Proc. Natl. Acad. Sci. USA Vol. 88, pp. 6575-6579, August 1991 Immunology

B-cell surface antigen B7 provides a costimulatory signal that induces T cells to proliferate and secrete interleukin 2 CLAUDE D. GIMMI*t, GORDON J. FREEMAN*, JOHN G. GRIBBEN*, KANJI SUGITA*, ARNOLD S. FREEDMAN*, CHIKAO MORIMOTO*, AND LEE M. NADLER* Departments of *Medicine and *Pathology, Division of Tumor Immunology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA 02115 Communicated by Baruj Benacerraf, April 15, 1991

ABSTRACT Occupancy of the T-cell receptor complex magnitude ofenhancement oflymphokine synthesis resulting does not appear to be a sufficient stimulus to induce a T-cell- from engagement of the CD28 pathway distinguishes it from mediated immune response. Increasing evidence suggests that other cell surface ligand pairs between T cells and antigen- cognate cell-cell interaction between an activated T cell and an presenting cells (APCs) such as LFA-1/ICAM-1 (19-23) and antigen-presenting cell may provide such a stimulus. A candi- CD2/LFA-3 (24-26). The observed increase in lymphokine date T-cell surface molecule for this costimulatory signal is the production by anti-CD28 crosslinking results initially from T-cell-restricted CD28 antigen. Following crosslinking with stabilization oflymphokine mRNAs (27, 28) and later from an anti-CD28 mAb, suboptimally stimulated CD28' T cells show increase in transcription (29). Only a subset ofT-cell-derived increased proliferation and markedly increased secretion of a lymphokines are induced by the CD28 pathway, including subset of lymphokines. Recently, the B-cell surface activation interleukin 2 (IL-2), interferon y, tumor necrosis factor a, and antigen B7 was shown to be a natural ligand for the CD28 granulocyte/macrophage colony-stimulating factor but not molecule, and both B7 and CD28 are members of the immu- IL-4 (27). This observation suggests that either alternative noglobulin superfamily. Here we report that B7-transfected signals or subpopulations of T cells may be necessary to CHO cells can induce suboptimally activated CD28+ T cells to induce IL-4 synthesis (1). proliferate and secrete high levels of interleukin 2. The re- The natural ligand for the CD28 molecule has recently been sponse is identical whether T cells are submitogenically stim- shown to be the B-cell activation antigen B7 (30). Heterotypic ulated with either phorbol myristate acetate or anti-CD3 to binding of cell lines transfected with CD28 and B7 and activate the T cells. This response is specific and can be totally specific inhibition of this binding by either anti-CD28 or abrogated with anti-B7 monoclonal antibody. As has previ- anti-B7 mAbs have convincingly demonstrated that B7 is a ously been observed for anti-CD28 monoclonal antibody, B7 CD28 ligand and this interaction can mediate T-B cell adhe- ligation induced secretion of interleukin 2 but not interleukin sion (30). The B7 molecule is a 44- to 54-kDa glycoprotein that 4. We have previously demonstrated that B7 expression is is also a member of the immunoglobulin superfamily (31). B7 restricted to activated B lymphocytes and interferon V-acti- expression is extremely restricted in that B7 is transiently vated monocytes. Since these two cellular populations are expressed on activated B cells and interferon y-treated mono- involved in antigen presentation as well as cognate interaction cytes (32-34). Since these two cellular populations are in- with T lymphocytes, B7 is likely to represent a central costim- volved in antigen presentation as well as cognate interaction ulatory signal that is capable of amplifying an immune re- with T lymphocytes, the interaction of B7 and CD28 is likely sponse. to represent a costimulatory signaling pathway. In the pres- ent report, we demonstrate that the B7 molecule is not only Although engagement of the T-cell receptor complex (TCR) an adhesion molecule for CD28 but also capable of upregu- by antigen in the context of proteins encoded by the major lating proliferation and markedly enhancing lymphokine pro- histocompatibility complex (MHC) is essential for the initial duction in suboptimally stimulated CD28+ T cells. stages of T-cell activation, it does not appear to be sufficient to induce all the events that accompany T-cell activation (1). MATERIALS AND METHODS Studies suggest that costimulation through additional T-cell surface molecules, independent of the TCR, leads to en- Cells. Human peripheral blood mononuclear cells were hanced proliferation and cytokine production (2-4). There- isolated from buffy coats obtained by leukopheresis of fore, signals that trigger these accessory molecules are likely healthy donors. After density gradient centrifugation the cells to be essential for the generation of an immune response and were further purified by depletion ofadherent cells on plastic. their dysregulation may be responsible for immune-mediated Residual B cells and monocytes were depleted by passage disease states (1, 5, 6). through nylon wool. The CD28+ subset of T cells was On the T cell, a candidate to receive this accessory cell enriched by separation from the reciprocal subset of CD11b+ contact signal is the 44-kDa homodimeric T-cell surface T cells (2, 12, 35), residual B cells, and monocytes by two protein CD28 (7-9). This T-cell-restricted molecule, which is treatments with complement lysis utilizing anti-3B8 (CD56), a member of the immunoglobulin superfamily, is expressed anti-Mol (CD11b), anti-Mo2 (CD14), and anti-B1 (CD20) on 95% of CD4+ cells, on 50% of CD8+ cells, and on mAbs. The efficiency of the purification process was ana- thymocytes that coexpress CD4 and CD8 (4, 10-12). Follow- lyzed in each case by indirect cell immunofluorescence and ing suboptimal activation of T cells with anti-CD3 monoclo- flow cytometry (Coulter EPICS flow cytometer) using T3 nal antibody (mAb) (4, 13), anti-CD2, or phorbol 12-myristate (CD3) and 4B10 (CD28) mAbs and fluorescein isothiocy- 13-acetate (PMA) (14, 15), crosslinking ofCD28 by anti-CD28 mAb results in enhanced T-cell proliferation (2, 8, 15-17) and Abbreviations: TCR, T-cell receptor complex; MHC, major histo- greatly augments synthesis ofmultiple lymphokines (18). The compatibility complex; mAb, monoclonal antibody; PMA, phorbol 12-myristate 13-acetate; IL, interleukin; APC, antigen-presenting cell. The publication costs of this article were defrayed in part by page charge tTo whom reprint requests should be addressed at: Division of payment. This article must therefore be hereby marked "advertisement" Tumor Immunology, Dana-Farber Cancer Institute, Mayer 730, 44 in accordance with 18 U.S.C. §1734 solely to indicate this fact. Binney Street, Boston, MA 02115. 6575 Downloaded by guest on September 30, 2021 6576 Immunology: Gimmi et al. Proc. Natl. Acad. Sci. USA 88 (1991) anate-labeled goat anti-mouse immunoglobulin (Tago). The and 2 x 104 fixed cells were added to the appropriate wells final T-cell preparation was >90%o CD3' and >88% CD28+ in a 96-well flat-bottomed microtiter plate (Nunclon; Nunc). in each case when compared with staining with an isotype- Proliferation Assay. CD28' T lymphocytes were incubated identical unreactive control antibody. Examination ofsmears in RPMI 1640 containing 10%o heat-inactivated human AB stained for nonspecific esterase (a-naphthyl acetate esterase; serum, 2 mM glutamine, 1 mM sodium pyruvate, penicillin Sigma) confirmed that the cell population contained z1% (100 units/ml), streptomycin sulfate (100 pug/ml), and gen- monocytes (36). tamicin sulfate (5 jig/ml). Cells were cultured at a concen- mAbs. 4B10 (IgGl) is an anti-CD28 mAb that immunopre- tration of 5 x 104 cells per 200 dul of medium in triplicate cipitates a 44-kDa disulfide-bonded dimer and enhances samples in a 96-well flat-bottomed microtiter plate at 370C for proliferation and lymphokine synthesis of suboptimally ac- 3 days in 5% CO2. Cells were cultured in medium and with the tivated T cells (data not shown). Indirect immunofluores- appropriate stimuli added. Cells were stimulated with PMA cence of CD28-transfected COS cells revealed :5% positive (Calbiochem) at 1 ng/ml and ionomycin (Sigma) at 100 ng/ml cells with similar intensities of staining using 4B10 or the (14, 15, 18). The anti-CD3 mAb was added at 1 ,tg/ml to the anti-CD28 mAbs YTH 913.12 and 9.3 (8). YTH 913.12 was 96-well flat-bottomed microtiter plates and incubated at room kindly provided by H. Waldmann (Cambridge, U.K.). Opti- temperature for 1 hr; the plates were then washed twice with mal stimulation with anti-CD28 mAb was obtained at a PBS before addition ofthe cells (3, 14, 39, 40). The concentration of 1 Ag/ml and this dose was used throughout anti-CD28 the experiments. A hybridoma secreting anti-CD3 mAb mAb 4B10 was added at 1 pug/ml. The fixed CHO-B7 and OKT3 (IgG2a) was obtained from the American Type Culture CHO-mock transfectants were added at 2 x 10" cells per well. Collection and the purified mAb was adhered to plastic plates Preliminary experiments showed that maximal stimulation at a concentration of 1 ,ug/ml. This concentration was found plateaued with the addition of 2 x 104 CHO-B7 cells. The to produce optimal stimulation in association with a second specificity ofthe stimulation with CHO-B7 cells was assayed signal of T-cell activation. 4B10 and OKT3 were purified by the addition ofanti-B7 mAb to the cultures at afinal ascites using a protein A-agarose column (Bio-Rad) as described dilution of 1:100. Over the wide range ofconcentrations (1:50 (37). The anti-B7 mAb 133 (IgM) was characterized in our to 1:2000) assayed, this dose was found to produce complete laboratory (31, 32), and was used as ascites at a final dilution blocking of CHO-B7 stimulation. of 1:100. Thymidine Incorporation Assay. Thymidine incorporation B7 Transfection. The B7 cDNA clone in the pCDM8 vector was used as an index of mitogenic activity. During the last 8 was digested with restriction endonucleases Dra I and Bgl II, hr ofthe 72-hour culture, the cells were incubated with 1 ,uCi and the fragment comprising nucleotides 86-1213, containing (37 kBq) of [methyl-3H]thymidine (ICN Flow, Costa Mesa, the coding region of B7, was isolated (31). The Dra I-Bgl II CA). The cells were harvested onto filters and the radioac- fragment was ligated into BamHI-digested, phosphatase- tivity on the dried filters was measured in a Packard Tri-Carb treated pLEN by a combination of sticky-end ligation, Kle- scintillation counter. now polymerase fill-in, and blunt-end ligation. pLEN is a Lymphokine Assay. Culture supernatants were collected 24 eukaryotic expression vector containing the human metal- hr after the initiation of the culture and IL-2 and IL4 lothionein IIA promoter, the simian virus 40 enhancer, and concentrations were assayed in duplicate using an ELISA kit the human growth hormone 3' untranslated region and poly- according to the manufacturer's instructions (Quantikine; R adenylylation site (38). pLEN was kindly provided by Met- & D Systems, Minneapolis). abolic Biosystem (Mountain View, CA). Fifty micrograms of Pvu I-linearized B7-pLEN construct was cotransfected with 5 ug of Pvu I-linearized SV2-Neo-Sp65 (44) into CHO-K1 RESULTS Chinese hamster ovary cells by electroporation using the B7-Transfected CHO Cells Stimulate Proliferation of Sub- BRL electroporator at settings of 250 V and 1600 mF. optimally Activated CD28+ T Cells. Crosslinking of CD28 on Transfectants were selected by growth in medium containing T cells by anti-CD28 mAb has been shown to stimulate T-cell the neomycin analogue G418 sulfate (400 jig/ml) and were proliferation and lymphokine synthesis (18). Since B7 is a cloned. Clones expressing cell surface B7, as assayed by natural adhesion ligand for CD28 (30), we attempted to indirect immunofluorescence with anti-B7 mAb, were re- determine whether binding of B7 to CD28 would deliver a cloned. These cells are referred to as CHO-B7 cells through- costimulatory signal to T cells. To this end, a CHO cell line out this paper. Mock-transfected CHO-K1 (CHO-mock) cells expressing high levels of B7 (CHO-B7) was constructed by were made by transfection ofPvu I-linearized SV2-Neo-Sp65 stable transfection of the B7 gene under the control of the alone. strong metallothionein promoter. The CHO-B7 cells were Cell Fixation. CHO cells were detached from tissue culture fixed with paraformaldehyde and used to stimulate CD28+ plates by incubation in Dulbecco's phosphate-buffered saline cells that had been suboptimally stimulated with PMA or without Ca2+ and Mg2' (PBS) with 0.5 mM EDTA for 30 min. anti-CD3. Cells were washed once in PBS and resuspended in PBS at As seen in Table 1, PMA (1 ng/ml) induced a 3- to 8-fold 1 per ml. An equal volume of freshly prepared 0.8% increase in T-cell proliferation over the medium-only con- paraformaldehyde in PBS was added and the cells were trols. Addition of paraformaldehyde-fixed CHO-B7 cells to gently mixed for 5 min at room temperature. An equal volume PMA-activated CD28+ T cells stimulated proliferation 17- to of 0.2 M lysine in PBS was added to block unreacted 40-fold. Addition of anti-CD28 mAb also enhanced CD28+ paraformaldehyde and the cells were pelleted by centrifuga- T-cell proliferation 26- to 58-fold compared with cells cul- tion. The cells were washed once in PBS, once in RPMI 1640 tured with PMA alone. The stimulation by CHO-B7 was (Whittaker Bioproducts) containing 10% heat-inactivated fe- -28% less than that observed with anti-CD28 mAb in all tal bovine serum (Sigma), resuspended in the same medium, experiments performed. CHO-mock cells did not induce and incubated for 1 hr in a humidified 37°C incubator. Cells proliferation over background, providing evidence that B7 were pelleted, washed in RPMI 1640 containing 10% heat- was specifically inducing the proliferative signal. Neither inactivated human AB serum (North American Biologicals, anti-CD28 mAb (4, 6, 8, 13) nor CHO-B7 cells were able to Miami), 2 mM glutamine, 1 mM sodium pyruvate, penicillin induce proliferation in the untreated CD28' T cells. Cultures (100 units/ml), streptomycin sulfate (100 Ag/ml), and gen- of the paraformaldehyde-fixed transfected CHO cells alone tamicin sulfate (5 tug/ml) (GIBCO). Cells were resuspended showed no proliferation over medium controls. Table 1 in this medium containing heat-inactivated human AB serum depicts results with T cells from three representative normal Downloaded by guest on September 30, 2021 Immunology: Girnmi et al. Proc. Natl. Acad. Sci. USA 88 (1991) 6577 Table 1. Effect of phorbol ester, anti-CD28, and CHO-B7 cells on proliferation of CD28+ T cells [3H]Thymidine incorporation, cpm (mean ± SEM) CD28' cell treatment Donor 1 Donor 2 Donor 3 Medium control 156 ± 65 130 ± 11 151 ± 29 PMA 1,404 ± 386 1,032 ± 176 537 ± 73 Anti-CD28 88 ± 18 109 + 6 188 ± 6 CHO-B7 104 ± 6 102 9 143 ± 37 PMA + CHO-B7 57,030 ± 1,017 34,560 ± 4,961 9,440 ± 1,103 PMA + anti-CD28 82,263 ± 1,137 45,023 ± 2,684 14,215 ± 1,682 PMA + CHO-mock 1,010 ± 228 728 ± 163 369 ± 36 PMA + CHO-B7 + anti-B7 1,041 ± 434 737 ± 78 559 ± 52 PMA + anti-CD28 + anti-B7 76,697 ± 1,241 48,776 ± 712 15,290 ± 2,011 donors, and similar results have been consistently observed B7-Transfected CHO Cells Induce IL-2 but Not IL-4 Secre- in seven independent experiments. tion. Stimulation of submitogenically activated T cells with To confirm that the increased proliferation observed in the anti-CD28 mAb has been shown to result in increased IL-2 PMA-treated CD28+ T cells was specifically mediated production (27). To determine whether-IL-2 secretion could through ligation to B7, anti-B7 mAb was added to the culture be induced by ligation of B7, the culture supernatants from system to block this binding. The addition of anti-B7 mAb cells cocultured with either PMA or anti-CD3 in the presence totally abrogated the proliferative response induced by the of CHO-B7 cells or anti-CD28 were collected from the above CHO-B7 cells (Table 1). In contrast, anti-B7 mAb had no experiments and assayed for lymphokine production. Results effect on the stimulation of proliferation induced by anti- of two representative donors from seven tested are depicted CD28 mAb. These results further confirm that B7 provides in Table 3. No IL-2 or IL4 was detected when the CD28+ T the costimulatory signal. cells were stimulated with PMA alone. When anti-CD28 mAb To determine whether binding of B7 to CD28 could aug- was added to PMA-stimulated CD28+ T cells, IL-2 secretion ment proliferation ofT cells that had received a first signal of was markedly increased. In contrast, there was no significant T-cell activation through the TCR, CD28+ T cells were first increase in IL-4 secretion over background, although positive submitogenically stimulated with anti-CD3 mAb fixed to controls demonstrated sensitivity and specificity of the as- plastic (3, 14). Activation via the TCR provides a more say. The addition of CHO-B7 cells similarly resulted in a physiologic model, since the cellular events following marked increase in IL-2 secretion but to a lesser extent, crosslinking of TCR by anti-CD3 mimic the transmembrane -45% of that observed with anti-CD28 mAb. As was ob- signaling that occurs following stimulation with antigen in served with anti-CD28 mAb, there was no increase in IL-4 association with MHC proteins. The results obtained using production. There was no IL-2 production when resting T anti-CD3 stimulation are shown in Table 2 for the same cells were cocultured with either anti-CD28 mAb or CHO-B7 normal donors depicted in Table 1. Activation with anti-CD3 cells (data not shown). CHO-mock cells did not induce IL-2 mAb fixed to plastic resulted in a small, 2- to 3-fold prolif- secretion by PMA-stimulated CD28+ cells. The addition of erative response above medium controls for the majority of anti-B7 mAb specifically and nearly completely blocked the donors examined. In contrast, donor 1 demonstrated a 10- stimulation ofIL-2 production by CHO-B7 cells. The anti-B7 fold stimulation. This increased proliferation was presumably mAb had no effect on IL-2 production in the activated cells due to the greater number ofcontaminating monocytes found stimulated with anti-CD28 mAb. in the preparation from this donor, as the presence of Previous studies have demonstrated that the addition of monocytes greatly increases the stimulatory potential of PMA and the calcium ionophore ionomycin strongly stimu- fixed anti-CD3 mAb (36). When CHO-B7 cells were added to lates proliferation ofCD28+ cells (18). Costimulation ofCD28 anti-CD3-activated T cells, a marked increase in stimulation cells with both PMA and ionomycin enhanced proliferation index, ranging from 23- to 180-fold, was observed. The by up to 75-fold compared with PMA alone in seven inde- addition of anti-CD28 mAb also led to a marked increase in pendent experiments. When anti-CD28 mAb or CHO-B7 proliferation, with a stimulation index ranging from 30- to cells were added to PMA- and ionomycin-stimulated CD28 75-fold over that observed with anti-CD3 mAb alone. This cells, proliferation was minimally augmented, between 1- and stimulation appeared to be B7-specific, since CHO-mock 2-fold (data not shown). However, to determine whether B7 cells did not augment proliferation. Addition of anti-B7 mAb ligation could further enhance IL-2 production, PMA- and completely blocked the proliferative response obtained with ionomycin-stimulated CD28 cells were cocultured with the CHO-B7 cells but again had no effect on the responses CHO-B7 cells or anti-CD28 mAb. As seen in Table 3, CD28+ seen with anti-CD28 mAb. This further confirms that the cells cultured with PMA and ionomycin secreted low levels stimulation occurred via binding of B7. of IL-2 and very low levels of IL-4. Addition of anti-CD28 Table 2. Effect of anti-CD3, anti-CD28, and CHO-B7 cells on proliferation of CD28' T cells [3HjThymidine incorporation, cpm (mean ± SEM) CD28' cell treatment Donor 1 Donor 2 Donor 3 Medium control 156 ± 65 130 ± 11 151 ± 29 Anti-CD3 1,953 ± 631 245 ± 14 347 ± 169 Anti-CD28 88 ± 18 109 ± 6 188 ± 6 CHO-B7 104 ± 6 102 ± 9 143 ± 37 Anti-CD3 + CHO-B7 46,543 ± 11,010 45,146 ± 4,391 35,106 ± 2,847 Anti-CD3 + anti-CD28 56,836 ± 10,440 18,383 ± 5,334 26,873 ± 7,833 Anti-CD3 + CHO-mock 1,618 ± 158 519 ± 135 282 ± 7 Anti-CD3 + CHO-B7 + anti-B7 174 ± 5 2,377 ± 1,072 321 ± 57 Anti-CD3 + anti-CD28 + anti-B7 54,646 ± 3,932 24,290 ± 14,630 30,326 ± 13,853 Downloaded by guest on September 30, 2021 6578 Immunology: Girnmi et al. Proc. Natl. Acad. Sci. USA 88 (1991) Table 3. Effect of B7 on IL-2 and IL-4 production in phorbol a soluble protein(s) that can be replaced by IL-1 and IL-6 and ester-stimulated CD28+ T cells the other requires direct cell-cell contact (6, 33, 41). There is Production, pg/ml evidence in both murine and human systems that this cell-cell contact costimulatory signal is expressed on both activated B Donor 1 Donor 2 cells and interferon y-stimulated monocytes. Therefore, this CD28+ cell treatment IL-2 IL-4 IL-2 IL-4 molecule should have the following characteristics: (i) it is Medium control <30 <30 <30 <30 not expressed on resting B cells or unstimulated monocytes PMA <30 <30 <30 <30 (32) but is induced when B-cell surface immunoglobulin is Anti-CD28 <30 ND <30 ND crosslinked or when monocytes are activated with interferon PMA + anti-CD28 10,500 65 11,300 60 'y (32-34); (ii) the stimulus will involve cell-cell contact (42, PMA + CHO-B7 3,800 <30 5,900 40 43); and (iii) the molecule must be distinct from MHC class PMA + CHO-mock <30 <30 <30 <30 II gene products. PMA + CHO-B7 + anti-B7 80 <30 <30 ND We postulate that following engagement of TCR with PMA + anti-CD28 + anti-B7 9,000 ND 11,200 ND antigen-MHC product, interaction of B7 on the APC with PMA + IM 80 50 <30 <30 CD28 on the T cell provides this costimulatory signal. Our PMA + IM + anti-CD28 11,500 <30 14,500 <30 prior studies (32, 34) have shown that B7 is not expressed on PMA + IM + CHO-B7 6,000 <30 10,000 ND resting B cells and unstimulated monocytes but appears PMA + IM + CHO-B7 + anti-B7 340 <30 <30 ND within 24 hr on the surface of B cells activated with anti- PMA + IM + anti-CD28 + anti-B7 10,000 ND 7,500 ND immunoglobulin and on monocytes stimulated with inter- IM, ionomycin; ND, not done. feron y. Our present results demonstrate that ligation of CD28 with B7 upregulates proliferation and markedly en- mAb to the PMA- and ionomycin-stimulated cells led to hances IL-2 production. These events are mediated by cell maximal IL-2 production, which was only slightly greater surface B7, since they are completely abrogated when cells than those levels observed when anti-CD28 mAb was added are cultured with anti-B7 mAb. Furthermore, as was previ- to cultures containing PMA alone. Addition of CHO-B7 cells ously observed with anti-CD28 mAb, ligation with B7 spe- to the PMA- and ionomycin-stimulated CD28' cells induced cifically induces IL-2 but not IL4 secretion. The pattern of higher levels of IL-2 secretion than was observed when the expression of B7, together with our present results, is con- CHO-B7 cells were added to CD28' cells stimulated with sistent with this hypothesis. Therefore, we argue that fol- PMA alone. Finally, specificity for B7 was again confirmed, lowing MHC-restricted, antigen-specific T-B cell interac- since anti-B7 mAb inhibited IL-2 production by 95%. tion, the ligation of CD28 by B7 provides an essential Similar results were obtained when anti-CD3 fixed to costimulatory signal that can amplify proliferation and lym- plastic was used to stimulate CD28 cells. As seen in Table 4, phokine synthesis ofa specific subset ofT cells. These T cells stimulation with anti-CD3 led to very low levels of IL-2 will proliferate in response to IL-2 and will most likely induce secretion and virtually no detectable IL-4. Addition of either B-cell proliferation by synthesis and secretion of B-cell anti-CD28 mAb or CHO-B7 cells led to maximal IL-2 secre- growth factors and IL-2. Finally, secretion of interferon 'y by tion without production of IL-4. Again, anti-B7 mAb could T cells will induce B7 on monocytes, thereby producing inhibit IL-2 secretion by CHO-B7 cells but had no effect on additional B7+ cells to interact with CD28+ T cells, thus anti-CD28 stimulation. amplifying the immune response. It should be noted that we have not formally proven that the B7 signal is transduced via CD28, since we did not specifically block the B7 ligand on T DISCUSSION cells. Until this is formally proven, this leaves the possibility APCs play a critical role in the initiation and amplification of that additional B7 ligands other than CD28 might be involved. an immune response, since they provide a specific signal This observation may have important clinical implications. through the TCR and one or more less-well-understood Considerable in vitro evidence suggests that the recognition costimulatory signals. Specifically, following presentation of of MHC product plus antigen in the absence of costimulatory MHC product-antigen complexes on the APC surface to the signals results in an abortive program of T-cell activation TCR, MHC-nonrestricted costimulatory signals are deliv- leading to a prolonged period of T-cell unresponsiveness ered by APCs, which then amplify the immune response. In (anergy) and, therefore, failure of clonal expansion. This fact, current evidence suggests that the absence of this defect appears to be due to a specific failure in IL-2 secretion second signal will most likely result in unresponsiveness of a that, in vivo, may result in tolerance (1). It is possible that the T cell that has been activated only through its TCR (1,5). Two CD28-B7 pathway mediates this response. Consequently, potential APC-induced costimulatory signals may actually the expression of B7 may dictate whether the T-cell response function independently. One such signal is mediated proceeds along the pathway of activation or tolerance. Ap- through proaches to block expression of B7 or to inhibit ligation of Table 4. Effect of B7 on IL-2 and IL-4 production in CD28 and B7 may provide a means to specifically induce anti-CD3-stimulated CD28' T cells tolerance. Conversely, it is also possible that unregulated expression of B7 on B cells or monocytes may be a sufficient Production, pg/ml stimulus to induce autoimmune disorders. Therefore, under- Donor 1 Donor 2 standing of the regulation of B7 expression and the pathways that mediate signaling via CD28 and possibly B7 may even- CD28' cell treatment IL-2 IL-4 IL-2 IL-4 tually provide avenues for specific intervention. Medium control <30 <30 <30 <30 Anti-CD3 80 <30 <30 45 We thank John Daley for technical assistance and Kerstin Hilde- Anti-CD3 + anti-CD28 1500 <30 1350 <30 brand for preparation of the manuscript. This work was supported by Anti-CD3 + CHO-B7 1050 <30 1350 40 National Institutes of Health Grants CA40216, AR33713, and Anti-CD3 + CHO-mock <30 <30 <30 <30 A129530. C.D.G. is supported by the Swiss National Science Foun- Anti-CD3 + CHO-B7 + anti-B7 130 <30 <30 ND dation and the Swiss Cancer League. J.G.G. is supported by the Fogarty International Center of the National Institutes of Health, Anti-CD3 + anti-CD28 + anti-B7 2200 <30 1500 ND Grant 1FO5 TW04469. A.S.F. is supported by Public Health Service ND, not done. Grant 5K08 CA01105. Downloaded by guest on September 30, 2021 Immunology: Gimmi et al. Proc. Natl. Acad. Sci. USA 88 (1991) 6579 1. Schwartz, R. H. (1990) Science 248, 1349-1356. Fitzgerald, K., Hodgdon, J., Protentis, J., Schlossman, S. & 2. June, C. H., Ledbetter, J. A., Gillespie, M. M., Lindsten, T. & Reinherz, E. (1984) Cell 36, 897-906. Thompson, C. B. (1987) Mol. Cell. Biol. 7, 4472-4481. 25. Dustin, M. I., Olive, D. & Springer, T. A. (1989) J. Exp. Med. 3. Weiss, A., Manger, B. & Imboden, J. (1986) J. Immunol. 137, 169, 503-517. 819-825. 26. Bierer, B. E., Peterson, A., Barbosa, J., Seed, B. & Burakoff, 4. Martin, P. J., Ledbetter, J. A., Morishita, Y., June, C. H., S. J. (1988) Proc. Natl. Acad. Sci. USA 85, 1194-1198. Beatty, P. G. & Hansen, J. A. (1986) J. Immunol. 136, 3282- 27. Thompson, C. B., Lindsten, T., Ledbetter, J. A., Kunkel, 3287. S. L., Young, H. A., Emerson, S. G., Leiden, J. M. & June, 5. Ramsdell, F. & Fowlkes, B. J. (1990) Science 248, 1342-1348. C. H. (1989) Proc. NatI. Acad. Sci. USA 86, 1333-1337. 6. Weaver, C. T. & Unanue, E. R. (1990) Immunol. Today 11, 28. Lindsten, T., June, C. H., Ledbetter, J. A., Stella, G. & 49-55. Thompson, C. B. (1989) Science 244, 339-343. 7. June, C. H., Ledbetter, J. A., Linsley, P. S. & Thompson, 29. Fraser, J. D., Irving, B. A., Crabtree, G. R. & Weiss, A. (1991) C. B. (1990) Immunol. Today 58, 271-276. Science 251, 313-316. 8. Hara, T., Fu, S. M. & Hansen, J. A. (1985) J. Exp. Med. 161, 30. Linsley, P. S., Clark, E. A. & Ledbetter, J. A. (1990) Proc. 1513-1524. NatI. Acad. Sci. USA 87, 5031-5035. 9. Aruffo, A. & Seed, B. (1987) Proc. NatI. Acad. Sci. USA 84, 31. Freeman, G. J., Freedman, A. S., Segil, J. M., Lee, G., Whit- 8573-8577. man, J. F. & Nadler, L. M. (1989) J. Immunol. 143, 2714-2722. 10. Turka, L. A., Ledbetter, J. A., Lee, K., June, C. H. & Thomp- 32. Freedman, A. S., Freeman, G., Horowitz, J. C., Daley, J. & son, C. B. (1990) J. Immunol. 144, 1646-1653. Nadler, L. M. (1987) J. Immunol. 139, 3260-3267. 11. Hansen, J. A., Martin, P. J. & Nowinski, R. C. (1980) Immu- K. nogenetics 10, 247-260. 33. Kawakami, K., Yamamoto, Y., Kakimoto, K. & Onoue, 12. Damle, N. K., Mohaghepour, N., Hansen, J. A. & Engelman, (1989) J. Immunol. 142, 1818-1825. E. G. (1983) J. Immunol. 131, 2296-2299. 34. Freeman, G. J., Freedman, A. S., Rhynhart, K. & Nadler, 13. Ledbetter, J. A., Martin, P. J., Spooner, C. E., Wofsy, D., L. M. (1990) Blood Suppl. 1, 76, 206. Tsu, T. T., Beatty, P. G. & Gladstone, P. (1985) J. Immunol. 35. Yamada, H., Martin, P. J., Bean, M. A., Braun, M. P., Beatty, 135, 2331-2336. P. G., Sadamoto, K. & Hansen, J. A. (1985) Eur. J. Immunol. 14. Manger, B., Weiss, A., Imboden, J., Laing, T. & Stobo, J. D. 15, 1164-1172. (1987) J. Immunol. 139, 2755-2760. 36. Jenkins, M. K., Ashwell, J. D. & Schwartz, R. H. (1988) J. 15. Wiskocil, R., Weiss, A., Imboden, J., Kamin-Lewis, R. & Immunol. 140, 3324-3330. Stobo, J. (1985) J. Immunol. 134, 1599-1603. 37. Van Wauwe, J. P., Demey, J. R. & Goossens, J. G. (1980) J. 16. Damle, N. K., Doyle, L. V., Grosmaire, L. S. & Ledbetter, Immunol. 124, 2708-2712. J. A. (1988) J. Immunol. 140, 1753-1761. 38. Friedman, J. S., Cofer, C. L., Anderson, C. L., Kushner, 17. Gmunder, H. & Lesslauer, W. (1984) Eur. J. Biochem. 142, J. A., Gray, P. P., Chapman, G. E., Stuart, M. C., Lazarus, 153-160. L., Shine, J. & Kushner, P. J. (1989) BiolTechnology 7, 18. June, C. H., Ledbetter, J. A., Lindstein, T. & Thompson, 359-362. C. B. (1989) J. Immunol. 143, 153-161. 39. Matsuyama, T., Yamada, A., Kay, J., Yamada, K. M., Ak- 19. Marlin, S. D. & Pringer, T. A. (1987) Cell 51, 813-819. iyama, S. K., Schlossman, S. F. & Morimoto, C. (1989) J. Exp. 20. Van Seventer, G. A., Shimizu, Y., Horgan, K. J. & Shaw, S. Med. 170, 1133-1148. (1990) J. Immunol. 144, 4579-4586. 40. Geppert, T. D. & Lipsky, P. E. (1987) J. Immunol. 138, 1660- 21. van Noesel, C., Miedema, F., Brouwer, M., de Rie, M. A., 1666. Aarden, L. A. & van Lier, R. A. W. (1988) Nature (London) 41. Williams, I. R. & Unanue, E. R. (1990) J. Immunol. 145, 85-93. 333, 850-851. 42. Mueller, D. L., Jenkins, M. K. & Schwartz, R. H. (1989) J. 22. Dustin, M. L. & Springer, T. A. (1989) Nature (London) 341, Immunol. 142, 2617-2628. 619-624. 43. Kohno, K., Shibata, Y., Matsuo, Y. & Minowada, J. (1990) 23. Springer, T. A. (1990) Nature (London) 346, 425-434. Cell. Immunol. 131, 1-10. 24. Meuer, S., Hussey, R., Fabbi, M., Fox, D., Acuto, O., 44. Streuli, M. & Saito, H. (1989) EMBO J. 8, 787-7%. Downloaded by guest on September 30, 2021