Você está na página 1de 14


Regulatory T-cell immunotherapy
for tolerance to self antigens and
alloantigens in humans
Maria-Grazia Roncarolo*‡ and Manuela Battaglia*§
Abstract | Substantial progress in understanding the biology of regulatory T cells and
their roles in health and disease has been achieved in the past 10 years. This has led to
an increasing interest in the possibility of using regulatory T cells as a biological therapy
to preserve and restore tolerance to self antigens and alloantigens. Immunotherapy by
the adoptive transfer of regulatory T cells may have several advantages over conventional
treatments. However, several hurdles have to be overcome before such a therapy can enter
clinical practice. This Review summarizes our current knowledge of regulatory T cells,
illustrates the ongoing regulatory T-cell-based clinical trials, analyses the strengths and
pitfalls of this new therapeutic approach, and highlights the future perspectives.

Regulatory T cells are fundamental in controlling vari- the scope of this Review. Here, we focus on TReg cells,
ous immune responses. Compelling data generated in being the only naturally occurring regulatory T-cell sub-
preclinical animal models indicate that adoptive transfer set described so far, and on TR1 cells, being the only
of regulatory T cells can prevent or cure several T-cell- inducible T-cell subset that is currently being used in
mediated diseases, including autoimmune diseases and the clinic. This does not exclude the possibility that
allograft rejection, by restoring immune tolerance to self other regulatory T-cell subsets may be effective in
antigens or alloantigens. Absence or defective function providing protection from immunopathology and be
of regulatory T cells has been correlated with autoim- suitable for immunotherapy in the near future.
munity in humans, whereas their presence has been There are several key differences between mouse
associated with tolerance. Several different regulatory and human regulatory T cells (TABLE 1) and this should
T-cell subsets have been described (reviewed in REF. 1). be kept in mind, as these differences may represent
The CD4+ regulatory T cells have been categorized into obstacles in transferring knowledge gained from mouse
two major subgroups based on their ontogeny: the natu- models to human diseases.
rally occurring forkhead box P3 (FOXP3)+CD4+CD25+ The possibility that regulatory T cells might be
regulatory T cells (referred to here as TReg cells), which used for the treatment of T-cell-mediated diseases has
develop in the thymus and are present in normal naive recently gained increasing momentum. The advan-
mice and healthy individuals from birth (BOX 1), and tages of the adoptive transfer of regulatory T cells over
the inducible regulatory T cells, which are generated conventional treatments are numerous. Some of these
*San Raffaele Telethon in the periphery under various tolerogenic condi- benefits include: the potential for antigen specificity
Institute for Gene Therapy tions. Many different subsets of inducible regulatory with the lack of general immunosuppression, the possi-
(HSR-TIGET), Via Olgettina 58, T cells have been reported. Among them, the T-cell bility of inducing ‘physiological’ long-lasting regulation
20132 Milan, Italy.

subset (known as T regulatory type 1 (TR1) cells) that in vivo, and the fact that regulatory T-cell-based immu-
Vita-Salute San Raffaele
University, 20132 Milan, Italy. produces high levels of interleukin-10 (IL-10) was notherapy could be a custom-made product, designed
San Raffaele Scientific originally described by our group2 (BOX 2). In addition, ad hoc for each patient, with very limited or absent
Institute, Immunology of inducible FOXP3+CD4+CD25+ human regulatory T cells side effects. However, it is still unclear to what extent
Diabetes Unit, can be generated in vitro from CD4+CD25– T cells in observations made in preclinical animal models will be
20132 Milan, Italy.
Correspondence to M.-G.R.
the presence of transforming growth factor-β (TGFβ)3. applicable to humans; which human regulatory T-cell
e-mail: m.roncarolo@hsr.it A detailed description of CD4+ regulatory T-cell subsets subset will be most appropriate for adoptive transfer;
doi:10.1038/nri2138 has been the subject of other reviews1,4 and is beyond which method should be used to safely and efficiently


© 2007 Nature Publishing Group

Box 1 | FOXP3+CD4+CD25+ naturally occurring regulatory T (TReg) cells hand, the absence of regulatory T cells and immuno-
pathology. Furthermore, the human genetic diseases
Mouse and human TReg cells are selected in the thymus and represent about 5–10% in which regulatory T-cell dysfunction occurs are
of total CD4+ T cells in the periphery. They are crucial for maintaining tolerance discussed.
by downregulating undesired immune responses to self and non-self antigens.
TReg cells are defined on the basis of constitutive expression of high levels of CD25,
Role of regulatory T cells in responses to self antigens.
the transcription factor FOXP3 (forkhead box P3)72, low or absent expression of
CD127 (REFS 49,50) and the inability to produce interleukin-2 (IL-2) and to proliferate Negative selection of developing autoreactive thymocytes
in vitro72. Human CD4+CD25– effector T cells lack detectable levels of FOXP3, is not a wholly efficient process and some T cells with
but FOXP3 mRNA and FOXP3 protein are upregulated after polyclonal stimulation73. high affinity for autoantigens can escape the deletion
Most in vitro studies indicate that TReg cells mediate suppression by an undefined process in the thymus and migrate to the periphery.
cell-contact-dependent mechanism. However, in vivo, suppression may be mediated Peripheral tolerance is required to suppress those auto-
by suppressive cytokines. TReg cells may facilitate the production of IL-10 and reactive T cells that escape thymic selection. Proliferative
transforming growth factor-β (TGFβ) by other T cells through ‘infectious tolerance’72. responses of human peripheral-blood mononuclear cells
Alternatively, under certain conditions, TReg cells might acquire the ability to produce (PBMCs) from healthy individuals to the self antigens
TGFβ and/or IL-10 (REF. 74).
heat-shock protein 60 (REF. 5), myelin oligodendrocyte
After activation through their T-cell receptor, TReg cells suppress proliferation of
glycoprotein (MOG)6, the type-1-diabetes-associated
CD4+ and CD8+ T cells by inhibiting IL2 transcription75. TReg cells can also suppress
other cell types, such as antigen-presenting cells. antigen glutamic-acid decarboxylase 65 (GAD65) and
IL-2 is crucial for the generation, expansion, survival and effector function of TReg cells the vitiligo-associated antigen tyrosinase7 are indeed
(reviewed in REF. 76). As TReg cells do not produce IL-2 themselves, their capacity to use significantly enhanced in vitro upon removal of natu-
IL-2 secreted by target T cells appears to be essential for their suppressive activity76. rally occurring TReg cells. Similarly, TR1 cells specific for
self MHC molecules8 and for pancreatic-islet-derived
peptides9 have been found in healthy individuals. These
expand and generate regulatory T cells ex vivo; which observations indicate that autoreactive T cells are present
human disease will be likely to benefit the most from and circulate in healthy subjects and that regulatory
regulatory T-cell transfer; and whether a combined T cells actively suppress their function.
therapy with other drugs is necessary. Some of these Also consistent with an important role for regulatory
questions will undoubtedly be addressed by the ongo- T cells in controlling autoreactivity is the demonstration
ing clinical trials with adoptive transfer of regulatory that a wide variety of autoimmune diseases are associ-
T cells, whereas others will need further preclinical ated with defects in regulatory T-cell function, raising
investigation and a better understanding of the basic the interesting possibility that this may be a common
biology of regulatory T cells. This Review provides an basis for the uncontrolled immune responses to self anti-
update on the most recent and exciting findings that gens10. TR1 cells have been isolated from the synovium of
might help to make regulatory T-cell therapy a feasible patients with rheumatoid arthritis, but they seem to be
approach for the cure of numerous human diseases. present in significantly lower numbers compared with
Important limiting financial and regulatory aspects are control individuals with non-autoimmune-mediated
also discussed. joint inflammation 11. Similarly, T Reg -cell defects in
peripheral blood from patients with multiple sclerosis,
Regulatory T cells control tolerance in humans type 1 diabetes, psoriasis, myasthenia gravis, rheumatoid
Several studies have shown that regulatory T cells have and juvenile idiopathic arthritis have been described
a crucial role in controlling tolerance to self antigens (reviewed in REF. 10). However, it should be noted
and alloantigens in humans. This section describes the that all these studies tracked the percentages of circu-
correlation between, on the one hand, the presence lating CD4+CD25+ T cells, which might also include
of regulatory T cells and tolerance and, on the other non-regulatory activated T cells.

Box 2 | T regulatory type 1 (TR1) cells

TR1 cells arise in the periphery after encounter with antigen in the presence of interleukin-10 (IL-10). The unique
cytokine production profile (IL-2low/–IL-4–IL-5+IL-10+TGFβ+) distinguishes TR1 cells from T helper 0 (TH0), TH1 and
TH2 cells2. To date, no specific cell-surface marker for TR1 cells has been identified. TR1 cells have a very low
proliferative capacity following activation in vitro through the T-cell receptor (TCR), in part due to autocrine
production of IL-10. However, their proliferative activity in vivo is unknown52.
TR1 cells regulate immune responses through the secretion of the immunosuppressive cytokines IL-10 and
transforming growth factor-β (TGFβ), and they suppress both naive and memory T-cell responses52 and downregulate
Negative selection
the expression of co-stimulatory molecules and pro-inflammatory cytokines by antigen-presenting cells (APCs).
The deletion of self-reactive
thymocytes in the thymus. Furthermore, TR1 cells favour the production of IgD, IgA and IgG by B cells52. Importantly, TR1 cells are inducible,
Thymocytes expressing T-cell antigen specific and need to be activated through their TCR to exert their suppressive functions. However, once
receptors that strongly activated, they mediate suppression in an antigen non-specific manner52.
recognize self peptide bound IL-10 is absolutely required for the differentiation and function of mouse TR1 cells. For human TR1-cell differentiation,
to self MHC molecules undergo IL-10 is necessary but probably not sufficient77,78. Other necessary factors include the presence of APCs, which provide,
apoptosis in response to the in addition to IL-10, other soluble and surface molecules that are crucial for TR1-cell generation. Many different
signalling generated by high- approaches have been explored for the induction of TR1 cells both ex vivo and in vivo52.
affinity binding.

586 | AUGUST 2007 | VOLUME 7 www.nature.com/reviews/immunol

© 2007 Nature Publishing Group

Table 1 | Differences between mouse and human regulatory T cells

Feature Mouse Human
+ +
CD4 CD25 naturally occurring regulatory T cells
Phenotype CD25hiCD127low/–FOXP3+ CD25hiCD127low/–FOXP3+
Anergic in vitro Anergic in vitro
IL-2-dependent in vivo proliferation ND
Granzyme-B expression on activation Granzyme-A expression on activation
FOXP3 One isoform79 Two isoforms73
Lack of FOXP3 expression by CD4+CD25– FOXP3 expression by CD4+CD25– T cells
T cells activated in vitro79 activated in vitro73
Deletion of Foxp3 in scurfy mice80 Mutations in FOXP3 in patients with IPEX25
Suppressive mechanism Cell–cell contact in vitro Cell–cell contact in vitro
Cytokine mediated in some in vivo models74 ND
Downregulation of Il2 transcription in Downregulation of IL2 transcription in target
target T cells75 T cells75
Apoptosis of effector T cells via a Killing of monocytes, DCs, CD4+ T cells,
granzyme-B-dependent81 and perforin- CD8+ T cells and B cells via a granzyme-A-
independent pathway82 and perforin-dependent pathway83
Suppression of innate immunity84 ND
TR1 cells
Phenotype IL-4–IL-5+IL-10+IFNγ –TGFβ+ IL-4–IL-5+IL-10+IFNγ lowTGFβ+
+ – low
CD4 CD25 CD45RB (REF. 69) CD4+CD25– (REF. 77)
Anergic in vitro2 Anergic in vitro2
FOXP3 Constitutively FOXP3– Constitutively FOXP3– (REF. 77)
ND Upregulation on activation in vitro
Differentiation in vitro Exogenous IL-10 (REF. 2) Exogenous IL-10 with IFNα and APCs78
Suppressive mechanism IL-10 and TGFβ mediated52 IL-10 and TGFβ mediated52
APC, antigen-presenting cell; DC, dendritic cell; FOXP3, forkhead box P3; IFN, interferon; IL, interleukin; IPEX, immunodysregulation,
polyendocrinopathy and enteropathy, X-linked syndrome; ND, not determined; R, receptor; TGFβ, transforming growth factor-β;
TR1 cell, T regulatory type 1 cell.

Role of regulatory T cells in responses to alloantigens. although studies in preclinical models have clearly
The role of human T Reg cells in the suppression of shown protection of GVHD after TReg-cell transfer
graft-versus-host disease (GVHD) after haematopoietic (see later), definitive clinical evidence that cell therapy
stem-cell transplantation (HSCT) has been investi- with TReg cells is effective in the cure or prevention of
gated in several studies by correlating the presence this disease is still lacking. Conversely, in solid-organ
of TReg cells with the incidence of GVHD, but these transplantation, the scenario is clearer. In patients trans-
studies have yielded conflicting results. A positive planted with lung18, liver19 or kidney grafts20, a positive
correlation between high numbers of circulating correlation between graft survival and the number of
Graft-versus-host disease FOXP3+CD4+CD25hi TReg cells and reduced GVHD has circulating TReg cells has been shown.
(GVHD). A frequent been found in some studies12–14, whereas in other stud- TR1 cells were originally described as an important
complication after allogeneic ies the outcome was the opposite15,16. A straightforward cellular component in maintaining peripheral tolerance
stem-cell transplantation
caused by the expansion of
comparison between all these studies is not feasible as in vivo in severe combined immunodeficient (SCID)
donor lymphocytes with helper patients were not homogeneously selected, different patients after allogeneic HLA-mismatched HSCT. CD4+
and cytotoxic reactivity against immunosuppressive therapies were administered, dif- host-reactive TR1-cell clones generated from transplant
host histocompatibility ferent sources of HSCs were used and analyses were recipients produced very high levels of IL-10 and IL-5 but
antigens. It can occur as two
carried out at different time points after transplantation. not IL-4 and IL-2 after alloantigen-specific stimulation
distinct syndromes: acute
GVHD (occurring within 100 Furthermore, an important limitation of these studies in vitro. The presence of these cells in vivo correlated with
days) and chronic GVHD is the analysis of peripheral-blood T cells, which might the absence of GVHD and with long-term graft tolerance
(occurring after 100 days). not mirror the situation in tissues. without the need for drug-induced immunosuppres-
Complete depletion of T cells An elegant study by Rieger and colleagues showed sion21. In addition, high levels of IL-10 produced spon-
from the transplant largely
eliminates GVHD but
that TReg cells are numerically deficient in the intestinal taneously by PBMCs from an immunocompetent host
significantly increases the risk mucosa, which is a preferential site of GVH reactivity, before bone-marrow transplantation (BMT) have been
of graft failure and infections. in patients with acute and chronic GVHD17. Overall, associated with a subsequent low incidence of GVHD


© 2007 Nature Publishing Group

and transplant-related mortality22. A high frequency feasible when the TReg cells are specific for the pathogenic
of donor cells producing IL-10 in response to host allo- antigen. Alternatively, transfer of polyclonal TReg cells
antigens was found to correlate with the absence of acute may cure ongoing disease only in lymphopenic hosts,
GVHD after BMT, whereas a low frequency of these cells in which the massive expansion of TReg cells may lead to
was strongly associated with severe GVHD23. We have the generation of a sufficient number of antigen-specific
also recently found increased frequencies of TR1-cell TReg cells. If this is the case also in humans, it may repre-
clones in patients with thalassaemia with a persistent sent a significant limitation for the clinical application
Mixed chimerism state of mixed chimerism after successful HSCT from HLA of TReg cells in autoimmune diseases, as, to date, human
A state of coexistence of the
identical or matched unrelated donors (G. Serafini and self-antigen-specific TReg cells have not been successfully
host and allogeneic donor
haematopoietic cells. M.-G.R., unpublished observations). CD4+ TR1 cells are expanded ex vivo. By contrast, TR1 cells can be generated
also present in patients who spontaneously developed in an antigen-specific manner ex vivo and/or directly
IPEX tolerance to a kidney or liver allograft24. Taken together, in vivo and therefore may represent an ideal therapeutic
(Immunodysregulation, these data indicate that TR1 cells can naturally regulate alternative. However, additional preclinical studies need
enteropathy, X-linked
immune responses and induce tolerance to alloantigens to be carried out to show that antigen-specific TR1 cells
syndrome). A disease caused in vivo in both BMT and solid-organ transplantation. can revert ongoing autoimmunity.
by mutations in FOXP3 These findings provide a strong rationale for the clinical
(forkhead box P3) and application of cell therapy with antigen-specific TR1 cells Adoptive transfer of regulatory T cells in animal models
characterized by refractory
generated ex vivo. of transplantation. The transfer of freshly isolated
enteritis and, in some patients,
autoimmune endocrinopathies, TReg cells together with the bone-marrow allograft has
autoimmune diabetes and Defective regulatory T cells in genetic diseases. IPEX been shown to ameliorate GVHD and facilitate engraft-
thyroiditis. Unlike scurfy mice, (immunodysregulation, polyendocrinopathy and entero- ment in mouse models of BMT32–34 (TABLE 2). GVHD
peripheral-blood mononuclear pathy, X-linked syndrome), WAS (Wiskott–Aldrich was also the first model in which it was shown that
cells from IPEX patients fail to
produce cytokines after in vitro
syndrome), APS type 2 (autoimmune polyglandular the adoptive transfer of donor TReg cells that were poly-
stimulation. syndrome type 2) and ALPS (autoimmune lympho- clonally expanded ex vivo is as efficient as the transfer
proliferative syndrome) are the few autoimmune diseases of freshly isolated TReg cells in curing disease that results
WAS caused by known genetic defects. Interestingly, patients from transplantation in mice35,36. Interestingly, contrary
(Wiskott–Aldrich syndrome).
with IPEX25, with WAS26 or with APS type 2 (REF. 27) have to what was observed in the experimental models of
A life-threatening X-linked
immunodeficiency caused by a clear defect in TReg-cell suppressive function. Patients autoimmunity, the transfer of TReg cells enriched for
mutation in the WAS protein. with ALPS type Ia are characterized by an expansion alloantigen specificity showed only moderately improved
It is characterized by of T cells, but a reduction in the CD4+CD25+ T-cell efficacy compared with the transfer of a polyclonal
thrombocytopaenia with subset, and this defect may be indicative of disturbed TReg-cell population37. These different results might be
small platelets, eczema,
recurrent infections caused
lymphocyte immunoregulation in ALPS28. Although ascribed to the higher frequency of alloantigen-specific
by immunodeficiency, and the molecular mechanisms underlying these defects in T cells compared with cells specific for self antigens
an increased incidence of TReg cells are still under scrutiny, these data suggest that within the TReg-cell subset. However, this hypothesis
autoimmune manifestations regulatory T-cell-based therapy might also represent a contrasts with the concept that TReg cells are enriched
and malignancies.
good immunomodulatory strategy for the treatment of for self-antigen specificity because they develop in the
APS these genetic diseases. thymus. An alternative explanation for the different
(Autoimmune polyglandular outcome in the experimental models of autoimmunity
syndrome; also known as Therapy with regulatory T cells in mice versus BMT might be the presence of a lymphopenic
APECED). APS type 1 is Adoptive transfer of regulatory T cells in animal models environment that supports the expansion of transferred
caused by the loss of central
tolerance due to mutations in
of autoimmune diseases. Several preclinical animal effector T cells and regulatory T cells in BMT recipients
autoimmune regulator (AIRE), studies have established that the adoptive transfer of and that it may be different to that in normal mice.
whereas APS type 2 is of regulatory T cells can prevent various autoimmune Thanks to these promising results generated in the ani-
unknown pathogenesis diseases (TABLE 2). As the development of autoimmunity mal models and to the lack of antigen-specific require-
besides an association with
in humans is rarely a foreseeable phenomenon, to be of ments for the transferred TReg cells, BMT is the setting
polymorphisms in CTLA4 and
an HLA-extended haplotype. use as therapeutic agents regulatory T cells must inhibit for the first human clinical trial with TReg cells generated
It is characterized by multiple ongoing T-cell responses and reverse established pathol- ex vivo, which will be carried out to test the ability of
endocrine diseases initiated ogy. However, among the numerous published studies TReg cells to suppress GVHD (see later).
by an autoimmune process. only three studies showed that TReg-cell transfer is effi- T R 1 cells generated ex vivo upon stimulation
cacious in reverting active disease29–31. Interestingly, of with alloantigens in the presence of IL-10 and TGFβ
(Autoimmune these three studies, only one used polyclonal wild-type have been shown to be potent regulators of GVH
lymphoproliferative TReg cells to show that reversal of disease can be medi- responses after allogeneic BMT. Infusion of unma-
syndrome). A syndrome ated by these cells29, whereas the other two studies used nipulated cultured T cells induced lethal GVHD in all
associated with diffuse
transgenic TReg cells expressing a T-cell receptor (TCR) transplant recipients, whereas only 25% of mice receiv-
autoimmune manifestations.
ALPS type Ia patients have specific for the pathogenic antigen30,31. Importantly, ing ex vivo generated TR1 cells died38. A clinical trial
mutations in TNFRSF6, which the positive results obtained with polyclonal wild-type with ex vivo generated alloantigen-specific TR1 cells in
encodes CD95; ALPS type Ib TReg cells29 were generated in lymphopenic hosts in BMT is ongoing at the San Raffaele Scientific Institute,
patients have mutations in which activation or expansion of TReg-cell subsets can be Milan, Italy (see later).
TNFSF6, which encodes CD95
ligand; ALPS type II patients
influenced by homeostatic proliferation. Overall these data In contrast to what has been reported in GVHD,
have mutations in CASP10, indicate that the suppression of an ongoing autoimmune to our knowledge there are no reports showing that
which encodes caspase-10. disease by the adoptive transfer of TReg cells might only be the transfer of freshly isolated TReg cells can prevent

588 | AUGUST 2007 | VOLUME 7 www.nature.com/reviews/immunol

© 2007 Nature Publishing Group

Table 2 | Adoptive transfer of regulatory T cells in experimental mouse models

Disease Regulatory T cells Mouse model Effect on disease Refs
Autoimmune disease
Type 1 diabetes BDC2.5 TCR-transgenic TReg cells NOD.Rag–/– mice reconstituted with Prevention 30
(from NOD mice) diabetogenic T cells
NOD.Cd28–/– mice (which lack TReg cells) Prevention 30
Diabetic NOD mice receiving syngeneic islets Prevention 30
NOD mice with new onset diabetes Remission (60%) 30
BDC2.5 TCR-transgenic TReg cells BDC2.5 TCR-transgenic mice treated Prevention 86
expanded by DCs in vitro with high doses of cyclophosphamide
NOD.SCID mice reconstituted with Prevention 86
diabetogenic T cells
Pre-diabetic NOD mice Prevention 86
NOD mice with new onset diabetes Remission (50%) 81
Antigen-specific NOD TReg cells NOD.Cd28 mice Prevention 87
expanded in vitro
GAD65-specific TR1 cells NOD.SCID mice reconstituted with Prevention 88,89
diabetogenic T cells
Multiple sclerosis TCR-transgenic MBP-specific Rag–/– TCR-transgenic (MBP-specific) mice Prevention of spontaneous 90
(EAE) TReg cells disease
TReg cells from naive C57BL/6 mice C57BL/6 mice immunized with MOG35–55 Prevention of induced 91
peptide disease
OVA-specific TR1 cells BALB/c mice immunized with mouse spinal- Prevention of induced 44
cord homogenate and with heat-killed disease
Mycobacterium tuberculosis
TR1 cells induced by B7H1– C57BL/6 mice immunized with MOG35–55 Prevention of induced 85
immunoglobulin fusion protein plus peptide disease
immobilized CD3-specific antibody
Rheumatoid arthritis TReg cells Collagen-induced arthritis Inhibited progression of 92
early stage disease
Inflammatory bowel TReg cells SCID mice reconstituted with CD45RBhi cells Reversal of established 29
disease disease
OVA-specific TR1 cells SCID mice reconstituted with CD45RBhi cells Prevention 2
Caecal-bacteria-specific TR1 cells SCID mice reconstituted with pathogenic Prevention 93
from C3H/HeJBir mice TH1 cells from C3H/HeJBir mice
Systemic lupus Thymus-derived TReg cells NZB × NZW mice Control of autoimmunity 94
Scurfy disease TReg cells FOXP3-deficient mice Rescue of the 95
syndrome in neonatal mice
GVHD TReg cells Mice having received an allogeneic BMT Inhibition of GVH lethality 96
TReg cells Mice having received an allogeneic BMT Delay and prevention 35
Donor TReg cells Mice having received an allogeneic BMT Inhibition of GVH lethality 36
Alloantigen-specific TR1 cells Mice having received an allogeneic BMT Inhibition of GVH lethality 38
Graft rejection TReg cells expanded in vivo Mice having received an allogeneic islet-cell Prevention of allograft 39
transplant rejection
TReg cells expanded ex vivo with Mice having received an allogeneic islet-cell Prevention of allograft 41
rapamycin transplant rejection
TR1 cells induced in vivo Mice having received an allogeneic islet-cell Prevention of allograft 42
transplant rejection
TCR-transgenic HA-specific Mice having received an allogeneic skin Prolonged survival of 40
TReg cells transplant (HA expressing) established graft
BMT, bone-marrow transplantation; DC, dendritic cell; EAE, experimental autoimmune encephalomyelitis; FOXP3, forkhead box P3; GAD65, glutamic-acid
decarboxylase 65; GVH, graft-versus-host; HA, haemagglutinin; MBP, myelin basic protein; MOG, myelin oligodendrocyte glycoprotein; NOD, non-obese diabetic;
NZB, New Zealand black mouse strain; NZW, New Zealand white mouse strain; OVA, ovalbumin; RAG, recombination activating gene; SCID, severe combined
immunodeficient; TCR, T-cell receptor; TR1 cell, T regulatory type 1 cell; TReg cell, regulatory T cell.


© 2007 Nature Publishing Group

Table 3 | Regulatory T-cell-based immunotherapy in humans

Regulatory T-cell subset Method Potential application
Pros Cons Selection Expansion/induction
in vitro
FOXP3+CD4+CD25+ regulatory T cells
• Circulating cells • Low numbers in the • Two-step procedure with • Beads coated with CD3- • GVHD (clinical
• Defined surface markers circulation magnetic beads to enrich and CD28-specific antibody trial ongoing),
• Proven efficacy in several • Multiple antigen specificity for CD4+CD25+ cells plus IL-2 and rapamycin autoimmunity, organ
preclinical models • Risk of pan • Further selection for low • Antigen specific (?) transplantation (?),
immunosuppression CD127 expression (?) IPEX (?)
T regulatory type 1 cells
• Inducible ex vivo • Hard to purify • Not feasible • IL-10 with or without IFNα • GVHD (clinical
• Antigen specific • IL-10-treated dendritic cells trial ongoing), organ
• Proven efficacy in several • Tolerogenic dendritic cells transplantation,
preclinical models autoimmunity
GVHD, graft-versus-host disease; IFNα, interferon-α; IL, interleukin; IPEX, immunodysregulation, polyendocrinopathy, enteropathy, X-linked syndrome.

the rejection of allogeneic solid-organ transplantation. Which regulatory T-cell subset should be used for
Several studies indicate that TReg cells generated in vivo adoptive immunotherapy? Several subsets of regulatory
in transplanted animals by various approaches (such T cells have been described so far and potentially each
as by treatment with vitamin D3 and mycophenolate of these subsets can be considered for clinical trials.
mofetil39) do transfer tolerance in secondary trans- Here, we focus on TReg cells and TR1 cells and indicate
plant recipients. However, the lack of data proving that their use has both advantages and disadvantages.
efficacy after transfer of freshly isolated T Reg cells TReg cells are circulating cells with a defined phenotype
might be due to unreported unsuccessful experiments (CD4+CD25+CD127low) that allows for their isolation and
and/or to the requirement of antigen-specific TReg cells purification. Furthermore, TReg cells have been tested in
for protection. Alloantigen-specific TReg cells may be several preclinical animal models and have been shown
absent in freshly isolated cells but may be generated to provide active and specific regulation in vivo. The
upon in vivo expansion in the presence of specific major hurdles in using TReg cells however is their low
alloantigens. Positive results generated with the trans- numbers in the circulation and their broad and still
Homeostatic proliferation
fer of antigen-specific transgenic TReg cells are con- poorly defined antigen specificity, as this may cause
Spontaneous proliferation sistent with this view40. By contrast, adoptive transfer pan immunosuppression when they are transferred
of T cells in lymphopenic of T Reg cells polyclonally expanded ex vivo with in vivo. Both these shortcomings could be overcome by
conditions, caused by chronic rapamycin, which blocks the proliferation of effector the strategies discussed below.
diseases or treatments such
T cells while sparing TReg cells, promotes tolerance to TR1 cells have the advantage, over TReg cells, of being
as thymectomy or irradiation.
Factors that support T-cell allogeneic pancreatic islet grafts41, suggesting that the inducible ex vivo and therefore cells of the desired anti-
homeostatic proliferation antigen specificity of TReg cells is dispensable in this gen specificity can be more easily generated. The major
include peptide–MHC–TCR model. At present, the lack of data clearly showing that limitation of this subset is that it is difficult to purify
interactions and cytokines transferred TReg cells protect from allograft rejection in TR1 cells owing to a lack of specific cell-surface markers.
such as interleukin-7 and
several preclinical animal models represents a major T-cell lines that are highly enriched for TR1 cells can be
hurdle to the use of TReg-cell-based immunotherapy in obtained but contamination with non-regulatory T cells
Rapamycin solid-organ transplantation. might represent a significant caveat. Furthermore, it is
An immunosuppressive We have shown that TR1 cells generated in vivo, in unclear whether the phenotype of TR1 cells generated
drug that, in contrast to
mice transplanted with allogeneic pancreatic islets and in vitro is stable once they are transferred in vivo in
calcineurin inhibitors (such
as cyclosporin A and FK506), treated with rapamycin and IL-10, transfer antigen- humans. Importantly, TR1 cells produce the immuno-
does not prevent T-cell specific tolerance to secondary transplant recipients42. suppressive cytokine IL-10 only after activation with
activation but blocks Antigen-specific TR1 cells can therefore be envisaged as their specific antigen, but once activated, they can
interleukin-2-mediated a possible approach for preventing allograft rejection in mediate some level of bystander suppression43. Bystander
clonal expansion by blocking
mTOR (mammalian target
humans. suppression would be limited to the site where TR1 cells
of rapamycin). It does not localize, as no pan immunosuppression has been
interfere with the function Therapy with regulatory T cells in humans observed in mouse models or patients with high num-
and expansion of naturally Despite the data generated in preclinical animal models bers of circulating TR1 cells21,42. Therefore, one could
occurring regulatory T cells.
that successfully show that regulatory T cells can prevent envisage generating TR1 cells specific for an antigen
Bystander suppression or cure several T-cell-mediated diseases, many questions that is not related to the disease, with the assumption
Suppression in which remain to be addressed for the translation of this approach that once TR1 cells are administered to the patients they
responses to a second, to the clinic. This section addresses which human regula- could be activated and suppress pathogenic T cells with
unrelated antigen can be tory T-cell subset should be used, which method for its different antigen specificity. This might represent an
inhibited when it is presented
together with the antigen
selection, expansion and induction should be adopted, interesting strategy, especially when the antigen caus-
towards which tolerance has and in which disease might regulatory T cells function ing the disease is unknown. However, the presence
been already established. successfully (TABLE 3). of an antigen capable of activating regulatory T cells

590 | AUGUST 2007 | VOLUME 7 www.nature.com/reviews/immunol

© 2007 Nature Publishing Group

is required, as shown by preclinical studies in which and are described elsewhere52. The only currently avail-
mice receiving ovalbumin-specific TR1 cells generated able GMP protocol to generate antigen-specific TR1 cells
ex vivo were cured from inflammatory bowel disease or for infusion after allogeneic BMT is the one described
experimental autoimmune encephalomyelitis (a mouse and used by our group. In this protocol, donor PBMCs
model of multiple sclerosis) only after administration of are cultured ex vivo with host PBMCs (depleted of CD3+
ovalbumin2,44. T cells and irradiated) in the presence of IL-10. After 10
days, donor T cells become anergic to the host antigens,
Which method should be used to select, expand and while preserving the ability to proliferate in response to
differentiate regulatory T cells? The isolation of human third party and recall antigens. Importantly, these IL-10-
TReg cells from standard leukapheresis products by using anergized T cells are enriched in alloantigen-specific
a two-step magnetic cell-separation protocol performed TR1-cell precursors that can be re-infused into the
under conditions of good manufacturing practice (GMP) patients at the time of donor-cell engraftment to prevent
has been shown to be feasible45. TReg cells represent only GVHD (R. Bacchetta and M.-G.R., unpublished obser-
5–10% of the peripheral-blood CD4+ T cells and there- vations). We are now in the process of developing a more
fore, once isolated, they need to be expanded without efficient GMP protocol in which monocyte-derived host
the loss of their regulatory properties. The reduced abil- DCs differentiated in the presence of IL-10 are used to
ity of TReg cells to proliferate in vitro can be reversed by stimulate donor CD4+ T cells. The resulting anergic
potent stimulation through the TCR in the presence of CD4+ T-cell lines contain 5–10% alloantigen-specific
high doses of IL-2. Such expanded TReg cells maintain TR1 cells and are highly suppressive in vitro (S. Gregori
expression of relevant lymph-node-homing receptors, and M.-G.R., unpublished observations).
such as CD62 ligand (CD62L) and CC-chemokine
receptor 7 (CCR7), and suppressive activity46. However, For which disease should cellular therapy be used?
these cell-culture conditions are also highly advanta- Allogeneic BMT (or HSCT) is the first clinical setting
geous for the expansion of effector T-cell populations, in which cellular therapy with TR1 cells has been applied
which can contaminate the purified TReg-cell popula- and in which expanded TReg cells will be tested, with
tion. In humans, even the CD4+CD25hi T-cell subset the aim of preventing acute GVHD. There are several
contains ~50% of in vivo recently activated T cells, which reasons why acute GVHD has been selected as the ini-
ultimately outgrow TReg cells after prolonged culturing tial clinical target for testing the feasibility and safety
in vitro47. We believe that, to this day, the risk of co- of regulatory T-cell immunotherapy. First, regulatory
expanding potentially harmful cells is the greatest concern T cells expanded from peripheral blood have a suffi-
regarding the clinical use of ex vivo expanded TReg cells. cient number of alloreactive T cells. Second, regulatory
We have shown that the addition of rapamycin to the T cells derived from a donor exhibit normal replicative
culture media significantly reduces the undesired expan- potential as they have not been exposed to immuno-
sion of effector T cells, as rapamycin selectively allows suppressive treatment. Third, regulatory T cells can
the proliferation of TReg cells but inhibits the proliferation be administered before or at the same time as effector
of effector T cells48. Alternatively, a more sophisticated T-cell activation induced by alloantigens. Fourth, acute
isolation procedure might limit contamination with GVHD has a predictable time of onset, which is usually
effector T cells; the expression of CD127 can, for exam- within 100 days after transplantation, and regulatory
ple, discriminate between CD25+ activated (CD127+) T-cell immunotherapy could therefore be timed to
and regulatory (CD127low/–) T cells49,50 and might prove prevent or cure active disease. Fifth, the severe lympho-
to be a useful marker for cell purification. penia of the host following pre-transplant conditioning
Although a polyclonal population of bona fide and post-transplant immunosuppressive treatment
Leukapheresis naturally occurring TReg cells can now be isolated and can favour the expansion of regulatory T cells, thereby
A laboratory procedure for expanded, one of the remaining questions is whether reducing the number of cells required for a therapeutic
separating high numbers of
their suppressive function would be antigen specific effect. However, some caveats temper the enthusiasm
leukocytes from peripheral
blood. in vivo. Growth of antigen-specific TReg cells could for testing regulatory T cells in the prevention of acute
be a preferable alternative to polyclonal expansion. GVHD. The alloresponse after transplantation is com-
Cytokine storm However, alloantigen-specific human TReg cells have plicated by an intense inflammatory reaction with a
A strong systemic immune been described in only one study, in which human ‘cytokine storm’ that might impede regulatory T-cell
response that results in the
release of more than 150
TReg-cell lines were generated by repetitive stimulation immunotherapy. Furthermore, most patients receive
inflammatory mediators of CD4+CD25+ T cells with autologous dendritic cells GVHD prophylaxis after transplantation and it is now
(cytokines, oxygen free (DCs) pulsed with HLA-derived allogeneic peptides51. well established that some immunosuppressive com-
radicals and coagulation Whether ex vivo expansion of human TReg cells specific pounds interfere with regulatory T-cell activity53. The
factors). Both pro-
for antigens other than alloantigens is possible is yet to choice of drugs is therefore critical to avoid any in vivo
inflammatory cytokines
(such as tumour-necrosis be determined. Realistically, on the basis of the cur- antagonism of regulatory T cells. Rapamycin is of sig-
factor, interleukin-1 (IL-1) and rently available data and technology, we believe that nificant interest in this regard as it preserves the func-
IL-6) and anti-inflammatory the use of polyclonal expanded T Reg cells is the only tion of adoptively transferred TReg cells in vivo54. One
cytokines (such as IL-10 and approach that can currently be proposed for imminent can therefore envisage replacing the standard immuno-
IL-1 receptor antagonist) are
elevated in the serum of
clinical trials. suppressive treatment of GVHD with rapamycin, or
patients experiencing a Several experimental protocols for the generation of presumably with any other non-calcineurin inhibitors,
cytokine storm. antigen-specific human TR1 cells have been published in conjunction with ex vivo expanded TReg cells.


© 2007 Nature Publishing Group

a Conditioning b Draining lymph node c Target organs

Gut Commensal
lumen bacteria Peptide– Donor Gut
class II T cell
epithelial Migration
cell Skin
Translocation mDC NO and
of microbial Upregulation HO1
products Expansion Cytotoxic
of MHC by activity
host cells Migration

IFNγ regulatory
T cell
IL-1 Liver

IL-10 activity
and Donor
TGFβ CD8+ Expansion
Neutrophil Peptide–
MHC class I T cell

NK cell
TR1 cell TReg cell Expansion

Figure 1 | Pathogenesis of graft-versus-host disease and control by regulatory T cells. The pathophysiology of acute
graft-versus-host disease (GVHD) can be envisaged as a three step process. a | The first step occurs before donor-cell
infusion. Prior to haematopoietic stem-cell transplantation (HSCT), the conditioning regimen (that is, irradiation and/or
chemotherapy) leads to damage and activation of host tissues, especially the intestinal mucosa. This allows the
translocation of microbial products, such as lipopolysaccharide, from the intestinal lumen to the circulation, which
stimulates the secretion of pro-inflammatory cytokines, such as interleukin-1 (IL-1) and tumour-necrosis factor (TNF),
from host tissues (particularly from macrophages). Activated macrophages produce chemokines that activate
neutrophils, which further increases inflammation. The release of these pro-inflammatory cytokines increases the
expression of MHC and adhesion molecules on host cells, enhancing their antigen-presenting capacity. b | The second
step is characterized by activation of donor T cells, which occurs mainly in the lymph nodes that drain GVHD target
organs, such as the gut, skin and liver. Donor T-cell activation results mainly in interferon-γ (IFNγ) production, which
further increases the expression of MHC and adhesion molecules, chemokines and CD95 on antigen-presenting cells
(APCs). This results in further increases in antigen presentation and recruitment and the expansion of host-specific
cytotoxic CD8+ and CD4+ T cells, and natural killer (NK) cells. c | In the final step, effector cells then migrate to the target
organs, where they mediate tissue injury that leads to multi-organ failure mediated mainly by the CD95–CD95 ligand
and the perforin–granzyme pathways. Regulatory T cells can intervene at different stages to control GVHD. It is likely
that regulatory T cells, like effector T cells, are first activated in the draining lymph nodes. Both naturally occurring
forkhead box P3 (FOXP3)+CD4+CD25+ regulatory T cells (TReg cells) and T regulatory type 1 (TR1) cells, although through
different mechanisms, block activation and expansion of effector T cells and/or modulate the functions of APCs,
monocytes, macrophages and neutrophils. Consequently, activation and proliferation of alloreactive donor effector
T cells is suppressed, resulting in diminished export from the lymph nodes to the target organs. In addition, TR1 cells,
through IL-10 and transforming growth factor-β (TGFβ) production, efficiently reduce the inflammatory state. IFNγ
produced in the presence of inducible regulatory T cells leads to the expression of inducible nitric-oxide synthase (iNOS)
and indoleamine 2,3-dioxygenase (IDO) with consequent production of NO and haem oxygenase 1 (HO1), which
modulate effector T-cell functions. iDC, immature DC; mDC, mature DC; TCR, T-cell receptor.

Alternatively, regulatory T-cell immunotherapy could Should studies carried out with the adoptive trans-
be carried out at later time points after transplantation fer of regulatory T cells to prevent GVHD prove to be
to control chronic GVHD. An obvious drawback to this, safe, systemic autoimmunity is likely to be the next
however, is that the infused regulatory T cells would have clinical setting in which the efficacy of regulatory T-cell
to modulate an ongoing allogeneic T-cell response. A ben- immunotherapy will be tested. As described, preclinical
efit, however, is that this application would circumvent the models have clearly shown that polyclonal regulatory
cytokine storm, which should be fully resolved in most of T cells can prevent autoimmune diseases, whereas only
the patients, and this would presumably provide a more self-antigen-specific regulatory T cells can cure active
permissive environment for regulatory T-cell function. autoimmunity. Therefore, antigen-specific regulatory

592 | AUGUST 2007 | VOLUME 7 www.nature.com/reviews/immunol

© 2007 Nature Publishing Group

T cells can hypothetically be generated for the treatment further supporting this hypothesis. After activation,
of autoimmune diseases in which the target antigen has TReg cells exert their suppressive function by inhibiting
been identified. Insulin and GAD65 have been recog- effector T-cell proliferation, cytokine production and
nized as crucial target antigens in type 1 diabetes, myelin migration. TReg cells also downmodulate DC function by
basic protein in multiple sclerosis, and type II collagen in preventing their maturation in a cell-contact-dependent
rheumatoid arthritis. In type 1 diabetes, several preclini- manner58. TReg cells can also influence monocyte and
cal models have shown that the transfer of regulatory macrophage function by inhibiting lipopolysaccharide-
T cells can block disease development. Individuals at induced monocyte survival through the CD95–CD95
risk of developing type 1 diabetes can now be efficiently ligand apoptotic pathway59. Alternatively, TReg cells restrain
recognized55 and autologous regulatory T cells could monocytes by reducing their activation state, and this
be collected or differentiated before disease onset and leads to reduced pro-inflammatory cytokine production
stored until required. Early intervention when residual and impaired APC function60. Interestingly, TReg cells also
β-cell islet function is still present, could also be possible inhibit neutrophil activity and promote their apoptosis61.
and may be beneficial. TR1 cells, by producing IL-10, induce anergy in effec-
Solid-organ transplantation might also benefit from a tor T cells and affect APC function by downregulating
non-toxic regulatory T-cell-based immunotherapy for the the expression of MHC and co-stimulatory molecules.
induction of donor-specific tolerance and for the reduc- In addition, IL-10 promotes the secretion of IL-1 receptor
tion of the requirement for constant immunosuppression. antagonist and soluble tumour-necrosis factor receptor
In patients with pancreatic islet transplants, cell therapy rather than pro-inflammatory cytokines by APCs.
with alloantigen-specific TR1 cells in combination with Of particular interest is the observation that both
Scurfy mice immunosuppressive regimens that are permissive for interferon-γ (IFNγ) and nitric oxide (NO), which are
Loss-of-function mutations their in vivo expansion is currently under investigation produced at high levels during GVHD, have paradoxical
of Foxp3 in scurfy mice inhibit
the development of naturally
at the San Raffaele Scientific Institute. functions62. The pathogenic role of IFNγ as a T helper 1
occurring regulatory T cells, Finally, patients with IPEX possibly represent the best (TH1)-cell-associated cytokine that is involved in the
resulting in a highly candidates for testing the efficacy of regulatory T-cell development of immune-mediated diseases has been
dysregulated immune system immunotherapy in humans. Indeed, scurfy mice, which well documented63. Similarly, NO is a cytotoxic mol-
and consequent aggressive
develop a related syndrome as a result of a loss-of-function ecule that may contribute to immune-mediated diseases.
autoimmunity. These mice
show hyperproduction of Foxp3 mutation, can be rescued from lethality by injection However, when regulatory T cells are present, these two
cytokines and increased of neonates with FOXP3-expressing TReg cells. However, molecules acquire immunoregulatory activity62. In the
numbers of memory T cells. contrary to mice, in humans the FOXP3 defect does not presence of regulatory T cells, IFNγ stimulates APCs
seem to be only restricted to TReg cells25, and therefore the to produce NO and haem oxygenase 1 (HO1) through
CD95–CD95 ligand
apoptotic pathway
transfer of functional TReg cells alone might not represent increased expression of the enzymes inducible nitric-
CD95 ligand (also known as a definitve curative therapeutic option. A better under- oxide synthase (iNOS) and indoleamine 2,3-dioxygenase
FAS ligand) binds to CD95 standing of the immune mechanisms underlying IPEX (IDO), which degrades the essential amino acid tryp-
(FAS), which results in the needs to be gained before proposing TReg-cell-based tophan and markedly affects T-cell proliferation and
formation of the death-
immunotherapy for patients with this disease. survival. NO then can diffuse into neighbouring T cells,
inducing signalling complex
and subsequent activation influencing their function and triggering apoptosis,
of caspases. Donor CD4+ Possible mechanisms of regulatory T-cell therapy whereas HO1 prevents T-cell proliferation and modu-
T cells with killing capability To control complex immune-mediated diseases, such as lates T-cell activation through the degradation of the
preferentially use this pathway acute GVHD after HSCT and autoimmune diabetes, the pro-oxidant haem into carbon monoxide, iron and
during acute graft-versus-host
immune system needs to be modulated at different lev- biliverdin (reviewed in REF. 62).
els. Here we describe the known immunological events
Interleukin-1 receptor that occur in the two abovementioned pathological set- Regulatory T cells and the control of autoimmune diabetes.
antagonist tings in an attempt to delineate, based on our present Type 1 diabetes is a tissue-specific autoimmune disease
(IL-1RA). A secreted protein
knowledge, the level at which regulatory T cells could in which inflammation plays a central role (FIG. 2). The
that binds to IL-1R, thereby
blocking IL-1R downstream intervene. mechanisms through which regulatory T cells can func-
signalling. IL-1RA inhibits the tion in vivo to block the development of type 1 diabetes
pro-inflammatory properties Regulatory T cells and the control of acute GVHD. The are under intense investigation. The model that we pro-
of IL-1α and IL-1β. pathophysiology of acute GVHD after HSCT can be con- pose is based on data generated mainly in the mouse
Nitric oxide
sidered as a three step process in which both the innate models, and its clinical significance is yet to be proved.
(NO). A small molecule and adaptive immune systems interact (FIG. 1). As draining TReg cells present in the pancreatic draining lymph nodes
synthesized, mainly in lymph nodes are the sites of donor T-cell priming by host regulate the priming of autoreactive T cells by limit-
macrophages, from arginine antigen-presenting cells (APCs), it seems probable that ing their expansion and differentiation. Interestingly,
by nitric oxide synthase (NOS)
protection from GVHD by adoptively transferred regula- TReg cells may function by interrupting the development
enzymes. Increased levels of
NO are found in inflammatory tory T cells might also occur at these sites. Therefore, an of effector T cells through limiting the access of auto-
and autoimmune diseases, and important prerequisite for regulatory T cells to function reactive T cells to DCs64. Thanks to the control of this
during allograft rejection. It is is their ability to migrate to the correct site where they can crucial step in lymph-node priming, TReg cells limit the
the effector cytotoxic molecule encounter alloantigens presented by host APCs and be chance of a T cell becoming an effector cell. TReg cells
responsible for macrophage-
mediated cytotoxicity but
activated34,56. Early robust expansion of regulatory T cells inhibit the expression of CXC-chemokine receptor 3
can also suppress T-cell in draining lymph nodes followed by their migration and (CXCR3) by TH cells with a consequent lack of infiltration
proliferation. localization to peripheral tissues was reported in mice57, of these cells into the pancreatic islets65. Furthermore,


© 2007 Nature Publishing Group

a Pancreas b Pancreatic lymph node TReg cells might restrain autoimmune aggression directly
Islet of in the islets by controlling the inflammatory reaction
Release Peptide–
Langerhans of β-cell MHC class II CD4+
(insulitis)66. TR1 cells produce IL-10 and TGFβ, which
antigens iDC T cell both have an important immunoregulatory role in
type 1 diabetes67. Transient expression of TGFβ in the
islets during the priming phase of diabetes inhibits
β-cell Antigen the onset of disease and stimulates the expansion or
uptake generation of intra-islet FOXP3-expressing TReg cells68.
by iDCs TR1 cell IL-10 modulates APC function, reduces inflammation
TReg cell and decreases T-cell activation. Furthermore, IL-10
produced by TR1 cells downregulates the expression of
Release Granzymes IFNγ Expansion of
of β-cell intercellular adhesion molecule 1 (ICAM1) on effector
IL-10 autoreactive
antigens and TGFβ CD4+ effector T cells, which prevents their migration to the target
T cells organ69 (M.B. and M.-G.R., unpublished observations).
Finally, the paradoxical effects of IFNγ and NO that are
DC TCR thought to be important for the immunomodulatory
effects of regulatory T cells during GVHD might also
CD8+ have a role in autoimmune diabetes70.
T cell of adhesion
MHC class I molecules and Ongoing clinical trials with regulatory T cells
receptors At present, clinical trials of immunotherapy based on
c Insulitis regulatory T cells are ongoing in BMT for the preven-
tion or cure of GVHD. A cellular therapy clinical trial
that involves the transfer of IL-10-anergized TR1 cells
and to patients with haematological cancers who are treated
TNF with HLA-haploidentical HSCT is underway. In this
CD95L- trial, PBMCs are collected from both the donor (before
Apoptotic β-cell killing stem-cell mobilization) and the host (before condition-
ing). Subsequently, high doses of purified T-cell-depleted
Figure 2 | Pathogenesis of type 1 diabetes and control by regulatory T cells. As for CD34+ HSCs are infused into the myeloablated host.
graft-versus-host disease, the pathophysiology of type 1 diabetes can be envisaged as a Once there are signs of neutrophil engraftment, donor
three step process. a | First, under still undefined pathogenic conditions, modified islet PBMCs are thawed and cultured ex vivo in the presence
β-cell antigens are released and presented by MHC class I molecules. These previously of irradiated host PBMCs and IL-10. IL-10-anergized
‘cryptic’ antigens are presented by tissue-resident antigen-presenting cells (APCs) and
donor T cells are then infused at increasing doses (start-
recognized by CD8+ T cells that cause damage to MHC-class-I-expressing cells either
through the release of cytotoxic cytokines (such as interferon-γ (IFNγ)) or through the
ing from 105 per kg of body weight) into the host with
perforin–granzyme pathway. b | The released islet β-cell components are taken up by the ultimate goal of providing immune reconstitution
immature dendritic cells (iDCs) in the pancreatic islets and transported to the draining without GVHD. The cell preparation contains donor
pancreatic lymph nodes, where the antigens are processed and presented to CD4+ T cells that are anergic to host antigens and that are
T cells. Lymph-node priming is thought to be the second crucial step leading to enriched in precursors of host-specific TR1 cells, but it
expansion of low frequency circulating autoreactive T cells. After clonal expansion, also includes memory T cells that can respond to infec-
CD4+ effector T cells express adhesion molecules, such as intercellular adhesion tious agents and possibly provide a graft-versus-leukaemia
molecule 1 (ICAM1) and lymphocyte function-associated antigen 1 (LFA1), and (GVL) effect71. This clinical trial is currently being car-
chemokine receptors, such as CC-chemokine receptor 4 (CCR4), CCR5 and ried out in patients undergoing HLA-haploidentical
CXC-chemokine receptor 3. This allows the effector cells to home to the pancreatic
HSCT but it has the potential to be extended to include
islets, tracing antigen gradients and chemokines induced by the early CD8+ T-cell-
mediated inflammatory response. c | Once in the pancreas, the activated CD4+ T cells
patients undergoing allogeneic unrelated BMT or pan-
recruit and activate inflammatory cells, causing insulitis. The effector phase of islet creatic islet transplantation, in which there is high risk
β-cell destruction is mediated by cytokines (mainly interleukin-1 (IL-1) and tumour- of GVHD and graft rejection, respectively.
necrosis factor (TNF)) through the induction of pro-apoptotic signalling selectively in The first clinical trial with TReg cells is ongoing
islet β-cells and/or by inducing the expression of CD95 by islet β-cells, which allows in Regensburg, Germany, and is led by M. Edinger.
direct killing by CD95 ligand (CD95L)-expressing effector T cells. There is also evidence In this study, patients with high risk of cancer relapse or
that the production of free radicals is involved in the pathogenic events leading to islet molecular relapse after allogeneic HSCT receive a pre-
β-cell destruction (not shown). Regulatory T cells can intervene at different stages to emptive donor lymphocyte infusion (DLI) after cessation of
control type 1 diabetes. As in graft-versus-host disease, it is likely that regulatory T cells treatment with the immunosuppressant cyclosporin A.
are first activated in the pancreatic lymph nodes. Activated naturally occurring forkhead
Eight to ten weeks after immunosuppressive treatment
box P3 (FOXP3)+CD4+CD25+ regulatory T (TReg) cells and T regulatory type 1 (TR1) cells,
through distinct regulatory mechanisms, block the activation and expansion of effector
is withdrawn, patients receive between 1 × 106 and
T cells either directly or indirectly through APCs. Expression of adhesion molecules and 5 × 106 freshly isolated donor TReg cells per kg of body
chemokine receptors by effector T cells is also suppressed by regulatory T cells, with weight followed by DLI of equal T-cell numbers. The
consequent reduced effector T-cell migration to the target organ. The aggressiveness data generated so far in five treated patients indicate
of insulitis is also directly inhibited in the pancreas by regulatory T cells. TR1 cells, that the isolation of sufficient numbers of TReg cells from
through IL-10 and transforming growth factor-β (TGFβ) production, can inhibit the donor blood is feasible, and no acute toxicities or other
onset of disease and reduce inflammation. TCR, T-cell receptor. adverse events or complications (such as opportunistic

594 | AUGUST 2007 | VOLUME 7 www.nature.com/reviews/immunol

© 2007 Nature Publishing Group

infections, GVHD and disease relapse) have been

reported (M. Edinger, personal communication). This
trial will provide crucial information on the purifica- Donor of
tion procedure for TReg cells and on the safety of adop- regulatory
T cells
tive transfer with unmanipulated TReg cells. Another
trial, designed by B. Blazar and C. June (University
of Minnesota, USA), will start this year. In this study,
donor TReg cells expanded with beads coated with CD3-
and CD28-specific antibodies and high doses of IL-2 will
be infused into the transplant recipient together with the
HSCT (B. Blazar, personal communication). This trial will
Leukapheresis Cryopreservation
provide valuable information on the transfer of polyclonal
TReg cells expanded ex vivo. Hopefully, these three clini-
cal studies will pave the way for the future of regulatory Thaw cells
T-cell-based immunotherapy in other diseases.

Strengths and pitfalls of regulatory T-cell therapy Cell selection

(using clinical
Strengths. The administration of immunosuppres- grade device and Vitality
sive drugs is the most common approach for treating GMP materials)
immune-mediated diseases. However, the long-term
administration of non-antigen-specific agents that
cannot distinguish between beneficial and destructive
immune responses is a major drawback. Furthermore,
• Absence of
immunosuppressive treatment must be lifelong, as, Sorted cells adventitious agents
if withdrawn, there is high risk of disease relapse due • Vitality
to lack of tolerance induction. By contrast, regulatory
T cells are physiological components of the immune
system and their adoptive transfer should re-establish the
immunological homeostasis altered under pathological
circumstances. Immunotherapy with TR1 cells can, for Regulatory T-cell • Absence of
expansion or adventitious agents Quality
example, be envisaged as an in vivo transfer of a ‘biologi- differentiation • Vitality
• Phenotype control
cal source’ of IL-10, which, if administered exogenously (using clinical GMP
materials) • Function
as a recombinant protein, would probably not have the
same therapeutic effects. Indeed, we have shown that
administration of exogenous IL-10 alone in a preclinical
model of allogeneic pancreatic islet transplantation does Cryopreservation
not block allograft rejection, whereas adoptive transfer
Thaw cells
of antigen-specific TR1 cells, which produce IL-10 upon
antigen encounter, prevents rejection and induces GMP certificate • Absence of
long-term tolerance42. released by adventitious agents
qualified person • Vitality
Importantly, regulatory T-cell-based therapy is cus- • Phenotype
tomized to the patient and can therefore be created to • Function
Graft-versus-leukaemia satisfy the specific needs of the patient with limited side
(GVL). An immune response effects. This is, however, a double-edged sword, as T-cell
mounted by the transplanted
products cannot be manufactured and distributed easily Infusion of
cells against the tumour cells
and freely as is the case for standard medicinal prod- patient with
of the host and it is one of regulatory T cells
the reasons that allogeneic ucts and can be prohibitively expensive in their current
transplants can be curative experimental phase.
for cancer.

Donor lymphocyte infusion

Pitfalls. The obstacles that limit regulatory T-cell-based
(DLI). Delayed administration immunotherapy at present are mainly technical and relate
of peripheral donor to cell manipulation. The regulatory T cells must be col-
lymphocytes after T-cell- lected from the peripheral blood, for example by leuka- Figure 3 | Development and assessment of human
depleted bone-marrow cell therapy products. The steps necessary for the
pheresis, and immediately processed or cryopreserved
transplantation in cancer development of regulatory T cells as a medicinal product
patients. DLI plays a central when necessary. Circulating TReg cells can be directly
are shown. From cell collection to cell processing,
role in both attacking the isolated from the circulating pool of CD4+ T cells but
expansion and differentiation, and final infusion into
tumour cells and providing because of their limited number they need to be further the patient, clinical good manufacturing practice (GMP)
immune reconstitution. expanded in vitro. TR1 cells are generated from CD4+
However, its use is limited by procedures need to be performed. Various quality controls
the risk of severe acute and
T cells, which need first to be isolated and primed in vitro. are required in each manufacturing step and the final
chronic graft-versus-host Cell isolation and manipulation are therefore prerequi- product can be released and infused into the patient only
disease. site for regulatory T-cell-based immunotherapy. Several after approval by a qualified person.


© 2007 Nature Publishing Group

other points need to be considered for the development In addition, the need for a patient-customized cell product
of regulatory T cells as a medicinal product and many prepared in GMP conditions may severely hamper a broad
different quality controls are needed (FIG. 3). application of this cellular drug. However, proof of robust
Cell isolation, manipulation, expansion and re- efficacy of this cellular product would generate increasing
infusion into the patients are all procedures that need interest and investments to overcome these hurdles.
to be performed under GMP conditions. As a result, Regulatory T-cell-based immunotherapy should not
this therapeutic approach is extremely expensive in its be envisaged as an all-or-nothing approach to re-establish
experimental phase and also requires certified GMP immunological tolerance on its own. As mentioned,
facilities. It is therefore unavoidable that only a few insti- pathological conditions such as GVHD and type 1 dia-
tutions can provide all the necessary infrastructure to betes are multifactorial diseases and monotherapy risks
make regulatory T-cell-based therapy a reality. However, being of limited success. We therefore propose that com-
should this immunotherapy meet its therapeutic targets, bining regulatory T-cell therapy with other therapeutic
we might envisage that not only academic institutions approaches might result in better efficacy. The selected
but also pharmaceutical companies will be interested in compound should allow survival, expansion and function
adopting this therapeutic approach. of regulatory T cells. Furthermore, it should prove of supe-
Safety of the infused ex vivo manipulated product rior efficacy over monotherapy with regulatory T cells.
is clearly a high priority. The risk of uncontrolled cell Rapamycin has been shown to be effective in inducing the
proliferation, pan immunosuppression and consequent expansion of TReg cells and to be permissive for TR1-cell
tumour development should be carefully monitored. induction by IL-10 (REF. 69). Similarly, other immuno-
Furthermore, superior efficacy of regulatory T-cell- suppressive compounds that do not act through the TCR
based immunotherapy over conventional therapy should signalling and calcineurin pathways may be combined
be clearly demonstrated. with regulatory T-cell immunotherapy to downregulate
inflammation and effector T-cell function.
Current picture and future perspectives In the future, ex vivo gene transfer with genes
Preliminary results from clinical trials with freshly encoding proteins that induce or enhance regulatory
ex vivo isolated TReg cells and ex vivo generated TR1 cells T-cell function and confer regulatory activity to effec-
in patients undergoing BMT show feasibility and sug- tor T cells can also be envisaged. Evidence that human
gest a good safety profile. Efficacy data will hopefully TReg cells constitutively express high levels of FOXP3
be generated in the coming years. These proof-of- and that mutations in FOXP3 results in severe autoim-
principle clinical trials in HSCT will pave the way for munity25 clearly indicates that the expression of this
further clinical studies in patients with genetic and transcription factor has a key role in regulatory T-cell
acquired autoimmune diseases. However, these diseases function. It is therefore possible that ectopic expression
may intrinsically have a higher level of complexity: they of FOXP3 in non-regulatory T cells may represent an
are not life threatening and therefore the risk–benefit alternative therapeutic approach for the generation of
assessment should be more stringent; the patients are stable and functional T cells with regulatory properties.
not lymphopenic but instead may have a high number Furthermore, the transformation of antigen-specific
of circulating activated effector T cells; and the regula- pathogenic T cells into regulatory T cells upon forced
tory T cells will need to suppress memory autoreactive expression of FOXP3 might overcome all of the current
T cells in an inflammatory environment. limitations in generating antigen-specific TReg cells ex vivo.
In some clinical settings, ex vivo expansion of Engineering of antigen-specific pathogenic T cells with
TReg cells will be necessary to achieve a sufficient number genes that encode immunosuppressive cytokines such as
of cells per kg of body weight. So far, sufficiently high IL-10 and TGFβ could also be an interesting possibility.
numbers of TReg cells can be generated for preclinical However, solid preclinical data supporting the safety and
studies but it remains to be determined whether the efficacy of these gene therapy approaches are required to
procedure can be scaled up under GMP conditions. propel it from the bench to the clinic.

1. Shevach, E. M. From vanilla to 28 flavors: multiple 6. Wing, K. et al. CD4 T cell activation by myelin 10. Baecher-Allan, C. & Hafler, D. A. Human
varieties of T regulatory cells. Immunity 25, 195–201 oligodendrocyte glycoprotein is suppressed by adult regulatory T cells and their role in autoimmune
(2006). but not cord blood CD25+ T cells. Eur. J. Immunol. disease. Immunol. Rev. 212, 203–216 (2006).
2. Groux, H. et al. A CD4+ T-cell subset inhibits antigen- 33, 579–587 (2003). 11. Yudoh, K., Matsuno, H., Nakazawa, F.,
specific T-cell responses and prevents colitis. Nature 7. Danke, N. A., Koelle, D. M., Yee, C., Beheray, S. & Yonezawa, T. & Kimura, T. Reduced expression
389, 737–742 (1997). Kwok, W. W. Autoreactive T cells in healthy of the regulatory CD4+ T cell subset is related
In this paper, both human and mouse IL-10- individuals. J. Immunol. 172, 5967–5972 to Th1/Th2 balance and disease severity in
producing TR1 cells were characterized for the first (2004). rheumatoid arthritis.Arthritis Rheum. 43,
time. 8. Kitani, A., Chua, K., Nakamura, K. & Strober, W. 617–627 (2000).
3. Walker, M. R. et al. Induction of FoxP3 and acquisition Activated self-MHC-reactive T cells have the 12. Miura, Y. et al. Association of Foxp3 regulatory
of T regulatory activity by stimulated human CD4+ cytokine phenotype of Th3/T regulatory cell 1 T cells. gene expression with graft-versus-host disease.
CD25– T cells. J. Clin. Invest. 112, 1437–1443 (2003). J. Immunol. 165, 691–702 (2000). Blood 104, 2187–2193 (2004).
4. Sakaguchi, S., Setoguchi, R., Yagi, H. & Nomura, T. 9. Arif, S. et al. Autoreactive T cell responses show This study shows for the first time a significant
Naturally arising Foxp3-expressing CD25+CD4+ proinflammatory polarization in diabetes but a reduction of FOXP3 mRNA levels in PBMCs from
regulatory T cells in self-tolerance and autoimmune regulatory phenotype in health. J. Clin. Invest. patients with GVHD compared with those without
disease. Curr. Top. Microbiol. Immunol. 305, 51–66 113, 451–463 (2004). GVHD.
(2006). This is the first study providing evidence that 13. Rezvani, K. et al. High donor FOXP3-positive
5. Taams, L. S. et al. Antigen-specific T cell suppression patients with type 1 diabetes have reduced regulatory T-cell (Treg) content is associated with a low
by human CD4+CD25+ regulatory T cells. Eur. J. numbers of circulating islet-specific TR1 cells risk of GVHD following HLA-matched allogeneic SCT.
Immunol. 32, 1621–1630 (2002). compared with HLA-matched healthy controls. Blood 108, 1291–1297 (2006).

596 | AUGUST 2007 | VOLUME 7 www.nature.com/reviews/immunol

© 2007 Nature Publishing Group

14. Zorn, E. et al. Reduced frequency of 36. Hoffmann, P., Ermann, J., Edinger, M., Fathman, C. G. 57. Nguyen, V. H. et al. In vivo dynamics of regulatory
FOXP3+CD4+CD25+ regulatory T cells in patients & Strober, S. Donor-type CD4+CD25+ regulatory T cell trafficking and survival predict effective
with chronic graft-versus-host disease. Blood T cells suppress lethal acute graft-versus-host disease strategies to control graft-versus-host disease
106, 2903–2911 (2005). after allogeneic bone marrow transplantation. following allogeneic transplantation. Blood
15. Clark, F. J. et al. Chronic graft-versus-host disease J. Exp. Med. 196, 389–399 (2002). 109, 2649–2656 (2007).
is associated with increased numbers of peripheral References 35, 36 and 96 are the first to show that 58. Misra, N., Bayry, J., Lacroix-Desmazes, S.,
blood CD4+CD25high regulatory T cells. Blood the adoptive transfer of CD4+CD25+ regulatory Kazatchkine, M. D. & Kaveri, S. V. Cutting edge:
103, 2410–2416 (2004). T cells significantly delays or even prevents GVHD human CD4+CD25+ T cells restrain the maturation
16. Meignin, V. et al. Numbers of Foxp3-expressing in preclinical mouse models of BMT. and antigen-presenting function of dendritic cells.
CD4+CD25high T cells do not correlate with the 37. Trenado, A. et al. Recipient-type specific CD4+CD25+ J. Immunol. 172, 4676–4680 (2004).
establishment of long-term tolerance after regulatory T cells favor immune reconstitution and 59. Venet, F. et al. Human CD4+CD25+ regulatory
allogeneic stem cell transplantation. Exp. Hematol. control graft-versus-host disease while maintaining T lymphocytes inhibit lipopolysaccharide-induced
33, 894–900 (2005). graft-versus-leukemia. J. Clin. Invest. 112, 1688–1696 monocyte survival through a Fas/Fas ligand-
17. Rieger, K. et al. Mucosal FOXP3+ regulatory T cells (2003). dependent mechanism. J. Immunol. 177,
are numerically deficient in acute and chronic GvHD. 38. Zeller, J. C. et al. Induction of CD4+ T cell alloantigen- 6540–6547 (2006).
Blood 107, 1717–1723 (2006). specific hyporesponsiveness by IL-10 and TGF-β. 60. Taams, L. S. et al. Modulation of monocyte/
18. Meloni, F. et al. Monocyte chemoattractant protein-1 J. Immunol. 163, 3684–3691 (1999). macrophage function by human CD4+CD25+
levels in bronchoalveolar lavage fluid of lung- 39. Gregori, S. et al. Regulatory T cells induced by regulatory T cells. Hum. Immunol. 66, 222–230
transplanted patients treated with tacrolimus as 1α,25-dihydroxyvitamin D3 and mycophenolate (2005).
rescue treatment for refractory acute rejection. mofetil treatment mediate transplantation tolerance. 61. Lewkowicz, P., Lewkowicz, N., Sasiak, A. &
Transplant Proc. 35, 1523–1526 (2003). J. Immunol. 167, 1945–1953 (2001). Tchorzewski, H. Lipopolysaccharide-activated
19. Demirkiran, A. et al. Low circulating regulatory T-cell 40. Lee, M. K. 4th. et al. Promotion of allograft survival by CD4+CD25+ T regulatory cells inhibit neutrophil
levels after acute rejection in liver transplantation. CD4+CD25+ regulatory T cells: evidence for in vivo function and promote their apoptosis and death.
Liver Transpl. 12, 277–284 (2006). inhibition of effector cell proliferation. J. Immunol. J. Immunol. 177, 7155–7163 (2006).
20. Salama, A. D., Najafian, N., Clarkson, M. R., 172, 6539–6544 (2004). 62. Wood, K. J. & Sawitzki, B. Interferon γ: a crucial role
Harmon, W. E. & Sayegh, M. H. Regulatory CD25+ 41. Battaglia, M., Stabilini, A. & Roncarolo, M. G. in the function of induced regulatory T cells in vivo.
T cells in human kidney transplant recipients. Rapamycin selectively expands CD4+CD25+FoxP3+ Trends Immunol. 27, 183–187 (2006).
J. Am. Soc. Nephrol. 14, 1643–1651 (2003). regulatory T cells. Blood 105, 4743–4748 63. Farrar, M. A. & Schreiber, R. D. The molecular cell
21. Bacchetta, R. et al. High levels of interleukin 10 (2005). biology of interferon-γ and its receptor. Annu. Rev.
production in vivo are associated with tolerance In this study, we showed, for the first time, that Immunol. 11, 571–611 (1993).
in SCID patients transplanted with HLA mismatched rapamycin has the ability to allow the selective 64. Tang, Q. et al. Visualizing regulatory T cell control
hematopoietic stem cells. J. Exp. Med. 179, 493–502 proliferation of TReg cells. of autoimmune responses in nonobese diabetic mice.
(1994). 42. Battaglia, M. et al. Rapamycin and interleukin-10 Nature Immunol. 7, 83–92 (2006).
This study describes the successful isolation of treatment induces T regulatory type 1 cells that 65. Sarween, N. et al. CD4+CD25+ cells controlling
CD4+ host-reactive T-cell clones from SCID patients mediate antigen-specific transplantation tolerance. a pathogenic CD4 response inhibit cytokine
transplanted with allogeneic HSCs, which produce Diabetes 55, 40–49 (2006). differentiation, CXCR-3 expression, and tissue
high amounts of IL-10 in the absence of IL-4 after 43. Waldmann, H., Adams, E., Fairchild, P. & Cobbold, S. invasion. J. Immunol. 173, 2942–2951 (2004).
antigen-specific stimulation in vitro. The presence Infectious tolerance and the long-term acceptance 66. Chen, Z., Herman, A. E., Matos, M., Mathis, D. &
of these IL-10-producing CD4+ T cells correlated of transplanted tissue. Immunol. Rev. 212, 301–313 Benoist, C. Where CD4+CD25+ T reg cells impinge on
with the absence of GVHD and long-term tolerance. (2006). autoimmune diabetes. J. Exp. Med. 202, 1387–1397
22. Baker, K. et al. High spontaneous IL-10 production 44. Barrat, F. J. et al. In vitro generation of (2005).
in unrelated bone marrow transplant recipients is interleukin 10-producing regulatory CD4+ T cells is 67. Roncarolo, M. G., Battaglia, M. & Gregori, S.
associated with fewer transplant-related complications induced by immunosuppressive drugs and inhibited The role of interleukin 10 in the control of
and early deaths. Bone Marrow Transplant. by T helper type 1 (Th1)- and Th2-inducing cytokines. autoimmunity. J. Autoimmun. 20, 269–272 (2003).
23, 1123–1129 (1999). J. Exp. Med. 195, 603–616 (2002). 68. Peng, Y., Laouar, Y., Li, M. O., Green, E. A. &
23. Weston, L. E., Geczy, A. F. & Briscoe, H. Production 45. Hoffmann, P. et al. Isolation of CD4+CD25+ Flavell, R. A. TGF-β regulates in vivo expansion of
of IL-10 by alloreactive sibling donor cells and its regulatory T cells for clinical trials. Biol. Blood Foxp3-expressing CD4+CD25+ regulatory T cells
influence on the development of acute GVHD. Marrow Transplant. 12, 267–274 (2006). responsible for protection against diabetes.
Bone Marrow Transplant. 37, 207–212 (2005). 46. Hoffmann, P., Eder, R., Kunz-Schughart, L. A., Proc. Natl Acad. Sci. USA 101, 4572–4577 (2004).
24. VanBuskirk, A. M. et al. Human allograft acceptance Andreesen, R. & Edinger, M. Large-scale in vitro 69. Battaglia, M. et al. Induction of tolerance in type 1
is associated with immune regulation. J. Clin. Invest. expansion of polyclonal human CD4+CD25high diabetes via both CD4+CD25+ T regulatory cells and
106, 145–155 (2000). regulatory T cells. Blood 104, 895–903 (2004). T regulatory type 1 cells. Diabetes 55, 1571–1580
25. Bacchetta, R. et al. Defective regulatory and effector 47. Levings, M. K. et al. Human CD25+CD4+ (2006).
T cell functions in patients with FOXP3 mutations. T suppressor cell clones produce transforming 70. Chen, C., Lee, W. H., Zhong, L. & Liu, C. P.
J. Clin. Invest. 116, 1713–1722 (2006). growth factor β, but not interleukin 10, and are Regulatory T cells can mediate their function
26. Marangoni, F. et al. WASP regulates suppressor activity distinct from type 1 T regulatory cells. J. Exp. Med. through the stimulation of APCs to produce
of human and murine CD4+CD25+FOXP3+ natural 196, 1335–1346 (2002). immunosuppressive nitric oxide. J. Immunol.
regulatory T cells. J. Exp. Med. 204, 369–380 (2007). 48. Battaglia, M. et al. Rapamycin promotes expansion 176, 3449–3460 (2006).
27. Kriegel, M. A. et al. Defective suppressor function of functional CD4+CD25+FOXP3+ regulatory T cells of 71. Battaglia, M., Gregori, S., Bacchetta, R. &
of human CD4+ CD25+ regulatory T cells in both healthy subjects and type 1 diabetic patients. Roncarolo, M. G. Tr1 cells: from discovery to their
autoimmune polyglandular syndrome type II. J. Immunol. 177, 8338–8347 (2006). clinical application. Semin. Immunol. 18, 120–127
J. Exp. Med. 199, 1285–1291 (2004). 49. Liu, W. et al. CD127 expression inversely correlates (2006).
28. Bleesing, J. J. et al. Immunophenotypic profiles with FoxP3 and suppressive function of human CD4+ 72. Sakaguchi, S. Naturally arising Foxp3-expressing
in families with autoimmune lymphoproliferative T reg cells. J. Exp. Med. 203, 1701–1711 (2006). CD25+CD4+ regulatory T cells in immunological
syndrome. Blood 98, 2466–2473 (2001). 50. Seddiki, N. et al. Expression of interleukin (IL)-2 tolerance to self and non-self. Nature Immunol.
29. Mottet, C., Uhlig, H. H. & Powrie, F. Cutting edge: and IL-7 receptors discriminates between human 6, 345–352 (2005).
cure of colitis by CD4+CD25+ regulatory T cells. regulatory and activated T cells. J. Exp. Med. 73. Allan, S. E. et al. The role of 2 FOXP3 isoforms in
J. Immunol. 170, 3939–3943 (2003). 203, 1693–1700 (2006). the generation of human CD4+ Tregs. J. Clin. Invest.
30. Tang, Q. et al. In vitro-expanded antigen-specific 51. Jiang, S., Camara, N., Lombardi, G. & Lechler, R. I. 115, 3276–3284 (2005).
regulatory T cells suppress autoimmune diabetes. Induction of allopeptide-specific human CD4+CD25+ 74. Singh, B. et al. Control of intestinal inflammation
J. Exp. Med. 199, 1455–1465 (2004). regulatory T cells ex vivo. Blood 102, 2180–2186 by regulatory T cells. Immunol. Rev. 182, 190–200
31. Tarbell, K. V. et al. Dendritic cell-expanded, islet- (2003). (2001).
specific CD4+ CD25+ CD62L+ regulatory T cells References 50 and 51 show that CD4+CD25+ 75. Shevach, E. M. et al. The lifestyle of naturally
restore normoglycemia in diabetic NOD mice. TReg cells are CD127low/– and open new venues occurring CD4+ CD25+ Foxp3+ regulatory T cells.
J. Exp. Med. 204, 191–201 (2007). for a better isolation and purification of TReg cells. Immunol. Rev. 212, 60–73 (2006).
32. Hanash, A. M. & Levy, R. B. Donor CD4+CD25+ 52. Roncarolo, M. G. et al. Interleukin-10-secreting 76. Scheffold, A., Hühn, J. & Höfer, T. Regulation of
T cells promote engraftment and tolerance following type 1 regulatory T cells in rodents and humans. CD4+CD25+ regulatory T cell activity: it takes
MHC-mismatched hematopoietic cell transplantation. Immunol. Rev. 212, 28–50 (2006). (IL-) two to tango. Eur. J. Immunol. 35, 1336–1341
Blood 105, 1828–1836 (2005). 53. Battaglia, M. & Roncarolo, M. G. Induction of (2005).
33. Joffre, O., Gorsse, N., Romagnoli, P., Hudrisier, D. & transplantation tolerance via regulatory T cells. 77. Levings, M. K. et al. Differentiation of Tr1 cells by
van Meerwijk, J. P. Induction of antigen-specific Inflamm. Allergy Drug Targets 5, 157–165 (2006). immature dendritic cells requires IL-10 but not
tolerance to bone marrow allografts with CD4+CD25+ 54. Zeiser, R. et al. Inhibition of CD4+CD25+ regulatory CD25+CD4+ Tr cells. Blood 105, 1162–1169 (2005).
T lymphocytes. Blood 103, 4216–4221 (2004). T-cell function by calcineurin-dependent interleukin-2 78. Levings, M. K. et al. IFN-α and IL-10 induce the
34. Taylor, P. A. et al. L-selectinhi but not the L-selectinlo production. Blood 108, 390–399 (2006). differentiation of human type 1 T regulatory cells.
CD4+25+ T-regulatory cells are potent inhibitors of 55. Achenbach, P., Bonifacio, E. & Ziegler, A. G. J. Immunol. 166, 5530–5539 (2001).
GVHD and BM graft rejection. Blood 104, 3804–3812 Predicting type 1 diabetes. Curr. Diab. Rep. 79. Ziegler, S. F. FOXP3: of mice and men. Annu. Rev.
(2004). 5, 98–103 (2005). Immunol. 24, 209–226 (2006).
35. Cohen, J. L., Trenado, A., Vasey, D., Klatzmann, D. & 56. Ermann, J. et al. Only the CD62L+ subpopulation 80. Khattri, R., Cox, T., Yasayko, S. A. & Ramsdell, F.
Salomon, B. L. CD4+CD25+ immunoregulatory of CD4+CD25+ regulatory T cells protects from An essential role for Scurfin in CD4+CD25+
T cells: new therapeutics for graft-versus-host disease. lethal acute GVHD. Blood 105, 2220–2226 T regulatory cells. Nature Immunol. 4, 337–342
J. Exp. Med. 196, 401–406 (2002). (2005). (2003).


© 2007 Nature Publishing Group

81. Gondek, D. C., Lu, L. F., Quezada, S. A., Sakaguchi, S. 89. You, S. et al. Presence of diabetes-inhibiting, glutamic 96. Taylor, P. A., Lees, C. J. & Blazar, B. R. The infusion
& Noelle, R. J. Cutting edge: contact-mediated acid decarboxylase-specific, IL-10-dependent, of ex vivo activated and expanded CD4+CD25+
suppression by CD4+CD25+ regulatory cells involves regulatory T cells in naive nonobese diabetic mice. immune regulatory cells inhibits graft-versus-host
a granzyme B-dependent, perforin-independent J. Immunol. 173, 6777–6785 (2004). disease lethality. Blood 99, 3493–3499 (2002).
mechanism. J. Immunol. 174, 1783–1786 (2005). 90. Hori, S., Haury, M., Coutinho, A. & Demengeot, J.
82. Zhao, D. M., Thornton, A. M., DiPaolo, R. J. & Specificity requirements for selection and effector Acknowledgements
Shevach, E. M. Activated CD4+CD25+ T cells functions of CD25+4+ regulatory T cells in anti-myelin The authors thank R. Bacchetta and S. Gregori (HSR-TIGET)
selectively kill B lymphocytes. Blood 107, 3925–3932 basic protein T cell receptor transgenic mice. for scientific discussions. M.-G.R. is supported by grants from
(2006). Proc. Natl Acad. Sci. USA 99, 8213–8218 (2002). the Telethon Foundation, RISET and the Juvenile Diabetes
83. Grossman, W. J. et al. Human T regulatory cells can 91. Kohm, A. P., Carpentier, P. A., Anger, H. A. & Research Foundation (JDRF). M.B. is supported by grants
use the perforin pathway to cause autologous target Miller, S. D. Cutting edge: CD4+CD25+ regulatory from Telethon and the JDRF.
cell death. Immunity 21, 589–601 (2004). T cells suppress antigen-specific autoreactive
84. Chen, D. et al. CD4+ CD25+ regulatory T-cells immune responses and central nervous system Competing interests statement
inhibit the islet innate immune response and inflammation during active experimental autoimmune The authors declare no competing financial interests.
promote islet engraftment. Diabetes 55, 1011–1021 encephalomyelitis. J. Immunol. 169, 4712–4716
(2006). (2002). DATABASES
85. Ding, Q. et al. B7H1-Ig fusion protein activates the 92. Morgan, M. E. et al. Effective treatment of The following terms in this article are linked online to:
CD4+ IFN-γ receptor+ type 1 T regulatory subset collagen-induced arthritis by adoptive transfer Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.
through IFN-γ-secreting Th1 cells. J. Immunol. of CD25+ regulatory T cells. Arthritis Rheum. fcgi?db=gene
177, 3606–3614 (2006). 52, 2212–2221 (2005). CD4 | CD25 | CD127 | FOXP3 | TGFβ
86. Tarbell, K. V., Yamazaki, S., Olson, K., Toy, P. & 93. Cong, Y., Weaver, C. T., Lazenby, A. & Elson, C. O.
Steinman, R. M. CD25+ CD4+ T cells, expanded Bacterial-reactive T regulatory cells inhibit FURTHER INFORMATION
with dendritic cells presenting a single autoantigenic pathogenic immune responses to the enteric flora. Maria-Grazia Roncarolo’s homepage: http://www.
peptide, suppress autoimmune diabetes. J. Exp. Med. J. Immunol. 169, 6112–6119 (2002). sanraffaele.org/EN_home/Research/Departments-Institutes_
199, 1467–1477 (2004). 94. Scalapino, K. J., Tang, Q., Bluestone, J. A., e_Research_Programs/The_San_Raffaele_Telethon_Institute_
87. Masteller, E. L. et al. Expansion of functional Bonyhadi, M. L. & Daikh, D. I. Suppression of disease for_Gene_Therapy_(HSR-TIGET)/index.html
endogenous antigen-specific CD4+CD25+ regulatory in New Zealand Black/New Zealand White lupus-prone RISET transplantation tolerance: http://www.risetfp6.org
T cells from nonobese diabetic mice. J. Immunol. mice by adoptive transfer of ex vivo expanded The European Agency for the Evaluation of Medicinal
175, 3053–3059 (2005). regulatory T cells. J. Immunol. 177, 1451–1459 Products: http://www.emea.europa.eu/pdfs/human/bwp/
88. Chen, C., Lee, W. H., Yun, P., Snow, P. & (2006). 4145098EN.pdf
Liu, C. P. Induction of autoantigen-specific Th2 95. Fontenot, J. D., Gavin, M. A. & Rudensky, A. Y. U.S. Food and Drug Administration:
and Tr1 regulatory T cells and modulation of Foxp3 programs the development and function of http://www.fda.gov/cber/gene.htm
autoimmune diabetes. J. Immunol. 171, 733–744 CD4+CD25+ regulatory T cells. Nature Immunol. Access to this links box is available online.
(2003). 4, 330–336 (2003).

598 | AUGUST 2007 | VOLUME 7 www.nature.com/reviews/immunol

© 2007 Nature Publishing Group