Selective CD28pYMNM mutations implicate phosphatidylinositol 3-kinase in CD86-CD28-mediated costimulation

Immunity,

Vol. 3, 417-426,

October,

1995, Copyright

0 1995 by Cell Press

Selective CD28pYMNM Mutations Implicate Phosphatidylinositol 3-Kinase in CD86-CD28-Mediated Costimulation Yun-Cai Cai,** Daniel Cefai,‘t Helga Schneider,‘* Monika Raab,‘t Nasrin Nabavi,§ and Christopher E. Rudd’t *Division of Tumor Immunology Dana-Farber Cancer Institute 44 Binney Street Boston, Massachusetts 02115 tDepartment of Pathology *Department of Medicine Harvard Medical School Boston, Massachusetts 02115 §Department of Experimental Oncology Hollings Cancer Center Medical University of South Carolina Charlestown, South Carolina 294252201

Summary CD28 costimulatory signals are required for lymphokine production and T cell proliferation. CD28 signaling recruits the intracellular proteins PI 3-kinase, ITK, and GRE2ISOS. PI 3-kinase and GRB-S/SOS bind the CD28 cytoplasmic pYMNM motif via SH2 domains. We generated CD28 pYMN# mutants and found that Ylsl mutation (Y1%D28F) disrupted both PI 3-kinase and GRB-2 binding, while M’” mutation (Mls4CD28C) disrupted only PI 3-kinase binding. Both mutants still bound ITK. We have assessed the ability of these selective mutants to support IL-2 production upon TCRLj CD3 ligation in the presence of CHO-CD88 (B7-2) cells. Both Y1%D28F and M’“CD28C mutants failed to generate IL-2. These data directly implicate PI 3-kinase in CDSI-media&d costimulation leading to IL-2 secretion. Wortmannin, an inhibitor of PI 3-kinase, induced cell apoptosis and as such was unsuitable for use in this study. Introduction Ligation of the antigen receptor on the surface of T cells (T cell receptor [TCR]qCD3) alone is insufficient to induce resting T lymphocytes to produce maximal levels of interleukin 2 (IL-2) and optimal proliferation (Schwartz, 1992). Additional ligand-receptor interactions on antigenpresenting cells (APCs) and T cells are required. The best characterized among these costimulatory pairs is CD28 and its binding partnersCD80 (87-l) and CDSS(B7-2; 870) (Jenkinsand Johnson, 1993; Linsleyand Ledbetter, 1993). While both ligands are expressed constitutively on macrophages and hendritic cells, CD86 is expressed more rapidly and at higher levels (Azuma et al., 1993; Freeman et al., 1993a; Hathcocket al., 1993,1994). Both ligands bind CD28 with similar avidity, but CD86 exhibits a higher rate of dissociation (Linsley et al., 1994). CD28-deficient mice show reduced T cell responses to allogeneic APCs, anti-

CD3 plus anti-CD28 or APC-B7-1, and exhibit reduced T cell-dependent humoral responses (Green et al., 1994; Shahinian et al., 1993). Similarly, CD80-negative mice exhibit reduced immune responses (Freeman et al., 1993b). In certain settings, the activation of T cell helper (Th)l cells without costimulation results in unresponsiveness to further challenge by antigen (Jenkins et al., 1988; Koulova et al., 1991; Schwartz, 1992). Much of CD28 costimulation is elicited indirectly via an effect on lymphokine production. Nonresponsiveness can be bypassed by IL-2, 11-4, or IL-7via signaling through a common y chain (Boussiotis et al., 1993, 1994). The importance of CD28 in T cell responses is further underlined by the findings that the generation of cytolytic responses against tumors can be regulated by ligation of this receptor (Baskar et al., 1993; Chen et al., 1992; Townsend and Allison, 1993). Furthermore, overexpression of CD80 together with an immunogenic antigen in pancreatic islets results in rapid autoimmune insulitis (Guerder et al., 1994), and the blockage of interaction between CD28 and 87 induces antigen tolerance and prolongs survival of transplants (Lenschow et al., 1992; Linsley et al., 1992a). CD28 also stabilizes mRNA for various lymphokines such as interleukin 2 (IL-2), and induces the expression of CTLA4 and the CD40 ligand (Fraser et al., 1991; Lindsten et al., 1989). Recent studies suggest that CD80 and CD86 differentially regulate cytokine production (Kuchroo et al., 1995; Lenschow et al., 1995; Freeman et al., 1995). CD28 could control any or all of these events by means of its interaction with phosphatidylinositol 3-kinase (PI 3-kinase), GRB9/SOS, or ITK (Truitt et al., 1994; Prasad et al., 1994; August and DuPont, 1994; Schneider et al., 1995a; August et al., 1994). The cytoplasmic tail of the receptor carries the phosphotyrosine-based motif pYMNM, which can bind with high affinity to the SH2 domains of PI 3-kinase and GRB-2 (Prasad et al., 1994; Schneider et al., 1995a). PI 3-kinase is comprised of a heterodimer consisting of an adaptor subunit (~85) with two SH2 domains that is coupled to a ~110 catalytic subunit (Carpenter et al., 1990; Dhand et al., 1994a; Parker and Waterfield, 1992). It is a bifunctional kinase, a serine kinase (Dhand et al., 1994b), and a lipid kinase that phosphorylates the D-3 position of the inositol ring of phosphatidylinositol, phosphatidylinositol4-phosphate, and phosphatidylinositol4,5 bisphosphate, generating PI 3-P, PI 3,4-P2 and PI 3,4,5-P3. Engagement of CD26 induces the recruitment of PI 3-kinasewith the receptor (Prasad et al., 1994; Truitt et al., 1994). Site-directed mutagenesis and peptide binding analysis showed that the CD28 pYMNM motif binds to the ~85 C- and N-terminal SH2 domains with an avidity comparable to that observed for PDGF-R and IRS-l. The C-terminal SH2 domain binds to the CD28 pYMNM motif with some lo-fold greater affinity than the N-terminal domain (Prasad et al., 1994). Given the importance of the PI 3-kinase binding to a similar motif

Immunity 418

in the PDGF-R (Fantl et al., 1992; Valius and Kazlauskas, 1993) PI 3-kinase is a potential candidate in providing second signals required for T cell proliferation (Rudd et al., 1994). CTLA4 has also been found to bind to PI 3-kinase by means of a pYVKM motif (Schneider et al., 1995b). SOS (Son of Sevenless) is a guanine nucleotide exchange factor that activates p21r”, by converting it from a GDP- to GTP-bound state (Bonfini et al., 1992; Simon et al., 1993). GRB-2 is an SH3-SH2-SH3 adapter subunit, which uses its SH3 domain to bind to proline residues within SOS, while using its single SH2 domain to bind the phosphotyrosine-based motifs of surface receptors (Simon et al., 1993). GRB-2 is recruited by CD28 by binding to the pYMNM motif in response to antibody ligation (Schneider et al., 1995a). CD28 may recruit GRB-P/SOS from the cytosol to the cell membrane where it should be well placed to regulate ~21’“. Indeed, anti-CD28-mediated cross-linking of CD28 has been reported to activate p21raS (Nunes et al., 1994) and downstream targets MAP kinase (MAPK; ERK 1,2) and Jun kinase (JNK) (August and DuPont, 1994; Nunes et al., 1994; Su et al., 1994). ~21’” is required for TCR-mediated IL-2 production (Rayter et al., 1992; Baldari et al., 1993). Despite the sharing of a common binding motif, PI 3-kinase does not prevent GRB-2 binding and vice versa. Based on peptide binding, the GRB-2 SH2 domain binds to CD28 with a considerably lower avidity than the PI 3-kinase SH2 domains. However, this may underestimate the avidity of binding, since secondary points of contact may exist. Consistent with this, mutation of Y-191 significantly reduced, but did not completely eliminate GRB-2 binding (Schneider et al., 1995a). CD28 has also been reported to bind to the nonreceptor tyrosine kinase ITK (EMT/TSK) (August et al., 1994). Nonreceptor tyrosine kinases can interact with cell surface receptors, as initially shown with CD4-~58’~” (Rudd et al., 1994). ITK is structurally related to members of the src family with an unique N terminus, an SH3 and SH2 domain, aswell asa kinasedomain. However, it sharesgreatest homology with members of the Tee family of tyrosine kinase. It is restricted to T cells, being expressed at the highest level on the thymus (Desiderioet al., 1993; Heyeck and Berg, 1993). CD28 ligation is required to activate and recruit ITK from the cytoplasm (August et al., 1994). The molecular basis of CD28 binding to ITK has yet to be established. The involvement of the phosphotyrosine-based CD28pYMNM motif in recruiting downstream molecules implicates an intervening protein-tyrosine kinase in the regulation of costimulation. p581Ckand p59bn can fufill this function by phosphorylating CD28 at Y-191, which in turn facilitates the recruitment of ~85 and GRB-2 (Raab et al., 1995). ZAP-70 and ITK failed to induce the formation of CD28-intracellular ligand complexes. Despite the involvement of src kinases, CD28 appears to act as an independent signaling unit. Anti-CD28 ligation of peripheral T cells can induce the recruitment of PI 3-kinase and GRB-2, without prior TCR ligation (Prasad et al., 1994; Schneider et al., 1995c; Truitt et al., 1994). Further, src kinases have

been reported to coprecipitate with CD28 in gentle detergents (Hutchcroft and Bierer, 1994). By binding to multiple intracellular signal-transducing molecules, CD28 may engage multiple signaling cascades in T cells. Different CD28 signaling pathways may selectivelyregulateevents, such asan increase in lymphokine production, and the expression of CTLA-4 and the CD40 ligand. Pages et al. (1994) demonstrated the importance of the pYMNM motif by showing that the mutation at Y-191 prevented the ability of anti-CD28 antibodies to produce IL-2. However, this mutation disrupts both PI 3-kinase and GRB-2 binding, and did not distinguish the roles played by PI-3 kinase, GRBQISOS, or ITK in TCRCDPSmediated costimulation. Other attempts to address this question using wortmannin, an inhibitor of PI 3-kinase, to block CD28 induced IL-2 production have resulted in contradictory findings (Lu et al., 1995; Ward et al., 1995). To address the role of CD2Sassociated PI 3-kinase in costimulation, we assessed the ability of CD28-PI 3-kinase binding mutants to induce IL-2 production in a costimulation system dependent on TCRTJCD3 and CD28 coligation. While wild-type CD28 supported IL-2 production, a SH2 binding mutant (Ylg1CD28F) and a selective PI 3-kinase binding mutant (M’%CD28C) failed to support IL-2 production, or to recruit the kinase. While the Mlg4CD28C mutant failed to bind to PI 3-kinase, it retained an ability to bind to GRB-2 and ITK. These data implicate CD28-PI 3-kinase in CD28 regulation of IL-2 production. Results Human CD28 Mediates AntXD8Dependent IL-2 Production The T cell hybridoma DC27 was transfected with human CD28 and assessed for the ability to respond to anti-CD3 in the presence of CHO cells expressing CD88 (97-2). The murine hybridoma has been previously described (Gabert et al., 1987). Transfected human CD28 was expressed at least 20 times greater than the endogenous murine CD28, as detected by immunofluorescence staining (Figure 1A). Biotinylation of cells followed by immunoprecipitation with anti-CD28 confirmed this observation (Figure 19). A broad band at 40-50 kDa corresponding to CD28 was precipitated by anti-mouse CD28 (Figure 1, lane 3) and antihuman CD28 (lane 4). Densitometric readings showed approximately 40 times greater human CD28 (hCD28) than endogenous murine CD28 (Figure 1, lane 4 versus lane 3). High levels of hCD28 expression allowed us to establish a model system to examine the role of the CD28 pYMNM motif in IL-2 production. Incubation of cells with anti-CD3 antibody and CHO-CD86 cells induced high levels of IL-2 production in hCD28-transfected cells (wild type) (Figure 2). The response depended on the expression of CD86, since the incubation of the T cells with anti-CD3 alone or anti-CD3 plus untransfected CHO cells failed to generate IL-2. Similarly, CHO-CD86 plus anti-CD3 caused little IL-2 production in vector-transfected cells. The low level of endogenous murine CD28 expression was insufficient to support IL-2 production at the lower levels of anti-CD3

CDZE-Mediated 419

Signaling

-

I

CHO

CHO-CD86 I I

I

400 rCD28

300 ” E 3 ;;;

200

i! 100

Figure 1, Analysis of Human CD28 Expression in Transfected DC27 Cells (Wild Type) (A) Analysis by flow cytometry of expression of murine CD3. murine CD28 (mCD28), and human CD28 (hCD28) using MAbs 2C11,37.51, and L293, respectively. The control corresponds to cell staining with secondary antibodies alone (FITC-conjugated goat anti-mouse MAb). (B) The SDS-PAGE profile of immunoprecipitates of biotinylated mCD28 and hCD28. DC27 cells were biotinylated and subjected to immunoprecipitation using MAbs as indicated (see Experimental Procedures). lmmunoprecipitations with protein G-Sepharose (PGS) and rabbit anti-hamster antibodies plus protein G-Sepharose (RaH) served as negative controls. Lane 1, PGS; lane 2, RaH; lane 3, anti-mCD28; lane 4. anti-hCD28.

Anti-CD% Figure 2. IL-2 Production Is Dependent on Anti-CD3 in Conjunction with CD86 IL-2 production in DC27 cells that were transfected with hCD28 (wild type) or vector was analyzed by incubation with anti-CD3, anti-CD3 plus CHO, and anti-CD3 plus CHO-CD86. The vector or human CD28transfected DC27 cells were cultured for 24 hr in the presence of anti-CD3 MAb (2C11, IO-” M) in medium alone (-). or together with Cl-IO or CHO-CD86 cells. Supernatants were harvested and analyzed for IL-2 in triplicates. Shown are typical results from one of six experiments.

antibody. At higher concentrations of anti-CD3 monoclonal antibody (MAb), a slight contribution of endogenous murine CD28 was observed in parental untransfected DC27 cells; however, analysis using higher anti-CD3 concentrations was complicated by a significant decrease in cell viability (Ucker et al., 1989; data not shown). Various anti-CD3 concentrations were tested in this system, with maximal levels of IL-2 observed at lo-” M. As a control for dependence on CD86 expression, anti-CD86 was found to inhibit CD86-mediated costimulation in a dose-dependent manner (Figure 3). Antibody concentration at about Kg/ml caused a 50% inhibition of the response.

Involvement of pYMNM Residues in CD28 Signaling

0.5

Y-191 and M-194

To assess the role of PI 3-kinase in costimulation, we transfected DC27 cells with various mutated forms of human CD28 that are defective in binding to PI 3-kinase. To allow study of the relative contributions of PI 3-kinase, GRB-2, and ITK binding, it was necessary to generate a form of CD28-carrying mutation that selectively disrupted PI 3-kinase binding. Y’%D28F within pYMNM abrogates binding to all SH2 domains, while the M’%D28C mutation should selectively disrupt PI 3-kinase binding (Prasad et al., 1994). To control for differences between individual cells, two transfectants were selected for the analysis of each mutation. The surface expression of the wild-type, Yqg1CD28F (YrsrF) and Mlg4CD28C (Mlg4C) were similar for most of the transfectants (Figure 4). One M’%D28C (Ml%) transfectant (termed Ml%-b) expressed CD28 at

I

0-l 0

0.156

0.31

0.625

Anti-CD86

1.25

2.5

5

@g/ml)

Figure 3. IL-2 Production Induced by Anti-CD3 plus CHO-CD86 Is Blocked by Anti-CD86 MAb The costimulation assay was performed as in Figure 2 in the presence of anti-CD86 antibody (2DlO) at various concentrations. Antibodies were added at the start of culture. Results are expressed as a percentage of IL-2 production in the absence of 2DlO MAb.

somewhat higher levels, although this did not alter itsfunctional phenotype (see below). An examination of other surface markers such as murine CD28 and CD3 showed that expression levels were similar to those of the wild-type

immunity

420

C I

anti-p85 0

Hot 30

*

DC 27 Transfectants

pas

123456

Figure 5. Y19’ and M lI” Residues Binding by PI 3-Kinase

IO-12

10-13

Anti-CD3 Figure 4. Mutations CHO-CDIB-Induced

10-1’

Concentration

at Y-191 and IL-2 Production

M-194

(M)

Abrogate

Anti-CD3

plus

(A)Cellsurfaceexpression ofwild-typeandmutatedformsof hCD28on transfected DC27 cells was analyzed by flow cytometry. Background fluorescence of secondary MAb alone was subtracted. Wl represents the wild-type hCD28 transfectant, while F191 and Cl94 represent the mutations at Tyr-191 to Phe and Met-194 to Cys, respectively. Two cell populations were tested for each transfectant (indicated as a, b). (6, C) Y’O’ and M’” residues in pYMNM motif are required for CD28induced IL-2 production. DC27 transfectants were incubated with CHO-CD86 and 2Cll (10~” M) for 24 hr, after which supernatants were assayed for IL-2 production (B). Each column includes standard deviation of triplicates. (C)depicts a separate experiment showing the inability of the F191 and Cl94 mutants to produce IL-2 in response to CHO-CD86 and different concentrations of anti-CD3 (lo-” to lo-l3 M).

transfectant (unpublished data). Hence, mutations at did not prevent expression of CD28 on the these residues cell surface. We next examined the ability of Y1VD28F (F191) and M’%ZD28C (C194) mutants to support IL-2 production. While anti-CD3 plus CHO-CD86 induced high levels of IL-2 production (>200 U/ml) in the wild-type transfectant, the Y1g’CD28F and Mig4CD28C mutants secreted lymphokine at markedly reduced levels (Figure 48). Each of the duplicate transfectants gave similar results. In both cases,

in pYMNM

Motif Are Required

for

Different DC27 transfectants (vector, WT, F191. and C194) were incubated with CHO or CHO-CD88 cells as indicated at 37% for 0 (lane 1) or 30 min (lanes 2-5), solubilized in NP-40 lysis buffer and immunoprecipitated with an anti-human CD28 antibody (L293). Precipitates were then labeled in an lipid kinase assay (A), or blotted using an anti-p85 antisera(C). Lanes 1 and 3, CHO-CD86 plus wild-type human CD28 transfectant; lane 2, CHO-CD86 plus vector-transfected DC27 cells; lanes 4 and 5, CHO-CD86 plus F191- and C19Ctransfected cells, respectively. Lane 6 corresponds to an anti-p85 precipitation, while lane 7 represents a rabbit anti-mouse precipitation. (8) shows the PI 3-kinase activities in (A) expressed as a percentage of wild type; the intensity of each PI-P spot was determined by Ultra&an densitometry.

IL-2 production was less than 10% of the wild-type control. Further, the two M199CD28C mutants showed a similar reduction in IL-2 production, despite differences in the expression of the receptor (Figure 4A). Similar results were observed when lo- to lOO-fold lower concentrations of anti-CD3 antibody were used (Figure 4C). Interestingly, the effect was noted at anti-CD3 concentrations as low as lo-l3 M, a concentration that is expected to approximate the concentration of peptide present in physiological APC-T cell responses (Sykulev et al., 1994). Y-191 and M-194 Residues Are Required for PI 3-Kinase Binding We next determined whether CHO-CD86 ligation of the transfected wild-type CD28 could recruit PI 3-kinase, and whether mutations at Y-191 and M-193 disrupted PI 3-kinase binding. As seen in Figure 5A, anti-CD28 precipitated a small amount of associated PI 3-kinase activity from unligated cells (lane 1). By contrast, exposure of wild type to CHO-CD86 resulted in marked increase in CD28associated PI 3-kinase as detected by kinase activity (Fig-

CD2s-Mediated Signaling 421

ure 5A, lane 3 versus lanes 1 and 2) and the presence of p85 as detected by immunoblotting (Figure 5C, lane 3 versus lanes 1 and 2). The increase in associated PI 3-kinase activity was therefore primarily due to PI 3-kinase recruitment by CD28. Importantly, mutation at Y-191 (Yqg’CD28F) reduced CD28-coprecipitated PI 3-kinase to background levels, as demonstrated by kinase analysis (Figure 5A, lane 4), and by anti-p85 immunoblotting (Figure 5C, lane 4). A small amount of kinase activity was observed with the Cl94 mutant (Figure 5A, lane 5); however, the amount of kinase was too low to be detected by anti-p85 immunoblotting (Figure 5C, lane 5). Anti-p85 precipitates served as a positive contol (Figures 5A and 5C, lane 6). To confirm the validity of the above observations, the interactions of wild-type CD28 and mutants with PI 3-kinase were reconstituted in a baculoviral expression system. The expression of large amount of proteins in insect cells allows the detection of low levels of association between proteins. We have previously shown that the expression of tyrosine kinases ~56”~ and p59b” with CD28 regulates PI 3-kinase, GRB-2, and ITK binding to CD28 (Raab et al., 1995). While coexpressed CD28 and PI 3-kinase failed to bind to each other, the presence of ~56’“” and its ability to phosphorylate Y-l 91 allowed for the formation of CD28-PI 3-kinase complexes (Raab et al., 1995). As shown in Figure 6A, under these conditions, wild-type CD28 was associated with PI 3-kinase (lane 1). By contrast, neither the Y1%D28F nor the M?ZD28C mutants were able to bind to PI 3-kinase (Figure 6A, lanes 2 and 3). These data confirm that the mutation at Y-191 and M-l 94 disrupted PI 3-kinase binding. To establish the specificity of the M-194 mutation in disrupting PI 3-kinase but not GRB-2 binding, the mutant was coexpressed with GRB-2 and Ick and assessed for binding. As shown in Figure 6B, the Y1%D28F mutant reduced binding to GRB-2 (lane 2), while M194CD28C had no effect when compared with wild type (lanes 1 and 3). In the case of ITK (Figure 6C), although we have had difficulty detecting the CD28-ITK interaction in T cells (data not shown), in coexpression studies mutations at neither Y-191 nor M-l 94 disrupted binding (lanes 3 and 4) when contrasted with wild-type CD28 (lane 2). The M-194 mutation therefore disrupts PI 3-kinase binding to CD28 without altering the binding of GRB-2 or ITK. Wortmannin, an Inhibitor of PI 3-kinase, Induces Apoptosis In T Cells Wortmannin has been reported to selectively inhibit PI 3-kinase when used at low concentrations (Arcaro and Wymann, 1993; Yano et al., 1993). However, previous studies using this drug to inhibit CD28 costimulation in T cells have yielded contradictory results (Lu et al., 1995; Ward et al., 1995). This underlines the importance of our mutational analysis outlined in this study. Nevertheless, we also attempted to use this inhibitor to assess whether it could block eD28-induced IL-2 production. The addition of a range of wortmannin concentrations from 20-200 nM inactivated anti-CD28-precipitated PI 3-kinase (Figure 7A;

B

11284so-

,zz,

4-m

f#km-

3s1

2

3 1

1

Figure 6. Mutational and ITK

Analysis

2

3

2

3

4

of CD28 Binding

lo PI 3-Kinase,

GRSP.

Sf21 cells were infected with different combinations of recombinant baculoviruses (CD28, ~85, ~56”~, GRB-2, or ITK) as indicated. ITK was expressed as an HA-tagged protein. CD28 proteins were immunoprecipitated from cell lysates with antiCD28, and were subjected to Western blot analysis. The membranes were blotted with either anti~85 (A), anti-GRBP (B), or anti-HA (C) antibodies. (A) F191 and Cl94 mutants failed lo bind ~85. Lane 1, coexpression of CD28, p561ck, and ~85; lane 2, F191 plus Lck and ~85; lane 3, Cl94 plus Lck and ~85. (8) Cl94 mutant binds to GAB-P. Lane 1, coexpression of CD28, pE16~, and GRB-2; lane 2, F191 plus Lck and GRB-2; lane 3, Cl94 plus lck and GRB-2. (C) V1 and Cl94 mutants bind to ITK. Lane 1, coexpression of CD28 plus HA-tagged ITK; lane 2. CD28 plus Lck and ITK; lane 3, F191 plus Lck and ITK; lane 4, Cl94 plus Lck and ITK.

Figure 78, top), as well as anti-p85-precipitated kinase (Figure 7A; Figure 78, bottom). Densitometric readings indicated that the reagent inhibited the kinase activity by some 90% (Figure 7A). Under these conditions, wortmannin caused a dramatic reduction of cell growth, as detected by [3H]thymidine incorporation and in cell viability, as assessed by trypan blue exclusion (Figure 7A). This effect was caused by concentration-dependent induction of apop tosis as revealed by DNA fragmentation (Figure 7C). Concentrations as low as 10 nM induced significant DNA fragmentation in the DC27 cells. Although of interest, the apoptotic effect of wortmannin made this drug inappropriate for use in examining the role of PI 3-kinase in costimulation. Discussion By binding to PI 3-kinase, GRB-2/SOS, and ITK, CD28 has the potential to engage multiple signaling pathways

Immunity 422

Am 60

p

60

p

4o 20

0

x)

60 100 200 500 600 1000

Wortmannln

B

(nM)

Figure

C

Wortmannin (nM)

I

7. Wortmannin

Induces

Wortmannin (nh4) I

1

Apoptosis

in the DC27 Cells

Wild-type human CD28 transfectant was incubated with or without wortmannin at different concentrations. (A) Cell proliferation (13H]-Tdr incorporation) and viability (trypan blue exclusion) were monitored after 24 hr culture of cells in the presence of various concentrations of wortmannin. (6) Inhibition of lipid kinase activity. After 2 hr of exposure to wortmannin, DC27 cells transfected with hCD28 were subjected to immunoprecipitation with either anti-CD28 (top) or anti-p85 (bottom), labeled in a lipid kinase assay. The intensity of each PI-P spot was also measured and expressed as a percentage of that in the absence of wortmannin (A). (C) DNA fragmentation was measured after 4 hr of cell culture in the absence or presence of different concentrations of wortmannin (O600 nM). Cells were solubilized and subjected to phenol/chloroform extraction and RNAse treatment. Digested DNA was then separated on agarose gel and visualized with ethidium bromide staining. During these experiments, a DMSO concentration equivalent to that contained in the highest concentration of wortmannin was used as a control.

in T cells. In turn, different signals may be linked to CD28regulated events, such as increased lymphokine production and the expression of CTLA4 and the CD40 ligand. Mutation of residue Y-191 has been reported to prevent anti-CD28-induced IL-2 in a T cell hybridoma (Pages et al., 1994). However, this study did not distinguish between the relative roles of PI 3-kinase, GRB9/SOS, and ITK binding in TCRICD3-CD28 coregulation of IL-2 production. Both PI 3-kinase and GRB-2 bind to CD28 pYMNM motif (Prasad et al., 1994; Truitt et al., 1994; Schneider et al., 1995a). Here, we have demonstrated that a mutation (M-l 94C) that selectively disrupts PI 3-kinase binding without altering GRB-2 or ITK binding markedly reduces IL-2 production. In this system, lymphokine production depended on signals from TCRI;ICD3 and CD28 using concentrations of MAb that approximate physiological ligand. Mutations at Y-191 and M-194 markedly reduced (by at least 90%) the ability of anti-CD3 plus CD86 to induce IL-2 secretion and, concordantly, to recruit the kinase (Figures

3-5). A small but reproducible induction of IL-2 was observed in the absence of PI 3-kinase binding. In both Tcells and insect cells, the mutation of M-194 in CD28 resulted in a marked decrease in binding (Figures 5 and 6). Significantly, while the M194CD28C mutant failed to bind to PI 3-kinase, it retained an ability to bind to GRB-2 and ITK (Figure 6). CD26 recruitment of PI 3-kinase therefore provides a cosignal in the signaling cascade linked to IL-2 production. A less likely possibility is that another unidentified SH2 domain carrying protein with the same binding requirements as PI 3-kinase exists in T cells. Previous studies using the inhibitor of PI 3-kinase, wortmannin, have yielded conflicting results (Ward et al., 1995; Lu et al., 1995). This discrepancy underlines the importance of our mutational analysis. While the basis of this discrepancy is unclear, inhibition was observed in cultures stimulated by the combination of anti-CD3 plus CD80, but not with combinations of CD28 plus phorbol ester (Ward et al.,1995). This is consistent with our results showing a dependency of anti-CD3 plus CHO-CD86 cosignaling on associated PI 3-kinase. The insensitivity to combinations of anti-CD28 plus phorbol ester to the inhibitor may be related to pleiotropic effects of phorbol ester on additional signaling pathways. Under the circumstances in which wortmannin inhibited costimulation, it was unclear whether TCRgCD3 or CD28 signaling was the target of inhibition. In our study, the lossof IL-2 production with the MlwCD28C mutant occurred without any disruption of the TCRrJCD3 pathway, thereby directly implicating PI 3-kinase in the CD28 signaling arm of CD28-TCR stimulation. In our system, concentrations of wortmannin from 20-200 nM inhibited CD28-associated PI 3-kinase as well as free PI 3-kinase from cells treated with the reagent (Figure 78). The lack of complete inhibition may be due to incomplete uptake or sequestration of the kinase in intracellular vesicles, since the same concentrations of reagent completely inhibited lipid kinase activity when added to in vitro precipitates (data not shown). Significantly, concentrations as low as 20 nM induced apoptosis as detected by extensive DNA fragmentation (Figure 7C). Wortmannin was therefore not suitable for a study of the role of PI 3-kinase in CDP&induced IL-2 production in our costimulation assays. Wortmannin has also been reported to induce apoptosis in PC1 2 cells (Yao and Cooper, 1995). By what intracellular mechanism might PI 3-kinase regulate IL-2 production? PI 3-kinase has been reported to bind and activate ~21’~~ (Sjolander et al., 1991; RodriguezViciana et al., 1994; Hu et al., 1995). This could connect PI 3-kinase to a signaling pathway mediated by p21rBs and its regulation of MAPK and JNK. p21’@, MAPK, and JNK are activated by CD28 ligation (August and DuPont, 1994; Nunes et al., 1994; Su et al., 1994). JNK phosphorylates cJun, thereby increasing AP-1 transcriptional activity within the IL-2 promoter (Derijard et al., 1994). In agreement, oncogenic Ras induces IL-2 production (Rayter et al., 1992; Baldari et al., 1993). Further, AP-1 transcriptional activity in transgenic mice carrying a reporter gene requires CD28-mediated signals (Rincon and Flavell, 1994). In this context, it is intriguing that GRB-2/SOS, another regulator of Ras, also binds to CD28 (Schneider et

CDZ8Mediated 423

Signaling

al.,1 995a). CD28 therefore engages two intracellular proteins with the potential to activate the ~21’~) pathway. GRB-2 binding does not prevent PI 3-kinase binding, and vice versa (Schneider et al. 1995a). A role for PI 3-kinase in IL-2 induction does not exclude a role for GRB-21SOS. PI 3-kinase and GRB-2/SOS may function cooperatively, or they may depend on each other. Alternatively, GRB-PI SOS may be involved in other functions attributed to CO28. We are presently examining whether mutations that interfere with GRB-2 binding also block IL-2 production. In an alternate pathway, PI 3-kinase may produce D-3 phospholipids PI 3,4-P2 and PI 3,4,5-P3 and regulate protein kinase C c (PKCQ (Nakanishi et al., 1993). PI 3-kinase binding to the PDGF-R is also required for the activation of p70SBK (Chung et al., 1994). We and others previously showed that the p59vn and p56rck SH3 domains bind to PI 3-kinase, and may recruit the enzyme to the TCRLJCD3 and CD4 receptors (Prasad et al., 1993a, 1993b; Vogel and Fujita, 1993). PI 3-kinase also becomes activated upon binding to the SH3 domain (Pleiman et al., 1994). Further, both TCR and CD28 ligation independently generate D-3 phosphoinositides (Ward et al., 1995). p59m and ~56”~ and CD28 binding differs quantitatively, where the CD28 binds several-fold more enzyme than the CD4-lck and TCR/CD3-p59vn complexes. In this sense, CD28 may adjust the threshold of signaling by supplementing suboptimal signals generated by the TCRgCD3-CD4 complex (Rudd et al., 1994; Schneider et al., 1995c). At low antigen concentrations, this additive signal might provide sufficient additional signal to surpass a threshold requirement that distinquishes between responsiveness from nonresponsiveness. In addition to PI 3-kinase and GRB-2, ITK has been reported to bind CD28 (August et al., 1994). The molecular basis of this interaction is unclear. Although we have had difficulty observing ITK binding to CD28 in T cells (data not shown), we could demonstrate binding in insect cells. In this system, p561Ckand p59b” can regulate the binding of ITK, placing ITK downstream of the src kinases in a signaling cascade (Raab et al., 1995). Unlike PI 3-kinase and GRB-2, we found that the disruption of the pYMNM motif still allowed for the binding of ITK to CD28 (Figure 6) possibly implicating other phosphotyrosine sites such as Y-206, Y-209, or Y218. It remains to be determined whether other CD28-mediated functions such as the induction of CTLA4 and CD40L expression are linked to ITK, to PI 3-kinase and GRB-2-mediated signal transduction, or both. CTLA4 has been found to bind to PI 3-kinase by means of a pYVKM motif (Schneider et al., 1995b). While CTLA4 has the capacity to generate negative signals (Bluestone et al., 1994) under other circumstances, CTLA4 may cooperate with CD28 in the production of IL-2 (Linsley et al., 1992b). The finding that PI 3-kinase is a component in the regulation of IL-2 supports that notion that CTLA-4 may in certain situations also play a positive role in T cell growth. Experimental

Procedures

Cells, Reagents, and Antibodies The murine T cell hybridoma, DC27 (gift of Dr. Ft. Zamoyska,

Medical

Research Council, London) was cultured in RPM1 1640 medium supplemented with 5% (v/v) fetal bovine serum (FES, Intergen, New York) and 50 uM 2-mercaptoethanol. CHO and CHO-CD86, which express high levels of transfected murine CD86 (Chen et al., 1994a), were grown in DMEM medium with 5% FBS. Anti-murine CD3 (2Cil) was from American Type Culture Collection; anti-murine CD28 (37.51) was provided by Dr. J. Allison (University of California, Berkeley) (Harding et al., 1992); anti-murine CD66 (87-2) (2DiO) is as described (Chen et al., 1994b); anti-CD28 (L293) was from Becton Dickinson (San Jose, California); anti-HA (12CA5) was from the Harvard Cell Culture Facility (Boston); and anti-GRBP was from Transduction Laboratories. Wortmannin was purchased from Sigma. Other reagents and their sources are as described (Prasad et al., 1994). Transfections of Human CD28 The plasmids of human CD28 wild-type and Yigl and M184 mutants have been described before (Prasad et al., 1994). The inserts were resubcloned into the SRa expression vector (Takebe et al., 1988). Human CD28 plasmids were cotransfected by electroporation at 260 V and 1600 uF with pSVneo, which contains a neomycin resistance gene. Cells were selected with 1.5 mglml of G418 for 2 weeks, and cells from different populations were assayed for human CD28 expression by FACS. Cell Flow Cytometry Analysis Cytometric analysis was carried out as described (Prasad et al., 1994). In brief, cells (5 x 105) were stained with antibodies for 1 hr on ice, washed twice, and were then incubated with secondary fluorescein isothiocyanate (FITC)-labeled antibodies under the same conditions. Cytometric analysis was conducted using an EPICS cell sorter (Coulter Corporation). Biotinylation and lmmunopreclpitation Analysis of Cell Surface Proteins Cells were washed with Biwa’s buffer (1 mM MgCk and 0.1 mM CaCk in phosphate-buffered saline [pH 7.21). Cells were then incubated at room temperature with 150 mM NaCI, 10 mM sodium borate, and 66 Kg/ml Sulfo-NHS-biotine (pH 8.8) (Pierce) for 20 min, followed by washing with 25 mM L-lysine in Biwa’s buffer. Cells were lysed in 1% NP40 (VN) lysis buffer with 1 mM PMSF and subjected to immunoprecipitation (Prasad et al., 1994). Biotinylated CD28 were detected with streptavidin coupled with horseradish peroxidase and chemiluminescence. CD28 Cross-Linking and lmmunopreclpitation CHO and CHO-CD86 cells were grown on 150 mm plates to confluence, washed once, and 150 x 1 d of DC27 transfectants were layered on top of these cells in 5 ml media. These plates were then incubated at 37OC for the indicated time before harvesting. Cells were washed oncewith cold phosphate-buffered saline, followed by lysis in 1% NP40 plus vanadate and PMSF. Cell lysates were precleared by incubation with pansorbin cells (Calbiochem) on ice for 30 min, followed by immunoprecipitation with pre-bound antibody-protein G/A sepharose beads for 2 hr. The beads were washed with lysis buffer three times before lipid kinase assay or electrophoresis was conducted. Lipid Klnase Assay and Western Blot Analysis For lipid kinase assay, immunoprecipitates were further washed with LICI and TNE (10 mM Tris [pH 7.51, 150 mM NaCI, 1 mM EGTA), followed by lipid kinase reaction and separation on TLC plates as described (Prasad et al.. 1994). For Western blot, immunoprecipitates were boiled in loading buffer, and proteins were separated on SDSPAGE and transfered onto nitrocellulose membrane. Following blotting with the first antibody, reactivities were detected with horseradish peroxidase-conjugated antibodies and chemiluminescence as described (Amersham). The relative intensity of each band on films was determined by LKB UltroScan. Expresslon and lmmunoprecipltation of Human CD28 in Sf21 Calls Expression in Sf21 cells by baculovirus system was performed as described (Raab et al., 1994). In brief, human CD28 (gifl of Dr. B. Seed, Massachusetts General Hospital, Boston), Lck, GRBP, and

Immunity 424

hemaglutinin-tagged ITK (HA-tagged ITK) (gift of Dr. L. Berg, Harvard University) constructs were subcloned into the pV11393 vector, and recombinant viruses were obtained from the S. frugiperda cell line by cotransfection with wild-type baculoviral DNA. Following infection of Sf21 cells with different combinations of recombinant viruses, human CD28 were immunoprecipitated from cell lysates with L293, and CD28associated proteins were analyzed by Western blot hybridization. Costimulation of T Cell Transfectants CHO or CHO-CD86 cells were plated at 5 x lo3 /well in 96-well plates and cultured overnight. Cells were washed twice and incubated with DC27 transfectants at 1 x 105/well plus 2Cll at indicated concentrations. The final volume was 200 ul/well. Each condition was assayed as a triplicate. Cells were then cultured for 24 hr before the supernatants were removed for the measurement of IL-2. Cell growth at this point was assayed by 1 hr pulse with 1 pCi of PHjthymidine, followed by harvesting and counting in scintillation fluid. IL-2 Assay The supernatants derived from cell cultures were stored at -70°C and thawed before IL-2 assay. CTLL-2 cells were grown in the presence of recombinant human 11-2. They were washed three times and plated at 2500 cells/well in a 96well plate. A series of P-fold diluted supernatants were added to CTLL-2 cells. Cells were cultured for 24 hr followed by an 18 hr pulsing with 1 t&i of PH)thymidine before harvesting and counting. The amount of IL-2 in these supernatants was determined by comparing with the standard curve established with recombinant IL-2. The half-maximum of stimulation is defined as 100 U/ml as described (Gillis et al., 1978). Trypan Blue Staining and DNA Fragmentation Cell viability was determined with 0.04% trypan blue staining and counting of triplicates under microscopy. For DNA fragmentation, cells were incubated with 1% NP40 plus PMSF lysis buffer for 30 min. Soluble materials were collected after 10 min centrifugation and subjected to phenol and chloroform extraction, followed by ethanol precipitation. The resulting DNA solutions were incubated with RNAse at 37°C for 30 min to remove RNA, and DNA was then resolved on 1.2% agarose gel. The gel was stained with ethidium bromide (Sigma) and photographed. Acknowledgments C. E. Ft. is a Scholar of the Leukemia Society of America. supported by the Mildred-Scheel Stiftung, Federal Republic many. D. C. is supported by funds from Coulter Immunology, We thank Dr. Stuart Schlossman for helpful discussions, Received

May 16, 1995; revised

August

H. S is of GerFlorida.

28, 1995

References Arcaro, A., and Wymann, M.P. (1993). Wortmannin is a potent phosphatidylinositol 3- kinase inhibitor: the role of phosphatidylinositol 3,4,5-trisphosphate in neutrophil responses. Biochem. J. 296, 297301. August, A., and DuPont, with phosphatidylinositol

6. (1994). CD28 of T lymphocytes associates 3-kinase. Int. Immunol. 6, 769-774.

August, A., Gibson, S., Kawakami, Y., Kawakami, T., Mills, G.B., and DuPont, B. (1994). CD28 is associated with and induces the immediate tyrosine phosphorylation and activation of the Tee family kinase ITK/ EMT in the human Jurkat leukemic T cell line. Proc. Natl. Acad. Sci. USA 91, 9347-9351. Azuma, M.. Ito, D., Yagita, H., Okumura, K.. Phillips, J.H.. Lanier, L.L.. and Somoza. C. (1993). B70antigen isasecond IigandforCTLA4 and CD28. Nature 366,76-79.

restores immunogenicity of tumor cells expressing truncated major histocompatibility complex class II molecules. Proc. Natl. Acad. Sci. USA 90,5667-5690. Bonfini, L.. Karlovich, The Son of sevenless ence 255, 603-606.

CA., Dasgupta, C., and Banerjee, U. (1992). gene product: a putative activator of Ras. Sci-

Boussiotis, V.A., Freeman, G.J., Gribben, J.G., Daley, J., Gray, G., and Nadler, L.M. (1993). Activated human B lymphocytes express three CTLA-4 counterreceptors that costimulate T cell activation. Proc. Natl. Acad. Sci. USA 90,11059-l 1063. Boussiotis, V.A., Barber, D.L., Nakari, T., Freeman, G.J., Gribben, J.G., Bernstein,G.M., D’Andrea,A.D., Ritz, J.,andNadler, L.M.(1994). Prevention of T cell anergy by signaling through the yc chain of the IL-2 receptor. Science 299, 1039-l 042. Carpenter, CL., Duckworth, B.C., Auger, K.R., Cohen, B., Schaffhausen, B.S., and Cantley, L.C. (1990). Purification and characterization of phosphoinositide 5kinase from rat liver. J. Biol. Chem. 265, 19704-19711. Chen, L., Ashe, S., Brady, W.A., Hellstroem, I., Hellstroem, K.E., Ledbetter, J.A., McGowan, P., and Linsley, P.S. (1992). Costimulation of antitumor immunity by the 87 counter-receptor for the T lymphocyte molecules CD26 and CTtA4. Cell 77, 1093-1102. Chen, C., Gault, A., Shen, L., and Nabavi, N. (1994a). Molecular cloning and expression of early T cell costimulatory molecule-l and its characterization as 87-2 molecule. J. Immunol. 752, 4929-4936. Chen, C., Denise, D.A., Gault, A., Connaughton, SE., Powers, G.D., Godfrey, D.I., and Nabavi, N. (1994b). Monoclonal antibody 2010 recognizes a novel T cell costimulatory molecule on activated murine B lymphocytes. J. Immunol. 752, 2105-2114. Chung, J., Grammer, T.C., Lemon, K.P., Kazlauskas, A., and Blenis, J. (1994). PDGF- and insulin-dependent pp70 S6k activation mediated by phosphatidylinositol-3-OH kinase. Nature 370, 71-75. Derijard, M., and light and domain.

B., Hibi. M., Wu, I.-H., Barrett, T., Su, B.. Deng, T., Karin, Davis, R.J. (1994). JNKl: a protein kinase stimulated by UV Ha-Ras that binds and phosphorylates the c&n activation Cell 76, 1025-1037.

Dhand, R., Hara, K., Hiles, I., Bax, B., Gout, I., Panayotou, G., Fry, M.J., Yonezawa, K., Kasuga, M., and Waterfield. M.D. (1994a). PI 3-kinase: structural and functional analysis of intersubunit interactions, EMBO J. 73. 51 l-521. Dhand, R., Hiles, I., Panayotou, G., Roche, S., Fry, M.J., Gout, I., Totty, N.F.,Truong, O.,Vicendo, P., Yonezawa, K.. Kasuga, M., Courtneidge, S.A., and Waterfield. M.D. (1994b). PI 3-kinase is a dual specificity enzyme: autoregulation by an intrinsic protein-serine kinase activity. EMBO J. 73, 522-533. Freeman,G.J.,Gribben, J.G., Boussiotis, V.A., Ng, J.W., RestivoV.A., Jr., Lombard, L.A., Gray, G.S.. and Nadler, L.M. (1993a). Cloning of 87-2: a CTLA-4 counter-receptor that costimulates human T cell proliferation. Science 262, 909-911. Freeman, G.J., Borriello. F.. Hodes. R.J.. Reiser, H., Hathcock, KS., Laszlo, G., McKnight. A.J., Kim, J., Du, L., Lombard, D.B., Gray, G.S., Nadler, L.M.. and Sharpe, A.H. (1993b). Uncovering of functional alternative CTLA-4 counter-receptor in B7-deficient mice. Science 262, 907-909. Freeman, G.J.. Boussiotis, V.A., Anumanthan, A., Bernstein, G.M., Ke. X.-Y., Rennert, P.D., Gray, G.S., Gribben, J.G., and Nadler, L.M. (1995). 87-l and 87-2 do not deliver identical costimulatory signals, since 87-2 but not 87-l preferentially costimulates the initial production of IL-4. Immunity 2. 523-532. Fantl, W.J., Escobedo, J.A., Martin, G.A., Turck, C.W., del Rosario, M.. McCormick, F., and Williams, L.T. (1992). Distinct phosphotyrosines on a growth factor receptor bind to specific molecules that mediate different signaling pathways. Cell 69, 413-423.

Baldari, C.T., Heguy, A., and Telford, J.L. (1993). ras protein activity is essential for T cell receptor signal transduction. J. Biol. Chem. 268, 2693-2698.

Fraser, J.D., Irving, B.A.,Crabtree,G.R..and Weiss, A.(1991). Regulation of interleukin-2 enhancer activity by the T cell accessory molecule CD28. Science 257, 313-316.

Baskar, S., Ostrand-Rosenberg. S., Nabavi, N.. Nadler, L.M., Freeman, G.J.. and Glimcher. L.H. (1993). Constitutive expression of 87

Gabert, J., Langlet, C.. Zamoyska, R., Parnes, A.-M., and Mallissen, B. (1987). Reconstitutionof

J.R., Schmitt-Verhulst, MHCclass 1 specific-

CDZB-Mediated 425

ity by transfer 554.

Signaling

of the T cell receptor

and Lyt-2 genes,

Cell 50, 545-

Gillis, S., Frem. M.M., Ou, W., and Smith, K.A. (1978). T cell growth factor: parameters of production and a quantitative microassay for activity. J. Immunol. 120, 2027-2032. Green, J.M., Noel, P.J., Sperling,A.I., Walunas,T.L.,Gray, S.G., Bluestone, J.A., and Thompson, C.B. (1994). Absence of B7-dependent responses in CDPB-deficient mice. Immunity 7, 501-508. Guerder, S., Picarella, D.E., Linsley, P.S., and Flavell, R.A. (1994). Costimulator 87-l confers antigen-presenting-cell function to parenchymal tissue and in conjunction with tumor necrosis factor alpha leads to autoimmunity in transgenic mice. Proc. Natl. Acad. Sci. USA 91, 5138-5142. Harding, F.A., McArthur, J.G., Gross, J.A., Raulet, D.H., and Allison, J.P. (1992). CD28-mediated signalling co-stimulates murine T cells and prevents induction of anergy in T cell clones. Nature 356, 607609. Hathcock, KS., Laszlo, G., Dickler, H.B., Bradshaw, J., Linsley, P., and Hodes, R.J. (1993). Identification of an alternative CTLA-4 ligand costimulatory for T cell activation. Science 262, 905-907. Hathcock, KS., Laszlo, G., Pucillo, C., Linsley, P., and Hodes, R.J. (1994). Comparative analysis of 87-l and 87-2 costimulatory ligands: expression and function. J. Exp. Med. 160, 631-640. Heyeck, S.D., and Berg, L.J. (1993). Developmental regulation of a murine T cell-specific tyrosine kinase gene, Tsk. Proc. Natl. Acad. Sci. USA 90, 669-673. Hu, Q., Klippel, A., Muslin, A.J., Fantl, W.J., and Williams, L.T. (1995). Ras-dependent induction of cellular responses by constitutively active phosphatidylinositol-3 kinase. Science 266, 100-l 02. Hutchcroft, J.E., and Bierer, B.E. (1994). Activation-dependent phosphorylation of the T-lymphocyte surface receptor CD28 and associated proteins. Proc. Natl. Acad. Sci. USA 91, 3260-3264. Jenkins, M.K., and Johnson, J.G. (1993). Molecules co-stimulation. Curr. Opin. Immunol. 5, 361-367.

involved

in T cell

similar avidities but distinct Immunity 7, 793-801.

kinetics

to CD28

and CTLA-4

receptors.

Lu, Y., Phillips, C.A., and Trevillyan, J.M. (1995). Phosphatidylinositol 5kinase activity, is not essential for CD28 costimulatory activity in Jurkat T cells: studies with selective inhibitor, wortmannin. Eur. J. Immunol. 25, 533-537. Nakanishi, H., Brewer, K.A., and Exton, J.H. (1993). Activation zeta isozyme of protein kinase C by phosphatidylinositol triphosphate. J. Biol. Chem. 268, 13-16.

of the 3,4,5-

Nunes, J.A., Collette, Y., Truneh, A., Olive, D., and Cantrell, D.A. (1994). The role of ~21’“” in CD28 signal transduction: triggering of CD28 with antibodies, but not the ligand 87-1, activates p21”*. J. Exp. Med. 180, 1067-1076. Pages, F., Ragueneau. M., Rottapel, R., Truneh, A., Nunes, J., Imbert, J., and Olive, D. (1994). Binding of phosphatidyl-3-OH kinase to CD28 is required for T-cell signalling. Nature 369, 327-329. Parker, 3-kinase:

P.J., and Waterfield, M.D. a novel effector. Cell Growth

(1992). Phosphatidylinositol Differ. 3, 747-752.

Pleiman, C.M., Hertz, W.M., and Cambier, phosphatidylinositol-3’ kinase by Src-family p85 subunit. Science 263, 1609-1612.

J.C. (1994). Activation of kinase SH3 binding to the

Prasad, K.V.S., Kapeller, R., Janssen, O., Repke, H., Duke-Cohan, J.S., Cantley, L.C., and Rudd, C.E. (1993a). Phosphatidylinositol (PI) 3-kinase and PI 4-kinase binding to the CD4-~56’~~ complex: the ~56”~ SH3 domain binds to PI 3-kinase but not PI 4-kinase. Mol. Cell. Biol. 13, 7708-7717. Prasad, K.V.S., Janssen, O., Kapeller, R., Raab. M., Cantley, L.C., and Rudd, C.E. (1993b). Src-homology 3 domain of protein kinase p5gsn mediates binding to phosphatidylinositol 3-kinase in T cells. Proc. Natl. Acad. Sci. USA 90, 7366-7370. Prasad, K.V.S.,Cai.Y.-C., Raab, M., Duckworth, B., Cantley. L., Shoelson, SE., and Rudd. C.E. (1994). T cell antigen CD28 interacts with the lipid kinase phosphatidylinositol3-kinase by a cytoplasmic Tyr(P)Met-Xaa-Met motif. Proc. Natl. Acad. Sci. USA 91, 2834-2838.

Jenkins, M.K., Ashwell, J.D.. and Schwartz, R.H. (1986). Comparative analysis of 87-2 costimulatory ligands: expression and function. J. Immunol. 140, 3324-3330.

Raab. M., Yamamoto, M., and Rudd, CD5 acts as a receptor and substrate ~56”‘. Mol. Cell. Biol. 74, 2862-2870.

Kuchroo, V.K.. Das, M.P., Brown, J.A., Ranger, A.M., Zamvil, S.S., Sobel, R.A., Weiner, H.L., Nabavi, N., and Glimcher, L.H. (1995). 87-l and 87-2 costimulatory molecules activate differentially the Thl/ThP developmental pathways: application to autoimmune disease therapy. Cell 80, 707-718.

Raab, M., Cai, Y.-C., Dunnell, SC., Heyeck, S.D., Berg, L.J., and Rudd, C.E. (1995). ~56”~ and p59b” regulateCD28 recruitment of phosphatidylinositol 3 kinase, growth factor receptor-bound GRB-2 and T cell-specific protein-tyrosine kinase ITK: implications for costimulation. Proc. Natl. Acad. Sci. USA 92, 8891-8695.

Koulova, B.L.. Clark, E.A., Shu, G., and DuPont, B. (1991). TheCD28 ligand 871861 provides costimulatory signal for alloactivation of CD4’ T cells. J. Exp. Med. 173, 759-762.

Rayter, S.I., Woodrow, M., Lucas, S.C., Cantrell, D.A., and Downward, J. (1992). ~21”~ mediates control of the IL-2 promoter function in T cell activation. EMBO J. 77, 4549-4556.

Lenschow, D.J., Zeng, Y., Thistlewaite, J.R., Montag, A., Brady, W., Gibson, M.G., Linsley, P.S., Bluestone, J.A. (1992). Long-term survival of Xenogeneic pancreatic islet grafts induced by CTLA4lg. Science 257, 789-792.

Rincon, M., and Flavell, R.A. (1994). AP-1 transcriptional quires both T cell receptor-mediated and co-stimulatory primary T lymphocytes. EMBO J. 73, 4370-4381.

Lenschow, D.J., Ho, S.C., Sattar. H., Rhee, L., Gray, G., Nabavi, N., Herold, K.C., and Bluestone. J.A. (1995). Differential effects of anti87-2 monoclonal antibody treatment on the development of diabetes in the nonobese diabetic mouse. J. Exp. Med. 187, 1145-1155. Lindstein, T., June, C.H., Ledbetter, J.A., Stella, G., and Thompson, C.B. (1989). Regulation of lymphokine messenger RNA stability by a surface-mediated T cell activation pathway. Science 244, 339-343.

C.E. (1994). The T cell antigen for the protein-tyrosine kinase

activity resignals in

Rodriguez-Viciana, P., Warne, P.H., Dhand, R., Vanhaesebroeck, B., Gout, I., Fry, M.J., Waterfield, M.D., and Downward, J. (1994). Phosphatidyl-3-OH kinase as a direct target of Ras. Nature 370, 527-532. Rudd, C.E., Janssen, O., Cai, Y.-C., da Silva, A.J., Raab, M.. and Prasad, K.V.S. (1994). Two-stepTCRuCD3-CD4 and CD28 signaling in Tcells: SH21SH3domains protein-tyrosineand lipid kinases. Immunol. Today 75, 225-234.

Linsley, P.S., and Ledbetter. J.A. (1993). The role of theCD28 receptor during T cell responses to antigen. Annu. Rev. Immunol. 11,191-212.

Schneider, H., Cai, Y.-C., Prasad, K.V.S., Shoelson, SE., and Rudd, C.E. (1995a). T cell antigen CD28 binds to the GRB-21SOS complex, regulators of ~21’“. Eur. J. Immunol. 25, 1044-1050.

Linsley, P.S., Wallace, P.M., Johnson, J., Gibson, M.G., Greene, J.L.. Ledbetter, J.A., Singh. C.. and Tepper, M.A. (1992a). Immunosupression in vivo by a soluble form of the CTLA-4 T cell activation molecule. Science 257, 792-795.

Schneider, H., Prasad, K.V.S., Shoelson, S.E., and Rudd, (I 995b). CTLA-4 binding to the lipid kinase phosphatiylinositol3-kinase in T cells. J. Exp. Med. 181, 351-355.

Linsley. P.S., Greene, J.A.L., Tan, P., Bradshaw, J., Ledbetter, J.A., Anasettei, C., and Damle, N.K. (1992b). Coexpression and functional cooperation of CTLA-4 and CD26 on activated T lymphocytes. J. Exp. Med. 776, 1595-1604. Linsley. P.S., Greene, and Peach, R. (1994).

J.L., Brady, Human 87-l

W., Bajorath, J., Ledbetter. J.A., (CD80) and 87-2 (CD86) bind with

C.E.

Schneider, H., Cai, Y.-C., Cefai, D., Raab, M., and Rudd, C.E. (199%). Mechanisms of CD28 signalling. Res. Immunol. 746, 149-154. Schwartz, R.H. (1992). Costimulation of T lymphocytes: the role of CD28, CTLA4, and 871881 in interleukin-2 production and immunotherapy. Cell 77, 1065-1068. Shahinian,

A., Pfeffer.

K., Lee, K.P.,

Kuendig,

T.M.,

Kishihara,

K..

Immunity 426

Wakeham, A., Kawai, K., Ohashi, P.S., Thompson, C.B., and Mak, T.W. (1993). Differential T cell costimulatory requirements in CD26deficient mice. Science 267, 609-612. Simon, M.A., Dodson, G.S., and Rubin, G.M. (1993). An SH3-SH2/ SH3 protein is required for p2lsasl activation and binds to sevenless and SOS proteins in vitro. Cell 73, 169-177. Sjolander, A., Yamamoto, K., Huber, B.E., and Papetina, Association of ~21” with phosphatidylinositol 3-kinase. Acad. Sci. USA 88, 7906-7912.

E.G. (1991). Proc. Natl.

Su, B., Jacinto, E., Hibi, M., Kallunki, T., Karin, M., and Ben-Neriah, Y. (1994). JNK is involved in signal integration during costimulation of T lymphocytes. Cell 77, 727-736. Sykulev, Y., Brunmark, M., Peterson, P.A., and tween antigen-specific allogeneic and syngeneic teins Proc. Natl. Acad.

A., Tsomides, T.J., Kageyama, S., Jackson, Eisen, H.N. (1994). High-affinity reactions beT cell receptors and peptides associated with major histocompatibility complex class I proSci. USA 91, 11487-l 1491.

Takebe, Y., Seiki, M., Fujisawa, J.-l., Hoy, P., Yokota, K., Arai, K.-l., Yoshida, M., and Arai, N. (1988). SRa promoter: an efficient andversatile mammalian cDNA expression system composed of the simian virus 40 early promoter and the R-U5 segment of human T cell leukemia virus type 1 long terminal repeat. Mol. Cell. Biol. 8, 466-472. Townsend, SE., and Allison, J.P. (1993). Tumor rejection after direct costimulation of CD8’ T cells by B7-transfected melanoma cells. Science 259, 368-370. Truitt, K.E., Hicks, CM., and Imboden, J.B. (1994). Stimulation CD28 triggers an association between CD28 and phosphatidylinositol3-kinase in Jurkat T cells. J. Exp. Med. 179, 1071-1076.

of

Ucker, D.S., Ashwell, J.D., and Nickas, G. (1989). Activation-driven T cell death: I. Requirements for de novo transcription and translation and association with genome fragmentation. J. Immunol. 743, 34613469. Valius, M., and Kazlauskas, A. (1993). Phospholipase C?l and phosphatidylinositol3 kinase are the downstream mediators of the PDGF receptor’s mitogenic signal. Cell 73, 321-334. Vogel, L.B., and Fujita, D.J. (1993). The SH3 domain of ~58~~ is involved in binding to phosphatidylinositol3’-kinase from T lymphocytes. Mol. Cell. Biol. 73, 7408-7417. Walunas, T.L., Lenschow, D.J., Bakker, C.Y.. Linsley, P.S., Freeman, G.J., Green, J.M., Thompson, C.B., and Bluestone, J.A. (1994). CTLA-4 can function as a negative regulator of T cell activation. Immunity 7, 405-413. Ward, S.G., Wilson, A., Turner, L., Westwick, J., and Sansom, D.M. (1995). Inhibition of CD28-mediated T cell costimulation by the phosphoinositide 3-kinase inhibitor wortmannin. Eur. J. Immunol. 25, 526532. Yano, H., Nakanishi, S., Kimura, K.. Hanai, N., Saitoh, Y., Fukui, Y., Nonomura, Y., and Matsuda, Y. (1993). Inhibition of histamine secretion by wortmannin through the blockade of phosphatidylinositol Skinase in RBL-2H3 cells. J. Biol. Chem. 268, 25846-25856. Yao, A., and Cooper, G.M. (1995). Requirement for phosphatidylinositol-3 kinase in the prevention of apoptosis by nerve growth factor. Science 267.2003-2006.