A positive regulatory role for Cbl family proteins in tumor necrosis factor-related activation-induced cytokine (trance) and CD40L-mediated Akt activation.

Tumor necrosis factor (TNF)-related activation-induced cytokine (TRANCE) is a TNF family member essential for osteoclast differentiation, and it induces the activation and survival of osteoclasts and mature dendritic cells. We recently demonstrated that TRANCE activates Akt via a mechanism involving TRANCE receptor (TRANCE-R)/RANK, TRAF6, and c-Src. Here, we show that TRANCE-R and CD40 recruit TRAF6, Cbl family-scaffolding proteins, and the phospholipid kinase phosphatidylinositol 3-kinase in a ligand-dependent manner. The recruitment of Cbl-b and c-Cbl to TRANCE-R is dependent upon the activity of Src-family kinases. TRANCE and CD40L-mediated Akt activation is defective in Cbl-b -/- dendritic cells, and CD40L-mediated Akt activation is defective in c-Cbl -/- B cells. These findings implicate Cbl family proteins as not only negative regulators of signaling but as positive modulators of TNF receptor superfamily signaling as well.

Tumor necrosis factor (TNF) 1 family proteins mediate diverse effects on cells of the hematopoietic lineage via their cognate receptors, members of the TNF receptor (TNFR) family (1). TNFR family proteins lack intrinsic enzymatic activity but are linked to intracellular signaling cascades through TNFRassociated factor (TRAF) proteins, and numerous TNF family proteins have been shown to activate nuclear factor B (NF-B) and mitogen-activated protein kinase cascades (2). TNF-related activation-induced cytokine (TRANCE; also called RANKL, ODF, and OPGL) is a TNF family member expressed on activated T cells and osteoblasts that regulates the function of dendritic cells and osteoclasts through its cognate receptor, TRANCE-R (also called RANK) (3). Recently, we demonstrated that in addition to activating NF-B (4) and c-Jun N-terminal kinase (5), TRANCE activates Akt, a serine/threonine kinase implicated in survival signals, through a mechanism involving TRAF6 and the nonreceptor tyrosine kinase c-Src (6).
c-Cbl is a cytoplasmic adapter molecule that has been implicated in the negative regulation of signaling from a variety of receptor tyrosine kinases, including growth factor receptors and antigen receptors in lymphocytes (7,8). A highly related protein, Cbl-b, has been identified (9). The domain structure of Cbl proteins consists of several functional domains, including an SH2-like phosphotyrosine binding domain, a RING finger, a proline-rich domain, and a leucine zipper. Originally identified as a viral proto-oncogene that acquires transforming potential with the 70z deletion (10), Cbl has been implicated in the negative regulation of tyrosine kinase signaling by shortening the duration of activating signals (7).
c-Cbl and Cbl-b have been shown to interact with a wide variety of activated signaling molecules including phosphatidylinositol 3-kinase (PI3-K), Src-family tyrosine kinases, Syk, and adaptor proteins Grb2 and Shc (11)(12)(13)(14). One mechanism by which they may negatively regulate signaling is by acting as an E3 ubiquitin ligase, which results in the degradation of activated molecules by the proteasome (15)(16)(17). This E3 ubiquitin ligase activity has been localized to the RING finger of c-Cbl, and the RING finger has been associated with the negative regulation of a number of tyrosine kinases, including the epidermal growth factor receptor (EGFR) (18,19), syk (20), and colony-stimulating factor-1 receptor (21), among others. Cbl-b has been associated with the negative regulation of Vav-mediated c-Jun N-terminal kinase activation (22) and EGFR signaling (23).
Mice with targeted deletions in c-Cbl (24) and, recently, Cbl-b (25,26) have been described; each displays a phenotype of decreased thresholds for lymphocyte activation and development of autoimmunity. Mice with a targeted deletion in c-Cbl display enhanced thymic positive selection, likely due to the persistence of activated costimulatory molecules that are ordinarily targeted for degradation by c-Cbl (24). Cbl-b Ϫ/Ϫ mice demonstrate T cell hyperactivation and hyperproliferation in response to antigen receptor stimulation, uncoupling of T cell receptor, and CD28 stimulation and develop spontaneous autoimmunity (25,26).
In this report, we identify a mechanism by which the TNF family members TRANCE and CD40L activate Akt through their cognate receptors TRANCE-R and CD40. We previously demonstrated that, upon ligand engagement, TRANCE-R forms a complex with c-Src and TRAF6 (6). We now show that PI3-K and Cbl proteins are also components of this signaling complex. Examination of the mechanism of the interaction of c-Cbl and Cbl-b with TRANCE-R shows a requirement for c-Src kinase activity, which may be regulated by TRAF6. Furthermore, a stable complex of TRANCE-R and Cbl-b is observable only in the presence of a proteasome inhibitor, suggesting that Cbl-b may negatively regulate TRANCE signaling by downregulating one or more components of the TRANCE-R-signaling complex. Finally, using cells derived from Cbl-b Ϫ/Ϫ and c-Cbl Ϫ/Ϫ mice, we show that Akt activation by TRANCE and CD40L in dendritic cells is dependent on Cbl-b, whereas Akt activation in B cells by CD40L is dependent on c-Cbl, suggesting a novel positive regulatory role for Cbl proteins in signaling.
Primary Cells-Mature dendritic cells were generated from bone marrow precursors as described (30). Osteoclasts were generated from bone marrow precursors as described (31). Lymphocytes were prepared from whole spleens by making single-cell suspensions followed by erythrocyte lysis and plating for 1 h on tissue culture plates to deplete adherent cells.
Cell Stimulation, Transfection, and Analysis-In vitro differentiated mature dendritic cells and osteoclasts and freshly isolated splenocytes were extensively washed to remove exogenous growth factors, cultured in medium with low serum (0.5% fetal calf serum for 2-4 h), then stimulated by adding TRANCE or CD40L as indicated. After stimulation, cells were washed with ice-cold phosphate-buffered saline, lysed, and subjected to immunoprecipitation and Western blotting as described (6). To control for equal loading of each time point, the protein concentration of each sample was determined, and samples were normalized for total protein content before further processing. 293T cells were transfected by calcium phosphate precipitation as described (4). The amount of transfected DNA was held constant to 1 g by addition of empty vector DNA where necessary. Cells were processed for analysis 24 h after transfection. Where indicated, MG-132 or an equivalent amount of vehicle (Me 2 SO) was added to a final concentration of 10 M 4 h before processing. Cells were processed and subject to immunoprecipitation and Western blotting as described (6). All transfection experiments were repeated at least three times, and representative results are shown.

TRANCE-R and CD40 Interact with PI3-K and c-Cbl upon
Ligand Stimulation-Since TRANCE activates Akt in dendritic cells (DCs) and osteoclasts and Akt activation is dependent on the activity of PI3-K, we investigated whether PI3-K was associated with TRANCE-R. To determine whether PI3-K is part of the TRANCE-R-signaling complex in primary cells, we immunoprecipitated TRANCE-R from TRANCE-treated DCs. The p85 regulatory subunit of PI3-K was recruited to TRANCE-R in a ligand-dependent manner (Fig. 1A, top), which correlates with Akt phosphorylation in the whole cell extract (6). Since PI3-K has been shown to associate with the cytoplasmic scaffolding protein c-Cbl in a variety of cell types, we probed the TRANCE-R immunoprecipitates for c-Cbl and found that it associates with TRANCE-R in a TRANCE-dependent fashion. This correlates with a ligand-dependent increase in TRANCE-R-associated TRAF6 (Fig. 1A, top). Immunoprecipitation of c-Cbl from dendritic cell lysates confirmed that TRAF6 inducibly associates with c-Cbl upon ligand stimulation (Fig. 1A, bottom).
To determine whether other TNFR family members known to signal through TRAF6 behave similarly to TRANCE-R, we treated DCs with CD40L, which like TRANCE promotes the FIG. 1. PI3-K and c-Cbl are recruited to TRANCE-R and CD40 upon ligand stimulation in dendritic cells. A, dendritic cells were treated for the indicated number of minutes with TRANCE (2 g/ml) and lysed. TRANCE-R (top) and c-Cbl (bottom) were immunoprecipitated, and the immunoprecipitates (IP) were probed with antibodies to PI3-K, c-Cbl, and TRAF6 as indicated. B, DCs were treated with soluble CD40L (1:100) for the indicated number of minutes, and CD40 was immunoprecipitated. The immunoprecipitates were probed with antibodies to c-Cbl and PI3-K (top and middle). An in vitro Src-family kinase assay was performed on the CD40 immunoprecipitates (bottom) with recombinant SAM68 as a substrate. C, DCs were treated as in A, TRANCE-R was immunoprecipitated, and the immunoprecipitates were probed with antibodies to PI3-K and c-Cbl as indicated (top). Whole cell extracts (WCE) were immunoblotted with antibodies to phospho-Akt (Akt P ) and PI3-K as indicated. D, DCs were pretreated with vehicle (Me 2 SO (DMSO)) or PP1 (10 M) for 90 min, then stimulated and lysed as in A. c-Cbl was immunoprecipitated, and the immunoprecipitates were probed with antibodies to phosphotyrosine, c-Cbl, and PI3-K as indicated.
survival and activation of myeloid dendritic cells (33). Immunoprecipitation of CD40 and Western blotting showed that there was a ligand-dependent increase in the p85 subunit of PI3-K and c-Cbl associated with CD40 ( Fig. 1B, top). As we have previously found with TRANCE-R (6), Src-family kinase activity co-precipitates with ligand-stimulated CD40, as assayed by the ability of the immunoprecipitates to phosphorylate recombinant SAM68, a known Src-family kinase substrate, in vitro (Fig. 1B, bottom).
We previously observed a peak of TRANCE-induced Akt phosphorylation after 20 min of stimulation (6). To determine if this activation correlates kinetically with PI3-K recruitment to TRANCE-R, we stimulated DCs with TRANCE for up to 60 min. Surprisingly, although PI3-K and c-Cbl continue to accumulate in the TRANCE-R complex in increasing amounts ( Since TRANCE-mediated Akt activation in DCs is dependent on the activity of Src-family kinases and can be inhibited by the Src-family kinase inhibitor PP1 (6), we endeavored to determine whether the association of c-Cbl and PI3-K is dependent on Src-family kinase activity. We pretreated DCs with PP1 or vehicle (Me 2 SO), stimulated the DCs with TRANCE, and immunoprecipitated c-Cbl. In the absence of PP1, c-Cbl was constitutively phosphorylated on tyrosine in unstimulated DCs, and its phosphorylation state was unaffected by TRANCE treatment. Pretreatment with PP1 completely blocked all c-Cbl phosphorylation (Fig. 1D, top). However, the phosphorylation state of c-Cbl did not affect its binding to PI3-K, as PI3-K was constitutively associated with c-Cbl regardless of PP1 treatment or TRANCE stimulation (Fig. 1D, bottom). To ensure that TRANCE stimulation activated signaling by TRANCE-R, we probed whole cell extracts with antibodies to IB-␣ and observed equivalent TRANCE-dependent degradation of IB-␣ in both Me 2 SO-and PP1-pretreated cells (not shown).
TRANCE-R and Cbl Proteins Interact Only in the Presence of Active Src-To further elucidate the mechanism of the c-Cbl interaction with TRANCE-R, we transiently transfected HEK 293T cells with constructs driving the expression of c-Cbl, c-Src, or FLAG-epitope-tagged TRANCE-R from a cytomegalovirus promoter. In the presence of overexpressed c-Src, an anti-c-Cbl antibody co-immunoprecipitated TRANCE-R ( Fig.  2A, lane a). In the absence of overexpressed c-Cbl, endogenous c-Cbl was sufficient to demonstrate a c-Src-dependent interaction with TRANCE-R ( Fig. 2A, lane c). Conversely, in the absence of overexpressed c-Src, neither overexpressed nor endogenous c-Cbl could co-precipitate TRANCE-R ( Fig. 2A, lanes b  and d). To differentiate between catalytic and structural roles for c-Src in the TRANCE-R⅐c-Cbl complex, we cotransfected c-SrcKD with TRANCE-R and c-Cbl. Although the wild-type c-Src construct used has been shown to phosphorylate c-Cbl in overexpression systems, c-SrcKD does not (34). Immunoprecipitation of TRANCE-R with the FLAG antibody and Western blotting revealed that the kinase-active form of c-Src was able to promote a strong interaction between c-Cbl and TRANCE-R (Fig. 2B, lane a). The kinase-inactive form of c-Src, however, could not promote a strong interaction between c-Cbl and TRANCE-R (Fig. 2B, lane b).
To determine if Cbl-b, another Cbl family protein, could also interact with TRANCE-R, we cotransfected Cbl-b, TRANCE-R, and c-Src or c-SrcKD. Immunoprecipitation of TRANCE-R with the FLAG antibody did not reveal the presence of Cbl-b protein in either case (Fig. 2C, lanes a and b). However, in the presence of MG-132, a proteasome inhibitor (35), we were able to coprecipitate Cbl-b with TRANCE-R in a c-Src kinase-dependent manner (Fig. 2C, lanes c and d). This suggests that Cbl-b may down-regulate one or more of the essential components of the TRANCE-R complex by ubiquitination, thus creating a transient interaction between TRANCE-R and Cbl-b. Indeed, when we expressed TRANCE-R in the presence of c-Src and fulllength Cbl-b, we observed a marked decrease in the amount of TRANCE-R protein in the cell extract (Fig. 2D, lane a). However, when we either substituted a truncated form of Cbl-b with a deletion of the N-terminal SH2 domain (Cbl-b⌬N, Fig. 2D teract with TRANCE-R only in the presence of active c-Src, we investigated whether this interaction is dependent on tyrosine phosphorylation of TRANCE-R. Sequence analysis of the cytoplasmic domain of mouse TRANCE-R revealed the presence of three tyrosine residues that could potentially serve as targets of c-Src: Tyr-345, Tyr-440, and Tyr-468. Alignment with human TRANCE-R shows that although Tyr-345 and Tyr-468 are conserved, Tyr-440 is not. There is an additional tyrosine in human TRANCE-R at position 422, corresponding to position 418 in mouse TRANCE-R (Fig. 3A). To determine if any of the tyrosine residues in mouse TRANCE-R are phosphorylated by c-Src, we employed site-directed mutagenesis to change each of the tyrosine residues to phenylalanine. We then cotransfected the tyrosine mutants of TRANCE-R with c-Src or c-SrcKD and immunoprecipitated TRANCE-R. Western blotting of the immunoprecipitates with an anti-phosphotyrosine antibody revealed that wild-type TRANCE-R is phosphorylated by or downstream of c-Src on Tyr-468, as only the Y468F mutant was not phosphorylated in the presence of c-Src. Neither wild-type TRANCE-R nor any of its tyrosine mutants was phosphorylated on tyrosine when cotransfected with c-SrcKD, suggesting that Tyr-468 is a specific target of c-Src activity (Fig. 3B). We then cotransfected TRANCE-R constructs containing tyrosine mutations with c-Src and Cbl-b or c-Cbl and found that Cbl-b and c-Cbl co-precipitated with all of the Y-F mutants of TRANCE-R, which suggests that the interaction between TRANCE-R and Cbl is not dependent on the tyrosine phosphorylation of TRANCE-R (Fig. 3, C-D). There was no interaction between a mutant of TRANCE-R with the cytoplasmic tail deleted and c-Cbl or Cbl-b, indicating that the interaction is specific to the cytoplasmic tail of TRANCE-R (data not shown).
Cbl Proteins Regulate TRANCE-and CD40-mediated Akt Activation-Gene-targeted mice with deletions in c-Cbl (24) and Cbl-b (25, 26) have been described recently. To determine the role of Cbl proteins in TRANCE-and CD40L-mediated activation of Akt, we used B lymphocytes, osteoclasts, and DCs derived from mice deficient in c-Cbl or Cbl-b. In cells derived from Cbl-b-deficient mice, we observed that neither TRANCE (Fig. 4A) nor CD40L (Fig. 4B) was able to strongly activate Akt in DCs within 20 min of stimulation, as opposed to what we observed in wild-type cells. In all cases, NF-B activation as measured by IB-␣ degradation was identical in wild-type and knockout cells. Interestingly, at later time points (Ͼ3 h), TRANCE and CD40L treatment did result in Akt activation in DCs, which was similar to a second wave of Akt activation observed in wild-type cells, suggesting that other gene products that activate Akt via a Cbl-b-independent mechanism are upregulated over this time period. Since Akt has been widely characterized as a survival factor, we investigated whether there was a defect in TRANCE-or CD40L-mediated survival in Cbl-b Ϫ/Ϫ DCs. Perhaps due to the intact secondary wave of Akt activation, there was no difference observed in TRANCEor CD40L-mediated survival in DCs derived from Cbl-b Ϫ/Ϫ mice over a 72-h period (data not shown). In DCs derived from c-Cbl Ϫ/Ϫ mice, we did not observe any differences in TRANCE or CD40L-induced Akt activation or survival (data not shown).
In contrast to the results obtained in DCs, in B lymphocytes from c-Cbl Ϫ/Ϫ mice there was a marked deficiency in CD40Linduced Akt activation, but Akt activation was intact in Cbl-b Ϫ/Ϫ B lymphocytes (Fig. 4C). In osteoclasts derived from c-Cbl Ϫ/Ϫ and Cbl-b Ϫ/Ϫ mice, we did not observe any defects in TRANCE-induced Akt activation (Fig. 4D). Consistent with intact TRANCE-mediated Akt activation in c-Cbl Ϫ/Ϫ and Cbl-b Ϫ/Ϫ osteoclasts, we did not observe any defects in the differentiation or survival of osteoclasts derived from these mice as determined by tartrate-resistant acid phosphatase assay (data not shown).
Taken together, these results suggest that Cbl-b and c-Cbl may have cell type-dependent, overlapping roles in Akt activation. It appears that c-Cbl is required for CD40L-dependent Akt activation in B cells, whereas Cbl-b is required for TRANCE and CD40L-dependent Akt activation in DCs. In osteoclasts, c-Cbl and Cbl-b appear to be able to substitute for one another in TRANCE-dependent Akt activation. In whole cell extracts, protein expression levels of c-Cbl and Cbl-b in dendritic cells, B lymphocytes, and osteoclasts do not account for these cell type-specific differences in function (data not shown). It has been reported recently that in c-Cbl Ϫ/Ϫ osteoclasts Cbl-b is compensatorily overexpressed (36). However, it is possible that the availability of c-Cbl and Cbl-b to the various receptor-signaling complexes differs in a cell type-dependent manner due to other, as yet unidentified components of the signaling complexes.

A Positive Signaling Role for Cbl-Although
Cbl-family proteins have been widely held to play a negative role in tyrosine kinase signaling, our results suggest a positive role as well. c-Cbl and Cbl-b have been demonstrated to associate with the p85 subunit of PI3-K both constitutively and in response to FIG. 4. Cbl proteins are required for cell type-specific TRANCE and CD40L-mediated Akt activation. A, DCs were derived from wild-type or Cbl-b deficient (Ϫ/Ϫ) mice, serum-starved, and treated with TRANCE (2 g/ml) for the indicated times. Lysates (50 g) were immunoblotted with a phospho-specific Akt antibody to indicate activation of Akt (Akt P ). Membranes were stripped and reprobed with antibodies to total Akt to normalize for protein loading and IB-␣ to demonstrate activation of the NF-B signaling pathway as indicated. Note the degradation of IB-␣ after 5 min and the appearance of newly synthesized protein by 180 min. B, as in A, but cells were treated with CD40L (1:100) instead of TRANCE. C, B lymphocytes were isolated from the spleens of wild-type, c-Cbl, or Cbl-b-deficient (Ϫ/Ϫ) mice, serumstarved, and treated with CD40L for the indicated time. Lysates were immunoblotted as in A and B. D, osteoclasts were derived from bone marrow of wild-type, Cbl-b, or c-Cbl-deficient (Ϫ/Ϫ) mice, serum-starved, and treated with TRANCE for the indicated times. Lysates were immunoblotted as in A, B, and C.
ligand stimulation in a number of cell types and receptor/ligand pairs (11,12,14,37). In dendritic cells, we observed constitutive association between c-Cbl and PI3-K, which is independent of Src-family kinase activity. Overexpression of Cbl-b has been shown to abrogate Akt activation downstream of EGFR in response to ligand stimulation (23), and the hyperactivation and increased survival of T cells in Cbl-b Ϫ/Ϫ mice suggests that Cbl-b negatively regulates T cell receptor and CD28-mediated signaling (25,26). However, using cells derived from gene-targeted mice, we found that in dendritic cells, Cbl-b is required for TRANCE-and CD40L-induced Akt activation, and in B lymphocytes, c-Cbl is required for CD40L-induced Akt activation. In osteoclasts, c-Cbl and Cbl-b appear to substitute for one another in TRANCE-induced Akt activation. Cbl proteins, therefore, may positively regulate PI3-K activation via TNFR family proteins in a receptor-and cell type-specific manner by recruiting PI3-K to the receptor complex, where it is phosphorylated by Src family kinases.
Potential Negative Roles for Cbl in TRANCE Signaling-This positive role, however, appears to be short-lived, as Akt activation by TRANCE and CD40L declines in dendritic cells after ϳ20 min of stimulation. It is possible that Cbl proteins are responsible for the termination of signaling by down-regulating Src kinases, PI3-K, TRAF6, or TRANCE-R and CD40 via internalization and/or ubiquitination. EGFR, a receptor tyrosine kinase, is rapidly autophosphorylated within several minutes of ligand binding, and its major signaling events take place rapidly (38). Cbl-mediated ubiquitination of EGFR becomes evident on the order of 20 -30 min after ligand binding and quenches the activation signal over the next 20 -30 min (16). Since c-Cbl binds exclusively to phosphorylated EGFR, only activated EGFR is ubiquitinated and targeted for destruction. For productive signaling to occur, there is necessarily a time lag between the activation of the kinase and its destruction. It is therefore likely that, by acting as a scaffold for the assembly of the PI3-K signaling complex and the TRANCE-R signaling complex, Cbl can make a short-lived positive contribution to signaling before down-regulating activated proteins. Since the TRANCE-R⅐c-Cbl⅐PI3-K complex is observed in DCs long after Akt activation is quenched (Fig. 1C), it is possible that Akt down-regulation is independent of Cbl in the receptor complex.
In support of the notion that Cbl indeed has a role in the negative regulation of TRANCE signaling, we were only able to observe Cbl-b binding to TRANCE-R in the presence of MG-132, a proteasome inhibitor. Additionally, we found that over-expression of full-length Cbl-b and c-Src resulted in a marked decrease in TRANCE-R protein in cell lysates, whereas eliminating either the N-terminal domain of Cbl-b or c-Src overexpression did not reduce TRANCE-R protein levels. This suggests that TRANCE-R and/or other essential activated components of the TRANCE-R-signaling complex are targeted for proteasome-mediated degradation by Cbl-b. Three likely candidates are TRAF6, c-Src, and PI3-K. Recently, Takayanagi et al. (39) demonstrated that TRAF6 is ubiquitinated and subsequently degraded by the proteasome in response to TRANCE stimulation in osteoclast precursor cells. Harris et al. (40) report that active c-Src is ubiquitinated and subsequently degraded, whereas the steady-state level of c-SrcKD is consistently higher than that of active c-Src. This may explain the slight increase in c-Src observed in the whole cell extract in the presence of MG-132 in Fig. 2C. Fang et al. (41) report that Cbl-b binds to and induces ubiquitination of the p85 subunit of PI3-K.
Roles of TRAF6 and c-Src in Receptor Assembly-We have shown that the C-terminal receptor binding domain of TRAF6 can interact with TRANCE-R (4) and c-Src (6). We have also found that the C-terminal half of TRAF6 interacts with c-Cbl and Cbl-b. This interaction promotes the activation of c-Src to tyrosine phosphorylate c-Cbl and Cbl-b, but phosphorylation is dependent on the N-terminal half of TRAF6 (Ref. 6 and data not shown). Since catalytically active c-Src is necessary to promote an interaction between TRANCE-R and Cbl proteins but phosphorylation of TRANCE-R on a specific tyrosine residue does not affect binding, it is likely that phosphorylation of Cbl proteins ultimately promotes this interaction. Nevertheless, the possibility remains that a component of the complex that has yet to be identified is the true target of c-Src that facilitates Cbl-TRANCE-R binding. Therefore, in addition to activating the PI3-K cascade, c-Src appears to play a vital role in the assembly of the signaling complex.
Physiological Consequences of Cbl in TRANCE and CD40L Signaling-If Cbl-b is essential for Akt activation by TRANCE and CD40L in dendritic cells, why is there no apparent defect in DC survival in Cbl-b Ϫ/Ϫ mice? When bone marrow-derived DCs reach maturity after 8 days in granulocyte-macrophage colony-stimulating factor culture, they begin to undergo apoptosis in the absence of survival stimuli (42). However, this process is observable on the order of many hours to days and is most likely due to high levels of pre-existing bcl-2 protein (27). In Cbl-b Ϫ/Ϫ DCs, we observed Akt activation after several hours of TRANCE or CD40L stimulation, consistent with a second wave of activation seen in wild-type DCs. It is possible

FIG. 5. Model of the proposed mechanism of Akt activation by TRANCE.
A, soluble TRANCE binds to membranebound TRANCE-R, leading to its aggregation. TRAF6 and c-Src are recruited to the TRANCE-R complex. B, Cbl recruits PI3-K to the TRANCE-R complex. C, c-Src, activated by its association with TRAF6, phosphorylates PI3-K, activating it to phosphorylate membrane phosphatidylinositides. Akt is recruited to these phosphatidylinositides via its pleckstrin homology domain and is activated. D, Cbl acts as an E3 ubiquitin ligase, leading to the ubiquitination and subsequent degradation of one or more components of the TRANCE-R-signaling complex, quenching the activating signal. PIP 2 , phosphatidylinositol 4,5-bisphosphate; PIP 3 , phosphatidylinositol 1,4,5-triphosphate; PTK, protein-tyrosine kinase. that this Cbl-b-independent Akt activation is due to the expression of new gene products induced by TRANCE or CD40L in DCs, since other signaling pathways activated by these cytokines appear to be intact in Cbl-b-deficient cells. NF-B activation, as measured by decreasing and subsequently increasing IB levels, follows identical kinetics in Cbl-b Ϫ/Ϫ and wild-type DCs. Furthermore, the up-regulation of IB observed in these cells at the 3-h time point suggests that the expression of other proteins that could potentially activate Akt is upregulated. Attempts to inhibit this second wave of Akt activation by blocking new gene transcription via the addition of cycloheximide were unsuccessful, as even extremely low doses of cycloheximide (Ͻ50 ng/ml) completely abrogated even the first wave of Akt activation (data not shown). Therefore, the contribution of Akt activation to TRANCE-and CD40L-mediated DC survival remains to be determined.
Although Akt is principally known as a regulator of cell survival, it is possible that it may serve other roles as well. Meili et al. (43) show that Akt plays an essential role in cell motility in chemoattractant responses in Dictyostelium. In particular, Akt has effects on actin-mediated cytoskeletal rearrangements. Since dendritic cells, when activated, are quickly mobilized to migrate from outer tissues to draining lymph nodes, it is possible that Akt activation by TNF family proteins or other inflammatory mediators such as interleukin-1 and lipopolysaccharide (6) play a role in DC migration. Indeed, CD40L-CD40 (44) and lipopolysaccharide-Toll-like receptor (45) interactions appear to be required for dendritic cell migration in vivo. We did not observe significant differences in DC migration in Cbl-b-deficient mice (data not shown), but again, the later wave of Akt activation in Cbl-b Ϫ/Ϫ DCs could be sufficient to allow DC migration within the experimental time frame. Given the importance of Akt in a variety of cell functions, further study is warranted to determine its role in DC biology.
Conclusion-TRANCE and CD40L activate Akt in a variety of cell types. In dendritic cells, the TRANCE-R and CD40 signaling complexes recruit TRAF6, Src family kinases, PI3-K, and Cbl in a ligand-dependent manner. The association of TRANCE-R and c-Cbl and Cbl-b is dependent on Src kinase activity, and TRAF6 can enhance Src-mediated Cbl phosphorylation. Cbl-b appears to down-regulate TRANCE-R expression in a Src-dependent manner. In c-Cbl-and Cbl-b-deficient mice, there are cell type-specific defects in Akt activation downstream of TRANCE-R and CD40, indicating that Cbl proteins may be required for TRANCE-and CD40L-dependent PI3-K activation. In Fig. 5, we propose a model in which Cbl brings PI3-K to the receptor complex, where it is activated by c-Src with rapid kinetics. Concurrently, but with slightly slower kinetics, Cbl acts as a ubiquitin ligase, leading to the degradation of one or more of the essential components of the signaling complex, quenching the activating signal. Cbl proteins thereby may act as both positive and negative regulators of TRANCE and CD40L signaling in a kinetically controlled manner.