TcR and TcR-CD28 Engagement of Protein Kinase B (PKB/AKT) and Glycogen Synthase Kinase-3 (GSK-3) Operates Independently of Guanine Nucleotide Exchange Factor VAV-1*

TcRζ/CD3 and TcRζ/CD3-CD28 signaling requires the guanine nucleotide exchange factor (GEF) Vav-1 as well as the activation of phosphatidylinositol 3-kinase, protein kinase B (PKB/AKT), and its inactivation of glycogen synthase kinase-3 (GSK-3). Whether these two pathways are connected or operate independently of each other in T-cells has been unclear. Here, we report that anti-CD3 and anti-CD3/CD28 can induce PKB and GSK-3α phosphorylation in the Vav-1–/– Jurkat cell line J. Vav.1 and in primary CD4-positive Vav-1–/– T-cells. Reduced GSK-3α phosphorylation was observed in Vav-1,2,3–/– T-cells together with a complete loss of FOXO1 phosphorylation. Furthermore, PKB and GSK-3 phosphorylation was unperturbed in the presence of GEF-inactive Vav-1 that inhibited interleukin-2 gene activation and a form of Src homology 2 domain-containing lymphocytic protein of 76-kDa (SLP-76) that is defective in binding to Vav-1. The pathway also was intact under conditions of c-Jun N-terminal kinase (JNK) inhibition and disruption of the actin cytoskeleton by cytochalasin D. Both events are down-stream targets of Vav-1. Overall, our findings indicate that the TcR and TcR-CD28 driven PKB-GSK-3 pathway can operate independently of Vav-1 in T-cells.

T-cell activation is induced by ligation of the antigen-receptor (TcR/CD3) as well as co-receptors such as CD28. TcR/ CD3 and CD4/CD8-lck initiate tyrosine phosphorylation, while TcR/CD3 and CD28 induce the production of D-3 lipids (1, 2). CD28 co-signals are needed for optimal cytokine production, proliferation, and effector function (3,4). CD28-deficient mice have reduced responses to antigen, highlighting the capacity of CD28 to lower the threshold of signaling (5). Primary responses exhibit more of a dependence on CD28 than do secondary responses, and the co-receptor can influence the differentiation of T helper 2 (Th2) versus T helper 1 (Th1) cell, increase cell survival, and prevent the induction of T-cell anergy (3,6,7).
In addition to activating the PI3K pathway, TcR/CD3 and CD28 signaling is dependent on the guanine nucleotide exchange factor, Vav-1. It has a calponin homology domain, an acidic motif, a zinc finger-like region, two SH3 domains, and a SH2 domain (31)(32)(33)(34). The SH2 domain of Vav-1 binds tyrosine residues within the adaptor SH2 domain-containing lymphocytic protein of 76 kDa (SLP-76) (35,36). These residues each reside within a YESP motif, both of which are phosphorylated upon receptor ligation. The Dbl homology (DH) domain has * This work was supported by a grant from the Wellcome Trust, London. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1  GEF activity for the activation of the small GTPases Rac1 and Cdc42 (37). Vav-1-deficient T-cells show defects in TcR capping and the induction of cytokine production (38,39). Phosphorylation of Vav-1 by CD28 depends on residues 173-181 of the receptor (15,40). CD28 has been reported to boost TcR signaling via Vav-1-SLP-76 cooperation (41). Furthermore, coexpression can drive NFAT translocation into the nucleus of COS cells in response to CD28 ligation (36). In addition, CD28 engagement can activate IB kinase (IKK) complex and NF-B activation in a Vav-1 dependent fashion (42). In keeping this, confocal microscopy has shown that endogenous VAV-1 and IKK␣ co-localize in response to CD28 stimulation (42). Vav-1 has been reported to cooperate with protein kinase C theta to activate c-Jun kinase (JNK) (43), while Vav-1-deficient Jurkat cells (termed J.Vav1) show defects in Ca 2ϩ mobilization as well as in the activation of JNK and transcription factors needed for interleukin-2 transcription (44). Given that TcR and TcR-CD28 generate signals that depend on Vav-1 and PI3K, an important question is whether these pathways are interconnected or operate independently of each other. Observations have been mixed in other systems. Vav-1 GEF activity has been reported to be dependent on inositol lipids due to pleckstrin homology domain binding and localization (45). By contrast, inhibition of PI3K does not inhibit Vav-3 phosphorylation in Vav-3-deficient cells (46). Other studies have reported a partial or full role for Vav-1 in activating PI3K (47)(48)(49). BcR activation of protein kinase B phosphorylation occurs normally in Vav-1-deficient B-cells (49).
In this study, we report that anti-CD3 and anti-CD3/CD28 can induce PKB and GSK-3 phosphorylation in the Vav-1 Ϫ/Ϫ Jurkat cell line J.Vav.1 and in primary CD4-positive Vav-1 Ϫ/Ϫ T-cells. Reduced but significant GSK-3 phosphorylation was also observed in Vav-1,2,3 Ϫ/Ϫ T-cells, despite a complete loss of FOXO1 phosphorylation. Furthermore, PKB and GSK-3 phosphorylation was unperturbed in the presence of a GEFinactive form of Vav-1, a mutant that markedly inhibits IL-2 gene activation. Wild-type levels of phosphorylation were also observed in the presence of a form of SLP-76 that is defective in Vav-1 binding and in cells treated with cytochalasin D to disrupt the cytoskeleton. These findings indicate that TcR and TcR-CD28 can induce PKB/GSK-3 signaling independently of Vav-1 in T-cells.
Luciferase Assay-For the luciferase assays, 0.5 ϫ 10 6 Jurkat cells were incubated in 100 l of RPMI 1640 medium containing 5% fetal calf serum plus the appropriate antibodies for 6 h at 37°C in a 96-well plate. Cells were lysed, and luciferase activity was determined using a Luminat LB9507 luminometer (EG&G Berthold) and the luciferase assay system protocol from Promega.
Anti-CD3 Surface Clustering-Jurkat cells were either untreated or preincubated with the indicated concentration of cytochalasin D for 1 h. Cells were then resuspended in 1 ml cold FACS buffer (1% bovine serum albumin and 0.01% sodium azide in phosphate-buffered saline, pH 7.0) and incubated with 2 g/ml anti-CD3 (OKT3) for 30 min at 4°C.  OCTOBER 27, 2006 • VOLUME 281 • NUMBER 43

JOURNAL OF BIOLOGICAL CHEMISTRY 32387
Cells were then washed twice with cold (0 h samples) or prewarmed (1-h samples) FACS buffer and FITC-conjugated anti-mouse IgG (Sigma) was added, followed by incubation immediately on ice or at 37°C for 1 h. Cold FACS buffer was added to terminate the stimulations, and cells were fixed with 2% paraformaldehyde and then mounted on coverslips. TcR distribution was visualized by fluorescence microscopy. At least 200 T-cells were counted for TcR cap formation in each experiment.
F-actin Content-Jurkat cells were treated as above except R␣M was used to cross-link anti-CD3 (OKT3) for 1 h at 37°C. Cold medium was added to terminate the stimulations, and paraformaldehyde-fixed cells were permeabilized with 0.03% saponin/FACS buffer followed by 0.3% saponin/FACS buffer. F-actin content was quantified by staining with FITC-conjugated phalloidin (Sigma) and analyzed using a FACSCalibur (BD Biosciences).
To assess whether Vav-1 expression was required for PKB and GSK-3␣ phosphorylation, blotting was conducted in the Vav-1-deficient Jurkat cell line J.Vav.1 (Fig. 1) (44). Anti-CD3, -CD28, and -CD3/ CD28 induced levels of PKB or GSK-3␣ phosphorylation in Vav-1 Ϫ/Ϫ cells that were comparable with that observed in WT cells (upper and middle panels, lanes  17-32 versus lanes 1-16). Densitometric readings of the phosphorylated bands documented similar levels of increased phosphorylation (histograms in the lower panels). The absence of Vav-1 was confirmed by anti-Vav-1 blotting (lowest panel, lanes 17-32 versus 1-16). These data show that anti-CD3 and anti-CD3/CD28 induction of PKB and GSK-3␣ phosphorylation in T-cells can occur independently of Vav-1 expression. The independence of PKB/GSK-3 phosphorylation on Vav-1 expression contrasts with the previously reported defects in Ca 2ϩ and JNK signaling in the same cells (44).
Despite the absence of an effect on the PKB-GSK-3 pathway, as described by others (31,38), expression of L213Q completely inhibited anti-CD3, anti-CD28, and anti-CD3/anti-CD28-induced IL-2 transcription (Fig. 3B). This confirmed the inhibitory effect of the mutant on T-cell function. These observations indicate that the TcR and TcR-CD28-PKB-GSK-3 pathway is intact in the presence of a Vav-1 DH domain mutant that can potently inhibit IL-2 transcription in T-cells. This indicates that the inhibitory effect of the DH domain mutant on IL-2 transcription occurs independently of the PKB-GSK-3 pathway.
Although the above data showing an independence of anti-CD3 and anti-CD3/CD28 up-regulation of PKB-GSK-3␣ from Vav-1, we also assessed whether the pathway could operate without the activation of downstream targets of Vav-1. Vav-1 can regulate JNK activation and actin cytoskeleton remodelling (31,32). While the regulation of these events could have branched prior to the regulation of PKB-GSK-3 on Vav-1, it was of interest to assess whether the pathway could occur in the absence of these fundamental events. To assess the role of JNK, cells were treated with the inhibitor SP600125 at concentrations that have previously been shown to inhibit JNK activity (53). In this instance, we limited ourselves to an examination of GSK-3 phosphorylation (Fig. 5) A similar level of normal GSK-3␣ phosphorylation was observed in cells that had been treated with cytochalasin D (cyt.D) (Fig. 6). Treatment resulted in a shift of F-actin staining (i.e. reduced staining) indicative of its disruptive effect on the cytoskeleton (52) (Fig. 6A). In addition, anti-CD3-induced clustering on the cell surface was affected by cytochalasin D treatment (Fig. 6B, compare upper and lower images and histogram). Despite this, anti-CD3/CD28 continued to induce normal levels of GSK-3␣ phosphorylation over a range of cytochalasin D concentrations (Fig. 6C). In fact, in three experiments a slight increase in phosphorylation was observed at the higher concentration of the drug. These data document the unusual observation that neither the inhibition of JNK nor the disruption of the cytoskeleton with cytochalasin D interfered with the TcR/ CD28 induction of GSK-3␣ phosphorylation.

DISCUSSION
The question of the relationship between the Vav-1-and PI3K-mediated pathways has been the topic of mixed observations. Vav-1 GEF activity has been reported to depend on inositol lipids (45), while others have found a role for Vav-1 in activating PI3K (46 -49). By contrast, it has been reported that the inhibition of PI3K does not inhibit Vav-3 phosphorylation in Vav-3-deficient cells (47), while BcR activation of PKB occurs normally in Vav-1-deficient B-cells (50). With this background, it was important to establish how these pathways interrelate in T-cells. This is especially important in the context of CD28 co-signaling since the receptor recruits and activates PI3K and is also influenced by Vav-1 (12-21, 30, 36, 41).
Our findings show that anti-CD3 and CD3/CD28 induced wild-type levels of PKB and GSK-3␣ phosphorylation in Vav-1 Ϫ/Ϫ Jurkat cells and primary CD4-positive T-cells ( Figs. 1 and 2). This indicates that Vav-1 is not specifically required for the activation of the PKB-GSK-3 pathway in T-cells. Similarly, wildtype levels of phosphorylation were observed in the presence of a Vav-1 DH domain mutant that is well established in its ability to potently inhibit IL-2 transcription (Fig. 3). This indicates that the DH domain has an involvement in TcR driven IL-2 transcription that is independent of the PKB- Previous studies have shown that Vav-1 Ϫ/Ϫ cells showed defects in TcR clustering, sustained Ca 2ϩ mobilization, JNK activation, and IL-2 production (38,39,34,44). These defects were evident despite the expression of other Vav isoforms indicating that Vav-1 can play a specialized or dominant role in regulating these events. This differs from our findings with the PKB-GSK-3 pathway. Unlike in the case of Vav-1 Ϫ/Ϫ T-cells, Vav-1,2,3 Ϫ/Ϫ cells showed a partial loss (i.e. Ͻ50%) of GSK-3 phosphorylation. This indicates that the involvement of the Vav family in a portion of thesignalingthatleadstoGSK-3phosphorylation. In this manner, other Vav isoforms (i.e. Vav-2,3) may substitute for Vav-1 in the Vav1 Ϫ/Ϫ T-cells to allow for normal PKB and GSK-3 phosphorylation. At the same time, significant residual (i.e. Ͼ50%) GSK-3 phosphorylation was observed in the absence of Vav family members. This contrasted with the complete loss of FOXO phosphorylation in these cells (Fig. 2). The nature of the pathway linked to this residual activation remains to be elucidated. Similarly, whether other Vav family members transmit signals along a single pathway (i.e. their loss affects the threshold of signaling) or whether they constitute a separate pathway from that which mediates the remaining activation of GSK-3 awaits further studies. The independence of PKB/GSK-3 activation from Vav-1 was observed with both anti-CD3 and anti-CD3/CD28 activation. We had expected to observe a differential dependence on Vav-1 with CD28 mediated co-stimulation; however, this was not the case. Co-stimulation was clearly demonstrated by the ability of anti-CD28 to potentiate PKB and GSK-3 phosphorylation as well as IL-2 production. Despite this, a similar increase in anti-CD3-and anti-CD3/CD28-induced PKB and GSK-3 phosphorylation was observed in WT and Vav-1 Ϫ/Ϫ cells (Figs. 1 and 2). The increase in PKB phosphorylation was mimicked by an increase in GSK-3 phosphorylation. The similarity in the level of increased phosphorylation with different modes activation was surprising given the existence of other pathways that regulate PKB and GSK-3. These include PKB autophosphorylation (55), phosphorylation by integrin-linked kinase (56,57), and unidentified kinases in lipid rafts (58).
Another noteworthy observation was that GSK-3␣ was found to be the main target of phosphorylation in T-cells (Figs.  1-3). This occurred, despite similar levels of GSK-3␣ and -␤ expression in Jurkat and primary T-cells. This was observed with both anti-CD3 and anti-CD3-CD28 co-ligation. The GSK-3␣ isoform therefore appears to be specially localized, or conformationally altered, to facilitate its preferential phosphorylation in T-cells. This observation contrasts with other mammalian cells (i.e. chondrocytes) where the ␤-isoform of GSK-3 is the primary target of phosphorylation (24,25). Our observation also implies that GSK-3␣ is the principle effector in the regulation of events such as NFAT localization in the nucleus of T-cells.
One apparent discrepancy is the difference between our results and those reported in a study on Vav-1-deficient double positive thymocytes (48). Vav-1-deficient thymocytes were found to show defects in PKB activation (48). Although the basis of this discrepancy is unclear, it could be related to a difference in the degree of surface receptor cross-linking or a difference between T-cells and thymocytes. For example, it is conceivable that lower degrees of receptor ligation are more dependent on Vav-1 expression than high degrees of co-engagement. It might also reflect different growth conditions etc. that may alter Vav-2 or -3 expression. Lower levels of Vav-2,3 may in turn increase the dependence on Vav-1.
Last, in addition to these direct assays, we examined the effect of inhibiting JNK or disrupting the cytoskeleton on PKB and GSK-3 phosphorylation. Although it could be argued that Vav-1 signaling may branch prior to the potential engagement of PKB/GSK-3, it might also have been linear in its connection to the kinases. Previous studies showed that JNK activity is either normal or defective in Vav Ϫ/Ϫ T-cells, while expression of dominant negative PI3K did not block c-Jun DNA binding mediated by JNK (44,59). Inhibition of JNK activity had no effect on anti-CD3 or anti-CD3/CD28 phosphorylation of GSK-3 (Fig. 5). Similarly, disruption of the actin cytoskeleton with cytochalasin D had no effect on GSK-3 activation (Fig. 6). This latter result is especially surprising given the number of signaling events that are thought to depend on an intact cytoskeleton. In this way, the cytoskeleton is thought to serve as a docking region that assembles signaling proteins for activation. Despite this, PKB and GSK-3 phosphorylation was normal in over four experiments where cells were treated with concentrations of cytochalasin D that clearly disrupted the cytoskeleton. This observation suggests that an intact cytoskeleton is not needed for the ability of TcR/CD3 ϫ CD28 co-ligation to induce GSK-3␣ phosphorylation and agrees with a previous study that found intact GSK-3 phosphorylation in response to PMA/ionomycin in cytochalasin D-treated cells (54). Further studies will be needed to establish the full range of pathways regulated by PI3K that operate independently of Vav-1.