International Immunology 2007; 1 of 8
ª The Japanese Society for Immunology. 2007. All rights reserved.
For permissions, please e-mail: email@example.com
Anti-thymocyte globulin (ATG) prevents autoimmune
encephalomyelitis by expanding myelin antigen-
specific Foxp31 regulatory T cells
Denise T. Chung1,*, Thomas Korn1,*, Julie Richard2, Melanie Ruzek2, Adam P. Kohm3,
Stephen Miller3, Sharon Nahill2and Mohamed Oukka1
1Center for Neurologic Diseases, Brigham and Women’s Hospital and Harvard Medical School, 77 Avenue Louis Pasteur,
HIM 780, Boston, MA 02115, USA
2U.S. Scientific Development, Genzyme Corporation, Framingham, MA 01701, USA
3Interdepartmental Immunobiology Center, Feinberg School of Medicine, Northwestern University, Chicago, IL 60611, USA
Keywords: anti-thymocyte globulin, T-reg cell, antigen-specific, EAE, multiple sclerosis
The T cell-depleting polyclonal antibody, anti-thymocyte globulin (ATG) has long been used in organ
transplantation to treat acute rejection episodes. More recently, it is also being used as part of an
induction regimen to protect allografts. It has been proposed that ATG might deplete effector T cells
(T-effs) while sparing regulatory T cells (T-regs). In order to test whether ATG is effective in
autoimmune disease, we used Foxp3gfp ‘knock-in’ mice in combination with a myelin
oligodendrocyte glycoprotein (MOG)35–55/IAbtetramer to study more closely the effect of ATG
treatment on antigen-specific T cell responses in vivo during MOG-induced experimental autoimmune
encephalomyelitis (EAE), an animal model for Multiple Sclerosis. ATG treatment enhanced the
expansion of MOG-specific T-regs (CD41Foxp31) in MOG-immunized mice. T-effs were depleted, but
on a single-cell basis, the effector function of residual T-effs was not compromised by ATG. Thus,
ATG tipped the balance of T-effs and T-regs and skewed an auto-antigen-specific immune reaction
from a pathogenic T cell response to a potentially protective T-reg response. In both acute and
relapsing remitting disease models, ATG treatment resulted in the attenuation from EAE, both in
a preventive and early therapeutic setting. We conclude that ATG treatment enforces the development
of a dominant immunoregulatory environment which may be advantageous for the treatment of T cell-
driven autoimmune diseases.
The ability of CD4+CD25+regulatory T cells (T-regs) to actively
suppress activation and expansion of self-reactive T cells is
critical for the protection from autoimmunity (1, 2). In the clini-
cal setting, it has been regarded as a potential therapeutic
approach to manipulate the balance between T-regs and ef-
fector T cells (T-effs) in order to treat active autoimmune dis-
eases in humans. T-reg-based treatment strategies have
been tested in animal models for transplantation (3–5) and
human autoimmune diseases including type I diabetes (T1D)
(6), multiple sclerosis (MS) (7) and rheumatoid arthritis (8).
Antibody treatments that have been used to shift the balance
between T-effs and T-regs in vivo include therapy of T1D (9–
11) and experimental autoimmune encephalomyelitis (EAE)
(12) with anti-CD3. The exact mode of action of non-mitogenic
anti-CD3 antibody remains elusive but it has been suggested
that it may induce anergy in recently activated T cells (13), al-
ter the balance between T-effs and T-regs by differential de-
pletion of T-effs or induce de novo generation of T-regs (14).
The polyclonal antibodies, anti-lymphocyte serum (ALS)
and anti-thymocyte globulin (ATG), that induce broad T cell
depletion have also been used to treat graft versus host dis-
ease (15), acute rejection in organ transplantation (16), and
autoimmune diabetes (17) and are FDA approved for the
treatment of renal transplant patients. However, at this point,
the exact mechanism of action is still unclear and has mostly
been regarded as generalized immunosuppression by T cell
depletion via complement-dependent lysis and Fas/Fas
ligand-mediated activation-induced cell death (18, 19). Only
*These authors contributed equally to this study.
Shared Senior authorship
Correspondence to: M. Oukka; E-mail: firstname.lastname@example.org Received 20 March 2007, accepted 30 May 2007
International Immunology Advance Access published August 13, 2007
by guest on June 1, 2013
recently, it has also been suggested that ATG might differen-
tially deplete T-effs but spare T-regs (20) or even induce the
generation of CD4+CD25+Foxp3+
However, functional analyses in these studies have been
marred by using CD25 as a marker for T-regs since CD25 is
also expressed on activated T-effs. Moreover, since acti-
vated T cells appear to be more susceptible to ATG-induced
depletion (18, 19), the question arises whether ATG depletes
the T cell repertoire in a skewed manner. Thus, it is essential
to study antigen-specific responses of both T-regs and T-effs
in order to address this issue.
In this report, we directly approach these questions by the
combination of technologies that we recently developed in
ourlaboratory. First,we used
(Foxp3gfp.KI) mouse that allowed us to faithfully track
Foxp3-expressing T cells in vivo. Second, we applied a mye-
lin oligodendrocyte glycoprotein (MOG)35–55/IAbtetramer in
order to visualize antigen-specific responses in vivo during
EAE. Overall, our data show that ATG treatment drives the
expansion of MOG-specific T-regs after preferential deple-
tion of T-effs, resulting in reduced EAE.
T-regs de novo (21).
a Foxp3gfp ‘knock-in’
C57Bl/6 and SJL/J mice (6–8 weeks old) were obtained from
Jackson Laboratory (Bar Harbor, ME, USA). Foxp3gfp
KI mice have been described previously (22, 23). Briefly, a
bicistronic eGFP reporter gene was introduced after the stop
codon of the endogenous Foxp3 locus. This strategy en-
abled us to track Foxp3-expressing regulatory cells in vivo.
The introduction of the eGFP reporter did not alter Foxp3
gene expression. Foxp3gfp. KI mice developed normally
and were not phenotypically different from their wild-type lit-
termates. All animals were kept in specific pathogen-free
conditions. All experiments and protocols were performed in
accordance to Institutional Animal Care and Use Committee
IACUC guidelines. All experiments with SJL/J mice were per-
formed in Genzyme Corporation (Framingham, MA, USA).
Induction of EAE and administration of ATG
Age-matched C57Bl/6 mice were immunized subcutane-
ously in the flanks with 100 lg of MOG35–55 (MEVG-
WYRSPFSRVVHLYRNGK) emulsified in CFA supplemented
with 5 mg ml?1Mycobacterium tuberculosis H37Ra (Difco,
Detroit, MI, USA). Pertussis toxin (List Biological Laborato-
ries, Campbell, CA, USA) was given intra-peritoneally (i.p.)
on days 0 and 2 in a dose of 200 ng per mouse.
Age-matched SJL/J mice were immunized subcutaneously
in the flanks with 60 lg of PLP139–151(HSLGKWLGHPDKF)
emulsified in CFA supplemented with 3 mg ml?1M. tuberculosis
H37Ra. Pertussis toxin (List Biological Laboratories) was given
intravenously on days 0 and 2 in a dose of 100 ng per mouse.
The severity of clinical disease was assessed as follows:
0 = no signs of disease, 1 = limp tail, 2 = abnormal gait,
2.5 = one hindlimb paralyzed, 3 = complete hindlimb paraly-
sis, 3.5 = complete hindlimb paralysis and one forelimb par-
alyzed, 4 = tetraplegia and 5 = moribund or death. Mice
were monitored for a period of 40 days.
In the B6 model, 100 lg ATG or rabbit IgG was given i.p.
on days ?7 and ?3 prior to immunization with MOG peptide
for the preventive setting. In the therapeutic setting 100 lg
of ATG or rabbit IgG was given i.p. when the mice devel-
oped a disease score of 1 followed by another 100 lg 2
In the SJL/J model, 450 lg ATG or PBS was given i.p. on
days ?7 and ?3 prior to immunization with PLP peptide for
the preventive setting. In the therapeutic setting, 450 lg
ATG or PBS was given i.p. when the mice developed a dis-
ease score of 1 followed by another 450 lg 4 days later.
In vitro proliferation assay and cytokine ELISA
After immunization with MOG35–55/CFA, mice were given 100
lg ATG or rabbit IgG i.p. on days 1 and 3, following immuni-
zation. Spleen and lymph node cells were isolated on day 8
after immunization. CD4+T cells were purified using CD4
MACS beads (Miltenyi Biotech, Auburn, CA, USA) following
the manufacturer’s instructions.
CD4+T cells (1 3 105) and 4 3 105irradiated (3300 rad)
syngeneic splenic antigen-presenting cells (APCs) were
seeded onto 96-well round-bottom plates in DMEM/10%
FCS supplemented with 5 3 10?5M b-mercaptoethanol, 1 mM
sodium pyruvate, non-essential amino acids, L-glutamine
and 100 U penicillin/100 lg streptomycin per ml and
stimulated with different concentrations of anti-CD3 (clone
145.2C11, Bioexpress, West Lebanon, NH, USA) antibody
or MOG35–55peptide for 72 h. Supernatants were collected
after 48 h and analyzed for IL-4, IL-10, IFN-c and IL-17 by
ELISA. Cells were pulsed with 1 lCi of [3H]thymidine
([3H]TdR) for 18 h and harvested on glass fiber filters. Incor-
porated [3H]TdR was determined using a beta plate scintilla-
One hundred micrograms of ATG or rabbit IgG was injected
i.p. into MOG35–55 immunized Foxp3gfp.KI mice on days
0 and 3 after MOG immunization. Mice were sacrificed on
day 8. Single-cell suspensions were prepared from the
spleen and lymph nodes. CD4+GFP+T cells from both treat-
ment groups were FACS sorted and used as regulatory
cells. Naive cells (CD4+CD62LhighCD25?) from regular 2D2
(24) mice were used as responder T cells. Twenty thousand
responder cells and regulatory cells from either treatment
group were cultured in different ratios with 80 000 irradiated
(3300 rad) syngeneic splenic APCs per well in the presence
of 10 lg ml?1MOG35–55peptide or 1 lg ml?1anti-mouse
CD3 for 72 h. Cells were pulsed with 1 lCi of [3H]TdR for
the last 18 h of culture and incorporated [3H]TdR was deter-
mined using a beta plate scintillation counter.
Isolation of T cells from the Central nervous system
Mice were perfused with cold PBS through the left cardiac
ventricle. The brain was dissected and the spinal cord was
flushed out by hydrostatic pressure. Central nervous system
(CNS) tissue was cut into pieces and digested with
2.5 mg ml?1of Collagenase D (Roche Diagnostics, Indianapolis,
IN, USA) and 1 mg ml?1of DNAse I (Sigma–Aldrich, St Louis,
MO, USA) in DMEM medium at 37?C for 40 min. Single-cell
2 ATG expansion of MOG-specific Foxp3 regulatory cells
by guest on June 1, 2013
suspensions were prepared using a 70-lm cell strainer fol-
lowed by percoll gradient centrifugation (70/37%). Mononu-
clear cells were removed from the interphase, washed and
re-suspended in culture medium.
MOG tetramer staining
T cells (5 3 105) from whole spleen and lymph nodes of im-
munized Foxp3gfp.KI mice were isolated on day 8 and stim-
ulated with 10 lg ml?1of MOG peptide for 4 days in
medium containing IL-2. Mononuclear cells from the CNS
were purified and isolated at the peak of disease as de-
tetramers and control tetramers (TMEV
70-86/IAs) were generated as described (23, 25). Prior to
ex vivo tetramer staining, 1 3 107mononuclear cells per ml
from the CNS were treated with 0.7 U ml?1of neuraminidase
(Sigma–Aldrich) in DMEM for 30 min at 37?C. Cells were in-
cubated with the tetramers (30 lg ml?1) for 2.5 h in the dark
at room temperature in DMEM containing 5 lM IL-2 and 2%
FCS (pH 8.0). Cells were stained with anti-CD4–APC and
7-amino-actinomycin D (7-AAD) (PharMingen, San Diego,
CA, USA). The gate was set on live (7-AAD?) CD4+cells
and the percentages of MOG-specific cells in the Foxp3/
GFP+(T-regs) and Foxp3/GFP?(T-effs) compartment were
Administration of ATG in vivo preferentially depleted
We wished to test the effect of ATG on T cell populations
in vivo. Therefore, we administered 100 lg of ATG i.p. to
Foxp3gfp.KI mice on day 0 and day 3 and sacrificed the
mice on day 8. We observed that the CD4+population in
the spleen was decreased by ;75% following ATG treat-
ment. However, the percentage of T-regs (GFP+Foxp3+)
within the CD4+population increased from 12 to 20% result-
ing in an increased T-reg to T-eff ratio (Fig. 1). Similar results
were obtained when analyzing the lymph nodes of ATG-
treated mice. Here, the percentage of CD4 cells within the
T cell population decreased by ;20% with the percentage
of T-regs (GFP+Foxp3+) within the CD4+population increasing
from 7 to 32% after ATG treatment (data not shown). Thus,
administration of ATG results in the preferential depletion of
T-effs in vivo increasing the ratio of T-regs versus T-effs.
ATG treatment inhibited proliferation and cytokine production
of CD4+T cells
In order to investigate whether ATG treatment affected
antigen-specific recall responses to a class II restricted anti-
gen, we isolated CD4+T cells from the spleen and lymph
nodes of MOG-immunized mice treated with ATG or rabbit
IgG followed by re-stimulation in vitro in the presence of syn-
geneic APCs. CD4+T cells from ATG-treated mice prolifer-
ated less when stimulated with anti-CD3 or MOG peptide
(Fig. 2A) compared with the control groups. This also
resulted in a significantly reduced production of IFN-c and
IL-17 from the ATG-treated group while splenocytes from
the rabbit IgG control group produced huge amounts of
these two cytokines. The altered cytokine pattern resulted in
markedly decreased IFN-c to IL-10 and IL-17 to IL-10 ratios
in splenocytes derived from ATG-treated mice as compared
with controls (Fig. 2B). Collectively, these data suggested that
the altered balance of T-effs/T-regs in the ATG-treated animals
was functionally relevant for the response to auto-antigens.
ATG did not alter the function of residual CD4+Foxp3/GFP?
T-effs and CD4+Foxp3/GFP+T-regs
Next, we asked whether ATG treatment affected the function
of remaining CD4+Foxp3/GFP?T-effs. To test this, we FACS
sorted CD4+Foxp3/GFP?T cells from MOG-immunized mice
treated with either ATG or rabbit IgG and re-stimulated them
with different concentrations of MOG peptide in the pres-
ence of syngeneic APCs. When plating identical numbers of
responder cells, no difference in cell proliferation and cyto-
kine production was observed (data not shown). These
results suggested that ATG treatment did not alter the func-
tion of residual effector cells.
We also tested whether ATG treatment changed the ability
of T-regs to suppress polyclonal and antigen-specific T cell
responses. We performed suppression assays using re-
sponder cells from 2D2 MOG TCR transgenic mice and
added FACS-sorted CD4+Foxp3/GFP+T-regs from either
ATG- or rabbit IgG-treated MOG-immunized Foxp3gfp.
KI mice in different ratios. There was no difference in the ability
of T-regs from ATG-treated and rabbit IgG-treated mice in
suppressing 2D2 responder cells upon anti-CD3 or MOG35–55
stimulation (Fig. 3A and B). These data emphasize that
ATG treatment did not impair the ability of T-regs to suppress
polyclonal and antigen-specific responses of naive T cells.
ATG treatment conferred protection from EAE in the B6–MOG
model and SJL–PLP model
Given the preferential depletion of T-effs by ATG, we won-
dered whether administration of ATG conferred protection
Fig. 1. Numbers of CD4+Foxp3/GFP+T-regs and CD4+Foxp3/GFP?T-
effs in rabbit IgG-treated and ATG-treated mice. One hundred
micrograms of rabbit IgG or ATG was administered i.p. on days
0 and 3. CD4+cells from the spleen were purified on day 8 and Foxp3
expression was determined based on the percentage of GFP+cells
by FACS analysis. The ratio between GFP?(T-effs) versus GFP+(T-
regs) is shown. Data are based on three independent experiments
with two mice in each group.
ATG expansion of MOG-specific Foxp3 regulatory cells3
by guest on June 1, 2013
from EAE. We found that when ATG was administered prior
to immunization with MOG peptide (preventive treatment
setting), mice were relatively protected from disease. Although
some ATG-treated mice still developed EAE, the disease on-
set was delayed and the disease was significantly milder
(Fig. 4A). More importantly, when ATG was administered in
a therapeutic setting (i.e. after the onset of clinical signs of
disease), the disease did not progress and remained stable
throughout the observation period (Fig. 4B). To further
confirm this beneficial effect on EAE in the B6 model, we
also tested ATG treatment in a remitting relapsing mouse
model of EAE, the PLP139–151-induced EAE in SJL mice. Sim-
ilar to the results observed in the B6 model, pre-treatment
with ATG effectively decreased the incidence and resulted
in a milder disease course (Fig. 4C). Furthermore, therapeu-
tic administration of ATG also reduced the severity of the first
disease episode in this model (Fig. 4D). These data suggest
that ATG is effective in the treatment of EAE. However, the
timing for the administration of ATG in a therapeutic setting
appeared to be critical. When ATG was given at the peak of
disease when the inflammatory infiltrate in the target tissue
was already established, a beneficial effect of ATG was no
longer observed (data not shown).
ATG treatment resulted in the generation/expansion of
In order to investigate the basis for the reduction of clinical
disease of ATG-treated animals, we determined the frequency
of antigen-specific T-regs and T-effs using MOG35–55/IAb
Spleen and lymph node cells from immunized Foxp3gfp
KI mice were stimulated in vitro with MOG peptide and IL-2
for 4 days prior to MOG tetramer staining. In this antigen-
specific recall culture of lymphocytes from rabbit IgG-treated
control mice, MOG tetramers detected a distinct population
of antigen-specific T-effs and T-regs. The expansion of T-effs
was always dominant over the expansion of T-regs (Fig. 5A).
In contrast,in ATG-treatedmice, theexpansionof
Fig. 2. Proliferation and cytokine ratio in CD4+cells from rabbit IgG-
and ATG-treated mice. (A) CD4+cells from MOG-immunized and
rabbit IgG- or ATG-treated mice were purified using CD4 MACS
beads and stimulated with different concentrations of anti-CD3 or
MOG35–55in the presence of syngeneic APCs. (B) Cytokines were
measured by ELISA 48 h after stimulation. The ratios of IFN-c and IL-
17 to IL-10 were calculated and are shown in the diagrams. Data are
representative of three independent experiments. A 95% level of
confidence was used to calculate the error bars for the proliferation
Fig. 3. MOG suppression assay. CD4+Foxp3/GFP+Tcells from Foxp3
knock-in mice immunized with 100 lg of MOG peptide and treated
with rabbit IgG or ATG were isolated and combined with naive
responder cells from 2D2 mice in different T-reg/T-eff ratios. Cells were
cultured in the presence of 10 lg ml?1MOG35–55for 72 h. Proliferation
of responder T cells as measured by [3H]TdR incorporation is shown
for suppression assays using T reg from the spleen (A) or the lymph
4 ATG expansion of MOG-specific Foxp3 regulatory cells
by guest on June 1, 2013
antigen-specific T-regs largely prevailed over T-effs resulting
in T-eff versus T-reg ratios of 1:15 in favor of T-regs (Fig. 5A).
Similar results were also observed in recall cultures of
splenocytes (data not shown). This suggested that ATG
treatment changed the balance between T-effs and T-regs in
favor of T-regs, allowing for a remarkable antigen-driven
expansion of T-regs thus attenuating disease.
In order to confirm this observation, we wished to investi-
gate whether ATG induced a similar expansion of antigen-
specific T-regs in vivo without further in vitro manipulation of
antigen-primed T cells. Therefore, mononuclear cells were
isolated from the CNS of MOG-immunized and IgG- or ATG-
treated mice when the animals had reached the individual
peak disease score, and stained with MOG35–55/IAbtet-
ramers directly ex vivo. It has to be noted that it was uncom-
mon for a mouse treated with ATG to develop severe EAE.
The ATG-treated mice also recovered very quickly. So, the
timing for MOG tetramer analysis was critical. We found that
the ratio of MOG-specific T-effs to T-regs in the CNS of rabbit
IgG-treated mice was 30:1 as opposed to 1:1 in the ATG-
treated group (Fig. 5B). It is likely that this favorable ratio of
antigen-specific T-effs/T-regs in the CNS of ATG-treated mice
is responsible for inhibiting the establishment of a severe in-
flammatory infiltrate in the CNS of ATG-treated mice. It has
recently been shown that the ratio of antigen-specific T-effs/
T-regs in the CNS of EAE mice was 1:13 at the peak of dis-
ease and increased to 1:4 during remission, but also that
the highly inflammatory cytokine milieu in an established au-
toimmune infiltrate in the CNS might impair T-reg function at
the peak of disease (25). Thus, by manipulating the T-eff/
T-reg ratio through administration of ATG before the onset of
massive inflammation, autoimmune tissue destruction might
In the present study, we tested a T cell-depleting polyclonal
antibody (ATG) for its potency to treat organ-specific autoim-
munity. Combining the use of Foxp3gfp.KI mice with the
technology of tetramer staining, we created a model system
that allowed us to track auto-antigen-specific T-effs and
T-regs during EAE and thus determine differential effects of
ATG on these T cell subsets. ATG, which is already used in
humans to prevent transplant rejection, was very efficient in
attenuating autoimmune encephalomyelitis both in a preven-
tive and early therapeutic treatment regimen. We provide ev-
idence, that ATG preferentially depletes T-effs while T-regs
are resistant to depletion in vivo. Thus, upon immunization
with MOG in CFA, the expansion of antigen-specific T-regs
is favored over the expansion of T-effs resulting in robust
protection from clinical disease. Furthermore, even if the
ATG treatment is given only at the onset of clinical signs,
the differential susceptibility of T-effs is still sufficient to re-
duce the severity of disease.
Peritransplantational depletion of T cells has proven to be
a valuable way to prevent transplant rejection. A series of
studies on kinetics and mechanism of ATG and ALS implied
that these reagents induced a window for graft acceptance
simply by inducing generalized suppression of T cell
responses. However, recent studies indicated that various
CD4+subsets might have different susceptibilities to ATG-
induced depletion (20, 21). Here, T-regs might be particu-
larly resistant to ATG. We have previously determined the
kinetics of antigen-specific T-effs and T-regs in autoimmunity
and established that in the widely used model of MOG35–55/
CFA-induced EAE, MOG-specific T-regs are generated by
expansion of naturally occurring Foxp3-expressing T-regs in
Fig. 4. Mean EAE score of the control group and ATG-treated group in the C57Bl/6 and SJL model. (A) Preventive treatment setting for C57Bl/6.
One hundred micrograms of ATG or rabbit IgG was administered i.p. on day ?7 and ?3 prior to MOG immunization (100 lg) (n = 16). (B)
Therapeutic treatment setting for C57Bl/6. One hundred micrograms of ATG or rabbit IgG was administered i.p. twice 1 day apart as soon as the
mice showed the first signs of EAE (score of 0.5–1) (n = 13). (C) Preventive treatment setting for SJL/J. Four hundred and fifty micrograms of ATG
or PBS was administered i.p. on day ?7 and ?3 prior to PLP immunization (60 lg) (n = 10). (D) Therapeutic treatment setting for SJL/J. Four
hundred and fifty micrograms ATG or PBS was administered i.p. twice 4 days apart as soon as the mice showed the first signs of EAE (n = 8–9). A
95% level of confidence was used to calculate the error bars.
ATG expansion of MOG-specific Foxp3 regulatory cells5
by guest on June 1, 2013
the peripheral immune compartment (23). These T-regs then
traffic to the target tissue. The differential kinetics of accu-
mulation of myelin-specific T-effs and T-regs in the CNS re-
flect the clinical course of the disease in that the T-reg/T-eff
ratio in the CNS is 1:13 at the peak of disease, but 1:4 dur-
ing recovery. It was also evident that T-reg-mediated regula-
tion was more efficient in suppressing primary activation of
responder cells, but not antigen-specific recall responses of
MOG-specific T-effs isolated from the CNS (23). As a result,
in order for T-reg-based therapeutic approaches to be
successful, it is required to achieve a favorable ratio of anti-
gen-specific T-regs and T-effs early during the activation of
T-effs and before the onset of massive inflammation in the
target tissue. This was the rationale for the application of
ATG in EAE.
ATG treatment led to the depletion of the peripheral CD4+
T cell compartment by 75%. Since T-effs (CD4+Foxp3?) were
preferentially depleted, T-regs (CD4+Foxp3+) were relatively
enriched in the remaining CD4+population. Near-complete
T cell depletion involves the issue of homeostatic prolifera-
tion of the remaining T cell population. It has been shown
that T-effs that proliferate homeostatically (due to lymphope-
nia) behave no longer as naive T cells, but as recently
activated T cells or—after cessation of homeostatic prolifera-
tion—as memory-like T cells (26). This makes them resistant
to tolerizing treatment approaches, which rely on the induc-
tion of anergy by blocking of co-stimulatory pathways. This
is a severe conceptual setback for the clinically desirable
combination of T cell depletion and immunomodulation.
However, in vivo, T-regs are known to proliferate homeo-
statically to the same extent as T-effs (27). Thus, immunomo-
dulation that is based on T-regs should not be impaired,
especially in the case of ATG treatment, as the T-reg fraction
within the CD4+population is increased. In line with this con-
cept, the unsorted CD4-positive T cell population isolated
from lymph nodes and spleen of ATG-depleted animals that
were immunized with MOG35–55/CFA showed reduced recall
responses to MOG in terms of proliferation and production
of inflammatory cytokines as compared with CD4+cells from
control-treated animals. When equal numbers of FACS-
sorted T-effs (CD4+Foxp3?) were tested separately from the
accompanying T-reg fraction, the recall responses to MOG
were identical between the ATG and the control groups.
Therefore, it can be ruled out that priming of antigen-specific
T-effs on a single-cell basis is defective in ATG-treated ani-
mals. Nevertheless, ATG-treated mice were protected from
disease induction by immunization with MOG35–55/CFA. This
clearly indicates that the ATG-induced shift in balance to-
ward a more favorable T-reg/T-eff ratio was clinically relevant
in a setting of pending autoimmunity proving the hypothesis
that T-reg-mediated suppression of autoaggressive T cell
responses is well maintained after T cell depletion by ATG.
It is very likely that the enrichment of T-regs in the CD4 pop-
ulation by ATG is the prerequisite for the massive expansion
of MOG-specific T-regs after immunization with MOG35–55/
CFA. Thus, MOG immunization in ATG-treated animals favors
the expansion of fully functional MOG-specific T-regs in the
peripheral immune compartment and the CNS. Although we
have not formally ruled out in this study that the increase in
frequency of MOG-specific T-regs upon ATG treatment fol-
lowed by MOG sensitization is due to the conversion of
Foxp3?into Foxp3+T cells, we consider this unlikely. In fact,
we have previously shown that there is no conversion of
Foxp3?into Foxp3+T cells under inflammatory conditions as
generated by immunization with CFA as an adjuvant (23).
In the present study, we determined that ATG treatment is
also efficient in an early therapeutic setting, i.e. after onset
of clinical signs of disease, which is particularly relevant for
Fig. 5. MOG tetramer staining of peripheral lymphocytes and CNS-
derived mononuclear cells in rabbit IgG-treated and ATG-treated
mice. (A) Mice were immunized with MOG/CFA as described in
Methods. One hundred micrograms of rabbit IgG or ATG was
administered on days 1 and 3. Spleen and lymph node cells were
harvested on day 8 after immunization. Cells were re-stimulated with
10 lg ml?1of MOG peptide in the presence of exogenous IL-2 for
4 days prior to tetramer staining. (B) Mononuclear cells were isolated
from the CNS of MOG/CFA-immunized and rabbit IgG-treated or
preventively ATG-treated mice at the peak of disease followed by ex
vivo MOG tetramer staining. The gates are set on the living (7-AAD
negative) CD4+T cell population.
6 ATG expansion of MOG-specific Foxp3 regulatory cells
by guest on June 1, 2013
the treatment of human autoimmune diseases. We propose
that in this condition, the effect of ATG is due to the deple-
tion of T cells that have infiltrated the target tissue or are in
the process of trafficking to the CNS before the firm estab-
lishment of a severe autoimmune infiltrate. Based on our ki-
netic analysis of MOG-specific T cells in the natural course
of EAE, both T-effs and T-regs are already present in the
CNS at the onset of clinical signs. More importantly, the
peak of IL-17 production by T-effs in the CNS is early after
the onset of clinical signs of disease and it is believed that
IL-17 initiates an inflammatory cascade that impairs T-reg
function in situ (22, 23). Thus, in a therapeutic setting there
is only a limited time frame for the depletion of T-effs. In-
deed, when ATG was not administered at the onset, but at
the peak of disease, the severity and duration of disease
In theory, T-reg-based treatments of autoimmune diseases
have an appealing elegance. However, much remains to be
learned about the sites and modes of action of T-regs in or-
gan-specific autoimmune diseases. It is particularly unclear
whether transfer of ex vivo generated bona fide T-regs into
inflamed tissue would be beneficial for controlling autoim-
mune tissue damage. So far, therapeutic approaches have
therefore focused on shifting the balance between endoge-
nous T-effs and T-regs. Recently, a superagonistic CD28 an-
expansion of CD4+CD25+T-regs as opposed to T-effs in a se-
ries of pre-clinical models for autoimmune disease (28, 29)
and transplantation (29) went into phase I clinical trial. Unfor-
tunately, catastrophic side effects were observed that were
most likely related to the mitogenic and activating properties
of this reagent. The most promising results both in pre-clinical
models and human trials were achieved with non-mitogenic
antibodies like CD3 or depleting reagents like ATG or ALS.
In this study, we provide evidence that ATG treatment is not
only valid in transplantation but might also be extended to
autoimmune diseases. We confirmed that ATG leads to a pref-
erential depletion of T-effs and established that the protection
from organ-specific autoimmunity is due to the enhanced ex-
pansion of auto-antigen-specific T-regs during the priming
phase of the disease. In later disease phases, ATG treatment
is still efficient as long as massive inflammation in the target
organ is not yet established. Thus, in conclusion, ATG
endorses the development and maintenance of dominant tol-
erance in autoimmunity, which might be relevant for the treat-
ment of human autoimmune disease such as rheumatoid
arthritis and MS.
Deutsche Forschungsgemeinschaft (KO 2964/1-1) to T.K.
Foxp3gfp.KI Foxp3gfp knock-in
experimental autoimmune encephalomyelitis
myelin oligodendrocyte glycoprotein
effector T cell
regulatory T cell
type I diabetes
1 Kim, J., Rasmussen, J. and Rudensky, A. 2006. Regulatory Tcells
prevent catastrophic autoimmunity throughout the lifespan of
mice. Nat. Immunol. 8:191.
2 Sakaguchi,S. 2005.Naturally
CD25+CD4+ regulatory T cells in immunological tolerance to self
and non-self. Nat. Immunol. 6:345.
3 Trenado, A., Charlotte, F., Fisson, S. et al. 2003. Recipient-type
specific CD4+CD25+ regulatory T cells favor immune reconsti-
tution and control graft-versus-host disease while maintaining
graft-versus-leukemia. J. Clin. Invest. 112:1688.
4 Taylor, P. A., Lees, C. J. and Blazar, B. R. 2002. The infusion of
ex vivo activated and expanded CD4(+)CD25(+) immune regu-
latory cells inhibits graft-versus-host disease lethality. Blood
5 Edinger, M., Hoffmann, P., Ermann, J. et al. 2003. CD4+CD25+
regulatory T cells preserve graft-versus-tumor activity while
inhibiting graft-versus-host disease after bone marrow trans-
plantation. Nat. Med. 9:1144.
6 Salomon, B., Lenschow, D. J., Rhee, L. et al. 2000. B7/CD28
costimulation is essential for the homeostasis of the CD4+CD25+
immunoregulatory T cells that control autoimmune diabetes.
7 Kohm, A. P., Carpentier, P. A. and Miller, S. D. 2003. Regulation of
experimental autoimmune encephalomyelitis (EAE) by CD4+CD25+
8 Morgan, M. E., Sutmuller, R. P., Witteveen, H. J. et al. 2003. CD25+
cell depletion hastens the onset of severe disease in collagen-
induced arthritis. Arthritis Rheum. 48:1452.
9 Herold, K. C., Bluestone, J. A., Montag, A. G. et al. 1992.
Prevention of autoimmune diabetes with nonactivating anti-CD3
monoclonal antibody. Diabetes 41:385.
10 Herold, K. C., Hagopian, W., Auger, J. A. et al. 2002. Anti-CD3
monoclonal antibody in new-onset type 1 diabetes mellitus.
N. Engl. J. Med. 346:1692.
11 Herold, K. C. and Taylor, L. 2003. Treatment of Type 1 diabetes
with anti-CD3 monoclonal antibody: induction of immune regula-
tion? Immunol. Res. 28:141.
12 Kohm, A. P., Williams, J. S., Bickford, A. L. et al. 2005. Treatment
CD4+ T cell unresponsiveness and functional reversal of estab-
lished experimental autoimmune encephalomyelitis. J. Immunol.
13 Smith, J. A., Tso, J. Y., Clark, M. R., Cole, M. S. and Bluestone, J. A.
1997. Nonmitogenic anti-CD3 monoclonal antibodies deliver a
partial Tcell receptor signal and induce clonal anergy. J. Exp. Med.
14 Bluestone, J. A. and Boehmer, H. 2006. Regulatory Tcells. Semin.
15 Lowsky, R., Takahashi, T., Liu, Y. P. et al. 2005. Protective
conditioning for acute graft-versus-host disease. N. Engl. J.
16 Beiras-Fernandez, A., Thein, E. and Hammer, C. 2003. Induction
of immunosuppression with polyclonal antithymocyte globulins: an
overview. Exp. Clin. Transplant. 1:79.
17 Ogawa, N., Minamimura, K., Kodaka, T. and Maki, T. 2006. Short
administration of polyclonal anti-Tcell antibody (ALS) in NOD mice
with extensive insulitis prevents subsequent development of
autoimmune diabetes. J. Autoimmun. 26:225.
18 Genestier, L., Fournel, S., Flacher, M., Assossou, O., Revillard, J. P.
and Bonnefoy-Berard, N. 1998. Induction of Fas (Apo-1, CD95)-
mediated apoptosis of activated lymphocytes by polyclonal
antithymocyte globulins. Blood 91:2360.
ATG expansion of MOG-specific Foxp3 regulatory cells7
by guest on June 1, 2013
19 Preville, X., Flacher, M., LeMauff, B. et al. 2001. Mechanisms
involved in antithymocyte globulin immunosuppressive activity in
a nonhuman primate model. Transplantation 71:460.
20 Minamimura, K., Gao, W. and Maki, T. 2006. CD4+ regulatory
T cells are spared from deletion by antilymphocyte serum, a
polyclonal anti-T cell antibody. J. Immunol. 176:4125.
21 Lopez, M., Clarkson, M. R., Albin, M., Sayegh, M. H. and Najafian,
N. 2006. A novel mechanism of action for anti-thymocyte globulin:
induction of CD4+CD25+Foxp3+ regulatory T cells. J. Am. Soc.
22 Bettelli, E., Carrier, Y., Gao, W. et al. 2006. Reciprocal de-
velopmental pathways for the generation of pathogenic effector
TH17 and regulatory T cells. Nature 441:235.
23 Korn, T., Reddy, J., Gao, W. et al. 2007. Myelin-specific regulatory
T-cells accumulate in the central nervous system, but fail to
suppress pathogenic effector T-cells at the peak of autoimmune
inflammation. Nat. Med. 13:423.
24 Bettelli, E., Pagany, M., Weiner, H. L., Linington, C., Sobel, R. A.
and Kuchroo, V. K. 2003. Myelin oligodendrocyte glycoprotein-
specific T cell receptor transgenic mice develop spontaneous
autoimmune optic neuritis. J Exp Med 197:1073.
25 Reddy, J., Bettelli, E., Nicholson, L. et al. 2003. Detection of
autoreactive myelin proteolipid protein 139–151-specific Tcells by
using MHC II (IAs) tetramers. J. Immunol. 170:870.
26 Wu, Z., Bensinger, S. J., Zhang, J. et al. 2004. Homeostatic
proliferation is a barrier to transplantation tolerance. Nat. Med.
27 Gavin, M. A., Clarke, S. R., Negrou, E., Gallegos, A. and
Rudensky, A. 2002. Homeostasis and anergy of CD4(+)CD25(+)
suppressor T cells in vivo. Nat. Immunol. 3:33.
28 Schmidt, J., Elflein, K., Stienekemeier, M. et al. 2003. Treatment
and prevention of experimental autoimmune neuritis with super-
agonistic CD28-specific monoclonal antibodies. J. Neuroimmunol.
29 Beyersdorf, N., Gaupp, S., Balbach, K. et al. 2005. Selective
targeting of regulatory T cells with CD28 superagonists allows
effective therapy of experimental autoimmune encephalomyelitis.
J. Exp. Med. 202:445.
8 ATG expansion of MOG-specific Foxp3 regulatory cells
by guest on June 1, 2013