour regression
Mouse radiation-induced CD25-specific antibody Induced tumour regression and immunity 24,25
leukaemia (RL1) against tumour challenge
Mouse DBA-induced leukaemia CD25-specific antibody Reduced tumour growth and/or induced 24
(EL4) tumour regression
Mouse spontaneous leukaemia CD25-specific antibody No effects on tumour growth 24
(AKSL2)
Mouse radiation-induced CD25-specific antibody No effects on tumour growth 24
leukaemia (RL8)
Mouse C3H-derived CD25-specific antibody Reduced tumour growth and/or induced 24
plasmacytoma (X5563) tumour regression
Mouse DBA2-derived CD25-specific antibody Reduced tumour growth and/or induced 24
mastocytoma (P815) tumour regression
Mouse mineral-oil-induced CD25-specific antibody Induced tumour regression and immunity 24
myeloma (MOPC70-A) against tumour challenge
Mouse methylcholanthrene- CD25-specific antibody Reduced tumour growth and/or induced 24,31,34
induced fibrosarcoma (CMS17) tumour regression
Mouse colorectal tumour (CT26) CD25-specific antibody Improved tumour immunity and reduced 36
tumour growth
CD25-specific antibody and Improved tumour immunity against tumour 35
irradiated CT26 challenge
Human cancer of ovary, breast Denileukin diftitox TReg-cell depletion, increased T-cell 37
and lung activation and induced tumour regression
Human renal-cell carcinoma Denileukin diftitox and DCs TReg-cell depletion, improved DC- 38
mediated TAA-specific immunity
Human melanoma Denileukin diftitox Negligible TReg-cell depletion, no clinical 39
efficacy
*The tumour-cell line is given in parentheses. DC, dendritic cell; DBA, dimethylbenzanthracene; IFN, interferon; TAA, tumour-associated antigen; TReg cell,
CD4+CD25+ regulatory T cell.

counteracted the CD28-mediated co-stimulatory signal neither a decrease in the number of peripheral-blood
that is necessary to fully activate T cells46. In the follow- TReg cells, nor a decrease in FOXP3 mRNA expression
ing years, this group and others confirmed these obser- in these cells post-treatment 55. Therefore, a direct
FIGO stages
vations in various tumour models, and synergistic effects link between attenuated CTLA4 signals in TReg cells
Based on the International have been documented when CTLA4-specific antibody and improved tumour immunity remains to be estab-
Federation of Gynaecology was used in combination with other treatments47–54 lished. Notably, Melief and colleagues 27 proposed
and Obstetrics (FIGO), ovarian (TABLE 3). that TReg cells and CTLA4 signalling represent two
cancer is classified into four
Current knowledge suggests that in vivo admin- independent and synergistic regulatory mechanisms
surgicopathological stages:
stage I, the tumour is confined istration of CTLA4-specific antibody blocks CTLA4 for suppression of T-cell activation in vivo. Five lines
to the ovaries; stage II, the signalling in CTLA4+ TReg cells and in turn promotes of experimental evidence exist to support this notion.
tumour is confined to the tumour immunity. However, in patients with stage IV First, in the absence of CD25+ cells, CTLA4 block-
pelvis; stage III, the tumour melanoma or renal-cell cancer that are treated with ade increased the induction of TAA-specific effector
is confined to the abdomen;
and stage IV, the tumour
the CTLA4-specific antibody, administration of the T cells. Second, depletion of CD25+ cells and CTLA4
extends outside the abdominal antibody did not inhibit the suppressive activity of blockade synergistically improved TAA-specific T-cell
cavity. TReg cells in vitro or in vivo. Furthermore, there was immunity27. Third, administration of CD25-specific

NATURE REVIEWS | IMMUNOLOGY VOLUME 6 | APRIL 2006 | 297

© 2006 Nature Publishing Group
REVIEWS

Table 3 | Treatment with CTLA4-specific antibody and tumour models

Tumour type* Treatment Observation Reference
Mouse colon tumour (51BLim10) CTLA4-specific antibody Tumour regression 46
Mouse prostate tumour (TRAMP) CTLA4-specific antibody Reduced tumour growth 47
Tumour resection and CTLA4-specific Reduced metastatic relapse 48
antibody
Irradiated tumour and CTLA4-specific Reduced tumour incidence and grade 49
antibody
GM-CSF+ tumour and CTLA4-specific Reduced tumour incidence and grade 49
antibody
Mouse mammary carcinoma (SM1) GM-CSF+ tumour and CTLA4-specific Reduced tumour growth and/or 50
antibody regression
Mouse B6-derived melanoma (B16) GM-CSF+ tumour and CTLA4-specific Reduced tumour growth 51
antibody
CD25- and CTLA4-specific antibodies Reduced tumour growth 27
Mouse methylcholanthrene-induced rMVAp53 vaccine and CTLA4-specific Tumour regression and increased 52
sarcoma (methA) antibody survival
Human metastatic melanoma Simultaneous peptide vaccine and CTLA4- Objective response in 3/14 patients, 53
specific antibody autoimmune disease in 6/14 patients
GM-CSF+ tumour and CTLA4-specific Tumour necrosis in 3/3 patients 54
antibody
Previous peptide vaccine and CTLA4-specific No tumour necrosis in 4/4 patients 54
antibody
Human metastatic ovarian carcinoma GM-CSF+ tumour and CTLA4-specific Stabilization of CA125 in 2/2 patients 54
antibody
*The mouse tumour model is given in parentheses. CA125, a protein that is released by ovarian cancer cells and functions as a surrogate marker for ovarian tumour
growth; CTLA4, cytotoxic T-lymphocyte-associated antigen 4; GM-CSF, granulocyte/macrophage colony-stimulating factor; methA, methylcholanthrene; rMVAp53,
recombinant-modified-vaccinia-virus-Ankara expressing wild-type murine p53; TRAMP, transgenic adenocarcinoma mouse prostate.

antibody did not induce severe autoimmune disease, studies in normal mice63, in mice bearing B16 melano-
unlike administration of CTLA4-specific antibody24,27, mas41 or neu-expressing tumours64, and in rats bearing
indicating that the targets of CD25-specific antibody a chemically induced colon cancer (PROb cell line)65.
and CTLA4-specific antibody might not be identical. However, as a chemotherapeutic agent, the effects of
Fourth, in vivo administration of CTLA4-specific cyclophosphamide on TReg cells might not be specific
antibody increased tumour immunity, but had no and selective. In addition to depleting TReg cells, cyclo-
detectable effects on peripheral TReg cells55. The fact phosphamide-based therapy can deplete CD4+CD25–
that TReg cells showed equivalent suppressive activity T cells and CD4 + CD25 mid T cells 66 . Nonetheless,
in Ctla4–/– mice and Ctla4+/+ mice further supports this although the long-term clinical efficacy of cyclophos-
notion56. phamide as a TReg-cell-depleting agent remains to be
Collectively, these studies show that signals from determined, current data support the idea that low doses
CTLA4 (possibly including signals from CTLA4 + of cyclophosphamide could be used clinically to reduce
TReg cells) negatively regulate tumour immunity and limit TReg-cell-mediated suppressive activity.
the efficacy of tumour immunotherapy. Collectively, studies in mouse models show that
depleting TReg cells or reducing their suppressive activ-
Treatment with cyclophosphamide. Cyclophosphamide ity improves spontaneous or immunotherapy-mediated
is an alkylating agent that mediates DNA crosslinking tumour clearance.
and is used to treat various tumours. High doses of
cyclophosphamide are required for effective tumour Regulatory T cells and human tumours
chemotherapy that might lead to immunosuppression. Although TReg cells have been studied extensively in
Strikingly, low doses of cyclophosphamide induced mouse cancer models, the role of regulatory T cells
improved immune responses in various animal tumour in human tumour immunity is less well studied.
models57–59 and in patients with metastatic melanoma60, TReg cells mediate peripheral tolerance by suppressing
and this might be a result of the depletion of TReg cells58. self-antigen-reactive T cells20,22,67. As most tumour anti-
These observations were initially confirmed when it gens are self-antigens68, TReg-cell-mediated suppression
was shown that cyclophosphamide treatment resulted of TAA-reactive lymphocytes has been proposed as
in immune-mediated regression of the immunogenic, a potential mechanism to explain the failure of anti-
cyclophosphamide-resistant lymphoma (L5178Y tumour immunity23,40. Both TReg cells and CD8+ regula-
cell line) in mice61,62. Depletion of TReg cells by cyclo- tory T cells have been reported in the context of human
phosphamide was further demonstrated by recent tumour immunity.

298 | APRIL 2006 | VOLUME 6 www.nature.com/reviews/immunol

© 2006 Nature Publishing Group
 REVIEWS

TReg cells. In 2001, June and colleagues observed increased Source of regulatory T cells in the tumour
numbers of TReg cells in patients with non-small-cell Functional regulatory T cells are found in the tumour
lung cancer and ovarian cancer compared to normal microenvironment; these include thymus-derived
donors69. Thereafter, a higher frequency of TReg cells in ‘natural’ TReg cells and locally induced T-regulatory 1
peripheral blood was reported in patients with various (TR1) cells (see TABLE 1 for definitions of regulatory
cancers including breast cancer70, colorectal cancer71,72, T cells and see FIG. 1).
oesophageal cancer73, gastric cancer73,74, hepatocellular
carcinoma75, leukaemia76, lung cancer69,72, lymphoma76,77,
a Trafficking
melanoma78,79, ovarian cancer69 and pancreatic cancer70. CCR4
This list continues to grow following the current interest • Thymus
• Bone marrow
of studying TReg cells in human tumours. These studies • Blood
demonstrated that peripheral TReg cells showed potent • Lymph node
Regulatory T cell
suppressive activity in vitro and indicated that a high
CCL22
frequency of TReg cells would temper TAA-specific
immunity in patients with cancer. Tumour
microenvironment
TReg cells have also been extensively studied in human
ovarian cancer. TReg cells within the ovarian-tumour
microenvironment expressed FOXP3, inhibited TAA-
specific CD8+ T-cell cytotoxicity and contributed to
b Tumour environmental differentiation
tumour growth in vivo in a human severe combined
immunodeficiency (SCID) chimeric model40. An accu- • IL-10
mulation of TReg cells in the tumour predicted a striking • TGFβ Induction of
• VEGF regulatory T cell
reduction of patient survival40. Notably, TReg cells that
were isolated from peripheral blood, ascites or solid
tumours of patients with ovarian cancer equally sup-
pressed T-cell activation in vitro40. Therefore, TReg cells
in the tumour microenvironment are not functionally Dysfunctional DC
Severe combined
immunodeficiency superior to their counterparts in blood. The reduced c Expansion
(SCID). SCID mice do not have TAA-specific T-cell immunity is probably a result
B cells or T cells. Tumour cells of the increased accumulation of TReg cells, rather than
can be grown in these mice the superior suppressive activity in the tumour micro-
without rejection.
environment. This is in contrast to some autoimmune
Regulatory T cell
Plasmacytoid dendritic cells diseases, in which onset is related to reduced TReg-cell
A subset of dendritic cells that suppressive capacity80–82. Understanding this striking dif-
is lineage–HLA-DR+CD11c– ference and phenomenon could provide a clue to address d Conversion TReg cell
TGFβ
mononuclear cells with a (CD4+CD25+FOXP3+)
the distinct immunopathogenesis of human cancers and
microscopic appearance
similar to plasmablasts. autoimmune diseases.
Plasmacytoid dendritic cells
are the main producers of CD8+ regulatory T cells. Plasmacytoid dendritic cells (DCs) CD4+CD25+ T cell
type I interferon. from human ovarian tumours induced IL-10+CD8+ Figure 1 | Regulatory T cells in the tumour
Central memory T cells
T cells, which inhibited TAA-specific T-cell immunity microenvironment. Regulatory T cells are found in the
Antigen-experienced CD8+ in vitro17,18. These tumour-associated plasmacytoid- tumour microenvironment. There are four potential
T cells that lack immediate DC-induced suppressive T cells had a central memory T-cell sources for regulatory T cells in the tumour
effector function but are able phenotype (CD8+CD45RO+ CC-chemokine receptor 7+ microenvironment. a | Trafficking. TReg cells
to mediate rapid recall
(CCR7+)) and suppressed TAA-specific T-cell effector (CD4+CD25+FOXP3+ T cells) from the thymus, lymph
responses. They also rapidly nodes, bone marrow and peripheral blood traffic to the
develop the phenotype and functions through IL-10 production. Repetitive stimula-
tumour. TReg cells express CC-chemokine receptor 4
function of effector-memory tion with myeloid DCs could not prevent their suppres- (CCR4), and abundant expression of CC-chemokine
cells after restimulation with sive functions. These data indicate that vaccination with ligand 22 (CCL22; the ligand for CCR4) in the tumour
antigen. Central memory myeloid DCs alone might not be sufficient to recover
T cells retain the migratory
microenvironment stimulates TReg-cell tumour infiltration.
properties of naive cells and
defective T-cell immunity. Furthermore, tumour- b | Differentiation. The tumour microenvironment
therefore circulate through associated, but not normal, primary CD8+CD45RO+ contains molecules that can suppress antigen-presenting
secondary lymphoid organs. CCR7+ T cells expressed IL-10 and suppressed T-cell cell (APC) differentiation and function. These
activation in vitro, which suggests that tumour plasma- dysfunctional APCs can in turn stimulate regulatory T-cell
Myeloid dendritic cells differentiation. c | Expansion. Dendritic cells (DCs) can
cytoid DCs might induce CD8+CCR7+IL-10+ regulatory
A subset of dendritic cells that stimulate regulatory T-cell expansion, and it is predicted
is lineage–HLA-DR+CD11c+ T cells in vivo18. Consistent with these observations,
that DCs in the tumour microenvironment and draining
mononuclear cells with a human plasmacytoid DCs can induce allogeneic CD8+
lymph nodes might also induce regulatory T-cell
monocytoid appearance. suppressive T cells19, and mouse CD8+ T cells with a expansion. d | Conversion. Normal T cells can be
Human myeloid dendritic cells central-memory phenotype also have regulatory T-cell
might develop from myeloid converted into TReg cells by transforming growth factor-β
precursors (for example,
capacity16. Therefore, CD8+ regulatory T cells exist (TGFβ), and high levels of TGFβ are often found in the
monocytes, macrophages and might be functionally operative in patients with tumour microenvironment. FOXP3, forkhead box P3;
and CD11c+ precursors). cancer (TABLE 1). IL, interleukin; VEGF, vascular endothelial growth factor.

NATURE REVIEWS | IMMUNOLOGY VOLUME 6 | APRIL 2006 | 299

© 2006 Nature Publishing Group
REVIEWS

TReg cells differentiate in the thymus. The mouse It has been suggested that thymus-derived TReg cells
thymus produces TReg cells as a separate lineage that express FOXP3 but not IL-10 and TGFβ, whereas
expresses FOXP3 (REFS 9,83). Because tumour-associated TR1 cells express IL-10 but not FOXP3 (REFS 20,100)
TReg cells express FOXP3 mRNA and protein28,34,40, it is (TABLE 1). However, human TReg-cell clones specific for the
possible that these cells traffic to tumours from the thy- TAA LAGE-1 (REF. 99) expressed both FOXP3 and IL-10.
mus, bone marrow84, lymph nodes and peripheral blood In a T-cell receptor (TCR)-transgenic model, in which
under the influence of tumour microenvironmental both conventional T cells and TReg cells were specific for
CC-chemokine ligand 22 (CCL22)40 (FIG. 1). influenza virus haemagglutinin, it was found that these
Thymic function is largely reduced after ado- haemagglutinin-specific TReg cells101,102 expressed both
lescence, whereas T Reg cells persist throughout the FOXP3 and IL-10. Therefore, TAA-specific TReg cells
human lifespan. Presumably, regulatory T cells seem to share common features with naturally occurring
can be induced and differentiate in the periphery, TReg cells and induced TR1 cells. Although the role of IL-10
such as in the tumour microenvironment. Tumour was not investigated in these studies101,102, the data predict
environmental factors, such as vascular endothelial that IL-10 derived from TAA-specific TReg cells might
growth factor (VEGF), IL-10 and transforming growth profoundly suppress APC and T-cell function100 (FIG. 2).
factor-β (TGFβ) suppressed DC differentiation and In addition to APCs and T cells, TReg cells also suppress
function, resulting in immature and/or partially dif- natural killer (NK)-cell function in a TGFβ-dependent
ferentiated DCs 22,85. Plasmacytoid DCs associated manner in tumour-bearing mice103. These data indicate
with ovarian cancer induced CD8 + FOXP3 – IL-10 + that TReg cells can dampen adaptive immunity as well as
regulatory T cells17,18, and dysfunctional myeloid DCs innate immunity.
also induced IL-10+ regulatory T cells in vitro86 and Notably, TGFβ and IL-10 can be produced by
in vivo in patients with cancer87,88. Tumours convert multiple cell types in the tumour microenvironment.
DCs into TGFβ-expressing immature myeloid DCs TReg cells might not be the main source of IL-10 and TGFβ
that are capable of promoting TReg-cell proliferation89. in vivo. Nonetheless, it is clear that TReg cells can medi-
Normal mature DCs stimulated the in vivo expan- ate suppression through the actions of IL-10 and TGFβ
sion of self-antigen-specific T Reg cells in mice 90,91. in vivo in mouse tumour models103–106. One possibility is
Analogously, thymus-derived TReg cells in the tumour that TReg cells might condition APCs and enable them
micro environment might clonally expand following to express functional IL-10 (REFS 107,108). Suppression
stimulation by tumour-associated DCs (FIG. 1). might then be mediated locally by APC-derived IL-10
In addition to trafficking, differentiation and expan- within the tumour microenvironment (FIG. 2).
sion of regulatory T cells, TReg cells can also be present
in a tumour as a result of conversion from CD4+CD25– Competitive consumption of IL-2. The IL-2 receptor con-
T cells92–95. TGFβ, which is present at high levels in sists of a heterotrimeric receptor complex that is made
the tumour microenvironment, might mediate this up of CD25 (α-chain), CD122 (β-chain) and CD132
conversion92,94,96. Similarly, IL-10 is often found in the (γ-chain). TReg cells express this heterotrimeric receptor
tumour microenvironment22, where it supports the dif- complex, which has a 100-fold higher affinity for IL-2
ferentiation of CD4+IL-10+TGFβ+ regulatory T cells28,97 than the dimeric form (made up of CD122 and CD132).
and induces TR1 cells12 (TABLE 1). However, it is unknown Therefore, competition for IL-2 between TReg cells and
LAGE-1 and NY-ESO-1 whether IL-10 is implicated in the differentiation and conventional T cells was suggested as a suppressive
LAGE-1 and NY-ESO-1 (New expansion of TReg cells. mechanism20,109,110. The main cellular source of IL-2
York oesophageal squamous-
cell carcinoma 1) are two
The tumour microenvironment might contain in vivo is conventional T cells. Most tumour-infiltrating
cancer- and testis-specific thymus-derived natural TReg cells, expanded and con- CD4+ T cells are TReg cells in established tumours28,34,40.
antigens that show 94% verted natural TReg cells, and locally differentiated and Therefore, the supply of IL-2 might be limited in the
sequence identity. The genes expanded TR1 cells. Cell surface CD25, rather than intra- tumour microenvironment and competitive consump-
encoding LAGE-1 and
cellular FOXP3, is used as a marker to isolate and sort tion of IL-2 could be a minor suppressive mechanism for
NY-ESO-1 are both located
on chromosome Xq28 and viable TReg cells for functional experiments. As there is TReg cells in established tumours (FIG. 2). The role of IL-2
are frequently co-expressed. no reliable cell-surface marker to distinguish these regu- in TReg-cell function is discussed later.
LAGE-1 and NY-ESO-1 have latory T-cell subsets, the statement that “so far, no one
been shown to be expressed has had a pure population of suppressor T cells in a test Perforin and granzyme pathway and TReg-cell-mediated
in 25–50% of various tumour
types, such as melanoma,
tube” remains as true today as in 1988 (REF. 98). direct killing. CD8+ T cells and NK cells use the perforin
breast carcinoma, prostate and granzyme pathways to kill infected cells and tumour
and bladder cancers. Suppressive mechanisms of regulatory T cells cells. Activated human TReg cells express granzyme A
Suppressive mechanisms of regulatory T cells have and kill T cells and APCs through perforin111. A study
Natural killer cell
been addressed using many in vitro and in vivo using granzyme B-deficient and perforin-deficient mice
Cytotoxic lymphocytes that
are distinguished from mouse models. Multiple suppressive mechanisms showed that TReg cells mediated suppression through a
CD8+ T cells by their lack including cell–cell contact and soluble factors have granzyme B-dependent but perforin-independent mecha-
of rearrangement of T-cell been proposed20,23,67,99. These mechanisms have not nism112. Nonetheless, it is unknown whether TReg cells
receptor genes. They have been exclusively investigated in tumour settings. in tumours express perforin and granzyme B, and if
abundant granule-containing
cytoplasm, and their functions
Therefore, I summarize the relevant information so, whether these pathways are functional for TReg cells.
are cell killing and cytokine and discuss their relevance in the context of tumour Another interesting question is whether TReg cells could
production. immunity (FIG. 2). kill tumour cells through similar mechanisms (FIG. 2).

300 | APRIL 2006 | VOLUME 6 www.nature.com/reviews/immunol

© 2006 Nature Publishing Group
 REVIEWS

a APC b c d
Regulatory Activated
T cell Regulatory
regulatory T cell
T cell

CD80 and/or CD86

Induction CTLA4
of B7-H4
expression Regulatory
Perforin and Granzyme B T cell

IDO Metabolized IL-10 TGFβ

tryptophan
Perforin
B7-H4 pore
Tryptophan
?
Decreased
expression of:
T cell • MHC molecules
• CD80 and CD86
• IL-12

T-cell cycle APC and T-cell T-cell apoptosis T-cell anergy APC T-cell anergy
arrest apoptosis dysfunction

Figure 2 | Possible suppressive mechanisms of tumour environmental regulatory T cells. Possible suppressive
mechanisms of regulatory T cells (TReg cells; CD4+CD25+FOXP3+) have been addressed in various models in vitro and in vivo.
Multiple suppressive mechanisms rather than a single mode of action are proposed. a | TReg cells induce B7-H4 expression
by antigen-presenting cells (APCs), and in turn these B7-H4+ APCs induce T-cell cycle arrest through B7-H4. b | Activated
human TReg cells directly kill target cells such as T cells and APCs through perforin- or granzyme B-dependent pathways.
c | Cytotoxic T-lymphocyte-associated antigen 4 (CTLA4)+ TReg cells induce indoleamine 2,3-dioxygenase (IDO) expression
by APCs, and IDO-expressing APCs in turn suppress T-cell activation by reducing tryptophan. d | Regulatory T cells,
including T Reg cells, might release interleukin-10 (IL-10) and transforming growth factor (TGFβ) in vivo, and directly inhibit
T-cell activation and suppress APC function, by inhibiting expression of MHC molecules, CD80, CD86 and IL-12.

CTLA4 induction of indoleamine 2,3-dioxygenase- APCs by stimulating B7-H4 expression. We therefore

expressing APCs. Indoleamine 2,3-dioxygenase (IDO) provide a novel cellular and molecular mechanism for
is an enzyme that degrades the essential amino acid TReg-cell-mediated immunosuppression at the level of
tryptophan. IDO+ human APCs reduce tryptophan, APCs107,108. Because high numbers of TReg cells40 and
and suppress T-cell activation and promote tolerance B7-H4+ macrophages107 are found in tumour environ-
in the tumour microenvironment113. Mouse CTLA4+ ments, this suppressive mechanism might also occur in
TReg cells mediate their suppressive activity by induc- patients with tumours (FIG. 2).
ing the expression of IDO in APCs through CTLA4 In summary, TReg cells target not only T cells but also
(REFS 113,114). Because human tumour TReg cells express APCs to efficiently temper TAA-specific immunity. It
CTLA4 (REF. 40) and because IDO+ APCs are found in is probable that multiple TReg-cell-mediated suppressive
human tumours and within tumour-draining lymph mechanisms might be operative in vivo.
nodes113, it seems reasonable to speculate that this
mechanism might occur in vivo (FIG. 2). TAA specificity and regulatory T cells
Our current knowledge of the antigen specificity of
Induction of B7-H4+ APCs. B7-H4 (also known as B7x, TReg cells has come largely from studies with antigen-
Indoleamine 2,3-
dioxygenase
B7S1) is a recently discovered member of the B7 family specific TCR-transgenic mice20,23. However, the TAA
(IDO). An intracellular of T-cell co-stimulatory molecules. B7-H4 negatively specificity of TReg cells is poorly understood. Here I focus
haem-containing enzyme that regulates T-cell responses 115–117. We observed that on the TAA specificity of TReg cells in the periphery.
catalyses oxidative catabolism ovarian-tumour-associated macrophages, but not
of tryptophan.
normal macrophages, expressed B7-H4 (REF. 107) . Are TReg cells TAA specific? Thymus-derived TReg cells
B7-H4 Tumour B7-H4+ macrophages inhibited TAA-specific maintain mouse and human T-cell self-tolerance23 and
A newly defined B7-family T-cell responses, which could be reversed by blocking many TAAs are self-antigens68; furthermore, TReg cells are
member. B7-H4 fusion protein B7-H4. Interestingly, TReg cells, but not normal T cells, specific for self-antigens, so it is possible that TReg cells
inhibits T-cell-mediated induced B7-H4 expression by APCs including mono- are specific for at least one subset of TAA68. In support
immune responses. The
receptor, regulation and
cytes, macrophages and myeloid DCs, and rendered of this, human melanoma-infiltrating TReg cells were
detailed function remain APCs immunosuppressive through B7-H4. These data cloned, and in one case the target antigen was found
to be defined. indicate that TReg cells convey suppressive activity to to be LAGE-1, which is a cancer- and testis-specific

NATURE REVIEWS | IMMUNOLOGY VOLUME 6 | APRIL 2006 | 301

© 2006 Nature Publishing Group
REVIEWS

antigen99. These cloned TReg cells suppressed LAGE-1- of intratumoural TReg cells induced potent T-cell tumour
specific T-cell activation. It is important to identify the immunity and resulted in regression of large established
ligands for TReg cells, to compare them to the ligands for tumours28. In our studies of advanced ovarian cancer40,
TAA-specific effector T cells, and to determine whether a large number of TReg cells accumulated in the tumour
there are ligands that are exclusively recognized by mass and associated malignant ascites, but there were
TReg cells, and furthermore, whether TReg cells recognize few in the tumour-draining lymph nodes, indicating that
mutated (novel) tumour antigens. TReg cells predominantly inhibit extranodal effector-cell
function at this stage. However, human tumorigenesis is a
Is TAA specificity acquired in tumours and draining long process, and TReg cells might be present in the drain-
lymph nodes? Regulatory T cells can be induced in ing lymph nodes before clinical manifestation of disease
tumour-draining lymph nodes and in the tumour micro- and might block TAA-specific T-cell priming during the
environment. Similar to TAA-specific effector T cells, priming process. In addition to TReg cells, other classes of
regulatory T cells can attain TAA specificity in the course regulatory T cells, such as CD8+ regulatory T cells, might
of TAA-specific induction. For example, dysfunctional also migrate to draining lymph nodes and block T-cell
APCs that express TAA might induce regulatory T-cell priming18. Nonetheless, TReg cells are able to hinder both
differentiation. In support of this, TReg cells isolated TAA-specific priming and effector function.
from mice that were immunized with self-antigens
strongly suppressed peptide-specific T-cell prolifera- Targeting regulatory T cells
tion and activation118. Some human vaccinations with Most studies of cancer immunotherapy so far have
TAA-expressing APCs induced TAA-specific regulatory focused on augmenting tumour immunity through
T cells87,88. In addition, it is possible that thymus-derived supplementing active immune elements such as den-
TReg cells might be activated by crossreactivity between dritic cells, TAA-specific T cells and cytokines. The
self-antigens and novel tumour antigens and might in concept of reversing immunosuppression in cancer
turn induce TAA specificity. has merit as a therapeutic approach22. Newer studies
targeting suppressive molecules and regulatory T cells
Is suppressor activity TAA specific or nonspecific? It is point the way to successful application of tumour
understandable that antigen (including self-antigen)- immunotherapy (FIG. 3).
specific TReg cells suppress antigen-specific T-cell responses.
Theoretically, self-antigen-specific TReg cells only sup- Blocking CTLA4 in vivo. CTLA4 blockade resulted in
press self-antigen (that is, tumour antigen)-specific improved tumour immunity and tumour regression
T-cell responses. In support of this, NY-ESO-1 (New York in several mouse models26,27,46–50,52. These observations
oesophageal squamous-cell carcinoma 1)-specific CD4+ led to interest in CTLA4 blockade for human cancer
T cells were induced from naive populations following immunotherapy. Two clinical trials, in which patients
CD4+CD25+ T-cell depletion, indicating that TReg cells with advanced cancer (TABLE 3) were treated with an
suppressed NY-ESO-1-specific CD4+ T-cell priming119. inhibitory human CTLA4-specific antibody, have been
Furthermore, ovarian-tumour TReg cells suppressed TAA- published53,54. Treatment with CTLA4-specific antibody
specific effector T-cell function in vitro and in vivo40. induced objective cancer regression in some patients
However, TReg cells isolated from TCR-transgenic mice that were vaccinated with HLA-A2-restricted peptides
inhibited responses of CD4+CD25– T cells to the same from the melanoma-associated antigen glycoprotein 100
antigen, but also inhibited a distinct antigen-specific (gp100) or with irradiated, autologous granulocyte/
T-cell response. Furthermore, TReg cells suppressed allo- macrophage colony-stimulating factor (GM-CSF)-
geneic T-cell activation in vitro and in vivo20. Therefore, secreting tumour cells53,54. Notably, CTLA4-specific anti-
although it remains controversial, it is possible that body treatment also resulted in severe, but manageable,
TReg-cell differentiation, expansion and activation are autoimmune responses in these patients53 (TABLE 3).
driven in an antigen-specific manner 67,101,109,118–121 that As discussed previously, the mechanistic link between
determines their antigenicity, whereas activated TReg cells the effects of treatment with CTLA4-specific antibody
show antigen-nonspecific suppressor activity 122. and TReg-cell function remains to be determined in these
human clinical trials. In fact, CTLA4-specific antibody
Do TReg cells suppress TAA-specific T-cell priming or did not induce depletion of peripheral TReg cells (or
effector function? TReg cells are homeostatically main- FOXP3 mRNA)55 or of peripheral CTLA4+ T cells in
tained in lymphoid organs20,23,67. Administration of tumour patients53. Nonetheless, in vivo administration
CD25-specific antibody, either 4 days before or 1 day of CTLA4-specific antibody breaks tumour tolerance in
after tumour inoculation, was efficient in promoting patients with cancer and regardless of the cellular source
tumour regression24,25,27,30,34,35. These data indicate that of CTLA4, these studies provide ‘proof of principle’ that
TReg cells mediate their suppressive effect by inhibiting blocking CTLA4 signals is an option for treating patients
T-cell priming in lymphoid organs. with cancer (FIG. 3).
By contrast, other evidence indicates that TReg cells
reduce the effector function of TAA-specific cells. Transfer Depletion of CD25+ T cells in vivo. Given that treat-
of TReg cells reduced the therapeutic efficiency of adop- ment with CD25-specific antibody efficiently depletes
tively transferred TAA-specific CD8+ effector T cells in TReg cells, improves tumour immunity and results in
a mouse model of melanoma42. Furthermore, depletion mouse tumour regression, it is predicted that depletion

302 | APRIL 2006 | VOLUME 6 www.nature.com/reviews/immunol

© 2006 Nature Publishing Group
 REVIEWS

RECIST criteria of TReg cells would be beneficial for cancer patients. lymphatic metastases were largely resolved. Based on the
A set of criteria, designated the Denileukin diftitox (Ontak) is a ligand–toxin fusion RECIST (response evaluation criteria in solid tumours) criteria,
response evaluation criteria in protein that consists of full-length IL-2 fused to the enzy- we observed an objective clinical response. However, our
solid tumours. matically active and translocating domains of diphtheria current observations indicate that more than four weekly
Objective clinical response
toxin123. This drug has been approved by the Food and infusions of denileukin diftitox might also deplete CD25+
Based on the International Drug Administration in the United States for treatment effector T cells37.
Union Against Cancer and of CD25+ cutaneous T-cell leukaemia and lymphoma. Another group has reported similar findings using
World Health Organization, an This fusion protein is internalized into CD25+ T cells denileukin diftitox in patients with renal tumours38.
objective clinical response to
by endocytosis. The ADP-ribosyltransferase activity Patients received a single dose of denileukin diftitox,
treatment is defined as a 50%
reduction in the sum of the
of diphtheria toxin is cleaved in the endosome and is followed by vaccination with DCs transfected with
products of the perpendicular translocated into the cytosol, where it inhibits protein total tumour RNA. They showed that peripheral blood
diameters of all lesions without synthesis, leading to apoptosis123. Denileukin diftitox TReg cells were eliminated in a dose-dependent manner.
the 25% growth of any lesion is therefore predicted to deplete TReg cells in patients Furthermore, denileukin diftitox administration fol-
or the appearance of new
lesions. RECIST defines an
with cancer. lowed by vaccination with tumour-RNA-transfected
objective clinical response We used denileukin diftitox to treat patients with DCs led to improved stimulation of tumour-specific
as a 30% reduction in the lung, ovarian or breast cancer, where TReg-cell numbers effector T cells compared with vaccination alone38.
sum of the maximum are greatly increased40,69. We demonstrated that a single It was also reported that different doses of deni-
diameters of lesions, along
dose of denileukin diftitox reduced the prevalence and leukin diftitox had variable effects on FOXP3 mRNA
with the appearance of no new
or progressive lesions.
absolute numbers of peripheral TReg cells and increased expression in CD4+ T cells in patients with metastatic
effector T-cell activation in all eight patients. This effect melanoma. Furthermore, there was no significant clini-
was evident for about one month. One patient with ovar- cal efficacy observed in this study39. Nonetheless, these
ian cancer had measurable disease, which permitted clinical trials suggest that in vivo depletion of functional
the assessment of potential clinical efficacy. CA125 is a human TReg cells is a possible option. Further clinical
protein that is often secreted into the blood by ovarian- study with large patient populations is essential to link
tumour cells and the amount of this protein in the blood TReg-cell depletion with improved immunity and potential
can provide information on the growth of ovarian can- clinical efficacy.
cer. The single denileukin diftitox infusion normalized Notably, depletion of TReg cells did not induce mean-
her blood CA125 levels, and after six additional weekly ingful tumour regression in some mouse models24,25,34
denileukin diftitox infusions, her bony, visceral and (TABLE 2). To reduce TReg-cell suppressor activity, when

Human tumour

Traditional Novel tumour Conventional

tumour therapy immunotherapy immunotherapy

• Surgical debulking Targeting regulatory T cells: • Tumour-associated antigens

• Radiation therapy • Depleting (Denileukin diftitox, CD25-specific antibody, cyclophosphamide) (tumour peptides)
• Chemotherapy • Blocking trafficking (CCL22-specific antibody) • APC vaccination
• Anti-angiogenic therapy • Blocking differentiation and signalling (blocking FOXP3 signal) (dendritic cells)
• Adoptive T-cell transfusion
Targeting suppressive molecules: (effector T cells)
• Blocking suppressive molecules (B7-H1, B7-H4, IDO, arginase) on APCs • Cytokine and/or chemokine
• Blocking suppressive molecules (CTLA4, PD1) on T cells administration
• Blocking soluble suppressive molecules (TGFβ, IL-10, VEGF, COX2) (IL-7, IL-15 and IL-21)

Combinatorial therapy

Figure 3 | Therapeutic targeting of suppressive mechanisms including regulatory T cells. After clinical and/or
pathological diagnosis, patients with cancer might be subjected to traditional tumour therapy, including surgical
debulking, radiation therapy, chemotherapy and antitumour angiogenic therapy. Depending on their clinical situation,
patients could receive a combination of these strategies. Traditional tumour therapy targets the tumour itself and remains
the ‘gold standard’ therapy. Conventional immunotherapy supplements the immune system and provides essential
immune elements, including tumour-associated antigen (TAA), antigen-presenting cells (APCs), effector T cells and
cytokines and/or chemokines with the aim of boosting TAA-specific immunity. Early clinical trials with conventional
tumour immunotherapy have been encouraging, but improvements in clinical efficacy are needed. Novel
immunotherapeutic strategies target the immunosuppressive network of tumours, including regulatory T cells,
suppressive molecules and dysfunctional APCs, with the aim of recovering TAA-specific immunity. To attain effective,
reliable and consistent clinical efficacy, it might be essential to combine traditional tumour therapy, conventional
immunotherapy and novel tumour immunotherapy. COX2, cyclooxygenase 2; CTLA4, cytotoxic T-lymphocyte-associated
antigen 4; FOXP3, forkhead box P3; IDO, indoleamine 2,3-dioxygenase; IL, interleukin; PD1, programmed cell death 1;
TGFβ, transforming growth factor-β; VEGF, vascular endothelial growth factor.

NATURE REVIEWS | IMMUNOLOGY VOLUME 6 | APRIL 2006 | 303

© 2006 Nature Publishing Group
REVIEWS

denileukin diftitox or other agents are considered for Immunotherapy and regulatory T cells
clinical trial, several possibilities need to be examined. The clinical efficacy of current methods of tumour
First, TReg-cell-mediated suppression might not be the immunotherapy and vaccination is not satisfactory. One
key tolerizing mechanism in certain tumour types or reason for this is that immunosuppressive mechanisms
at certain tumour stages, and at these times, other sup- predominate in patients with advanced cancer21,22,68,132.
pressive mechanisms might predominate22. Second, it is Another possibility is that certain immunotherapy and
important to examine the clinical efficacy of the route vaccination protocols might promote regulatory T-cell
of administration of TReg-cell-depletion agents. Systemic function121 and tumour trafficking instead of selectively
administration could result in insufficient concentra- improving the function of antitumour effector cells.
tions to deplete TReg cells in the tumour microenviron- Here, I use IL-2 treatment and APC vaccination as
ment. Furthermore, because many human tumours are examples to discuss this possibility.
not suitable for surgery, intratumoural administration
of TReg-cell-depleting agents could be an interesting IL-2 and TReg cells. IL-2 has been used to treat patients
alternative28,34. Third, the treatment might substantially with HIV and some cancers, such as myelogenous leu-
deplete other immune cells, including CD25+ effector kaemia, metastatic melanoma and renal-cell carcinoma127.
T cells. Therefore, although depletion of TReg cells is an Interestingly, in patients with HIV infection, the long-term
attractive option, careful studies are needed to optimize effect of IL-2 treatment is associated with decreased T-cell
treatment. These include dose escalation, identification proliferation and activation in vivo, which is accompanied
of the window of clinical benefit, the use of combinato- by an increase of CD4+CD25+ T-cell numbers133.
rial regimens and identification of which tumour types In patients with melanoma, renal cancer134, Ewing’s
best respond to such depletion (FIG. 3). sarcoma, alveolar rhabdomyosarcoma66 and ovarian
cancer (W.Z. et al., unpublished observations), IL-2
Chemotherapy. As discussed previously, low doses administration increased the number of peripheral
of cyclophosphamide reduces TReg-cell numbers and TReg cells. Furthermore, IL-2 stimulated CXC-chemokine
suppressive capacity in normal and tumour-bearing receptor 4 (CXCR4) and CCR4 expression on TReg cells
mice41,58,61–63,65. Other immunosuppressants, including and promoted their migration towards tumour micro-
calcineurin inhibitors, cyclosporine (Sandimmune) and environmental CXCL12 and CCL22 (W.Z. et al., unpub-
tacrolimus (FK506), might also reduce TReg-cell numbers lished observations)40,84,135. In line with these findings,
and function124,125. Cyclosporine and tacrolimus inhibit recent studies have shown that IL-2 contributes to
T-cell activation by suppressing IL-2 and are often used TReg-cell differentiation, expansion, maintenance and sup-
to prevent the rejection of transplanted organs. As IL-2 pressor activity in mice126–131. These data indicate that the
is crucial for TReg-cell expansion and function126–131, it clinical benefit of IL-2 treatment in patients with cancer
is predicted that cyclosporine and tacrolimus abrogate and AIDS requires urgent re-evaluation. IL-7, IL-15 and
TReg-cell function by altering IL-2 production and signals. IL-21 share the common cytokine-receptor γ-chain (γc)
Further pre-clinical studies are needed to assess whether and certain properties with IL-2, and have been shown
low doses of these agents are an option to promote to stimulate innate immunity and increase CD8+ T cell-
tumour immunity in patients (FIG. 3). mediated antitumour activity136–138. Furthermore, addition
of IL-7 with IL-15 abrogated the suppressive activity of
Other potential strategies to control TReg-cell function. TReg cells in vitro139. Therefore, IL-2 could be therapeuti-
The first potential strategy is one that could target cally replaced by other γc cytokines, and γc cytokines such
molecules that are involved in TReg-cell trafficking. In as IL-7, IL-15 and IL-21, alone or in combination, might
ovarian cancer, tumour microenvironmental CCL22- be useful as novel immunotherapeutic agents.
mediated trafficking of TReg cells into the tumour 40 and
blockade of CCL22 reduces TReg-cell tumour traffick- APC and TAA-expressing vector vaccination and regula-
ing 40. Therefore, altering TReg-cell tumour trafficking tory T cells. DCs have been used widely as an adjuvant
would be an attractive strategy for blocking TReg-cell to boost TAA-specific immunity in cancer patients.
suppressive capacity. A second possible strategy However, immature and semi-mature DCs induced
involves altering TReg-cell differentiation and func- regulatory T cells in mice140–142 and humans87,88,143, and
tion. FOXP3 is crucial for TReg-cell differentiation and their maturation status might determine their immuno-
function9–11, and compounds that inhibit the expres- genic versus tolergenic properties142. As the maturation
sion, function and signalling of FOXP3 might have status is defined by the cytokine profile142, it needs
therapeutic potential. IL-2, GITR and B7-H4 are also careful manipulation to obtain fully matured DCs with
implicated in TReg-cell differentiation or function and therapeutic efficacy. However, recent studies show that
targeting these molecules could also provide additional mature DCs activate and expand autoantigen-specific
therapeutic opportunities (FIG. 3). TReg cells, and in a mouse model of diabetes these cells
In summary, depletion of TReg cells, blockade of inhibited diabetes mediated by reactivity to multiple
TReg-cell trafficking, differentiation and function repre- antigens in mice90,91. Furthermore, mature DCs were
sent new ways to augment tumour immunity. This strat- shown to induce tumour-specific IL-10-expressing
egy might be used in combination with various current TR1 cells144. These studies indicate the possibility that
therapeutic approaches, including traditional tumour certain fully matured DCs promote regulatory T-cell
therapy and conventional immunotherapy (FIG. 3). function in vivo.

304 | APRIL 2006 | VOLUME 6 www.nature.com/reviews/immunol

© 2006 Nature Publishing Group
 REVIEWS

Additionally, the tumour antigen LAGE-1 was blocking trafficking into tumours, or reducing their dif-
identified as a ligand for TReg cells in patients with ferentiation and suppressive mechanisms, represent new
melanoma99, and this raises the question of whether vac- strategies for cancer treatment. However, although it has
cination with DCs or vector(s) bearing certain tumour been stated that “all anyone is talking about these days
antigens, for example LAGE-1, would promote TReg-cell is regulatory T cells”145, regulatory T cells might not be
clonal expansion and function in patients with cancer. the only or important suppressive mechanism for cer-
Therefore, additional studies are required to evaluate tain tumour stages and/or certain tumours. Therefore,
current and future vaccination trials. The ideal vaccina- it is essential to define the molecular basis and nature
tion needs to maximize the potential beneficial effects of regulatory T cells in each individual human tumour
while minimizing detrimental effects. microenvironment, including their antigen specificity,
induction and mechanisms of action. Therapeutically,
Concluding remarks agents that can selectively and efficiently target regula-
It is evident that tumours actively develop different tory T cells deserve extensive research. Furthermore, we
mechanisms to escape tumour immunity and defeat con- need to bear in mind that although targeting regulatory
ventional tumour immunotherapy22. Regulatory T cells T cells is an attractive option in treating human tumours,
have an important role in suppressing TAA-specific it is probable that to attain effective, reliable and con-
immunity. Regulatory T cells inhibit TAA-specific prim- sistent clinical efficacy, a complicated combinatorial
ing in tumour draining lymph nodes, and are further therapeutic regimen of traditional tumour therapy and
recruited into the tumour microenvironment, where conventional immunotherapy combined with novel
they suppress TAA-specific effector functions. Therefore, tumour immunotherapy, including targeting regulatory
manipulation of regulatory T cells, including depletion, T cell will be warranted (FIG. 3).

1. Shevach, E. M. Fatal attraction: tumours beckon 14. Cosmi, L. et al. Human CD8+CD25+ thymocytes share 26. van Elsas, A. et al. Elucidating the autoimmune
regulatory T cells. Nature Med. 10, 900–901 phenotypic and functional features with CD4+CD25+ and antitumour effector mechanisms of a treatment
(2004). regulatory thymocytes. Blood 102, 4107–4114 (2003). based on cytotoxic T lymphocyte antigen-4 blockade
2. Gershon, R. K. & Kondo, K. Infectious immunological 15. Chang, C. C. et al. Tolerization of dendritic cells by in combination with a B16 melanoma vaccine:
tolerance. Immunology 21, 903–914 (1971). TS cells: the crucial role of inhibitory receptors ILT3 comparison of prophylaxis and therapy. J. Exp. Med.
3. Gershon, R. K. & Kondo, K. Cell interactions in the and ILT4. Nature Immunol. 3, 237–243 (2002). 194, 481–489 (2001).
induction of tolerance: the role of thymic lymphocytes. 16. Rifa’i, M., Kawamoto, Y., Nakashima, I. & Suzuki, H. 27. Sutmuller, R. P. et al. Synergism of cytotoxic
Immunology 18, 723–737 (1970). Essential roles of CD8+CD122+ regulatory T cells in T lymphocyte-associated antigen 4 blockade and
4. Fujimoto, S., Greene, M. & Sehon, A. H. the maintenance of T cell homeostasis. J. Exp. Med. depletion of CD25+regulatory T cells in antitumour
Immunosuppressor T cells in tumour bearing host. 200, 1123–1134 (2004). therapy reveals alternative pathways for suppression
Immunol. Commun. 4, 201–217 (1975). 17. Zou, W. et al. Stromal-derived factor-1 in human of autoreactive cytotoxic T lymphocyte responses.
5. Berendt, M. J. & North, R. J. T-cell-mediated tumours recruits and alters the function of J. Exp. Med. 194, 823–832 (2001).
suppression of anti-tumour immunity. An explanation plasmacytoid precursor dendritic cells. Nature Med. 28. Yu, P. et al. Intratumour depletion of CD4+ cells
for progressive growth of an immunogenic tumour. 7, 1339–1346 (2001). unmasks tumour immunogenicity leading to the
J. Exp. Med. 151, 69–80 (1980). 18. Wei, S. et al. Plasmacytoid dendritic cells induce rejection of late-stage tumours. J. Exp. Med. 201,
6. Bursuker, I. & North, R. J. Generation and decay of the CD8+ regulatory T cells in human ovarian carcinoma. 779–791 (2005).
immune response to a progressive fibrosarcoma. Cancer Res. 65, 5020–5026 (2005). 29. Steitz, J., Bruck, J., Lenz, J., Knop, J. &
II. Failure to demonstrate postexcision immunity after 19. Gilliet, M. & Liu, Y. J. Generation of human CD8 Tuting, T. Depletion of CD25+CD4+ T cells
the onset of T cell-mediated suppression of immunity. T regulatory cells by CD40 ligand-activated and treatment with tyrosinase-related protein
J. Exp. Med. 159, 1312–1321 (1984). plasmacytoid dendritic cells. J. Exp. Med. 195, 2-transduced dendritic cells enhance the interferon
7. North, R. J. & Bursuker, I. Generation and decay of 695–704 (2002). α-induced, CD8+T-cell-dependent immune defense
the immune response to a progressive fibrosarcoma. References 17–19 report that human plasmacytoid of B16 melanoma. Cancer Res. 61, 8643–8646
I. Ly-1+2– suppressor T cells downregulate the dendritic cells can induce CD8+ regulatory T cells. (2001).
generation of Ly-1–2+ effector T cells. J. Exp. Med. 20. Shevach, E. M. CD4+ CD25+ suppressor T cells: 30. Jones, E. et al. Depletion of CD25+ regulatory cells
159, 1295–1311 (1984). more questions than answers. Nature Rev. Immunol. results in suppression of melanoma growth and
8. Sakaguchi, S., Sakaguchi, N., Asano, M., Itoh, M. & 2, 389–400 (2002). induction of autoreactivity in mice. Cancer Immun.
Toda, M. Immunologic self-tolerance maintained by An outstanding Review that summarizes recent 2, 1 (2002).
activated T cells expressing IL-2 receptor α-chains findings in the field of regulatory T cells and gives 31. Tanaka, H., Tanaka, J., Kjaergaard, J. & Shu, S.
(CD25). Breakdown of a single mechanism of self- an outlook for future direction. Depletion of CD4+ CD25+ regulatory cells augments
tolerance causes various autoimmune diseases. 21. Dunn, G. P., Old, L. J. & Schreiber, R. D. The the generation of specific immune T cells in tumour-
J. Immunol. 155, 1151–1164 (1995). immunobiology of cancer immunosurveillance and draining lymph nodes. J. Immunother. 25, 207–217
This paper proposed CD25 as a surface marker immunoediting. Immunity 21, 137–148 (2004). (2002).
of CD4+ regulatory T cells. This paper summarizes the most important 32. Nagai, H. et al. In vivo elimination of CD25+
9. Hori, S., Nomura, T. & Sakaguchi, S. Control of experimental information to refine and revisit regulatory T cells leads to tumour rejection of B16F10
regulatory T cell development by the transcription the concept of tumour immune surveillance. melanoma, when combined with interleukin-12 gene
factor Foxp3. Science 299, 1057–1061 (2003). 22. Zou, W. Immunosuppressive networks in the tumour transfer. Exp. Dermatol. 13, 613–620 (2004).
10. Fontenot, J. D., Gavin, M. A. & Rudensky, A. Y. environment and their therapeutic relevance. Nature 33. Prasad, S. J. et al. Dendritic cells loaded with
Foxp3 programs the development and function of Rev. Cancer 5, 263–274 (2005). stressed tumour cells elicit long-lasting protective
CD4+CD25+ regulatory T cells. Nature Immunol. 4, 23. Sakaguchi, S. Naturally arising Foxp3-expressing tumour immunity in mice depleted of CD4+CD25+
330–336 (2003). CD25+CD4+ regulatory T cells in immunological regulatory T cells. J. Immunol. 174, 90–98 (2005).
11. Khattri, R., Cox, T., Yasayko, S. A. & Ramsdell, F. An tolerance to self and non-self. Nature Immunol. 6, 34. Ko, K. et al. Treatment of advanced tumours with
essential role for Scurfin in CD4+CD25+ T regulatory 345–352 (2005). agonistic anti-GITR mAb and its effects on tumour-
cells. Nature Immunol. 4, 337–342 (2003). 24. Onizuka, S. et al. Tumour rejection by in vivo infiltrating Foxp3+CD25+CD4+ regulatory T cells.
References 9–11 provide evidence that FOXP3 administration of anti-CD25 (interleukin-2 receptor α) J. Exp. Med. 202, 885–891 (2005).
is crucial for regulatory T-cell development and monoclonal antibody. Cancer Res. 59, 3128–3133 35. Golgher, D., Jones, E., Powrie, F., Elliott, T. &
function. (1999). Gallimore, A. Depletion of CD25+ regulatory cells
12. Groux, H. et al. A CD4+ T-cell subset inhibits antigen- 25. Shimizu, J., Yamazaki, S. & Sakaguchi, S. Induction uncovers immune responses to shared murine tumour
specific T-cell responses and prevents colitis. Nature of tumour immunity by removing CD25+CD4+ rejection antigens. Eur. J. Immunol. 32, 3267–3275
389, 737–742 (1997). T cells: a common basis between tumour immunity (2002).
The first study showing that functional regulatory and autoimmunity. J. Immunol. 163, 5211–5218 36. Casares, N. et al. CD4+/CD25+ regulatory cells
T cells can be induced. (1999). inhibit activation of tumour-primed CD4+ T cells
13. Weiner, H. L. Induction and mechanism of action References 24 and 25 are the first reports that with IFN-γ-dependent antiangiogenic activity,
of transforming growth factor-β-secreting Th3 depletion of CD25+ cells, probably CD4+CD25+ as well as long-lasting tumour immunity elicited by
regulatory cells. Immunol. Rev. 182, 207–214 T cells, improves or promotes tumour immunity peptide vaccination. J. Immunol. 171, 5931–5939
(2001). in mice. (2003).

NATURE REVIEWS | IMMUNOLOGY VOLUME 6 | APRIL 2006 | 305

© 2006 Nature Publishing Group
REVIEWS

37. Barnett, B., Kryczek, I., Cheng, P., Zou, W. & regression as well as severe, but manageable, 74. Sasada, T., Kimura, M., Yoshida, Y., Kanai, M. &
Curiel, T. J. Regulatory T cells in ovarian cancer: autoimmune diseases in patients with tumours. Takabayashi, A. CD4+CD25+ regulatory T cells
biology and therapeutic potential. Am. J. Reprod. 55. Maker, A. V., Attia, P. & Rosenberg, S. A. Analysis of in patients with gastrointestinal malignancies:
Immunol. 54, 369–377 (2005). the cellular mechanism of antitumour responses and possible involvement of regulatory T cells in disease
38. Dannull, J. et al. Enhancement of vaccine-mediated autoimmunity in patients treated with CTLA-4 progression. Cancer 98, 1089–1099 (2003).
antitumour immunity in cancer patients after blockade. J. Immunol. 175, 7746–7754 (2005). 75. Ormandy, L. A. et al. Increased populations of
depletion of regulatory T cells. J. Clin. Invest. 115, 56. Tang, Q. et al. Distinct roles of CTLA-4 and TGF-β regulatory T cells in peripheral blood of patients
3623–3633 (2005). in CD4+CD25+ regulatory T cell function. with hepatocellular carcinoma. Cancer Res. 65,
39. Attia, P., Maker, A. V., Haworth, L. R., Eur. J. Immunol. 34, 2996–3005 (2004). 2457–2464 (2005).
Rogers-Freezer, L. & Rosenberg, S. A. Inability 57. Rollinghoff, M., Starzinski-Powitz, A., 76. Karube, K. et al. Expression of FoxP3, a key
of a fusion protein of IL-2 and diphtheria toxin Pfizenmaier, K. & Wagner, H. Cyclophosphamide- molecule in CD4CD25 regulatory T cells, in adult
(Denileukin Diftitox, DAB389IL-2, ONTAK) to sensitive T lymphocytes suppress the in vivo T-cell leukaemia/lymphoma cells. Br. J. Haematol.
eliminate regulatory T lymphocytes in patients with generation of antigen-specific cytotoxic T lymphocytes. 126, 81–84 (2004).
melanoma. J. Immunother. 28, 582–592 (2005). J. Exp. Med. 145, 455–459 (1977). 77. Marshall, N. A. et al. Immunosuppressive regulatory
40. Curiel, T. J. et al. Specific recruitment of regulatory 58. Glaser, M. Augmentation of specific immune response T cells are abundant in the reactive lymphocytes
T cells in ovarian carcinoma fosters immune privilege against a syngeneic SV40-induced sarcoma in mice by of Hodgkin lymphoma. Blood 103, 1755–1762
and predicts reduced survival. Nature Med. 10, depletion of suppressor T cells with cyclophosphamide. (2004).
942–949 (2004). Cell Immunol. 48, 339–345 (1979). 78. Viguier, M. et al. Foxp3 expressing CD4+CD25high
41. Turk, M. J., Guevara-Patino, J. A., Rizzuto, G. A., 59. Yoshida, S., Nomoto, K., Himeno, K. & Takeya, K. regulatory T cells are overrepresented in human
Engelhorn, M. E. & Houghton, A. N. Concomitant Immune response to syngeneic or autologous metastatic melanoma lymph nodes and inhibit the
tumour immunity to a poorly immunogenic melanoma testicular cells in mice. I. Augmented delayed footpad function of infiltrating T cells. J. Immunol. 173,
is prevented by regulatory T cells. J. Exp. Med. 200, reaction in cyclophosphamide-treated mice. Clin. Exp. 1444–1453 (2004).
771–782 (2004). Immunol. 38, 211–217 (1979). 79. Gray, C. P., Arosio, P. & Hersey, P. Association of
42. Antony, P. A. et al. CD8+ T cell immunity against a 60. Berd, D. & Mastrangelo, M. J. Effect of low dose increased levels of heavy-chain ferritin with increased
tumour/self-antigen is augmented by CD4+ T helper cyclophosphamide on the immune system of cancer CD4+ CD25+ regulatory T-cell levels in patients
cells and hindered by naturally occurring T regulatory patients: depletion of CD4+, 2H4+ suppressor-inducer with melanoma. Clin. Cancer Res. 9, 2551–2559
cells. J. Immunol. 174, 2591–601 (2005). T-cells. Cancer Res. 48, 1671–1675 (1988). (2003).
References 40, 41 and 42 provide a direct 61. Awwad, M. & North, R. J. Cyclophosphamide- 80. Viglietta, V., Baecher-Allan, C., Weiner, H. L. &
functional link between CD4+CD25+ T cells and induced immunologically mediated regression Hafler, D. A. Loss of functional suppression by CD4+
tumour immunopathogenesis in tumour patients of a cyclophosphamide-resistant murine tumour: CD25+ regulatory T cells in patients with multiple
and in tumour-bearing mice, respectively. a consequence of eliminating precursor L3T4+ sclerosis. J. Exp. Med. 199, 971–979 (2004).
43. Shimizu, J., Yamazaki, S., Takahashi, T., Ishida, Y. & suppressor T-cells. Cancer Res. 49, 1649–1654 81. Kriegel, M. A. et al. Defective suppressor function of
Sakaguchi, S. Stimulation of CD25+CD4+ regulatory (1989). human CD4+ CD25+ regulatory T cells in autoimmune
T cells through GITR breaks immunological 62. Awwad, M. & North, R. J. Cyclophosphamide (Cy)- polyglandular syndrome type II. J. Exp. Med. 199,
self-tolerance. Nature Immunol. 3, 135–142 facilitated adoptive immunotherapy of a Cy-resistant 1285–1291 (2004).
(2002). tumour. Evidence that Cy permits the expression of 82. Ehrenstein, M. R. et al. Compromised function of
44. Read, S., Malmstrom, V. & Powrie, F. Cytotoxic adoptive T-cell mediated immunity by removing regulatory T cells in rheumatoid arthritis and reversal
T lymphocyte-associated antigen 4 plays an essential suppressor T cells rather than by reducing tumour by anti-TNFα therapy. J. Exp. Med. 200, 277–285
role in the function of CD25+CD4+ regulatory cells burden. Immunology 65, 87–92 (1988). (2004).
that control intestinal inflammation. J. Exp. Med. 63. Lutsiak, M. E. et al. Inhibition of CD4+25+ 83. Fontenot, J. D. et al. Regulatory T cell lineage
192, 295–302 (2000). T regulatory cell function implicated in enhanced specification by the forkhead transcription factor
45. Takahashi, T. et al. Immunologic self-tolerance immune response by low-dose cyclophosphamide. foxp3. Immunity 22, 329–341 (2005).
maintained by CD25+CD4+ regulatory T cells Blood 105, 2862–2868 (2005). 84. Zou, L. et al. Bone marrow is a reservoir for CD4+
constitutively expressing cytotoxic T lymphocyte- 64. Ercolini, A. M. et al. Recruitment of latent pools CD25+ regulatory T cells that traffic through CXCL12/
associated antigen 4. J. Exp. Med. 192, 303–310 of high-avidity CD8+T cells to the antitumour CXCR4 signals. Cancer Res. 64, 8451–8455 (2004).
(2000). immune response. J. Exp. Med. 201, 1591–1602 85. Gabrilovich, D. Mechanisms and functional
46. Leach, D. R., Krummel, M. F. & Allison, J. P. (2005). significance of tumour-induced dendritic-cell defects.
Enhancement of antitumour immunity by CTLA-4 65. Ghiringhelli, F. et al. CD4+CD25+ regulatory T cells Nature Rev. Immunol. 4, 941–952 (2004).
blockade. Science 271, 1734–1736 (1996). suppress tumour immunity but are sensitive to 86. Jonuleit, H., Schmitt, E., Schuler, G., Knop, J. &
47. Kwon, E. D. et al. Manipulation of T cell co-stimulatory cyclophosphamide which allows immunotherapy of Enk, A. H. Induction of interleukin 10-producing,
and inhibitory signals for immunotherapy of prostate established tumours to be curative. Eur. J. Immunol. nonproliferating CD4+ T cells with regulatory
cancer. Proc. Natl Acad. Sci. USA 94, 8099–8103 34, 336–344 (2004). properties by repetitive stimulation with allogeneic
(1997). 66. Zhang, H. et al. Lymphopenia and interleukin-2 immature human dendritic cells. J. Exp. Med. 192,
48. Kwon, E. D. et al. Elimination of residual metastatic therapy alter homeostasis of CD4+CD25+ 1213–1222 (2000).
prostate cancer after surgery and adjunctive cytotoxic regulatory T cells. Nature Med. 11, 1238–1243 87. Dhodapkar, M. V., Steinman, R. M., Krasovsky, J.,
T lymphocyte-associated antigen 4 (CTLA-4) blockade (2005). Munz, C. & Bhardwaj, N. Antigen-specific inhibition
immunotherapy. Proc. Natl Acad. Sci. USA 96, 67. Von Herrath, M. G. & Harrison, L. C. Regulatory of effector T cell function in humans after injection of
15074–15079 (1999). Lymphocytes: antigen-induced regulatory T cells in immature dendritic cells. J. Exp. Med. 193, 233–238
49. Hurwitz, A. A. et al. Combination immunotherapy of autoimmunity. Nature Rev. Immunol. 3, 223–232 (2001).
primary prostate cancer in a transgenic mouse model (2003). 88. Chakraborty, N. G., Chattopadhyay, S., Mehrotra, S.,
using CTLA-4 blockade. Cancer Res. 60, 2444–2448 68. Khong, H. T. & Restifo, N. P. Natural selection Chhabra, A. & Mukherji, B. Regulatory T-cell response
(2000). of tumour variants in the generation of ‘tumour and tumour vaccine-induced cytotoxic T lymphocytes
50. Hurwitz, A. A., Yu, T. F., Leach, D. R. & Allison, J. P. escape’ phenotypes. Nature Immunol. 3, 999–1005 in human melanoma. Hum. Immunol. 65, 794–802
CTLA-4 blockade synergizes with tumour-derived (2002). (2004).
granulocyte-macrophage colony-stimulating factor for An outstanding review of tumour immune-evasion References 86–88 and 143 provide evidence that
treatment of an experimental mammary carcinoma. mechanisms. APCs can induce regulatory T cells in humans.
Proc. Natl Acad. Sci. USA 95, 10067–10071 (1998). 69. Woo, E. Y. et al. Regulatory CD4+CD25+ T cells in 89. Ghiringhelli, F. et al. Tumour cells convert immature
51. van Elsas, A., Hurwitz, A. A. & Allison, J. P. tumours from patients with early- stage non-small cell myeloid dendritic cells into TGF-β-secreting cells
Combination immunotherapy of B16 melanoma using lung cancer and late-stage ovarian cancer. Cancer Res. inducing CD4+CD25+ regulatory T cell proliferation.
anti-cytotoxic T lymphocyte-associated antigen 4 61, 4766–4772 (2001). J. Exp. Med. 202, 919–929 (2005).
(CTLA-4) and granulocyte/macrophage colony- The first demonstration that CD4+CD25+ T cells 90. Yamazaki, S. et al. Direct expansion of functional
stimulating factor (GM-CSF)-producing vaccines are increased in patients with tumours and can CD4+CD25+ regulatory T cells by antigen-processing
induces rejection of subcutaneous and metastatic suppress T-cell activation in vitro. dendritic cells. J. Exp. Med. 198, 235–247 (2003).
tumours accompanied by autoimmune 70. Liyanage, U. K. et al. Prevalence of regulatory 91. Tarbell, K. V., Yamazaki, S., Olson, K., Toy, P. &
depigmentation. J. Exp. Med. 190, 355–366 (1999). T cells is increased in peripheral blood and tumour Steinman, R. M. CD25+ CD4+ T cells, expanded
52. Espenschied, J. et al. CTLA-4 blockade enhances the microenvironment of patients with pancreas or with dendritic cells presenting a single autoantigenic
therapeutic effect of an attenuated poxvirus vaccine breast adenocarcinoma. J. Immunol. 169, peptide, suppress autoimmune diabetes. J. Exp. Med.
targeting p53 in an established murine tumour model. 2756–2761 (2002). 199, 1467–1477 (2004).
J. Immunol. 170, 3401–3407 (2003). 71. Somasundaram, R. et al. Inhibition of cytolytic 92. Chen, W. et al. Conversion of peripheral CD4+CD25–
53. Phan, G. Q. et al. Cancer regression and T lymphocyte proliferation by autologous CD4+/ naive T cells to CD4+CD25+ regulatory T cells
autoimmunity induced by cytotoxic T lymphocyte- CD25+ regulatory T cells in a colorectal carcinoma by TGF-β induction of transcription factor Foxp3.
associated antigen 4 blockade in patients with patient is mediated by transforming growth factor-β. J. Exp. Med. 198, 1875–1886 (2003).
metastatic melanoma. Proc. Natl Acad. Sci. USA Cancer Res. 62, 5267–5272 (2002). 93. Curotto de Lafaille, M. A., Lino, A. C.,
100, 8372–8377 (2003). 72. Wolf, A. M. et al. Increase of regulatory T cells in the Kutchukhidze, N. & Lafaille, J. J. CD25– T cells
54. Hodi, F. S. et al. Biologic activity of cytotoxic peripheral blood of cancer patients. Clin. Cancer Res. generate CD25+Foxp3+ regulatory T cells by
T lymphocyte-associated antigen 4 antibody blockade 9, 606–612 (2003). peripheral expansion. J. Immunol. 173, 7259–7268
in previously vaccinated metastatic melanoma and 73. Ichihara, F. et al. Increased populations of (2004).
ovarian carcinoma patients. Proc. Natl Acad. Sci. USA regulatory T cells in peripheral blood and tumour- 94. Fantini, M. C. et al. Cutting edge: TGF-β induces a
100, 4712–4717 (2003). infiltrating lymphocytes in patients with gastric and regulatory phenotype in CD4+CD25– T cells through
References 53 and 54 report clinical trials that esophageal cancers. Clin. Cancer Res. 9, 4404–4408 Foxp3 induction and downregulation of Smad7.
show that CTLA4 blockade can induce tumour (2003). J. Immunol. 172, 5149–5153 (2004).

306 | APRIL 2006 | VOLUME 6 www.nature.com/reviews/immunol

© 2006 Nature Publishing Group
 REVIEWS

95. Liang, S. et al. Conversion of CD4+ CD25– cells into 113. Mellor, A. L. & Munn, D. H. IDO expression by 132. Finn, O. J. Cancer vaccines: between the idea and the
CD4+ CD25+ regulatory T cells in vivo requires B7 dendritic cells: tolerance and tryptophan catabolism. reality. Nature Rev. Immunol. 3, 630–641 (2003).
co-stimulation, but not the thymus. J. Exp. Med. 201, Nature Rev. Immunol. 4, 762–774 (2004). An outstanding and comprehensive review of the
127–137 (2005). 114. Fallarino, F. et al. Modulation of tryptophan development of vaccination and immunotherapy for
96. Wan, Y. Y. & Flavell, R. A. Identifying Foxp3-expressing catabolism by regulatory T cells. Nature Immunol. 4, cancer. It proposes that cancer vaccines must
suppressor T cells with a bicistronic reporter. Proc. 1206–1212 (2003). overcome immune suppression.
Natl Acad. Sci. USA 102, 5126–5131 (2005). 115. Sica, G. L. et al. B7-H4, a molecule of the B7 family, 133. Sereti, I. et al. IL-2-induced CD4+ T-cell expansion in
97. Seo, N., Hayakawa, S., Takigawa, M. & Tokura, Y. negatively regulates T cell immunity. Immunity 18, HIV-infected patients is associated with long-term
Interleukin-10 expressed at early tumour sites induces 849–861 (2003). decreases in T-cell proliferation. Blood 104, 775–780
subsequent generation of CD4+T-regulatory cells and 116. Zang, X. et al. B7x: a widely expressed B7 family (2004).
systemic collapse of antitumour immunity. member that inhibits T cell activation. Proc. Natl 134. Ahmadzadeh, M. & Rosenberg, S. A. IL-2
Immunology 103, 449–457 (2001). Acad. Sci. USA 100, 10388–10392 (2003). administration increases CD4+CD25hiFoxp3+
98. Moller, G. Do suppressor T cells exist? Scand J. 117. Prasad, D. V., Richards, S., Mai, X. M. & Dong, C. regulatory T cells in cancer patients. Blood 107,
Immunol. 27, 247–250 (1988). B7S1, a novel B7 family member that negatively 2409–2414 (2006).
99. Wang, H. Y. et al. Tumour-specific human CD4+ regulates T cell activation. Immunity 18, 863–873 135. Kryczek, I. et al. CXCL12 and vascular endothelial
regulatory T cells and their ligands: implications for (2003). growth factor synergistically induce neoangiogenesis
immunotherapy. Immunity 20, 107–118 (2004). 118. Nishikawa, H. et al. Definition of target antigens for in human ovarian cancers. Cancer Res. 65, 465–472
The first study demonstrating TAA-specific ligand naturally occurring CD4+CD25+ regulatory T cells. (2005).
for CD4+CD25+ T cells in human cancer. J. Exp. Med. 201, 681–686 (2005). 136. Klebanoff, C. A. et al. IL-15 enhances the in vivo
100. Hawrylowicz, C. M. & O’Garra, A. Potential role of 119. Nishikawa, H., Jager, E., Ritter, G., Old, L. J. & antitumour activity of tumour-reactive CD8+ T cells.
interleukin-10-secreting regulatory T cells in allergy and Gnjatic, S. CD4+ CD25+ regulatory T cells control Proc. Natl Acad. Sci. USA 101, 1969–1974 (2004).
asthma. Nature Rev. Immunol. 5, 271–283 (2005). the induction of antigen-specific CD4+ helper T cell 137. Zeng, R. et al. Synergy of IL-21 and IL-15 in regulating
101. Hsieh, C. S. et al. Recognition of the peripheral responses in cancer patients. Blood 106, 1008–1011 CD8+ T cell expansion and function. J. Exp. Med.
self by naturally arising CD25+ CD4+ T cell receptors. (2005). 201, 139–148 (2005).
Immunity 21, 267–277 (2004). 120. Apostolou, I., Sarukhan, A., Klein, L. & 138. Melchionda, F. et al. Adjuvant IL-7 or IL-15 overcomes
102. Zhou, G., Lu, Z., McCadden, J. D., Levitsky, H. I. & von Boehmer, H. Origin of regulatory T cells with immunodominance and improves survival of the CD8+
Marson, A. L. Reciprocal changes in tumour known specificity for antigen. Nature Immunol. 3, memory cell pool. J. Clin. Invest. 115, 1177–1187
antigenicity and antigen-specific T cell function during 756–763 (2002). (2005).
tumour progression. J. Exp. Med. 200, 1581–1592 121. Zhou, G., Drake, C. G. & Levitsky, H. I. Amplification 139. Ruprecht, C. R. et al. Co-expression of CD25
(2004). of tumour-specific regulatory T cells following and CD27 identifies FoxP3+ regulatory T cells in
103. Ghiringhelli, F. et al. CD4+CD25+ regulatory T cells therapeutic cancer vaccines. Blood 107, 628–636 inflamed synovia. J. Exp. Med. 201, 1793–1803
inhibit natural killer cell functions in a transforming (2006). (2005).
growth factor-β-dependent manner. J. Exp. Med. 202, 122. Thornton, A. M. & Shevach, E. M. Suppressor effector 140. Verginis, P., Li, H. S. & Carayanniotis, G. Tolerogenic
1075–1085 (2005). function of CD4+CD25+ immunoregulatory T cells is semimature dendritic cells suppress experimental
The first in vivo study demonstrating a role of antigen nonspecific. J. Immunol. 164, 183–190 autoimmune thyroiditis by activation of thyroglobulin-
CD4+CD25+ T cells in blunting the NK-cell arm of (2000). specific CD4+CD25+ T cells. J. Immunol. 174,
the innate immune system in tumour immunity. 123. Foss, F. M. DAB389 IL-2 (ONTAK): a novel fusion toxin 7433–7439 (2005).
104. Chen, M. L. et al. Regulatory T cells suppress tumour- therapy for lymphoma. Clin. Lymphoma 1, 110–116 141. Wakkach, A. et al. Characterization of dendritic cells
specific CD8 T cell cytotoxicity through TGF-β signals (2000). that induce tolerance and T regulatory 1 cell
in vivo. Proc. Natl Acad. Sci. USA 102, 419–424 124. Shibutani, S. et al. Effects of immunosuppressants differentiation in vivo. Immunity 18, 605–617 (2003).
(2005). on induction of regulatory cells after intratracheal 142. Lutz, M. B. & Schuler, G. Immature, semi-mature
105. Peng, Y., Laouar, Y., Li, M. O., Green, E. A. & delivery of alloantigen. Transplantation 79, 904–913 and fully mature dendritic cells: which signals induce
Flavell, R. A. TGF-β regulates in vivo expansion of (2005). tolerance or immunity? Trends Immunol. 23,
Foxp3-expressing CD4+CD25+ regulatory T cells 125. Kawai, M., Kitade, H., Mathieu, C., Waer, M. & 445–449 (2002).
responsible for protection against diabetes. Proc. Pirenne, J. Inhibitory and stimulatory effects of 143. Dhodapkar, M. V. & Steinman, R. M. Antigen-bearing
Natl Acad. Sci. USA 101, 4572–4577 (2004). cyclosporine A on the development of regulatory immature dendritic cells induce peptide-specific CD8+
106. Green, E. A., Gorelik, L., McGregor, C. M., Tran, E. H. T cells in vivo. Transplantation 79, 1073–1077 regulatory T cells in vivo in humans. Blood 100,
& Flavell, R. A. CD4+CD25+ T regulatory cells control (2005). 174–177 (2002).
anti-islet CD8+ T cells through TGF-β-TGF-β receptor 126. Furtado, G. C., Curotto de Lafaille, M. A., 144. Lundqvist, A., Palmborg, A., Pavlenko, M.,
interactions in type 1 diabetes. Proc. Natl Acad. Sci. Kutchukhidze, N. & Lafaille, J. J. Interleukin 2 Levitskaya, J. & Pisa, P. Mature dendritic cells
USA 100, 10878–10883 (2003). signalling is required for CD4+ regulatory T cell induce tumour-specific type 1 regulatory T cells.
107. Kryczek, I. et al. B7-H4 expression identifies a novel function. J. Exp. Med. 196, 851–857 (2002). J. Immunother. 28, 229–235 (2005).
suppressive macrophage population in human ovarian 127. Malek, T. R. & Bayer, A. L. Tolerance, not immunity, 145. Wickelgren, I. Immunology. Policing the immune
carcinoma. J. Exp. Med. (in the press). crucially depends on IL-2. Nature Rev. Immunol. 4, system. Science 306, 596–599 (2004).
108. Kryczek, I. et al. Induction of B7-H4 on antigen 665–674 (2004).
presenting cells through interleukin 10: novel 128. Bayer, A. L., Yu, A., Adeegbe, D. & Malek, T. R. Acknowledgements
suppressive mode for regulatory T cells. J. Immunol. Essential role for interleukin-2 for CD4+CD25+ I thank J. L. Wilson for discussing and shaping this manuscript,
(in the press). T regulatory cell development during the neonatal and my collaborators for their intellectual input and hard work.
109. von Boehmer, H. Mechanisms of suppression by period. J. Exp. Med. 201, 769–777 (2005). The work described in this Review was supported by grants
suppressor T cells. Nature Immunol. 6, 338–344 129. Setoguchi, R., Hori, S., Takahashi, T. & Sakaguchi, S. from the United States National Institutes of Health and the
(2005). Homeostatic maintenance of natural United States Department of Defense.
110. de la Rosa, M., Rutz, S., Dorninger, H. & Scheffold, A. Foxp3+CD25+CD4+ regulatory T cells by
Interleukin-2 is essential for CD4+CD25+ regulatory interleukin (IL)-2 and induction of autoimmune disease Competing interests statement
T cell function. Eur. J. Immunol. 34, 2480–2488 by IL-2 neutralization. J. Exp. Med. 201, 723–735 The author declares no competing financial interests.
(2004). (2005).
111. Grossman, W. J. et al. Human T regulatory cells can 130. Thornton, A. M., Donovan, E. E., Piccirillo, C. A. & DATABASES
use the perforin pathway to cause autologous target Shevach, E. M. Cutting edge: IL-2 is critically required The following terms in this article are linked online to:
cell death. Immunity 21, 589–601 (2004). for the in vitro activation of CD4+CD25+ T cell Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.
112. Gondek, D. C., Lu, L. F., Quezada, S. A., Sakaguchi, S. suppressor function. J. Immunol. 172, 6519–6523 fcgi?db=gene
& Noelle, R. J. Cutting edge: contact-mediated (2004). B7-H4 | CD25 | CTLA4 | FOXP3 | GITR | IDO | LAGE-1 |
suppression by CD4+CD25+ regulatory cells involves 131. Antony, P. A. & Restifo, N. P. CD4+CD25+ T regulatory NY-ESO 1 | TGFβ
a granzyme B-dependent, perforin-independent cells, immunotherapy of cancer, and interleukin-2. Access to this links box is available online.
mechanism. J. Immunol. 174, 1783–1786 (2005). J. Immunother. 28, 120–128 (2005).

NATURE REVIEWS | IMMUNOLOGY VOLUME 6 | APRIL 2006 | 307

© 2006 Nature Publishing Group