Portal de Periódicos da CAPES

NTB-A, a New Activating Receptor in T Cells That Regulates Autoimmune Disease

2004; Elsevier BV; Volume: 279; Issue: 18 Linguagem: Inglês

10.1074/jbc.m312313200

1083-351X

Patricia Valdez, Hua Wang, Dhaya Seshasayee, Menno van Lookeren Campagne, Austin Gurney, Wyne P. Lee, Iqbal S. Grewal,

Immune Response and Inflammation

The CD28 co-stimulatory pathway is well established for T cell activation; however, results from CD28 -/- mice suggest the existence of additional co-stimulatory pathways. Here we report the further characterization of a new member of the CD2 superfamily, NTB-A, important in T cell co-stimulation. NTB-A is expressed on T cells, and its expression is up-regulated on activated cells. Triggering of NTB-A with monoclonal antibodies in the absence of CD28 signals leads to T cell proliferation and interferon-γ secretion but not interleukin-4. Cross-linking of NTB-A also induces phosphorylation of NTB-A and the association of SAP (SLAM-associated protein), the protein absent in X-linked lymphoproliferative disease. T helper cells differentiated by cross-linking NTB-A and CD3 developed predominantly into Th1 cells not Th2 cells. In vivo blocking of NTB-A interactions with its ligands by using soluble NTB-A-Fc fusion protein inhibits B cell isotype switching to IgG2a and IgG3, commonly induced by Th1-type cytokines. Most important, treatment of mice with NTB-A-Fc delays the onset of antigen-induced experimental allergic encephalomyelitis in myelin basic protein-T cell receptor transgenic mice, suggesting a role in T cell-mediated autoimmune disease. Regulation of interferon-γ secretion, and not interleukin-4 in vitro, as well as inhibition of Th1 cell-induced isotype switching and attenuation of experimental allergic encephalomyelitis indicate that NTB-A is important for Th1 responses. The observation that cross-linking of NTB-A induces T cell activation, expansion, and Th1-type cytokine production suggests NTB-A is a novel co-stimulatory receptor. The identification of NTB-A as a regulator of T cell response paves the way to provide novel therapeutic approaches for modulation of the immune response. The CD28 co-stimulatory pathway is well established for T cell activation; however, results from CD28 -/- mice suggest the existence of additional co-stimulatory pathways. Here we report the further characterization of a new member of the CD2 superfamily, NTB-A, important in T cell co-stimulation. NTB-A is expressed on T cells, and its expression is up-regulated on activated cells. Triggering of NTB-A with monoclonal antibodies in the absence of CD28 signals leads to T cell proliferation and interferon-γ secretion but not interleukin-4. Cross-linking of NTB-A also induces phosphorylation of NTB-A and the association of SAP (SLAM-associated protein), the protein absent in X-linked lymphoproliferative disease. T helper cells differentiated by cross-linking NTB-A and CD3 developed predominantly into Th1 cells not Th2 cells. In vivo blocking of NTB-A interactions with its ligands by using soluble NTB-A-Fc fusion protein inhibits B cell isotype switching to IgG2a and IgG3, commonly induced by Th1-type cytokines. Most important, treatment of mice with NTB-A-Fc delays the onset of antigen-induced experimental allergic encephalomyelitis in myelin basic protein-T cell receptor transgenic mice, suggesting a role in T cell-mediated autoimmune disease. Regulation of interferon-γ secretion, and not interleukin-4 in vitro, as well as inhibition of Th1 cell-induced isotype switching and attenuation of experimental allergic encephalomyelitis indicate that NTB-A is important for Th1 responses. The observation that cross-linking of NTB-A induces T cell activation, expansion, and Th1-type cytokine production suggests NTB-A is a novel co-stimulatory receptor. The identification of NTB-A as a regulator of T cell response paves the way to provide novel therapeutic approaches for modulation of the immune response. T cell activation requires two distinct signals as follows: the first signal through the T cell receptor (TCR), 1The abbreviations used are: TCR, T cell receptor; EAE, experimental allergic encephalomyelitis; MBP, myelin basic protein; IFN, interferon; IL, interleukin; XLP, X-linked lymphoproliferative; NP-KLH, nitrophenol-conjugated keyhole limpet hemocyanin; ELISA, enzyme-linked immunosorbent assay; m, murine; FACS, fluorescence-activated cell sorter; SAP, SLAM-associated protein. which maintains antigen specificity, and a second co-stimulatory signal, from another receptor such as CD28 (1Carreno B.M. Collins M. Annu. Rev. Immunol. 2002; 20: 29-53Google Scholar). Interaction of CD28 with its ligands B7.1 (CD80) and B7.2 (CD86) expressed on activated antigen-presenting cells is thought to provide T cells with the necessary signals for proper activation and development into effector T cells. Although CD28 is considered a primary receptor for the second signal, CD28-deficient mice can still mount an antigen-specific T cell response, suggesting the existence of other co-stimulatory molecules (2Lucas P.J. Negishi I. Nakayama K. Fields L.E. Loh D.Y. J. Immunol. 1995; 154: 5757-5768Google Scholar, 3Shahinian A. Pfeffer K. Lee K.P. Kundig T.M. Kishihara K. Wakeham A. Kawai K. Ohashi P.S. Thompson C.B. Mak T.W. Science. 1993; 261: 609-612Google Scholar). For instance, CD8+ cytotoxic T cells are still functional in CD28-deficient mice, and these mice also exhibit delayed type hypersensitivity in response to viral infection. Several other receptors in the CD28 family have been subsequently cloned in recent years. ICOS is a positive regulatory receptor in the family, which can provide co-stimulatory signals to T cells, although ICOS does not activate naive T cells and is thought to be involved primarily in T helper cell development (4Tafuri A. Shahinian A. Bladt F. Yoshinaga S.K. Jordana M. Wakeham A. Boucher L.M. Bouchard D. Chan V.S. Duncan G. Odermatt B. Ho A. Itie A. Horan T. Whoriskey J.S. Pawson T. Penninger J.M. Ohashi P.S. Mak T.W. Nature. 2001; 409: 105-109Google Scholar, 5Sharpe A.H. Freeman G.J. Nat. Rev. Immunol. 2002; 2: 116-126Google Scholar, 6Dong C. Juedes A.E. Temann U.A. Shresta S. Allison J.P. Ruddle N.H. Flavell R.A. Nature. 2001; 409: 97-101Google Scholar). The current data in the field indicate that there may be other positive regulatory receptors yet to be identified. In addition to CD28 family members, other receptors such as those belonging to the CD2 superfamily can also provide the second signal for co-stimulation. Included in the CD2 super-family are CD2 (LFA-2), CD48, CD58 (LFA-3), CD84, CD150 (SLAM), and CD244 (2B4) (7Seed B. Aruffo A. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 3365-3369Google Scholar, 8Staunton D.E. Thorley-Lawson D.A. EMBO J. 1987; 6: 3695-3701Google Scholar, 9de la Fuente M.A. Pizcueta P. Nadal M. Bosch J. Engel P. Blood. 1997; 90: 2398-2405Google Scholar, 10Cocks B.G. Chang C.C. Carballido J.M. Yssel H. de Vries J.E. Aversa G. Nature. 1995; 376: 260-263Google Scholar, 11Aversa G. Chang C.C. Carballido J.M. Cocks B.G. de Vries J.E. J. Immunol. 1997; 158: 4036-4044Google Scholar, 12Aversa G. Carballido J. Punnonen J. Chang C.C. Hauser T. Cocks B.G. De Vries J.E. Immunol. Cell Biol. 1997; 75: 202-205Google Scholar, 13Davis S.J. van der Merwe P.A. Immunol. Today. 1996; 17: 177-187Google Scholar, 14Valiante N.M. Trinchieri G. J. Exp. Med. 1993; 178: 1397-1406Google Scholar). The CD2 family of receptors and ligands are expressed throughout the immune system and play an important role in immune responses. Although CD2 can act as a co-stimulatory molecule on T cells, CD2-deficient mice exhibit normal T cell responses, suggesting a redundant role for CD2 in T cell co-stimulation (15Killeen N. Stuart S.G. Littman D.R. EMBO J. 1992; 11: 4329-4336Google Scholar). The importance of CD2 becomes much clearer when analyzed in the context of CD28 deficiency. CD2-/-/CD28-/- double knockout mice exhibit a much more severe T cell activation defect when compared with CD28-deficient mice, underscoring the important contribution of CD2 in T cell co-stimulation (16Green J.M. Karpitskiy V. Kimzey S.L. Shaw A.S. J. Immunol. 2000; 164: 3591-3595Google Scholar). Still, in the absence of both of these receptors, T cells can be activated under certain circumstances. Clearly, other molecules in addition to CD28 and CD2 are involved in early T cell activation. Another CD2 superfamily member, CD150, can also provide a co-stimulatory signal to T cells (10Cocks B.G. Chang C.C. Carballido J.M. Yssel H. de Vries J.E. Aversa G. Nature. 1995; 376: 260-263Google Scholar, 11Aversa G. Chang C.C. Carballido J.M. Cocks B.G. de Vries J.E. J. Immunol. 1997; 158: 4036-4044Google Scholar, 12Aversa G. Carballido J. Punnonen J. Chang C.C. Hauser T. Cocks B.G. De Vries J.E. Immunol. Cell Biol. 1997; 75: 202-205Google Scholar). CD150 is included in a subfamily of receptors that are characterized by extracellular immunoglobulin-like repeats and intracellular binding sites for SAP (SH2D1A), the protein associated with X-linked lymphoproliferative (XLP) disease (9de la Fuente M.A. Pizcueta P. Nadal M. Bosch J. Engel P. Blood. 1997; 90: 2398-2405Google Scholar, 17Mathew P.A. Garni-Wagner B.A. Land K. Takashima A. Stoneman E. Bennett M. Kumar V. J. Immunol. 1993; 151: 5328-5337Google Scholar, 18Sandrin M.S. Gumley T.P. Henning M.M. Vaughan H.A. Gonez L.J. Trapani J.A. McKenzie I.F. J. Immunol. 1992; 149: 1636-1641Google Scholar, 19Latour S. Gish G. Helgason C.D. Humphries R.K. Pawson T. Veillette A. Nat. Immun. 2001; 2: 681-690Google Scholar, 20Morra M. Howie D. Grande M.S. Sayos J. Wang N. Wu C. Engel P. Terhorst C. Annu. Rev. Immunol. 2001; 19: 657-682Google Scholar, 21Nichols K.E. Koretzky G.A. June C.H. Nat. Immun. 2001; 2: 665-666Google Scholar, 22Sayos J. Wu C. Morra M. Wang N. Zhang X. Allen D. van Schaik S. Notarangelo L. Geha R. Roncarolo M.G. Oettgen H. De Vries J.E. Aversa G. Terhorst C. Nature. 1998; 395: 462-469Google Scholar, 23Veillette A. Science's STKE. 2002; http://www.stke.org/cgi/content/full/OC_sigtrans;2002/E8Google Scholar). The basis for the lymphoproliferation defect seen in XLP patients has been attributed to defects in T cell and NK cell function, although there is no clear understanding of how SAP deficiency contributes to this disease at this time. A better understanding of SAP will come from studying the function of receptors associated with SAP. The most well studied SAP-associated receptor is CD150. CD150 can provide a co-stimulatory signal to T cells, inducing proliferation and IFN-γ expression. There is also evidence for CD150 involvement in Th1 cell responses, suggesting the importance of this molecule in inflammation and autoimmune disease (10Cocks B.G. Chang C.C. Carballido J.M. Yssel H. de Vries J.E. Aversa G. Nature. 1995; 376: 260-263Google Scholar, 11Aversa G. Chang C.C. Carballido J.M. Cocks B.G. de Vries J.E. J. Immunol. 1997; 158: 4036-4044Google Scholar, 12Aversa G. Carballido J. Punnonen J. Chang C.C. Hauser T. Cocks B.G. De Vries J.E. Immunol. Cell Biol. 1997; 75: 202-205Google Scholar, 24Castro A.G. Hauser T.M. Cocks B.G. Abrams J. Zurawski S. Churakova T. Zonin F. Robinson D. Tangye S.G. Aversa G. Nichols K.E. de Vries J.E. Lanier L.L. O'Garra A. J. Immunol. 1999; 163: 5860-5870Google Scholar). Recently, CD84, another CD2/CD150 family member, was found to possess co-stimulatory activity on T cells (25Martin M. Romero X. de la Fuente M.A. Tovar V. Zapater N. Esplugues E. Pizcueta P. Bosch J. Engel P. J. Immunol. 2001; 167: 3668-3676Google Scholar). Clearly this family of receptors plays an important role in co-stimulation and lymphocyte regulation, based on its role in lymphocyte activation and association with XLP. Given the important role of CD2 superfamily members in the immune system, we sought to identify novel members of this family. In our search, we have identified a protein referred to as TCOM (for T cell co-stimulatory molecule), which shares high amino acid similarity with CD2/CD150 family members. TCOM is identical to NTB-A and SF2000, cloned by Bottino et al. (26Bottino C. Falco M. Parolini S. Marcenaro E. Augugliaro R. Sivori S. Landi E. Biassoni R. Notarangelo L.D. Moretta L. Moretta A. J. Exp. Med. 2001; 194: 235-246Google Scholar) and Fraser et al. (27Fraser C.C. Howie D. Morra M. Qiu Y. Murphy C. Shen Q. Gutierrez-Ramos J.C. Coyle A. Kingsbury G.A. Terhorst C. Immunogenetics. 2002; 53: 843-850Google Scholar), respectively; therefore, we refer to our molecule as NTB-A. We show that NTB-A enhances the Th1 cell phenotype, and cross-linking of NTB-A induces phosphorylation of NTB-A and SAP association. An NTB-A-Fc fusion protein inhibits Th1 cytokine-induced isotype switching and attenuates EAE, indicating that NTB-A is important for in vivo T cell responses including Th1 cell-mediated autoimmune disease. Generation of Anti-NTB-A Monoclonal Antibody and NTB-A Fusion Proteins—NTB-A-His consists of the first 226 amino acids of NTB-A, which compose the extracellular region of the protein, fused to six His residues. NTB-A-Fc consists of the first 226 amino acids of NTB-A fused to the Fc portion of murine IgG1. For monoclonal antibody generation, mice were immunized with NTB-A-His protein. Several hybridomas were screened for recognition of NTB-A. Specificity of the antibody was confirmed by ELISA binding to NTB-A-Fc as well as by binding to 293 cells transfected with full-length NTB-A. Homotypic Binding of NTB-A—Human NTB-A-His or 19A-His was coated onto Nunc Maxisorb plates at 2 μg/ml and washed off. After blocking in 10% fetal calf serum, NTB-A-Fc or 19A-Fc was titrated on the plate starting at 20 μg/ml. Binding of the Fc fusion proteins was detected using Fc-specific goat anti-human IgG (Kirkegaard & Perry Laboratories). For detection of cell surface binding, full-length human NTB-A or human 19A was transfected into 293T cells. Forty-eight hours after transfection, the cells were washed in PBS and incubated with 5 μg/ml NTB-A-Fc plus 5% normal goat serum for 1 h. Cells were then washed and fixed in 4% paraformaldehyde for 15 min. Biotinylated goat anti-human Fc antibody (Jackson ImmunoResearch) was added to the cells at 1:400 dilution and incubated for 30 min, followed by washing in PBS. Cy3-conjugated streptavidin (Jackson ImmunoResearch) was used at 1:400 dilution for detection. Cell staining was analyzed using a ×40 objective and SPOT camera. NTB-A Co-stimulation Assay—Costar 96-well U-bottom plates were coated overnight with 0.1 μg/ml anti-CD3 antibody (Pharmingen), 1 μg/ml anti-CD28 (Pharmingen), 1 μg/ml IgG1, or 1 μg/ml anti-NTB-A antibody generated against the extracellular domain of human NTB-A. Human CD4+ T cells were purified from whole blood by negative selection using the CD4 T cell MACs separation kit (Miltenyi Biotec). The resulting CD4 T cell population was 95% pure. Naive CD4 T cells were separated by positive selection of CD45RA+ cells. The remaining CD4 T cells were saved as the memory CD4 T cell fraction. Naive and memory T cells were plated at 0.5 × 106 T cells per well in RPMI. On the 3rd day, supernatant was removed for cytokine analysis (by ELISA or Luminex, Upstate Biotechnology, Inc.), and the cells were pulsed for 16 h with [3H]thymidine. For Jurkat cell stimulation, Costar plates were coated as described above and 0.5 × 106 Jurkat cells were stimulated for 24 h. At 24 h, the cells were analyzed by FACS for expression of the activation marker, CD69 (Pharmingen). Co-stimulation of mouse T cells was carried out by coating a plate with 0.1 μg/ml anti-CD3 antibody (Pharmingen), 1 μg/ml anti-CD28 (Pharmingen), 1 μg/ml IgG1, or 1 μg/ml anti-NTB-A that cross-reacted to mouse form. CD4 T cells were isolated from the spleen of a C57/B6 mouse by negative selection using the CD4 T cell isolation kit from Miltenyi Biotec. 0.5 × 106 CD4 T cells were activated for 3 days, and on the 3rd day, supernatant was removed for cytokine analysis, and the cells were pulsed with [3H]thymidine. T Helper Cell Differentiation—Naive human CD4 T cells were isolated from peripheral blood by negative selection and Automacs sorting. Cells were either stimulated with anti-CD3 coated at 2 μg/ml plus anti-CD28 coated at 1 μg/ml (Pharmingen) or with anti-CD3 coated at 2 μg/ml plus anti-NTB-A coated at 1 μg/ml. For Th0 conditions, cells were cultured with IL-2, anti-INF-γ, anti-IL-12, and anti-IL-4 antibody. For Th1 conditions, cells were cultured with IL-2, IL-12, and anti-IL-4 in RPMI. For Th2 conditions, cells were cultured with IL-2, IL-4, anti-IL-12, and anti-INF-γ in RPMI. Concentrations of antibodies and cytokines were as follows: IL-2, 2 ng/ml; IL-4, ng/ml; IL-12, 5 ng/ml; anti-IFN-γ,1 μg/ml; anti-IL-12, 1 μg/ml; anti-IL-4, 1 μg/ml (R & D Systems). Three days after the initial activation, the T cells were washed and left to rest with only IL-2 in the media. After 7 days, cells were re-stimulated for 24 h with anti-CD3 only coated at 2 μg/ml in 96-well plates. The cell number per well was 0.5 × 106. Supernatants were then removed for cytokine analysis by Luminex (Upstate Biotechnology, Inc.). NTB-A Immunoprecipitation and Western Blots—107 Jurkat T cells were treated with pervanadate for 15 min at 37 °C and lysed in RIPA buffer with 1% Nonidet P-40 and protease and phosphatase inhibitors. Lysates were pre-cleared with 10 μg of mIgG1 and protein G; then NTB-A was immunoprecipitated using 10 μg of a mixture of IgG1 monoclonal antibodies raised against the extracellular domain of NTB-A. LAIR was also immunoprecipitated using the monoclonal antibody DX26 (Pharmingen). Beads were washed with RIPA buffer with 0.1% Nonidet P-40. Association of SAP was detected by Western blot (Santa Cruz Biotechnology), along with SHP-1 and SHP-2 (Upstate Biotechnology, Inc.). Human T cells were isolated by AutoMACS sorting of CD4 positive cells from blood. 4 × 107 T cells were used per condition. Cells were first incubated with 8 μg of anti-CD3 (Pharmingen) and 8 μg of anti-NTB-A antibodies. Cells were then washed, left to rest in PBS, and then stimulated with 4 μg of goat anti-mouse Fc antibody (Jackson ImmunoResearch) for 2 min to cross-link the receptors. Immunoprecipitation and Western blot were carried out as described above. In Vivo T-dependent B Cell Responses—C57/B6 mice were immunized intraperitoneally with 100 μg of NP(23)-KLH emulsified 1:1 in Complete Freund's Adjuvant. Mice were injected intraperitoneally daily with 100 μg of mNTB-A-Fc fusion protein or mIgG1 as a control. After 10 days, serum was obtained, and NP-specific immunoglobulins were measured by ELISA. Nunc maxisorb plates were coated overnight with NP(23)-bovine serum albumin at 10 μg/ml. After washing and blocking with 10% fetal calf serum, dilutions of serum were then added to the plate for 2 h. After washing the serum from the plate, biotinylated anti-isotype-specific antibodies were added to the plate. The antibodies used included biotinylated anti-IgM, anti-IgG1, anti-IgG2a, and anti-IgG3 (Pharmingen). Detection was carried out with streptavidin-horse-radish peroxidase and TMB (3,3′,5,5′-tetramethylbenzidine) substrate. Induction of EAE—Two groups of MBP TCR transgenic mice were immunized with 10 μg of Ac1–11/Complete Freund's Adjuvant subcutaneously on day 0. One group of 8 mice was injected intraperitoneally with 100 μg of mNTB-A-Fc daily, and the other group of 9 mice was injected with 100 μg of mIgG1 daily. On days 1 and 2, mice were given 200 ng of pertussis toxin intraperitoneally. Mice were evaluated daily for signs of disease. Disease scores were as follows: 0, no disease; 1, tail droop or hind limb weakness; 2, paralysis involving one limb; 3, paralysis involving two limbs; 4, death. Characterization of NTB-A Expression—In a search for novel CD2 family members, we identified a protein referred to as TCOM (for T cell co-stimulatory molecule). Shortly thereafter, this novel protein was described by Bottino et al. (26Bottino C. Falco M. Parolini S. Marcenaro E. Augugliaro R. Sivori S. Landi E. Biassoni R. Notarangelo L.D. Moretta L. Moretta A. J. Exp. Med. 2001; 194: 235-246Google Scholar) and Fraser et al. (27Fraser C.C. Howie D. Morra M. Qiu Y. Murphy C. Shen Q. Gutierrez-Ramos J.C. Coyle A. Kingsbury G.A. Terhorst C. Immunogenetics. 2002; 53: 843-850Google Scholar) as NTB-A and SF2000, respectively. NTB-A has been show to play a role in NK cell activity; however, a physiological role in T cells has yet to be examined. Like other family members NTB-A has two predicted immunoglobulin-like repeats in the extracellular domain as well as three tyrosine motifs in the intracellular domain. Two of these domains bear the TXYXX(V/I) motif, which is commonly found in the cytoplasmic domains of CD2 family members, and act as a SAP-binding domain in CD150. The murine form of NTB-A was also cloned based on sequence homology with the human form and is identical to Ly108 (28Peck S.R. Ruley H.E. Immunogenetics. 2000; 52: 63-72Google Scholar). The amino acid similarity between the mouse and human NTB-A is 46%, and the two TXYXX(V/I) motifs are highly conserved between species. By using both the human and murine form of NTB-A, our studies define a novel role for NTB-A in T cell activation. Expression of human NTB-A on cell lines was evaluated using a monoclonal antibody raised against the extracellular domain of NTB-A. Our monoclonal antibody detected NTB-A expression on several T cells lines including Jurkat, MOLT3, and MOLT4 (Fig. 1A). Some, but not all, B cell lines were also positive for NTB-A expression, including Raji, Ramos, and Daudi cell lines. No cell surface expression of NTB-A was detected on other cell lines including a breast adenocarcinoma (MDA231), a monocyte line (THP-1), and a kidney epithelial cell line (293). In addition to cells lines, cell surface expression of NTB-A on human peripheral T cells was also detected (Fig. 1B). Stimulation of T cells with anti-CD3 and anti-CD28 antibodies led to up-regulation of NTB-A surface expression at 24 h (data not shown), with even higher expression seen at 48 h post-activation (Fig. 1B). Human peripheral B cells also express NTB-A and NTB-A expression increased upon stimulation with anti-CD40 and IL-4 (Fig. 1C). High expression of NTB-A on CD4+CD45RO+ memory T cells was also observed (Fig. 1D). Our data indicate that NTB-A is expressed on both naive and activated lymphocytes. Homotypic Binding of NTB-A—Homotypic interaction has been demonstrated for CD150 (29Mavaddat N. Mason D.W. Atkinson P.D. Evans E.J. Gilbert R.J. Stuart D.I. Fennelly J.A. Barclay A.N. Davis S.J. Brown M.H. J. Biol. Chem. 2000; 275: 28100-28109Google Scholar, 30Punnonen J. Cocks B.G. Carballido J.M. Bennett B. Peterson D. Aversa G. de Vries J.E. J. Exp. Med. 1997; 185: 993-1004Google Scholar), and it is suggested that other closely related family members (CD84 and Ly9) may also exhibit homotypic binding (25Martin M. Romero X. de la Fuente M.A. Tovar V. Zapater N. Esplugues E. Pizcueta P. Bosch J. Engel P. J. Immunol. 2001; 167: 3668-3676Google Scholar). To determine whether NTB-A can also act as a self-ligand, we looked for NTB-A:NTB-A association in a number of assays. First, the ability of NTB-A to bind to cell-surface-expressed NTB-A was analyzed by transfection of 293T cells with the full-length form of human NTB-A or human 19A, a member of the CD150 family of receptors (Fig. 2A). The transfected cells were incubated with an NTB-A-Fc fusion protein. Binding of NTB-A-Fc to the cell surface was observed only in 293T cells transfected with NTB-A and not in cells transfected with 19A, suggesting that NTB-A can bind to cell-surface-expressed NTB-A and that this interaction is specific. No significant binding of NTB-A-Fc was found to cells transfected with other CD2 family members, including CD150, BLAME, or 2B4 (data not shown). In an ELISA, we find that NTB-A-Fc can specifically associate with NTB-A-His protein coated on a plate (Fig. 2B). 19A-Fc does not associate with NTBA-His bound to a plate, and NTBA-Fc does not bind to 19A-His bound to a plate. BIAcore assays were also carried out to determine the affinity of His-tagged NTB-A protein for itself. Homotypic binding of His-NTB-A was found to be a very weak interaction (Kd >100 μm) (data not shown). Most interesting, the CD150-CD150 interaction has been reported to be very weak as well at 200 μm (29Mavaddat N. Mason D.W. Atkinson P.D. Evans E.J. Gilbert R.J. Stuart D.I. Fennelly J.A. Barclay A.N. Davis S.J. Brown M.H. J. Biol. Chem. 2000; 275: 28100-28109Google Scholar). Our data indicate that NTB-A can bind to itself very weakly, and because the interaction is to weak, we cannot rule out the existence of other ligands for NTB-A. Co-stimulation of T Cells by NTB-A—To determine whether NTB-A is capable of providing a co-stimulatory signal to T cells, we isolated naive and memory human peripheral CD4 T cells and stimulated them with anti-NTB-A antibody and a suboptimal dose of anti-CD3 antibody. Anti-CD3 treatment alone only induced minimal proliferation in the naive T cells, whereas higher proliferation was seen in the memory T cells because these cells do not require co-stimulation for activation (Fig. 3A). Treatment with anti-NTB-A or anti-CD28 antibody alone did not induce proliferation or IFN-γ secretion in naive or memory T cells. The combination of anti-NTB-A with a suboptimal dose of anti-CD3 in the absence of anti-CD28 antibody led to a substantial level of proliferation that was higher than anti-CD3 alone and slightly lower than the proliferation induced with anti-CD3 plus anti-CD28 treatment. Increase in IFN-γ expression corresponded to the increase in proliferation induced by anti-CD3 plus anti-NTB-A treatment. The effect of anti-NTB was more dramatic in the naive T cells, presumably due to the fact that memory cells respond very strongly to anti-CD3 treatment alone. IFN-γ secretion induced by anti-CD3 treatment was not augmented by the addition of anti-NTB-A. The ability of NTB-A to activate Jurkat T cells was assessed by culturing Jurkat cells on plates coated with anti-CD3, anti-CD28, anti-NTB-A, or a combination of these antibodies. Twenty four hours after activation, Jurkat cells that were stimulated with anti-CD3 and anti-NTB-A up-regulated expression of the activation marker CD69 (Fig. 3B). Co-stimulation of murine CD4 T cells was also carried out by using antibody that cross-reacts to the murine form of NTB-A (Fig. 3C). We find similar results as with human T cells, including induced proliferation and IFN-γ expression with NTB-A co-stimulation, indicating the murine and human forms of NTB-A function similarly in T cells. Thus, our results suggest that NTB-A is capable of co-stimulating T cells to a level similar to CD28. NTB-A Promotes Th1 Cell Phenotype—Given that NTB-A cross-linking can induce proliferation and INF-γ secretion, we next sought to determine whether NTB-A influences Th1 or Th2 cell differentiation. Purified human CD4 T cells were stimulated with anti-CD3 and anti-NTB-A under Th1-inducing conditions or under Th2-inducing conditions. This stimulation was independent of a CD28 signal because no anti-CD28 agonistic reagents were used where anti-NTB-A antibody was used. For comparison, cells were also stimulated with anti-CD3 and anti-CD28 under appropriate Th1 or Th2 conditions. After 3 days of stimulation, the cells were washed and rested for 4 days with IL-2. After that time, all cells were re-stimulated with anti-CD3 only. Th1 cells stimulated with anti-NTB-A secreted more IFN-γ compared with Th1 cells stimulated with anti-CD28, and Th0 cells also had slightly higher levels of INF-γ (Fig. 4). Levels of IL-4 and IL-5, Th2-type cytokines, were comparable in all samples, indicating that NTB-A had no effect distinguishable from CD28 on the differentiation of Th2 cells. These results suggest that NTB-A helps to enhance the Th1 cell phenotype by inducing more Th1-type cytokine expression. Most interesting, effects of NTB-A in enhancing differentiation of T cells to Th1 were completely independent of CD28 signals. Phosphorylation of NTB-A and SAP Association—We have demonstrated that cross-linking of CD3 and NTB-A on T cells leads to proliferation and IFN-γ expression. We next sought to analyze the molecular events associated with CD3 and NTB-A cross-linking. A hallmark of CD2 family members is the presence of two to four SAP-binding motifs (TXYXX(V/I)) in the intracellular domain; SAP can bind to the both the phosphorylated and non-phosphorylated forms of CD150, although SAP has a higher affinity for the phosphorylated form (31Li S.C. Gish G. Yang D. Coffey A.J. Forman-Kay J.D. Ernberg I. Kay L.E. Pawson T. Curr. Biol. 1999; 9: 1355-1362Google Scholar, 32Poy F. Yaffe M.B. Sayos J. Saxena K. Morra M. Sumegi J. Cantley L.C. Terhorst C. Eck M.J. Mol. Cell. 1999; 4: 555-561Google Scholar). In addition to binding SAP, CD150 has also been reported to bind SHP-2 (22Sayos J. Wu C. Morra M. Wang N. Zhang X. Allen D. van Schaik S. Notarangelo L. Geha R. Roncarolo M.G. Oettgen H. De Vries J.E. Aversa G. Terhorst C. Nature. 1998; 395: 462-469Google Scholar, 24Castro A.G. Hauser T.M. Cocks B.G. Abrams J. Zurawski S. Churakova T. Zonin F. Robinson D. Tangye S.G. Aversa G. Nichols K.E. de Vries J.E. Lanier L.L. O'Garra A. J. Immunol. 1999; 163: 5860-5870Google Scholar, 33Howie D. Simarro M. Sayos J. Guirado M. Sancho J. Terhorst C. Blood. 2002; 99: 957-965Google Scholar, 34Lewis J. Eiben L.J. Nelson D.L. Cohen J.I. Nichols K.E. Ochs H.D. Notarangelo L.D. Duckett C.S. Clin. Immunol. 2001; 100: 15-23Google Scholar, 35Shlapatska L.M. Mikhalap S.V. Berdova A.G. Zelensky O.M. Yun T.J. Nichols K.E. Clark E.A. Sidorenko S.P. J. Immunol. 2001; 166: 5480-5487Google Scholar). Both human and mouse NTB-A have two SAP-binding motifs in the intracellular domain. To determine whether NTB-A also associates with these CD150-associated proteins, Jurkat T cells were treated with pervanadate to induce phosphorylation of NTB-A. NTB-A was immunoprecipitated from cell lysates, and the association of SAP, SHP-1, and SHP-2 was determined by Western blot. SAP association with NTB-A was not detected in untreated cells. However, SAP did associate with NTB-A in pervanadate-treated cells, suggesting a phosphorylation-dependent association of SAP with NTB-A. SHP-1 and SHP-2, although expressed in Jurkat cells, were not found associated with NTB-A in either untreated or in pervanadate-treated cells (Fig. 5A). This lack of association with SHP-1 or SHP-2 suggests a signaling mechanism distinct from CD150 in T cells. The inhibitory receptor LAIR does associate with SHP-1 and SHP-2 in pervanadate-treated Jurkat cells, ruling out the possibility that SHP-1 and SHP-2 are not capable of being immunoprecipitated in our system (Fig. 5A). To study the signaling events following co-stimulation with NTB-A in T cells, we next analyzed the phosphorylation of NTB-A induced by cross-linking CD3 and NTB-A in freshly isolated human T cells (Fig. 5B). NTB-A is phosphorylated upon cross-linking of CD3 and NTB-A in human T cells as shown by immunoprecipitation of NTB-A and Western blotting with an anti-phosphotyrosine antibody. Most interesting, SAP does not associate with NTB-A in untreated human T cells, but SAP does associate with co-stimulation-induced phosphorylated NTB-A. Similar to Jurkat cells, SHP1 and SHP2 do not associate with NTB-A in human T cells, although both proteins are expressed. These findings suggest a mechanism by which cross-linking of NTB-A induces NTB-A phosphorylation and SAP association in T cells. It is possible that these signaling events then lead to proliferation and IFN-γ secretion. NTB-A-Fc Inhibits Th1 Cell-induced Isotype Switching—Our data thus far suggest that NTB-A may act to induce Th1 T cell responses. We next analyzed whether NTB-A plays a role in T-dependent B cell responses. To block a possible ligand interaction with NTB-A in vivo, we used a fusion protein consisting of the extracellular portion of murine NTB-A fused to an Fc domain (mNTB-A-Fc). C57/B6 mice were immunized with NP(23)-KLH and treated daily with mNTB-A-Fc or mIgG1 as a control. Ten days after immunization, serum was analyzed for the expression of NP-specific immunoglobulin isotypes (Fig. 6). Generally, Th2-type cytokines such as IL-4 induce the increase of switching to the IgG1 isotype and suppress the switching to IgG2a and IgG3. Th1-type cytokines such as IFN-γ suppress switching to IgG1 and increase switching to IgG2a and IgG3. Mice treated with mNTB-A-Fc demonstrated decreased levels of IgG2a and IgG3 and increased levels of IgG1, which is indicative of the presence of Th2-type cytokines such as IL-4. These results suggest that mNTB-A-Fc suppressed B cell isotype switching to IFN-γ-induced isotypes, possibly by inhibiting the activation of Th1 cells during immunization. Although we have demonstrated a homotypic interaction for NTB-A, the weak binding affinity suggests that there may be an additional ligand for NTB-A. The addition of the fusion protein in vivo may act to block this ligand in vivo and prevent the activation and development of Th1 cells. NTB-A-Fc Delays Onset of EAE—The increase in IFN-γ expression induced by NTB-A suggests that NTB-A may act to skew T cell responses to a Th1 response. To determine the effect of NTB-A in Th1-mediated autoimmune disease, we analyzed the effect of an NTB-A-Fc fusion protein in a murine model for multiple sclerosis, EAE. MBP TCR transgenic mice were immunized with Ac1-11 (NH2-terminal peptide from MBP) and treated daily with either NTB-A-Fc or mIgG1 control protein. Disease progression was monitored each day, and the onset of disease began at day 5 in the control mice (Fig. 7). Mice treated with mNTB-A-Fc showed a delayed onset of disease at day 8, and 20% of the mice never developed disease. The ability of mNTB-A-Fc to delay the onset of EAE in MBP TCR transgenic mice suggests that the fusion protein may dampen effector functions of T cells because EAE is a predominantly a Th1 cell-mediated disease. Based on the ability of NTB-A-Fc to decrease Th1 cytokine-induced isotype switching, we hypothesize that NTB-A-Fc delays EAE development by inhibition of the generation of Th1 responses. Recognition of antigen by TCR is not sufficient for proper T cell activation. Additional co-stimulatory molecules are required, and this provides the basis for the "two signal model" of lymphocyte activation. Members of the CD28/B7 family of co-stimulatory molecules are generally considered to provide the major pathway for co-stimulation. However, data including analysis of CD28-deficient T cells suggest that other co-stimulatory molecules exist that can provide the second signal for T cells. The CD2 superfamily has been implicated in T cell co-stimulation. Included in this superfamily is the CD150 subfamily of receptors, which associate with SAP. Human patients who lack expression of SAP have major defects in cell-mediated immunity and succumb to massive lymphocyte infiltration to multiple organs, a syndrome called X-linked proliferative disease (XLP). Thus, the CD2/CD150 family members may play a significant part in cellular immunity. We have identified a new T cell function for the CD2 family member, NTB-A. We find that NTB-A can co-stimulate T cells to proliferate and secrete IFN-γ. NTB-A cross-linking leads to phosphorylation of the receptor and association of SAP, but not SHP-1 or SHP-2, in human T cells. Our data suggest that NTB-A plays a role in Th1 cell responses because Th1 cells stimulated with NTB-A secrete more Th1- and not Th2-type cytokines. An NTB-A-Fc fusion protein inhibits IFN-γ-induced isotype switching. NTB-A-Fc also delays development and severity of a mouse model for multiple sclerosis. NTB-A contains two ITSM (TXYXX(V/I)) motifs in its intracellular domain. These motifs have been shown to bind SAP in other receptors including CD150 and 2B4. Co-stimulation induced by cross-linking of CD3 and NTB-A leads to early phosphorylation of NTB-A and association of SAP but not SHP-1 or SHP-2. SAP is composed largely of an SH2 domain, and it was previously thought to block interaction of SH2 domain-containing proteins such as SHP-2 (22Sayos J. Wu C. Morra M. Wang N. Zhang X. Allen D. van Schaik S. Notarangelo L. Geha R. Roncarolo M.G. Oettgen H. De Vries J.E. Aversa G. Terhorst C. Nature. 1998; 395: 462-469Google Scholar). Unlike CD150, NTB-A does not associate with SHP-2. Lack of association of NTB-A with SHP-2 in T cells suggests that the regulation of NTB-A signaling and IFN-γ production is distinct from CD150. Recently, it has been suggested that SAP may act to recruit other molecules such as FynT (19Latour S. Gish G. Helgason C.D. Humphries R.K. Pawson T. Veillette A. Nat. Immun. 2001; 2: 681-690Google Scholar), and this could help drive the specificity of the signal through different SAP-associated receptors. It will be interesting to determine whether FynT also associates with NTB-A. Given that XLP patients are deficient in SAP, understanding the mechanism by which SAP helps regulate signals through NTB-A may give us insight into the pathology of XLP disease. Our preliminary data suggest that T cells from XLP patients cannot be activated through NTB-A, suggesting an important role for SAP and NTB-A in T cell activation. We have demonstrated a critical role for NTB-A in T cell activation and disease by using an NTB-A-Fc fusion protein. NTB-A-Fc was able to suppress T-dependent B cell class switching to Th1 cytokine-induced isotypes, most likely by blocking interaction with an NTB-A ligand and interfering with activation of T cells to Th1-type cells. The contribution of NTB-A to early activation events versus later effector functions is not clear. We do find that naive cells can respond to NTB-A cross-linking and that initial cross-linking with NTB-A skews development to the Th1 cell type. Stimulation of Th1 or Th2 cell clones with NTB-A has little effect (data not shown), suggesting that NTB-A may play a greater role in the early phase of activation. When tested in a Th1-mediated autoimmune disease model, EAE, NTB-A-Fc delayed the onset of disease. These results further suggest that NTB-A may play an important role in T cell activation and effector functions. The block in disease by NTB-A-Fc is not complete, and it may be the case that NTB-A-Fc does not block NTB-A activity completely. As we have shown, NTB-A exhibits weak homotypic interaction, similar to CD150 and other family members. It is possible that NTB-A-Fc may provide a weak activation signal while blocking a stronger signal at the same time. Thus, disease is not completely blocked during EAE progression. Another possibility for the lack of a complete block in disease is that other molecules important for T cell functions are present. Other co-stimulatory molecules such as ICOS have been implicated in autoimmunity (36Sperling A.I. Bluestone J.A. Nat. Immun. 2001; 2: 573-574Google Scholar, 37Ozkaynak E. Gao W. Shemmeri N. Wang C. Gutierrez-Ramos J.C. Amaral J. Qin S. Rottman J.B. Coyle A.J. Hancock W.W. Nat. Immun. 2001; 2: 591-596Google Scholar, 38Rottman J.B. Smith T. Tonra J.R. Ganley K. Bloom T. Silva R. Pierce B. Gutierrez-Ramos J.C. Ozkaynak E. Coyle A.J. Nat. Immun. 2001; 2: 605-611Google Scholar, 39Gonzalo J.A. Tian J. Delaney T. Corcoran J. Rottman J.B. Lora J. Al-garawi A. Kroczek R. Gutierrez-Ramos J.C. Coyle A.J. Nat. Immun. 2001; 2: 597-604Google Scholar). How NTB-A regulates T cell function along with other co-stimulatory molecules will be an interesting focus. Our results suggest an important role for NTB-A in T cell-mediated functions, including T cell activation and T cell-mediated autoimmune disease. The identification of NTB-A as a new SAP-associated receptor may help to further understand the XLP disease process in human patients. Our findings implicate a role for NTB-A in T cell activation, suggesting a possible mechanism by which SAP-deficient T cells fail in their response to pathogens, including virally infected cells and tumor cells. The identification of NTB-A and new SAP-associated receptors, as well as understanding the signaling mechanisms involved, will help shed new light on XLP disease. Given its role as a co-stimulatory molecule, inducing proliferation and IFN-γ secretion by T cells, as well as the ability of the fusion protein to suppress Th1-induced isotype switching and EAE, NTB-A is a potential target for therapeutic intervention in Th1-type autoimmune disease as well as inflammation. We thank Betty Li for technical support; the Genentech Antibody Technology group, DNA sequencing lab, FACS lab, and protein purification group for their support; Andy Chan and lab members for discussion; Laura de Forge for cytokine analysis; Henry Lowman for Biacore measurements; Peter Gribling, Kathy Nguyen, and Yifan Zhang for in vivo work; Amy Shen and Zhonghua Lin for technical assistance; and Charles A. Janeway, Jr., for MBP-TCR transgenic mice.