ARTHRITIS & RHEUMATISM
Vol. 54, No. 10, October 2006, pp 3135–3143
© 2006, American College of Rheumatology
CTLA-4Ig Suppresses Reactive Oxygen Species by Preventing
Synovial Adherent Cell–Induced Inactivation of Rap1,
a Ras Family GTPase Mediator of Oxidative Stress in
Rheumatoid Arthritis T Cells
P. H. J. Remans,1C. A. Wijbrandts,1M. E. Sanders,1R. E. Toes,2F. C. Breedveld,2P. P. Tak,1
J. M. van Laar,2and K. A. Reedquist1
Objective. Oxidative stress contributes to the in-
flammatory properties of rheumatoid arthritis (RA)
synovial T lymphocytes. This study was undertaken to
investigate the mechanisms leading to production of
reactive oxygen species (ROS) and oxidative stress in
RA synovial T lymphocytes.
Methods. ROS production in T lymphocytes from
the peripheral blood (PB) of healthy donors and from
the PB and synovial fluid (SF) of RA patients was
measured by ROS-dependent fluorescence of 6-carboxy-
2?,7?-dichlorofluorescein. Rap1 GTPase activation was
assessed by activation-specific probe precipitation. Pro-
liferation of RA PB and SF T lymphocytes was assayed
RA PB T cells were preincubated with autologous SF or
with PB or SF adherent cells. Experiments were per-
formed in the absence or presence of transwell mem-
branes or CTLA-4Ig fusion proteins. Short- and long-
term stimulations of healthy donor PB T lymphocytes
were performed with inflammatory cytokines, in the
absence or presence of activating anti-CD28 antibodies.
Results. T lymphocyte ROS production and Rap1
inactivation were mediated by cell–cell contact with RA
3H-thymidine incorporation. In some experiments,
synovial adherent cells, and this correlated with T cell
mitogenic hyporesponsiveness. CTLA4-Ig blockade of
synovial adherent cell signaling to CD28 T cells reversed
the inhibition of Rap1 activity and prevented induction
of ROS. Introduction of active RapV12 into T cells also
prevented induction of ROS production. Coincubation
of T cells with stimulating anti-CD28 antibodies and
inflammatory cytokines synergistically increased T cell
Conclusion. Cell–cell contact between T cells and
RA synovial adherent cells mediates Rap1 inactivation
and subsequent ROS production in T lymphocytes
following exposure to inflammatory cytokines. This pro-
cess can be blocked by CTLA4-Ig fusion protein.
In vitro studies, animal disease models, and clin-
ical studies have all provided strong evidence that T
lymphocytes perpetuate inflammation and eventual joint
destruction in the rheumatoid arthritis (RA) synovial
joint (1–3). T lymphocytes derived from RA synovial
tissue display markers of recent activation, including
up-regulation of HLA class II proteins and very late
activation antigen 4 integrins (4,5). Consistent with this
activated phenotype, RA synovial T lymphocytes can
stimulate monocyte tumor necrosis factor ? (TNF?)
production in a cell–cell-dependent manner (6). Para-
doxically, RA synovial T lymphocytes are noncycling and
hyporesponsive to subsequent mitogenic stimuli, includ-
ing T cell receptor (TCR) ligation (7–12). Although
there is still no direct evidence to indicate that T cell
activation in RA synovial tissue is mediated by antigen-
specific stimulation, a pivotal role of T lymphocytes in
the pathogenesis of RA is highlighted by the recent
clinical success of CTLA-4Ig fusion protein, which dis-
Dr. Reedquist’s work was supported by a Dutch Arthritis
Foundation project grant (NR 04-1-301).
1P. H. J. Remans, MD, C. A. Wijbrandts, MD, M. E. Sanders,
P. P. Tak, MD, PhD, K. A. Reedquist, PhD: Academic Medical
Center, University of Amsterdam, Amsterdam, The Netherlands;
2R. E. Toes, MD, PhD, F. C. Breedveld, MD, PhD, J. M. van Laar,
MD, PhD: Leiden University Medical Center, Leiden, The Nether-
Address correspondence and reprint requests to P. H. J.
Remans, MD, or K. A. Reedquist, PhD, Division of Clinical Immu-
nology and Rheumatology, Academic Medical Center, University of
Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands.
E-mail: firstname.lastname@example.org or email@example.com.
Submitted for publication September 23, 2005; accepted in
revised form June 28, 2006.
rupts costimulatory protein interactions between T cell
CD28 and CD80/86 on antigen-presenting cells (13,14).
Chronic oxidative stress triggered by reactive
oxygen species (ROS) is thought to underlie the patho-
physiologic role of T cells in many autoimmune diseases,
including RA, systemic lupus erythematosus, and human
immunodeficiency virus infection (1). In RA synovial T
cells, chronic oxidative stress leads to misfolding of the
cysteine-rich signaling proteins required for prolifera-
tion, including linker for activated T cells (LAT) and the
TCR-associated ?-chain (15–17). This effect can be
mimicked pharmacologically in normal T lymphocytes
(16) and in murine T cell hybridomas after prolonged
exposure to TNF? (18). Restoration of the redox bal-
ance, or overexpression of mutant ROS-insensitive LAT
(19) and TCR? (20), can partially relieve the induction
of mitogenic hyporesponsiveness by oxidative stress.
Oxidative stress may also contribute to inflamma-
tion by enhancing NF-?B–dependent transcription of
inflammatory cytokines and apoptotic and antiapoptotic
proteins of the Bcl-2 family, conferring resistance to
apoptotic stimuli. In contrast, transiently produced hy-
drogen peroxide and superoxide anions act as important
second messengers in TCR signaling (21). Acute stimu-
lation of T cells with physiologically relevant concentra-
tions of ROS can enhance MAP kinase activation,
proliferation, and transcription by NF-?B, activator pro-
tein 1, and the interleukin-2 (IL-2) promoter. Since
many cellular effects of ROS can be suppressed by
antioxidants, physiologic ROS generation is believed to
“fine-tune” T cell antigen responses (22,23).
We have recently demonstrated that oxidative
stress in T lymphocytes from the synovial fluid (SF) (24)
and synovial tissue (25) of RA patients is a result of T
cell–intrinsic intracellular ROS production. Persistent
ROS production in RA SF T cells correlates with
constitutive activation of the small GTPase Ras, and
blocks activation of the related GTPase Rap1 (24). Ras
and Rap1 are both signaling proteins that play central
roles in integrating the intracellular signaling pathways
leading to the functional outcomes of T cell stimulation.
Activation of Ras is sufficient and necessary for intra-
cellular ROS production, while activation of Rap1 can
block agonist- and Ras-dependent ROS production. In
contrast, inhibition of Rap1 prevents down-regulation of
ROS production in RA SF T cells, as has been observed
in a human T cell line following agonist stimulation (24).
Because these studies provided evidence that altered
Rap1 signaling is responsible for the development of
oxidative stress in RA synovial T cells, we conducted this
study to explore which factors in the synovial micro-
environment may alter T cell Rap1 activation to induce
T cell oxidative stress in RA.
PATIENTS AND METHODS
Patient characteristics. Peripheral blood (PB) from
healthy volunteers and PB and SF from RA patients in our
clinic were obtained using protocols approved by the medical
ethics committee of the Academic Medical Center, University
of Amsterdam. Paired PB and SF samples were obtained from
28 patients with RA (18 women and 10 men) meeting the
American College of Rheumatology (formerly, the American
Rheumatism Association) criteria for RA (26). The mean
duration of RA was 4.4 years (?SEM 7.6 years) and 21 of the
patients were seropositive for rheumatoid factor. Among the
patients, 25 were receiving disease-modifying antirheumatic
drugs and 3 were taking prednisolone.
Isolation and culture of PB and SF cells. PB mono-
nuclear cells (PBMCs) and SF mononuclear cells (SFMCs)
were obtained by Ficoll gradient centrifugation. PB and SF T
cells were purified from the PBMCs and SFMCs using a
negative isolation procedure, performed in accordance with
the manufacturer’s specifications (T Cell Negative Isolation
Kit; Dynal Biotech, Oslo, Norway). Purified T cells were
?90% CD3?, as assessed by fluorescence-activated cell sort-
ing (FACS) analysis. In some experiments, autologous PB and
SF adherent cells (?70% CD14? monocytes) were obtained
by allowing PBMCs and SFMCs to adhere to plastic tissue
culture dishes overnight, followed by washing. PB T cells were
then incubated alone or with purified adherent cells at a ratio
of 3:1 for 72 hours. Cells were maintained in RPMI 1640
medium supplemented with 10% fetal bovine serum, glu-
tamine, HEPES buffer, penicillin, and streptomycin (all from
Gibco, Grand Island, NY).
T cell proliferation assays. In control experiments, PB
and SF T lymphocytes were stimulated at a concentration of
5 ? 105cells/well in 96-well plates with activating anti-CD3
(Central Laboratory of The Netherlands Red Cross Blood
Transfusion Service, Amsterdam, The Netherlands) and anti-
CD28 antibodies (15E5, provided by Dr. R. A. van Lier,
University of Amsterdam). Seventy-two hours poststimulation,
cells were pulsed with 1 ?Ci/well3H-thymidine (New England
Nuclear, Boston, MA) for an additional 20 hours. Cells were
subsequently harvested on filter mats (Skatron Instruments,
Lier, Norway), and incorporation of radioactivity was mea-
sured using a liquid scintillation counter (Skatron Instru-
ments). Where indicated, purified PB T lymphocytes were
preincubated for 72 hours at a ratio of 3:1 with autologous PB
or SF adherent cells or with 50% autologous SF, prior to
repurification and stimulation. Alternatively, T cell preincuba-
tion with PB or SF adherent cells was conducted in the
presence or absence of transwell membranes (Costar, Cam-
bridge, MA), or in the presence of 10 ?g/ml control Ig or
recombinant CTLA-4Ig (kindly provided by Dr. R. A. van
Measurement of ROS production in T cells. Purified
PB and SF T cells were resuspended at 5 ? 106cells/ml in
phenol red–free Dulbecco’s modified Eagle’s medium (Gibco)
and loaded for 20 minutes at 37°C with 28 ?M of the
ROS-reactive dye 6-carboxy-2?,7?-dichlorofluorescein (DCF)
3136 REMANS ET AL
(Molecular Probes, Eugene, OR). Cells were then either left
unstimulated or stimulated with anti-CD3 antibody 1XE,
TNF? (10 ng/ml), transforming growth factor ? (TGF?) (4
?g/ml), IL-1? (125 pg/ml), interferon-? (IFN?) (100 units/ml)
(all cytokines from R&D Systems, Minneapolis, MN), or 50%
autologous RA SF. Cells were analyzed for ROS production
on a FACScan (Becton Dickinson, San Jose, CA), with results
expressed as the mean fluorescence intensity of oxidated DCF
in the FL1 channel at 0, 10, and 20 minutes poststimulation.
For nucleofected PB T cells (see below), T cells were first
gated by forward and side scatter to identify viable cells, and
then analyzed by CD20 staining (CyChrome-conjugated anti-
CD20 antibody; BD PharMingen, San Diego, CA) to detect
transfected cells. Nucleofected T cells were 51% positive
(?SD 9%, n ? 4) for CD20 expression at 24 hours postnucleo-
fection, and 38% positive (?SD 15%) at 96 hours postnucleo-
fection. The viability of the CD20-positive T cells was ?90% at
both time points, as assessed by annexin V staining.
Detection of Rap1 activation. Repurified PB and SF T
cells were resuspended at 5 ? 106cells/ml in serum-free RPMI
1640 medium, equilibrated for 10 minutes at 37°C, and then
left unstimulated or stimulated for 5 minutes with phorbol
myristate acetate (PMA) (100 ng/ml) and ionomycin (1 ?g/ml)
(both compounds from Sigma, St. Louis, MO). Cells were then
lysed in lysis buffer containing 1% Nonidet P40, 10% glycerol,
20 mM Tris (pH 7.6), 150 mM NaCl, 4 mM MgCl2, 2 mM
NaVO4, 10 mM NaF, and 2 mM phenylmethylsulfonyl fluoride.
Insoluble cell material was removed by centrifugation.
GST fusion proteins of Ral guanine nucleotide disso-
ciation stimulator Ras-binding domain (for precipitation of
active Rap1) were precoupled to glutathione-Sepharose beads
(Amersham, Arlington Heights, IL) and incubated with cellu-
lar lysates which were subjected to rocking for 1 hour at 4°C.
Subsequently, bound activated Rap1 was precipitated, washed
5 times in lysis buffer, and eluted in Laemmli’s sample buffer.
Proteins were separated by sodium dodecyl sulfate–
polyacrylamide gel electrophoresis and transferred to polyvi-
nylidene difluoride membranes (Bio-Rad, Hercules, CA).
Following blocking of blots in 2% milk/Tris buffered
saline (TBS)/Tween, GTPases were detected by immunoblot-
ting overnight with anti-Rap1 antibodies (1:500 in TBS/Tween)
(BD Transduction Laboratories, Lexington, KY). Blots were
Figure 1. Production of T cell reactive oxygen species (ROS) and inactivation of Rap1 by cell–cell contact with synovial fluid (SF) adherent cells
(SFAdCs) obtained from patients with rheumatoid arthritis (RA). A, ROS production was determined in autologous RA SF T cells (SFTCs)
incubated alone or in RA peripheral blood (PB) T cells (PBTCs) incubated alone or in combination with 50% autologous SF, PBAdCs, or SFAdCs.
Purified T cells were labeled with a ROS-reactive dye and assessed by fluorescence-activated cell sorting analysis. ? ? P ? 0.05 versus PBTCs alone.
B, SFAdC induction of T cell ROS production by cell–cell contact was determined as described in A, with some RA PBTCs and SFAdCs separated
by a transwell membrane during the incubation. ? ? P ? 0.05 versus untreated PBTCs. Values are shown as box plots, where the boxes represent
the 25th to 75th percentiles, the lines within the boxes represent the median, and the lines outside the boxes represent the 10th and 90th percentiles,
from 5 independent experiments. Results are expressed as the change in reactivity over 20 minutes, normalized to 100% for PBTCs incubated alone.
C, To observe blocking of Rap1 activation by SFAdCs in T lymphocytes, RA PBTCs were incubated in medium alone or with 50% autologous SF
or SFAdCs or with unfractionated PB mononuclear cells (PBMCs) cultured with or without 50% SF. After stimulation with medium alone or with
phorbol myristate acetate and ionomycin (PMA/I) for 5 minutes, purified cells were lysed and active GTP-bound Rap1 was precipitated with
immobilized Ral guanine nucleotide dissociation stimulator Ras-binding domain GST fusion protein, followed by immunoblotting. Results are
representative of 1 of 5 independent experiments. D, Inactivation of T cell Rap1 by cell–cell contact with SFAdCs is evident on immunoblotting.
RA PB T cells were coincubated overnight with SFAdCs, either together or separated by a transwell membrane. Results are representative of 1 of
3 independent experiments.
CTLA-4Ig SUPPRESSION OF RA SYNOVIAL T CELL OXIDATIVE STRESS3137
then washed 6 times in TBS/Tween, incubated with anti-mouse
horseradish peroxidase–conjugated antibody (1:3,000 in 2%
milk/TBS/Tween; Bio-Rad), washed again 6 times, and devel-
oped in ECL Detection Solution (Amersham).
Introduction of complementary DNA (cDNA) into
human PB T lymphocytes by nucleofection. To examine the
influence of active Rap1 on SF monocyte–induced ROS pro-
duction in PB T cells, 5 ? 106purified human T lymphocytes
were incubated in nucleofection buffer alone (Human T Cell
Nucleofection Kit; Amaxa, Cologne, Germany) or buffer con-
taining 3 ?g pCMV-CD20 (to detect transfected cells) and 7
?g empty vector pMT2-HA or pMT2-HA-RapV12 (encoding
constitutively active Rap1). Cells were nucleofected according
to the manufacturer’s directions, using program setting U-14.
Cells were allowed to rest overnight prior to stimulation and
ROS detection or incubation with PB and SF adherent cells.
Statistical analysis. Data were analyzed using SPSS for
Windows (version 11.5.1; SPSS, Chicago, IL). Because the data
were nonparametric, Wilcoxon’s signed rank test was per-
formed to compare paired data from control samples and
treated samples from the same patient. P values less than 0.05
were considered significant.
Regulation of T lymphocyte ROS production and
Rap1 activity by synovial adherent cells. Similar to
previous results (24), RA SF T cells displayed a high
basal rate of intracellular ROS production as compared
with autologous PB T lymphocytes (P ? 0.05) (Figure
1A). Initial experiments revealed a time-dependent de-
crease in intracellular ROS production in purified RA
SF T lymphocytes cultured ex vivo (results not shown),
indicating that the T cell signaling pathways regulating
ROS production are maintained by synovial microenvi-
ronment factors in vivo. To gain insight into the identity
of these factors, we examined the effects of autologous
SF on ROS production in PB T cells. Coincubation of
PB T cells with autologous SF cells failed to induce a
significant increase in ROS production (Figure 1A).
We next examined the effects of antigen-
presenting cells isolated from PB and SF on PB T cells.
Antigen-presenting cells were obtained by allowing
PBMCs and SFMCs to adhere to plastic tissue culture
dishes overnight, followed by washing. The cells con-
sisted of ?70% monocytes, as detected by CD14 stain-
ing. Coincubation of PB T cells with autologous SF
adherent cells induced a significant increase in T cell
ROS production as compared with that in untreated PB
T cells (P ? 0.05), with levels similar to those observed
in SF T cells (Figure 1A).
We then sought to determine whether the syno-
vial adherent cell–stimulated T cell ROS production was
mediated by secreted products or by cell–cell contact. To
address this issue, purified PB T cells were preincubated
with SF adherent cells, either together or separated by a
transwell membrane. Separation of T cells and SF
adherent cells reduced T cell ROS production to the
levels observed in unstimulated PB T lymphocytes (P ?
0.05) (Figure 1B).
Consistent with a proposed role for inactivation
of Rap1 in mediating RA SF T cell oxidative stress (24),
blockage of T cell Rap1 activation was observed only
following coincubation with SF adherent cells. Acti-
vated, GTP-bound Rap1 could be readily detected in
RA PB T lymphocytes following PMA/ionomycin stim-
ulation (Figure 1C). Preincubation of T cells with autol-
ogous SF had no inhibitory effect on subsequent Rap1
activation. Maintenance of PB T cells in the presence of
PB adherent cells also failed to influence Rap1 activa-
tion. However, preincubation of PB T cells with SF
adherent cells or exposure of PB adherent cells to SF led
to a complete block of Rap1 activation. Similar to our
initial findings, separation of PB T cells and SF adherent
cells by a transwell prevented inhibition of Rap1 signal-
ing (Figure 1D). Together, these results demonstrate a
Figure 2. Prevention of SFAdC–induced ROS production via activa-
tion of Rap1. Purified RA PB T cells were left untreated or nucleo-
fected with cDNA encoding CD20 to mark nucleofected cells, and
were incubated with either empty control vector pMT2-HA or
pMT2-HA encoding active RapV12. Nucleofected cells were exposed
to autologous SFAdCs prior to T cell purification and detection of
ROS production in CD20-positive T cells. Results are the mean and
SD of 5 independent experiments. ? ? P ? 0.05 versus empty vector.
See Figure 1 for definitions.
3138 REMANS ET AL
strong relationship between T cell Rap1 inactivation and
ROS production by activated SF adherent cells.
Prevention of synovial adherent cell–induced T
cell ROS production via Rap1 activation. We next sought
to determine whether Rap1 signaling was sufficient to
maintain the redox balance in T cells exposed to SF
adherent cells. We nucleofected RA PB T lymphocytes
with cDNA encoding CD20 to detect nucleofected cells,
followed by incubation with either empty vector or
vector encoding HA-tagged active RapV12. Nucleo-
fected T cells were allowed to recover for 24 hours and
then coincubated a further 72 hours with autologous SF
adherent cells. As compared with nonnucleofected PB T
cells, T cells nucleofected with empty vector were not
protected against induction of oxidative stress (Figure
2). In contrast, T cells nucleofected with active RapV12
displayed an ?80% decrease in intracellular ROS pro-
duction (P ? 0.05). Thus, rescue of Rap1 signaling is
sufficient to restore the redox balance in T cells exposed
to SF adherent cells.
Induction of mitogenic hyporesponsiveness in PB
T lymphocytes via cell–cell contact with RA synovial
adherent cells. We next examined whether induction of
ROS production and inactivation of Rap1 in T cells
correlated with T cell mitogenic hyporesponsiveness, a
hallmark of oxidatively stressed T lymphocytes. As ex-
pected (9,10,15,16), freshly isolated SF T lymphocytes
stimulated with anti-CD3/CD28 antibodies were mito-
genically hyporesponsive as compared with autologous
PB T lymphocytes (Figure 3A). Incubation of PB T cells
with autologous SF for 72 hours prior to CD3/CD28
antibody stimulation had no effect on T cell prolifer-
ative responses (Figure 3B). In contrast, incubation of
PB T cells with SF adherent cells, but not PB adherent
cells, blocked T cell proliferative responses, reducing
them to the levels observed in purified SF T lympho-
Inhibition of T cell proliferation by SF adherent
cells could not be attributed to secreted factors, since
SF, which contains a multitude of inflammatory cyto-
Figure 3. RA SFAdC induction of T cell mitogenic hyporesponsiveness via cell–cell contact. A, Hyporespon-
siveness of RA SF T cells to mitogenic stimulus was determined in purified RA SF and PB T cells that were either
left unstimulated (?) or stimulated for 72 hours with anti-CD3 and anti-CD28 antibodies (?), followed by
labeling with3H-thymidine for 20 hours. B, Induction of T cell hyporesponsiveness was determined in purified
RA PB T lymphocytes preincubated with SFAdCs, as compared with preincubation with autologous 50% SF or
PBAdCs. C, Inhibition of T cell proliferation by cell–cell contact with SFAdCs was determined by preincubation
of T cells with PBAdCs and SFAdCs in the absence or presence of an intervening transwell membrane, prior to
T cell repurification and stimulation. Bars show the mean and SD incorporated radioactivity from 3 independent
experiments performed in triplicate. See Figure 1 for definitions.
CTLA-4Ig SUPPRESSION OF RA SYNOVIAL T CELL OXIDATIVE STRESS3139
kines, had no effect on T cell proliferation, and separa-
tion of PB T cells from SF adherent cells by a transwell
membrane during incubation completely blocked the
inhibitory effect on T cell proliferation (Figure 3C).
These experiments indicated that induction of T cell
mitogenic hyporesponsiveness in RA requires cell–cell
contact with SF adherent cells.
Induction of T cell oxidative stress by synovial
adherent cells via CD28-dependent inactivation of
Rap1. Our results suggested that T cell surface protein
interactions with SF adherent cell surface ligands were
responsible for actively suppressing Rap1 signaling. We
therefore turned our attention to the possible role of
CD28 in this process, since we and other investigators
have previously demonstrated that CD28 costimulatory
signaling can block T cell Rap1 activation via RapGAP
I (27,28). In addition, CD28 signaling pathways, as
opposed to those of CD3, have been reported to be
intact in RA SF T lymphocytes (29).
To examine whether the T cell CD28–dependent
interactions with synovial adherent cells were responsi-
ble for the inactivation of Rap1 and subsequent ROS
induction, PB T lymphocytes were coincubated with
autologous SF adherent cells in the presence of control
human Ig or chimeric CTLA-4Ig recombinant fusion
protein, to disrupt CD28 interactions with adherent
cell CD80/86. The presence of CTLA-4Ig, but not
control Ig protein, led to an almost complete inhibition
of ROS production in T cells exposed to SF adherent
cells (P ? 0.05) (Figures 4A and B). The effects of
CTLA-4Ig treatment on T cell ROS production corre-
lated with a rescue of Rap1 signaling, since PB T cells
coincubated with CTLA-4Ig, but not those coincu-
bated with control Ig, displayed no defects in Rap1
activation after exposure to SF adherent cells (Fig-
Synergistic induction of T lymphocyte ROS pro-
duction by inflammatory cytokines and CD28 stimula-
tion. Our results indicated that CD28-dependent inacti-
vation of Rap1 is required for SF adherent cell induction
Figure 4. Use of CTLA-4Ig for the prevention of rheumatoid arthritis
(RA) synovial fluid adherent cell (SFAdC) reactive oxygen species
(ROS) induction and Rap1 inactivation. A, The effects of CTLA-4Ig
were determined in experiments with purified RA peripheral blood
(PB) T lymphocytes preincubated for 72 hours in medium alone or in
medium containing autologous SFAdCs (T cell:SFAdC ratio 3:1) and
either control IgG-Fc or CTLA-4Ig (both at 10 ?g/ml), followed by T
cell purification, 6-carboxy-2?,7?-dichlorofluorescein (DCF) reactive
dye loading, and ROS detection as described in Figure 1, with results
expressed as the change in mean fluorescence intensity (?MFI). B,
ROS production was determined in SFAdCs preincubated with
CTLA-4Ig as compared with that in medium alone or with control Ig.
Bars show the mean and SD rate of ROS production over 20 minutes,
normalized to 100% for incubation in medium alone. ? ? P ? 0.05
versus control Ig. C, CTLA-4Ig was shown to prevent SFAdC-induced
inhibition of T cell Rap1 activity, in experiments comparing PB T
lymphocytes preincubated with medium alone with SFAdCs in the
presence of control Ig or CTLA-4Ig. Reisolated T cells were left
unstimulated (?) or stimulated for 5 minutes with phorbol myristate
acetate and ionomycin (PMA/I) (?), and GTP-bound active Rap1 was
precipitated and detected by immunoblotting. Results are representa-
tive of 1 of 5 independent experiments.
Figure 5. Synergistic induction of ROS production in PB T lympho-
cytes by inflammatory cytokines and CD28 stimulation. A, To show
that inflammatory cytokines alone are insufficient to induce ROS
production in RA PB T cells (PBTCs), isolated PBTCs were incubated
for 72–96 hours in medium alone or in medium containing tumor
necrosis factor ? (TNF?) (10 ng/ml), transforming growth factor ?
(TGF?) (4 ?g/ml), or 50% autologous SF. SF T cells (SFTCs) were
maintained in 50% SF. Cells were then harvested and loaded with
DCF, and ROS detection was measured as described in Figure 1, with
results expressed as ?MFI. B, Costimulation of PBTCs with inflam-
matory cytokines and activating anti-CD28 antibodies is shown to
induce production of ROS in healthy donor PBTCs. PBTCs were
maintained for 72–96 hours in medium alone or in medium containing
TNF? (10 ng/ml), interleukin-1? (IL-1?) (125 pg/ml), interferon-?
(IFN?) (100 units/ml), or TGF? (4 ?g/ml) in the absence or presence
of activating anti-CD28 antibody. To determine production of ROS,
the T cells were harvested and loaded with DCF, prior to measurement
as described in Figure 1. Bars show the mean and SD of 3 independent
experiments. ? ? P ? 0.05 versus control (without antibody) and
versus medium alone. See Figure 4 for other definitions.
3140 REMANS ET AL
of ROS production in T lymphocytes. To test whether
CD28 stimulation was sufficient to induce intracellular
ROS production, we stimulated PB T cells with different
cytokines in the presence or absence of CD28. Short-
term stimulation of PB T cells isolated from healthy
donors was carried out with TNF?, TGF?, or SF, but
none of these stimulations induced T cell ROS produc-
tion (Figure 5A). Similarly, long-term stimulation of PB
T cells with TNF?, IL-1?, IFN?, or TGF? for 3–7 days
also failed to induce significant ROS production above
basal levels (Figure 5B and results not shown). Strik-
ingly, coincubation of PB T cells with activating anti-
CD28 antibodies and each of these inflammatory cyto-
kines led to a synergistic increase in T cell ROS
production (P ? 0.05) (Figure 5B).
Recent studies from our group, utilizing FACS-
based detection of intracellular ROS production in RA
SF T lymphocytes (24) and ROS-dependent visualiza-
tion of diaminobenzidine precipitation in RA synovial
tissue (25), demonstrated that endogenous, intracellular
ROS production is sufficient to induce oxidative stress in
RA synovial T lymphocytes. Our present results provide
a molecular and cellular basis for the induction of
oxidative stress in RA synovial T cells, suggesting a
model in which inactivation of Rap1 plays a central role
in establishing oxidative stress and altered T cell behav-
ior in RA synovial tissue.
Upon arrival in synovial tissue, T cells are ex-
posed to a mix of inflammatory cytokines and cell–cell
interactions, many of which have been reported to
activate Ras in T cells. In RA, engagement of CD28 by
CD80/86 expressed on monocytes, B lymphocytes, den-
dritic cells, and stromal cells will lead to prolonged Rap1
inactivation and a subsequent inability of the T cell to
down-regulate ROS production. Although long-term
stimulation of T cells with inflammatory cytokines was
not sufficient to induce T cell oxidative stress in the
present study, preliminary experiments by our group
have found that long-term exposure to TNF? constitu-
tively activates Ras in T cells. T cell activation of Ras by
presentation of inflammatory cytokines, in combination
with CD28-dependent inactivation of Rap1 by CD80/86-
expressing synovial cells, might be responsible for the
high levels of intracellular ROS production observed in
synovial T cells. Consistent with this idea, investigations
by Isomaki et al and Clark et al established that long-
term stimulation of murine T lymphocytes with TNF?
rendered the cells defective in TCR-dependent prolifer-
ative responses, which is due, in part, to T cell ROS
We propose that deregulation of Rap and Ras
are critical events leading to the disturbed intracellular
redox balance underlying antigenic hyporesponsiveness
and inflammatory gene transcription in RA synovial T
cells. Rap1 plays a central role in integrating the TCR
and costimulatory signals leading to T cell immune
responses (30). In T lymphocytes, studies on Rap1 have
focused on its role in regulating integrin-dependent
adhesion (30,31). Our results suggest that improper, or
long-term, inactivation of Rap1 can also influence T cell
function through deregulation of the T cell redox balance.
Consistent with our previous findings, the ability
of RapV12 to prevent oxidative stress in T cells exposed
to SF adherent cells was not secondary to the effects on
T cell integrin activity, since a RapV12/E38 mutant that
does not regulate ROS production in Jurkat T cells and
yet, similar to RapV12, can stimulate integrin-
dependent adhesion (24) failed to suppress T cell ROS
production (results not shown). CD28-dependent inac-
tivation of Rap1 is mediated by Lck tyrosine kinase
activation of RapGAP I (27,28). Intriguingly, in mice
transgenically overexpressing RapGAP I in the T cell
compartment, an age-dependent accumulation of acti-
vated T lymphocytes was observed (32), although sus-
ceptibility of these mice to spontaneous or induced
chronic inflammatory diseases has not been examined.
Our results underscore the importance of CD28
costimulation in the activation of T cells in the RA
synovium. In particular, CD28 stimulation up-regulates
intracellular ROS production. ROS regulation in RA
synovial T lymphocytes may contribute to inflammation,
since, in vitro and in pharmacologic and genetic studies
in rodent models of arthritis, there are strong indications
that ROS-dependent activation of NF-?B in T lympho-
cytes contributes to pathogenesis (33–35). Conversely,
the resultant oxidative stress is associated with inhibition
of TCR-proximal proliferative signals, via misfolding of
LAT and TCR? (19,20).
The findings in our present model suggest that
CTLA-4Ig therapy could block oxidative stress in syno-
vial T cells in RA patients, and it will be of interest to
determine whether intracellular ROS production in RA
synovial T cells may be a predictor of, or correlate with,
clinical responses to CTLA-4Ig therapy. In many animal
models of arthritis, CD28 acts as a classic essential
costimulatory protein in permitting the TCR-dependent
responses to collagen that are required for initiation and
progression of joint inflammation (36–39). Trials with
CTLA-4Ig blockade of CD28 signaling in RA have been
CTLA-4Ig SUPPRESSION OF RA SYNOVIAL T CELL OXIDATIVE STRESS 3141
extremely promising, but the mechanism of its therapeu-
tic activity in humans has yet to be assessed (13,14).
Several mechanisms might explain how CTLA-
4Ig therapy exerts its clinical benefits despite rescuing
the proliferative responsiveness of potentially autoreac-
tive T lymphocytes. First, CTLA-4Ig would be expected
to both decrease oxidative stress–dependent NF-?B
inflammatory gene transcription and block CD28 signal-
ing critical for TCR-dependent T cell activation and
proliferation. Second, restoration of TCR-dependent
IL-2 production may simultaneously act to enhance
regulatory T cell function, which is defective in RA (40).
Finally, restoration of Rap1 function may allow integrin-
dependent emigration of T cells from the synovium,
independent of TCR-dependent proliferative signals.
Intriguingly, in a subset of RA patients,
CD4?,CD28? T lymphocyte numbers are greatly ex-
panded in the synovium (41). CD4?,CD28? T cell
clones, which display some similarities to natural killer
cells, are often autoreactive and sensitive to TCR trig-
gering, and are associated with extraarticular organ
involvement in RA (41,42). Lack of CD28 expression
may protect these cells from induction of oxidative
stress, contributing to TCR-dependent activation. Alter-
natively, T cell costimulatory proteins other than CD28
may redundantly regulate Rap1 function and ROS pro-
duction in these cells.
For future studies, it will be of interest to see if
synovial CD28? T cells also undergo oxidative stress, or
whether oxidative stress leads to CD28 down-regulation
in these cells. The recent development of techniques to
quantitatively detect ROS-producing T lymphocytes in
RA synovial tissue in situ, in conjunction with functional
analysis of T cells following CTLA-4Ig treatment of RA
patients, will allow more detailed characterization of the
T cell subsets that are under oxidative stress in the RA
synovium. Of particular importance will be determining
how these T cells respond to therapeutic treatment.
1. Cope AP. Studies of T-cell activation in chronic inflammation.
Arthritis Res 2002;4 Suppl 3:S197–211.
2. Firestein GS. The T cell cometh: interplay between adaptive
immunity and cytokine networks in rheumatoid arthritis. J Clin
3. Firestein GS, Zvaifler NJ. How important are T cells in chronic
rheumatoid synovitis? II. T cell-independent mechanisms from
beginning to end [review]. Arthritis Rheum 2002;46:298–308.
4. Janossy G, Panayi G, Duke O, Bofill M, Poulter LW, Goldstein G.
Rheumatoid arthritis: a disease of T-lymphocyte/macrophage im-
munoregulation. Lancet 1981;2:839–42.
5. Laffon A, Garcia-Vicuna R, Humbria A, Postigo AA, Corbi AL,
de Landazuri MO, et al. Upregulated expression and function of
VLA-4 fibronectin receptors on human activated T cells in rheu-
matoid arthritis. J Clin Invest 1991;88:546–52.
6. Brennan FM, Hayes AL, Ciesielski CJ, Green P, Foxwell BM,
Feldmann M. Evidence that rheumatoid arthritis synovial T cells
are similar to cytokine-activated T cells: involvement of phospha-
tidylinositol 3-kinase and nuclear factor ?B pathways in tumor
necrosis factor ? production in rheumatoid arthritis. Arthritis
7. Kingsley GH, Pitzalis C, Panayi GS. Abnormal lymphocyte reac-
tivity to self-major histocompatibility antigens in rheumatoid
arthritis. J Rheumatol 1987;14:667–73.
8. Verwilghen J, Vertessen S, Stevens EA, Dequeker J, Ceuppens JL.
Depressed T-cell reactivity to recall antigens in rheumatoid arthri-
tis. J Clin Immunol 1990;10:90–8.
9. Keystone EC, Poplonski L, Miller RG, Gorczynski R, Gladman D,
Snow K. Reactivity of T-cells from patients with rheumatoid
arthritis to anti-CD3 antibody. Clin Immunol Immunopathol
10. Pope RM, McChesney L, Talal N, Fischbach M. Characterization
of the defective autologous mixed lymphocyte response in rheu-
matoid arthritis. Arthritis Rheum 1984;27:1234–44.
11. Mirza NM, Relias V, Yunis EJ, Pachas WN, Dasgupta JD.
Defective signal transduction via T-cell receptor-CD3 structure in
T cells from rheumatoid arthritis patients. Hum Immunol 1993;
12. Allen ME, Young SP, Michell RH, Bacon PA. Altered T lympho-
cyte signaling in rheumatoid arthritis. Eur J Immunol 1995;25:
13. Moreland LW, Alten R, Van den Bosch F, Appelboom T, Leon M,
Emery P, et al. Costimulatory blockade in patients with rheuma-
toid arthritis: a pilot, dose-finding, double-blind, placebo-con-
trolled clinical trial evaluating CTLA-4Ig and LEA29Y eighty-five
days after the first infusion. Arthritis Rheum 2002;46:1470–9.
14. Kremer JM, Westhovens R, Leon M, Di Giorgio E, Alten R,
Steinfeld S, et al. Treatment of rheumatoid arthritis by selective
inhibition of T-cell activation with fusion protein CTLA4Ig.
N Engl J Med 2003;349:1907–15.
15. Maurice MM, Lankester AC, Bezemer AC, Geertsma MF, Tak
PP, Breedveld FC, et al. Defective TCR-mediated signaling in
synovial T cells in rheumatoid arthritis. J Immunol 1997;159:
16. Gringhuis SI, Leow A, Papendrecht-van der Voort EA, Remans
PH, Breedveld FC, Verweij CL. Displacement of linker for
activation of T cells from the plasma membrane due to redox
balance alterations results in hyporesponsiveness of synovial fluid
T lymphocytes in rheumatoid arthritis. J Immunol 2000;164:
17. Berg L, Ronnelid J, Klareskog L, Bucht A. Down-regulation of the
T cell receptor CD3 ? chain in rheumatoid arthritis and its
influence on T cell responsiveness. Clin Exp Immunol 2000;120:
18. Isomaki P, Panesar M, Annenkov A, Clark JM, Foxwell BM,
Chernajovsky Y, et al. Prolonged exposure of T cells to TNF
down-regulates TCR ? and expression of the TCR/CD3 complex at
the cell surface. J Immunol 2001;166:5495–507.
19. Gringhuis SI, Papendrecht-van der Voort EA, Leow A, Nivine
Levarht EW, Breedveld FC, Verweij CL. Effect of redox balance
alterations on cellular localization of LAT and downstream T-cell
receptor signaling pathways. Mol Cell Biol 2002;22:400–11.
20. Clark JM, Annenkov AE, Panesar M, Isomaki P, Chernajovsky Y,
Cope AP. T cell receptor ? reconstitution fails to restore responses
of T cells rendered hyporesponsive by tumor necrosis factor ?.
Proc Natl Acad Sci U S A 2004;101:1696–701.
21. Tatla S, Woodhead V, Foreman JC, Chain BM. The role of
reactive oxygen species in triggering proliferation and IL-2 secre-
tion in T cells. Free Radic Biol Med 1999;26:14–24.
22. Hehner SP, Breitkreutz R, Shubinsky G, Unsoeld H, Schulze-
3142REMANS ET AL
Osthoff K, Schmitz ML, et al. Enhancement of T cell receptor
signaling by a mild oxidative shift in the intracellular thiol pool.
J Immunol 2000;165:4319–28.
23. Devadas S, Zaritskaya L, Rhee SG, Oberley L, Williams MS.
Discrete generation of superoxide and hydrogen peroxide by T cell
receptor stimulation: selective regulation of mitogen-activated
protein kinase activation and fas ligand expression. J Exp Med
24. Remans PH, Gringhuis SI, van Laar JM, Sanders ME, Papendre-
cht-van der Voort EA, Zwartkruis FJ, et al. Rap1 signaling is
required for suppression of Ras-generated reactive oxygen species
and protection against oxidative stress in T lymphocytes. J Immu-
25. Remans PH, van Oosterhout M, Smeets TJ, Sanders M, Frederiks
WM, Reedquist KA, et al. Intracellular free radical production in
synovial T lymphocytes from patients with rheumatoid arthritis.
Arthritis Rheum 2005;52:2003–9.
26. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF,
Cooper NS, et al. The American Rheumatism Association 1987
revised criteria for the classification of rheumatoid arthritis.
Arthritis Rheum 1988;31:315–24.
27. Reedquist KA, Bos JL. Costimulation through CD28 suppresses T
cell receptor-dependent activation of the Ras-like small GTPase
Rap1 in human T lymphocytes. J Biol Chem 1998;273:4944–9.
28. Carey KD, Dillon TJ, Schmitt JM, Baird AM, Holdorf AD, Straus
DB, et al. CD28 and the tyrosine kinase lck stimulate mitogen-
activated protein kinase activity in T cells via inhibition of the
small G protein Rap1. Mol Cell Biol 2000;20:8409–19.
29. Maurice MM, van der Voort EA, Leow A, Levarht N, Breedveld
FC, Verweij CL. CD28 co-stimulation is intact and contributes to
prolonged ex vivo survival of hyporesponsive synovial fluid T cells
in rheumatoid arthritis. Eur J Immunol 1998;28:1554–62.
30. Cantrell DA. GTPases and T cell activation. Immunol Rev 2003;
31. Bos JL. Linking Rap to cell adhesion. Curr Opin Cell Biol
32. Dillon TJ, Carey KD, Wetzel SA, Parker DC, Stork PJ. Regulation
of the small GTPase Rap1 and extracellular signal-regulated
kinases by the costimulatory molecule CTLA-4. Mol Cell Biol
33. Collantes E, Valle BM, Mazorra V, Macho A, Aranda E, Munoz
E. Nuclear factor-?B activity in T cells from patients with rheu-
matic diseases: a preliminary report. Ann Rheum Dis 1998;57:
34. Seetharaman R, Mora AL, Nabozny G, Boothby M, Chen J.
Essential role of T cell NF-?B activation in collagen-induced
arthritis. J Immunol 1999;163:1577–83.
35. Gerlag DM, Ransone L, Tak PP, Han Z, Palanki M, Barbosa MS,
et al. The effect of a T cell-specific NF-?B inhibitor on in vitro
cytokine production and collagen-induced arthritis. J Immunol
36. Knoerzer DB, Karr RW, Schwartz BD, Mengle-Gaw LJ. Collagen-
induced arthritis in the BB rat: prevention of disease by treatment
with CTLA-4-Ig. J Clin Invest 1995;96:987–93.
37. Webb LM, Walmsley MJ, Feldmann M. Prevention and amelio-
ration of collagen-induced arthritis by blockade of the CD28
co-stimulatory pathway: requirement for both B7-1 and B7-2. Eur
J Immunol 1996;26:2320–8.
38. Tada Y, Nagasawa K, Ho A, Morito F, Ushiyama O, Suzuki N, et
al. CD28-deficient mice are highly resistant to collagen-induced
arthritis. J Immunol 1999;162:203–8.
39. Ijima K, Murakami M, Okamoto H, Inobe M, Chikuma S, Saito I,
et al. Successful gene therapy via intraarticular injection of ade-
novirus vector containing CTLA4IgG in a murine model of type II
collagen-induced arthritis. Hum Gene Ther 2001;12:1063–77.
40. Ehrenstein MR, Evans JG, Singh A, Moore S, Warnes G, Isenberg
DA, et al. Compromised function of regulatory T cells in rheuma-
toid arthritis and reversal by anti-TNF? therapy. J Exp Med
41. Schmidt D, Goronzy JJ, Weyand CM. CD4? CD7? CD28? T
cells are expanded in rheumatoid arthritis and are characterized by
autoreactivity. J Clin Invest 1996;97:2027–37.
42. Martens PB, Goronzy JJ, Schaid D, Weyand CM. Expansion of
unusual CD4? T cells in severe rheumatoid arthritis. Arthritis
CTLA-4Ig SUPPRESSION OF RA SYNOVIAL T CELL OXIDATIVE STRESS3143