The Role of ICOS in the CXCR5?Follicular B Helper
T Cell Maintenance In Vivo1
Hisaya Akiba,2* Kazuyoshi Takeda,* Yuko Kojima,‡Yoshihiko Usui,*§Norihiro Harada,*†
Tomohide Yamazaki,* Juan Ma,* Katsunari Tezuka,¶Hideo Yagita,* and Ko Okumura*
ICOS is a new member of the CD28 family of costimulatory molecules that is expressed on activated T cells. Its ligand B7RP-1
is constitutively expressed on B cells. Although the blockade of ICOS/B7RP-1 interaction inhibits T cell-dependent Ab production
and germinal center formation, the mechanism remains unclear. We examined the contribution of ICOS/B7RP-1 to the generation
of CXCR5?follicular B helper T (TFH) cells in vivo, which preferentially migrate to the B cell zone where they provide cognate
help to B cells. In the spleen, anti-B7RP-1 mAb-treated or ICOS-deficient mice showed substantially impaired development of
CXCR5?TFHcells and peanut agglutinin?germinal center B cells in response to primary or secondary immunization with SRBC.
Expression of CXCR5 on CD4?T cells was associated with ICOS expression. Adoptive transfer experiments showed that the
development of CXCR5?TFHcells was enhanced by interaction with B cells, which was abrogated by anti-B7RP-1 mAb treatment.
The development of CXCR5?TFHcells in the lymph nodes was also inhibited by the anti-B7RP-1 mAb treatment. These results
indicated that the ICOS/B7RP-1 interaction plays an essential role in the development of CXCR5?TFHcells in vivo. The Journal
of Immunology, 2005, 175: 2340–2348.
switching and affinity maturation of the Abs produced by B cells,
and the development of memory B cells or plasma cells occur
within the GC. These B cell maturational processes require cog-
nate help provided by CD4?T cells (1). It is well known that both
CD28/B7 (CD80 and CD86) and CD40/CD40L (CD154) interac-
tions are required for optimal activation of CD4?T cells and B
cells to support GC formation (2, 3). Recently, ICOS, a homologue
of CD28, was cloned (4–6) and shown to be also involved in the
GC development (7, 8). Expression of ICOS is restricted to acti-
vated T cells (4, 9). Its ligand B7RP-1 (also known as B7h, B7-H2,
GL50, and LICOS) (9–13) was identified as the third member of
the B7 family, which is constitutively expressed on B cells, mac-
rophages, and dendritic cells. In certain circumstances, ICOS sig-
naling regulates either Th1 or Th2 cell differentiation (7, 8). More-
over, ICOS/B7RP-1 interaction may be involved in T/B cell
interaction because ICOS is expressed on GC T cells and B7RP-1
is expressed on resting B cells (4). Transgenic mice expressing a
soluble B7RP-1-Ig fusion protein, which could engage ICOS sig-
naling, were characterized by lymphoid hyperplasia in the spleen,
-dependent humoral immune responses are characterized
by the development of germinal center (GC)3in B cell
follicles of the secondary lymphoid organs. Isotype
lymph nodes (LN), and Peyer’s patches, and high levels of serum
IgG (9). Moreover, it has been shown that ICOS- or B7RP-1-
deficient mice developed fewer and smaller GC in response to
immunization (14–19). These mice had consistently lower levels
of serum IgG and showed a defect in IgG1 Ab production in re-
sponse to T-dependent Ags, whereas the responses against T-in-
dependent Ags were normal. These results have indicated that the
ICOS/B7RP-1 interaction regulates the GC development and Ab
production, but the exact mechanism remains unclear.
We speculated that the ICOS/B7RP-1 interaction might be re-
quired to up-regulate the expression of a chemokine receptor
CXCR5 on CD4?T cells. CXCR5 confers responsiveness to B
lymphocyte chemokine (CXCL13), which is produced by follicu-
lar stroma cells in the spleen, LN, and Peyer’s patches (20, 21).
CXCR5 is constitutively expressed by circulating B cells and is
required for their migration into B cell follicles in the secondary
lymphoid organs (22, 23). A subset of CD4?T cells also express
CXCR5, which mediates their migration to the B cell follicles
where they provide cognate help to B cells (24–27). Thus,
CXCR5?CD4?T cells are referred to as follicular B helper T
(TFH) cells (25, 28). Previous studies have implicated OX40/OX40
ligand (OX40L) interaction, a pair of the TNFR/TNF family mem-
bers (29), in the expression of CXCR5 on CD4?T cells (30, 31).
It has been also reported that OX40L-transgenic mice, expressing
a large amount of OX40L on dendritic cells, developed an in-
creased number of CD4?T cells in the B cell follicles of second-
ary lymphoid organs in response to immunization (32). Therefore,
in this study, we compared the contributions of ICOS/B7RP-1 and
OX40/OX40L to the development of CXCR5?TFHcells and GC
B cells. Our present results indicated that the ICOS/B7RP-1 inter-
action plays an essential role of CXCR5?TFHcells in the spleen
and LN, but the GC formation in LN is not always dependent on
CXCR5?TFHcells. In contrast, a substantial contribution of
OX40/OX40L interaction to the development of CXCR5?TFH
cells and GC B cells was observed only in LN of certain strains of
mice, depending on differential expression of OX40 on CXCR5?
*Department of Immunology and†Department of Respiratory Medicine, and‡Divi-
sion of Biomedical Imaging Research, Juntendo University School of Medicine, To-
kyo, Japan;§Department of Ophthalmology, Tokyo Medical University, Tokyo, Ja-
pan; and¶Central Pharmaceutical Research Institute, Japan Tobacco, Osaka, Japan
Received for publication June 21, 2004. Accepted for publication May 30, 2005.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1This work was supported by grants from the Ministry of Education, Culture, Sports,
Science and Technology, Japan.
2Address correspondence and reprint requests to Dr. Hisaya Akiba, Department of
Immunology, Juntendo University School of Medicine, 2-1-1 Hongo, Bunkyo-ku,
Tokyo 113-8421, Japan. E-mail address: firstname.lastname@example.org
3Abbreviations used in this paper: GC, germinal center; LN, lymph node; PNA,
peanut agglutinin; TFH, follicular B helper T cell; OX40L, OX40 ligand.
The Journal of Immunology
Copyright © 2005 by The American Association of Immunologists, Inc.0022-1767/05/$02.00
Materials and Methods
Female BALB/c, C57BL/6, C57BL/10, B10.D2, CBA/N, C3H/He, DBA/1,
A/J, and C.B-17/scid (SCID) mice were purchased from Charles River
Japan and Japan SLC. OX40L-deficient mice on C57BL/6 or BALB/c
background were obtained from Drs. N. Ishii and K. Sugamura (Tohoku
University School of Medicine, Sendai, Japan) (33, 34). CD40-deficient
mice on BALB/c background were gift from Dr. H. Kikutani (Osaka Uni-
versity, Osaka, Japan) (35). CD28-deficient mice on C57BL/6 background
were purchased from The Jackson Laboratory. ICOS-deficient mice on
C57BL/6 background have been described previously (36). These mice
were bred and maintained in the Oriental Yeast Company. All mice were
6–8 wk old at the start of experiments and kept under specific pathogen-
free conditions during the experiments.
Abs and reagents
Anti-mouse B7RP-1 (HK5.3) and anti-mouse OX40L (RM134L) mAbs
were generated in our laboratory as previously described (37, 38). Control
rat IgG was purchased from Sigma-Aldrich. FITC-conjugated anti-CD4
(RM4-5), biotin- or allophycocyanin-conjugated anti-CD45R/B220 (RA3-
6B2), and PE-conjugated anti-ICOS (15F9) mAbs were purchased from
eBioscience. Biotin-conjugated anti-OX40 (OX86) mAb, biotin- or PE-
conjugated anti-CXCR5 (2G8) mAb, rat IgG isotype controls, hamster IgG
control, and PE- or allophycocyanin-labeled streptavidin were purchased
from BD Pharmingen. Biotin- or FITC-conjugated peanut agglutinin
(PNA) was purchased from Vector Laboratories. SRBC were purchased
from Nippon Bio-supply Center.
Groups of five mice were i.p. immunized with 2 ? 108SRBC in 0.2 ml of
PBS to induce GC response in the spleen or immunized with 5 ? 107
SRBC/50 ?l in the footpads to induce GC response in the popliteal LN. In
some groups, mice were i.p. administrated with 300 ?g of anti-B7RP-1
mAb, anti-OX40L mAb, or control rat IgG at the time of immunization
(day 0) and on days 2 and 4. Spleen cells or popliteal LN cells were
collected at day 6 or 7, and the induction of PNA?B220?GC B cells and
CXCR5?TFHcells was analyzed by flow cytometry. To measure the sec-
ondary response, mice were i.p. injected with 2 ? 108SRBC/0.2 ml on day
30 after the first immunization. Some mice were treated with 300 ?g of
anti-B7RP-1 mAb, anti-OX40L mAb, or control rat IgG on days 30 and 32.
Three days after the second immunization, spleen cells were analyzed by
Flow cytometric analysis
Spleen or LN cells (0.5–1 ? 106) were first preincubated with unlabeled
anti-CD16/32 mAb (BD Pharmingen) to avoid nonspecific binding of
mAbs to Fc?R. The GC B cells were determined by staining with FITC-
labeled PNA and allophycocyanin-labeled anti-B220 mAb. The TFHcells
were determined by staining with FITC-labeled anti-CD4 and biotin-la-
beled anti-CXCR5 mAbs, followed by PE-labeled streptavidin. The ex-
pression of ICOS, OX40, and CXCR5 on CD4?T cells was determined by
staining with FITC-labeled anti-CD4 mAb and PE- or biotin-labeled mAbs
for respective molecules, followed by allophycocyanin-labeled streptavi-
din. After washing with PBS, the stained cells (live-gated on the basis of
forward and side scatter profiles and propidium iodide exclusion) were
analyzed on a FACSCalibur (BD Biosciences), and data were processed
using the CellQuest program (BD Biosciences).
The spleens from anti-B7RP-1- or rat IgG-treated BALB/c mice and ICOS-
deficient or wild-type C57BL/6 mice on day 7 after SRBC immunization
were embedded in Tissue-Tek OCT compound (Sakura Finetechnical), and
were frozen in liquid nitrogen bath. PNA staining was conducted as pre-
viously described (39). Briefly, 3-?m cryostat sections were air dried and
fixed with 8% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for 30
min at 4°C. Nonspecific binding sites were blocked by incubation for 30
min at room temperature in PBS containing 2% BSA fraction V (Sigma-
Aldrich). After removing the solution, the sections were incubated with 5
?g/ml biotin-conjugated PNA for 1 h at 37°C. After washing with PBS, the
slides were treated with 0.3% hydrogen peroxide in methanol for 20 min at
room temperature to block endogenous peroxidase activity. The sections
were incubated with avidin-biotin-peroxidase complex followed by further
amplification with boyinyl tyramide (Catalyzed Signal Amplification sys-
tem; DAKO) according to the manufacturer’s instructions. Subsequently,
the peroxidase color reaction was performed by a 10-min application of
freshly prepared 0.05% 3,3?-diaminobenzidine-0.01% hydrogen peroxide
solution (WAKO). The sections were counterstained with hematoxylin. For
immunohistochemistry, 3-?m cryostat sections were fixed with acetone for
10 min at 4°C. The sections were incubated with 2% BSA in PBS for 30
min at room temperature before incubation with Abs to reduce nonspecific
binding of Abs. After removing the solution, the sections were incubated
with 20 ?g/ml biotin-conjugated anti-CD45R/B220 mAb for 1 h at 37°C.
After washing with PBS, the sections were labeled with Alexa-Fluor 594-
conjugated streptavidin (Molecular Probes) at 1:100 for 1 h at 37°C, and
finally incubated with 20 ?g/ml FITC-conjugated anti-CD4 mAb for 1 h at
ELISA for SRBC-specific serum Abs
BALB/c mice were i.p. immunized with SRBC on days 0 and 15 and
administrated with 300 ?g of anti-B7RP-1 mAb, anti-OX40L mAb, or
control rat IgG on days ?1, 0, and 2. Serum anti-SRBC Abs were mea-
sured on day 22 by isotype-specific ELISA. Soluble SRBC Ags were pre-
pared as described (40) and protein concentration was determined using the
Bio-Rad Protein Assay reagent (Bio-Rad). Soluble SRBC Ags (5 ?g/ml in
carbonate buffer, pH 9.6) were coated onto 96-well Immulon 2HB plates
(Thermo Labsystems). After blocking with 1% BSA in PBS, SRBC-spe-
cific IgM and IgG isotypes were determined by incubating serially diluted
serum samples for 2 h at 37°C. After washing with 0.05% Tween 20 in
PBS, wells were incubated with biotin-conjugated isotype-specific mAbs,
including anti-mouse IgG1 (Serotec) or anti-mouse IgG2a, IgG2b, or IgG3
(BD Pharmingen), washed, and then developed with Vectastain ABC kit
(Vector Laboratories) and o-phenylendiamine (WAKO). After terminating
the reaction with 2N H2SO4, OD at 490/595 nm was measured on a mi-
croplate reader (Bio-Rad).
Cytokine production by CD4?T cell subsets
Spleen cells were collected on day 7 after immunization of BALB/c mice
with SRBC. ICOS?CXCR5?, ICOS?CXCR5?, and ICOS?CXCR5?
CD4?T cells were isolated by FACS sorting and 2 ? 105cells per well
were cultured in RPMI 1640 medium containing 10% FCS, 10 mM
HEPES, 2 mM L-glutamine, 0.1 mg/ml penicillin and streptomycin, and 50
?M 2-ME on 5 ?g/ml immobilized anti-CD3 mAb (2C11). To determine
the production of cytokines, cell-free supernatants were collected at 48 h
and assayed for IL-2, IL-4, IL-5, and IL-10 by ELISA using OptEIA kits
(BD Pharmingen) and IFN-? using Mouse IFN-? ELISA Ready-SET-Go!
kit (eBioscience) according to the manufacturer’s instructions.
Adoptive transfer experiments
CD62L?CD4?naive T cells were purified from the spleen of BALB/c
mice by passage through nylon wool columns (WAKO) and by using au-
toMACS columns with CD4?T cell isolation kit and anti-CD62L-coupled
microbeads (Miltenyi Biotec) according to the manufacturer’s instructions.
Small resting B cells were also purified from the spleen of BALB/c mice
as previously described (38). Briefly, spleen cells were treated with a mix-
ture of hybridoma supernatants (anti-Thy-1.2, anti-CD4, and anti-CD8) and
low-tox rabbit complement (Cedarlane Laboratories). After Percoll (Am-
ersham Biosciences) gradient centrifugation, small B cells were collected
from the 60/70% interface. The purified CD4?
CD4?CD62L?; 2 ? 106cells) with or without the purified B cells (?95%
B220?; 1 ? 106cells) were i.v. injected into SCID mice (day ?1). After
24 h, mice were i.p. immunized with 2 ? 108SRBC and then i.p. admin-
istrated with 300 ?g of anti-B7RP-1 mAb or control rat IgG on days 0, 2,
and 4. Seven days after the immunization, spleen cells were analyzed by
T cells (?95%
The results are expressed as the mean ? SD of five mice in each group.
Significant differences between two experimental groups were analyzed by
the unpaired Student’s t test. Values of p ? 0.01 were considered
Effect of anti-B7RP-1 and anti-OX40L mAbs on primary GC
formation and TFHcell development
Because a previous report demonstrated an impaired GC formation
in ICOS-deficient mice in response to immunization of SRBC (17),
we followed the same protocol and used SRBC as an Ag in this
study, which can induce robust polyclonal GC responses in an
2341The Journal of Immunology
25. Schaerli, P., K. Willimann, A. B. Lang, M. Lipp, P. Loetscher, and B. Moser.
2000. CXC chemokine receptor 5 expression defines follicular homing T cells
with B cell helper function. J. Exp. Med. 192: 1553–1562.
26. Kim, C. H., L. S. Rott, I. Clark-Lewis, D. J. Campbell, L. Wu, and E. C. Butcher.
2001. Subspecialization of CXCR5?T cells: B helper activity is focused in a
germinal center-localized subset of CXCR5?T cells. J. Exp. Med. 193:
27. Campbell, D. J., C. H. Kim, and E. C. Butcher. 2001. Separable effector T cell
populations specialized for B cell help or tissue inflammation. Nat. Immunol. 2:
28. Moser, B., P. Schaerli, and P. Loetscher. 2002. CXCR5?T cells: follicular hom-
ing takes center stage in T-helper-cell responses. Trends Immunol. 23: 250–254.
29. Croft, M. 2003. Co-stimulatory members of the TNFR family: keys to effective
T-cell immunity? Nat. Rev. Immunol. 3: 609–620.
30. Flynn, S., K. M. Toellner, C. Raykundalia, M. Goodall, and P. Lane. 1998. CD4
T cell cytokine differentiation: the B cell activation molecule, OX40 ligand, in-
structs CD4 T cells to express interleukin 4 and upregulates expression of the
chemokine receptor, Blr-1. J. Exp. Med. 188: 297–304.
31. Walker, L. S., A. Gulbranson-Judge, S. Flynn, T. Brocker, C. Raykundalia,
M. Goodall, R. Forster, M. Lipp, and P. Lane. 1999. Compromised OX40 func-
tion in CD28-deficient mice is linked with failure to develop CXC chemokine
receptor 5-positive CD4 cells and germinal centers. J. Exp. Med. 190:
32. Brocker, T., A. Gulbranson-Judge, S. Flynn, M. Riedinger, C. Raykundalia, and
P. Lane. 1999. CD4 T cell traffic control: in vivo evidence that ligation of OX40
on CD4 T cells by OX40-ligand expressed on dendritic cells leads to the accu-
mulation of CD4 T cells in B follicles. Eur. J. Immunol. 29: 1610–1616.
33. Murata, K., N. Ishii, H. Takano, S. Miura, L. C. Ndhlovu, M. Nose, T. Noda, and
K. Sugamura. 2000. Impairment of antigen-presenting cell function in mice lack-
ing expression of OX40 ligand. J. Exp. Med. 191: 365–374.
34. Hoshino, A., Y. Tanaka, H. Akiba, Y. Asakura, Y. Mita, T. Sakurai, A. Takaoka,
S. Nakaike, N. Ishii, K. Sugamura, et al. 2003. Critical role for OX40 ligand in
the development of pathogenic Th2 cells in a murine model of asthma. Eur.
J. Immunol. 33: 861–869.
35. Kumanogoh, A., X. Wang, I. Lee, C. Watanabe, M. Kamanaka, W. Shi,
K. Yoshida, T. Sato, S. Habu, M. Itoh, et al. 2001. Increased T cell autoreactivity
in the absence of CD40-CD40 ligand interactions: a role of CD40 in regulatory
T cell development. J. Immunol. 166: 353–360.
36. Iiyama, R., T. Kanai, K. Uraushihara, T. Totsuka, T. Nakamura, T. Miyata,
H. Yagita, A. Kushi, K. Suzuki, K. Tezuka, and M. Watanabe. 2003. The role of
inducible co-stimulator (ICOS)/B7-related protein-1 (B7RP-1) interaction in the
functional development of Peyer’s patches. Immunol. Lett. 88: 63–70.
37. Iwai, H., Y. Kozono, S. Hirose, H. Akiba, H. Yagita, K. Okumura, H. Kohsaka,
N. Miyasaka, and M. Azuma. 2002. Amelioration of collagen-induced arthritis by
blockade of inducible costimulator-B7 homologous protein costimulation. J. Im-
munol. 169: 4332–4339.
38. Akiba, H., H. Oshima, K. Takeda, M. Atsuta, H. Nakano, A. Nakajima,
C. Nohara, H. Yagita, and K. Okumura. 1999. CD28-independent costimulation
of T cells by OX40 ligand and CD70 on activated B cells. J. Immunol. 162:
39. Imai, Y., T. Fukui, A. Ikegaya, T. Ishikawa, Y. Ono, and K. Kurohane. 2002.
Lack of Shiga-like toxin binding sites in germinal centres of mouse lymphoid
tissues. Immunology 105: 509–514.
40. Kelly, B. S., J. G. Levy, and L. Sikora. 1979. The use of the enzyme-linked
immunosorbent assay (ELISA) for the detection and quantification of specific
antibody from cell cultures. Immunology 37: 45–52.
41. Shinall, S. M., M. Gonzalez-Fernandez, R. J. Noelle, and T. J. Waldschmidt.
2000. Identification of murine germinal center B cell subsets defined by the ex-
pression of surface isotypes and differentiation antigens. J. Immunol. 164:
42. Ferguson, S. E., S. Han, G. Kelsoe, and C. B. Thompson. 1996. CD28 is required
for germinal center formation. J. Immunol. 156: 4576–4581.
43. Kawabe, T., T. Naka, K. Yoshida, T. Tanaka, H. Fujiwara, S. Suematsu,
N. Yoshida, T. Kishimoto, and H. Kikutani. 1994. The immune responses in
CD40-deficient mice: impaired immunoglobulin class switching and germinal
center formation. Immunity 1: 167–178.
44. Schaerli, P., P. Loetscher, and B. Moser. 2001. Cutting edge: induction of fol-
licular homing precedes effector Th cell development. J. Immunol. 167:
45. Garside, P., E. Ingulli, R. R. Merica, J. G. Johnson, R. J. Noelle, and
M. K. Jenkins. 1998. Visualization of specific B and T lymphocyte interactions
in the lymph node. Science 281: 96–99.
46. Obermeier, F., H. Schwarz, N. Dunger, U. G. Strauch, N. Grunwald,
J. Scholmerich, and W. Falk. 2003. OX40/OX40L interaction induces the ex-
pression of CXCR5 and contributes to chronic colitis induced by dextran sulfate
sodium in mice. Eur. J. Immunol. 33: 3265–3274.
47. Ekkens, M. J., Z. Liu, Q. Liu, J. Whitmire, S. Xiao, A. Foster, J. Pesce,
J. VanNoy, A. H. Sharpe, J. F. Urban, and W. C. Gause. 2003. The role of OX40
ligand interactions in the development of the Th2 response to the gastrointestinal
nematode parasite Heligmosomoides polygyrus. J. Immunol. 170: 384–393.
2348ICOS REGULATES TFHCELL MAINTENANCE IN VIVO