The B Cell Antigen Receptor Activates the Akt (Protein Kinase B)/Glycogen Synthase Kinase-3 Signaling Pathway Via Phosphatidylinositol 3-Kinase¹

Engagement of the B cell Ag receptor (BCR)⁴ by Ags initiates signals that can promote apoptosis, anergy, survival, activation, or proliferation, depending upon the differentiation state of the B cell, the magnitude, and duration of the BCR signal, and whether or not the B cell receives additional signals from CD40 or cytokine receptors (1, 2). When clustered by multivalent Ags or by anti-Ig Abs, the BCR activates three key signaling enzymes: Ras, phospholipase C-γ (PLC-γ), and phosphatidylinositol (PI) 3-kinase (1, 3). Although PI 3-kinase had been shown to contribute to the ability of the BCR to regulate the proliferation of normal B cells, as well as a human B lymphoma cell line (4, 5), until recently little was known about the downstream targets of PI 3-kinase signaling in B cells.

Recent work has revealed that PI 3-kinase plays a key role in multiple cellular processes. PI 3,4,5-trisphosphate (PIP₃), a plasma membrane phospholipid produced by PI 3-kinase, is a ligand for the pleckstrin homology (PH) domains that are found in a variety of cytosolic proteins, many of which are involved in signal transduction (6). By binding PH domains, PIP₃ can recruit PH domain-containing proteins to the plasma membrane. This is of particular importance for cytosolic signaling enzymes that act on membrane-associated substrates. In addition to recruiting signaling proteins to the plasma membrane, the binding of PIP₃ to a protein’s PH domain may induce conformational changes that affect its enzymatic activity, its ability to act as a substrate for kinases, or its ability to interact with other proteins.

In B cells, PIP₃-PH domain interactions play an essential role in the ability of the BCR to stimulate PLC-γ-dependent signaling. The Btk tyrosine kinase, which phosphorylates and activates PLC-γ, is recruited to the plasma membrane via the binding of its PH domain to PIP₃ (7, 8, 9, 10). PIP₃-PH domain interactions are also involved in activation of the Rac1 GTPase, which is required for sustained hydrolysis of PI 4,5-bisphosphate by PLC-γ (11). Rac1-GTP activates PI 4-phosphate 5-kinases, which are responsible for providing a continuous supply of PI 4,5-bisphosphate to act as a substrate for PLC-γ (12, 13, 14). Both Sos and Vav can act as exchange factors that activate Rac1, and their ability to do so is dependent on the binding of PIP₃ to their PH domains (15, 16). In addition to Btk, SOS, and Vav, a large number of other PH domain-containing proteins have been identified, many of them cytosolic signaling proteins (6). Thus, it is likely that the activation of PI 3-kinase by the BCR and the subsequent production of PIP₃ regulates many other signaling pathways in addition to the PLC-γ pathway.

There is now considerable evidence that a 60-kDa serine/threonine protein kinase called Akt (or protein kinase B) is a major downstream target of PI 3-kinase signaling. Akt is activated by a number of receptors that activate PI 3-kinase, including the receptors for insulin, platelet-derived growth factor, epidermal growth factor, IL-2, IL-3, IL-4, GM-CSF, and stem cell factor (17, 18, 19, 20, 21, 22, 23, 24). In these systems, it has been clearly shown that activation of PI 3-kinase is both necessary and sufficient for Akt to be activated. Receptor-induced activation of Akt is blocked by PI 3-kinase inhibitors and by expression of dominant negative forms of PI 3-kinase (17, 18, 19, 20, 21, 22, 23, 24). Conversely, expression of constitutively active forms of PI 3-kinase in cells results in activation of Akt (25).

Considerable progress has been made toward understanding how PI 3-kinase activates Akt (reviewed in Ref. 26). Akt has a PH domain that binds the PI 3-kinase-derived lipids PIP₃ and PI 3,4-bisphosphate (PI(3, 4)P₂) (27). This interaction recruits Akt to the plasma membrane and induces a conformational change that allows it to be phosphorylated and activated by upstream kinases (28). In response to receptor engagement, human Akt is phosphorylated on threonine 308 and serine 473. Phosphorylation of both of these residues is required for maximal activation of Akt. Changing threonine 308 to an alanine completely blocks activation of Akt, while changing serine 473 to an alanine significantly compromises Akt activation (19, 29, 30). A kinase called PDK1 has been isolated that phosphorylates threonine 308, but not serine 473 (31). PDK1 also contains a PH domain, and the production of PIP₃ by PI 3-kinase is required to recruit PDK1 to the plasma membrane so that it can phosphorylate Akt (32). The kinase that phosphorylates serine 473 of Akt has been termed PDK2. Although it has yet to be identified definitively, there is some evidence that the integrin-linked kinase (ILK) can phosphorylate this site on Akt (33).

Akt is a multifunctional mediator of PI 3-kinase-dependent signaling. First, PI 3-kinase activity is essential for the prevention of apoptosis in a number of cell types (34, 35, 36) and this appears to be mediated by Akt (reviewed in Refs. 26, 37, 38). Expression of a constitutively active form of Akt is sufficient to protect cells from apoptosis caused by growth factor withdrawal, while dominant-negative forms of Akt can cause apoptosis (36, 39). Several groups have proposed that Akt prevents apoptosis by phosphorylating Bad (22, 40), a death-promoting member of the Bcl-2 family. This remains controversial, however, since other kinases can apparently phosphorylate Bad (24) and the physiological relevance of Bad phosphorylation has yet to be established. Thus, there are likely to be additional mechanisms by which Akt can prevent apoptosis. Consistent with this idea, Cardone et al. (41) showed that Akt can phosphorylate caspase-9 and prevent its proteolytic activation. Recent work has shown that Akt may also prevent apoptosis by phosphorylating forkhead family transcription factors (42). Phosphorylation of these transcription factors causes them to be exported from the nucleus and prevents them from inducing the expression of proapoptotic genes, such as the Fas ligand.

Another important function of Akt may be to promote cell cycle progression. Expressing a constitutively active form of Akt in a T cell line has been shown to stimulate the activity of E2F, a transcription factor that induces the expression of cyclin D3 and other genes required for the G₁ to S transition (43). Consistent with a role for Akt in cell growth and proliferation, Akt may also increase the rate of protein synthesis by either directly or indirectly stimulating phosphorylation of the p70^S6K kinase (21) and 4E-BP1 (44). p70^S6K phosphorylates the ribosomal S6 protein, and this preferentially increases the translation of mRNAs containing 5′-terminal oligopolypyrimidine tracts (45). Phosphorylation of 4E-BP1 releases the eIF-4E translation initiation factor and allows it to promote translation (46). Recent work has shown that activated forms of Akt also selectively enhance the translation of cyclin D1 (47). Thus, Akt may promote cell cycle progression at both the transcriptional and posttranscriptional levels.

Another major downstream target of Akt is glycogen synthase kinase-3 (GSK-3), a constitutively active serine/threonine kinase whose activity is inhibited by Akt. Regulation of GSK-3 activity has been studied primarily in the context of insulin receptor signaling. Insulin receptor signaling leads to a reduction in GSK-3 activity (20, 48, 49), and this appears to be mediated by Akt, since a dominant-negative form of Akt can block the ability of insulin to decrease GSK-3 activity (48). In vitro, Akt can phosphorylate GSK-3α on serine 21 and GSK-3β on serine 9, and these modifications correlate with inhibition of GSK-3 enzyme activity (20). Thus, GSK-3α and GSK-3β may be physiological substrates of Akt.

GSK-3-mediated phosphorylation may regulate a number of different cellular processes. First, GSK-3 may regulate the subcellular localization of the NF-AT_c transcription factor. The BCR and other receptors that cause increases in intracellular Ca²⁺ concentrations activate the Ca²⁺-dependent phosphatase calcineurin, which dephosphorylates NF-AT_c (50). This reveals a nuclear localization signal that allows NF-AT_c to translocate into the nucleus, bind to AP-1-like transcription factors, and stimulate transcription. GSK-3 opposes this process by phosphorylating NF-AT_c. This causes a conformational change that reveals a nuclear export signal, resulting in NF-AT_c being rapidly exported from the nucleus (51, 52). Thus, Akt-dependent inhibition of GSK-3 activity may be required for NF-AT_c to accumulate in the nucleus and promote transcription. A number of other potential GSK-3 substrates have been identified (53), including the microtubule-associated protein tau (54), the ε subunit of the eIF-2B translation initiation factor (55), and β-catenin (56). Phosphorylation of β-catenin promotes its degradation (57). Inhibition of GSK-3 by Wnt family receptors has been shown to play an important role in development by stabilizing β-catenin and allowing β-catenin/LEF-1 transcription factor complexes to accumulate in the nucleus (56). Recent work has also suggested that GSK-3-mediated phosphorylation of yet unidentified substrates can promote apoptosis and that Akt-dependent inhibition of GSK-3 activity promotes cell survival (58). Thus, like Akt, GSK-3 may be involved in a variety of important cellular functions, including transcription, translation, and cell survival.

Although BCR engagement activates PI 3-kinase (4) and causes a significant increase in the levels of PI 3-kinase-derived lipids in B cells (59), the effects of BCR signaling on Akt and GSK-3 have not been examined. Since Akt and GSK-3 may regulate important cell functions, we asked whether these enzymes were targets of BCR signaling. In this report, we show that both Akt and GSK-3 are regulated by the BCR in a PI 3-kinase-dependent manner. Thus, BCR ligation initiates a PI 3-kinase/Akt/GSK-3 signaling pathway.

Goat Abs specific for mouse IgM, mouse IgG, and human IgM were obtained from Bio-Can (Mississauga, Ontario, Canada). Goat Abs against murine κ light chain were from Southern Biotechnology Associates (Birmingham, AL). The 6E4 mAb to chicken Ig light chain (60) was a gift of M. Ratcliffe (McGill University, Montreal, Quebec, Canada). Wortmannin and LY294002 were from Biomol (Plymouth Meeting, PA). Microcystin-LR was obtained from Alexis (San Diego, CA). Avidin-conjugated agarose was from Pierce (Rockford, IL).

The WEHI-231 and A20 murine B cell lines, as well as the RAMOS human B cell line, were grown in RPMI 1640 supplemented with 10% heat-inactivated FCS (Intergen, Purchase, NY), 50 μM 2-ME, 1 mM pyruvate, and 2 mM glutamine. To reduce the signaling contribution from serum-derived growth factors, A20 cells (and in some experiments, WEHI-231 cells) were washed twice with RPMI 1640 and then cultured for 12–18 h at 37°C in serum-free medium (RPMI 1640 supplemented with 0.5 mg/ml BSA, 50 μM 2-ME, 1 mM pyruvate, and 2 mM glutamine) before being stimulated. The wild-type DT40 chicken B cell line, as well as the Lyn-deficient (61), Syk-deficient (61), Btk-deficient (62), and PLC-γ2-deficient (63) variants of this cell line, were grown in RPMI 1640 supplemented with 10% heat-inactivated FCS, 1% heat-inactivated chicken serum (Life Technologies, Grand Island, NY), 50 μM 2-ME, and 4 mM glutamine. Small resting B cells were isolated from the spleens of C57BL/6 mice by Percoll density centrifugation after Ab- and complement-mediated lysis of T cells. Briefly, spleen cells were incubated at 37°C for 45 min with a 1:4 dilution of guinea pig complement (Life Technologies) and 1:4 dilutions of culture supernatants from the HO13.4 anti-Thy1 hybridoma (American Type Culture Collection (ATCC), Manassas, VA), the 3.155 anti-CD8 hybridoma (ATCC), and the 2B6 2D8 anti-CD4 hybridoma (from D. Hanson, Washington University, St. Louis, MO). The small resting B cells were recovered from the interface of 60% and 75% isotonic Percoll layers. Flow cytometry showed that the resulting population of cells was >90% IgM⁺ B cells.

Akt cDNA was cloned by PCR using HeLa cell cDNA as a template and primers that added the influenza hemagglutinin (HA) epitope tag at the C terminus (18). The HA-Akt was originally cloned into the pECE expression vector. For expressing the HA-Akt in WEHI-231 cells, we subcloned the HA-Akt cDNA into the pLXSN retroviral expression vector. The HA-Akt was excised from pECE by digesting with EcoRI and BamHI and then ligated into pLXSN that had been digested with the same enzymes. The resulting HA-Akt-pLXSN plasmid was transfected into the BOSC23 packaging cell line (64) using the calcium phosphate precipitation method, and the resulting retrovirus particles were used to infect WEHI-231 cells. A detailed protocol for retrovirus-mediated gene transfer into WEHI-231 cells has recently been described by Krebs et al. (65). Two days postinfection, the WEHI-231 cells were transferred into medium containing 1.8 mg/ml G418 (Life Technologies) and cloned by limiting dilution. G418-resistant colonies were expanded and HA-Akt expression was assessed by immunoblotting with a biotinylated form of the 12CA5 anti-HA mAb (Boehringer Mannheim, Laval, Quebec, Canada), followed by HRP-conjugated streptavidin (Amersham Pharmacia Biotech, Piscataway, NJ).

The cells were washed once with modified HEPES-buffered saline (25 mM sodium HEPES (pH 7.2), 125 mM NaCl, 5 mM KCl, 1 mM CaCl₂, 1 mM Na₂HPO₄, 0.5 mM MgSO₄, 1 mg/ml glucose, 2 mM glutamine, 1 mM sodium pyruvate, 50 μM 2-ME) and then resuspended in this buffer at 10⁷/ml for tissue culture cells or 2.5 × 10⁷/ml for splenic B cells. The cells were warmed to 37°C for 30 min and then stimulated with anti-Ig Abs. In some experiments, the cells were incubated with the PI 3-kinase inhibitors wortmannin or LY294002 for 20 min at 37°C before the addition of anti-Ig Abs. Reactions were terminated by adding ice-cold PBS containing 1 mM Na₃VO₄ and then centrifuging the cells in the cold for 15 s at full speed in a microfuge. The cell pellets were washed once, without resuspending, with cold PBS/Na₃VO₄ and then solubilized. For immunoblotting or immunoprecipitation experiments, the cell pellets were solubilized in buffer A (20 mM Tris-HCl (pH 8), 1% Triton X-100, 137 mM NaCl, 2 mM EDTA, 10% glycerol, 1 mM PMSF, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 1 mM Na₃VO₄, 25 mM β-glycerophosphate). After 10 min on ice, detergent-insoluble material was removed by centrifugation, and protein concentrations for the detergent-soluble supernatant fractions were determined using the bicinchoninic acid assay (Pierce). These samples were stored at −80°C until analyzed. For in vitro kinase assays, the cell pellets were solubilized in buffer B (50 mM Tris-HCl (pH 7.7), 0.5% Nonidet P-40, 2.5 mM EDTA, 20 mM β-glycerophosphate, 10 mM NaF, 1 mM Na₂MoO₄, 1 mM Na₃VO₄, 0.25 μM PMSF, 1 μM pepstatin, 0.5 μg/ml leupeptin, 10 μg/ml soybean trypsin inhibitor, 1 μg/ml microcystin-LR), detergent-insoluble material was removed by centrifugation, and the detergent-soluble supernatant fractions were used immediately for in vitro kinase assays.

Transfected variants of the AtT20 endocrine cell line that express the complete BCR (100.33 cells) or the complete BCR plus the Syk tyrosine kinase (Syk13 cells) have been described previously (66, 67). These cells were cultured in DMEM supplemented with 10% FCS and grown to near confluence in 10-cm tissue culture dishes. Before being stimulated, the cells were washed with PBS and cultured overnight under low serum conditions (DMEM with 0.2% FCS). The cells were then washed with PBS and incubated for 15 min at 37°C in modified HEPES-buffered saline (see above) to further reduce any signaling due to serum growth factors. The cells were washed again with PBS, and 10 ml of 37°C modified HEPES-buffered saline was added to each dish. BCR signaling was initiated by adding goat anti-mouse IgM Abs to a final concentration of 20 μg/ml. Reactions were terminated by aspirating the medium, washing the cells twice with ice-cold PBS containing 1 mM Na₃VO₄, and then solubilizing the cells with 1 ml buffer A (see above). After 20 min on a rocker in the cold, the cell lysate was collected, detergent-insoluble material was removed by centrifugation, and protein concentrations were determined using the BCA assay.

Cell lysates (60 μg protein) were separated on 10% SDS-PAGE gels and transferred to nitrocellulose membranes. For detection of phosphorylated Akt, the membranes were blocked for 2 h at room temperature with 5% (w/v) nonfat dry milk powder in TBST (10 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.1% Tween 20). The filters were then rinsed three times with TBST before being incubated overnight in the cold with anti-phospho-Akt Abs (New England Biolabs, Mississauga, Ontario, Canada), which were diluted 1:1000 in TBST plus 5% BSA. One Ab specifically recognizes Akt that is phosphorylated on serine 473, while the other Ab specifically recognizes Akt that is phosphorylated on threonine 308. After the overnight incubation, the filters were washed three times for 10 min with TBST and then incubated for 1 h at room temperature with HRP-conjugated goat anti-mouse IgG Abs (Bio-Rad, Richmond, CA) diluted 1:20,000 in TBST plus 5% nonfat dry milk powder. The membranes were washed extensively with TBST, and immunoreactive bands were visualized by enhanced chemiluminescence detection (Amersham Pharmacia Biotech). To reprobe these blots, bound Abs were eluted by incubating the blots for 1 h with 10 mM Tris-HCl (pH 1.6)/150 mM NaCl. The membranes were blocked, as described above, and then incubated overnight in the cold with an anti-Akt Ab (New England Biolabs) diluted 1:1250 in TBST plus 5% BSA. Immunoreactive bands were visualized as described for the anti- phospho-Akt Abs.

Detection of phosphorylated GSK-3β by immunoblotting was performed in a similar manner with the following three exceptions. First, the concentration of Tween 20 in the TBST was reduced to 0.05% for all buffers. Second, the primary Ab was a rabbit polyclonal Ab that specifically recognizes GSK-3β that is phosphorylated on serine 9 (Chemicon, Temecula, CA). This Ab was diluted 1:300 in TBST. Finally, the blots were reprobed with an anti-GSK-3β mAb (Chemicon; diluted 1:1000 in TBST), which was visualized using HRP-conjugated sheep anti-mouse IgG (Amersham; diluted 1:10,000 in TBST) and enhanced chemiluminescence detection.

Akt in vitro kinase assays were performed as described by Scheid and Duronio (24). Akt was immunoprecipitated from cell lysates using 2 μg of anti-PKB-α Ab (Upstate Biotechnology, Lake Placid, NY; catalogue no. 06-558) that had been prebound to 20 μl of protein G-Sepharose beads (Sigma, St. Louis, MO). After mixing the cell lysates with the immobilized anti-PKB-α Abs for 1 h at 4°C, the beads were washed once with fresh buffer B, three times with buffer B containing 500 mM NaCl, and once with Akt kinase reaction buffer (20 mM HEPES (pH 7.4), 1 mM EGTA, 1 mM Na₃VO₄, 1 mM DTT, 0.25 mM PMSF, 1 μg/ml microcystin-LR). The beads were then resuspended in 20 μl of Akt kinase reaction buffer containing 60 μM Crosstide (Upstate Biotechnology) as a substrate. Reactions were initiated by adding 10 μl of Akt kinase reaction buffer containing 500 μM ATP, 75 mM MgCl₂, and 10 μCi [γ-³²P]ATP. After 15 min at 30°C, 20 μl of the reaction mixture was spotted onto a 2 × 2-cm square of Whatman (Tewksbury, MA) p81 phosphocellulose chromatography paper. The filter papers were washed extensively with 1% phosphoric acid to remove free [γ-³²P]ATP. Radioactivity that remained bound to the filters was quantitated by liquid scintillation counting in the presence of scintillation fluor.

Cell lysates were mixed with 2 μg of anti-GSK-3α Ab (Upstate Biotechnology) for 1 h at 4°C. Immune complexes were recovered by adding 20 μl of protein G-Sepharose beads and mixing for an additional hour. The beads were then washed four times with fresh buffer B and once with GSK-3 kinase reaction buffer (5 mM HEPES (pH 7.4), 0.5 mM EDTA, 1 mM Na₃VO₄, 1 μg/ml microcystin-LR) before being resuspended in 30 μl GSK-3 kinase reaction buffer containing 75 μM phosphoglycogen synthase peptide-2 (Upstate Biotechnology). Reactions were initiated by adding 10 μl of GSK-3 kinase reaction buffer containing 500 μM ATP, 75 mM MgCl₂ and 5 μCi [γ-³²P]ATP. After 15 min at 30°C, 25 μl of the reaction mixture was spotted onto a 2 × 2-cm square of Whatman p81 chromatography paper. The filter papers were washed with 1% phosphoric acid and counted.

Phosphorylation of Akt on threonine 308 and serine 473 are essential for its activation (19, 29, 30). Thus, immunoblotting with an anti-peptide Ab that specifically recognizes Akt that is phosphorylated on serine 473 can be used as an indirect measure of Akt activation (68, 69). Such phosphorylation state-specific Abs are now commonly used to assess the activation state of a number of kinases that are activated by phosphorylation, including the ERK, JNK, and p38 mitogen-activated protein kinases. To determine whether BCR engagement might activate Akt, we stimulated B cell lines with anti-Ig Abs for various times and then performed anti-phospho-Akt immunoblots. Fig. 1, A–D, shows that Akt was not phosphorylated to a significant extent in unstimulated cells, but that BCR ligation caused a substantial increase in phosphorylation of Akt on serine 473. This response was observed in the immature murine B cell line WEHI-231, the IgG⁺ murine B cell line A20, the RAMOS human B cell line, and the DT40 chicken B cell line. In WEHI-231 cells, phosphorylation of Akt peaked at 1–5 min and then declined to lower levels at 15–60 min (Fig. 1,A). In contrast, BCR-induced phosphorylation of Akt was sustained at high levels for 30–60 min in the A20, RAMOS, and DT40 cell lines. BCR engagement also stimulated phosphorylation of Akt on serine 473 in mature resting B cells isolated from mouse spleen (Fig. 1 E). In anti-IgM-stimulated splenic B cells, Akt phosphorylation was very strong at 1.5–15 min and then declined to lower levels at 30–60 min. Although there were differences in the duration of the response, our results clearly show that Akt phosphorylation is a consistent characteristic of BCR signaling.

To confirm that the protein being recognized by the anti-phospho-Akt Ab was in fact Akt, as opposed to an unrelated comigrating protein, we expressed an epitope-tagged version of Akt in WEHI-231 cells. Anti-Akt immunoblotting of total cell lysates from this stable WEHI-231 clone showed that the HA-tagged Akt was expressed at similar levels as the endogenous Akt protein (Fig. 2, lower panel). The HA tagged-Akt was then immunoprecipitated using the 12CA5 anti-HA mAb and analyzed by sequential immunoblotting with the anti-phospho-Akt Ab and the anti-Akt Ab. Fig. 2 shows that BCR ligation induced phosphorylation of the HA-Akt on serine 473. The anti-HA-precipitated protein that was recognized by both the anti-phospho-Akt Ab and the anti-Akt Ab had the same electrophoretic mobility as the HA-Akt present in the cell lysates. This result confirms that BCR engagement leads to phosphorylation of Akt at a site that plays a role in Akt activation. It is worth noting that, in response to BCR ligation, the HA-tagged Akt was phosphorylated on serine 473 to a much lesser extent than the endogenous Akt, even though they were expressed at similar levels. The C-terminal HA tag is very close to serine 473 (18) and may interfere with either the ability of PDK2 to phosphorylate Akt on this residue or with the ability of the anti-phospho-Akt Ab to bind to this site. Thus, in all additional experiments, we analyzed only the endogenous Akt protein.

The current model for Akt activation involves recruitment of Akt to the plasma membrane by PIP₃ and/or PI(3, 4)P₂, followed by phosphorylation of Akt at threonine 308 by PDK1 and at serine 473 by PDK2 (26, 70). PIP₃ is produced only by PI 3-kinase while PI(3, 4)P₂ is thought to be produced primarily by the dephosphorylation of PIP₃. To test whether BCR-induced phosphorylation of Akt on serine 473 is dependent on activation of PI 3-kinase by the BCR, we asked whether this response could be blocked by the PI 3-kinase inhibitor wortmannin (71). Fig. 3 shows that a 20-min pretreatment with 30 nM wortmannin completely blocked BCR-induced phosphorylation of Akt on serine 473 in both splenic B cells and in the A20 B cell line. Thus, the ability of PDK2 to phosphorylate Akt on this residue in response to BCR engagement is a PI 3-kinase-dependent process.

Although phosphorylation of Akt on serine 473 correlates with maximal activation of Akt, it was important to directly demonstrate that BCR engagement increases the enzymatic activity of Akt. To do this, we immunoprecipitated Akt from RAMOS cell lysates and then performed in vitro kinase assays using a substrate called “Crosstide,” a peptide that contains the sequence from GSK-3 that is phosphorylated by Akt (20). We found that stimulating RAMOS cells for 2–15 min with anti-IgM Abs routinely caused a 6- to 8-fold increase in Akt enzyme activity (Fig. 4). Akt activation was maximal at 2 and 5 min after addition of anti-IgM Abs and began to decline at 15 min. We also found that BCR-induced activation of Akt was dependent on PI 3-kinase activity. The PI 3-kinase inhibitors wortmannin and LY294002 both completely blocked the ability of the BCR to activate Akt (Fig. 4). This confirms that Akt is a downstream target of PI 3-kinase signaling in B cells.

Previous work has shown that activation of Akt requires that it be phosphorylated on threonine 308 by PDK1 and on serine 473 by PDK2. Phosphorylation of threonine 308 may be more important for Akt activation than phosphorylation of serine 473, since changing threonine 308 to an alanine completely ablates Akt activation, while changing serine 473 to an alanine reduces Akt activation substantially, but not completely (19, 29, 30). Our finding that BCR ligation increases the enzymatic activity of Akt suggested that BCR engagement leads to phosphorylation of Akt on threonine 308 as well as on serine 473. At the end of our study, an Ab that specifically recognizes Akt that is phosphorylated on threonine 308 became available, allowing us to directly test whether BCR ligation stimulates phosphorylation of Akt at this site. We found that BCR engagement resulted in increased phosphorylation of Akt on threonine 308 in both RAMOS cells (Fig. 5,A) and small resting B cells from mouse spleen (Fig. 5,C). In the RAMOS cells, phosphorylation of Akt on threonine 308 was sustained for at least 60 min after addition of anti-IgM to the cells (Fig. 5,A), similar to what was observed for phosphorylation of serine 473 in these cells (see Fig. 1,C). BCR-induced phosphorylation of Akt on threonine 308 was completely blocked by both wortmannin and LY294002 (Fig. 5 B), indicating that this response is dependent on PI 3-kinase. Since PDK1 is the only known kinase that can phosphorylate Akt on threonine 308 (31), our results suggest that BCR engagement activates PDK1.

Phosphorylation of Akt on serine 473 is mediated by a kinase that has been termed PDK2, although its identity has not been firmly established. Using the phosphorylation of Akt on serine 473 as our assay, we investigated some of the upstream requirements for regulation of PDK2 by the BCR. In particular, we focused on the role of the different BCR-associated tyrosine kinases in promoting the phosphorylation of Akt on serine 473.

The BCR activates multiple members of the Src family of tyrosine kinases, as well as the Syk and Btk tyrosine kinases (1). Considerable evidence suggests that these three families of tyrosine kinases play distinct roles in coupling the BCR to various signaling pathways (72, 73). To investigate the role of these tyrosine kinases in activation of the PDK2/Akt pathway by the BCR, we made use of three variants of the DT40 chicken B cell line, one in which the genes encoding Lyn were disrupted (61), one in which the genes encoding Syk were disrupted (61), and one in which the genes encoding Btk were disrupted (62). Since Lyn is the only Src family tyrosine kinase expressed in DT40 cells (61), the Lyn-deficient cells are therefore devoid of all Src kinase activity.

Fig. 6,A shows that the BCR-induced phosphorylation of Akt on serine 473 was completely dependent on the presence of the Src family kinase Lyn. In the Lyn-deficient DT40 cells, Akt was not phosphorylated at all on serine 473 in response to BCR engagement. In contrast, the BCR could induce some phosphorylation of Akt on serine 473 in the Syk-deficient cells, although this response was less robust and was not sustained for as long as in the wild-type cells (Fig. 6,B). In the wild-type DT40 cells, BCR engagement resulted in strong phosphorylation of Akt that persisted for at least 30 min (see Fig. 1,B). However, in the Syk-deficient DT40 cells, BCR engagement consistently resulted in modest phosphorylation of Akt that peaked at 5 min and returned to basal levels by 15 min (Fig. 6 B). Thus, in DT40 cells, Lyn is absolutely required for activation of the PDK2/Akt pathway by the BCR, while Syk is required for maximal phosphorylation of Akt, as well as for sustaining this response.

The Btk tyrosine kinase is a downstream target of PI 3-kinase in that production of PIP₃ by PI 3-kinase is required for Btk to translocate to the plasma membrane and phosphorylate physiological substrates, such as phospholipase C-γ (7, 8, 9, 10, 62). Thus, we considered the possibility that Btk might link PI 3-kinase to PDK2, the kinase that phosphorylates Akt on serine 473. Fig. 6 C, however, clearly shows that BCR-induced phosphorylation of Akt on serine 473 is completely normal in Btk-deficient DT40 cells. Thus, Btk is not involved in the activation of PDK2 by the BCR.

We also showed that DT40 cells lacking PLC-γ2, the only isoform of PLC-γ expressed in these cells (63), exhibited normal Akt phosphorylation in response to BCR engagement (Fig. 6 D). This supports the idea that BCR-induced phosphorylation of Akt on serine 473 is dependent solely on the activation of PI 3-kinase and does not involve increases in intracellular Ca²⁺ or the activation of kinases that are regulated by PLC-derived second messengers (e.g., protein kinase C enzymes).

The DT40 experiments described above suggested that Src family tyrosine kinases, such as Lyn, are both necessary and sufficient for BCR-induced activation of the PDK2/Akt, while Syk is required for sustaining and amplifying this response. To provide further evidence to support this conclusion, we made use of several variants of the AtT20 endocrine cell line that express a transfected murine BCR (μ, λ, Ig-α, Ig-β). Of the tyrosine kinases relevant for BCR signaling, the BCR-expressing AtT20 cells (100.33 cells) express only Fyn (67), a Src family tyrosine kinase. We have also established a BCR-expressing AtT20 cell line, Syk 13, that expresses a transfected Syk gene in addition to the endogenous Fyn. Using the 100.33 and Syk13 cells, we asked whether the Src family tyrosine kinase Fyn was sufficient for BCR-induced activation of the PDK2/Akt pathway and whether Syk influenced this response.

In the Fyn-expressing 100.33 cells, we found that BCR engagement led to a transient increase in Akt phosphorylation on serine 473 (Fig. 7). Upon addition of anti-IgM Abs to these cells, Akt phosphorylation increased slowly, peaking at 15–30 min and then declining. In contrast, BCR-induced Akt phosphorylation in the Syk13 cells (which express both Fyn and Syk) reached maximal levels after 3 min and remained high for at least 60 min. Thus, while the Src family tyrosine kinase Fyn was sufficient for BCR-induced Akt phosphorylation in AtT20 cells, Syk was required to sustain Akt phosphorylation at high levels. These data are consistent with our results using the kinase-deficient DT40 cell variants. In both DT40 cells and AtT20 cells, a Src family tyrosine kinase is sufficient for BCR-induced phosphorylation of Akt on serine 473, while Syk is required for sustaining this response. Thus, Src kinases and Syk are both involved in BCR-induced activation of the PDK2/Akt pathway, but appear to play different roles in this process.

We also used the 100.33 and Syk13 variants of the AtT20 cells to examine the kinase requirements for BCR-induced phosphorylation of Akt on threonine 308 (Fig. 7). The results obtained were very similar to those observed for phosphorylation of Akt on serine 473. The Src family tyrosine kinase Fyn was sufficient for BCR-induced phosphorylation of Akt on threonine 308, while Syk was required to sustain this response. Although the Ab that specifically recognizes Akt that is phosphorylated on threonine 308 did not react with Akt from the DT40 chicken B cell line, the results from the AtT20 cells suggest that regulation of both PDK1 and PDK2 by the BCR involves Syk, as well as a Src family tyrosine kinase.

In insulin-responsive cells, GSK-3 is a downstream target of Akt. Insulin receptor signaling decreases GSK-3 activity, and this response can be blocked by expression of a dominant-negative form of Akt (48). In vitro, both the α and β isoforms of GSK-3 can be phosphorylated by Akt, and this results in a decrease in GSK-3 enzyme activity (20). Taken together, these data suggest that GSK-3 may be a physiological substrate of Akt. Since GSK-3 is involved in a number of important cellular functions, including regulation of NF-AT_c, our finding that the BCR activated Akt prompted us to examine the effects of BCR engagement of GSK-3 activity and phosphorylation.

We immunoprecipitated GSK-3α using an Ab that specifically recognizes only the α isoform of GSK-3 and then performed in vitro kinase assays using a peptide derived from glycogen synthase as a substrate. We found that treating A20 cells with anti-IgG Abs for 2–15 min caused an ∼50% decrease in GSK-3α activity (Fig. 8,A). BCR ligation also decreased GSK-3α activity in the RAMOS human B cell line (Fig. 8,B), although to a lesser extent (25–35%) than in the A20 cells. The inhibition of GSK-3 activity caused by the BCR is similar in magnitude to that caused by insulin (20, 49) or epidermal growth factor (74), both of which have been shown to reduce total cellular GSK-3 activity by 40–50%. The ability of the BCR to decrease GSK-3α activity in the A20 cells was completely blocked by the PI 3-kinase inhibitor wortmannin (Fig. 8 A). Thus, GSK-3α is a downstream target of PI 3-kinase signaling in B cells.

To determine whether GSK-3β is also a target of BCR signaling, we made use of a recently developed anti-peptide Ab that specifically recognizes GSK-3β that is phosphorylated on serine 9. This serine residue is phosphorylated by Akt in vitro, and phosphorylation of this residue correlates with an inhibition of GSK-3β activity (20). Fig. 9 shows that BCR engagement resulted in substantial phosphorylation of GSK-3β. This response was observed in the A20 murine B cell line, in the RAMOS human B cell line, and in murine splenic B cells. BCR-induced phosphorylation of GSK-3β was inhibited by pretreating the cells with the PI 3-kinase inhibitor wortmannin (Fig. 9, C and D). Thus, both GSK-3α and GSK-3β are PI 3-kinase-dependent targets of BCR signaling.

In this report, we have shown that the BCR regulates the phosphorylation state and the activity of the Akt and GSK-3 kinases in a PI 3-kinase-dependent manner. BCR engagement resulted in phosphorylation of Akt on both threonine 308 and serine 473, modifications that are required for maximal activation of Akt. Moreover, Akt enzyme activity increased 6- to 8-fold after BCR ligation. GSK-3β was also phosphorylated in response to BCR ligation, in this case on serine 9, a residue that can be phosphorylated by Akt. Phosphorylation of GSK-3β at this site correlates with inhibition of its enzymatic activity (20). Finally, we directly showed that BCR engagement caused a significant decrease in the activity of the very closely related GSK-3α isoform. All of these responses were blocked by the PI 3-kinase inhibitors wortmannin or LY294002, indicating that both Akt and GSK-3 are downstream targets of PI 3-kinase-derived lipids. We have previously shown that BCR engagement leads to an increase in the levels of PIP₃ and PI(3, 4)P₂ in B cells (59). Thus, our data indicate that the BCR activates a PI 3-kinase/Akt/GSK-3 signaling pathway.

Based on our current knowledge (26, 70), the likely series of events in this PI 3-kinase/Akt/GSK-3 pathway are: 1) Recruitment of Akt and PDK1 to the plasma membrane via the binding of their PH domains to PIP₃ and/or PI(3, 4)P₂, 2) activation and/or membrane recruitment of PDK2, 3) phosphorylation of Akt on threonine 308 by PDK1 and on serine 473 by PDK2, which results in activation of Akt, and 4) release of Akt into the cytoplasm, where it can phosphorylate GSK-3α on serine 21 and GSK-3β on serine 9, resulting in a decrease in GSK-3 activity. Recently, we have expressed in WEHI-231 cells an estrogen receptor-Akt chimeric protein that can be inducibly activated by addition of 4-hydroxytamoxifen to the medium (75). Preliminary experiments have shown that activation of this estrogen receptor-Akt chimeric protein results in phosphorylation of GSK-3α, supporting the idea that GSK-3α is a downstream target of Akt (B. Hong and M. Gold, unpublished observations).

The use of DT40 chicken B cells and BCR-expressing AtT20 endocrine cells with different subsets of the BCR-regulated tyrosine kinases allowed us to assess the role of Src family kinases, Syk, and Btk in activation of the PI 3-kinase/PDK2/Akt pathway. Together, these studies showed that Src family tyrosine kinases are both necessary and sufficient for activation of this pathway, as judged by phosphorylation of Akt on serine 473. Akt phosphorylation on serine 473 was completely absent in DT40 cells lacking Lyn, the only Src family tyrosine kinase expressed in these cells. Conversely, both DT40 and AtT20 variants that express a Src family tyrosine kinase in the absence of Syk were competent for BCR-induced Akt serine 473 phosphorylation. This latter result indicates that activation of Src kinases by the BCR is sufficient for PI 3-kinase and Akt to be recruited to the plasma membrane and for PDK2 to be activated and/or localized to the membrane. Several lines of evidence support the conclusion that Src family kinases are important for the BCR to recruit PI 3-kinase to the plasma membrane, where its substrates are located. First, Lyn has been shown to associate tightly with the cytoplasmic domain of CD19 (76), and BCR-induced activation of Lyn results in tyrosine phosphorylation of CD19 and the subsequent binding of PI 3-kinase to CD19 (77). PI 3-kinase may also be recruited to the plasma membrane by using its Src homology (SH) 2 domains to bind to other membrane-associated docking proteins that are phosphorylated on appropriate tyrosine residues after BCR engagement. Cbl and Gab1 are two such docking proteins that bind PI 3-kinase in activated B cells (78, 79, 80). Tezuka et al. (81) have shown in DT40 cells that BCR-induced tyrosine phosphorylation of Cbl is dependent on Lyn but not on Syk. Finally, Src kinase activation is accompanied by a significant conformational change that reveals the SH3 domain of the kinase (82). Cambier and colleagues (83) have shown that the SH3 domains of Src kinases (including Lyn and Fyn) can bind to proline-rich regions in the p85 subunit of PI 3-kinase, and, in doing so, increase the enzymatic activity of PI 3-kinase. Thus, there are multiple ways in which Src kinases can act independently of Syk to recruit and activate PI 3-kinase.

While the Src kinases appear to be essential for initiating the PI 3-kinase/PDK2/Akt pathway, the Syk tyrosine kinase is required for maximal phosphorylation of Akt on serine 473 and for sustaining this response. In the Syk-deficient DT40 cells, the maximal level of Akt phosphorylation of serine 473 was considerably lower than in the wild-type DT40 cells. While this difference in the maximal level of Akt phosphorylation was not as evident in the AtT20 cells, there was clearly a Syk-dependent difference in the kinetics of Akt phosphorylation in both the AtT20 cells and the DT40 cells. In the absence of Syk, Akt phosphorylation was transient, whereas it was sustained at high levels for longer periods of time in the cells expressing Syk. The precise nature of this requirement for Syk is not clear. One possibility is that both Syk and the Src kinases recruit PI 3-kinase to the plasma membrane, but by different mechanisms. Src kinase-mediated recruitment of PI 3-kinase to the plasma membrane (by the mechanisms described in the previous paragraph) may be transient, while Syk-mediated phosphorylation of docking proteins, such as Gab1, may be sustained for longer periods of time. Alternatively, since the readout in these experiments was phosphorylation of Akt on serine 473, PDK2, the kinase that phosphorylates this site, may be subject to regulation by both Src kinases and Syk. Another possibility is that two different kinases can phosphorylate Akt at serine 473, one that is regulated by Src kinases, one that is regulated by Syk. Further characterization of the kinases that phosphorylate Akt on serine 473 is required to address this possibility.

Analysis of Akt phosphorylation on threonine 308 in the BCR-expressing AtT20 cells showed that the kinase requirements for activation of PDK1 by the BCR were similar to that for PDK2. As for phosphorylation of serine 473 by PDK2, Src family tyrosine kinases were sufficient for PDK1-mediated phosphorylation of Akt on threonine 308, while Syk was required for sustained phosphorylation of threonine 308. Thus, at least in AtT20 cells, PDK1 and PDK2 appear to be coordinately regulated by both Syk and the Src family tyrosine kinase Fyn. The simplest way to explain this coordinate regulation of PDK1 and PDK2 by Fyn and Syk is that both of these tyrosine kinases are involved in the membrane recruitment and/or activation of PI 3-kinase by the BCR. However, we cannot rule out the possibility that Fyn or Syk regulate the subcellular localization and/or activity of PDK1 and PDK2 in other ways.

In contrast to the role of Src kinases and Syk, we found that the Btk tyrosine kinase was not involved in PDK2-mediated phosphorylation of Akt. BCR-induced phosphorylation of Akt on serine 473 was completely normal in Btk-deficient DT40 cells. Moreover, BCR ligation caused significant and sustained phosphorylation of Akt in AtT20 cells that express Fyn and Syk but not Btk.

While we have shown that the BCR regulates the activity of Akt and GSK-3, the role of these enzymes in B cell development and activation remain to be elucidated. One of the major functions of Akt is to prevent apoptosis. Activation of Akt by PI 3-kinase has been shown to protect neuronal cells from apoptosis (36, 39). Inhibition of GSK-3 may also promote cell survival (58). However, our observations suggest that BCR-induced activation of Akt and the subsequent inhibition of GSK-3 are not sufficient to protect B cells from apoptosis. BCR ligation causes apoptosis in WEHI-231 cells, RAMOS cells, and DT40 cells, despite the fact that the BCR activates Akt and inhibits GSK-3. This implies that the main role of Akt in BCR signaling is not the prevention of apoptosis.

As opposed to the BCR, it is CD40 that usually delivers B cell survival signals. Although CD40 has been reported to activate PI 3-kinase in human B cells (4), our preliminary results show that CD40 does not induce Akt phosphorylation in resting B cells from mouse spleen and induces only a very small amount of Akt phosphorylation (∼10% as much as the BCR) in WEHI-231 cells. Experiments are in progress to test whether CD40 increases Akt enzyme activity in B cells. Nevertheless, our current results suggest that activation of Akt does not play a significant role in the ability of CD40 to prevent B cell apoptosis. Moreover, these results indicate that it is necessary to evaluate whether CD40 activates PI 3-kinase in murine B cells.

While BCR-induced Akt activation may not mediate survival signals, a potential role for Akt in B cells may be to enhance NF-AT_c-dependent transcription by inhibiting GSK-3. GSK-3 has been identified as the kinase that phosphorylates NF-AT_c, resulting in its rapid export from the nucleus (51, 52). Inhibition of GSK-3 may therefore be necessary for NF-AT_c to accumulate in the nucleus and promote transcription. The BCR has been shown to cause NF-AT_c to translocate to the nucleus (84, 85), and NF-AT_c is likely to be a key mediator of Ca²⁺-dependent signaling initiated by the BCR. Inhibition of GSK-3 also promotes the formation and nuclear localization of β-catenin/LEF-1 transcription factor complexes (57). The role of such complexes in BCR signaling has not been investigated.

Another potential role for Akt in BCR signaling is to increase protein synthesis. Akt either directly or indirectly stimulates phosphorylation of 4E-BP1 (44) and the p70^S6K kinase (21), both of which regulate protein translation. 4E-BP1 sequesters the eIF-4E translation initiation factor in an inactive complex. Phosphorylation of 4E-BP1 releases eIF-4E and allows it to promote translation (46). p70^S6K phosphorylates the ribosomal S6 protein, which results in increased translation of mRNAs containing 5′-terminal oligopolypyrimidine tracts (45). Since the mRNAs encoding ribosomal proteins contain 5′-terminal oligopolypyrimidine tracts, ribosome formation and general protein translation would also be increased. Finally, Akt may also play a role in regulation of the cell cycle by virtue of its ability to increase the activity of the E2F transcription factor and to promote the transcription and translation of D-type cyclins (43, 47).

Our finding that BCR ligation induces phosphorylation of Akt on both threonine 308 and serine 473 suggests that the BCR activates both PDK1 and PDK2. Thus, if PDK1 and PDK2 have other substrates besides Akt, these proteins are also likely to be targets of BCR signaling. Recent work has shown that PDK1 phosphorylates and activates the p70^S6K kinase (86), the Sgk protein kinase (87), and the ζ isoform of protein kinase C (PKC) (88). Bras et al. (89) have shown that BCR ligation causes PKC ζ to translocate from the cytosol to the membrane fraction of A20 B lymphoma cells, suggesting that the BCR activates PKC ζ. However, the roles of PI 3-kinase and PDK1 in this process have not been examined. The identity of PDK2 has not been firmly established, although there is some evidence that the ILK can phosphorylate Akt (33). We are now investigating whether the BCR activates ILK.

In summary, we have provided evidence that the BCR activates a PI 3-kinase/PDK1,2/Akt/GSK-3 signaling pathway, and we have directly shown that BCR engagement regulates the activity of the Akt and GSK-3 kinases. Since Akt and GSK-3, as well as PDK1 and PDK2, can regulate a number of cellular processes, such as cell survival, cell cycle progression, transcription, and translation, this signaling pathway is likely to have important roles in BCR signaling.

We thank Jason Dinglasan for technical assistance and Tomohiro Kurosaki for the DT40 cell lines.