Spatiotemporal Basis of CTLA-4
Negative Regulation of T Cell Activation
Tadashi Yokosuka,1,* Wakana Kobayashi,1Masako Takamatsu,1Kumiko Sakata-Sogawa,2,5Hu Zeng,1,6
Akiko Hashimoto-Tane,1Hideo Yagita,4Makio Tokunaga,3,5and Takashi Saito1,6,*
1Laboratory for Cell Signaling
2Research Unit for Single Molecule Imaging
3Research Unit for Molecular Systems Immunology
RIKEN Research Center for Allergy and Immunology, Yokohama, Kanagawa 230-0045, Japan
4Department of Immunology, Juntendo University School of Medicine, Tokyo 113-8421, Japan
5Department of Biological Information, Tokyo Institute of Technology, Yokohama, Kanagawa 226-8501, Japan
6WPI Immunology Frontier Research Center, Osaka University, Suita, Osaka 565-0871, Japan
*Correspondence: firstname.lastname@example.org (T.Y.), email@example.com (T.S.)
Tcell activation ispositively andnegatively regulated
by a pair of costimulatory receptors, CD28 and
CTLA-4, respectively. Because these receptors
share common ligands, CD80 and CD86, the expres-
sion and behavior of CTLA-4 is critical for T cell cos-
timulation regulation. However, in vivo blocking of
CD28-mediated costimulation by CTLA-4 and its
mechanisms still remain elusive. Here, we demon-
strate the dynamic behavior of CTLA-4 in its real-
time competition with CD28 at the central-supramo-
lecular activation cluster (cSMAC), resulting in the
dislocalization of protein kinase C-q and CARMA1
scaffolding protein. CTLA-4 translocation to the
T cell receptor microclusters and the cSMAC is
mulation at the cSMAC is required for its inhibitory
function. The CTLA-4-mediated suppression was
demonstrated by the in vitro anergy induction in
regulatory T cells constitutively expressing CTLA-4.
These results show the dynamic mechanism of
CTLA-4-mediated T cell suppression at the cSMAC.
T cell activation requires two signals from the T cell receptor
(TCR) and costimulatory receptor. There are sets of costimula-
signal regulators, all which cooperatively modify T cell activation
(Alegre et al., 2001; Rudd and Schneider, 2003; Sharpe, 2009).
Cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) was first
discovered as a negative regulator and plays a pivotal role in
terminating T cell activation in vivo. CTLA-4-deficient (Ctla4?/?)
mice display the activation of all T cell subsets and massive
T cell proliferation and die from severe systemic autoimmune
diseases with the infiltration of activating T cells (Tivol et al.,
1995; Waterhouse et al., 1995). The modulation of CTLA-4 func-
tion is already applied clinically, in cancer immunotherapy with
CTLA-4 antibodies (Abs) (Egen et al., 2002), and in the treatment
for autoimmune diseases with CTLA-4 immunoglobulin (Ig)
fusion proteins (Linsley and Nadler, 2009).
In an in vitro analysis, CTLA-4 was suggested to mediate T cell
suppression by ectodomain competition with the positive costi-
mulatory receptor CD28 for binding to their common ligands
CD80 and CD86. Other mechanisms are thought to involve the
phosphatase recruitment and the trafficking regulation and
expression of TCRs and lipid rafts (Chikuma et al., 2003).
Although it has been shownthat some phosphatases areassoci-
ated with the cytoplasmic tail of CTLA-4 (Baroja et al., 2002;
Chuang et al., 2000; Lee et al., 1998; Sloan-Lancaster et al.,
1997), it remains unclear whether T cell signaling is inhibited
by these phosphatases. In contrast, it is clear that the ectodo-
main of CTLA-4 mediates the inhibition of CD28-mediated costi-
CD28, the expression and dynamism of CTLA-4 are critical in
the competition for ligand binding. However, the actual competi-
tion between them has not been experimentally demonstrated.
When an immune response is initiated by the communication
between T cells and antigen-presenting cells (APCs), a charac-
teristic pattern of receptors and their downstream molecules,
which is identified by the immunological synapse (IS), is formed
at the T cell-APC interface (Grakoui et al., 1999; Monks et al.,
1998). The IS is made up of a central supramolecular activation
cluster (cSMAC) containing TCR-CD3 peptide, major histocom-
patibility complex (MHCp), and a peripheral SMAC (pSMAC)
containing adhesion molecules, leukocyte function-associated
(ICAM-1). Both CTLA-4 and CD28 (Egen and Allison, 2002;
Pentcheva-Hoang et al., 2004) and other costimulatory recep-
tors, such as inducible T cell costimulator (ICOS) (Fos et al.,
2008) and programmed death 1 (PD-1) (Pentcheva-Hoang
et al., 2007), integrins (Nguyen et al., 2008), and TCR itself all
accumulate at the IS in a T cell-APC conjugate. Thus, the IS is
considered to be a huge aggregation of TCR-based signalo-
somes that induce T cell responses, including the costimulation.
We and others have recently reported that T cell activation is
spatiotemporally induced by TCR microclusters (MCs), which
326 Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc.
consist of receptors and their downstream signaling molecules,
such as zeta-associated protein 70 (Zap70) and SH2-domain-
containing leukocyte protein of 76 kDa (SLP-76), and are formed
Saito and Yokosuka, 2006; Yokosuka et al., 2005). We have also
found that CD28-mediated T cell costimulation is mediated by
the TCR-CD28 MCs through the specific recruitment of protein
kinase C-q (PKC-q) (Yokosuka et al., 2008; Yokosuka and Saito,
2009). A magnified image has revealed that a cSMAC is divided
into two distinct regions, the TCR-CD3 high-density region
(CD3hi) and the TCR-CD3 low-density region (CD3lo). CD28
and PKCq are recruited to the CD3loregion, suggesting that
the CD3loregion contributes to CD28-mediated T cell costimula-
tion through PKCq recruitment as the signaling cSMAC (Yoko-
suka et al., 2008; Yokosuka and Saito, 2009).
Here, we present a molecular imaging analysis of the compe-
tition between CTLA-4 and CD28, which reveals the dynamic
mechanism of the CTLA-4-mediated negative regulation of
T cell activation. CTLA-4, which is stored in the secretary
granules, directly moves to the activating T cell-APC or bilayer
interface and translocated to the CD3losignaling cSMAC in
a ligand-binding manner, resulting in the inhibition of CD28-
PKCq-CARD-containing MAGUK protein 1 (CARMA1) recruit-
ment. Furthermore, the translocation of CTLA-4 to TCR MCs
and the cSMAC was tightly regulated by its ectodomain size.
We also found that this inhibitory mechanism operates not only
in CD4+effector Tcells but also in regulatory T (Treg) cells, which
lying the in vitro anergy of Treg cells.
Ligand-Dependent Accumulation of CTLA-4
at the Activated T Cell-APC Interface
The surface expression of CTLA-4 on T cells is controlled by
transcription and intracellular localization. CTLA-4 is induced
upon T cell activation and predominantly localized in perforin-
containing secretary granules (Iida et al., 2000). Those granules
are translocated to the activation site by TCR signaling alone
and CTLA-4 is stabilized by the interaction with CD80 on APCs
(Egen and Allison, 2002; Linsley et al., 1996). To confirm this,
we transduced AND-TCR (specific for moth cytochrome C
[MCC] 88-103) Tcell hybridomas byenhanced green fluorescent
protein (EGFP)-tagged CTLA-4 (CTLA-4-EGFP), conjugated
these T cells with an I-Ek-expressing dendritic cell line, DC-1,
prepulsed with MCC88-103, and observed the cells by confocal
microscopy. As previously reported, the CTLA-4-containing
granules in the cytoplasm were translocated to the T cell-DC-1
activating interface and CTLA-4 densely accumulated in the
presence of CD80 (Figures 1A–1D).
Generation of CTLA-4 Clusters behind TCR MC
To analyze the precise movement of CTLA-4 during the initial
AND-Tg T cells settled on a McConnell’s glass-supported planar
bilayer (Grakoui et al., 1999) by total internal reflection fluores-
cence microscopy (TIRFM) (Tokunaga et al., 2008). The planar
bilayer contained glycophosphatidylinositol
ICAM-1, CD80, CD86, and I-Ek, which had been pulsed with
MCC88-103 and their densities were adjusted to equal those
on lipopolysaccharide-stimulated B cells, i.e., ?200–250, 80,
and 80–150 molecules/mm2, respectively. On the planar bilayer,
the TCRs first accumulated to form clusters at the nascent
contact region of the T cell-bilayer interface; then these TCR
MCs migrated toward the center and developed a cSMAC
within 5 min (Figure 2) (Campi et al., 2005; Varma et al., 2006;
TCRs to form TCR-CTLA-4 MCs and then accumulated into the
cSMAC (Figures 2A and 2B; Figures S1A–S1C and Movie S1
available online). Because there was a delay in the appearance
Figure 1. CTLA-4 Is Recruited to the Acti-
vated T Cell-APC Interface
AND-TCR T cell hybridomas introduced by
CTLA-4-EGFP(green) were conjugated
MCC88-103-prepulsed DC-1 expressing (+) or
not expressing (?) CD80 and were real-time
imaged by confocal microscopy at 2 or 30 min
after conjugation in (A). The asterisk marks DC-1;
scale bars represent 5 mm. Fold fluorescence
intensities of CTLA-4 on the diagonal yellow lines
in (A) were measured in ten representative T cell–
DC-1 conjugates in (B). Area (C) and intensity of
CTLA-4 (D) on the surface (10% area shown in C)
oron thesurfaceplus cytoplasm (20%areashown
in C) in each cell after T cell-DC-1 conjugate
(30 min) were compared with those of entire
T cells (CD80?, n = 33; CD80+, n = 33). *p <
0.0001 with Student’s t test. A representative
two independent experiments is shown.
Dynamics of CTLA-4-Mediated T Cell Suppression
Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc. 327
of CTLA-4 when compared with that of TCR, the CTLA-4 MCs
were detected in the area relatively close to the cSMAC.
Although TCR MCs and a cSMAC were seen on the T cell,
CTLA-4 was not stabilized at the T cell-bilayer interface and
CTLA-4-containing granules floated behind the interface in the
absence of CD80-GPI (Figures S1D and S1E and Movie S2).
TCR MCs and finally with the cSMAC, in a manner dependent on
Accumulation of CTLA-4 at the Signaling Subregion
within the cSMAC
As we previously demonstrated, CD28 and PKCq colocalize with
TCR MCs at the initiation of T cell-bilayer contact and then accu-
mulate at the cSMAC particularly in the CD3loregion, which is
segregated from the CD3hiwithin the cSMAC (Dustin, 2009;
Yokosuka et al., 2008). Photobleaching analysis of these two
distinct regions demonstrated the quick recovery of TCR-CD3
in the CD3loregion and the irrecoverable damage in the CD3hi
(Figure S2), suggesting a dynamic accumulation of receptors in
the CD3loand a processing degradation of TCR in the CD3hi.
Accordingly, the CD3loregion might serve as the signaling
cSMAC, compared to the CD3hi. Because the accumulation of
CD28 at the TCR MCs precedes the arrival of the CTLA-4-con-
taining granules at the T cell-bilayer interface (Figures 2B
and 3A, top), this dynamics of CTLA-4 movement contributes
to the block of CD28-CD80 binding in the later stage, rather
than during the initial contact. We next visualized TCR-CD3,
CD28, and CTLA-4 after TCR MC formation by using AND-Tg
T cells expressing both fluorescently tagged CD28 and
CTLA-4. Similar to CD28, CTLA-4 was translocated to the
CTLA-4 was also segregated from the CD3hiregion and devel-
oped an annular structure (Figure 3A and Movies S3 and S4).
In parallel, CD28 was excluded from the cSMAC and clustered
outside the pSMAC in the presence of a high density of CD80
(80 molecules/mm2) in 67% of the cells. However, there were
no CD28 clusters with low-density CD80 (5 molecules/mm2) in
83% of the cells. In contrast, CTLA-4 clearly formed annular
clusters in the cSMAC especially in the CD3loregion in all cells
in the presence of both high and low densities of CD80 (Figures
3A and 3B). To visualize the colocalization of CD28 or CTLA-4
labeled CD80- or CD86-GPI into the bilayers. With high-density
CD80 and CD86, CTLA-4 and CD28 colocalized with CD80- or
low-density CD80 and CD86 (Figures S3A and S3B). In this
system, CD80 and CD86 accumulated strongly in the annular
CTLA-4 clusters regardless of CD80 and CD86 density.
To confirm the differential localization of endogenous CD28
and CTLA-4 at the IS, we analyzed the total intensity of TCR-
CD3, CD28, and CTLA-4 at the cSMAC in AND-Tg effector
T cells settled on a bilayer containing different densities of
CD80-GPI (Figure 3C). CD28 accumulated more strongly at the
cSMAC upon stimulation with a high density of CD80 (80 mole-
cules/mm2), whereas CTLA-4 tended to accumulate in the
same region even at a lower density of CD80 (5 or 1.25 mole-
cules/mm2). Furthermore, the analysis of AND-Tg Ctla4?/?
T cells demonstrated that the lack of CTLA-4 facilitated CD28
accumulation at the cSMAC even in the presence of low-density
CD80 (Figure 3D). From these data, we propose that the CD3lo
signaling cSMAC is the major site at which CTLA-4 competes
with CD28 for CD80 and CD86 binding.
Costimulatory signals generally become more effective with
weak TCR stimulation. When we pretreated the planar bilayers
with lower doses of antigen peptide at physiological concentra-
tions, the cSMACs became smaller or the cSMAC itself disap-
peared. We have previously shown that the annular clustering
of CD28 and PKCq still occurs on a bilayer with low doses of
antigen peptide even in the absence of a cSMAC and that the
Figure 2. Ligand-Dependent Clustering of CTLA-4 at the T Cell-Bilayer Interface
planar bilayer containing I-Ek-, ICAM-1-, and CD80-GPI. The cells were visualized at video rate by TIRFM (time shown above images) in (A). Images on the
diagonal yellow lines in (A) are shown as kymographs (B). The real-time image is available in Movie S1. Scale bars represent 5 mm. A representative three
independent experiments is shown.
Dynamics of CTLA-4-Mediated T Cell Suppression
328 Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc.
generation of CD28 clusters is largely dependent on the amount
of CD80 present. The membrane translocation of CTLA-4 also
correlates with the strength of the TCR signal, but annular
CTLA-4 clusters were observed on bilayers containing sufficient
CD80 and CD86 (80 molecules/mm2) and low-dose MCC88-103
(0.01 mM) (Figures S3C and S3D). These results indicate that,
even under physiological conditions, CTLA-4 can compete
with CD28 for ligand binding at the cSMAC, resulting in the
exclusion of CD28 outside the cSMAC.
Structural Regulation of CTLA-4 Accumulation at TCR
MCs and the cSMAC
We analyzed the structure-function relationship of CTLA-4 in the
MC and cSMAC formation by using several mutant CTLA-4 to
Figure 3. CTLA-4 Interferes with CD28 Accumulation at the cSMAC
(A and B) AND-Tg T cells expressing both CD28-ECFP (cyan) and CTLA-4-EYFP (yellow) were stained by DyLight 549-conjugated H57 Fab (red) and plated onto
an MCC88-103-prepulsed planar bilayer containing I-Ek-, Cy5-labeled ICAM-1-, and CD80-GPI at the indicated concentrations. The cells were real-time imaged
by confocal microscopy at 2 or 30 min after contact. Histograms show fold fluorescence intensities of TCRb (red), CD28 (blue), CTLA-4 (yellow), and ICAM-1
(green) on the diagonal yellow lines in the DIC images. CD28 and CTLA-4 clusters are categorized into inside or outside pSMAC or none (B, CD80-GPI 80 mole-
cules/mm2, n = 28; 5/mm2, n = 42; 0/mm2, n = 30). Scale bars represent 5 mm. A representative four independent experiments is shown.
(C) AND-Tg effector T cells were plated onto an MCC88-103-prepulsed planar bilayer containing I-Ek-, ICAM-1-, and CD80-GPI at different densities and fixed at
30 min after contact. The cells were stained for CD33, CD28, and CTLA-4 and the fluorescence intensities at the cSMAC were compared with the intensities
across the entire cell-bilayer interface (CD80-GPI 80 molecules/mm2, n = 116; 20/mm2, n = 74; 5/mm2, n = 121; 1.25/mm2, n = 74; 0.32/mm2, n = 57; 0/mm2,
n = 95). *p < 0.0001 with Student’s t test. A representative two independent experiments is shown.
(D)AND-Tg Ctla4+/+or Ctla4?/?effector T cells expressing CD28-EGFPwere plated onto an MCC88-103-prepulsed planar bilayer containing I-Ek-, ICAM-1-, and
CD80-GPI and were imaged by confocal microscopy every 2 s from 3 to 6 min after contact. Fold fluorescence intensities of CD28-EGFP at the cSMAC (white
squares) were compared with those across the entire cell-bilayer interface in the graph (black, Ctla4+/+; green, Ctla4?/?). Scale bars represent 5 mm. A represen-
tative three independent experiments is shown.
Dynamics of CTLA-4-Mediated T Cell Suppression
Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc. 329
Figure 4. Structural Regulation of CTLA-4 in Its Translocation at TCR Microclusters and the cSMAC
(A) AND-Tg T cells expressing Y139A mutant CTLA-4-EGFP (green) were stained by DyLight 649-conjugated H57 Fab (red) and plated onto an MCC88-103-pre-
pulsed planar bilayer containing I-Ek-, ICAM-1-, and CD80-GPI. The cells were real-time imaged by confocal microscopy 20 min after contact. Histograms show
Dynamics of CTLA-4-Mediated T Cell Suppression
330 Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc.
clarify the mechanism of CTLA-4 inhibitory activity. The translo-
cation of CTLA-4 to the TCR MCs and the cSMAC is completely
dependent on ligand binding, as shown by the failure of CTLA-4
This was confirmed by an analysis with the mutant CTLA-4 in
MYPPPY motif (Y139A; tyrosine 139 is substituted by alanine),
which is a critical residue for ligand binding (Figure 4A). Based
on the cocrystal structure of CTLA-4 and CD80 or CD86, it has
been suggested that CTLA-4 homodimers form a lattice struc-
ture at the IS, interlocking tightly with homodimeric CD80 or
CD86 (Schwartz et al., 2001; Stamper et al., 2001). In our
experiments, no significant defect was observed in CTLA-4
accumulation at the cSMAC with the C157S mutant CTLA-4,
which normally mediates the intermolecular disulfide bonding
of CTLA-4 homodimer, and further with the N113A, N145A,
C157S (NNC) mutant involving the N-glycosylation sites, which
are critical for C157-independent dimerization (Figures 4B–4D),
45.4% ± 11.2% of CTLA-4 was accumulated at the cSMAC in
the WT, 42.7% ± 11.3% in C157S, and 53.9% ± 9.0% in NNC
mutant CTLA-4 (Darlington et al., 2005; Linsley et al., 1995).
CTLA-4 is constitutively internalized and degraded by binding
with the endocytic adaptor complex AP-2 via its tyrosine motif
the Y165G mutant CTLA-4 shows spontaneous high expression
at the cell surface. When this mutant CTLA-4 was used, CTLA-4
translocation of CD28 depending on the surface expression of
CTLA-4 (Figure 4E). These results suggest that CTLA-4 directly
blocks CD28-mediated T cell costimulation at the TCR MCs if
CTLA-4 is sufficiently expressed on the T cell surface.
of CTLA-4, the phosphorylation of this tyrosine upon TCR stim-
ulation induces the recruitment of two phosphatases, Src
homology 2 (SH2)-domain-containing tyrosine phosphatase 1
and 2 (SHP1 and SHP2) (Lee et al., 1998; Sloan-Lancaster
et al., 1997). Furthermore, the other phosphatase, PP2A, is
known to associate at the lysine-based charged motif of the
CTLA-4 tail in the absence of TCR-CTLA-4 colligation (Baroja
et al., 2002; Chuang et al., 2000). Although we tried to visualize
the translocation of these phosphatases at TCR MCs or the
cSMAC, none of these three phosphatases formed detectable
clusters (Figure S4).
CTLA-4-Mediated Inhibition of PKCq and CARMA1
Clusters at the cSMAC
The mechanisms of CD28-mediated costimulation have been
extensively analyzed in its downstream signaling molecules.
Although it is difficult to separate the TCR- and CD28-specific
pathways, some kinases and adaptors downstream of the TCR
are also important in CD28-mediated costimulation signals
(Acuto and Michel, 2003; Rudd and Schneider, 2003). PKCq is
implicated in CD28-mediated costimulation and its recruitment
to the IS is well documented (Huang et al., 2002; Monks et al.,
1998). We have previously demonstrated the initial translocation
of PKCq to TCR-CD28 MCs and the sequential accumulation at
the annular CD28 cluster within the cSMAC through the specific
association between CD28 andPKCq (Yokosuka etal., 2008).To
examine whether endogenous CTLA-4 affects the translocation
of not only CD28 but also PKCq at the cSMAC, we settled
Ctla4?/?or Cd28?/?AND-Tg T cells expressing PKCq–EGFP
on bilayers containing different densities of CD80-GPI and
analyzed the dynamic localization of PKCq in the IS in each
T cell. With high-density CD80 (80 molecules/mm2), no PKCq
cluster was observed at the Cd28?/?T cell-bilayer interface
but clusters clearly formed in both wild-type (WT; 94.7%)
and Ctla4?/?cells (85.0%). With low-density CD80 (8 mole-
cules/mm2), the majority (87%) of WT T cells, which expressed
endogenous CTLA-4, did not form PKCq clusters, whereas
55% of Ctla4?/?T cells formed clear clusters under the same
conditions (Figures 5A and 5B). These results indicate that the
translocation of PKCq to the cSMAC is interrupted by CTLA-4
of ligands and receptors.
To further analyze the CTLA-4-mediated blocking of the phys-
ical association of CD28 and PKCq and the biological T cell
responses by CD28-PKCq assembly, we prepared AND-TCR
T cell hybridomas and DC-1 expressing different amounts of
CD28 and CD80, respectively (Figures S5A and S5B). We first
examined the physical CD28-PKCq association in T cell-DC-1
conjugates. The CD28-PKCq association was partially inhibited
by CTLA-4 in the CD28 intermediate (CD28int) T cell–CD80int
DC-1 conjugates. However, this inhibition was not observed in
conjugates in which either CD28 or CD80 was highly expressed
(Figure 5C). These results are similar to those observed for PKCq
cluster formation in WT T cells on a bilayer containing high-
density CD80 (Figures 5A and 5B). Interleukin 2 (IL-2) production
is largely dependent on T cell costimulation and T cells produce
more IL-2 by exposure of the stronger costimulation, depending
on the expression levels of CD28 and CD80 (Figure S5C). Like
the CTLA-4-mediated blockage of the CD28-PKCq physical
association, IL-2 production was partially inhibited (36%) only
in CD28intT cells after stimulation by CD80intDC-1 (Figure 5D).
This inhibition was dependent on ligand binding given that no
change was observed after stimulation with immobilized CD3
and CD28 Abs or on T cells expressing the C139A mutant
CTLA-4. This inhibition was lost when CD80hiDC-1 was used
fold fluorescence intensities of TCR-b (red) and CTLA-4 (green) on the diagonal yellow lines in the DIC images. Scale bars represent 5 mm. A representative five
independent experiments is shown.
or actin (bottom). A representative two independent experiments is shown.
(C and D) AND-Tg T cells expressing WT, C157S, or NNC mutant CTLA-4-EGFP (green) were stained by DyLight 649-conjugated H57 Fab (red), plated onto
a same planar bilayer as in (A) (CD80-GPI negative, top three rows; positive, bottom three rows), and imaged at 30 min after contact. Histograms show fold fluo-
rescence intensities of TCR-b (red) and CTLA-4 (green) on the diagonal yellow lines in the DIC images. Fluorescence intensities of CTLA-4-EGFP at the cSMAC
were compared with that across the entire cell–bilayer interface in (D) (n = 30). A representative three independent experiments is shown.
(E) AND-Tg T cells expressing CD28-ECFP (cyan) and/or Y165G mutant CTLA-4-EYFP (yellow) were stained by DyLight 549-conjugated H57 Fab (red), plated
onto a same planar bilayer as in (A), and imaged at 2 min after contact. Histograms show fold fluorescence intensities of TCRb (red), CD28 (blue), and CTLA-4
(yellow) on the diagonal yellow lines in the DIC images. Scale bars represent 5 mm. A representative three independent experiments is shown.
Dynamics of CTLA-4-Mediated T Cell Suppression
Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc. 331
Dynamics of CTLA-4-Mediated T Cell Suppression
332 Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc.
(data not shown). It was also cancelled by the stronger TCR
stimulation with DC-1 prepulsed with high-dose MCC peptide
(Figure 5E). These data suggest that in a ligand-binding-depen-
dent manner, CTLA-4 inhibits the accumulation of PKCq at
the cSMAC and the CD28-PKCq physical association, resulting
in the inhibition of IL-2 production. They also suggest that
the negative regulation of T cell activation is fine-tuned by the
expression of the receptors and ligands and by the TCR signal
Because CD28-induced PKCq activation is known to lead to
activate nuclear factor kB (NF-kB) pathway, next we examined
the dynamic recruitment of a scaffolding protein, CARMA1,
which is phosphorylated by PKCq as a downstream molecule
of PKCq. CARMA1 forms the signalosome CARMA1-B cell CLL
lymphoma 10 (Bcl10)-mucosa-associated lymphoid tissue
transformation protein 1 (MALT1) complex (CBM complex),
which is critical for NF-kB activation. Expectedly, CARMA1 clus-
ters were formed and accumulated next to the CD3hiwithin the
cSMAC in the annular structure similar to PKCq (Figure 5F,
top). The formation of CARMA1 clusters was dependent on
CD28-CD80 binding and was diminished if WT, but not Y139A
mutant, CTLA-4 was overexpressed in T cells (Figures 5F and
5G; CARMA1 clusters were formed in 76% cells without CTLA-
4 overexpression, 20% with WT, and 74% with Y139A mutant
CTLA-4). CTLA-4-mediated blockade of CARMA1 clustering
was also shown by endogenous CTLA-4 (Figure 5H). CARMA1
clusters were substantially induced in Ctla4?/?T cells and
controlled by the expression of its ligand (Figure 5I). These
data coincided with the idea that PKCq functions in NF-kB acti-
vation via the CBM complex and suggested that CTLA-4 inhibits
the NF-kB signaling at the cSMAC by the blockade of CD28-
mediated CARMA1 complex formation.
CTLA-4 Ectodomain Size Controls Its Localization
in the IS and Inhibition of T Cell Activation
mulation at the cSMAC, we introduced the elongated CTLA-4
chimeras bearing long ectodomains. The IgV domain of murine
CTLA-4 was attached to human CD22 with the deletion of IgV
and several N-terminal Ig domains and tagged by EGFP, and
these chimeras were retrovirally introduced into AND-Tg
T cells from Ctla4?/?mice (Figure S6A). These mCTLA-4-
hCD22 chimeras were well expressed on the cell surface and
bound to CD80 similarly (Figure S6B). On a planar bilayer, the
chimeras with shorter ectodomains (C5-6 and C6) were translo-
cated at both TCR MCs and the cSMAC, whereas the taller ones
(C2-6 and C3-6) were neither at TCR MCs nor the cSMAC
(Figures 6A and 6B). C3-6 chimera was still colocalized with
CD80-GPI (Figure S6C) but clearly excluded outside pSMAC
(Figure 6C). Furthermore, those taller chimeras failed to exclude
PKCq from the cSMAC (Figures 6D and 6E). To examine the
blocking efficiency of each chimera in physical association
between CD28 and PKCq as well as IL-2 production, we used
AND-TCR T cell hybridomas expressing equal levels of the
flag-tagged chimeras in a similar analysis as Figures 5C–5E.
The much shorter chimeras C5-6 and C6 blocked the CD28-
PKCq association more effectively than the taller ones
(Figure S6E). Consistently, IL-2 production was more effectively
suppressed by the shorter chimeras C5-6 and C5 (Figure S6F).
This regulation is fine-tuned by the strength of the stimulus
through TCR and CD28 because the inhibition was no more
prominent when the higher dose peptide or CD80hiAPCs were
used. These results suggest that the ectodomain size of
CTLA-4 (at most three Ig domains) is critical for its translocation
into TCR MCs and the cSMAC and that the accumulation of
CTLA-4 at the cSMAC is required for the effective inhibition of
T cell activation.
Lack of PKCq Accumulation at the cSMAC in Treg Cells
Among all T cell subpopulations, Treg cells are the only popula-
tion that constitutively express a high amounts of CTLA-4
(Sakaguchi et al., 2008). Foxp3 is a master regulator of Treg cells
and also important for the in vitro Treg cell anergy. Foxp3 is sug-
gested to repress IL-2 transcription by competing with NF-AT
for binding to the IL-2 promoter (Schubert et al., 2001). However,
an analysis of Foxp3-deficient mice predicted another possible
mechanism underlying in vitro anergy. We then presumed that
CTLA-4 expressed by Treg cells might induce a negative signal
that attenuates their activation and induces in vitro anergy, and
Figure 5. CTLA-4 Blocks the Translocation of CD28-PKCq-CARMA1 at the cSMAC and Regulates T Cell Responses
(A and B) AND-Tg WT, Ctla4?/?, or Cd28?/?effector T cells expressing PKCq-EGFP were plated onto an MCC88-103-prepulsed planar bilayer containing I-Ek-,
ICAM-1-, and CD80-GPI at 80 or 8 molecules/mm2and real-time imaged by confocal microscopy 30 min after contact. The PKCq clusters were categorized into
three groups: none (no visible cluster formed), random (clusters were formed but diffused randomly), or annular (clusters were formed along the circle). (B, WT
CD80-GPI 80molecules/mm2,n=56and 8/mm2,n=60; Ctla4?/?80/mm2,n=127and 8/mm2,n=51; Cd28?/?80/mm2,n= 111and 8/mm2,n=54). *p< 0.0001with
Student’s t test; the scale bar represents 5 mm. A representative two independent experiments is shown.
(C) AND-TCR T cell hybridomas expressing PKCq-EYFP and CD28 at intermediate (CD28int) or high (CD28hi; Figure S5A) amounts were transduced by WT or
Y139A mutant CTLA-4-ECFP. The cells were stimulated by MCC88-103-prepulsed DC-1 expressing CD80 at low (CD80lo), intermediate (CD80int), or high densi-
ties (CD80hi; Figure S5B) for 5 min. Cell lysates immunoprecipitated by anti-CD28 (top two rows) or whole cell lysates (lower three rows) were blotted for PKCq
(rows 1 and 3), CD28 (rows 2 and 4), or actin (bottom row). A representative five independent experiments is shown.
(D and E) CD28intAND-TCR T cell hybridomas were stimulated for 24 hr by CD80lo, CD80int, or CD80hiDC-1 cells with or without 0.1 mM MCC88-103, anti-CD33
with or without anti-CD28, or PMA plus ionomycin in (D) or by CD80intDC-1 with different concentrations of MCC88-103 in (E). The graphs represent IL-2 produc-
tion measured by ELISA (white, control cells; gray, WT CTLA-4-trasnfected cells; black, Y139A mutant CTLA-4-transfected cells). The graph shows the means ±
SD (n = 3) of three independent experiments. A representative five independent experiments is shown.
(F and G) AND-Tg T cells expressing CARMA1-mGFP (top, green) plus WT (middle) or Y139A CTLA-4-ECFP (bottom, cyan) were stained by DyLight 649-conju-
gated H57 Fab (red), plated onto a same planar bilayer as in (A) (CD80-GPI 150/mm2), and real-time imaged by confocal microscopy 30 min after contact. The
percentage of the cells forming CARMA1 clusters is shown in (G) (n = 50) A representative two independent experiments is shown.
(HandI)AND-TgCtla4+/+(left)or?/?Tcells(right) expressing CARMA1-mGFP(green)werestainedbyDyLight649-conjugated H57Fab(red), platedontoasame
planar bilayer as in (A) (CD80-GPI 150/mm2, top; 38/mm2, bottom), and real-time imaged by confocal microscopy 30 min after contact. The percentage of the cells
forming CARMA1 clusters is shown in (I) (n = 50) A representative four independent experiments is shown.
Dynamics of CTLA-4-Mediated T Cell Suppression
Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc. 333
Figure 6. The Size of CTLA-4 Ectodomain Controls Its Accumulation at TCR MCs and the cSMAC and Inhibitory Function
(A and B)AND-Tg Ctla4?/?T cells werereconstituted by mCTLA-4-hCD22 chimeras (Figure S6A) tagged by EGFP (green).The cells were stained by DyLight 649-
conjugated H57 Fab (red), plated onto an MCC88-103-prepulsed planar bilayer containing I-Ek- and ICAM-1- (top) plus CD80-GPI (bottom four rows), and real-
time imaged by confocal microscopy 2 (left) or 30 min after contact (right). Percentage of the cells forming mCLTA-4-hCD22 clusters within the cSMAC in (A) is
shown in (B) (n = 50). Scale bars represent 5 mm. A representative three independent experiments is shown.
(C) T cells in (A) were plated onto the same planar bilayer as in (A) containing Cy5-labeled ICAM-1-GPI (cyan) and imaged by confocal microscopy at 30 min after
contact. Scale bars represent 5 mm. A representative two independent experiments is shown.
Dynamics of CTLA-4-Mediated T Cell Suppression
334 Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc.
examined whether CTLA-4 blocks the CD28 clustering at the
cSMAC in Treg cells. Treg cells were collected from AND-Tg
mice crossed with Foxp3GFPreporter mice (Fontenot et al.,
2005). Whereas Foxp3GFP+and Foxp3GFP?T cells expressed
similar amounts of CD28, Foxp3GFP+T cells expressed much
higher amounts of CTLA-4 than Foxp3GFP?cells (Figure S7A).
As expected, Foxp3GFP?T cells displayed clear clustering of
CD28 at the cSMAC, particularly at the outer region, in
a ligand-dependent manner (Figure 7A). In contrast, Foxp3GFP+
T cells formed tiny and barely visible CD28 clusters (Fig-
ure S7B; 24.9% ± 6.2% of CD28 intensity was accumulated at
the cSMAC in Foxp3GFP?T cells; 18.1% ± 5.0% in Foxp3GFP+
T cells), Instead, Foxp3GFP+T cells, but not Foxp3GFP?cells,
induced the intensive accumulation of CTLA-4 at the cSMAC
(Figure S7B; 22.2% ± 9.4% of CTLA-4 intensity was accumu-
lated at the cSMAC in Foxp3GFP?T cells; 38.1% ± 11.7% in
Foxp3GFP+T cells), suggesting that CTLA-4 blocks CD28 clus-
ters at the cSMAC in Treg cells. Next, we examined whether
CTLA-4 blocked the accumulation of PKCq at the cSMAC in
AND-Tg Foxp3GFP+and Foxp3GFP?T cells on a planar bilayer.
In the presence of CD80, Foxp3GFP?T cells showed the forma-
tion of annular PKCq clusters at the cSMAC, but Foxp3GFP+
T cells did not (Figure 7B and Figure S7C).
Naive T cells at the periphery can also acquire Foxp3 expres-
sion in the experimental setting in vitro and during in vivo
responses. AND-Tg naive T cells apparently express Foxp3 after
TCR stimulation in the presence of transforming growth factor
b (TGF-b) and IL-2 (Figure S7D) (Chen et al., 2003), and these
induced Treg (iTreg) cells inhibited the proliferation of responder
naive T cells in a mixed lymphocyte culture assay (Figure S7E).
The iTreg cells also showed in vitro anergic response to the
secondary antigen stimulation, but not to the stimulation with
phorbol 12-myristate 13-acetate (PMA) plus ionomycin as
compared with the normal response by AND-Tg effector
T cells generated without TGF-b (neutral) (Figure S7F). In the
planar bilayer system, both AND-Tg neutral effector cells and
iTreg cells formed TCR MCs, which colocalized with PKCq at
the initial contact (Figure S7G). In contrast, whereas neutral
effector cells formed annular PKCq clusters surrounding the
CD3hiregion, iTreg cells showed fewer PKCq clusters, similar
to the Foxp3GFP+Treg cells (Figures 7C and 7D). We further
examined the functional importance of CTLA-4 localization
at the cSMAC in in vitro Treg cell anergy. The iTreg cells devel-
oped from AND-Tg Ctla4?/?cells showed higher response
to the secondary stimulation than Ctla4+/+cells (Figure 7E).
When mCTLA-4-hCD22 chimeras were expressed on iTreg
cells, similarly to the effector T cells in Figure 6, the shorter
mCTLA-4-hCD22 chimera C6, but not the taller C3-6, was trans-
located into TCR MCs and the cSMAC (Figures 7F and 7G)
and more effectively blocked the proliferation of iTreg cells
(Figure 7H). These results suggest that the constitutive
expression of CTLA-4 in Treg cells results in the reduced
accumulation of CD28 and PKCq at the cSMAC and that the
cSMAC is the regulatory region for CD28-CTLA-4 competition
in Treg cells.
We have demonstrated here the dynamic competition between
positive and negative costimulatory receptors CD28 and
CTLA-4 for the regulation of T cell activation and the importance
and Saito, 2009). Our data suggest that T cell costimulation is
under finely tuned regulation strictly based on the expression
of CD28 and CTLA-4 on T cells, CD80, CD86, and MHCp on
APCs. We have shown that this fine-tuning is spatiotemporally
operated by the competition between CD28 and CTLA-4 at the
IS, especially at the CD3lo‘‘signaling cSMAC.’’ Under physiolog-
ical conditions, the expression of receptors and ligands changes
in the activation status of both T cells and APCs and in the
subsets of T cells.
The IS is considered as an important structure for T cell-APC
communication and for the functional maturation and activation
of T cells. Although the geometric patterning of the IS has been
intensively characterized, the real function of the IS, particularly
its segregation into SMACs, remains unclear. We and others
have revised the concept of the IS by showing that the TCR
MC is the site responsible for T cell activation (Bunnell et al.,
2002; Campi et al., 2005; Yokosuka et al., 2005). TCR MCs,
interface between a T cell and an APC or a planar bilayer and are
translocated to the center of the interface to form a cSMAC.
Thereafter, small TCR MCs containing Zap70 and SLP-76 are
continuously generated in the nascent contact region at the
periphery of the IS and are sequentially translocated to the
center; therefore, proximal TCR signals are induced at the TCR
MCs and not at the cSMACs.
In contrast, we recently discovered two separate subregions
of the cSMAC, the CD3loand CD3hiregions, which differ in their
densities of TCR-CD3 complexes (Yokosuka and Saito, 2009,
2010). Upon T cell costimulation, CD28 accumulates in the
CD3loregion in a ligand-dependent manner and specifically
recruits PKCq to the CD28hiCD3loregion (Yokosuka et al.,
2008; Yokosuka and Saito, 2009). This molecular dynamism
strongly suggests a function for the CD3loregion in T cell costi-
mulation and is also consistent with the historically well-known
role of PKCq as a marker for the cSMAC. We have also demon-
strated here that CTLA-4 initially translocates to the TCR MCs,
ultimately accumulates in the CD3loregion of the cSMAC, and
interferes with CD28 and PKCq translocation.
The function of the cSMAC has been extensively debated. An
analysis of CD2AP-deficient T cells suggested that the cSMAC
negatively regulates T cell activation by TCR degradation (Lee
et al., 2003). The cSMAC has been reported to be a site of endo-
cytosis or degradation and negative regulation, on the basis of
the observation that the cSMAC is rich in lysobisphosphatidic
acid, the membrane phsophatase CD45 (Varma et al., 2006),
and a component of the endosomal sorting complex required
for transport (ESCRT) (Vardhana et al., 2010). The BCR-medi-
ated endocytosis of antigens at the cSMAC is also obvious in
B cells (Fleire et al., 2006). The cSMAC in T cells might function
(D and E) T cells in (A) were fixed at 30 min after contact, stained for PKCq (cyan), and imaged by confocal microscopy. The percentage of the cells bearing PKCq
clusters in (D) is shown in (E) (n = 50). Scale bars represent 5 mm. A representative two independent experiments is shown.
Dynamics of CTLA-4-Mediated T Cell Suppression
Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc. 335
Figure 7. CTLA-4 Translocation and Loss of CD28 and PKCq Accumulation at the cSMAC in Treg Cells
(A and B) AND-Tg Foxp3GFP?or+effector T cells (Figure S7A) were stained by DyLight 488- (A, green) or 549-conjugated H57 Fab (B, red) and plated onto an
MCC88-103-prepulsed planar bilayer containing I-Ek-, ICAM-1-, and CD80-GPI. The cells were fixed 30 min after contact, stained for CD28 (A, red) and CTLA-4
(A, cyan) or PKCq (B, cyan), and imaged by confocal microscopy. Histograms show the fold fluorescence intensities of TCR-b (green), CD28 (red), and CTLA-4
(cyan) in (A), or GFP (green), TCR-b (red), and PKCq (cyan) in (B) on the diagonal yellow lines in the DIC images. Statistical analysis is shown in Figures S7B and
S7C. Scale bars represent 5 mm. A representative three independent experiments is shown.
(C and D) AND-Tg neutral effector T cells (top) or iTreg cells (bottom) expressing PKCq-EGFP (green) were stained by DyLight 649-conjugated H57 Fab (red),
plated on the same planar bilayer as in (A), and real-time imaged by confocal microscopy 2 (Figure S7G) or 30 min after contact. PKCq clusters 30 min
after contact were categorized as in Figure 5B in (C) (neutral, n = 53; iTreg, n = 44). Scale bars represent 5 mm. A representative three independent experiments
Dynamics of CTLA-4-Mediated T Cell Suppression
336 Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc.
in the termination of the TCR signals via the endocytotic and
dephosphorylation machinery. In contrast, weak but distinct
amounts of phosphoproteins have been detected at the cSMAC
during weak and delayed responses, suggesting continuing
TCR-activation signals at the cSMAC (Cemerski et al., 2008).
Taken together, these data suggest that the cSMAC has dual
functions in the positive and negative regulation of TCR signals
through spatially differential subregions.
The critical event in the signaling at the cSMAC is the specific
recruitment of PKCq and CARMA1 and the consequent blocking
oftheir clustering byCTLA-4. After theinitiation ofa Tcell-bilayer
contact, PKCq is slowly recruited at the TCR MCs and exhibits
delayed accumulation in the cSMAC compared with the accu-
mulation of other TCR-proximal signaling molecules. CARMA1
clustering is induced further later than that of PKCq (data not
shown). Considering that PKCq phosphorylates CARMA1 and
Bcl10, it is possible that CD28-PKCq accumulation at the
cSMAC is involved in a CARMA1-containing signaling complex
that leads to the activation of NF-kB. Because CD28 and
CTLA-4 reciprocally regulate PKCq accumulation at the cSMAC,
they could be regulators of NF-kB signals. Taken together, our
findings define several functional subregions of the IS that allow
theactivation of distinct signaling pathways: (1) TCRMCs, asthe
TCR-proximal signaling complexes, mainly regulate calcium
flux, the mitogen-activated protein kinase (MAPK), or extracel-
lular signal-regulated kinase (Erk) signaling cascade, and actin
polymerization; (2) CD3losignaling cSMAC is a costimulatory
signaling site that may regulate NF-kB signaling; and (3) the
CD3hiregion of cSMAC terminates TCR signals by endocytosis
The mechanisms that underlie the CTLA-4-mediated inhibition
of T cell activation have been a matter of great debate. CTLA-4
contains the YVKM motif and the lysine-based charged motif,
which are associated by several phosphatases, such as SHP1,
SHP2 (Lee et al., 1998; Sloan-Lancaster et al., 1997), and
PP2A (Baroja et al., 2002; Chuang et al., 2000). We analyzed
the possible colocalization of CTLA-4 and SHP1, SHP2, or
PP2A in both primary T cells and T cell hybridomas. However,
we found that none of them formed clusters or was observed
at CTLA-4 clusters. It may suggest that such phosphatase
recruitment could be highly transient or could occur in small
populations or at low levels of receptors. In addition to the
ectodomain competition, a ligand-independent inhibitory func-
tion via cytoplasmic signaling has been also demonstrated
(Vijayakrishnan et al., 2004). The downstream of CTLA-4 has
still been open to dispute particularly in the induced clustering
of CTLA-4 by its binding to CD80 or CD86 and in the actual
requirement of associated phosphatases in CTLA-4 function
The structural characterization of CTLA-4 has provided an
additional basis for the ligand binding. Whereas both CTLA-4
and CD28 form homodimers, only CTLA-4 is bivalent, suggest-
ing the multivalent binding of CTLA-4-CD80 or CD86 with ampli-
fied affinity. Structural analysis of the CTLA-4-CD80 or CD86
demonstrated the crystal lattice structures of these complexes
(Schwartz et al., 2001; Stamper et al., 2001). However, our anal-
ysis demonstrated the CTLA-4 MCs and the annular CTLA-4
clusters at the cSMAC, suggesting that these CTLA-4 clusters
could be constructed by a minimal crystal lattice, which then
excludes CD28 from the TCR MCs and the cSMAC. On the
contrary, the ectodomain size of CTLA-4 is structurally critical
for its behavior in the IS. The size-based segregation of cell-
surface molecules is an original conception after the IS was
discovered and then the exclusion of CD45 to the outside of
pSMAC, distal SMAC, further confirms this idea. It has been
demonstrated that the elongated MHCp ectodomain reduced
TCR triggering without affecting TCR-MHCp ligation (Choudhuri
et al., 2005). Our study using mCTLA-4-hCD22 chimeras with
elongated ectodomain suggested that the spatial segregation
of tall CTLA-4-CD80 from relatively small TCR-MHCp and
CD28-CD80 resulted in less efficient inhibition of T cell activa-
tion. The tandem three Ig domain is a maximal size for the
functional CTLA-4 for both cSMAC recruitment and T cell
suppression and this size restriction appears to be a general
rule for most of costimulatory receptors.
The CTLA-4-mediated attenuation of T cell activation has
already been applied in clinical therapies, with anti-CTLA-4
Abs or CTLA-4 Ig fusion proteins. The Abs are used to attenuate
the inhibition of Tcell activation, asin tumor immunotherapies by
enhancing cytotoxicity and cytokine secretion, and the Ig fusion
proteins are used in allergic and autoimmune diseases to inhibit
CD28-CD80 and CD86 binding (Egen et al., 2002; Linsley and
Nadler, 2009). Whereas the biological relevance and clinical
impacts are quite dramatic in these applications of CD28-
CTLA-4 costimulation, the precise mechanisms involved are still
unclear. Our finding of the differential localization of CD28 and
CTLA-4 should extend our understanding of T cell costimulation
and allow the development of more effective strategies for
Abs and reagents were purchased from the following suppliers: fluorescein
isothiocyanate (FITC)-anti-mCD4 (RM4-5), phycoerythrin (PE)-anti-mCD28
(37.51), PE-anti-mCD62L (MEL-16), PE-anti-mCD80 (16-10A1), PE-anti-
mFoxp3 (FJK-16s), and PE-isotype-matched control IgG were from
eBioscience; Alexa Fluor 488-anti-mCD33 (17A2) was from BD; goat poly-
clonal anti-CD28 (M-20) and rabbit polyclonal anti-PKCq (C-18) were from
Santa Cruz Biotechnology; mouse monoclonal anti-actin (AC-40) was
from Sigma, and HRP-labeled anti-GFP was from Miltenyi Biotec. Phorbol
12-myristate 13-acetate (PMA) was from Sigma; ionomycin was from Calbio-
chem. Anti-mCD86 (PO3) was previously described (Nuriya et al., 1996).
(E) AND-Tg Ctla4+/+(gray) or?/?(black) iTreg cells were stimulated by irradiated B10.BR whole splenocytes with 0.3 mM or without MCC88-103 for 2 days and
counted for cell number. A representative two independent experiments is shown.
(F and G) iTreg cells were reconstituted by C3-6 (row 1and 3) or C6 mCTLA-4-hCD22 (row 2 and 4)taggedwith EGFP (green),stained by DyLight 649-conjugated
H57 Fab (red), plated onto the same planar bilayer as in (A), and real-time imaged by confocal microscopy 2 (left) or 30 min after contact (right). The percentage of
the cells forming mCLTA-4-hCD22 clusters within the cSMAC is shown in (G) (n = 50). Scale bars represent 5 mm. A representative two independent experiments
(H)AND-TgiTreg cells in(F, noreconstitution, white;C3-6, gray; C6,black) werestimulated as in(E)with1,0.1, or 0mMMCC88-103 (left)or immobilizedanti-CD3
(2C11) and CD28 (PV-1) for 2 days and counted for cell number. A representative two independent experiments is shown.
Dynamics of CTLA-4-Mediated T Cell Suppression
Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc. 337
B cell hybridomas producing anti-mCD33 (145-2C11) and anti-mCTLA-4
(UC10.4F10.11) were provided by J. Bluestone (University of California, San
Francisco, CA); anti-mCD28 (PV-1) was provided by R. Abe (Science Univer-
sity ofTokyo, Japan); anti-mTCRb(H57-597) wasprovidedby R.T. Kubo(Cytel
Corp., CA); anti-I-Ek(14-4-4), anti-mCD80 (16-10A1), and anti-mICAM-1 (YN1/
1.7.4) was provided by M.L. Dustin (New York University, NY), plasmid for
murine CD80-human IgG fusion protein were provided by T. Uede (Hokkaido
University, Japan). H57 was digested with immobilized papain and the Fab
fragment was confirmed by SDS-PAGE (Takara Bio). The fluorescent tags,
DyLight 488, 549, and 649, were conjugated to Abs with DyLight labeling
kits (Thermo Scientific).
Mice and Cells
AND-Tg mice in a Rag2?/?background were provided by R.N. Germain
(National Institutes of Health, MD); Ctla4?/?mice by T.W. Mak (Campbell
Family Cancer Research Institute, Canada). Cd28?/?and Foxp3GFPreporter
SLC. The DC line (DC-1) was provided by J. Kaye. The T cell hybridoma
expressing AND-TCR was established previously (Yokosuka et al., 2008).
Primary Cell Culture and Transduction
Expression constructs were transiently transduced into Phoenix packaging
cells (provided by G. Norlan, Stanford University, CA) with LipofectAmine
Plus (Invitrogen). Retroviral supernatants were concentrated 10-fold by centri-
fugation at 8000 3 g for 12 hr. CD4+T cells were purified from AND-Tg mice
in Rag2?/?, Cd28?/?Rag2?/?, Ctla4?/?Rag2?/?, and Foxp3GFP+Rag2+/?back-
grounds and stimulated with 5 mM MCC88-103 (ANERADLIAYLKQATK) and
irradiated spleen cells from B10.BR mice. One day after stimulation, the cells
were suspended in retroviral supernatant with 10 mg/ml polybrene (Sigma) and
200 U/ml human recombinant IL-2 (Ajinomoto) and centrifuged at 1000 3 g for
90 min at 37?C. On day 2 or later, the cells were sorted with FACSAria (BD) for
obtaining populations with homogeneous fluorescence intensity, which were
maintained in RPMI1640 medium containing 10% fetal calf serum (FCS) and
Cellswereimaged withTIRFMaspreviouslydescribed (Tokunagaetal.,2008).
Two solid-state laser systems (488 nm, 20 mW, Sapphire 488-20-OPS,
Coherent; and 558 nm, 10 mW, YA11, Megaopto, Japan) and an inverted
microscope (IX-81, Olympus, Japan) were used. The images were captured
with a Back-Thinned Electron Multiplier Charge-Coupled Device Camera
(C9100-13; Hamamatsu Photonics).
Supplemental Information includes Supplemental Experimental Procedures,
seven figures, and four movies and can be found with this article online at
for secretarial assistance. This work was supported by a Grant-in-Aid for
and Technology of Japan (T.Y., K.S.S., M.T., and T.S.).
Received: February 23, 2010
Revised: July 11, 2010
Accepted: August 12, 2010
Published online: September 23, 2010
support for TCR signalling. Nat. Rev. Immunol. 3, 939–951.
Alegre, M.L., Frauwirth, K.A., and Thompson, C.B. (2001). T-cell regulation by
CD28 and CTLA-4. Nat. Rev. Immunol. 1, 220–228.
Baroja, M.L., Vijayakrishnan, L., Bettelli, E., Darlington, P.J., Chau, T.A., Ling,
V., Collins, M., Carreno, B.M., Madrenas, J., and Kuchroo, V.K. (2002). Inhibi-
tion of CTLA-4 function by the regulatory subunit of serine/threonine phospha-
tase 2A. J. Immunol. 168, 5070–5078.
Bunnell, S.C., Hong, D.I., Kardon, J.R., Yamazaki, T., McGlade, C.J., Barr,
V.A., and Samelson, L.E. (2002). T cell receptor ligation induces the formation
of dynamically regulated signaling assemblies. J. Cell Biol. 158, 1263–1275.
Campi, G.,Varma,R.,and Dustin, M.L.(2005). Actin and agonist MHC-peptide
complex-dependent T cell receptor microclusters as scaffolds for signaling.
J. Exp. Med. 202, 1031–1036.
Cemerski, S., Das, J., Giurisato, E., Markiewicz, M.A., Allen, P.M.,
signaling and degradation at the center of the immunological synapse is deter-
mined by antigen quality. Immunity 29, 414–422.
Chen,W., Jin,W.,Hardegen,N., Lei,K.J.,Li,L.,Marinos, N., McGrady, G.,and
Wahl, S.M. (2003). Conversion of peripheral CD4+CD25- naive T cells to
CD4+CD25+ regulatory T cells by TGF-beta induction of transcription factor
Foxp3. J. Exp. Med. 198, 1875–1886.
Chikuma, S., Imboden, J.B., and Bluestone, J.A. (2003). Negative regulation of
T cell receptor-lipid raft interaction by cytotoxic T lymphocyte-associated
antigen 4. J. Exp. Med. 197, 129–135.
Choudhuri, K., Wiseman, D., Brown, M.H., Gould, K., and van der Merwe, P.A.
peptide-MHC ligand. Nature 436, 578–582.
Chuang, E., Fisher, T.S., Morgan, R.W., Robbins, M.D., Duerr, J.M., Vander
Heiden, M.G., Gardner, J.P., Hambor, J.E., Neveu, M.J., and Thompson,
C.B. (2000). The CD28 and CTLA-4 receptors associate with the serine/threo-
nine phosphatase PP2A. Immunity 13, 313–322.
Darlington, P.J., Kirchhof, M.G., Criado, G., Sondhi, J., and Madrenas, J.
(2005). Hierarchical regulation of CTLA-4 dimer-based lattice formation and
its biological relevance for T cell inactivation. J. Immunol. 175, 996–1004.
Dustin, M.L. (2009). The cellular context of T cell signaling. Immunity 30,
Egen, J.G., and Allison, J.P. (2002). Cytotoxic T lymphocyte antigen-4
accumulation in the immunological synapse is regulated by TCR signal
strength. Immunity 16, 23–35.
Egen, J.G., Kuhns, M.S., and Allison, J.P. (2002). CTLA-4: New insights into
its biological function and use in tumor immunotherapy. Nat. Immunol. 3,
Fleire, S.J., Goldman, J.P., Carrasco, Y.R., Weber, M., Bray, D., and Batista,
F.D. (2006). B cell ligand discrimination through a spreading and contraction
response. Science 312, 738–741.
Fontenot, J.D., Rasmussen, J.P., Gavin, M.A., and Rudensky, A.Y. (2005).
A function for interleukin 2 in Foxp3-expressing regulatory T cells. Nat. Immu-
nol. 6, 1142–1151.
Fos, C., Salles, A., Lang, V., Carrette, F., Audebert, S., Pastor, S., Ghiotto, M.,
Olive, D., Bismuth, G., and Nune `s, J.A. (2008). ICOS ligation recruits the
p50alpha PI3K regulatory subunit to the immunological synapse. J. Immunol.
Grakoui, A., Bromley, S.K., Sumen, C., Davis, M.M., Shaw, A.S., Allen, P.M.,
and Dustin, M.L. (1999). The immunological synapse: A molecular machine
controlling T cell activation. Science 285, 221–227.
H.M. (2002). CD28 plays a critical role in the segregation of PKC theta within
the immunologic synapse. Proc. Natl. Acad. Sci. USA 99, 9369–9373.
Iida, T., Ohno, H., Nakaseko, C., Sakuma, M., Takeda-Ezaki, M., Arase, H.,
Kominami, E., Fujisawa, T., and Saito, T. (2000). Regulation of cell surface
expression of CTLA-4 by secretion of CTLA-4-containing lysosomes upon
activation of CD4+ T cells. J. Immunol. 165, 5062–5068.
Lee, K.M., Chuang, E., Griffin, M., Khattri, R., Hong, D.K., Zhang, W., Straus,
D., Samelson, L.E., Thompson, C.B., and Bluestone, J.A. (1998). Molecular
basis of T cell inactivation by CTLA-4. Science 282, 2263–2266.
Dynamics of CTLA-4-Mediated T Cell Suppression
338 Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc.
Lee, K.H., Dinner, A.R., Tu, C., Campi, G., Raychaudhuri, S., Varma, R., Sims, Download full-text
T.N., Burack,W.R., Wu, H., Wang, J., et al.(2003). The immunological synapse
balances T cell receptor signaling and degradation. Science 302, 1218–1222.
Linsley, P.S., and Nadler, S.G. (2009). The clinical utility of inhibiting CD28-
mediated costimulation. Immunol. Rev. 229, 307–321.
Linsley, P.S., Nadler, S.G., Bajorath, J., Peach, R., Leung, H.T., Rogers, J.,
Bradshaw, J., Stebbins, M., Leytze, G., Brady, W., et al. (1995). Binding stoi-
chiometry of the cytotoxic T lymphocyte-associated molecule-4 (CTLA-4).
A disulfide-linked homodimer binds two CD86 molecules. J. Biol. Chem.
Linsley, P.S., Bradshaw, J., Greene, J., Peach, R., Bennett, K.L., and Mittler,
R.S. (1996). Intracellular trafficking of CTLA-4 and focal localization towards
sites of TCR engagement. Immunity 4, 535–543.
Monks, C.R., Freiberg, B.A., Kupfer, H., Sciaky, N., and Kupfer, A. (1998).
Three-dimensional segregation of supramolecular activation clusters in
T cells. Nature 395, 82–86.
Nguyen, K., Sylvain, N.R., and Bunnell, S.C. (2008). T cell costimulation via the
integrin VLA-4 inhibits the actin-dependent centralization of signaling micro-
clusters containing the adaptor SLP-76. Immunity 28, 810–821.
Nuriya, S., Yagita, H., Okumura, K., and Azuma, M. (1996). The differential role
of CD86 and CD80 co-stimulatory molecules in the induction and the effector
phases of contact hypersensitivity. Int. Immunol. 8, 917–926.
Pentcheva-Hoang, T., Egen, J.G., Wojnoonski, K., and Allison, J.P. (2004).
B7-1 and B7-2 selectively recruit CTLA-4 and CD28 to the immunological
synapse. Immunity 21, 401–413.
Pentcheva-Hoang, T., Chen, L., Pardoll, D.M., and Allison, J.P. (2007).
Programmed death-1 concentration at the immunological synapse is deter-
mined by ligand affinity and availability. Proc. Natl. Acad. Sci. USA 104,
Rudd, C.E., and Schneider, H. (2003). Unifying concepts in CD28, ICOS and
CTLA4 co-receptor signalling. Nat. Rev. Immunol. 3, 544–556.
Saito, T., and Yokosuka, T. (2006). Immunological synapse and microclusters:
the site for recognition and activation of T cells. Curr. Opin. Immunol. 18,
Sakaguchi, S., Yamaguchi, T., Nomura, T., and Ono, M. (2008). Regulatory
T cells and immune tolerance. Cell 133, 775–787.
Schubert, L.A., Jeffery, E., Zhang, Y., Ramsdell, F., and Ziegler, S.F. (2001).
Scurfin (FOXP3) acts as a repressor of transcription and regulates T cell acti-
vation. J. Biol. Chem. 276, 37672–37679.
Schwartz, J.C., Zhang, X., Fedorov, A.A., Nathenson, S.G., and Almo, S.C.
(2001). Structural basis for co-stimulation by the human CTLA-4/B7-2
complex. Nature 410, 604–608.
Sharpe, A.H. (2009). Mechanisms of costimulation. Immunol. Rev. 229, 5–11.
Shiratori, T., Miyatake, S., Ohno, H., Nakaseko, C., Isono, K., Bonifacino, J.S.,
and Saito, T. (1997). Tyrosine phosphorylation controls internalization of
CTLA-4 by regulating its interaction with clathrin-associated adaptor complex
AP-2. Immunity 6, 583–589.
Sloan-Lancaster, J., Zhang, W., Presley, J., Williams, B.L., Abraham, R.T.,
Lippincott-Schwartz, J., and Samelson, L.E. (1997). Regulation of ZAP-70
intracellular localization: visualization with the green fluorescent protein.
J. Exp. Med. 186, 1713–1724.
Stamper, C.C., Zhang, Y., Tobin, J.F., Erbe, D.V., Ikemizu, S., Davis, S.J.,
Stahl, M.L., Seehra, J., Somers, W.S., and Mosyak, L. (2001). Crystal structure
of the B7-1/CTLA-4 complex that inhibits human immune responses. Nature
Tivol, E.A., Borriello, F., Schweitzer, A.N., Lynch, W.P., Bluestone, J.A., and
Sharpe, A.H. (1995). Loss of CTLA-4 leads to massive lymphoproliferation
and fatal multiorgan tissue destruction, revealing a critical negative regulatory
role of CTLA-4. Immunity 3, 541–547.
Tokunaga, M., Imamoto, N., and Sakata-Sogawa, K. (2008). Highly inclined
thin illumination enables clear single-molecule imaging in cells. Nat. Methods
Vardhana, S., Choudhuri, K., Varma, R., and Dustin, M.L. (2010). Essential role
of ubiquitin and TSG101 protein in formation and function of the central supra-
molecular activation cluster. Immunity 32, 531–540.
Varma, R., Campi, G., Yokosuka, T., Saito, T., and Dustin, M.L. (2006). T cell
receptor-proximal signals are sustained in peripheral microclusters and termi-
nated in the central supramolecular activation cluster. Immunity 25, 117–127.
Vijayakrishnan, L., Slavik, J.M., Ille ´s, Z., Greenwald, R.J., Rainbow, D., Greve,
B., Peterson, L.B., Hafler, D.A., Freeman, G.J., Sharpe, A.H., et al. (2004). An
autoimmune disease-associated CTLA-4 splice variant lacking the B7 binding
domain signals negatively in T cells. Immunity 20, 563–575.
Waterhouse, P., Penninger, J.M., Timms, E., Wakeham, A., Shahinian, A., Lee,
disorders with early lethality in mice deficient in Ctla-4. Science 270, 985–988.
Yokosuka, T., and Saito, T. (2009). Dynamic regulation of T-cell costimulation
through TCR-CD28 microclusters. Immunol. Rev. 229, 27–40.
Yokosuka, T., and Saito, T. (2010). The immunological synapse, TCR micro-
clusters, and T cell activation. Curr. Top. Microbiol. Immunol. 340, 81–107.
Yokosuka, T., Sakata-Sogawa, K., Kobayashi, W., Hiroshima, M., Hashimoto-
Tane, A., Tokunaga, M., Dustin, M.L., and Saito, T. (2005). Newly generated
T cell receptor microclusters initiate and sustain T cell activation by recruit-
ment of Zap70 and SLP-76. Nat. Immunol. 6, 1253–1262.
Yokosuka, T., Kobayashi, W., Sakata-Sogawa, K., Takamatsu, M., Hashi-
moto-Tane, A., Dustin, M.L., Tokunaga, M., and Saito, T. (2008). Spatiotem-
kinase C theta translocation. Immunity 29, 589–601.
Zhang, Y., and Allison, J.P. (1997). Interaction of CTLA-4 with AP50, a clathrin-
coated pit adaptor protein. Proc. Natl. Acad. Sci. USA 94, 9273–9278.
Dynamics of CTLA-4-Mediated T Cell Suppression
Immunity 33, 326–339, September 24, 2010 ª2010 Elsevier Inc. 339