sign in sign up
<<
Home , T helper cell

The Journal of

B7RP-1 Blockade Ameliorates through Regulation of Follicular Helper T Cells

Yi-Ling Hu,* Daniela P. Metz,* James Chung,† Gerald Siu,1 and Ming Zhang2*

Autoimmune are marked by the presence of class-switched, high-affinity with pathogenic potential. Co- stimulation plays an important role in the activation of T cells and the development of -dependent responses. ICOS

plays an indispensable role in the development of follicular helper T cells (TFH cells), which provide cognate help to germinal

center (GC) B cells. We show that the levels of TFH cells and GC B cells in two different models of autoimmunity, the New Zealand

Black/New Zealand White (NZB/NZW) F1 mouse model of systemic and the collagen-induced arthritis model of , are dependent on the maintenance of the ICOS/B7RP-1 pathway. Treatment with an anti-B7RP-1 Ab

ameliorates manifestations and leads to a decrease in TFH cells and GC B cells as well as an overall decrease in the frequency of ICOS؉ T cells. Coculture experiments of Ag-primed B cells with CXCR5؉ or CXCR5؊ T cells show that blocking B7RP-1 does not directly impact the production of IgG by B cells. These findings further support the role of ICOS in autoimmunity

and suggest that the expansion of the TFH cell pool is an important mechanism by which ICOS regulates Ab production. The Journal of Immunology, 2009, 182: 1421–1428.

ffective activation of T cells via their Ag requires ing anti-B7RP-1 reagents effectively inhibited the initiation and additional stimuli provided by cell-surface costimulatory maturation of an Ab-mediated Ag response (15). In addition, E molecules (1, 2). The prototypical costimulatory receptor ICOS/B7RP-1 blockade has been shown to be efficacious in CD28 is expressed constitutively on T cells, and its interaction mouse models of lupus nephritis (16), rheumatoid arthritis (17), with the ligands CD80 and CD86 was determined to be critical for and (18). the activation of naive T cells (1, 2). Subsequently, a second mol- The maturation of the Ag-specific B cell response occurs within ecule was identified, referred to as ICOS, which was closely re- specific structures within the secondary lymph organs. The forma- lated structurally to CD28 (3–6). Initial experiments indicated that tion of these structures, referred to as GCs, is the anatomic hall-

the interaction of ICOS with its ligand B7RP-1, like the CD28/ mark of TH cell activity during a T cell-dependent B cell immune CD80-CD86 interactions, is critical for a wide variety of T cell response (19). During the initiation of the immune response, den-

responses (1–9). In vitro experiments demonstrated that both dritic cells present Ag to TH cells, which in turn induce cognate B

by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. CD28 and ICOS effectively enhance all basic T cell responses, cells to proliferate. As the response matures, the proliferating B including proliferation, up-regulation of molecules mediating cell- cells form the within a B cell-rich region referred cell interaction, synthesis of , and B cell help for Ab to as the follicle, where ongoing and Ig

secretion (10, 11). Although CD28 and ICOS induce the secretion class switching occurs (19, 20). These B cells compete for TH of an overlapping spectrum of T cell cytokines, only CD28 elicits support from T cells adjacent to the follicle (21, 22); B cells ex- abundant IL-2 generation, whereas ICOS is particularly effective pressing immunoglobulins with mutations that improve the affinity in inducing IL-10 release, a implicated in the gener- for Ag either return to the GC for additional rounds of mutation or ation of B cell memory and plasma cells (11). B7RP-1 and migrate to the periphery, whereas B cells that obtain mutations that ICOS-deficient mice as well as ICOS null patients showed sim- render the Ig autoreactive are deleted (20). The T cells on the http://classic.jimmunol.org ilar of defective T cell-dependent Ab response, Ig margin of the follicle are specialized for B cell help and are re- 3 class switching and germinal center (GC) formation (12–14); ferred to as follicular Th (TFH) cells (20). These cells express high pharmacological blockade of the ICOS/B7RP-1 interaction us- constitutive levels of ICOS, further suggesting an important role for the ICOS/B7RP-1 pathway in the B cell mediated immune response (20–22). *Department of Inflammation and †Department of Medical Sciences, Amgen, Inc., Thousand Oaks, CA 91320 We speculated that the ICOS/B7RP-1 interaction might be play-

Downloaded from ing a critical role in the induction of disease by enhancing the Received for publication June 24, 2008. Accepted for publication November 23, 2008. development and/or function of autoimmune TFH cells, leading to The costs of publication of this article were defrayed in part by the payment of page the development of autoimmune GC B cells and the production of charges. This article must therefore be hereby marked advertisement in accordance autoimmune Abs. To address this issue, we studied the role of TFH with 18 U.S.C. Section 1734 solely to indicate this fact. and the ICOS/B7RP-1 pathway in murine models of B cell-medi- 1 Current address: Immunoinflammation Centre of Excellence for Drug Discovery, ated using pharmacological blockade. We GlaxoSmithKline, Gunnels Wood Road, Stevenage, Hertfordshire, SG1 2NY U.K. demonstrate that blockade of the ICOS/B7RP-1 pathway with a 2 Address correspondence and reprint requests to Dr. M. Zhang, Amgen Inc., M/S B29-2-C, Amgen Centre Drive, Thousand Oaks, CA 91320. E-mail address: mAb against B7RP-1 leads both to reduction of TFH and GC B cell [email protected] number as well as to amelioration of disease manifestations in 3 Abbreviations used in this paper: GC, germinal center; TFH cell, follicular ; murine models of rheumatoid arthritis and systemic lupus ery- CIA, collagen-induced arthritis; CII, collagen II; NZB/NZW, New Zealand Black/New thematosus (SLE). Although purified T cells are capable of in- Zealand White; PNA, peanut agglutinin; SLE, systemic lupus erythematosus. FH ducing Ig production from GC B cells in in vitro culture, ICOS/ Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 B7RP-1 blockade does not inhibit this process. These data indicate

www.jimmunol.org 1422 B7RP-1 REGULATES TFH CELLS IN AUTOIMMUNITY

that the ICOS/B7RP-1 pathway is critical for B cell-mediated sciences unless otherwise indicated. Flow cytometry was performed with a pathogenic autoimmune processes by affecting development but FACSCalibur (BD Biosciences), and the data were analyzed using the FlowJo program (Tree Star). not the function of TFH cells, and that the blockade of this pathway may therefore be beneficial in the prevention and treatment of ELISAs these diseases. Anti-CII Ab ELISA. Serum samples were collected on day 14 and 34 from CII-immunized DBA/1J mice and anti-CII Ab level was measured by Materials and Methods ELISA. In brief, plates coated with bovine CII (Chondrex) were incubated Mice for2hatroom temperature with serial dilutions of serum samples obtained from the test mice. Bound CII-specific Ig was detected with HRP-conju- Specific -free 8-wk-old male DBA/1J mice and female New Zea- gated goat anti-mouse IgG1, IgG2a, and IgG2b Abs (Southern Biotech-

land Black/New Zealand White (NZB/NZW) F1 mice were purchased from nology Associates). The substrate reaction was performed with Sureblue The Jackson Laboratories. Specific pathogen-free 6- to 8-wk-old female TMB microwell peroxidase substrate (Kirkegaard & Perry Laboratories) BALB/c mice were purchased from Charles River Laboratories. Mice were and the OD was read using Spectrum Max (Molecular Devices). A standard housed under pathogen-free conditions at the Laboratory Animal Research serum composed of a mixture of sera from arthritic mice was added to each Facility at Amgen. Animal experiments were conducted as approved by plate in serial dilutions and a standard curve was constructed. Institutional Animal Care and Use Committee of Amgen. Anti-dsDNA Ab ELISA. Serum anti-dsDNA IgG level in NZB/NZW F1 mice was measured by ELISA using a mouse anti-dsDNA ELISA kit (Al- reagents and the induction and assessment of pha Diagnostic) following the manufacturer’s instructions. collagen-induced arthritis Anti-SRBC Ab ELISA. In brief, plates coated with 10 ␮g/ml soluble SRBC The generation of the chimeric mouse anti-mouse B7RP-1 mAb clone Ag were incubated for2hatroom temperature with diluted cell culture 1B7-V2 as well as the rat anti-mouse B7RP-1 mAb 18A5 was described supernatant. Bound SRBC-specific Ig was detected with HRP-conjugated previously. Mouse IgG1 control was supplied by Amgen. Col- goat anti-mice IgG and IgM Abs (Southern Biotechnology Associates). lagen-induced arthritis (CIA) was induced as described earlier (24). In The substrate reaction was performed with Sureblue TMB microwell per- brief, male DBA/1J mice (7- to 10-wk-old) were injected intradermally at oxidase substrate (Kirkegaard & Perry Laboratories) and the OD was read the base of the tail with 200 ␮g of bovine collagen II (CII) (Chondrex) using Spectrum Max (Molecular Devices). A standard serum composed of emulsified in CFA (Difco Laboratories). Twenty-one days after primary a mixture of sera from SRBC immunized mice was added to each plate in immunization, the mice were boosted in the same way. The day of the serial dilutions and a standard curve was constructed. boost was designated as day 0. Mice were randomly divided into five groups (15 mice in each group) and treated with intraperitoneal injections T cell proliferation and cytokine production by CII rechallenge of 1B7-V2 at doses of 15, 3, and 0.6 mg/kg, IgG1 isotype control at a in vitro dose of 15 mg/kg, or PBS once per week after primary immunization and Draining lymph nodes were removed 7 days after CII boost. Single-cell every other day after the boost until day 15. Mice were examined daily for suspensions were prepared in RPMI 1640 supplemented with 10% FBS, 2 the onset of arthritis symptoms. The swelling of all four paws was graded Ϫ mM glutamine, 1 mM sodium pyruvate, 5 ϫ 10 5 M 2-ME, and antibi- from 0 to 3 as follows: grade 0, no swelling; grade 1, swelling of finger otics. The cells (5 ϫ 105/well) were cultured in 96-well flat- joints or focal redness; grade 2, mind swelling of wrist or ankle joints; bottom plates (BD Biosciences) in the presence or absence of 50 ␮g/ml grade 3, severe swelling of entire paw or deformity or ankylosis. Each paw bovine CII for 5 days. For the last 16 h, 1 ␮Ci of [3H]thymidine (MP was graded and the four scores were totaled so that the maximal score per Biomedicals) was added to each well. The cells were then harvested and 3H mouse was 12. Incidence was expressed as the percentage of mice that had incorporation measured using a 1450 Microbeta liquid scintillation and a score of at least 1 in the total number of mice examined. luminescence counter (PerkinElmer). Supernatants from similar cultures

by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. ␥ Induction and assessment of disease symptoms in the NZB/NZW were collected after 96 h for the assessment of IFN- production by MSD (Meso Scale Discovery) according to the protocol recommended by the F1 model of SLE manufacturer.

Two-month-old female NZB/NZW F1 mice were randomly divided into five groups (15 mice in each group), treated i.p. either with doses of cocultures 1B7-V2 at 8, 1.5, and 0.3 mg/kg, mouse IgG1 isotype control at 8 mg/kg, cells from SRBC-immunized BALB/c mice were sorted for B cells, or PBS control weekly for 23 wk. Peripheral blood was collected at 4, 5, CXCR5ϩCD4ϩ T cells or CXCR5ϪCD4ϩ T cells by FACSAria (BD Bio- 6, and 7 mo of age for measuring serum anti-dsDNA Ab; urine was ex- sciences). B cells (2 ϫ 105/well) were cultured alone or with CXCR5ϩ pressed from the bladder and tested for proteinuria colorometrically using CD4ϩ or CXCR5ϪCD4ϩ T cells (2 ϫ 104/well) with 10 ␮g/ml SRBC Albustix (Bayer) every 2 wk starting at 5 mo of age. The incidence of soluble membrane Ag in 96-well U-bottom microplate dishes (Becton proteinuria was expressed as the percentage of mice where urine http://classic.jimmunol.org Dikinson) in RPMI 1640 supplemented with 10% FBS, 2 mM glutamine, was at least 300 mg/dl in two consecutive measurements. Survival was 1 mM sodium pyruvate, 5 ϫ 10Ϫ5 M 2-ME, and antibiotics. 1B7-V2, calculated at 10 mo of age. mouse CTLA4Ig (R&D Systems), anti-mouse CD40L Ab (R&D Systems) and isotype-matched control was added (10 ␮g/ml) into the culture to block SRBC immunization model ICOS/B7RP-1, B7/CD28, or CD40/CD40L engagements. The SRBC Ag- BALB/c mice (6- to 8-wk old) were immunized with a single i.p. injection containing culture medium was replaced by fresh medium (containing 10 ␮ of 2 ϫ 108 SRBC (Lampire Biological Laboratories) in 0.2 ml of PBS. The g/ml of above Abs or CTLA4Ig) after 2 days of culture, and the super- immunized mice were randomly divided into three groups (4 mice in each natants were collected at day 12 to measure SRBC-specific Ig by ELISA. Downloaded from group) and treated with 15 mg/kg 1B7-V2, 15 mg/kg murine IgG1 isotype control, or PBS control on day 0, day 2, and day 5 by i.p. injection. On day Statistical analysis 7, mice were sacrificed and the spleen cells collected to measure TFH cell The results are expressed as the mean Ϯ SD of 4 mice in each group in and GC B cell induction by flow cytometry. In some experiments, spleen Figs. 1, 2, 3, and 6. Significant difference between experimental groups was cells from SRBC-immunized mice were collected to sort B cells, ϩ ϩ Ϫ ϩ analyzed by unpaired Student’s t test using Prism software. In Fig. 5C, CXCR5 CD4 cells and CXCR5 CD4 cells. survival was estimated using a Kaplan-Meier curve. Flow cytometric analysis Results and draining lymph nodes were collected and the cells harvested Blockade of ICOS-B7RP-1 inhibits the development of TFH and by grinding the tissues through a cell strainer. The cells were preincubated GC B cells in murine models of autoimmune disease with unlabeled anti-CD16/32 to block nonspecific binding of the staining ␥ mAbs to Fc R. The GC B cells were identified by staining with FITC- Due to the critical role of TFH cells in the induction and maturation labeled peanut agglutinin (PNA) (Vector Laboratories) and PE-labeled anti- of the Ab response and the dependence of many autoimmune dis- B220. The TFH cells were identified by staining with FITC-labeled anti- CD4 and PE-labeled anti-CXCR5. The ICOS and CD69 expression on T eases on the production of autoantibodies (20), we hypothesized cells were determined by staining FITC-labeled anti-CD3 and PE-labeled that disease induction is the result of the generation of TFH cells, anti-ICOS or anti-CD69. All staining Abs were purchased from BD Bio- which in turn induce the production of autoreactive GC B cells and The Journal of Immunology 1423 by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd.

FIGURE 2. Effect of anti-B7RP-1 on inhibition GC B cells in mice

autoimmune disease models. CIA mice (A) and NZB/NZW F1 mice (B) were treated as described. Germinal center B cells (PNAϩB220ϩ cells) were determined by staining draining lymph nodes in the CIA mice or FIGURE 1. Effect of anti-B7RP-1 on inhibition follicular B helper T http://classic.jimmunol.org splenocytes from NZB/NZW F1 mice with PE-labeled anti-B220 and cell (TFH) induction in mice autoimmune disease models. The induction of FITC-labeled PNA. The dot plot represents one of four mice in each group ϩ ϩ TFH cells (CXCR5 CD4 T cells) in mice autoimmune disease models (upper panel), the data are also expressed as the mean percentages of was determined by staining with FITC-labeled anti-CD4 and PE-labeled PNAϩ cells within B220 cells from four mice in each group (lower panel). anti-CXCR5. The dot plot represents one of four mice in each group (upper panel). The data are also expressed as the mean percentages of CD4ϩ ϩ CXCR5 T cells from four mice in each group (lower panel). A,TFH cells B7RP-1 blockade was confirmed by B7RP-1 receptor occupancy

Downloaded from induction in the draining lymph nodes of CIA mice at day 7 with 15 mg/kg assay in draining lymph nodes at day 7 as described previously. 1B7-V2 or isotype control treatment. B,TFH cells induction in the spleen Cells from draining lymph nodes were collected on day 7 after CII of NZB/NZW F mice at the age of 8 mo with 8 mg/kg 1B7-V2 or isotype 1 boost and analyzed for T and GC B cell formation using flow control treatment. FH cytometry. As shown Figs. 1A and 2A, isotype-control-treated Ϯ mice generated significant amounts of both TFH (7.6 0.52% of the total population) and GC B (8.52 Ϯ 1.8%) cells in response to the production of pathogenic autoantibodies. Furthermore, we hy- collagen administration. In contrast, the 1B7-V2-treated mice (15 Ϯ pothesize that the blockade of ICOS/B7RP-1 would inhibit this mg/kg) developed significantly less TFH (2.25 0.51%) and GC process and would thus be efficacious in treating these diseases by B (3.8 Ϯ 0.17%) cells (Fig. 2A). Similar experiments were con-

inhibiting the development of these cell types. To address this, we ducted in the NZB/NZW F1 model of SLE (Figs. 1B and 2B). For used the anti-mouse B7RP-1 mAb 1B7-V2 in the CIA model and these studies, NZB mice were crossed to NZW mice and the re-

the NZB/NZW F1 lupus model. For the CIA studies, DBA/1J mice sulting F1 offspring allowed to mature; once-weekly treatment with were immunized with type II collagen to induce arthritis and dif- 1B7-V2, PBS, or control mouse IgG1 was initiated at two months

ferent doses of 1B7-V2, PBS or control mouse IgG1 were admin- of age, and generation of TFH and GC B cells was assessed at 8 mo. Ϯ istrated from primary immunization until 15 days after Ag boost. 1B7-V2-treated mice (8 mg/kg) have significantly less TFH (3.9 1424 B7RP-1 REGULATES TFH CELLS IN AUTOIMMUNITY

FIGURE 3. Anti-B7RP-1 treatment inhibits T cell activation in vivo in autoimmune disease models. A, 1B7-V2 inhibits CII-specific T cell proliferation (upper panel) and IFN-␥ production (lower panel) in CIA mice. CIA mice were immunized and treated as described in Fig. 1A. Draining lymph node cells from 1B7-V2 or isotype control treated mice were collected at day 7, cultured in the absence (black bar) or presence (white bar) of 50 ␮g/ml of CII. Cell proliferation was measured by [3H]thymidine incorporation at day 5 and IFN-␥ production was measured by MSD at day 4. B, 1B7-V2 treatment leads to

decreases in ICOS and CD69 expression in NZB/NZW F1 lupus mice. NZB/NZW F1 mice were treated as described in Fig. 1B. Spleen cells were stained by FITC-labeled anti-CD3 and PE-labeled anti-ICOS or anti-CD69. The dot plot represents one of four mice in each group (upper panel), the data are also expressed as the mean percentages of ICOS cells or CD69ϩ cells within CD3ϩ T cells from four mice in each group (lower panel).

Ϯ Ϯ 0.12% vs 13.27 2.9%) (Fig. 1B) and GC B (1.7 0.35% vs trol treated mice. In addition, NZB/NZW F1 mice treated with 6.27 Ϯ 0.51%) (Fig. 2B) cells than isotype-matched-control- 1B7-V2 also show statistically significant decreases in ICOSϩ and treated mice, consistent with the data presented above. Taken to- CD69ϩ T cells in spleen after 1B7-V2 treatment (Fig. 3B). These by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. gether, these data suggest that TFH and GC B cells are dependent data show the extent to which the T cell responses in the 1B7-V2- on the ICOS/B7RP-1 pathway in autoimmune disease models. treated mice are blunted.

ICOS/B7RP-1 blockade inhibits T cell activation ICOS/B7RP-1 blockade inhibits anti-CII Ab production and To assess whether ICOS/B7RP-1 pathway blockade is inhibiting T disease symptoms in CIA cell activation, draining lymph node cells were harvested from To assess the production of anti-collagen Abs in the CIA model, mice from the CIA experiment treated with either 1B7-V2 (15 serum levels of anti-CII IgG1, IgG2A, and IgG2B were measured mg/kg) or isotype-matched control and rechallanged in vitro with by ELISA 14 and 34 days after collagen boost, respectively. As soluble CII Ag, and activation status assessed. As shown in Fig. can be seen in Fig. 4A, high doses (15 mg/kg) of 1B7-V2 dramat- http://classic.jimmunol.org 3A, lymph node cells from 1B7-V2-treated mice showed signifi- ically suppressed serum levels of anti-CII IgG1, IgG2A, and cantly reduced responses as determined by both cell proliferation IgG2b, whereas intermediate doses suppressed IgG1 and IgG2B and IFN-␥ production compared with those from the isotype-con- but not IgG2A production. Blockade of the ICOS/B7RP-1 pathway

FIGURE 4. Anti-B7RP-1 treatment Downloaded from decreases serum anti-CII IgG titers and reduces the development of arthritis in CIA mice. Fig. 4A. Serum anti-CII IgG1, IgG2a, and IgG2b titers were measured by ELISA at day 14 (white bar) and day 34 (black bar). Data are shown as the mean Ϯ SEM of 15 mice in each group. B, Anti-B7RP-1 treatment reduces the development of CIA in both disease in- cidence and clinical score (mean Ϯ SEM). CIA was induced as described in Material and Methods. The data were from 15 mice in each group. The PBS-treated group shows similar pat- tern with isotype control treated group (data not show). The Journal of Immunology 1425

leads to suppression of clinical symptoms in this model as can be seen in Fig. 4B. Of the mice treated with the isotype control, 70% developed arthritis after 40 days with a mean arthritis score per paw of 3.5; in contrast, a dose-dependent decrease in both fre- quency and arthritis score was observed in the 1B7-V2-treated mice, with the high-dose group resulting in only 5% incidence at 40 days with a mean arthritis score of 0.25.

ICOS/B7RP-1 blockade inhibits development of signs and symptoms in a murine model of SLE

To test whether decreases in TFH and GC B cells resulting from blockade of B7RP-1 in the SLE model would also correlate with clinical efficacy, we followed the disease status in the NZB/NZW

F1 mice treated with 1B7-V2 or isotype-matched control. As can be seen in Fig. 5, 1B7-V2-treated mice had statistically significant de- creases in disease as assessed by anti-dsDNA Abs, proteinuria, or survival as compared with the isotype-matched dosed control mice.

Acute blockade of the ICOS/B7RP-1 pathway does not inhibit

TFH help of B cells FIGURE 5. Anti-B7RP-1 treatment decreases serum anti-dsDNA ti- The above data indicate that the blockade of the ICOS/B7RP-1 ters and reduces the development of lupus nephritis in NZB/NZW F1 pathway leads to the reduction of TFH and GC B cells in two mice. NZB/NZW F mice were divided into two groups (n ϭ 15), i.p. 1 different models of autoimmune disease. To address whether the injected with 8 mg/kg 1B7-V2 or mouse IgG1 isotype control once per control of production by T cells is at the level of week from age 2 mo for 23 wk. Serum anti-dsDNA titer was measured FH by ELISA at age of 7 mo (the data from age of 5 and 6 mo showed development or B cell helper function, we conducted experiments exactly similar pattern with 7 mo, data not shown). The development of in mice immunized with SRBC. This system provides a robust proteinuria was measured every 2 wk from age of 5 mo, and the data immune response and therefore, unlike standard Ag-challenge here are the result of age 8.5 mo. The survival curve was generated by models as well as the disease models above, generates sufficient a standard Kaplan-Meier curve. 1B7-V2-treated mice showed signifi- numbers of Ag-specific for use in in vitro studies. cant decreases in anti-dsDNA Ab (p Ͻ 0.01) and proteinuria and a Mice were immunized with SRBC on day 0 and treated with either significant enhancement in survival (p Ͻ 0.01). 1B7-V2 or isotype control on days 0, 2, and 5. Spleen cells from SRBC immunized mice were harvested on day 8 and analyzed for

TFH and GC B cells using flow cytometry as described above for by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. http://classic.jimmunol.org Downloaded from

FIGURE 6. Effect of anti-B7RP-1 on TFH cell and GC B cell induction in mice SRBC immunized model. BALB/c mice were immunized by i.p. injection

of SRBC at day 0. 1B7-V2, isotype control or PBS was administrated by i.p. injection at day 0, 2, and 5. Spleen cells were collected on day 8. The TFH cell (A) and GC B cell (B) induction were determined as described as above. The total spleen cell count (C), total spleen B cell count (D) and total spleen T cell count (E) were also calculated. 1426 B7RP-1 REGULATES TFH CELLS IN AUTOIMMUNITY

Table I. Enhanced ICOS expression parallels the increased TFH cells in lupus micea,b

ϩ ϩ ϩ ICOS T Cell CD4 CXCR5 TFH Cells Young mice 11 Ϯ 0.9% 4.53 Ϯ 1.07% Lupus mice 43 Ϯ 6.7% 9.83 Ϯ 1.08%

a Young NZB/W F1 mice (3-mo-old) or old proteinuria positive NZB/W F1 mice (7.5-mo-old) were sacrificed and the spleen cells were collected. The ICOS express- ing T cells were stained by FITC-labeled anti-ICOS and PE-labeled anti-CD3. The TFH cells were stained with PE-labeled anti-CXCR5 and FITC-labeled anti-CD4. The data are expressed as the percentage of total T cells (ICOS) or total CD4 T cells ϩ (CXCR5 TFH cells). b Four mice in each group.

IgM but not IgG production, consistent with the hypothesis that al-

though the non-TFH T cells are capable of helping B cells proliferate and secrete Ag-specific IgM, they are not capable of supporting ma-

ture, switched B cells. Interestingly, incubation of B cells and TFH from SRBC-challenged mice leads to high levels of both anti-SRBC

IgM and IgG, indicating that TFH cells are indeed capable of support- ing the activation of both immature and mature B cells. To determine whether the blockade of the ICOS/B7RP-1

pathway inhibits the ability of the TFH cell to help mature B cell function, the cell culture experiments were conducted with and without 1B7-V2 blockade. Somewhat surprisingly, Ab blockade of the ICOS/B7RP-1 pathway did not appreciably affect the secretion ϩ ϩ of either IgM or IgG in cultures containing B cells with either FIGURE 7. B7RP-1 blockade cannot inhibit CXCR5 CD4 TFH cells supporting B cells Ig production and class switching in vitro. BALB/c mice normal T cells or TFH cells, suggesting that the B helper function were immunized with SRBC. The spleens were collected at day 8 and of the TFH cell does not require ICOS expression. Blocking the sorted for B cells, CXCR5ϩCD4ϩ cells and CXCR5ϪCD4ϩ cells by CD40:CD40L costimulatory pathway with anti-CD40L also did FACS. B cell alone (2 ϫ 105 B cell per well), B cell (2 ϫ 105 per well) plus not lead to a decrease in the secretion of SRBC specific Abs. Ad- ϩ ϩ Ϫ ϩ CXCR5 CD4 cells (2 ϫ 104 per well) or B cell plus CXCR5 CD4 dition of CTLA4-Ig, on the other hand, dramatically inhibited the cells (2 ϫ 104 per well) were cultured without any Ab treatment, treated production of anti-SRBC IgG (and IgM) Abs. While ICOS does with 10 ␮g/ml of 1B7-V2, CTLA4Ig, anti-CD40L or isotype control. All by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. not exert a significant influence on the production of Abs, its im- ␮ the cultures were stimulated with 10 g/ml SRBC soluble membrane Ag. pact on the T cells appears likely through inhibition of their After 12 days of culture, the anti-SRBC specific IgG and IgM in the culture FH development. The proportion of ICOS expressing T cells and T supernatant were determined by ELISA. A representative example out of FH three independent experiments is shown. cells were assessed at 3 mo and at 7.5 mo in the NZB/NZW F1 mice (Table I). The increased age and disease severity of the mice correspond with an increase in both the proportion of ICOS ex-

Figs. 1 and 2. Similar to previously published reports (15), com- pressing T cells and the level of TFH cells. plete blockade of the ICOS/B7RP-1 pathway in mice leads to a Ϯ decrease in the percentage of TFH cells (3.5 0.89%) compared Discussion http://classic.jimmunol.org with isotype control treated mice (12.7 Ϯ 1.5%) or PBS control Aberrant regulation of the adaptive immune response results in the treated mice (10.2 Ϯ 0.96%) (Fig. 6A). Similarly, blockade of this production of autoantibodies that are the hallmark of both organ- pathway in mice also led to a substantial decrease in GC B cells specific and systemic autoimmune diseases such as SLE and rheu- (7.44 ϩ 0.15%) compared with the isotype-control (18.16 Ϯ 3.0%) matoid arthritis (25, 26). Costimulation plays an important role in or PBS-treated mice (14.97 Ϯ 1.82%) (Fig. 6B). The proportional the Ag-driven immune response that lead to potentially pathogenic decrease in the absolute number of T cells (Fig. 6E) is less pro- isotype-switched, high affinity Abs (1, 11). ICOS: B7RP-1 has ϩ Downloaded from nounced than the decrease seen in the percentage of CD4 T cells been shown to be critical for T cell effector function and T cell- that are CXCR5ϩ thus showing a disproportionately greater effect dependent B cell responses (1–8). Disruption of the pathway as null of B7RP-1 blockade on the TFH population. The relative decrease observed in knockout mice and in ICOS patients demonstrate in the percentage of GC B cells (Fig. 6B), total splenocyte number the critical role that ICOS plays in GC formation and the differ- (Fig. 6C) and total splenic B cell number (Fig. 6D) are comparable. entiation of naive B cells into class switched memory B cells and Consistent with our data above, these data indicate that blockade of IgG secreting cells (7, 8, 10–14).

the ICOS/B7RP-1 pathway inhibits development of TFH and GC B Growing evidence in animal models and in human diseases has cells, even with an extremely potent Ag. To determine the effects implicated ICOS/B7RP-1 as an important costimulatory pathway of blockade of the ICOS/B7 pathway on direct T-B interaction, in the pathogenesis of autoimmunity (25, 27–31). High levels of

TFH and whole B cells were purified from mice previously immu- both ICOS expression (25, 30, 31) and ICOS/B7RP-1-mediated nized with SRBC, incubated in vitro, and assessed for anti-SRBC down-regulation of B7RP-1 expression (25) have been reported on Ab production (Fig. 7). Incubation of B cells from the SRBC- lymphocytes purified from SLE patients. In addition, there have challenged mice alone does not lead to significant production of been reports that peripheral T cells from SLE patients are more

either anti-SRBC IgM or IgG. Incubation of B cells and non-TFH sensitive to stimulation by a combination of anti-CD3 and anti- T cells from SRBC-challenged mice leads to significant anti-SRBC ICOS and are capable of enhancing autoantibody production in The Journal of Immunology 1427

vitro by autologous B cells (25). ICOS has been shown to be in- Acknowledgments creased in RA patients as well (31). Disruption of the ICOS: We thank Hailing Hsu and Jo Viney for many helpful comments and B7RP-1 pathway has been shown to ameliorate disease in numer- discussions.

ous animals models of autoimmunity including the NZB/NZW F1 (17), CIA (17), and OVA-induced lung model Disclosures (23). Although these data attest to the importance of the ICOS/ The authors have no financial conflict of interest. B7RP-1 pathway in autoimmunity, the precise mechanisms and specifically the role of various ICOS expressing T cell subsets References remains unclear. 1. Greenwald, R. J., G. J. Freeman, and A. H. Sharpe. 2005. B7 family revisited. The mechanism of ICOS-mediated B cell help function of T Annu. Rev. Immunol. 23: 515–548. 2. Sharpe, A. H., and G. J. Freeman. 2002. The B7-CD28 superfamily. Nat. Rev. cells is through both cytokines and cell-cell interactions. Initially Immunol. 2: 116–126. felt to drive the TH2 response, it is now clear that ICOS is impor- 3. Yoshinaga, S. K., J. S. Whoriskey, S. D. Klare, U. Sarmiento, J. Guo, T. Horan, G. Shih, M. Zhang, M. A. Coccia, T. Kohno, et al. 1999. T cell costimulation tant for the production of a range of cytokines including those that through B7RP-1 and ICOS. 402: 827–832. lead to B cell activation and differentiation including IL-4, IL-10, 4. Hutloff, A., A. M. Dittrich, K. C. Beier, B. Eijaschewitsch, R. Kraft, and IL-21 (15, 34). ICOS signaling can also induce CD40L ex- I. Anagnostopoulos, and R. A. Kroczek. 1999. ICOS is an inducible T-cell co- stimulator structurally and functionally related to CD20. Nature 397: 263–266. pression on T cells (4, 12). Additionally, administration of acti- 5. Swallow, M. W., J. J. Wallin, and W. C. Sha. 1999. B7h, a novel costimulatory vating anti-CD40 reagents was found to enhance class switching in homolog of B7.1 and B7.2, is induced by TNF␣. 11: 423–432. ICOS null mice, suggesting a direct link between the CD40 and 6. Coyle, A. J., S. Lehar, C. Lioyd, J. Tian, T. Delaney, S. Manning, T. Nguyen, T. Burwell, H. Schneider, J. A. Gonzalo, et al. 2000. The CD28-related molecules ICOS pathways (12). Work by Sha and colleagues indicated that ICOS is required for effective T cell-dependent immune response. Immunity 13: CD40 signaling rescues activation-dependent down-regulation of 95–105. 7. Dong, C., A. E. Juedes, U. A. Temann, S. Shresta, J. P. Allison, N. H. Ruddle, B7RP-1 expression (34); in addition, these data implied a link and R. A. Flavell. 2001. ICOS co-stimulatory receptor is essential for T-cell between these two pathways which results in a feedback loop that activation and function. Nature 409: 97–101. serves to amplify both T and B cell interaction. A subset of T 8. Nurieva, R. I., X. M. Mai, K. Forbush, M. J. Bevan, and C. Dong. 2003. B7h is FH required for T cell activation, differentiation, and effector function. Proc. Natl. cells transiently expresses high levels of CD40L, supporting this Acad. Sci. USA 100: 14163–14168. hypothesis (20). 9. Burmeister, Y., T. Lischke, A. C. Dahler, H. W. Mages, K. P. Lam, A. J. Coyle, R. A. Kroczek, and A. Hutloff. 2008. ICOS control the pool size of effector- The coculture experiments with Ag-primed B cells and TFH memory and regulatory T cells. J. Immunol. 179: 774–782. cells show that the generation of the TFH cell pool, rather than the 10. Kroczek, R. A., H. W. Mages, and A. Hutloff. 2004. Emerging paradigms of direct activation of B cells by this T cell subset are the primary T-cell co-stimulation. Curr. Opin. Immunol. 16: 321–327. 11. Van Berkel, M. E. A. T., and M. A. Oosterwegel. 2006. CD28 and ICOS: similar mechanism regulating IgG Ab production. Although it has been or separate costimulators of T cells? Immunol. Lett. 105: 115–122. previously demonstrated that ICOS stimulation enhanced T cell 12. McAddam, A. J., R. J. Greeenwald, M. A. Levin, T. Chernova, N. Malenkovich, V. Ling, G. J. Freeman, and A. H. Sharpe. 2001. ICOS is critical for CD40- proliferation and cytokine secretion (3, 5, 6), it was unclear if, once mediated antibody class switching. Nature 409: 102–105. activated, constant ICOS stimulation was necessary to maintain the 13. Mak, T. W., A. Shahinian, S. K. Yoshinaga, A. Wakeham, L. M. Boucher, activation state – specifically in the manner to allow B cell help. M. Pintilie, G. Duncan, B. U. Gajewska, M. Gronski, U. Eriksson, et al. 2003. Costimulation through the inducible costimulator ligand is essential for both T The data presented above demonstrate that although ICOS/B7RP-1

by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. helper and B cell function in T cell-dependent B cell response. Nat. Immunol.4: blockade leads to complete inhibition of the secondary immune 765–772. response, it does not affect the ability of the T cell to induce 14. Grimbacher, B., A. Hutloff, M. Schlesier, E. Glocker, K. Warnatz, R. Drager, FH H. Eibel, B. Fischer, A. A. Schaffer, H. W. Mages, et al. 2003. Homozygous loss Ag-specific IgG secretion in in vitro culture. Interestingly, block- of ICOS is associated with adult-onset common variable immunodeficiency. Nat. ing CD40L failed to inhibit the T cell induced IgG production Immunol. 4: 261–268. FH 15. Akiba, H., K. Takeda, Y. Kojima, Y. Usui, N. Harada, T. Yamazaki, J. Ma, whereas blocking CD28 did. Thus, the ICOS/B7RP-1 pathway ap- K. Tezuka, H. Yagita, and K. Okumura. 2005. The role of ICOS in the CXCR5ϩ pears to be required primarily for the development of the TFH cell follicular B helper T cell maintenance in vivo. J. Immunol. 175: 2340–2348. and that, once generated, the T cell does not require ICOS sig- 16. Iwai, H., M. Abe, S. Hirose, F. Tsushima, K. Tezuka, H. Akiba, H. Yagita, FH K. Okumura, H. Kohsaka, N. Miyasaka, and M. Azuma. 2003. Involvement of naling to maintain its ability to activate and enhance B cell func- inducible costimulator-B7 homologous protein costimulatory pathway in murine lupus nephritis. J. Immunol. 171: 2848–2854. http://classic.jimmunol.org tion. The mechanisms underlying a productive TFH-B cell inter- 17. Iwai, H., Y. Kozono, S. Hirose, H. Akiba, H. Yagita, K. Okumura, H. Kohsaka, action require better understanding. N. Miyasaka, and M. Azuma. 2002. Amelioration of collagen-induced arthritis by Follicular Th cells express high levels of ICOS and are known blockade of inducible costimulator-B7 homologous protein costimulation. J. Im- to be important in the generation of a mature B cell response (34, munol. 169: 4332–4339. 18. Scott, B. G., H. Yang, E. Tuzun, C. Dong, R. A. Flavell, and P. Christadoss. 2004. 35), but their role in autoimmune disease pathogenesis has not ICOS is essential for the development of experimental autoimmune myasthenia gravis. J. Neuroimmunol. 153: 16–25. been well characterized. Our data suggest that the TFH cell in fact 19. Klein, U., and R. Dalla-Favera. 2008. Germinal center: role in B-cell physiology Downloaded from plays a critical role in the generation of autoantibodies as well as and malignancy. Nat. Rev. Immmunol. 8: 853–863. disease progression in two different autoimmune disease models. 20. Vinuesa, G. C., S. G. Tangye, B. Moser, and C. R. Mackay. 2005. Follicular B We show that of the levels of T cells and GC B cells in models helper T cells in antibody responses and autoimmunity. Nat. Rev. Immunol. 5: FH 853–863. of autoimmune disease are dependent on the ICOS/B7RP-1 path- 21. Breitfeld, D., L. Ohl, E. Kremmer, J. Ellwart, F. Sallusto, M. Lipp, and R. Forster. way and that B7RP-1 blockade effectively leads to a decrease in 2000. Follicular B helper T cells express CXC chemokine receptor 5, localize to both cell types with associated decreases in autoantibodies and B cell follicles, and support immunoglobulin production. J. Exp. Med. 192: 1545–1551. other disease manifestations. These data confirm the critical role of 22. Schaerli, P., K. Willimann, A. B. Lang, M. Lipp, P. Loetscher, and B. Moser. ICOS in the development of T cells and support blockade of the 2000. CXC chemokine receptor 5 expression defines follicular homing T cells FH with B cell helper function. J. Exp. Med. 4: 1553–1562. pathway an attractive strategy to treat human autoimmune dis- 23. Terato, K., K. A. Hasty, R. A. Reife, M. A. Cremer, A. H. Kang, and J. M. Stuart. eases. The therapeutic potential for intervening in this pathway 1992. Induction of arthritis with monoclonal to collage. J. Immunol. remains to be tested in the clinic. It seems quite likely that T 148: 2103–2108. FH 24. Ternthan, D. E., A. S. Townes, and A. H. Kang. 1977. Autoimmunity to type II cells play a prominent role in the production of pathogenic auto- collagen: an experimental model of arthritis. J. Exp. Med. 146: 857–868. antibodies in the diseased state and the role that ICOS/B7RP-1 25. Okamoto, T., S. Satio, H. Yamanaka, T. Tomatsu, N. Kamatani, H. Ogiuchi, T. Uchiyama, and J. Yagi. 2003. Expression and function of the co-stimulator plays in the development of this unique cell population suggests H4/ICOS on activated T cells of patients with rheumatoid arthritis. J. Rheumatol. that blocking Abs be investigated for their therapeutic potential. 30: 1157–1163. 1428 B7RP-1 REGULATES TFH CELLS IN AUTOIMMUNITY

26. Singh, R. R. 2005. SLE: translating lessons from model systems to human dis- function of inducible co-stimulator in patients with systemic lupus erythemato- ease. Trends Immunol. 26: 572–579. sus: possible involvement in excessive -r and anti-double-stranded 27. Nurieva, R. I., P. Treuting, J. Duong, R. A. Flavell, and C. Dong. 2003. Inducible DNA antibody production. Arthritis Res. Ther. 8: R62. costimulator is essential for collagen-induced arthritis. J. Clin. Invest. 111: 32. Tesciuba, A. G., S. Subudhi, R. P. Rother, S. J. Faas, A. M. Frantz, D. Elliot, 701–706. J. Weinstock, L. A. Matis, J. A. Bluestone, and A. I. Sperling. 2001. Inducible 28. Yu, D., A. H. M. Tan, X. Hu, V. Athanasopoulos, N. Simpson, D. G. Silva, costimulator regulates Th2-mediated inflammation, but not Th2 differentiation, in A. Hutloff, K. M. Giles, P. J. Leedman, et al. 2007. Roquin represses autoim- a model of allergic airway disease. J. Immunol. 167: 1996–2003. munity by limiting inducible T-cell co-stimulator messenger RNA. Nature 450: 33. Sharpe, A. I., and J. A. Bluestone. 2001. ICOS costimulation: it’s not just for TH2 299–304. cells anymore. Nat. Immunol. 2: 573–574. 29. Vinnuesa, C. G., M. C. Cook, C. Angelucci, V. Athanasopoulos, L. Rui, 34. Liang, L., E. M. Porter, and W. C. Sha. 2002. Constitutive expression of the B7h K. M. Hill, D. Yu, H. Domaschenz, B. Whittle, T. Lambe, et al. 2005. A RING- ligand for inducible costimulator on naive B cells is extinguished after activation type ubiquitin ligase family member required to repress follicular helper T cells by distinct B cell receptor and 4 receptor-mediated pathways and can and autoimmunity. Nature 435: 452–458. be rescued by CD40 signaling. J. Exp. Med. 196: 97–108. 30. Hutloff, A., K. Buchner, K. Reiter, H. J. Baelde, M. Odendahl, A. Jacobi, 35. Dong, C., U. A. Temann, and R. A. Flavell. 2001. Cutting edge: critical role of T. Dorner, and R. A. Kroczek. 2004. Involvement of inducible costimulator in the inducible costimulator in germinal center reaction. J. Immunol. 166: 3659–3662. exaggerated memory B cells and generation in systemic lupus ery- 36. Bossaller, L., J. Burger, R. Draeger, B. Grimbacher, R. Knoth, A. Plebani, thematosus. Arthritis Rheum. 50: 3211–3220. A. Durandy, U. Baumamm, M. Schleiser, A. W. Welcher, et al. 2006. ICOS 31. Kawamoto, M., M. Harigai, M. Hara, Y. Kawaguchi, K. Tezuka, M. Tanaka, deficiency is associated with a severe reduction of CXCR5ϩCD4ϩ germinal T. Sugiura, Y. Katsumata, C. Fukasswa, H. Ichida, et al. 2006. Expression and center Th cells. J. Immunol. 177: 4927–4932. by guest on October 1, 2021. Copyright 2009 Pageant Media Ltd. http://classic.jimmunol.org Downloaded from