Vav Regulates Peptide-specific Apoptosis in Thymocytes

Gene-targeted mice made deficient in Vav by homologous recombination (19), and H-Y and P14 TCR Tg, and P14 Tg β₂-microglobulin (β2m)^−/− mice have been described previously (22, 24–27). Mice were screened for the TCR transgene using mAbs against the Tg TCRVβ8 chain. The vav mutation was identified using genomic Southern blotting and PCR. Care of animals was in accordance with institutional guidelines.

The PKC blockers RO-31-8220 (which blocks all PKC isoforms) and GF109023X (which inhibits the activity of only the Ca²⁺-dependent PKC isoforms PKC-α, -β, and -γ) were purchased from Calbiochem Corp. (La Jolla, CA [28]). The phosphatidylinositol 3′-kinase (PI3′K)-specific inhibitor Wortmannin and the fungal metabolite CytD (which blocks actin polymerization) were obtained from Sigma Chemical Co. (St. Louis, MO). Phalloidin binds specifically to polymerized filamentous actin (F-actin) and was directly labeled with FITC (Sigma Chemical Co.).

Freshly isolated thymocytes from C57BL/6 mice were cultured in RPMI 1640 medium containing 10% FCS and 10⁻⁵ M β-mercaptoethanol in the absence or presence of dexamethasone (Sigma Chemical Co.), anti-CD95 (clone Jo91; PharMingen, San Diego, CA), anti-CD3ε (clone 145-2C11; PharMingen), anti-CD3ε plus anti-CD28 (clone 37.51), or PMA (12.5 ng/ml) plus Ca²⁺ ionophore A23617 (100 ng/ml), for different time periods and at different concentrations as indicated in figure legends. Optimal concentrations and activation regimes for the induction of apoptosis were determined in pilot studies (29, and data not shown).

Blood samples (20 μl) were collected in heparinized capillary tubes and washed once in immunofluorescence staining buffer (1% FCS, 0.01% NaN₃ in PBS). Single cell suspensions of thymocytes, spleen cells, and mesenteric lymph node cells were prepared as described (29), resuspended in PBS, and incubated with the appropriate mAbs for 30 min at 4°C. The following mAbs were used: anti-CD4 (FITC- or PE-labeled; PharMingen); anti-CD8 (FITC- or PE-labeled, or biotinylated; Serotec, Inc., Raleigh, NC); anti-pan TCR-α/β (FITC- or PE-labeled; PharMingen); anti-TCRVβ8.1+8.2 (clone KJ16, PE-labeled or biotinylated; PharMingen); anti–H-2K^b (FITC-labeled or biotinylated; PharMingen); anti–H-2K^d (FITC-labeled; PharMingen); anti-CD5 (biotinylated; PharMingen); anti-CD28 (biotinylated; PharMingen); anti-HSA (biotinylated; PharMingen); anti-CD69 (biotinylated; PharMingen); anti-CD3ε (FITC- or PE-labeled; PharMingen); anti–H-Y TCR clonotype (T3.70; FITC-labeled or biotinylated); anti-TCRVβ8 (biotinylated; PharMingen); and anti-P14 TCRVα2 (biotinylated; PharMingen). All staining combinations were as indicated in figure legends. Biotinylated mAbs were visualized with Streptavidin-RED670 (GIBCO BRL, Gaithersburg, MD), and samples were analyzed using a FACScan^® and Lysis II software (Becton Dickinson, Mountain View, CA).

Thymocytes were purified from vav^+/+, vav^+/−, and vav^−/− P14 Tg mice. P14 Tg mice express an α/β TCR (TCRVα2Vβ8) specific for the strong agonist p33 peptide of LCMV. Thymocytes were cultured on EL4 cells (H-2^b/b) in RPMI medium containing 5% FCS and 10⁻⁵ M β-mercaptoethanol (2 ml final vol/well of 24-well flat-bottomed plates; Costar Corp., Cambridge, MA). EL4 cells (H-2^b/b) were pulsed with different concentrations of the deleting p33 peptide or the weak agonist p33 peptide analogue 8.1 for 2 h before coculture with thymocytes (26, 27, 30). Thymocytes were harvested after 10, 16, and 22 h of incubation and stained with anti-CD4–PE, anti-CD8–FITC, and the chromogenic dye 7-amino-actinomycin D (7-AAD; Calbiochem Corp.). 7-AAD is a vital dye that enters dead cells after the loss of membrane integrity and allows quantitation of apoptotic cells by cytometry (FL-3 channel on a FACScan^® [31]). Apoptosis was also determined by trypan blue exclusion and by staining with propidium iodide and annexin V (R&D Systems, Inc., Minneapolis, MN). Propidium iodide and annexin V staining of thymocytes was determined by cytofluorometry using a FACScalibur^® (Becton Dickinson). Percent survival was calculated as [(total number of viable CD4⁺CD8⁺ thymocytes cultured at different LCMV-peptide concentrations)/(total number of viable CD4⁺CD8⁺ thymocytes from the same mouse cultured with EL4 cells at 37°C in the absence of peptide)] × 100.

The mitochondrial transmembrane potential (ΔΨ_m) results from the asymmetric distribution of protons across the inner mitochondrial membrane, giving rise to a chemical (pH) and electric gradient (32, 33). The inner side of the inner mitochondrial membrane is negatively charged. As a consequence, the cationic lipophilic fluorochrome 3,3′-dihexyloxacarbocyanine iodide [DiOC₆(3)] is distributed on the mitochondrial matrix, correlating with ΔΨ_m. DiOC₆(3) can be used to measure variations in the ΔΨ_m on a per cell basis. Cells induced to undergo apoptosis manifest an early reduction in the incorporation of Δψ_m-sensitive dyes, indicating a disruption of Δψ_m. For DiOC₆(3) staining, 10⁶ thymocytes were incubated with DiOC₆(3) (final concentration 20 nM in PBS) for 20 min at 37°C. DiOC₆(3) staining was analyzed immediately using a FACScan^®.

Thymocytes were isolated from nonselecting β2m^−/−vav^+/+, positively selecting vav^+/−, and vav^−/− P14 Tg mice and cultured on a monolayer of confluent and adherent MC57G fibroblasts (H-2^b/b) in RPMI medium containing 5% FCS and 10⁻⁵ M β-mercaptoethanol. MC57G APCs were pulsed with different concentrations of the agonist LCMV–p33 peptide or the control AV peptide for 2 h before coculture with thymocytes. Nonadherent thymocytes (2 × 10⁶ P14 Tg thymocytes/well) were harvested after 15, 30, 60, and 180 min of coculture with adherent fibroblasts and assayed for MAPK and SAPK activation, nuclear factor κB (NF-κB) activation (IκB phosphorylation), and phosphotyrosine signaling (19). In brief, tyrosine phosphorylation was monitored in total cell lysates and immunoprecipitates using an antiphosphotyrosine mAb (Upstate Biotechnology, Inc., Lake Placid, NY). Activation of SAPKs and MAPKs was detected using phospho-SAPK– and phospho-MAPK–specific Abs (New England Biolabs Inc., Beverly, MA). The levels of SAPK and MAPKs (extracellular signal–regulatory kinase [ERK]1/ERK2) were determined by immunoblotting (New England Biolabs Inc.). Phosphorylated IκB was detected using an mAb specific for phosphoserine (Ser32) of IκBα (New England Biolabs Inc.). Ser32 phosphorylation of IκBα targets IκB for degradation and is essential for the release of active NF-κB (34).

Thymocytes and peripheral lymph node T cells were activated with anti-CD3ε (1 μg/ml) for different time periods as indicated in figure legends. Cells (10⁷/lane) were harvested and lysed, and Vav was immunoprecipitated using an anti-Vav1 Ab reactive against amino acids 576–589 (Santa Cruz Biotechnology Inc., Santa Cruz, CA). Immunocomplexes were separated by SDS-PAGE, transferred onto a membrane, and Western blotted to detect the presence of Vav and the PKC isoforms PKC-α, -β (-βI and -βII), -γ, -δ, -ε, -ζ, -θ, -ι, -λ, -μ, and receptor for activated C-kinase 1 (RACK1). In parallel, total cell lysates (2 × 10⁶ cells/lane) were Western blotted to ascertain the relative protein levels of these PKC isoforms in thymocytes. All PKC isoform Abs were purchased from Transduction Laboratories (Lexington, KY).

Peptide-specific Ca²⁺ mobilization ([Ca²⁺]_i) in P14 thymocytes was determined as described (35). In brief, freshly isolated P14 Tg thymocytes (2 × 10⁶) were loaded with 3 mM INDO-1 (Molecular Probes, Inc., Eugene, OR) in IMDM (pH 7.4) for 1 h at 37°C. Thymocytes were then incubated with peptide-presenting EL4 cells loaded with either p33 or AV peptide and centrifuged (1,500 rpm, 4°C) to allow conjugate formation between thymocytes and EL4 cells. Increases in intracellular free Ca²⁺ were recorded in real time on live-gated thymocyte–EL4 conjugates using a FACScalibur^® (Becton Dickinson).

Thymocytes were preincubated with PMA (12.5 ng/ml), and EL4 cells were loaded with either p33 or AV peptide (10⁻⁶ M) for 30 min on ice. Actin polymerization was initiated by placing the cells at 37°C for different time periods as indicated in figure legends. Activation was stopped by addition of 4% paraformaldehyde. After fixation, cells were incubated for 30 min with FITC-labeled phalloidin, which specifically binds to polymerized F-actin. Cells were washed three times in PBS, and phalloidin staining was analyzed using a FACScan^®. Percentage of peptide-specific actin polymerization (percent F-actin) was calculated as [(mean value of phalloidin fluorescence intensity of p33-activated cells at a certain time point)/(mean value of phalloidin fluorescence intensity of AV-activated cells at the same time point)] × 100.

To analyze the role in vivo of Vav in thymocyte selection, we introduced two rearranged TCR-α/β transgenes, H-Y and P14, into a vav^−/− background. The H-Y TCR recognizes a male-specific peptide in the context of MHC class I. Thymocytes expressing the H-Y TCR are positively selected in female H-2^b mice (22, 24, 25). The P14 TCR is specific for a peptide epitope of the LCMV glycoprotein p33 in the context of the MHC haplotype H-2D^b (23). In positively selecting vav^+/− mice, the development of P14⁺ and H-Y⁺ thymocytes is skewed towards the CD8⁺ lineage (Fig. 1, A and B). In vav^−/− mice, this bias in favor of mature CD8⁺ thymocytes did not occur in either H-Y Tg or P14 Tg thymocytes (Fig. 1, A and B), and development was in fact blocked at the immature CD4⁺ CD8⁺ stage of differentiation. Consistent with a block in positive selection, TCR Tg vav^−/− thymocytes expressed high levels of HSA and did not upregulate surface expression of the maturation and selection markers CD69, CD5 (Fig. 1,C), H-2K^b, or CD45RB (not shown). Expression of the P14 Tg TCRVβ8 and TCRVα2 chains was significantly lower in immature CD4⁺CD8⁺ thymocytes of vav^−/− mice compared with CD4⁺CD8⁺ thymocytes from vav^+/− mice (Fig. 1,C). Similarly, expression of the H-Y Tg– specific TCRVβ8 and TCRVα3 chains detected by the T3.70 Ab was significantly lower in H-Y Tg vav^−/− mice (not shown). Importantly, P14 Tg vav^−/− (and H-Y Tg vav^−/− [not shown]) thymocytes displayed a phenotype that is similar to that of P14 Tg (and H-Y) vav^+/+ β2m^−/− mice, which have a defect in positive selection of MHC class I–restricted thymocytes (Fig. 1 C). These results show that Vav regulates the positive selection of MHC class I–restricted thymocytes and confirm previous data demonstrating the crucial role of Vav in positive thymocyte selection (16–18, 21).

To examine the role of Vav in peptide-specific clonal deletion in vivo and in vitro, we investigated the negative selection of H-Y and LCMV TCR Tg thymocytes. In the H-Y Tg system, thymocytes expressing the H-Y TCR are positively selected in female H-2^b mice but negatively selected in male H-2^b mice. Negative selection in male H-Y Tg mice results in a small thymus due to deletion of CD4⁺CD8⁺ thymocytes (24; Fig. 1 B). However, male vav^−/− mice contained a large number of CD4⁺CD8⁺ thymocytes, indicating that negative selection of H-Y Tg thymocytes was severely impaired in the absence of Vav.

To further elucidate the requirement for Vav in negative selection, an in vitro model of peptide-specific negative selection was used (30). In the LCMV–p33 peptide system, APCs are loaded with different concentrations of the LCMV glycoprotein strong agonist peptide p33 or the weak agonist p33 peptide analogue 8.1. Although vav^+/− P14 Tg thymocytes readily underwent apoptosis in a dose-dependent manner in response to treatment with either the p33 or 8.1 peptide (Fig. 2,A), vav^−/− P14 Tg thymocytes were completely resistant to peptide-mediated apoptosis, even at very high peptide concentrations (Fig. 2,B). Since P14 vav^−/− thymocytes are blocked in positive T cell selection, we used β2m^−/− vav^+/+ P14 TCR Tg mice (H-2^b/b) as an additional control since these mice have a block in the positive selection of the P14 TCR due to deletion of the MHC class I ligand (β2m^−/−). The kinetics and extent of p33-induced apoptosis of β2m^−/− CD4⁺CD8⁺ P14 Tg thymocytes were similar to those of vav^+/− P14 Tg mice, indicating that the defect of P14 Tg vav^−/− thymocytes to undergo peptide-specific apoptosis was not due to differences in the composition of thymocyte populations (Fig. 2 C). These results show that Vav is required for the negative selection of peptide-specific thymocytes.

Immature CD4⁺CD8⁺ thymocytes are highly susceptible to cell death induced by many apoptotic stimuli (36). We evaluated the ability of vav^+/− and vav^−/− thymocytes to undergo apoptosis in response to the following stimuli: dexamethasone, PMA/Ca²⁺ ionophore, anti-CD95 (FAS), anti-CD3ε, and anti-CD3ε/anti-CD28 (29, 37). Although treatment with anti-CD3ε/anti-CD28 induced the cell death of vav^+/− CD4⁺CD8⁺ thymocytes, vav^−/− thymocytes were strikingly resistant to anti-CD3ε/anti-CD28–triggered apoptosis (Fig. 3). No significant differences were observed between vav^−/−, vav^+/−, and vav^+/+ thymocytes in the extent or kinetics of cell death in response to anti-CD95 (FAS), PMA plus Ca²⁺ ionophore (Fig. 3), or dexamethasone (not shown), implying that the cellular apoptotic machinery is functional in the absence of Vav. These data indicate that Vav is a crucial signal transduction molecule involved in TCR-mediated thymocyte apoptosis.

Disruption of the mitochondrial transmembrane potential (Δψ_m) due to the opening of mitochondrial membrane pore complex has been identified as the earliest common denominator of apoptosis (36). Alterations in mitochondria lead to the release of “apoptosis-inducing factor” (AIF) or cytochrome c, resulting in the activation of effector caspases such as caspase 3 (Cpp32) (32, 38–41). To assess whether Vav links TCR signaling to the opening of mitochondrial pores, we determined the mitochondrial transmembrane potential (Δψ_m) by cytometry using the fluorochromic dye DiOC₆(3) (33). Stimulation of P14 Tg vav^+/− thymocytes with 10⁻⁶ M p33 peptide led to measurable changes in Δψ_m compared with P14 Tg vav^+/− thymocytes treated with similar concentrations of the nondeleting control AV peptide (compare Fig. 4, A and C). However, p33 peptide–triggered disruption of Δψ_m did not occur in vav^−/− thymocytes (compare Fig. 4, B and D). Fig. 4 E shows that this effect was dose dependent only in P14 Tg vav^+/− thymocytes. Moreover, peptide-specific activation of the downstream effector caspase 3 (42) did not occur in P14 Tg vav^−/− thymocytes even when very high concentrations of the deleting p33 peptide were used (not shown). Disruption of Δψ_m and caspase 3 activation occurred normally in vav^−/− thymocytes treated with dexamethasone or CD95 (not shown).

We (19) and other groups (17, 20, 21) have used anti-CD3 cross-linking to establish that Vav links TCR/ CD3 activation to the actin cytoskeleton and Ca²⁺ flux. However, these experiments showed that Vav had no apparent role in phosphotyrosine signaling, or in the activation of MAPK, SAPK, or NF-κB. However, the role of Vav in peptide–MHC-mediated physiological responses and peptide–MHC-specific signaling remained obscure. To determine the biochemical defect in vav^−/− T cells, we analyzed peptide-triggered signaling pathways in P14 TCR Tg thymocytes from vav^+/+, vav^−/−, and various control strains. Since P14 vav^−/− thymocytes are blocked in positive T cell selection, we used β2m^−/− vav^+/+ P14 TCR Tg mice (H-2^b/b) as an additional control. These mice have a block in the positive selection of the P14 TCR due to the deletion of the MHC class I ligand (β2m^−/− [27, 43]). The use of β2m^−/− vav^+/+ P14 Tg mice ensured that thymocytes expressing the Tg TCR chain had not been previously activated by the selecting MHC ligand. TCR levels and expression of maturation markers were similar on the thymocytes of P14 vav^−/− and P14 β2m^−/− vav^+/+ mice (Fig. 1 C).

P14 vav^−/−, P14 vav^+/−, and P14 β2m^−/− vav^+/+ thymocytes were cultured with the deleting p33 peptide or the nondeleting control AV peptide for different time periods. No apparent differences were observed between vav^−/−, vav^+/−, and β2m^−/− vav^+/+ P14 Tg thymocytes in either the extent or kinetics of phosphotyrosine signaling, MAPK activation, or NF-κB activation after p33 peptide–specific activation (not shown). Intriguingly, p33 peptide–MHC-mediated activation of either vav^−/−, vav^+/−, or β2m^−/− vav^+/+ P14 Tg thymocytes did not lead to any detectable SAPK/JNK activity (not shown). However, p33 peptide– specific Ca²⁺ mobilization was significantly decreased in vav^−/− P14 thymocytes compared with P14 β2m^−/− vav^+/+ thymocytes (compare Fig. 5, C and D) and P14 vav^+/− thymocytes (not shown). Moreover, phalloidin staining experiments showed that vav^−/− thymocytes exhibited a decrease in actin polymerization and the formation of F-actin after activation with the deleting p33 peptide (Fig. 6). Peptide– MHC-triggered Ca²⁺ mobilization and actin polymerization were comparable between P14 vav^+/− and P14 β2m^−/− vav^+/+ thymocytes (not shown). These results provide the first genetic evidence that Vav links peptide-specific activation of the TCR to cytoskeletal reorganization and Ca²⁺ mobilization.

Treatment with the actin polymerization blocker CytD is known to mimic the functional and biochemical defects observed in peripheral vav^−/− T cells after CD3 cross-linking (20). It has also been reported that CytD addition impairs peptide-induced Ca²⁺ flux in human CD4⁺ T cell clones (35). Therefore, we analyzed whether CytD treatment could abrogate Ca²⁺ mobilization in freshly isolated, wild-type P14 thymocytes. Incubation of P14 thymocytes with CytD (Fig. 5,E) or chelation of extracellular Ca²⁺ by EGTA (Fig. 5 F) significantly decreased p33 peptide–triggered Ca²⁺ mobilization. The extent of the CytD-mediated decrease in peptide-specific Ca²⁺ flux was dependent on the concentration of the p33 peptide (not shown). Treatment of P14 thymocytes with CytD had no apparent effect on the extent of tyrosine phosphorylation, NF-κB activation, or MAPK activation (not shown). These results suggest that peptide-specific actin polymerization has an important role in the extent and duration of TCR-mediated Ca²⁺ mobilization in thymocytes.

To further elucidate the role of Vav-regulated changes in the actin cytoskeleton during negative selection, we tested whether the actin polymerization inhibitor CytD could interfere with peptide-specific apoptosis of P14 TCR Tg thymocytes. Interestingly, CytD blocked the peptide-specific cell death of vav^+/+ P14 thymocytes in a dose-dependent manner (Fig. 7). Similarly, anti-CD3/anti-CD28–mediated, but not dexamethasone- or CD95-mediated, apoptosis of vav^+/+ thymocytes could be inhibited by CytD (not shown). These data suggest that TCR-mediated actin polymerization plays a role in peptide-specific apoptosis of thymocytes.

The primary biochemical defects associated with signaling in peptide-specific vav^−/− thymocytes were impaired Ca²⁺ flux and reduced actin polymerization. Addition of Ca²⁺ ionophore, which restores Ca²⁺ flux in P14 vav^+/− and vav^−/− thymocytes (not shown), did not restore p33 peptide–mediated apoptosis even when used at very high concentrations (100 ng/ml; Fig. 8,A). In contrast, activation of vav^−/− P14 Tg cells with the deleting p33 peptide plus activation of PKC via the phorbol ester PMA significantly shifted the dose–response curve of cell death in a dose- dependent manner (Fig. 8 B). Thus, activation of PKC via PMA, but not rescue of Ca²⁺ flux by Ca²⁺ ionophore, was able to restore peptide-specific apoptosis in P14 vav^−/− thymocytes, suggesting that the Vav and PKC signaling cascades cooperate in the induction of TCR-mediated thymocyte apoptosis. However, it should be noted that our results do not preclude a role for Ca²⁺ elevation in thymocyte selection and clonal deletion in vivo.

Recently, it has been shown that in order for Vav to act as a GDP/GTP exchange factor (GEF) for Rac, RhoA, or CDC42, PI3′K activity and the binding of PI3′K-generated phospholipid products to the PH domain of Vav are required (44). Therefore, we investigated whether the specific PI3′K inhibitor Wortmannin could block peptide-specific apoptosis. As shown in Fig. 8, C and D, the inhibition of PI3′K did not affect the apoptosis of P14 thymocytes triggered by the deleting p33 peptide, suggesting that the role of Vav in thymocyte apoptosis is independent of PI3′K phospholipid–dependent Vav activation. However, inhibition of PKC by the compound RO-31-8220 (which blocks all PKC isoforms) prevented peptide-specific apoptosis in wild-type thymocytes (Fig. 8, C and D). Peptide-specific apoptosis of P14 thymocytes was not blocked by the pharmacological inhibitor GF109203X, which prevents activation of Ca²⁺-dependent PKC isoforms (Fig. 8,C). PMA-triggered apoptosis was blocked by the global PKC blocker RO-31-8220 but not by CytD or Wortmannin (Fig. 8 E). These data indicate that Vav links TCR signaling to activation of a Ca²⁺-independent PKC isoform required for the induction of peptide-specific thymocyte apoptosis.

To further evaluate the link between Vav and PKC, we immunoprecipitated Vav from wild-type thymocytes and analyzed the binding of various PKC isoforms to Vav. Surprisingly, Vav coimmunoprecipitated with the Ca²⁺-independent PKC isoform PKC-θ but not with any other PKC isoform tested (Fig. 9, left). The association between Vav and PKC-θ did not change after CD3 cross-linking, suggesting that Vav/PKC-θ binding is constitutive in thymocytes. Constitutive Vav/PKC-θ association was also observed in mature peripheral T cells (not shown). Our results do not preclude associations between Vav and low abundance PKC isoforms.

Our genetic and functional analyses of vav^−/− mice show that Vav is a crucial regulator both of positive and negative selection of CD8⁺ thymocytes and the peptide-triggered apoptosis of developing T cells. Vav is involved as a regulator of TCR-mediated cytoskeletal reorganization and Ca²⁺ mobilization after peptide-specific activation by APCs. Our results suggest that peptide-triggered thymocyte apoptosis is mediated via Vav activation, changes in the actin cytoskeleton, and subsequent activation of a PKC isoform. This hypothesis is based on the following findings: (a) Vav-deficient thymocytes do not undergo peptide–MHC-mediated cell death in vitro or in vivo; (b) vav^−/− thymocytes exhibit a defect in actin polymerization, and inhibition of cytoskeletal changes by CytD blocks peptide–MHC-mediated and anti-CD3/anti-CD28–mediated thymocyte apoptosis. However, inhibition of actin polymerization does not inhibit dexamethasone-, CD95-, or PKC-mediated thymocyte apoptosis; (c) activation of PKC restores the susceptibility to apoptosis of p33 peptide–triggered P14 Tg vav^−/− thymocytes, and inhibition of a Ca²⁺-independent PKC isoform inhibits TCR-mediated thymocyte apoptosis; and (d) Vav associates constitutively with PKC-θ but not with any other PKC isoform. Thus, the PKC isoform PKC-θ is a good candidate for the effector kinase of negative thymocyte selection.

The PKCs are a family of 11 phospholipid-dependent serine/threonine kinases. These closely related isoenzymes differ in their structural and biochemical properties, tissue distribution, subcellular localization, and substrate specificity (45, 46). According to primary structure and binding to Ca²⁺ or phorbol esters, different PKC subgroups exist: conventional PKCs (α, βI, βII, and γ) bind Ca²⁺ and are activated by PMA; novel PKCs (δ, ε, θ, and η) are activated by PMA but do not bind Ca²⁺; atypical PKCs (ζ, λ, ι, and μ) bind to diacylglycerol (DAG) but are not activated by PMA or Ca²⁺ ionophores. Activation of PKC molecules by lipid second messengers requires membrane recruitment (47). T cell activation requires PKC activity. Previously, it has been reported using PCR and Northern blotting that PKC-α, -β, -δ, -ε, -ζ, and -θ are expressed in thymocytes (48–50). Our study shows for the first time that all PKC isoforms are expressed in thymocytes, albeit at different levels. Although most PKC isoforms are expressed ubiquitously, the isoform PKC-θ is predominantly expressed in the hematopoietic system, particularly in T cells (48–50), and has been placed upstream of IL-2 transactivation and AP1 (Fos/Jun) activity in T lymphoma cells (51, 52). Recently, it has been shown in Jurkat cells that calcineurin and PKC-θ cooperate in inducing IL-2 gene transcription (53), a function that is reminiscent of the coordination of IL-2 expression by calcineurin and Vav/Rac1 (20, 54–57). Interestingly, the small Ras-like molecules Rac1, CDC42, and Rho, whose activation is regulated by Vav, may act upstream of PKC (58, 59), suggesting that Vav, Rac1, and PKC-θ may form a complex. Although Vav and PKC-θ clearly associate in thymocytes and peripheral T cells, Rac1, CDC42, and RhoA do not coimmunoprecipitate with these complexes.

Vav and Vav-regulated Rac1 and CDC42 have been previously shown to mediate changes in the actin cytoskeleton and to activate SAPK/JNK (13–15, 60, 61). Similarly, PKC-θ has been placed upstream of SAPK/JNK in T cell lines (53). However, TCR/CD28-mediated SAPK/JNK activation appears normal in vav^−/− thymocytes and mature T cells after anti-CD3 cross-linking (19, 20). Similarly, a genetic mutation in the SAPK/JNK signaling pathway had no effect on the induction of thymocyte apoptosis (29), suggesting that Vav and perhaps PKC-θ–mediated negative selection are independent of SAPK/JNK. Importantly, although SAPK/JNK activity is readily induced in thymocytes in response to anti-CD3 and anti-CD28 Abs (62), we failed to detect SAPK/JNK activation in thymocytes after specific peptide–MHC-mediated stimulation. Thus, Vav may define a novel signaling pathway that leads to negative thymocyte selection and IL-2 production in peripheral T cells. This signaling pathway appears to be independent of MAPK, SAPK/JNK, or NF-κB activation.

Rac1, RhoA, or CDC42 activation by Vav requires PI3′K activity in vitro (44). However, inhibition of PI3′K has no apparent effect on the peptide-triggered thymocyte apoptosis of thymocytes, suggesting either that the role of Vav in thymocyte apoptosis is independent of its GTP exchange activity for Rac and CDC42, or that Vav can function as an exchange factor even in the absence of PI3′K activity in vivo. Similarly, it has been shown that a pharmacological PKC inhibitor, but not the PI3′K inhibitor Wortmannin, can block anti-CD3/anti-CD28–mediated thymocyte apoptosis (63). These latter results are consistent with our observation that Vav and PKC regulate anti-CD3/anti-CD28–mediated apoptosis and imply that PKC activation is a crucial signal in the pathway of TCR-triggered apoptosis in developing thymocytes. Whether PKC-θ and/or other PKC isoforms are the crucial downstream kinases in thymocyte selections needs to be determined in genetic experiments.

Both we (19) and Holsinger et al. (20) have previously shown that Vav regulates Ca²⁺ mobilization and actin reorganization in peripheral T cells after anti-CD3ε Ab cross-linking. The results of the present study provide the first genetic evidence that Vav is also a crucial regulator of the TCR-mediated Ca²⁺ flux and actin polymerization required for thymocyte selection of peptide–MHC-stimulated thymocytes. In particular, actin polymerization was found to be crucial for the induction of apoptosis. Various cytoskeletal proteins and regulators of cytoskeletal changes, such as gelsolin (64–66), β-catenin (67), PAK2 (68), actin (67, 69–76), fodrin (77), and Gas2 (76), are proteolytically cleaved after the induction of apoptosis. Cleavage of cytoskeletal proteins, which occurs after caspase activation, has an important role in the death effector phase of apoptosis, particularly in the regulation of membrane alterations, morphological changes, and DNA fragmentation (68, 69). In contrast, inhibition of actin polymerization by CytD resulted in impaired TCR-mediated apoptosis. These cytoskeletal changes were found to act upstream of the opening of the mitochondrial membrane pore complex and caspase activation. However, CytD could not block apoptosis after PKC activation (which induces rapid changes in the actin cytoskeleton) and CytD could not protect thymocytes from dexamethasone- or CD95-mediated apoptosis, suggesting that actin changes per se are not the principle mechanism for conveying a cell death signal in thymocytes. Rather, we propose that TCR-mediated and Vav-regulated changes in the actin cytoskeleton lead to the recruitment and/or assembly of a death effector molecule(s) that relays the TCR signal to the apoptotic machinery.

It has been shown in the D10 T cell line that only PKC-θ translocates to the site of contact between T cells and APCs. All other PKC isoforms (α, β1, δ, η, and ζ) are excluded from the contact site, suggesting that PKC-θ has a specific role in T cell activation (78). Since Vav regulates TCR clustering and PKC-θ is found in these clusters, Vav-regulated actin polymerization may be required for recruitment of PKC-θ to the TCR and activation of PKC at the cell membrane. How Vav and possibly PKC-θ link TCR-mediated signals to mitochondrial apoptosis and caspase 3 activation remains to be determined. Importantly, our results show that Vav and actin polymerization also regulate the extent and duration of peptide-specific Ca²⁺ mobilization, an event thought to be involved in thymocyte apoptosis and clonal selection (79–83).

Our data show that Vav regulates positive and negative selection of MHC class I–restricted thymocytes. We have previously reported that the progression of CD4⁻CD8⁻ T cell precursor cells to CD4⁺CD8⁺ thymocytes is impaired in vav^−/− mice on a C57BL/6 (H-2^b/b) background but not on a CD1 (H-2^q/q) background (19). Thus, Vav clearly has a role in preTCR-mediated expansion of early thymocytes, but the absence of Vav can be partially compensated by other signaling molecules. However, this compensatory mechanism does not appear to be operational or sufficient to rescue positive and negative selection of TCR Tg thymocytes in vav^−/− mice. Thus, Vav either plays a different role in TCR signaling at different stages of development and/or different compensatory mechanisms are available to CD4⁻CD8⁻ versus CD4⁺CD8⁺ thymocytes that make up for the Vav mutation in the expansion from CD4⁻CD8⁻ to CD4⁺CD8⁺ cells but cannot compensate for selection of CD4⁺CD8⁺ thymocytes.

Our data provide the first genetic evidence for a role for Vav in peptide–MHC-triggered cytoskeletal reorganization in vivo. These results indicate that Vav-regulated actin reorganization is a crucial prerequisite for antigen receptor–mediated selection and apoptosis in peptide-specific thymocytes. We have shown that Vav and the actin cytoskeleton regulate the extent and duration of Ca²⁺ mobilization after peptide-specific activation, and that Vav functions upstream of mitochondrial pore opening and caspase activation. The defect in peptide-specific apoptosis in vav^−/− thymocytes can be overcome by activation of PKC, and the inhibition of PKC blocks peptide-specific cell death in wild-type thymocytes, indicating that PKC activation is the trigger for thymocyte apoptosis in response to peptide. Of all PKC isoforms tested, Vav associated only with the PKC-θ isoform expressed primarily in T cells, suggesting that PKC-θ is the crucial kinase involved in thymocyte selection and clonal deletion.

We thank K. Bachmaier, K. Tedford, A. Oliveira-dos-Santos, T. Sasaki, H. Nishina, A. Tafuri-Bladt, and L. Zhang for helpful comments, and M. Saunders for scientific editing.

This work was supported by grants from the Medical Research Council of Canada (to J.M. Penninger and A. Bernstein); from the National Cancer Institute of Canada (to A. Bernstein); and from the Deutsche Forschungsgemeinschaft (SFB 465; to K.-D. Fischer).

7-AAD

7-amino-actinomycin D

β2m

β2-microglobulin

CytD

cytochalasin D

F-actin

filamentous actin

JNK

c-Jun NH₂-terminal kinase

LCMV

lymphocytic choriomeningitis virus

MAPK

mitogen-activated protein kinase

NF-κB

nuclear factor κB

PI3′K

phosphatidylinositol 3′-kinase

PKC

protein kinase C

SAPK

stress-activated protein kinase

Tg

transgenic