The Journal of Immunology
Arsenic Trioxide Prevents Murine Sclerodermatous
Niloufar Kavian,*,†,1Wioleta Marut,*,1Ame ´lie Servettaz,*,‡He ´le `ne Laude,*,x
Carole Nicco,* Christiane Che ´reau,* Bernard Weill,*,†and Fre ´de ´ric Batteux*,†
Chronic graft-versus-host disease (GVHD) follows allogeneic hematopoietic stem cell transplantation. It results from alloreactive
processes induced by minor MHC incompatibilities triggered by activated APCs, such as plasmacytoid dendritic cells (pDCs), and
leading to the activation of CD4 T cells. Therefore, we tested whether CD4+and pDCs, activated cells that produce high levels of
reactive oxygen species, could be killed by arsenic trioxide (As2O3), a chemotherapeutic drug used in the treatment of acute
promyelocytic leukemia. Indeed, As2O3exerts its cytotoxic effects by inducing a powerful oxidative stress that exceeds the lethal
threshold. Sclerodermatous GVHD was induced in BALB/c mice by body irradiation, followed by B10.D2 bone marrow and
spleen cell transplantation. Mice were simultaneously treated with daily i.p. injections of As2O3. Transplanted mice displayed
severe clinical symptoms, including diarrhea, alopecia, vasculitis, and fibrosis of the skin and visceral organs. The symptoms were
dramatically abrogated in mice treated with As2O3. These beneficial effects were mediated through the depletion of glutathione
and the overproduction of H2O2that killed activated CD4+T cells and pDCs. The dramatic improvement provided by As2O3in
the model of sclerodermatous GVHD that associates fibrosis with immune activation provides a rationale for the evaluation of
As2O3in the management of patients affected by chronic GVHD.
Chronic GVHD emerges from alloreactive processes between
donor-derived immune cells and host cell populations induced by
minor MHC incompatibilities between donor and recipient. Its
pathophysiology is poorly understood in contrast to that of acute
GVHD (1, 3). Donor CD4+T cells are involved in the induction of
chronic GVHD, but the effector mechanisms through which they
mediate tissue inflammation are unclear; the recognition of MHC
class II alloantigens on host dendritic cells (DCs) is sufficient to
prime donor CD4+T cells and induce GVHD (3–5). Moreover,
among APCs, plasmacytoid DCs (pDCs) were shown to be
pathogenic in GVHD. In the absence of other APCs, pDCs alone
can stimulate donor T cells to trigger GVHD, and total-body
irradiation is crucial for their maturation and the subsequent
priming of alloreactive CD4+T cells (6).
The Journal of Immunology, 2012, 188: 5142–5149.
hronic graft-versus-host disease (GVHD) is a major factor
of morbidity following allogeneic hematopoietic stemcell
transplantation, with variable clinical presentations (1, 2).
Chronic GVHD often mimics autoimmune diseases (7).
Sclerodermatous-GVHD (Scl-GVHD) makes up 10–15% of cases
of chronic GVHD (8). This clinical form of GVHD resembles
systemic sclerosis, because it includes fibrotic changes and
chronic inflammation of the skin, lung, and gastrointestinal tract.
Several animal models have been developed to help define the
pathophysiology of chronic GVHD. One of them is based on the
transfer of donor immune cells into sublethally irradiated host
mice with mismatched minor MHC histocompatibility Ags, re-
sulting in full donor lymphoid chimerism (9). This model recapitu-
lates the clinical features of Scl-GVHD, with fibrosis of the skin
and visceral organs 14 d following the graft.
Activated T cells with a high rate of production of reactive
GVHD. Therefore, we investigated whether a cytotoxic molecule
that acts by enhancing ROS production beyond a lethal threshold
could be of any help in treating chronic GVHD.
Arsenic trioxide (As2O3) is an inorganic trivalent salt that
exhibits potent antitumor effects in vitro and in vivo, especially in
the treatment of hematological malignancies, such as acute pro-
myelocytic leukemia refractory to all-trans retinoic acid (10).
Several reports suggested that As2O3can affect many cellular
functions, such as proliferation, apoptosis, differentiation, and
angiogenesis, in various cell lines. An important cellular event
occurring after As2O3treatment is the elevation of intracellular
ROS levels (11). This ROS generation appears to be regulated
through several pathways, including NADPH oxidase, mitochon-
drial electron transport chain, and inhibition of antioxidant en-
zymes (12–14). The ROS-mediated apoptosis triggered by As2O3
can impact hematological tumor cells; under certain circum-
stances, it can also affect nontumor cells, such as keratinocytes,
fibroblasts, or activated autoimmune lymphocytes (15–17).
In this study, we tested the therapeutic effects of As2O3 in
a murine model of Scl-GVHD generated by grafting B10.D2
(H-2d) bone marrow and spleen cells into sublethally irradiated
BALB/c mice (H-2d). We show that As2O3limits both the acti-
*Laboratoire EA 1833, Faculte ´ de Me ´decine, Sorbonne Paris Cite ´, Universite ´ Paris
Descartes, 75679 Paris Cedex 14, France;†Laboratoire d’Immunologie Biologique,
Ho ˆpital Cochin, Assistance Publique Ho ˆpitaux de Paris, 75014 Paris, France;
‡Faculte ´ de Me ´decine de Reims, Service de Me ´decine Interne, Maladies Infectieuses,
Immunologie Clinique, Ho ˆpital Robert Debre ´, 51092 Reims Cedex, France; and
xLaboratoire de Virologie, Ho ˆpital Cochin, Assistance Publique Ho ˆpitaux de Paris,
75014 Paris, France
1N.K. and W.M. contributed equally to this work.
Received for publication December 12, 2011. Accepted for publication March 9,
This work was supported by grants from Assistance Publique Ho ˆpitaux de Paris (to
N.K.) and Redcat (to W.M.).
Address correspondence and reprint requests to Prof. Fre ´de ´ric Batteux, Laboratoire
d’Immunologie EA1833, Faculte ´ de Me ´decine Paris Descartes, 75679 Paris Cedex
14, France. E-mail address: firstname.lastname@example.org
Abbreviations used in this article: As2O3, arsenic trioxide; BMT, bone marrow trans-
plantation; cDC, conventional dendritic cell; DC, dendritic cell; GSH, glutathione;
GVHD, graft-versus-host disease; MFI, mean fluorescence intensity; NAC, N-acetyl-
cysteine; pDC, plasmacytoid dendritic cell; ROS, reactive oxygen species; Scl-
GVHD, sclerodermatous graft-versus-host disease.
vation of the immune system and fibrosis in mice with Scl-GVHD.
As2O3induces apoptosis in alloreactive CD4+T cells and activated
pDCs, thus limiting the development of GVHD reaction in mice.
Materials and Methods
Animals, cells, and chemicals
Specific pathogen-free, 6-wk-old female BALB/c and male B10.D2 mice
water ad libitum. They were given humane care, according to the guidelines
of our institution (Universite ´ Paris Descartes). All cells were cultured as
previously reported (18). All chemicals were from Sigma-Aldrich (Saint-
Quentin Fallavier, France).
Experimental procedure in Scl-GVHD mice
Induction of GVHD in BALB/c mice. GVHD following bone marrow
transplantation (BMT) was induced in BALB/c mice (H-2d; Janvier Lab-
oratory, Le Genest Saint Isle, France) by grafting cells from 7–8-week-old
female B10.D2 mice (H-2d; Janvier Laboratory), as previously described
by Jaffee and Claman (19). Briefly, recipient mice were lethally irradiated
with 750 cGy from a Gammacel [137Cs] source. Three hours later, they
were injected i.v. with donor spleen cells (2 3 106/mouse) and bone
marrow cells (1 3 106/mouse) suspended in RPMI 1640. A control group
of BALB/c recipient mice received BALB/c spleen and bone marrow cells
(syngeneic BMT, referred to as control animals). Transplanted animals
were maintained in sterile microisolator cages (Lab Products, Langen-
selbold, Germany) and supplied with autoclaved food and sterile water.
Animals were sacrificed by cervical dislocation 4 wk after BMT.
Treatment of Scl-GVHD mice with As2O3. Scl-GVHD and control mice
were randomized and treated for 3 wk with i.p. injections of either As2O3or
vehicle alone beginning on day 7 post-BMT (10 mice/group). A stock
solution was prepared extemporaneously, as described above. As2O3was
given 5 d/wk at a dose of 5 mg/g body weight, as described by Bobe ´ et al.
(17). Control mice received i.p. injections of PBS 5 d/wk. Four weeks after
BMT, the animals were sacrificed by cervical dislocation.
Assessment of collagen accumulation
Skin thickness. Skin thickness of the shaved back of mice was measured
1 d before sacrifice with a caliper and expressed in millimeters.
Histopathological analysis. Fixed lung and skin pieces were embedded in
paraffin. A 5-mm-thick tissue section was stained with H&E or Picrosirius
Red. Slides were examined by standard bright-field microscopy (Olympus
BX60, Rungis, France) by a pathologist who was blinded to the assignment
of the animal group.
Collagen content in skin and lung. Skin and lung pieces were diced using
a scalpel, put into tubes, thawed, and mixed with pepsin (1:10 weight ratio)
and 0.5 M acetic acid overnight at room temperature under stirring. The
assay of collagen content was based on the quantitative dye-binding Sircol
method (Biocolor, Belfast, Ireland).
Disease severity score
To determine the incidence and severity of disease, we assigned a score to
each mouse using the following criteria: 0: no external sign; 1: piloerection
on back and underside, 1: hunched posture or lethargy; 1: weight loss
.10%; 0.5: alopecia ,1 cm2; 1: alopecia .1 cm2; 1: vasculitis (one or
more purpural lesions); and 1: eyelid sclerosis (blepharophymosis). The
severity score is the sum of these values and ranges from 0 (unaffected) to
a maximum of 6. The incidence and severity score was recorded every
week by two blinded scientists.
Flow cytometric analysis of spleen cell subsets
Cell suspensions from spleens were prepared after hypotonic lysis of
erythrocytes in potassium acetate solution. Cells were incubated with the
appropriate labeled Ab at 4˚C for 30 min in PBS with 0.1% sodium azide
and 5% normal rat serum. Flow cytometry was performed using a FACS-
Canto flow cytometer (BD Biosciences, San Jose, CA), according to
standard techniques. The mAbs used in this study were anti–B220-PE-
Cy7, anti–CD11b-biotin, anti–CD11c-FITC, anti–CD4-PerCP, and anti–
CD8-PE-Cy7 (BD Biosciences). Data were analyzed with FlowJo software
(Tree Star, Ashland, OR).
Determination of IL-4 and IL-17 production by splenocytes
Spleen cells were isolated by gentle disruption of the tissues, and the
erythrocytes were lysed by hypotonic shock in potassium acetate solution.
Spleen cells were cultured in RPMI 1640 supplemented with antibiotics,
GlutaMAX (Invitrogen Life Technologies), and 10% heat-inactivated FCS
(Invitrogen Life Technologies) (complete medium). CD4 T cells were
isolated from spleen cell suspensions by positive selection using CD4
microbeads and LS columns (Miltenyi Biotec, Paris, France), according to
the manufacturer’s instructions. CD4 T cell suspensions were then seeded
in 96-well flat-bottom plates and cultured (2 3 105cells) in complete
medium for 48 h in the presence of 5 mg/ml Con A. Supernatants were
collected, and cytokine concentrations were determined by ELISA (for
IL-4: eBioscience, San Diego, CA; for IL-17: DuoSet; R&D Systems).
Results are expressed in ng/ml.
Assays of serum anti-DNA topoisomerase 1 autoantibodies
Levels of anti-DNA topoisomerase 1 IgG Abs were detected by ELISA on
microtiter plates (ImmunoVision, Springdale, AR). A 1:50 serum dilution
was used for the ELISA.
Effects of As2O3on B10.D2 CD4 T cells in vitro
mouse were prepared after hypotonic lysis of erythrocytes. The BALB/c
cell suspension was irradiated at 30 Gy. A B10.D2 cell suspension was
also prepared and irradiated at 30 Gy for syngeneic control.
B10.D2 cells were labeled with PKH26 dye, according to the manu-
facturer’s instructions (Sigma-Aldrich).
Measurement of H2O2concentration in CD4 T cells. B10.D2 cells labeled
with PKH26 were incubated with irradiated BALB/c cells and complete
medium for 24 h. After this incubation period, cells were washed without
serum and incubated with 5 mM CM-H2DCFDA (Sigma-Aldrich) at 37˚C
for 20 min. After washing, cells were incubated with 10 mM As2O3, with
or without 4 mM N-acetylcysteine (NAC), for 5 h at 37˚C. Cells were then
washed and labeled with anti-CD4 mAb.
Measurement of glutathione concentration in CD4 T cells. B10.D2 cells
labeled with PKH26 were incubated with or without irradiated BALB/c
cells and complete medium, with or without 10 mM As2O3and with or
without 4 mM NAC, for 24 h. Cells were washed and labeled with 100 mM
monochlorobimane at 37˚C for 20 min, followed by labeling with anti-
CD4 mAb for 20 min.
Determination of apoptosis in CD4 T cells. B10.D2 cells labeled with
PKH26 were incubated with or without irradiated BALB/c spleen cells and
complete medium alone, with 10 mM As2O3alone, or with 4 mM NAC for
24 h. Cells were then washed and stained with anti-CD4 mAb (eBio-
science) at 4˚C for 20 min and Yopro-1 (Sigma-Aldrich) at room tem-
perature for 10 min.
Effects of As2O3on B10.D2 pDCs in vitro
A suspension of spleen cells from a male B10.D2 mouse was prepared after
hypotonic lysis of erythrocytes.
Measurement of H2O2concentration in pDCs. B10.D2 cells were incubated
in completemedium,washedwithoutserum,andincubatedwith 5mMCM-
H2DCFDA (Sigma-Aldrich) at 37˚C for 20 min. After washing, cells were
incubated with 10 mM As2O3, with or without 4 mM NAC, for 5 h at 37˚C.
Cells were then washed and labeled with anti-B220, anti-CD11b, and anti-
CD11c mAb (BD, Le Pont de Claix, France).
Measurement of glutathione concentration in pDCs. B10.D2 cells were
incubated with complete medium with or without 10 mM As2O3and with or
without 4 mM NAC for 24 h. Cells were then washed and labeled with 100
mM monochlorobimane at 37˚C for 20 min, followed by labeling with anti-
B220, anti-CD11b, and anti-CD11c mAb (BD) for 20 min.
Determination of apoptosis in pDCs. B10.D2 cells were incubated with
complete medium alone, with 10 mM As2O3alone, or with 10 mM As2O3
and 4 mM NAC for 24 h. Cells were then washed and stained with anti-
B220, anti-CD11b, and anti-CD11c mAb (BD) at 4˚C for 20 min and
Yopro-1 (Sigma-Aldrich) at room temperature for 10 min.
For all flow cytometry analyses, pDCs were defined as B220+CD11cint
CD11blow. Data were acquired on a FACSCanto II flow cytometer (BD
Biosciences) and analyzed with FlowJo software (Tree Star).
Measurement of in vitro IFN-a production by pDCs
pDCs were isolated from the spleen of a BALB/c mouse using the pDC
isolation kit and MS columns (Miltenyi Biotec). pDCs were then coated in
six-well plates and incubated with 10 mg/ml Gardiquimod (InvivoGen,
Toulouse, France) and increasing doses of As2O3(from 0 to 25 mM) or
with medium alone for 24 h. Supernatants were harvested, and IFN-a was
assayed as described by Dubois et al. (20). Briefly, L929 cells (5 3 104/
well; kindly provided by P. Lebon, Laboratoire de Virologie, Ho ˆpital
The Journal of Immunology5143
Cochin, Paris, France) were coated in 96-well plates and cultured in RPMI
1640 at 37˚C with 5% CO2until confluence. Culture media were dis-
carded, and plates were incubated for an additional 24 h with 50 ml 2-fold
serial dilutions (from 1:2 to 1:256 in RPMI 1640) of supernatant samples
or with serial dilutions of a standard solution of murine IFN-a. Thereafter,
vesicular stomatitis virus was added; the final concentration was 1:400 of
a stock solution previously shown to cause complete lysis of L929 cells at
a dilution of l:2000. We considered the dilution that destroyed half of the
cell layer at 24 h, and IFN-a units were determined by comparison with
cells incubated with 2-fold serial dilutions of mouse IFN-a (kindly pro-
vided by P. Lebon).
All of the quantitative data are expressed as mean 6 SEM and were an-
alyzed with Prism 5 (GraphPad), using one-way ANOVA or the Student t
test, as appropriate. A p value ,0.05 was considered statistically signif-
As2O3prevented clinical symptoms of systemic sclerosis
induced by GVHD
We investigated the effects of As2O3on the development of Scl-
GVHD, a fibrotic variant of GVHD. As2O3was administered daily
from day 7 following BMT, and mice were sacrificed 21 d later.
Lethally irradiated BALB/c mice transplanted with bone marrow
and spleen cells from B10.D2 mice developed skin fibrosis, as
shown by the measurement of ear skin thickening in Fig. 1A.
Control BALB/c animals with syngeneic grafts did not develop
skin fibrosis or GVHD (Fig. 2). In addition to skin thickening, the
engrafted animals displayed alopecia (100% of mice), vasculitis
(.80% of mice), or diarrhea (100% of mice). On day 21, the
disease severity score of Scl-GVHD mice was $5, whereas that of
control animals (syngeneic graft) was 0 (Fig. 2B). As2O3effec-
tively prevented severe GVHD; the mean severity score of treated
mice at day 14 was 2.5 6 0.2 versus 5 6 0.4 in untreated Scl-
GVHD mice (p , 0.001, Fig. 2B). Scl-GVHD mice treated with
As2O3displayed a reduction in skin thickness .40% compared
with untreated Scl-GVHD mice (p = 0.007, Fig. 1B). Moreover,
type 1 collagen content in the skin of Scl-GVHD mice treated
with As2O3was 45% lower than in untreated Scl-GVHD mice
(p = 0.002, Fig. 1C). Picrosirius Red staining of ear sections
showed important collagen deposits in the ears from Scl-GVHD
mice but not in those of Scl-GVHD mice treated with As2O3(Fig.
1A). Also, type 1 collagen concentration was higher in the lungs
of Scl-GVHD mice than in Scl-GVHD mice treated with As2O3
(data not shown). Altogether, these results show that As2O3limits
the deposition of collagen and prevents the development of alo-
pecia, vasculitis, and diarrhea in Scl-GVHD mice.
As2O3altered spleen cell subsets in Scl-GVHD mice
GVHD is caused by a donor T cell antihost reaction. We inves-
tigated whether the clinical improvement observed in Scl-GVHD
mice treated with As2O3correlated with quantitative or qualitative
alterations in spleen cell subsets. Flow cytometric analysis of
splenocytes showed that As2O3decreased the percentage of CD4+
T cells (p = 0.049 versus untreated Scl-GVHD mice). This re-
duction involved the CD44+CD62L2effector memory CD4 sub-
set, because the ratio of CD4+CD44highCD62Llow/CD4+CD44low
CD62Lhighcells was 15.7 6 4.6 in Scl-GVH mice versus 3.6 6
0.7 in Scl-GVHD mice treated with arsenic (p = 0.001, Fig. 3A–
C). Furthermore, the percentage of splenic pDCs, defined as
B220+CD11c+CD11blow, was three times lower in arsenic-treated
Scl-GVHD mice compared with untreated Scl-GVHD mice
(p = 0.021, Fig. 3D).
As2O3modified the splenic production of cytokines and the
serum levels of autoantibodies in Scl-GVHD mice
In addition, we explored the splenic production of IL-4 and IL-17,
two cytokines implicated in the development of GVHD in mice
(21). Scl-GVHD mice produced more IL-4 and IL-17 than did
control mice (IL-4: 0.24 6 0.023 for Scl-GVHD mice and 0.12 6
0.021 for control mice, p = 0.011; IL-17: 0.64 6 0.083 for Scl-
GVHD mice and 0.37 6 0.088 for control mice, p = 0.042; Fig.
3E, 3F). As2O3significantly reduced the production of the two
cytokines (p = 0.043 and p = 0.022 for IL-4 and IL-17, respec-
tively, versus nontreated mice, Fig. 3E, 3F). We then investigated
mice with Scl-GVHD. GVHD was induced by allo-
geneic transplant of B10.D2 bone marrow and spleen
cells into irradiated BALB/c mice. Control mice re-
ceived syngeneic transplants. As2O3was administered
from day 7 post-BMT at a dose of 5 mg/g. Mice were
sacrificed on day 28. Data from two independent ex-
periments were pooled (n = 10/group). (A) Representa-
tive Picrosirius Red staining of the skin (objective 310).
(B) Skin thickness measured with a caliper and ex-
pressed in millimeters. (C) Collagen content in skin of
mice expressed in mg/punch biopsy. Values are mean 6
SEM of data from all mice in the experimental or
control groups. *p , 0.05, paired Mann–Whitney U
As2O3reduces the fibrotic changes in
5144ARSENIC TRIOXIDE PREVENTS MURINE SCLERODERMATOUS GVHD
the presence of anti-DNA topoisomerase 1 Abs, a hallmark of
the Scl-GVHD model, which are generally detected 3–9 wk fol-
lowing the onset of disease (9). On the day of sacrifice, 80% of
Scl-GVHD mice and only 20% of Scl-GVHD mice treated with
As2O3were positive for anti–DNA-topoisomerase 1 Abs (p = 0.048,
As2O3triggered apoptosis of activated B10.D2 CD4+T cells
by enhancing ROS production
In vitro, ROS production was higher in B10.D2 CD4+T cells
stimulated with BALB/c splenocytes than in B10.D2 CD4+
T cells stimulated with B10.D2 splenocytes (syngeneic controls)
(mean fluorescence intensity [MFI] = 1721 6 40 versus 1186 6
92, p = 0.032). Treatment of stimulated B10.D2 CD4+T cells with
As2O3further increased their production of H2O2that reached an
MFI of 2172 6 103 (p = 0.031 versus stimulated B10.D2 CD4+
T cells without As2O3, Fig. 4A). The enhancement of ROS pro-
duction was abrogated by incubation with 4 mM NAC (MFI: 1636 6
87, p = 0.021 versus As2O3alone, Fig. 4A). The level of gluta-
thione (GSH) in B10.D2 CD4+T cells was in accordance with
those results. Syngeneic stimulated cells displayed elevated levels
of GSH (mean of 45 6 2% of GSH+cells), whereas allogeneic
stimulated cells had a slight decrease in their GSH content (mean
31 6 2.5% of positive cells) (p = 0.041, Fig. 4B). Incubation with
10 mM As2O3dramatically decreased the GSH content in stimu-
lated B10.D2 CD4+T cells (Fig. 4B). In addition, Yopro-1 staining
of activated B10.D2 CD4+T cells indicated that arsenic dramat-
ically triggered apoptosis in those cells. Apoptosis correlated with
ROS production measured by flow cytometry and was downreg-
ulated by incubation with 4 mM NAC (p = 0.0009, Fig. 4C).
of Scl-GVHD in mice. GVHD was induced by allo-
geneic transplant of B10.D2 bone marrow and spleen
cells into irradiated BALB/c mice. Control mice re-
ceived syngeneic transplants. As2O3was administered
from day 7 post-BMT at a dose of 5 mg/g. Mice were
sacrificed on day 28. Data from two independent
experiments were pooled. (A) Representative photo-
graphs of mice on day 28 post-BMT (n = 10/group).
(B) Disease severity scores (mean 6 SEM).
As2O3improves the clinical symptoms
pDCs. Mice were treated with As2O3from day 7 post-BMT at a dose of 5 mg/g (n = 10/group). Spleen cells were harvested on day 28 post-BMT. (A)
Numbers of CD4+naive T cells in untreated and treated Scl-GVHD mice. (B) Numbers of CD4+effector memory T cells. (C) Effector memory/naive CD4+
T cell ratio. (D) Numbers of splenic pDCs in untreated and treated Scl-GVHD mice. (E) Numbers of splenic cDCs in untreated and treated Scl-GVHD mice.
(F) Percentages of CD86+pDCs in untreated and treated Scl-GVHD mice. Values are mean 6 SEM of data from all mice in the experimental or control
groups. *p , 0.05, **p , 0.01, ***p , 0.001, paired Mann–Whitney U test.
Treatment of Scl-GVHD mice with As2O3altered the balance between memory and naive CD4+T cells and decreased numbers of splenic
The Journal of Immunology5145
As2O3also triggered apoptosis of pDCs by enhancing H2O2
We observed the same effects on ROS production and apoptosis
in B10.D2 pDCs incubated with 10 mM As2O3. Indeed, the basal
detection MFI of CM-H2DCFDA by flow cytometry was 5405 6
12 for B10.D2 pDCs and 6527 6 25 when treated with As2O3.
Adding NAC reduced H2O2production by pDCs, as shown by the
CM-H2DCFDA MFI of 3675 6 15 (Fig. 5). Monochlorobimane
staining was also reduced in pDCs treated with arsenic compared
with untreated pDCs (MFI = 707 6 10 versus 1583 6 80),
whereas coaddition of NAC and As2O3increased the intracellular
content of GSH (MFI = 1541 6 8). The effects of As2O3were also
studied on conventional DCs (cDCs). There was a tendency to-
ward an increase in H2O2production and a decrease in GSH
content in cDCs upon treatment with arsenic, but those results did
not reach significance (for H2O2levels: MFI = 1,118 6 90 for
cDCs alone, 1,445 6 86 for cDCs + As2O3, 1,325 6 76 for
cDCs + As2O3+ NAC; p = 0.08 for cDCs versus cDCs + As2O3,
p = 0.09 for cDCs + As2O3+ NAC versus cDCs + As2O3; for
GSH levels: MFI = 24,308 6 395 for cDCs alone, 21,719 6 384
for cDCs + As2O3, 23,000 6 376 for cDCs + As2O3+ NAC;
p = 0.089 for cDCs versus cDCs + As2O3, p = 0.10 for cDCs +
As2O3+ NAC versus cDCs + As2O3).
As2O3blocked IFN-a production of B10.D2 pDCs
To assess the specificity of As2O3against pDCs, we investigated
its effect on the production of IFN-a by splenic pDCs, with and
without activation by the TLR7 agonist Gardiquimod. Fig. 6
shows that, in the absence of stimulation of pDCs, As2O3has a
significant effect on the production of IFN-a only at the highest
concentrations tested (10 and 25 mM). In contrast, after stimula-
tion of TLR7 and incubation with 10 and 25 mM As2O3, the levels
of IFN-a in cell supernatants were strongly decreased (p = 0.002
for 10 mM As2O3). Incubation with 5 mM As2O3decreased the
concentration of IFN-a in the supernatants .2-fold, whereas
lower doses of As2O3had no effect on IFN-a production (Fig. 7).
This study shows that As2O3selectively deletes activated CD4+
T cells and pDCs that have low levels of GSH and overproduce
H2O2and, thus, ameliorates Scl-GVHD in mice.
We tested the effects of As2O3, a chemotherapeutic drug used in
hematological malignancies, on the development of Scl-GVHD in
mice. This model of chronic GVHD shares typical features with
systemic sclerosis, including skin and visceral fibrosis and auto-
immune manifestations (22). As2O3dramatically improved the
clinical outcome of sublethally irradiated BALB/c mice trans-
planted with B10.D2 hematopoietic cells; weight loss, fibrosis,
vasculitis, and alopecia were markedly reduced in treated versus
The percentages of effector memory CD4+T cells (CD4+
CD44highCD62Llow) decreased in mice with Scl-GVHD that were
treated with arsenic. The pathophysiology of chronic GVHD
remains poorly understood, although a large amount of evidence
suggests that, in contrast to acute GVHD, which is dependent on
CD8+T cells, the manifestations observed in chronic GVHD are
dependent on the activation of minor histocompatibility Ag-
specific donor CD4+T cells (4, 23–26). After transplantation of
B10.D2 lymphoid cells into irradiated BALB/c mice, naive donor
CD4+T cells initiate the disease. As a result, donor CD4+T cells
infiltrate the skin, recruit macrophages and monocytes, and induce
fibrosis and destructive changes. Because activated CD4+T cells
play a pivotal role in the induction of the disease and are de-
creased by in vivo treatment with arsenic, we investigated the
mechanism of action of arsenic on those cells. We show that B10.
D2 CD4+T cells stimulated with irradiated BALB/c spleen cells
display lower GSH contents and produce higher levels of H2O2
than do unstimulated CD4+T cells. These results are in agreement
with previous studies showing that, upon activation, T cells
overproduce ROS (23, 24). Then, we showed that the high lev-
els of ROS production by activated CD4+T cells make them
hypersensitive to arsenic-induced apoptosis. Indeed, in vitro treat-
ment with arsenic induced an important decrease in GSH content
and a subsequent increase in H2O2levels beyond a lethal thresh-
old, inducing cell apoptosis. These data confirm the role of the
oxidant/antioxidant balance as a crucial factor that determines cell
susceptibility to arsenic (25). Several studies demonstrated that
GSH can bind arsenic from attacking its target by formation of
a transient As(GS)3complex and that GSH depletion in acute
promyelocytic leukemia cells synergizes with As2O3in the in-
duction of apoptosis (27).
We next investigated whether an alteration in the profile of
cytokine production by splenocytes frommicewithScl-GVHD and
treated with As2O3could reflect changes in splenic T cell pop-
ulations. Splenic IL-17 production was lower in arsenic-treated
Scl-GVHD mice than in untreated mice. These data are consis-
tent with a large number of recent studies conducted in mice that
concluded that Th17 cells are implicated in GVHD development
from day 7 post-BMTat a dose of 5 mg/g (n = 10/group). Spleen cells were harvested on day 28 post-BMT, and CD4 T cells were purified as described in
Materials and Methods. (A) IL-4 secretion measured in supernatants of CD4 T cells by ELISA (ng/ml). (B) IL-17 secretion measured in supernatants of
CD4 T cells by ELISA (ng/ml). (C) Anti–DNA-topoisomerase 1 autoantibody concentrations in the sera (A.U.). Values are mean 6 SEM of data from all
mice in the experimental or control groups. *p , 0.05, paired Mann–Whitney U test.
In vivo treatment with As2O3reduced the production of IL-4, IL-17, and autoantibodies in Scl-GVHD mice. Mice were treated with As2O3
5146ARSENIC TRIOXIDE PREVENTS MURINE SCLERODERMATOUS GVHD
in mice. Among them, a study reported that amplification of IL-17
production by the use of the stem cell mobilization factor G-CSF
leads to a cutaneous fibrosis occurring late after the graft, as in
Scl-GVHD (21). Consistent with those data, another recent article
stated that the use of an anti–IL-17 mAb can ameliorate skin
symptoms in chronic GVHD (20, 26). Moreover, Nishimori et al.
(28) recently showed the beneficial effects of the synthetic retinoid
Am80, which belongs to the same family as all-trans retinoic
acid, on chronic GVHD. These effects are mediated through the
downregulation of Th17 cells. Because other synthetic retinoids,
such as N-(4-hydroxyphenyl)retinamide, can induce apoptosis
through increased production of ROS, it is possible that Am80
also acts on Th17 cells through the induction of ROS, because
As2O3does in our model (29).
irradiated B10.D2 splenocytes (↓B10, syngeneic control) or BALB splenocytes (↓BALB) and treated or not with 10 mM As2O3with or without NAC. Flow
cytometry analysis was gated on CD4+T cells. Results are representative of four experiments carried out in duplicates. Data were analyzed with FlowJo
software. (A) Increase in H2O2generation by As2O3, measured by flow cytometry using CM-H2DCFDA. (B) GSH content in B10.D2 CD4 T cells,
measured by flow cytometry using monochlorobimane staining. (C) Induction of apoptosis by As2O3, measured by flow cytometry using Yopro staining.
Mean values were compared using paired Mann–Whitney U tests.
As2O3induced apoptosis of activated B10.D2 CD4+T cells in culture through ROS production. B10D2 spleen cells were incubated with
The Journal of Immunology 5147
The decrease in splenic CD4+effector memory cells in Scl-
GVHD mice treated with As2O3also correlates with a reduction
in the Th2 cytokine IL-4 produced in vitro by activated spleno-
cytes. Following the graft of B10.D2 spleen and bone marrow
cells into sublethally irradiated BALB/c mice, Zhou et al. (30)
observed an increase in the expression of type 2 cytokines in the
skin of Scl-GVHD mice compared with syngeneic grafts.
Moreover, in other chronic GVHD models, type 2 polarized im-
mune responses are required for the induction of skin GVHD in
mice and the development of fibrosis in the skin and visceral
organs (31). Thus, the decreased production of IL-4 observed
in our study probably contributes to the improvement of skin
As a whole, the alterations in splenic production of cytokines in
our model are consecutive to a reduced immune activation after
treatment with arsenic and, thus, contribute to the amelioration of
Chronic GVHD is associated with other autoimmune manifes-
tations, such as the production of autoantibodies in relation to the
production of Th2 cytokines. As described by others, we observed
the production of anti–DNA-topoisomerase 1 Abs in mice with
Scl-GVHD. The levels of these autoantibodies were decreased by
As2O3in our model. Similar effects of As2O3were reported in
a lupus mouse model (MRL/lpr mice), with a decrease in the
production of autoantibodies (anti-dsDNA and rheumatoid fac-
tors) (17). Consistent with these data, we conclude that As2O3
24 h. Flow cytometry analysis was gated on pDCs subsets defined as B220+CD11cintCD11blow. Results are representative of four experiments carried out in
duplicates. Data were analyzed with FlowJo software. (A) Induction of ROS formation by As2O3, measured by flow cytometry using CM-H2DCFDA. (B)
Arsenic induces a decrease in the GSH content in pDCs, measured by flow cytometry using monochlorobimane staining. (C) Induction of apoptosis in pDCs
by As2O3, measured by flow cytometry using Yopro staining.
Effects of in vitro treatment with As2O3on B10.D2 pDCs. B10.D2 spleen cells were incubated with 10 mM As2O3with or without NAC for
production by those cells. pDCs were seeded in six-
well plates and coincubated with 10 mg/ml Gardi-
quimod and/or increasing doses (0–25 mM) of As2O3.
*p , 0.05.
As2O3targets pDCs and blocks IFN-a
5148ARSENIC TRIOXIDE PREVENTS MURINE SCLERODERMATOUS GVHD
prevents the development of an autoimmune reaction in Scl- Download full-text
GVHD mice by specifically targeting CD4+T cells.
In contrast, the role of APCs in chronic GVHD was recently
emphasized by the observation that costimulation of donor T cells
through CD80 or CD86 on either host or donor APCs is necessary
to induce chronic GVHD (32). pDCs are especially involved in the
physiopathology of GVHD because the adoptive transfer of ma-
ture pDCs exacerbates GVHD, and they can stimulate donor
T cells to trigger GVHD in the absence of other APCs (6, 33). Yet,
the state of maturation of pDCs seems to be crucial in their ability
to trigger GVHD. On the one hand, total-body irradiation induces
inflammation and maturation of pDCs, which can subsequently
activate T cells (6). On the other hand, Banovic et al. (34) showed
that precursors of pDCs, but not mature pDCs, can attenuate the
symptoms of GVHD, emphasizing the differential functions of
pDCs depending on the environment and the model of GVHD.
In our hands, the pDC subset was decreased in treated mice
compared with untreated mice. As observed for CD4+T cells, the
increased levels of H2O2production, along with the decreased
GSH content, render pDCs sensitive to As2O3-induced apoptosis.
The action of arsenic on pDCs was confirmed in vitro by the
decrease in IFN-a production following stimulation of pDCs by
a TLR-7 agonist. This selective effect of arsenic on pDCs could
also lead to interesting insights about IFN-a–related diseases, such
as systemic lupus erythematosus. Indeed, Bobe ´ et al. (17) showed
the beneficial effects of As2O3in MRL/lpr lupus-prone mice. In
accordance with our results, they reported a cytotoxic effect of
As2O3on activated CD4+T cells through the reduction in GSH
levels, but the beneficial effects observed in their model could also
be mediated through regulation of IFN-a production by pDCs.
In summary, our work highlights the beneficial effects of As2O3
in chronic GVHD. As2O3could be a therapeutic tool in hemato-
logic and solid malignancies, as well as in chronic GVHD (10).
We thank P. Lebon for kindly providing L229 cells and IFN-a standard
solution and A. Colle for typing the manuscript.
The authors have no financial conflicts of interest.
1. Lee, S. J., J. P. Klein, A. J. Barrett, O. Ringden, J. H. Antin, J. Y. Cahn,
M. H. Carabasi, R. P. Gale, S. Giralt, G. A. Hale, et al. 2002. Severity of chronic
graft-versus-host disease: association with treatment-related mortality and re-
lapse. Blood 100: 406–414.
2. Baird, K., and S. Z. Pavletic. 2006. Chronic graft versus host disease. Curr. Opin.
Hematol. 13: 426–435.
3. Coghill, J. M., S. Sarantopoulos, T. P. Moran, W. J. Murphy, B. R. Blazar, and
J. S. Serody. 2011. Effector CD4+ T cells, the cytokines they generate, and
GVHD: something old and something new. Blood 117: 3268–3276.
4. Korngold, R., and J. Sprent. 1978. Lethal graft-versus-host disease after bone
marrow transplantation across minor histocompatibility barriers in mice. Pre-
vention by removing mature T cells from marrow. J. Exp. Med. 148: 1687–1698.
5. Duffner, U. A., Y. Maeda, K. R. Cooke, P. Reddy, R. Ordemann, C. Liu,
J. L. Ferrara, and T. Teshima. 2004. Host dendritic cells alone are sufficient to
initiate acute graft-versus-host disease. J. Immunol. 172: 7393–7398.
6. Koyama, M., D. Hashimoto, K. Aoyama, K. Matsuoka, K. Karube, H. Niiro,
M. Harada, M. Tanimoto, K. Akashi, and T. Teshima. 2009. Plasmacytoid
dendritic cells prime alloreactive T cells to mediate graft-versus-host disease as
antigen-presenting cells. Blood 113: 2088–2095.
7. Filipovich, A. H., D. Weisdorf, S. Pavletic, G. Socie, J. R. Wingard, S. J. Lee,
P. Martin, J. Chien, D. Przepiorka, D. Couriel, et al. 2005. National Institutes of
Health consensus development project on criteria for clinical trials in chronic
graft-versus-host disease: I. Diagnosis and staging working group report. Biol.
Blood Marrow Transplant. 11: 945–956.
8. Vogelsang, G. B., L. Lee, and D. M. Bensen-Kennedy. 2003. Pathogenesis and
treatment of graft-versus-host disease after bone marrow transplant. Annu. Rev.
Med. 54: 29–52.
9. Ruzek, M. C., S. Jha, S. Ledbetter, S. M. Richards, and R. D. Garman. 2004. A
modified model of graft-versus-host-induced systemic sclerosis (scleroderma)
exhibits all major aspects of the human disease. Arthritis Rheum. 50: 1319–
10. Sanz, M. A., D. Grimwade, M. S. Tallman, B. Lowenberg, P. Fenaux,
E. H. Estey, T. Naoe, E. Lengfelder, T. Bu ¨chner, H. Do ¨hner, et al. 2009. Man-
agement of acute promyelocytic leukemia: recommendations from an expert
panel on behalf of the European LeukemiaNet. Blood 113: 1875–1891.
11. Miller, W. H., Jr., H. M. Schipper, J. S. Lee, J. Singer, and S. Waxman. 2002.
Mechanisms of action of arsenic trioxide. Cancer Res. 62: 3893–3903.
12. Lynn, S., J. R. Gurr, H. T. Lai, and K. Y. Jan. 2000. NADH oxidase activation is
involved in arsenite-induced oxidative DNA damage in human vascular smooth
muscle cells. Circ. Res. 86: 514–519.
13. Smith, K. R., L. R. Klei, and A. Barchowsky. 2001. Arsenite stimulates plasma
membrane NADPH oxidase in vascular endothelial cells. Am. J. Physiol. Lung
Cell. Mol. Physiol. 280: L442–L449.
14. Pelicano, H., L. Feng, Y. Zhou, J. S. Carew, E. O. Hileman, W. Plunkett,
M. J. Keating, and P. Huang. 2003. Inhibition of mitochondrial respiration: a novel
strategy to enhance drug-induced apoptosis in human leukemia cells by a reactive
oxygen species-mediated mechanism. J. Biol. Chem. 278: 37832–37839.
15. Tse, W. P., C. H. Cheng, C. T. Che, and Z. X. Lin. 2008. Arsenic trioxide, arsenic
pentoxide, and arsenic iodide inhibit human keratinocyte proliferation through
the induction of apoptosis. J. Pharmacol. Exp. Ther. 326: 388–394.
16. Chen, Y. C., S. Y. Lin-Shiau, and J. K. Lin. 1998. Involvement of reactive oxygen
species and caspase 3 activation in arsenite-induced apoptosis. J. Cell. Physiol.
17. Bobe ´, P., D. Bonardelle, K. Benihoud, P. Opolon, and M. K. Chelbi-Alix. 2006.
Arsenic trioxide: A promising novel therapeutic agent for lymphoproliferative
and autoimmune syndromes in MRL/lpr mice. Blood 108: 3967–3975.
18. Kavian, N., A. Servettaz, C. Mongaret, A. Wang, C. Nicco, C. Che ´reau,
P. Grange, V. Vuiblet, P. Birembaut, M. D. Diebold, et al. 2010. Targeting
ADAM-17/notch signaling abrogates the development of systemic sclerosis in
a murine model. Arthritis Rheum. 62: 3477–3487.
19. Jaffee, B. D., and H. N. Claman. 1983. Chronic graft-versus-host disease
(GVHD) as a model for scleroderma. I. Description of model systems. Cell.
Immunol. 77: 1–12.
20. Dubois, M. F., V. Mezger, M. Morange, C. Ferrieux, P. Lebon, and O. Bensaude.
1988. Regulation of the heat-shock response by interferon in mouse L cells. J.
Cell. Physiol. 137: 102–109.
21. Hill, G. R., S. D. Olver, R. D. Kuns, A. Varelias, N. C. Raffelt, A. L. Don,
K. A. Markey, Y. A. Wilson, M. J. Smyth, Y. Iwakura, et al. 2010. Stem cell
mobilization with G-CSF induces type 17 differentiation and promotes sclero-
derma. Blood 116: 819–828.
22. McCormick, L. L., Y. Zhang, E. Tootell, and A. C. Gilliam. 1999. Anti-TGF-
beta treatment prevents skin and lung fibrosis in murine sclerodermatous graft-
versus-host disease: a model for human scleroderma. J. Immunol. 163: 5693–
23. Chu, Y. W., and R. E. Gress. 2008. Murine models of chronic graft-versus-host
disease: insights and unresolved issues. Biol. Blood Marrow Transplant. 14:
24. Shlomchik, W. D. 2007. Graft-versus-host disease. Nat. Rev. Immunol. 7: 340–
25. Kappel, L. W., G. L. Goldberg, C. G. King, D. Y. Suh, O. M. Smith, C. Ligh,
A. M. Holland, J. Grubin, N. M. Mark, C. Liu, et al. 2009. IL-17 contributes to
CD4-mediated graft-versus-host disease. Blood 113: 945–952.
26. Kansu, E. 2004. The pathophysiology of chronic graft-versus-host disease. Int. J.
Hematol. 79: 209–215.
27. Davison, K., S. Co ˆte ´, S. Mader, and W. H. Miller. 2003. Glutathione depletion
overcomes resistance to arsenic trioxide in arsenic-resistant cell lines. Leukemia
28. Nishimori, H., Y. Maeda, T. Teshima, H. Sugiyama, K. Kobayashi, Y. Yamasuji,
S. Kadohisa, H. Uryu, K. Takeuchi, T. Tanaka, et al. 2012. Synthetic retinoid
Am80 ameliorates chronic graft-versus-host disease by down-regulating Th1 and
Th17. Blood 119: 285–295.
29. Asumendi, A., M. C. Morales, A. Alvarez, J. Are ´chaga, and G. Pe ´rez-Yarza.
2002. Implication of mitochondria-derived ROS and cardiolipin peroxidation in
N-(4-hydroxyphenyl)retinamide-induced apoptosis. Br. J. Cancer 86: 1951–
30. Zhou, L., D. Askew, C. Wu, and A. C. Gilliam. 2007. Cutaneous gene expression
by DNA microarray in murine sclerodermatous graft-versus-host disease,
a model for human scleroderma. J. Invest. Dermatol. 127: 281–292.
31. Wynn, T. A. 2004. Fibrotic disease and the T(H)1/T(H)2 paradigm. Nat. Rev.
Immunol. 4: 583–594.
32. Anderson, B. E., J. M. McNiff, D. Jain, B. R. Blazar, W. D. Shlomchik, and
M. J. Shlomchik. 2005. Distinct roles for donor- and host-derived antigen-
presenting cells and costimulatory molecules in murine chronic graft-versus-
host disease: requirements depend on target organ. Blood 105: 2227–2234.
33. MacDonald, K. P., V. Rowe, C. Filippich, R. Thomas, A. D. Clouston,
J. K. Welply, D. N. Hart, J. L. Ferrara, and G. R. Hill. 2003. Donor pretreatment
with progenipoietin-1 is superior to granulocyte colony-stimulating factor in
preventing graft-versus-host disease after allogeneic stem cell transplantation.
Blood 101: 2033–2042.
34. Banovic, T., K. A. Markey, R. D. Kuns, S. D. Olver, N. C. Raffelt, A. L. Don,
M. A. Degli-Esposti, C. R. Engwerda, K. P. MacDonald, and G. R. Hill. 2009.
Graft-versus-host disease prevents the maturation of plasmacytoid dendritic
cells. J. Immunol. 182: 912–920.
The Journal of Immunology5149