B7S1, a Novel B7 Family Member that Negatively Regulates T Cell Activation

Immunity, Vol. 18, 863–873, June, 2003, Copyright 2003 by Cell Press

B7S1, a Novel B7 Family Member that Negatively Regulates T Cell Activation Durbaka V.R. Prasad, Sabrina Richards, Xoi Muoi Mai, and Chen Dong* Department of Immunology University of Washington School of Medicine Seattle, Washington 98195

Summary T cell activation by antigen-presenting cells (APC) is regulated by positive and negative costimulatory molecules in the B7 family. Here we describe a novel addition in this family, designated as B7S1, which is uniquely anchored to the cell membrane via a GPI linkage. B7S1 is expressed on professional APC and widely distributed in nonlymphoid tissues. A soluble B7S1-Ig fusion protein binds to activated but not naive T cells. B7S1Ig inhibits T cell activation and IL-2 production. A monoclonal antibody that blocks binding of B7S1 to its receptor enhances T cell proliferation in vitro and exacerbates experimental autoimmune encephalomyelitis in vivo. This study identifies a novel negative regulator of T cell activation and further reveals complex costimulatory regulation of immune responses. Introduction T cell activation requires two signals: one via TCR recognition of antigenic peptides presented by MHC molecules and the other from costimulatory molecules on the antigen-presenting cells (APC) (Mueller et al., 1989). The best-characterized costimulatory molecules are CD80 and CD86, also known as B7.1 and B7.2, respectively, which are expressed by professional APC as a result of innate activation (Janeway and Medzhitov, 2002). The receptors for CD80 and CD86 are CD28 and CTLA4. CD28 is expressed by naive and activated T cells, and plays a major role in T cell activation. Mice lacking CD28 or both CD80 and CD86 are impaired in T cell immune responses in vitro and in vivo (Lenschow et al., 1996). CTLA4, on the other hand, is induced after T cell activation and binds to the same ligands with a higher affinity (Chambers and Allison, 1999). CTLA4 carries an ITIM motif in its cytoplasmic region and functions as a negative regulator of T cell activation; CTLA4 knockout mice develop profound autoimmune diseases (Thompson and Allison, 1997). Therefore, CD28 and CTLA4 engaged by CD80 and CD86 molecules on APC play essential roles in maintaining the threshold of T cell activation. In the past several years, the B7 ligand and the corresponding CD28 receptor families have been vastly expanded. Inducible costimulator (ICOS), a third member of the CD28 family, is expressed on activated but not naive T cells, and recognizes its own ligand B7h (also named as B7RP-1 etc) (Hutloff et al., 1999; Yoshinaga et al., 1999). B7h is constitutively expressed in certain *Correspondence: [email protected]

APC such as B cells and macrophages, and can be induced in nonlymphoid tissues and cells by inflammatory stimuli (Swallow et al., 1999; Yoshinaga et al., 1999). Recently, we as well as others have generated ICOSdeficient mice and indicated ICOS as an important regulator of T cell activation, differentiation, and function (Dong et al., 2001; McAdam et al., 2001; Nurieva et al., 2003; Tafuri et al., 2001). PD-1, another ITIM-containing receptor expressed on activated T cells, binds to B7H1/PDL1 and PDL2/B7DC that are broadly expressed in APC and nonlymphoid tissues (Dong et al., 1999; Freeman et al., 2000; Latchman et al., 2001; Tseng et al., 2001). PD-1 plays an important role in maintaining immune tolerance as PD-1-deficient mice develop multiple autoimmune diseases on different genetic backgrounds (Nishimura et al., 1999, 2001). B7-H3 is the newest addition to the B7 family whose receptor has not been identified (Chapoval et al., 2001; Sun et al., 2002). B7-H3 was first reported to be expressed by human dendritic cells and to stimulate human T cell proliferation and IFN␥ production (Chapoval et al., 2001). Recently, we identified the mouse B7-H3 homolog that is broadly expressed in lymphoid and nonlymphoid tissues; a soluble mouse B7-H3-Ig fusion protein binds to activated but not naive T cells (Sun et al., 2002). Therefore, the new members of the B7 family are broadly expressed on professional APC and in nonlymphoid tissues, and through their receptors on activated T cells regulate T cell activation and effector function. Here we describe a novel member of the B7 family named as B7 superfamily member 1 (B7S1). B7S1 is expressed on professional APC and is broadly distributed in nonlymphoid tissues. Its expression on B cells is downregulated following their activation. Unlike the other known members of the B7 family, B7S1 is anchored to the cell membrane via a glycosyl phosphatidylinositol (GPI) linkage. A soluble B7S1-Ig protein binds to activated but not naive T cells. B7S1-Ig reduces T cell proliferation by inhibiting IL-2 production. An anti-B7S1 blocking antibody enhances T cell proliferation and IL-2 secretion in vitro. In vivo blockade of B7S1 also led to greater T-dependent immune responses and exacerbation of experimental autoimmune encephalomyelitis (EAE). Therefore, B7S1 is a novel negative costimulator and regulates the threshold of T cell activation. Results Identification of B7S1 as a New Member of the B7 Family To identify novel members of the B7 family with potential function in immune regulation, we performed a homology search in mouse and human EST databases using amino acid sequences of B7h and B7-H3. Human FLJ22418 molecule was found to share significant homology with these two B7-like molecules (data not shown). In addition, a mouse EST clone was found to contain nucleotide sequence encoding amino acids that are similar to the most N terminus of FLJ22418. We

Immunity 864

Figure 1. Identification of B7S1 as a Novel Member of the B7 Superfamily (A) Nucleotide sequence of B7S1 cDNA encompassing the open reading frame. (B) Alignment of deduced amino acid sequence of mouse B7S1 cDNA with its human homologs. The N-terminal leader peptide and C-terminal hydrophobic regions are indicated by straight lines; Ig-like domains, by dotted lines. The residues conserved in all members of the B7 family are indicated by underlines. (C) Phylogenic analysis of the B7 family members. The scores at each branch represent the degree of sequence variability between proteins (0 for identical sequences). Percentage of amino acid identity between B7S1 protein and the other members of the B7 family is indicated.

obtained this clone and completely sequenced it (Figure 1A). The deduced peptide sequence from the mouse open reading frame shares striking homology with the human protein (Figure 1B). They all contain an N-terminal hydrophobic region that can serve as a leader peptide, two immunoglobulin (Ig)-like domains, and a hydrophobic C terminus (Figure 1B). Therefore, this novel protein shares common structural features with known members of the B7 family. In a phylogenic analysis, we found it to be similar to all existing B7 family members with 20%–30% amino acid identity (Figure 1C). It is noteworthy that this degree of homology is also observed among other B7 molecules and even between B7.1 and B7.2 that share the same receptors (Carreno and Collins, 2002). However, all critical cysteine residues as well as the DxGxYxC motif in the first Ig-like domain are conserved in these proteins (Figure 1B). We also compared the amino acid sequence of B7S1 with those in the CD28 family—CD28, ICOS, and PD-1, which also possess extracellular Ig-like domains, and found no significant homology (data not shown). Therefore this molecule is a novel member of the B7 family and we name it B7 superfamily member 1 (B7S1).

We constructed a mouse B7S1-Ig fusion protein as we previously reported for mouse B7-H3 (Sun et al., 2002). In brief, the nucleotide sequence encoding the extracellular portion of the mouse B7S1 molecule (amino acids 31-254) was amplified and cloned into the DESIg insect expression vector. This new plasmid was stably transfected into the S2 Drosophila cells that can be induced to secrete large amounts of B7S1-Ig fusion proteins. B7S1-Ig protein produced in this fashion was then purified by a Protein A column. We used B7S1-Ig fusion protein to immunize a female Lewis rat as described in the Experimental Procedures section. The harvested draining lymph node cells were fused with Ag8.653. After selection, hybridoma clones were screened by ELISA for production of antibodies that react against B7S1-Ig but not to human IgG1. Three IgG monoclonal antibodies were generated by this means, which also stained 293 cells transiently transfected with a mouse B7S1 expression vector but not to the mock transfectant (Figure 2A; data not shown). We have characterized two of them, the clone 9 and 54 antibodies, in more detail and found they did not bind to 293 cells transfected with a B7-H3 expression plasmid

B7S1, an Inhibitor of T Cell Activation 865

Figure 2. B7S1 Anchors to the Cell Membrane via a GPI Linkage (A) Generation of monoclonal antibodies to mouse B7S1. Supernatants from anti-B7S1 hybridoma clones 9 and 54 were used to stain B7S1-, mock-, or B7-H3-transfected 293 cells, and the staining was revealed by an anti-rat IgG-FITC. (B) B7S1 expression is sensitive to PI-PLC treatment. 293 cells transfected with a B7S1 expression vector (for B7S1) or EL-4 cells (for Thy1 and ICOS) were treated with PI-PLC at 37⬚C for 30 min. Cells with (PI-PLC) or without the treatment (NT) were stained with antibodies to these antigens. Mean fluorescence (MF) of each staining is indicated.

(Figure 2A). In addition, PDL2-Ig did not block the binding of these antibodies (data not shown). These results indicate that we have generated B7S1-specific antibodies. Although the C terminus of B7S1 protein is hydrophobic, it is not followed by any residues that can be predicted as intracellular amino acids. This structural characteristic is usually shared by cell surface GPI-anchored proteins (Udenfriend and Kodukula, 1995). We therefore tested whether cell surface expression of B7S1 is sensitive to phosphatidylinositol-specific phospholipase C (PI-PLC) treatment. Indeed, PI-PLC reaction resulted in ⬎50% reduction of B7S1 expression on transfected 293 cells (Figure 2B). Similarly, expression of native B7S1 protein on splenic B cells (see below) was also sensitive to the treatment of PI-PLC (data not shown). As a positive control, expression of GPI-anchored Thy1 molecule on EL-4 cell line can be reduced to a similar extent by PI-PLC while transmembrane protein ICOS is largely insensitive (Figure 2B). B7S1 thus represents the first GPI-linked protein in the B7 family although the functional significance of this configuration is unknown at this stage.

Expression of B7S1 Molecule by Professional APC All known members of the B7 family are expressed by professional APC. In addition, the new members of this family, i.e., B7h, PDL1, PDL2, and B7-H3 are broadly distributed in nonlymphoid tissues and cells. To assess the possible immune regulation by B7S1, we first examined the expression of B7S1 mRNA by a Northern blot analysis. A cDNA probe consisting of the coding sequence for the extracellular region of B7S1 is utilized. B7S1 expression can be detected in lymphoid tissues thymus and spleen, and in a number of nonlymphoid organs (Figure 3A). Ubiquitous expression of B7S1 in nonlymphoid tissues suggests its role in modulating immune responses in these tissues. In most tissues, two mRNA species were detected, but the nature of this size difference is unclear at this stage. We then further examined B7S1 expression in immune cells by use of the anti-B7S1 monoclonal antibody. In thymus, B7S1 is expressed by a minor population of cells that also express CD4 and CD8 (Figure 3B). In spleen, B7S1 is not expressed by CD4⫹, CD8⫹, or CD11C⫹ cells, but is constitutively on all B220⫹ B cells (Figure 3B). Therefore in secondary lymphoid organs, B7S1 appears to exhibit B cell-specific expression.

Immunity 866

Figure 3. Expression of B7S1 in Tissues and by Immune Cells (A) A PCR fragment consisting of two Ig-like domains of the mouse B7S1 gene was used to hybridize mouse tissue Northern blot (Seegene, Inc., Korea). (B) Expression of B7S1 by thymocytes, spleen cells, and peritoneal macrophages. Cells were stained with a biotinylated anti-B7S1 antibody in conjunction with other indicated markers.

However, B7S1 expression can be detected on peritoneal CD11b⫹ macrophages elicited by thioglycollate (Figure 3B) and on some bone marrow-derived CD11C⫹ dendritic cells (data not shown). This analysis indicates that B7S1 is expressed by a variety of professional APC and may possess regulatory roles on T cells. Members of the B7 family are differentially regulated in professional APC by various stimuli. For instance, CD80 and CD86 expression can be induced by innate

activation (Janeway and Medzhitov, 2002; Lenschow et al., 1996); B7h is downregulated on B cells by IgM engagement and upregulated on fibroblasts by TNF␣ (Liang et al., 2002; Swallow et al., 1999). We therefore examined whether B7S1 is regulated in APC. LPS treatment of peritoneal macrophages or bone marrowderived dendritic cells did not significantly alter the B7S1 expression (data not shown). On the other hand, stimulation of purified splenic B cells by different stimuli, includ-

B7S1, an Inhibitor of T Cell Activation 867

Figure 4. B7S1-Ig Binds to Activated T Cells Lymph node cells from a C57BL/6 mouse were activated with ConA for 48 hr, and cells before and after activation were analyzed for B7S1-Ig binding together with antibodies for CD4 and CD8.

ing LPS, IL-4, anti-IgM, and anti-CD40, all resulted in 50%–70% reduction of B7S1 expression: LPS reduced B7S1 expression by 54%, IL-4 by 60%, and anti-IgM or anti-CD40 by 71%. This result indicates that B7S1 is downregulated after B cell activation. Inhibition of T Cell Activation by B7S1-Ig Expression of B7S1 on professional APC suggests a role of B7S1 in regulation of T cell immune responses. We employed our B7S1-Ig fusion protein to assess if B7S1 has a putative receptor on T cells. CD4⫹ and CD8⫹ T cells from C57BL/6 lymph nodes are not strongly bound by the biotinylated B7S1-Ig; after ConA activation for 2 days, all of them do (Figure 4). B7S1-Ig bound to CD28⫺/⫺ cells in the same degree (data not shown). B7.1-Ig did not block binding of B7S1-Ig to activated T cells (data not shown), suggesting that it does not recognize CD28 or CTLA4. In addition, B7S1-Ig could bind well to ICOS⫺/⫺ cells activated by ConA (data not shown). It does not stain 293 cells transfected with PD-1, which could be recognized by an anti-PD-1 antibody (data not shown). B7-H3-Ig and B7S1-Ig did not reciprocally block the binding of the other fusion protein to their corresponding receptor on activated T cells (data not shown). All these data indicate a putative receptor for B7S1 on activated T cells, which is distinct from CD28, CTLA4, ICOS, PD-1, and the receptor for B7-H3. To assess the function of B7S1 on T cell activation and function, we stimulated purified CD4⫹ cells from C57BL/6 (Figure 5A) or OT-II TCR transgenic (Figure 5B) mice with different doses of anti-CD3 in the absence or presence of B7.1-Ig or B7S1-Ig and measured cell proliferation. Anti-CD3 plus a human IgG1 control resulted in the similar proliferation of stimulated T cells as the ones treated with anti-CD3 only (data not shown). B7.1-Ig, as expected, strongly enhanced T cell stimulation (Figure 5A). B7S1-Ig on the other hand inhibited T cell proliferation (Figures 5A and 5B). To rule out the possibility that this inhibitory effect was artificially de-

rived from our insect culture system, we employed an irrelevant protein containing the human IgG1 tag expressed and prepared in the same fashion as B7S1-Ig and found it had no effect on T cell proliferation (data not shown). B7S1-Ig inhibited proliferation of OT-II transgenic T cells in a dose-dependent manner (Figure 5C). B7S1-Ig treatment also moderately reduced CD25 and CD44 upregulation on activated OT-II T cells (data not shown). In the presence of CD28 costimulation, B7S1-Ig inhibited T cell proliferation, most potently when low dose of anti-CD3 was used (Figure 5D). Interestingly, strong TCR and CD28 costimulation can partially overcome this inhibition (Figure 5D). These results indicate B7S1 as a negative regulator of T cell activation and B7S1-Ig treatment render the cells less responsive to TCR and CD28 signaling. The hallmark of T cell activation is the production of IL-2, which drives T cell clonal expansion. We thus examined whether IL-2 production is affected by B7S1Ig costimulation. While B7.1-Ig strongly potentiated IL-2 production, B7S1-Ig inhibited (Figure 5E). B7S1-Ig also inhibited IL-2 production by cells treated with anti-CD3 and anti-CD28 (data not shown). To assess whether inhibition of T cell proliferation by B7S1-Ig was the result of IL-2 reduction, we added exogenous IL-2 to the OT-II T cells treated with anti-CD3 with or without B7S1-Ig. Addition of IL-2 fully restored the proliferation of T cells costimulated with B7S1-Ig (Figure 5F). Therefore, B7S1 inhibits T cell activation via reducing the IL-2 production. Effector Th cells differentiated in this fashion exhibited no cytokine defect (data not shown). Our work suggests that B7S1 primarily inhibits T cell activation and IL-2 production. IL-2 gene induction in activated T cells is the result of multiple signaling pathways from cell surface receptors leading to activation of NFAT, NF-␬B, and AP1 transcription factors (Jain et al., 1995; Serfling et al., 1995). We assess whether B7S1-Ig costimulation resulted in an inhibition of IL-2 transcription machinery. OT-II T cells were treated with anti-CD3 in the presence or absence

Immunity 868

Figure 5. B7S1-Ig Inhibits T Cell Proliferation and IL-2 Production (A–C) CD4 T cells isolated from C57BL/6 (A) or OT-II (B and C) mice were treated with indicated doses of various stimuli, and T cell proliferation measured by 3H-thymidine incorporation. (D) CD4 T cells from C57BL/6 mice were stimulated with indicated doses of anti-CD3 with or without anti-CD28 (2 ␮g/ml) in the presence or absence of B7S1-Ig and their proliferation measured. (E) OT-II T cells were treated by indicated means for 24 hr and IL-2 expression measured by ELISA. (F) Exogenous IL-2 restored proliferation by B7S1-treated OT-II cells. OT-II cells were treated as in (B) at the presence of exogenous IL-2 (30 units/ml) and cell proliferation assayed.

of B7S1-Ig costimulation and nuclear extracts prepared 16 hr after stimulation. Nuclear localization of NFATc1 and c-rel transcription factors, both of which bind to the IL-2 promoter and are important for its induction, is the result of T cell activation (Kontgen et al., 1995; Peng et al., 2001; Zhou et al., 2002). B7S1-Ig costimulation did not reduce the amount of these two factors in the nucleus as examined by Western blotting (data not shown). However, the expression of JunB, a component of the AP-1 family induced after T cell activation, was reduced by 49% after B7S1 costimulation. On the other hand, there was little change in c-Jun expression (11% reduction) with B7S1-Ig treatment. JunB has been previously shown to bind to the IL-2 promoter (Boise et al., 1993) and JunB overexpression resulted in greater IL-2 production (our unpublished data). Since JunB is induced after T cell activation (Jain et al., 1995), the mechanism

of which is unknown, B7S1 costimulation may result in inefficient JunB induction. Overall, this analysis suggests B7S1-Ig treatment may lead to a selective signaling and transcription defect. Enhanced T Cell Activation by an Anti-B7S1 Blocking Antibody To assess the physiological importance of B7S1 binding to its corresponding receptor in immune responses, we employed an anti-B7S1 blocking antibody for our in vitro and in vivo analysis. We found that clone 54 antibody inhibited binding of B7S1-Ig to activated T cells (Figure 6A), while clone 9 did not significantly (data not shown). We first assessed the function of this blocking antibody in vitro by activating splenocytes from C57BL/6 mice with different doses of anti-CD3. In this experiment, positive and negative costimulation is provided by dif-

B7S1, an Inhibitor of T Cell Activation 869

T cells (Figure 6C). This work substantiates the above data using B7S1-Ig and indicates that B7S1 is a physiological negative regulator of T cell activation and IL-2 expression. To examine if negative regulation by B7S1 has any important immune function in vivo, we immunized C57BL/6 mice at their base of tail with KLH protein emulsified in CFA. A control rat IgG or the B7S1 blocking antibody was injected into experimental mice every other day for a total of 3 times. 8 days after the immunization, the mice were sacrificed and anti-KLH antibody in the serum was measured. We found that treatment with anti-B7S1 blocking antibody led to greater anti-KLH IgM (Figure 7A) and IgG (data not shown) production, indicative of a stronger immune responses in vivo. We also collected spleen cells from immunized mice and restimulated them in vitro with or without 10 ␮g/ml KLH. Cells from anti-B7S1 treated mice consistently exhibited greater proliferation and IL-2 production (Figures 7B and 7C), also suggesting that stronger T cell priming occurred in vivo in the presence of anti-B7S1 blocking antibody. To assess the importance of B7S1 in T cell activation and tolerance, we immunized C57BL/6 with MOG35-55 peptide as we described before to induce EAE disease (Dong et al., 2001). Control or anti-B7S1 blocking antibody was injected into mice during T cell priming phase, i.e., between first and second immunization. Mice treated with anti-B7S1 blocking antibody consistently developed greatly accelerated and much more robust EAE than those treated with a control rat antibody (Figure 7D). When we examined infiltrating mononuclear cells in the brain of experimental mice, we found that anti-B7S1 antibody treatment in mice resulted in greater CD4 and CD8 cell infiltration, and also remarkably increased CD11b⫹ macrophages (Figure 7E). This work strongly supports an essential function of B7S1 in negative regulation of T cell activation. Discussion

Figure 6. Anti-B7S1 Blocking Antibody Enhanced T Cell Proliferation and IL-2 Production In Vitro (A) Biotinylated B7S1-Ig was incubated with a rat control Ig (no blocking) or clone 54 anti-B7S1 (blocking) before staining with ConA-activated mouse lymph nodes cells. (B and C) Spleen cells from C57BL/6 mice were incubated with indicated doses of anti-CD3 at the presence of 5 ␮g/ml control rat IgG or purified clone 54 antibody. Cell proliferation (B) was measured by 3H-thymidine uptake after 72 hr and IL-2 (C) assayed by ELISA 24 hr after the treatment.

ferent splenic APCs, mostly B cells. While a control rat IgG did not alter T cell proliferation, treatment with antiB7S1 blocking antibody greatly enhanced it (Figure 6B). We also measured IL-2 production within the first 24 hr of treatment, and found that B7S1 blocking antibody also greatly increased the levels of IL-2 production by

T cell activation requires two signals: one from TCR and the other from costimulatory molecules, most notably those in the B7 family. Recent expansion of the B7 family has revealed complex costimulatory regulation of T cell activation (Frauwirth and Thompson, 2002; Carreno and Collins, 2002; Sharpe and Freeman, 2002). The new members of the B7 family are more broadly expressed, bind to receptors on activated T cells, and exert positive and negative regulation of T cell activation program. Here we describe for the first time one such molecule— B7S1. We show that it functions as a negative regulator of T cell activation and IL-2 production. B7S1, like other B7 family members, possesses one pair of Ig-like domains in its extracellular region (Figure 1B). However, it lacks an obvious transmembrane region. We showed that its cell surface expression is sensitive to PI-PLC treatment (Figure 2B) and concluded that B7S1 is anchored to the cell membrane via a GPI linkage. Thus B7S1 is the first GPI-linked protein in the B7 family, and the rest are type I transmembrane glycoproteins. The functional significance for the unique cell membrane anchorage of B7S1 is unclear. Although originally thought

Immunity 870

Figure 7. Anti-B7S1 Blocking Antibody Enhanced T-Dependent Immune Responses and EAE Disease In Vivo (A–C) C57BL/6 mice (three in each group) immunized with KLH in CFA were treated with a rat control Ig or anti-B7S1 blocking antibody. Eight days after the immunization, experimental mice were sacrificed and anti-KLH serum IgM was measured individually by ELISA and averaged for each group (A). Spleen cells from each immunized mouse were restimulated in vitro with or without KLH and T cell proliferation (B) and IL-2 production (C) was measured an average for the control antibody or anti-B7S1 groups is indicated. (D and E) C57BL/6 mice (five mice in each group) immunized with MOG peptide to induced EAE were treated with a rat control Ig or antiB7S1 blocking antibody. (D) EAE disease in these mice was scored as described in the Experimental Procedures. The result shown is a representative of two independent experiments with similar results. (E) Mononuclear cells in CNS from mice with EAE were typed by staining with anti-CD4, CD8, or CD11b.

to be more mobile on cell membrane than transmembrane proteins, GPI-linked proteins are preferentially localized in the lipid rafts. Lipid rafts on T cells are thought to play important roles in the organization of immunological synapses. However, its role in APC has not been well established. Recently, MHC class II molecules have been shown to be localized in lipid rafts, which allows efficient antigen presentation (Anderson et al., 2000). With a GPI linkage, B7S1 may be in close proximity to the MHC and this spatial organization may allow efficient negative costimulation to occur. With the monoclonal antibodies we generated against mouse B7S1 molecule, we found B7S1 is expressed by most professional APC, including some bone marrowderived dendritic cells, all peritoneal macrophages and splenic B cells (Figure 3B). Its expression on B cells can be downregulated by different stimulation. This observation suggests costimulatory regulation of T cells by

B7S1 is influenced by the activation status of B cells. It is noteworthy that other B7 family members are regulated differentially in B cells. CD80 and CD86 are well known to be upregulated after B cell activation while B7h is downregulated only after IgM crosslinking (Liang et al., 2002). All these suggest a combinatorial model for costimulation of T cells: each B7 ligand is regulated differentially, which reflects the natural history and activation states of APC, and the combination of these ligands regulate the threshold of T cell activation. On the other hand, the new B7 family members- B7h, PDL1/2, B7H3, and B7S1 are also widely distributed in nonlymphoid tissues. Since their receptors are expressed on activated T cells, they may possess important function in modulating effector T cell function once activated T cells migrate into the nonlymphoid tissues. It is also possible that at this effector stage, the combinatorial signals presented by B7 ligands that are tissue-specific and regu-

B7S1, an Inhibitor of T Cell Activation 871

lated by inflammatory cytokines may influence the nature and extent of T cell function. A B7S1-Ig fusion protein we prepared bound to activated but not naive CD4⫹ and CD8⫹ cells. Therefore, B7S1 joins B7h, PDL1/2, and B7-H3 to recognize receptors induced after T cell activation. The nature of the B7S1 receptor is unknown at this stage, but appears distinct from known members of the CD28 receptor family. B7S1-Ig can bind to CD28⫺/⫺ and ICOS⫺/⫺ cells; this binding is not inhibited by a B7.1-Ig or B7-H3-Ig. In addition, it does not react with a PD-1 transfectant. The CD28 family is thus expected to expand in the near future. Also with the fusion protein, we tested the function of B7S1 on T cell activation and differentiation. B7S1-Ig very potently inhibited CD4⫹ T cell proliferation (Figures 5A–5D). We further showed that B7S1 reduction of T cell proliferation is through an IL-2-dependent mechanism. B7S1-Ig greatly reduced the IL-2 production by activated T cells and addition of exogenous IL-2 restored the proliferation of T cells by B7S1-Ig-treated cells (Figures 5E and 5F). These data demonstrated that B7S1 can serve as a negative regulator of T cell activation and IL-2 production. How B7S1 works is unclear at this stage, but we found that JunB induction in activated T cells is most significantly reduced when T cells are costimulated by B7S1. JunB is induced after T cell activation and functions to regulate IL-2 gene transcription. It is interesting that B7S1 at this stage does not globally inhibit all signaling pathways but instead selectively blocks a specific mechanism. Further analysis of B7S1 signal transduction will therefore reveal more complex signaling mechanisms of costimulatory molecules. We have examined the physiological significance of B7S1 costimulation by use of a blocking antibody. Treatment of anti-CD3-activated splenocytes with this antibody greatly enhanced IL-2 production and T cell proliferation (Figures 6B and 6C). This work, in agreement with our results using B7S1-Ig (Figures 5A–5F), indicates that B7S1 expressed on APC does function physiologically to limit the amount of IL-2 expression by activated T cells and hence the extent of their clonal expansion. More importantly, blocking of B7S1 function led to greater T cell priming and function in vivo. In an immunization experiment, we found that treatment of immunized mice with anti-B7S1 blocking antibody led to greater antigen-specific Ig production (Figure 7A). Furthermore, enhanced IL-2 production and T cell proliferation by restimulated spleen cells harvested from antiB7S1 treated mice also support that T cell priming was indeed enhanced in vivo (Figures 7B and 7C). Thus blocking B7S1 may possess potential therapeutic values in enhancing wanted immune responses. We further assess the importance of B7S1 negative costimulation using an autoimmunity mouse model— MOG-induced EAE in C57BL/6 mice. Anti-B7S1 blocking antibody treatment led to an extreme EAE disease. It is noteworthy that the antibody was injected into experimental mice at T cell priming phase, i.e., after the first immunization but before the second. The disease normally becomes observable only after the second immunization. Therefore, anti-B7S1 effect may be due to the enhanced expansion of autoreactive T cells. In support of this idea, we did find greater CD4⫹ and CD8⫹ T cell infiltration into the brain tissue in mice treated with anti-

B7S1 antibody (Figure 7E). Macrophages were also greatly increased in number in these mice as well. These results indicate an essential regulation of B7S1 in negative regulation of T cell-dependent autoimmune reaction. From all data presented in this paper, the primary cause of this greatly enhanced autoimmunity may be primarily due to excessive T cell activation and clonal expansion in vivo. However, we cannot rule out that B7S1 may be additionally important to inhibit certain effector or to generate regulatory T cells in vivo that function to contain autoimmunity, although we do not have any supporting evidence from our current work along this line. Nonetheless, the EAE experiment strongly suggests that B7S1 contributes in certain degrees to the maintenance of peripheral tolerance and to the containment of autoimmune diseases. Overall, we identify a novel member of the B7 family that serves as a negative regulator of T cell proliferation and IL-2 production. B7S1 is downregulated after B cell activation, which may result in lowering the T cell activation threshold. Therefore this study further demonstrates complex regulation of T cell activation via positive and negative costimulatory molecules as suggested by others (Frauwirth and Thompson, 2002; Carreno and Collins, 2002; Dong et al., 2002; Sharpe and Freeman, 2002). Therapeutic intervention of these costimulatory pathways may help modulate immune responses in many diseases. Experimental Procedures B7S1 Cloning and Analysis Mouse B7S1 EST clone was purchased from Incyte and completely sequenced. The Ig-like domains of the deduced B7S1 protein were predicted by CD-Search program of NCBI. The alignment of mouse and human B7S1 protein sequences was performed using Jellyfish software, Biowire. The phylogeny tree was constructed with GeneWorks software. Comparison of amino acid sequences of B7S1 with other B7 family members was performed using the blast2 program of NCBI and DS Gene 1.1. Mouse tissue Northern blot was purchased from Seegene, Inc., Seoul, Korea. A PCR fragment consisting of two Ig-like domains of the mouse B7S1 gene, which was subsequently cloned into the insect expression vector (see below), was used for the analysis. The final wash of the blot was performed with 0.1⫻ SSC/0.1% SDS at 65⬚C. B7S1 Expression The sequence coding the extracellular portion of the protein was PCR amplified and cloned into the DES-Ig plasmid we previously described (Sun et al., 2002). The PCR primers, with restriction sites, were: forward, AGATCTATTTCAGGCAAGCACTTCATCAC, and reverse, ACTAGTGAGCAGCTGCAGCTGACTCCGCC. The new B7S1Ig expression vector was stably transfected into Drosophila S2 cells which upon CuSO4 induction, secrete B7S1-Ig fusion protein. The B7S1-Ig fusion protein produced by this fashion was further purified by a Protein A column. B7S1, B7-H3 and PD-1 expression vectors were transfected into 293 cells by Lipofectamin 2000, Clontech. Generation of Anti-B7S1 Monoclonal Antibodies A female Lewis rat (3–4 months old) was immunized with 100 ␮g B7S1-Ig in complete Freund’s adjuvant (CFA) at the foodpad, axial, and linguinal areas and boosted every 3rd day 5 times in the same protein quantity with antigen in IFA. The draining lymph node cells were harvested 1 day after the last boost and fused with Ag8.653 cells by PEG1500 as previously described (Kubagawa et al., 1999). ELISA was performed to identify the clones that produced IgG antibodies reacting with B7S1-Ig fusion protein but not with control human IgG1. These clones were further subcloned and a total of 3 clones were generated.

Immunity 872

Flow Cytometry Analysis B7S1-Ig and anti-B7S1 were purified by Protein A and G, respectively, and biotinylated with Sulfo-NHS-LC-Biotin (Pierce). These reagents were used in conjunction with anti-CD4, CD8, CD11b, CD11c, and B220 antibodies (Pharmingen) for analysis of various population of immune cells by flow cytometry. Before staining cells with B7S1-Ig, we preblocked nonspecific binding with a human IgG1 as described earlier (Sun et al., 2002). PI-PLC Treatment 293 cells transfected with a B7S1 expression vector, splenocytes, and EL4 cells were treated with PI-PLC (Sigma) in Hank’s solution for 30 min at 37⬚C as previously described (Dong et al., 1994), followed by flow cytometry analysis. B Cell Preparation and Activation Spleen B cells were prepared by anti-Thy1.2-mediated complement lysis followed by plate depletion of adherent cells. These cells were treated with plate-bound anti-IgM (10 ␮g/ml) or anti-CD40 (10 ␮g/ ml, Pharmingen), LPS (10 ␮g/ml), or IL-4 (500 units/ml) for 3 days, and B7S1 expression examined via flow cytometry analysis, together with the B220 antibody (Pharmingen). T Cell Activation and Differentiation CD4⫹ T cells from C57BL/6 or OT-II mice were purified as described (Dong et al., 2000, 2001). The cells were treated with plate-bound anti-CD3 in the absence or presence of human IgG1 (Sigma), B7.1Ig (R&D), or B7S1-Ig. After coating, plates were washed 3 times with PBS before cell culture. IL-2 production was measured 24 hr after T cell activation, and cell proliferation was measured 72 hr after the treatment with 3H-thymidine in the last 8 hr. CD4 T cell differentiation and restimulation of effector T cells were performed according to what we have previously described (Dong et al., 2000, 2001). In the in vitro experiments using the blocking antibody, we treated total splenocytes with different doses of anti-CD3 at the presence of a control or B7S1 blocking antibody (5 ␮g/ml). IL-2 and T cell proliferation was assayed as above. Nuclear Extract Preparation and Immunoblotting Analysis Nuclear extracts of T cells activated for 24 hr were prepared as previously described (Dong et al., 1998) and expression of several key transcription factors was analyzed by Western blot analysis using antibodies from Santa Cruz. Quantitation of autoradiography intensity was performed by NIH Image software. KLH Immunization C57BL/6 mice were treated with KLH protein (Sigma) emulsified in CFA at the base of tail. On day 0, 3, and 6, the mice were given intraperitonally a control rat IgG or B7S1 blocking antibody (100 ␮g each). On day 8, the immunized mice were sacrificed and three mice from each group were analyzed individually for their immune responses. Splenocytes were restimulated with KLH to measure T cell proliferation and IL-2 expression. Anti-KLH antibodies of different isotypes in the serum of immunized mice were measured by ELISA. EAE Induction and Analysis EAE was induced in C57BL/6 mice with MOG 35-55 peptide as we previously described (Dong et al., 2001) by immunizing twice with the peptide in CFA and two subsequent treatment of pertussis toxin, 1 day after each immunization. The control or blocking antibodies were injected into experimental mice on day 1, 4, and 7 of the first immunization. The mice were observed daily for clinical signs and scored on a scale of 0–5 with gradations of 0.5 for intermediate scores: 0, no clinical signs; 1, loss of tail tone; 2, wobbly gait; 3, hindlimb paralysis; 4, hind- and forelimb paralysis; 5, death. Preparation and stimulation of mononuclear cells from brain tissues were done as described (Dong et al., 2001). Acknowledgments We thank Dr. Max Cooper and Ms. Lanier Gartland for their kind help in hybridoma generation and provision of Ag8 cells. We are

also grateful to Dr. Toby Kollman for providing bone marrow-derived dendritic cells, Dr. Thea Brabb for her help with mononuclear isolation from CNS, Dr. Thang Nguyen for critical reading of the manuscript, and to the entire Dong lab for their help and discussion. This work is supported in part by a start-up fund from the Department of Immunology at University of Washington and by a grant from National Institute of Health (to C.D.). C.D. is a recipient of an Arthritis Investigator award of the Arthritis Foundation and a Basil O’Connor Starter Scholar award from March of Dimes Foundation. Received: April 9, 2003 Revised: May 1, 2003 Accepted: May 7, 2003 Published: June 17, 2003 References Anderson, H.A., Hiltbold, E.M., and Roche, P.A. (2000). Concentration of MHC class II molecules in lipid rafts facilitates antigen presentation. Nat. Immunol. 1, 156–162. Boise, L.H., Petryniak, B., Mao, X., June, C.H., Wang, C.Y., Lindsten, T., Bravo, R., Kovary, K., Leiden, J.M., and Thompson, C.B. (1993). The NFAT-1 DNA binding complex in activated T cells contains Fra-1 and JunB. Mol. Cell. Biol. 13, 1911–1919. Carreno, B.M., and Collins, M. (2002). The B7 family of ligands and its receptors: new pathways for costimulation and inhibition of immune responses. Annu. Rev. Immunol. 20, 29–53. Chambers, C.A., and Allison, J.P. (1999). Costimulatory regulation of T cell function. Curr. Opin. Cell Biol. 11, 203–210. Chapoval, A.I., Ni, J., Lau, J.S., Wilcox, R.A., Flies, D.B., Liu, D., Dong, H., Sica, G.L., Zhu, G., Tamada, K., et al. (2001). B7–H3: a costimulatory molecule for T cell activation and IFN-gamma production. Nat. Immunol. 2, 269–274. Dong, C., Wang, J., Neame, P., and Cooper, M.D. (1994). The murine BP-3 gene encodes a relative of the CD38/NAD glycohydrolase family. Int. Immunol. 6, 1353–1360. Dong, C., Yang, D.D., Wysk, M., Whitmarsh, A.J., Davis, R.J., and Flavell, R.A. (1998). Defective t cell differentiation in the absence of Jnk1. Science 282, 2092–2095. Dong, H., Zhu, G., Tamada, K., and Chen, L. (1999). B7–H1, a third member of the B7 family, co-stimulates T-cell proliferation and interleukin-10 secretion. Nat. Med. 5, 1365–1369. Dong, C., Yang, D., Tournier, C., Whitmarsh, A., Xu, J., Davis, R., and Flavell, R. (2000). JNK is required for effector T-cell function but not for T-cell activation. Nature 405, 91–94. Dong, C., Juedes, A.E., Temann, U.-A., Shresta, S., Allison, J.P., Ruddle, N.H., and Flavell, R.A. (2001). ICOS co-stimulatory receptor is essential for T-cell activation and function. Nature 409, 97–102. Frauwirth, K.A., and Thompson, C.B. (2002). Activation and inhibition of lymphocytes by costimulation. J. Clin. Invest. 109, 295–299. Freeman, G.J., Long, A.J., Iwai, Y., Bourque, K., Chernova, T., Nishimura, H., Fitz, L.J., Malenkovich, N., Okazaki, T., Byrne, M.C., et al. (2000). Engagement of the PD-1 immunoinhibitory receptor by a novel B7 family member leads to negative regulation of lymphocyte activation. J. Exp. Med. 192, 1027–1034. Hutloff, A., Dittrich, A.M., Beier, K.C., Eljaschewitsch, B., Kraft, R., Anagnostopoulos, I., and Krocsek, R.A. (1999). ICOS is an inducible T-cell co-stimulator structurally and functionally related to CD28. Nature 397, 263–266. Jain, J., Loh, C., and Rao, A. (1995). Transcriptional regulation of the IL-2 gene. Curr. Opin. Immunol. 7, 333–342. Janeway, C.A., Jr., and Medzhitov, R. (2002). Innate immune recognition. Annu. Rev. Immunol. 20, 197–216. Kontgen, F., Grumont, R.J., Strasser, A., Metcalf, D., Li, R., Tarlinton, D., and Gerondakis, S. (1995). Mice lacking the c-rel proto-oncogene exhibit defects in lymphocyte proliferation, humoral immunity, and interleukin-2 expression. Genes Dev. 9, 1965–1977. Kubagawa, H., Chen, C.C., Ho, L.H., Shimada, T.S., Gartland, L., Mashburn, C., Uehara, T., Ravetch, J.V., and Cooper, M.D. (1999).

B7S1, an Inhibitor of T Cell Activation 873

Biochemical nature and cellular distribution of the paired immunoglobulin-like receptors, PIR-A and PIR-B. J. Exp. Med. 189, 309–318. Latchman, Y., Wood, C.R., Chernova, T., Chaudhary, D., Borde, M., Chernova, I., Iwai, Y., Long, A.J., Brown, J.A., Nunes, R., et al. (2001). PD-L2 is a second ligand for PD-1 and inhibits T cell activation. Nat. Immunol. 2, 261–268. Lenschow, D.J., Walunas, T.L., and Bluestone, J.A. (1996). CD28/ B7 system of T cell costimulation. Annu. Rev. Immunol. 14, 233–258. Liang, L., Porter, E.M., and Sha, W.C. (2002). Constitutive expression of the B7h ligand for inducible costimulator on naive B cells is extinguished after activation by distinct B cell receptor and interleukin 4 receptor-mediated pathways and can be rescued by CD40 signaling. J. Exp. Med. 196, 97–108. McAdam, A.J., Greenwald, R.J., Levin, M.A., Chernova, T., Malenkovich, N., Ling, V., Freeman, G.J., and Sharpe, A.H. (2001). ICOS is critical for CD40-mediated antibody class switching. Nature 409, 102–105. Mueller, D.L., Jenkins, M.K., and Schwartz, R.H. (1989). Clonal expansion versus functional clonal inactivation: a costimulatory signalling pathway determines the outcome of T cell antigen receptor occupancy. Annu. Rev. Immunol. 7, 445–480. Nishimura, H., Nose, M., Hiai, H., Minato, N., and Honjo, T. (1999). Development of lupus-like autoimmune diseases by disruption of the PD-1 gene encoding an ITIM motif-carrying immunoreceptor. Immunity 11, 141–151. Nishimura, H., Okazaki, T., Tanaka, Y., Nakatani, K., Hara, M., Matsumori, A., Sasayama, S., Mizoguchi, A., Hiai, H., Minato, N., et al. (2001). Autoimmune dilated cardiomyopathy in PD-1 receptor-deficient mice. Science 291, 319–322. Nurieva, R.I., Treuting, P., Duong, J., Flavell, R.A., and Dong, C. (2003). Inducible costimulator is essential for collagen-induced arthritis. J. Clin. Invest. 111, 701–706. Peng, S.L., Gerth, A.J., Ranger, A.M., and Glimcher, L.H. (2001). NFATc1 and NFATc2 together control both T and B cell activation and differentiation. Immunity 14, 13–20. Serfling, E., Avots, A., and Neumann, M. (1995). The architecture of the interleukin-2 promoter: a reflection of T lymphocyte activation. Biochim. Biophys. Acta 1263, 181–200. Sun, M., Richards, S., Prasad, D.V., Mai, X.M., Rudensky, A., and Dong, C. (2002). Characterization of mouse and human B7-H3 genes. J. Immunol. 168, 6294–6297. Swallow, M.M., Wallin, J.J., and Sha, W.C. (1999). B7h, a novel costimulatory homolog of B7.1 and B7.2, is induced by TNFa. Immunity 11, 423–432. Tafuri, A., Shahinian, A., Bladt, F., Yoshinaga, S.K., Jordana, M., Wakeham, A., Boucher, L.-M., Bouchard, D., Chan, V.S.F., Duncan, G., et al. (2001). ICOS is essential for effective T-helper-cell responses. Nature 409, 105–109. Thompson, C.B., and Allison, J.P. (1997). The emerging role of CTLA-4 as an immune attenuator. Immunity 7, 445–450. Tseng, S.Y., Otsuji, M., Gorski, K., Huang, X., Slansky, J.E., Pai, S.I., Shalabi, A., Shin, T., Pardoll, D.M., and Tsuchiya, H. (2001). B7-DC, a new dendritic cell molecule with potent costimulatory properties for T cells. J. Exp. Med. 193, 839–846. Udenfriend, S., and Kodukula, K. (1995). How glycosylphosphatidylinositol-anchored membrane proteins are made. Annu. Rev. Biochem. 64, 563–591. Yoshinaga, S.K., Whoriskey, J.S., Khare, S.D., Sarmiento, U., Guo, J., Horan, T., Shih, G., Zhang, M., Coccia, M.A., Kohno, T., et al. (1999). T-cell co-stimulation through B7RP-1 and ICOS. Nature 402, 827–832. Zhou, X.Y., Yashiro-Ohtani, Y., Nakahira, M., Park, W.R., Abe, R., Hamaoka, T., Naramura, M., Gu, H., and Fujiwara, H. (2002). Molecular mechanisms underlying differential contribution of CD28 versus non-CD28 costimulatory molecules to IL-2 promoter activation. J. Immunol. 168, 3847–3854.