The B Cell Antigen Receptor Regulates the Transcriptional Activator β-Catenin Via Protein Kinase C-Mediated Inhibition of Glycogen Synthase Kinase-3¹

Signaling by the B cell Ag receptor (BCR)³ can promote B cell survival, proliferation, differentiation, apoptosis, or anergy depending on the maturation state of the B cell and the context provided by signals from other receptors. Although the BCR activates multiple signaling pathways, the role of individual signaling pathways in mediating responses to BCR engagement is not completely understood.

Activation of the phosphatidylinositol 3-kinase (PI3K) pathway is a key element in BCR signaling (1, 2). Upon BCR engagement, PI3K is recruited to the plasma membrane via the binding of its Src homology 2 domains to phosphotyrosine-containing sequences on membrane-associated scaffolding proteins such as CD19, Gab1, BCAP, and Cbl (3, 4, 5, 6). Once at the plasma membrane, PI3K generates phosphatidylinositol 3,4,5-trisphosphate (PIP₃) by phosphorylating the membrane phospholipid phosphatidylinositol 4,5-bisphosphate (1, 2). Subsequent dephosphorylation of PIP₃ yields phosphatidylinositol 3,4-bisphosphate (PI(3, 4)P₂). Both PIP₃ and PI(3, 4)P₂ act as anchors that recruit pleckstrin homology (PH) domain-containing proteins to the plasma membrane (7). BCR engagement has been shown to increase the levels of both PIP₃ and PI(3, 4)P₂ (8). This allows PH domain-containing signaling enzymes such as Btk, phospholipase C (PLC)-γ2, and Akt/protein kinase B to be recruited to the plasma membrane where they are activated (1, 2).

We and others have shown that BCR engagement activates Akt (9, 10, 11, 12, 13, 14). Akt is the primary mediator of the anti-apoptotic signals generated by PI3K (15), and recent work has shown that Akt kinase activity is essential for the survival of the DT40 chicken B cell line (16). Akt phosphorylates a number of proteins that regulate cell survival (17, 18, 19, 20, 21). In addition, Akt can also phosphorylate the serine/threonine kinases glycogen synthase kinase-3 (GSK-3)α and GSK-3β (22).

GSK-3α and GSK-3β are constitutively active in resting cells, but receptor-stimulated phosphorylation of GSK-3α at Ser²¹ or GSK-3β at Ser⁹ inhibits their enzymatic activity (23). These negative regulatory sites on GSK-3α and GSK-3β can be phosphorylated not only by Akt, but also by several other kinases including the p90^Rsk kinase, integrin-linked kinase, and several protein kinase C (PKC) isoforms (22, 24, 25, 26). In B cells, we have shown that BCR engagement induces the phosphorylation of GSK-3α and GSK-3β on these negative regulatory sites and inhibits the activity of GSK-3α (9).

An important target of GSK-3 is β-catenin (27), a transcriptional coactivator that has important roles in early development (28, 29). In unstimulated cells, GSK-3 constitutively phosphorylates β-catenin on N-terminal serine residues, targeting β-catenin for rapid ubiquitination and proteasome-mediated degradation (30, 31). In immature progenitor cells of various lineages, including pro-B cells (32), Wnt hormones regulate developmental processes by inhibiting GSK-3 (33). This Wnt-induced inhibition of GSK-3 is mediated by the Disheveled protein (29). Inhibition of GSK-3-dependent phosphorylation of β-catenin allows β-catenin to accumulate in the cytoplasm and then translocate into the nucleus (34). Once in the nucleus, β-catenin promotes transcription by binding to members of the lymphoid enhancer factor-1 (LEF-1)/T factor (TCF) family of DNA-binding proteins (35, 36). β-Catenin displaces Groucho/transducin-like enhancer of split transcriptional repressors from LEF-1/TCF and provides a transactivation domain that can recruit CBP/p300 and promote transcription (37, 38, 39). In mammalian cells, β-catenin up-regulates the transcription of both cyclin D1 and c-myc, genes whose products promote cell growth and proliferation (40, 41, 42).

Because the inhibition of GSK-3 kinase activity by the Wnt signaling pathway results in an increase in β-catenin levels, we hypothesized that the inhibition of GSK-3 that occurs after BCR engagement would also result in increased β-catenin levels. In this report, we show that BCR signaling causes an increase in β-catenin levels in the nuclear fraction of B cells as well as an increase in β-catenin-dependent transcription. Although Akt phosphorylates GSK-3 in other cell types (43, 44), we found that both BCR-induced phosphorylation of GSK-3 and BCR-induced up-regulation of β-catenin were mediated primarily by PLC-dependent activation of PKC and not by Akt. Thus, the BCR regulates β-catenin via a PLC-γ2/PKC/GSK-3 signaling pathway.

Goat Abs against mouse IgM (μ-chain specific), mouse IgG (γ-chain specific), and Armenian hamster Ig were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). Goat anti-mouse-κ L chain Abs were purchased from Southern Biotechnology Associates (Birmingham, AL). The hybridoma producing the HM79-16 hamster anti-mouse Igβ mAb was a gift from Dr. T. Nakamura (University of Tokyo, Tokyo, Japan) (45). The HM79-16 mAb was purified from the hybridoma supernatant using a protein G-Sepharose column. Abs specific for Akt, Akt phosphorylated on Ser⁴⁷³ (anti-P-Ser⁴⁷³ Akt), Akt phosphorylated on Thr³⁰⁸ (anti-P-Thr³⁰⁸ Akt), GSK-3α/GSK-3β phosphorylated on Ser²¹ and Ser⁹, respectively (anti-P-GSK-3α/GSK-3β), the Forkhead-related transcription factor (FKHR), and FKHR phosphorylated on Ser²⁵⁶ (anti-P-FKHR) were all purchased from Cell Signaling Technologies (Beverly, MA). The anti-GSK-3 Ab was from Chemicon International (Temecula, CA). The mAb to β-catenin was obtained from BD Transduction Laboratories (Lexington, KY). The anti-Sam 68 Ab (sc-733) and the anti-protein kinase D (PKD) Ab (sc-639) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA).

The WEHI-231 and A20 murine B cell lines and the Ramos human B cell line were obtained from American Type Culture Collection (Manassas, VA). The K40-B1 pro-B cell line (46) was a gift from Dr. A. DeFranco (University of California, San Francisco, CA). All cell lines were maintained in RPMI 1640 supplemented with 10% heat-inactivated FCS, 50 μM 2-ME, 1 mM pyruvate, 2 mM glutamine, 15 U/ml penicillin, and 50 μg/ml streptomycin (complete medium). WEHI-231 cells expressing a myristoylated Akt-estrogen receptor chimeric protein (mER-Akt) were maintained in complete medium supplemented with 0.25 μg/ml puromycin (Calbiochem, La Jolla, CA). To generate this cell line, cDNA encoding the mER-Akt protein (a gift from Dr. R. Roth, Stanford University, Stanford, CA) (47) was subcloned into the pMX-puro-IRES-EGFP retroviral vector, a derivative of pMX-puro (DNAX, Palo Alto, CA) (48). The resulting plasmid was transfected into the BOSC 23 packaging cell line and the viral particles produced were used to infect WEHI-231 cells as described (49). Stable bulk populations of WEHI-231 cells expressing the mER-Akt protein were selected using puromycin. 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 (9).

To reduce basal signaling caused by serum components, WEHI-231 cells were grown in complete medium with the FCS reduced to 1% for 12–18 h before stimulation while A20 cells were grown in complete medium with 0.5 mg/ml BSA instead of FCS. The cells were washed once with modified HEPES-buffered saline (9), resuspended in this buffer at 1 × 10⁷ and warmed to 37°C for 10–30 min. Where indicated, the cells were incubated with wortmannin (Biomol, Plymouth Meeting, PA), Ly294002 (Biomol), acetyl-leucine-leucine-norleucinol (ALLN; Sigma-Aldrich, St. Louis, MO), safingol (Calbiochem), U73122 (Biomol), or U73343 (Biomol) for 20–30 min before stimulation. The cells were then stimulated with either anti-Ig Abs, phorbol dibutyrate (PdBu), 4-hydroxytamoxifen (4-HT) (Sigma-Aldrich), LiCl, or bisindolylmaleimide I (BIM I; Calbiochem) for the indicated times. The reactions were terminated by adding ice-cold PBS containing 1 mM Na₃VO₄ and then centrifuging the cells for 1 min at 1100 × g in a cold microfuge. For cell lines, the cell pellets were solubilized in Triton X-100 lysis buffer (20 mM Tris-HCl (pH 8), 1% Triton X-100, 137 mM NaCl, 2 mM EDTA, 10% glycerol, PMSF, 1 μg/ml aprotinin, 10 μg/ml leupeptin, 1 mM Na₃VO₄, 25 mM β-glycerophosphate). Splenic B cells were solubilized in buffer B (20 mM HEPES (pH 7.4), 0.42 M NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 25% glycerol, 0.5 mM DTT, 1 mM PMSF, 1 μg/ml aprotinin, 10 μg/ml leupeptin, 1 mM Na₃VO₄, 25 mM β-glycerophosphate) containing 0.5% Igepal (ICN Pharmaceuticals, Costa Mesa, CA). After 10 min on ice, insoluble material was removed by centrifugation and the protein concentrations determined using the bicinchoninic acid assay (Pierce, Rockford, IL).

Nuclear and cytoplasmic fractions were prepared as described by Dignam et al. (50). After stimulation, 5 × 10⁶ cells were resuspended in 200 μl buffer A (10 mM HEPES (pH 7.4), 10 mM KCl, 1.5 mM MgCl₂, 20% glycerol, 0.5 mM DTT, 1 mM PMSF, 1 μg/ml aprotinin, 10 μg/ml leupeptin, 1 mM Na₃VO₄, 25 mM β-glycerophosphate). After 15 min on ice, the nonionic detergent Igepal was added to a final concentration of 0.5%. The samples were then centrifuged for 1 min at 1100 × g in a cold microfuge. The supernatant was removed and used as the cytosolic fraction. The pellets were rinsed once with buffer A and then extracted with 100 μl buffer B for 20 min on ice. The insoluble material was removed by centrifuging for 3 min at full speed in a cold microfuge. The supernatant was collected and used as the nuclear fraction. Protein concentrations for the cytosolic and nuclear fractions were determined using the bicinchoninic acid assay.

Total cell extracts or cytoplasmic and nuclear fractions (20 μg protein unless otherwise indicated) were separated on SDS-PAGE gels and then transferred to nitrocellulose membranes. The membranes were blocked for 1–2 h with 5% (w/v) skim milk powder in TBST and then incubated overnight at 4°C with the primary Ab. All Abs were diluted in TBST/1% BSA, with the exception of the Abs to PKD and Sam 68 which were diluted in TBST/1% skim milk powder. The membranes were then washed with TBST and incubated with the appropriate HRP-conjugated secondary Ab (Bio-Rad, Hercules, CA) for 1 h at room temperature. Immunoreactive bands were visualized using ECL (Amersham Pharmacia Biotech, Baie d’Urfe, Quebec, Canada). To reprobe the membranes, bound Abs were eluted by incubating the membrane in 10 mM Tris-HCl (pH 2), 150 mM NaCl for 30 min. The membranes were then reblocked and probed as described above. To quantitate results, scans of ECL exposures were saved as TIFF files and analyzed using ImageQuant 1.2 software (Molecular Dynamics, Sunnyvale, CA).

A GST fusion protein containing the C-terminal portion of the cytoplasmic domain of E-cadherin (GST-ECT) was used to precipitate unbound β-catenin. The pGEX-4t1-ECT plasmid encoding this fusion protein (a gift from Drs. H. Aberle and R. Kemler, Max Planck Institute for Immunobiology, Freiburg, Germany) (51) was transformed into the Escherichia coli strain DH5α. Fusion protein production was induced by growing the bacteria in the presence of 100 μM isopropyl-β-d-thiogalactopyranoside for 12 h at 37°C. The bacteria were then resuspended in sonication buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% Triton X-100, 1 mg/ml lysozyme, 0.1 mg/ml DNase I, 10 μg/ml leupeptin, 10 μg/ml soybean trypsin inhibitor, 1 μg/ml aprotinin, 1 mM PMSF) and lysed by sonication. The lysate was centrifuged at 30,000 rpm for 45 min in the cold. The supernatant containing the fusion protein was collected and stored as aliquots at −70°C. To precipitate β-catenin, 30 μl of bacterial lysate containing the GST-ECT fusion protein was incubated with 15 μl packed glutathione-Sepharose beads (Amersham Pharmacia Biotech) for 1 h at 4°C. After washing, the beads were mixed with Triton X-100 cell lysates (0.5 mg protein) for 1 h at 4°C. The beads were then pelleted and washed three times with Triton X-100 lysis buffer. Bound proteins were eluted using SDS-PAGE sample buffer.

The TOPtk and FOPtk plasmids were obtained from Dr. M. Waterman (University of California, Irvine, CA) (52). Transient transfection of WEHI-231 cells was performed using the DMRIE-C lipid reagent (Invitrogen, Burlington, Ontario, Canada). Briefly, lipid:DNA complexes were formed by adding the DMRIE-C lipid reagent (12 μl) and the DNA (4 μg) to 1 ml of OPTI-MEM medium and incubating for 45 min at 21°C. WEHI-231 cells (4 × 10⁶) that had been resuspended in 0.2 ml complete medium lacking antibiotics were added to this mixture. After 4 h at 37°C, 2 ml of complete medium lacking antibiotics was added and the cells were grown for an additional 20 h at 37°C. The transfected cells were then resuspended in fresh complete medium and divided into multiple wells of a 24-well dish. The cells were cultured with medium only, 10 μg/ml goat anti-mouse IgM Ab or 20 mM LiCl for 4 h at 37°C. After washing with PBS, the cells were lysed in Reporter Lysis Buffer (Promega, Madison, WI). Luciferase activity was determined using the Promega luciferase assay system (Promega). Readings were made using a MicroLumat Plus luminometer (EG&G Berthold, Bad Wildbad, Germany) set for a 10 s acquisition window. Relative luciferase units were derived by normalizing the luciferase activity to the amount of protein in each sample.

GSK-3 is a constitutively active kinase that normally keeps β-catenin levels low by phosphorylating β-catenin such that it becomes a target for ubiquitin-mediated degradation. Because BCR signaling inhibits GSK-3 activity (9), we hypothesized that BCR engagement would lead to an increase in β-catenin levels. To test this, we prepared cytosolic and nuclear fractions from WEHI-231 B lymphoma cells that had been stimulated with anti-IgM Abs for various times (Fig. 1,A). We found that BCR engagement caused an increase in β-catenin levels in both cellular compartments within 5–15 min. Because β-catenin is a transcriptional activator, in all subsequent experiments we focused on the levels of β-catenin in the nuclear fraction of the cells. We found that BCR engagement consistently caused an increase in the levels of β-catenin in the nuclear fraction of WEHI-231 cells, and that this increase persisted for at least 1 h (Fig. 1 B).

To show that the BCR-induced increase in the amount of β-catenin in the nuclear fractions of the cells was not due to contaminating cytosolic β-catenin, we validated our cell fractionation technique by reprobing the blots with Abs to the cytosolic protein PKD (53) and the nuclear protein Sam 68 (Fig. 1,A, middle and lower panels). PKD is a protein kinase while Sam 68 is a 68-kDa nuclear RNA-binding protein that is phosphorylated by Src during mitosis (54, 55). We found that the Sam 68 was entirely in the nuclear fractions, while only a small amount of the total PKD was in the nuclear fractions. Thus, the nuclear fractions were only minimally contaminated with cytosolic proteins. Because the amount of β-catenin in the nuclear fractions was equal to or greater than that in the cytosol fractions, neither the basal nor the BCR-stimulated levels of β-catenin in the nuclear fractions can be accounted for by cytosolic contamination. Further analysis showed that the levels of Sam 68 in the nuclear fractions increased slightly after BCR engagement, but that this was not statistically significant as assessed using Student’s t test (p > 0.09) (Fig. 1,C). Therefore, we used the levels of Sam 68 as a loading control and calculated the relative levels of β-catenin in the nuclear fraction of each sample by dividing the intensity of the β-catenin band by the intensity of the corresponding Sam 68 band for that sample (Fig. 1, B and D). Even if Sam 68 levels do increase slightly after BCR engagement, then we are underestimating the BCR-induced increases in the levels of β-catenin in the nucleus. In any case, Fig. 1 D shows that BCR signaling caused a statistically significant 2- to 2.5-fold increase in the levels of β-catenin in the nuclear fractions of WEHI-231 B lymphoma cells.

To determine whether up-regulation of β-catenin is a consistent characteristic of BCR signaling, we extended our analysis to include two other murine B cell lines as well as a human B cell line and murine splenic B cells (Fig. 2). The A20 and K40-B1 murine B cell lines represent different stages of B cell development than WEHI-231 cells. Whereas WEHI-231 cells express membrane IgM and represent immature/transitional B cells that are susceptible to Ag-induced clonal deletion, A20 cells express surface IgG and represent mature B cells that have undergone Ig class switching. K40-B1 cells represent pro-B cells that express the Igα/Igβ heterodimer on their surface in the absence of Ig chains (46). The Igα/Igβ heterodimer on K40-B1 cells is associated with several proteins including calnexin and Src family kinases (56); and therefore, may be part of a pro-BCR. Moreover, cross-linking this putative pro-BCR with anti-Igβ Abs induces a subset of the signaling reactions characteristic of the mature BCR (Ref. 56 ; S. L. Christian and M. R. Gold, unpublished observations). We found that engaging the BCR on A20 cells resulted in an increase in the levels of β-catenin in the nuclear fractions of these cells (Fig. 2,A). Similarly, clustering the putative pro-BCR on K40-B1 cells with anti-Igβ Abs also caused an increase in nuclear β-catenin levels (Fig. 2,B). To extend this analysis beyond murine B cell lines, we showed that BCR engagement caused an increase in nuclear β-catenin levels in the Ramos human B cell line (Fig. 2,C). Finally, to confirm that this response also occurs in normal B cells, we showed that engaging the BCR on murine splenic B cells caused an increase in β-catenin levels (Fig. 2,D). Both the time course and the magnitude of the BCR-induced increases in β-catenin levels were similar in all of the cells we examined (Figs. 1 and 2). The up-regulation of β-catenin occurred within 2–5 min and was sustained for at least 30 min, with the maximal increases in the range of 2- to 3-fold. This response is similar in magnitude to the increase in nuclear β-catenin levels caused by LPS in macrophages (57). Thus, up-regulation of nuclear β-catenin levels is a consistent characteristic of BCR signaling that occurs in normal B cells and in B cell lines representing multiple stages of B cell development.

In other cell types, receptor-induced increases in nuclear β-catenin levels are a consequence of decreased degradation of β-catenin in the cytosol. In unstimulated cells, the constitutive phosphorylation of β-catenin by GSK-3 keeps β-catenin levels low by targeting it for ubiquitination and proteasome-mediated degradation. However, inhibition of GSK-3 by Wnt signaling, for example, allows β-catenin to accumulate in the cytosol and then rapidly translocate into the nucleus. Because BCR engagement caused a rapid increase in the levels of both cytosolic and nuclear β-catenin, we asked whether this was also due to the inhibition of proteasome-mediated degradation of β-catenin. To test this, we treated WEHI-231 cells with the proteasome inhibitor ALLN to allow ubiquitinated proteins to accumulate. We then used a GST fusion protein containing the cytoplasmic domain of E-cadherin (GST-ECT) to selectively pull down cytoplasmic β-catenin present in Triton X-100 cell extracts. This GST-ECT fusion protein has been widely used to selectively isolate free cytoplasmic β-catenin that can potentially translocate into the nucleus, as opposed to β-catenin that is bound to cadherins, transmembrane proteins that function as adhesion molecules (51).

We found that treating WEHI-231 cells with the proteasome inhibitor ALLN for 20 min was sufficient to cause a 4-fold increase in the levels of free β-catenin (Fig. 3, compare 0 min ethanol lane to 0 min ALLN lane). This argues that proteasome-mediated degradation normally prevents the accumulation of β-catenin in B cells. The ALLN-induced increase in β-catenin levels was similar in magnitude to that caused by anti-IgM in the ethanol-treated control cells (Fig. 3). Moreover, in the presence of ALLN, anti-IgM treatment did not cause a further increase in β-catenin levels (Fig. 3). These results are consistent with the idea that the BCR-induced increase in β-catenin levels is due to inhibition of proteasome-mediated degradation of β-catenin. When the cells were treated with ALLN, we also observed slower migrating forms of β-catenin. These may be phosphorylated or ubiquitinated forms of β-catenin which are normally degraded very rapidly by proteasomes.

Because β-catenin is a transcriptional coactivator, we examined whether the BCR-induced increase in β-catenin protein levels in the nuclear fraction of B cells correlated with an increase in β-catenin-mediated transcription. In other cell types, β-catenin promotes transcription by binding to LEF-1/TCF DNA-binding proteins and the p300/CBP coactivator. To assess β-catenin-dependent transcription, we performed reporter gene assays using the TOPtk and FOPtk plasmids. The TOPtk plasmid contains multiple LEF-1/TCF binding sites, as well as the minimal thymidine kinase promoter, upstream of the luciferase gene (52). The binding of β-catenin-LEF-1/TCF complexes to the TOPtk promoter has been shown to stimulate transcription of the luciferase reporter gene. The FOPtk plasmid is identical except that it contains mutated LEF-1/TCF binding sites; and therefore, does not bind LEF-1/TCF proteins that can recruit β-catenin. Fig. 4 shows that cross-linking the BCR on WEHI-231 cells caused a 2.2-fold increase in luciferase activity in cells transfected with the TOPtk plasmid, but did not cause a significant increase in luciferase activity in cells transfected with the FOPtk plasmid. Thus, BCR signaling can stimulate β-catenin-dependent transcription driven by the LEF-1/TCF-binding sites in the TOPtk promoter. This suggests that the BCR-induced increase in β-catenin protein levels in B cells could result in increased transcription of genes that contain LEF-1/TCF binding sites. We also found that treating WEHI-231 cells with a GSK-3 inhibitor, LiCl, resulted in a 2-fold increase in luciferase activity (Fig. 4). As discussed in the following section, this indicates that inhibition of GSK-3 is sufficient to increase β-catenin-dependent transcription in WEHI-231 cells.

In Wnt-responsive cells, GSK-3 normally phosphorylates β-catenin and targets it for degradation while Wnt signaling increases β-catenin levels by inhibiting GSK-3. Because BCR signaling inhibits GSK-3 (9), the BCR may also regulate β-catenin via GSK-3. To determine whether GSK-3 normally targets β-catenin for degradation in B cells, we asked whether inhibiting GSK-3 activity with LiCl would be sufficient to cause an increase in β-catenin levels. Lithium has been shown to specifically inhibit GSK-3 kinase activity by displacing the Mg²⁺ cofactor (58, 59). We found that treating WEHI-231 cells with 20 mM LiCl for 15–30 min resulted in an increase in nuclear β-catenin levels (Fig. 5,A). As a specificity control, we showed that treating the cells with 20 mM KCl did not increase β-catenin levels. Another inhibitor of GSK-3, BIM I (60), also increased the levels of β-catenin in the nuclear fraction of WEHI-231 cells (Fig. 5 B). Thus, inhibition of GSK-3 is sufficient to allow the accumulation of β-catenin in the nuclear fraction of B cells. This indicates that GSK-3 normally prevents the accumulation of β-catenin in B cells and is consistent with the idea that BCR up-regulates β-catenin by inhibiting GSK-3.

We also examined whether inhibition of GSK-3 activity is sufficient to increase β-catenin-mediated transcription in WEHI-231 cells. We found that LiCl treatment caused a 2-fold increase in β-catenin-dependent transcription from the TOPtk promoter while having little or no effect on transcription driven by the control FOPtk promoter (Fig. 4). Thus, inhibition of GSK-3 is sufficient to increase both β-catenin protein levels and β-catenin-dependent transcription in B cells.

Our next goal was to determine the mechanism by which the BCR regulates GSK-3; and therefore, regulates β-catenin. Receptor-induced inhibition of GSK-3 kinase activity is due to phosphorylation of Ser²¹ of GSK-3α or Ser⁹ of GSK-3β (23). The serine/threonine kinase Akt was a good candidate for BCR-induced GSK-3 phosphorylation since the BCR activates Akt (9) and Akt has been shown to phosphorylate these negative regulatory sites on GSK-3α/GSK-3β in insulin-stimulated cells (44).

To assess the contribution of Akt to the regulation of GSK-3 and β-catenin in B cells, we expressed a conditionally active form of Akt (mER-Akt) in the WEHI-231 cell line. The mER-Akt protein lacks the PH domain of Akt but contains a myristoylation sequence at the N terminus to localize it to the inner leaflet of the plasma membrane. This altered form of Akt is fused to a mutant form of the estrogen receptor that is responsive to the estrogen analog 4-HT. In response to 4-HT, the estrogen receptor portion of the mER-Akt protein undergoes a conformational change that exposes the Akt activation sites, allowing the mER-Akt protein to be phosphorylated and activated by PDK1 and PDK2 (47). We found that 4-HT treatment of WEHI-231 cells expressing the mER-Akt protein resulted in activation of this 90-kDa Akt fusion protein as indicated by its phosphorylation on the key Akt regulatory sites that correspond to Thr³⁰⁸ and Ser⁴⁷³ of wild-type Akt (Fig. 6,A). In addition, the 60-kDa endogenous Akt was also activated when the mER-Akt-expressing cells were treated with 4-HT (Fig. 6,A), possibly due to an interaction between the mER-Akt protein and the endogenous Akt. The 4-HT-induced activation of the endogenous Akt was similar in magnitude to that caused by BCR engagement (Fig. 6 A). If the mER-Akt can also phosphorylate cytoplasmic substrates such as GSK-3, then the total 4-HT-induced Akt activation could be greater than the BCR-induced activation of endogenous Akt.

To test whether 4-HT treatment of mER-Akt-expressing cells could stimulate the phosphorylation of known Akt substrates, we examined the phosphorylation of the FKHR, a protein that is found in both the cytoplasm and nucleus. Akt has been shown to phosphorylate FKHR on Ser²⁵³ and Thr³² (18). To assess FKHR phosphorylation, we used an Ab that recognizes FKHR that is phosphorylated on Ser²⁵³. We found that FKHR appeared to be phosphorylated on this site even in unstimulated WEHI-231 cells (Fig. 6,B). However, 4-HT treatment caused the appearance of a slower migrating form of FKHR (Fig. 6 B). Because such changes in electrophoretic mobility frequently correlate with increased phosphorylation, this bandshift could reflect Akt-dependent phosphorylation of FKHR on Thr³². Note that BCR engagement caused the appearance of additional forms of FKHR that migrated even more slowly. This may reflect the phosphorylation of FKHR on additional sites that are targeted by other BCR signaling pathways. Indeed, Ras-dependent phosphorylation of FKHR has been reported (61). In any case, these data indicate that 4-HT treatment of mER-Akt-expressing cells can stimulate Akt-dependent phosphorylation events.

We then asked whether 4-HT-induced Akt activation could lead to phosphorylation of GSK-3α/GSK-3β on their negative regulatory sites. Phosphorylation of GSK-3α and GSK-3β at Ser²¹ and Ser⁹, respectively, was assessed using phosphorylation state-specific Abs. Fig. 6,C shows that while BCR engagement caused a significant increase in the phosphorylation of both GSK-3α and GSK-3β, 4-HT treatment caused only a very small increase in GSK-3 phosphorylation, even though it activated Akt to the same extent as BCR engagement (Fig. 6 A). Thus, while Akt activation is sufficient to inhibit GSK-3 activity in muscle cells (43), in B cells the amount of Akt activation stimulated by the BCR does not cause significant phosphorylation of GSK-3 on the negative regulatory sites. This indicates that BCR-induced phosphorylation and inhibition of GSK-3 is mediated by a kinase other than Akt.

Because the inhibition of GSK-3 is sufficient to increase β-catenin levels in WEHI-231 cells (Fig. 5), the inability of Akt activation to cause significant phosphorylation of GSK-3 in these cells suggested that Akt activation would be unable to stimulate the up-regulation of β-catenin. Indeed, we found that 4-HT treatment of mER-Akt-expressing WEHI-231 cells did not cause an increase in β-catenin levels (Fig. 6 D). Thus, Akt does not regulate the GSK-3/β-catenin pathway in B cells.

In addition to Akt, several PKC isoforms can phosphorylate GSK-3β in vitro (26). Moreover, treating cells with phorbol esters, compounds that activate both classical and novel PKC isoforms, can inhibit GSK-3 activity (33). Therefore, we asked whether PKC enzymes might link the BCR to the GSK-3/β-catenin pathway.

First, we asked whether treating B cells with phorbol esters could stimulate GSK-3 phosphorylation. Fig. 7,A shows that treating WEHI-231 cells with a low concentration (20 nM) of PdBu caused significant phosphorylation of both GSK-3α and GSK-3β. This PdBu-induced GSK-3 phosphorylation was equal to or greater than that caused by anti-IgM. Moreover, the PdBu-induced GSK-3 phosphorylation was not dependent on Akt, because this concentration of PdBu did not activate Akt, as judged by phosphorylation of Akt on Ser⁴⁷³ (Fig. 7 B), a very sensitive measurement of Akt activation.

Because PdBu treatment of WEHI-231 cells could stimulate the phosphorylation of GSK-3α/GSK-3β on their negative regulatory sites, we asked whether it could also cause an increase in β-catenin levels. Fig. 7 C shows that low concentrations of PdBu (10–30 nM) could increase the level of β-catenin in the nuclear fraction of WEHI-231 cells to the same extent as anti-IgM treatment. Thus, PKC activation is sufficient to induce GSK-3 phosphorylation and to increase nuclear β-catenin levels. Moreover, since the amount of PKC activation caused by anti-IgM treatment of WEHI-231 cells is similar to that caused by 10 nM PdBu (62), this suggests that PKC activation is sufficient to mediate the effects of BCR engagement on GSK-3 and β-catenin.

To test whether PKC activation is necessary for BCR-mediated regulation of GSK-3 and β-catenin, we used safingol to inhibit PKC activity. Safingol functions by competitively inhibiting the conserved diacylglycerol (DAG)-binding C1 domain of PKC enzymes. Treating WEHI-231 cells with safingol inhibited BCR-induced phosphorylation of GSK-3α/GSK-3β (Fig. 8,A) and the BCR-induced increase in nuclear β-catenin levels (Fig. 8,B) to a similar extent (Fig. 8,C). This is consistent with a model in which GSK-3 phosphorylation and inactivation is responsible for the increase in β-catenin levels. Moreover, these data argue that the BCR regulates GSK-3 and β-catenin via the PKC-mediated inhibition of GSK-3. In support of this idea, we found that inhibiting PKC activity with safingol completely blocked BCR-induced up-regulation of β-catenin protein levels in murine splenic B cells (Fig. 8 D).

Our data suggest that the BCR regulates GSK-3 and β-catenin via conventional (PKC-α, -β, -γ) or novel (PKC-δ, -ε, -η, -θ) PKC isoforms. These PKC isoforms are inhibited by safingol and are activated by phorbol esters such as PdBu which mimic the action of DAG. During BCR signaling, these PKC isoforms would be activated via the production of DAG by PLC-γ2 (63). Therefore, preventing the activation of PLC-γ2 should inhibit the ability of the BCR to activate this putative PKC/GSK-3/β-catenin pathway. Indeed, we found that treating WEHI-231 cells with U73122, an inhibitor of PLC activity, blocked both the BCR-induced increase in GSK-3 phosphorylation (Fig. 9,A) and the BCR-induced increase in β-catenin levels (Fig. 9 B). In contrast, an inactive structural analog of U73122, U73343, had no effect on BCR-induced GSK-3 phosphorylation or β-catenin up-regulation (data not shown). The finding that both PLC and PKC activities are required for the BCR to regulate GSK-3 and β-catenin suggests that the BCR regulates β-catenin via a PLC-γ2/PKC/GSK-3 pathway.

We have previously shown that BCR-induced GSK-3 phosphorylation, as well as the subsequent inhibition of GSK-3 activity, is dependent on PI3K (9). Because Akt is an important downstream target of PI3K, we had initially assumed that the BCR would regulate GSK-3 via Akt. However, we have now shown that the BCR regulates GSK-3 via a PLC-γ2/PKC pathway. This does not preclude a role for PI3K because PI3K regulates the membrane recruitment and activation of Btk. Phosphorylation of PLC-γ2 by Btk is required for maximal activation of PLC-γ2 (63) and previous work has shown that inhibition of PI3K reduces anti-IgM-induced production of PLC-γ2-derived second messengers by ∼60% (64). Therefore, we would predict that inhibition of PI3K would partially block the ability of the BCR to regulate GSK-3 and β-catenin. Indeed, we found that two structurally distinct PI3K inhibitors, wortmannin and Ly294002, inhibited both the BCR-induced increase in GSK-3 phosphorylation and the BCR-induced increase in nuclear β-catenin levels by 60–75% (Fig. 10). Thus, PI3K does contribute to the ability of the BCR to regulate β-catenin, but this may reflect the role of PI3K in the activation of PLC-γ2 as opposed to Akt.

In this report we have shown for the first time that the transcriptional activator β-catenin is a target of BCR signaling. We show that BCR engagement increases β-catenin protein levels as well as β-catenin-dependent transcription. We also provide evidence that these responses are mediated by the BCR-induced inhibition of GSK-3 activity that we have described previously (9). Finally, we investigated the mechanism by which the BCR regulates GSK-3 activity. GSK-3 activity is inhibited when it is phosphorylated on negative regulatory sites, and we show in this study that the BCR-induced phosphorylation of GSK-3 on these sites is mediated by a PI3K/PLC-γ2/PKC signaling pathway (Fig. 11).

BCR signaling appears to regulate β-catenin levels by controlling the rate at which β-catenin is degraded. Studies in several cell types have shown that β-catenin is rapidly degraded in unstimulated cells via a GSK-3-dependent mechanism (31). Free β-catenin in the cytoplasm binds to a protein complex that contains GSK-3 as well as a scaffolding protein called axin and the adenamatous polyposis coli protein. Axin and adenamatous polyposis coli bind both GSK-3 and β-catenin, facilitating the phosphorylation of β-catenin by GSK-3. Phosphorylation of β-catenin by GSK-3 targets it for ubiquitination and proteasome-mediated degradation. Our finding that the proteasome inhibitor ALLN increases β-catenin levels in B cells argues that proteasome-mediated degradation keeps β-catenin levels low in unstimulated B cells. Wnt hormones, the best-studied regulators of β-catenin, cause increases in β-catenin levels by inhibiting GSK-3. This allows β-catenin to accumulate because it is no longer efficiently targeted for degradation. Our data suggest that the BCR also increases β-catenin levels by preventing the degradation of β-catenin. We found that BCR engagement did not cause a further increase in β-catenin levels in ALLN-treated cells in which β-catenin degradation was already inhibited. In addition, BCR signaling caused β-catenin levels to increase within 5–15 min, kinetics that are more consistent with inhibition of β-catenin degradation as opposed to an increase in transcription or translation.

In epithelial cells and other cell types in which Wnt signaling has been studied, the constitutive phosphorylation of β-catenin by GSK-3 keeps β-catenin levels low by targeting it for ubiquitination and proteasome-mediated degradation. Our data indicate that GSK-3 kinase activity is also responsible for keeping β-catenin levels low in unstimulated B cells. GSK-3 is constitutively active in B cells (9), and we found that two different inhibitors of GSK-3, LiCl and BIM I, both caused β-catenin levels to increase in B cells. LiCl inhibits GSK-3 kinase activity by displacing the Mg²⁺ cofactor (59), while BIM I competitively inhibits the ATP-binding site of GSK-3 (60). Although BIM I also inhibits the activity of several PKC isoforms (65), its ability to increase β-catenin levels in B cells most likely reflects its inhibitory effect on GSK-3 since inhibiting PKC activity blocks increases in β-catenin levels (Fig. 8). In any case, the use of these two different GSK-3 inhibitors indicates that GSK-3 is a central regulator of β-catenin levels in B cells.

Inhibition of GSK-3 is thought to be the major mechanism by which receptor signaling increases β-catenin levels. Our data suggest that the BCR also regulates β-catenin via the inhibition of GSK-3. We have previously shown that BCR signaling leads to the phosphorylation of GSK-3α/GSK-3β on their negative regulatory sites and the concomitant inhibition of GSK-3 activity (9). As described above, we found that inhibition of GSK-3 is sufficient to cause an increase in β-catenin levels in B cells. Moreover, we found that agents that blocked the BCR-induced phosphorylation of GSK-3 on its negative regulatory sites also blocked the ability of the BCR to increase β-catenin levels. Thus, it appears that the inhibition of GSK-3 is both necessary and sufficient for the BCR to up-regulate β-catenin.

A number of different kinases can phosphorylate the negative regulatory sites on GSK-3 and inhibit GSK-3 activity. These include Akt, the p90^Rsk kinase, the p70 S6 kinase, protein kinase A, integrin-linked kinase, and the α, βI, βII, δ, and ζ isoforms of PKC (22, 24, 25, 26, 66, 67, 68, 69). Thus, different receptors may regulate GSK-3 via different signaling pathways (23). For example, insulin receptor-induced inhibition of GSK-3 is mediated by Akt (44), while the α1-adrenergic receptor inhibits GSK-3 via PKC-ζ (69). PKC activity is also required for Wnt-induced inhibition of GSK-3 (33). Although the BCR activates Akt (9), we found that selectively activating the mER-Akt chimeric protein along with endogenous Akt was not sufficient to induce significant GSK-3 phosphorylation or up-regulation of β-catenin in WEHI-231 cells. In contrast, overexpressing a membrane-bound form of constitutively active Akt does lead to the inhibition of GSK-3 in L6 skeletal muscle cells (43). This suggests that the degree of coupling between Akt and GSK-3 could be cell type specific.

Whereas Akt does not play a major role in linking the BCR to GSK-3 and β-catenin, our results indicate that the BCR regulates GSK-3 and β-catenin primarily via PKC. Treating B cells with the PKC activator PdBu was sufficient to induce GSK-3 phosphorylation and cause an increase in nuclear β-catenin levels. This was not dependent on Akt because PdBu did not stimulate the phosphorylation of Akt on sites that are required for its activation. In support of the idea that the BCR regulates GSK-3 and β-catenin via PKC, we showed that the PKC inhibitor safingol blocked BCR-induced GSK-3 phosphorylation and BCR-induced up-regulation of β-catenin to the same extent. Safingol prevents the binding of DAG to the C1 domains of conventional and novel PKC isoforms. Thus, activation of a DAG-dependent PKC isoform is both necessary and sufficient for the BCR to stimulate the phosphorylation of GSK-3α/GSK-3β and the accumulation of β-catenin.

Although other groups have shown that PKC can phosphorylate and inhibit GSK-3 (33, 70, 71), this is the first report showing that PKC-induced inhibition of GSK-3 is sufficient to cause the accumulation of β-catenin. In epithelial cells, PKC activity is necessary for Wnt-induced accumulation of β-catenin but phorbol ester-induced PKC activation is not sufficient to increase β-catenin levels (70). This suggests that other signals are required for the up-regulation of β-catenin, but that these signals are already present in WEHI-231 cells.

BCR engagement leads to the activation of PKC-α, -β, -δ, -ε, and -ζ (72, 73, 74, 75, 76). Because each PKC isoform is likely to have a unique set of substrates, we are currently attempting to identify which PKC isoform is responsible for the BCR-induced phosphorylation and inhibition of GSK-3. Our data suggest that either a conventional PKC isoform (PKC-α, -βI, -βII, -γ) or a novel PKC isoform (PKC-δ, -ε, -η, -θ) mediates GSK-3 phosphorylation because these PKC isoforms are responsive to increases in DAG. In addition, our finding that U73122 blocked BCR-induced phosphorylation of GSK-3 and up-regulation of β-catenin supports the idea that these responses are mediated via a PLC-γ2/PKC pathway. We also found that PI3K contributes to the ability of the BCR to regulate GSK-3 and β-catenin. This presumably reflects the role of PI3K in the BCR-induced activation of PLC-γ2 (63).

β-Catenin is a transcriptional activator. During early development it regulates the expression of genes that determine cell fate while in differentiated cells it regulates the expression of genes that promote proliferation. Our data suggest that β-catenin can also function as a transcriptional activator in B cells. We found that BCR engagement could stimulate β-catenin-dependent transcription, as judged by a luciferase reporter gene assay. Moreover, treating WEHI-231 cells with the GSK-3 inhibitor LiCl was sufficient to stimulate β-catenin-dependent transcription to a similar extent as anti-IgM treatment. Thus, at least in WEHI-231 cells, other BCR signaling pathways are not required to induce the expression or activation of proteins that cooperate with β-catenin to drive transcription. This is in contrast to Jurkat T cells where LiCl inhibits GSK-3 activity but is not sufficient to increase β-catenin-dependent transcription of the same luciferase reporter gene construct (77).

In all of the systems studied thus far, β-catenin promotes transcription by cooperating with LEF-1/TCF family proteins. The LEF-1/TCF proteins bind specific DNA sequences while β-catenin provides a transactivation domain that can recruit CBP/p300. Although Wnt-responsive bone marrow pro-B cells express LEF-1/TCF family proteins, these proteins have not been detected in mature B cells (32). WEHI-231 cells may express a novel member of this family that can bind to the LEF-1/TCF sites in the TOPtk promoter and cooperate with β-catenin to promote transcription. We are currently trying to identify such proteins to understand how β-catenin regulates transcription in WEHI-231 cells.

We found that BCR signaling increased the levels of β-catenin in WEHI-231 cells, A20 cells, Ramos cells, and murine resting splenic B cells. WEHI-231 and Ramos cells resemble immature/transitional B cells while A20 cells are presumably derived from memory cells. The function of β-catenin in immature/transitional B cells or in mature B cells is not known. In differentiated cells such as epithelial cells, up-regulation of β-catenin promotes cell cycle entry and proliferation by driving the transcription of the genes encoding cyclin D1 and c-Myc genes (40, 41, 42). LiCl, which we showed up-regulates β-catenin in WEHI-231 cells, has been shown to promote the proliferation of fetal liver B cell progenitors (32). Thus, it is possible that β-catenin could be involved in BCR-induced proliferation.

Although the role of β-catenin in mature B cells is not clear, work by Grosschedl and colleagues indicates that β-catenin is important for the proliferation and survival of pro-B cells. In pro-B cells, Wnt signaling increases β-catenin levels and stimulates proliferation (32). Moreover, disrupting the genes encoding the β-catenin binding partner LEF-1 results in increased apoptosis of pro-B cells (32). We found that clustering the Igβ on the surface of the K40-B1 pro-B cell line resulted in the up-regulation of β-catenin. This suggests that the putative pro-B cell receptor, which consists of the Igα/Igβ subunit associated with calnexin and several unidentified proteins (56), could also deliver survival and proliferative signals via β-catenin. If this is the case, then the pro-B cell receptor may have a role similar to that of the pre-B cell receptor which delivers Syk- and Btk-dependent signals that promote the survival and further differentiation of pre-B cells.

In summary, we have shown that the BCR regulates the levels of β-catenin in the nuclei of B cells via a PI3K/PLC-γ2/PKC/GSK-3 pathway. Because β-catenin can promote survival, differentiation, or proliferation, depending on the cell type, it may play an important role in the regulation of B cell development and activation by the BCR.

We thank Sherwin Ting and May Dang-Lawson for technical assistance, Marian Waterman, Anthony DeFranco, Richard Roth, Tetsuya Nakamura, Hermann Aberle, and Rolf Kemler for providing key reagents, and Randall Moon for advice and encouragement. We thank Linda Matsuuchi for critical reading of the manuscript.