Requirement for a hsp90 Chaperone-dependent MEK1/2-ERK Pathway for B Cell Antigen Receptor-induced Cyclin D2 Expression in Mature B Lymphocytes*

A requirement for cyclin D2 in G1-to-S phase progression has been definitively established in mature B cells stimulated via the B cell antigen receptor (BCR). However, the identity of constituents of the BCR signaling cascade that leads to cyclin D2 accumulation remains incomplete. We report that inhibition of mitogen-activated protein kinase/extracellular signal-regulated kinase kinase (MEK)-1/2 blocked BCR-induced activation of extracellular signal-regulated kinase (ERK). Inhibition of the MEK1/2-ERK pathway was sufficient to abrogate BCR-induced cyclin D2 expression at the mRNA and protein levels. Disruption of endogenous heat shock protein 90 (hsp90) function with geldanamycin abrogated BCR-induced cyclin D2 expression and proliferation. Geldanamycin effects were attributed to a selective depletion of cellular Raf-1 that interrupted BCR-coupled activation of MEK1/2 and ERK. By contrast, signaling through the phosphatidylinositol 3-kinase and protein kinase C pathways was not affected, suggesting that disruption of hsp90 function did not cause a general impairment of BCR signaling. These results suggest that the MEK1/2-ERK pathway is essential for BCR signaling to cyclin D2 accumulation in ex vivo splenic B lymphocytes. Furthermore, these findings imply that hsp90 function is required for BCR signaling through the Raf-1-MEK1/2-ERK pathway but not through the phosphatidylinositol 3-kinase- or protein kinase C-dependent pathways.

Binding of antigen to the B cell antigen receptor (BCR) 1 can lead to proliferation, differentiation, or lymphocyte death by clonal deletion (1). These distinct cell fates are determined, at least in part, by the developmental stage of the lymphocyte and through the coordinate regulation of signaling pathways by the BCR in response to binding of monovalent versus polyvalent antigen. BCR signaling is triggered by the activation of the Src family protein-tyrosine kinases (PTKs), Bruton's tyrosine kinase (Btk), and Syk (1,2). These PTKs, in turn, coordinate the activation of multiple signal transduction cascades, including phospholipase C (PLC)-␥1 and -␥2, phosphatidylinositol 3-kinase, the Vav/Rho family pathway, and mitogen-activated protein kinase (MAPK) pathways (1,3). Recent efforts have focused on understanding how individual signaling pathways lead to distinct cell fates (1). In particular, the nature of signal transduction that controls BCR-induced proliferation in mature B cells remains incomplete. In mammalian cells, the D-type cyclins are considered end point targets of mitogenic signaling pathways and function as growth factor sensors (4,5). The requirement of D-type cyclins in G 1 phase progression has been definitively established in mammalian cells (5)(6)(7)(8). D-type cyclins function as positive regulatory subunits for a subset of cyclin-dependent kinases (CDKs) 4 and 6 (9,10). Several findings contribute to the emerging view that CDK4/6, together with CDK2, represent the primary protein kinases that drive cells through G 1 phase, presumably through the phosphorylation of the retinoblastoma gene product (pRb) (11,12). A current model holds that sequential phosphorylation of pRb by CDK4/6 and CDK2 disrupts its association with E2F family proteins, leading to the coordinated transcription of E2F-responsive genes directly involved in G 1 /S transition and DNA metabolism (11).
We and others have shown previously that cyclin D2, and to a lesser extent cyclin D3, are induced in response to BCR cross-linking (13,14). Definitive evidence in support of cyclin D2 in BCR-induced proliferation came from studies with cyclin D2-deficient mice, which exhibit normal levels of splenic B cells in comparison with wild-type littermates; however, ex vivo splenic B cells fail to proliferate in response to BCR crosslinking (15). These observations suggest that the principal initial determinant of cell cycle progression following BCR engagement in mature B cells lies in the sustained accumulation of cyclin D2. Despite these findings, little is known about the nature of signaling pathways that couple BCR to cyclin D2 accumulation in splenic B cells. In many nonlymphoid cell types, activation of the Ras-coupled mitogen-activated protein kinase/extracellular signal-regulated kinase kinase (MEK)-extracellular signal-regulated kinase (ERK) pathway promotes cell cycle entry and proliferation (reviewed in Refs. 4 and 5). This is achieved, in part, through Ras signaling de novo transcription of the cyclin D1 gene (16,17). Ras, together with the phosphatidylinositol 3-kinase/Akt pathway, can increase Dtype cyclin protein availability by enhancing mRNA translation; this pathway also negatively regulates ubiquitin-dependent proteasomal degradation of cyclin D1 (18,19).
Evidence has recently been obtained supporting a role for MEK1/2 and ERK in BCR-induced proliferation of quiescent splenic B cells (20). This signaling pathway is also activated during BCR-induced apoptosis in WEHI-231 B cells and autoreconstituted splenic B cells, which exhibit an immature B cell phenotype; however, specific inhibitors of MEK1/2 do not block BCR-induced G 1 phase arrest or apoptosis in these lymphocyte models (20). These findings suggest that activation of the MEK1/2-ERK pathway is a requisite for cyclin D2 expression and hence G 1 -to-S phase progression following BCR crosslinking on mature B cells. This view has been challenged by studies in mice with deletional mutations of the B cell signalosome. Splenic B cells from B cell linker (BLNK)-and Vavdeficient mice fail to proliferate in response to BCR crosslinking due to impaired cyclin D2 expression (21,22). BCR signaling in these mutant mice exhibits impaired PLC-␥2 activation and Ca 2ϩ mobilization, whereas activation of the MEK1/2-ERK pathway is normal. In the present study, we sought to determine whether activation of MEK1/2 and ERK is required for BCR-induced cyclin D2 expression in ex vivo cultures of splenic B cells. Results of this study indicate that Raf-1, MEK1/2, and ERK are necessary for BCR-induced expression of cyclin D2. Our findings also suggest that the Raf-1/MEK1/2-ERK pathway requires functional heat shock protein 90 (hsp90) in order to signal induction of cyclin D2 gene expression by BCR.
Preparation of Splenic B Lymphocytes-Balb/c mice at 8 -12 weeks were obtained from Taconic Laboratories (Germantown, NY) and housed at Boston College. Mice were cared for and handled at all times in accordance with National Institutes of Health and Boston College institutional guidelines. Mature B lymphocytes were isolated from spleens by depletion of T cells with anti-Thy-1.2 plus rabbit complement (Accurate Chemical and Scientific Corp., Westbury, NY) (13); macrophages (and other adherent cells) were removed by plastic adherence. Red blood cells and nonviable cells were removed by sedimentation on Lympholyte M (Accurate Chemical and Scientific Corp.). B cells were cultured in RPMI 1640 (BioWhittaker, Walkersville, MD) supplemented with 10 mM HEPES, pH 7.5, 2 mM L-glutamine, 5 ϫ 10 Ϫ5 M 2-ME, 100 units/ml penicillin, 100 g/ml streptomycin, and 10% heatinactivated fetal calf serum (BioWhittaker). B cells were maintained in a 37°C humidified incubator at 5% CO 2 .
Immunoblotting and Immunoprecipitation-B lymphocytes were solubilized at 2 ϫ 10 7 /100 l in 0.1% Triton X-100-containing buffer as described (13). Cellular proteins were separated through polyacrylamide-SDS gels and transferred to Immobilon-P membrane (Millipore Corp., Bedford, MA). Immunodetection of proteins was carried out as previously described (13). Immunoprecipitation was performed in 0.5% Nonidet P-40-solubilized cell lysates according to the method described by Zhang et al. (23). Lysates were incubated with 1.5 g of the indicated Abs (5 h), followed by the addition of 50 l of a 1:1 slurry of protein A/G-agarose (90 min). The protein A/G-agarose was collected by centrifugation and washed eight times with 1 ml of ice-cold Nonidet P-40 buffer. The immunoprecipitates were dissociated in SDS-PAGE sample buffer, separated through a 10% polyacrylamide-SDS gel, and immunoblotted. Where indicated, the resulting autoradiograms were ana-lyzed by densitometry using a Molecular Dynamics Personal Densitometer equipped with ImageQuant software (Molecular Dynamics, Inc., Sunnyvale, CA).
RNA Isolation and RT-PCR-Total cellular RNA was isolated from 2 ϫ 10 7 splenic B cells using Ultraspec Reagent (Biotecx Laboratories, Inc., Houston, TX) denaturing solution following the manufacturer's protocol. Purified RNA was treated with 1 unit of DNase I using DNAfree (Ambion, Inc., Austin, TX). First-strand cDNA synthesis was carried out with 2 g of total RNA using the RETROscript TM first strand synthesis kit (Ambion, Inc.) following the manufacturer's protocol. The concentration of cDNA in each sample was adjusted to produce similar amounts of product following quantitative PCR with primers specific for ␤ 2 -microglobulin gene (␤ 2 -microglobulin sense primer corresponds to 5Ј-cggtcgcttcagtcgtcagc-3Ј; antisense primer corresponds to 5Ј-cccagtagacggtcttggg-3Ј). The PCRs were performed in 25-l final reaction volumes, each consisting of 5 l of template cDNA using reagents from GeneAmp R PCR kit (Roche Molecular Systems, Inc., Branchburg, NJ) supplemented with 10 Ci of ␣-[ 32 P]dCTP (6000 Ci/mmol; PerkinElmer Life Sciences). Amplification was conducted with an Eppendorf Mastercycler gradient (Brinkmann Instrument, Inc. Westbury, NY). Temperature cycling parameters for the ␤ 2 -microglobulin gene were as follows: 95°C for 3 min, followed by 17 cycles of 95°C for 30 s, 55°C for 1 min for annealing, and 72°C for 1 min. A 72°C final elongation step for 10 min was included following the last cycle. PCRs were optimized with respect to Mg 2ϩ concentration, pH, annealing temperature, and the number of cycles as described (24). PCR products (20 l) were electrophoresed on a 8% polyacrylamide gel and then visualized by autoradiography. The identity of each PCR product was determined based on its expected size. Autoradiographic bands were quantitated with the aid of a Molecular Dynamics PhosphorImager. Under these conditions, the number of PCR cycles and the amount of cDNA used were such that increasing amounts of cDNA template yielded proportionately increasing amounts of ␤ 2 -microglobulin-specific PCR product (data not shown). The variation of ␤ 2 -microglobulin PCR product between adjusted samples was Ϯ5%, as judged by densitometric analysis of the resulting autoradiograms. The number of cycles and amount of cDNA used were then adjusted accordingly to yield the same signal output. Cyclin D2 cDNA was then amplified from the normalized ␤ 2 -microglobulin cDNA samples using the identical PCR and temperature cycling parameters. The sequence of the cyclin D2 primer set corresponds cyclin D2 sense primer 5Ј-agctgtccctgatccgcaag-3Ј and antisense primer 5Ј-gtcaacatcccgcacgtctg-3Ј.

Inhibition of MEK1/2 Blocks BCR-induced ERK Activity in
Splenic B Cells-PD98059 has been shown to act in vivo as a highly selective inhibitor of MEK1 activation and the ERK1/2 cascade, without affecting the related MEK family kinases, MKK3, MKK6, SEK/MKK4, or their immediate downstream effector kinases, JNK and p38 MAPK (25). Likewise, U0126 is a highly selective inhibitor of MEK1 and MEK2 and downstream ERK1/2, without affecting p38 MAPK, JNK, or numerous other protein kinases (e.g. protein kinase C, CDK4, or Raf-1) (26). The three major classes of mammalian MAPKs are rapidly and transiently activated following BCR cross-linking in B cells (1,20,27). To determine the specificity of PD98059 and U0126 as inhibitors of BCR-induced ERK1/2 activity in ex vivo splenic B lymphocyte cultures, whole cell extracts were prepared from quiescent B cells, and cells were mitogenically stimulated with 10 g/ml F(abЈ) 2 fragments of mouse anti-IgM (anti-Ig) pretreated for 60 min in the presence or absence of PD98059 or U0126. The concentrations of MEK1/2 inhibitors used herein were selected based on a recent report demonstrating that 10 M PD98059 or 10 M U0126 inhibited proliferation of splenic B cells stimulated with anti-Ig (20). 2 Immunoblotting was carried out with highly specific anti-p44/p42 ERK, anti-SAPK/JNK, and anti-p38 MAPK Abs that recognize the corresponding dually phosphorylated threonine and tyrosine motifs, indicative of catalytically active MAPKs. Treatment with either 10 M PD98059 or 10 M U0126 blocked BCR-induced ERK1/2 phosphorylation in splenic B cells when measured at 20 min (Fig. 1A). U0126 appeared to be more effective than PD98059 as an inhibitor of BCR-induced ERK1/2 phosphorylation at the concentration tested. The membrane was reprobed with anti-p44/p42 ERK Ab to confirm that the differences in phosphorylation were not due to changes in total cellular ERK. By contrast, BCR-induced phosphorylation of SAPK/JNK or p38 MAPK was not measurably reduced by inhibition of MEK1/2 (Fig. 1A). The extracts were also probed with anti-␤actin mAb to verify that an equal amount of cellular protein was being compared between conditions (Fig. 1A). It is noteworthy that basal and BCR-induced phosphorylation of individual MAPKs were not affected by pretreatment of parallel B cells with the corresponding solvent controls (data not shown). Collectively, these data indicate that PD98059 and U0126 selectively inhibit ERK activation in response to BCR crosslinking. These results are consistent with a more detailed report confirming the selectively of these reagents as inhibitors of the MEK1/2-ERK pathway in splenic B cells (20).
BCR Induction of Cyclin D2 mRNA Is Dependent on MEK1/ 2-ERK Activity-Several laboratories, including our own, have shown that cross-linking the BCR induces cyclin D2 protein expression (13,14). By contrast, little is known about the regulation of cyclin D2 mRNA expression in mature B cells. A report by Reid and Snow (28) indicates that mitogenic stimulation of splenic B cells with phorbol diester plus ionomycin leads to expression of cyclin D2 mRNA within 4 h, with maximal levels occurring by 18 h. Based on these results, RNA was isolated from splenic B cells stimulated with 10 g/ml anti-Ig for 4 and 18 h, and cyclin D2 gene expression was measured by RT-PCR. Control B cells express a relatively low level of cyclin D2 mRNA, which increased ϳ2and 7-fold in response to BCR cross-linking at 4 and 18 h, respectively (Fig. 1B). To test whether activation of the MEK1/2-ERK pathway was required for BCR-induced cyclin D2 mRNA accumulation, B cells were cultured in medium alone or stimulated with 10 g/ml anti-Ig, after a 60-min preincubation in the presence of 10 M U0126. Anti-Ig-induced cyclin D2 mRNA accumulation was completely suppressed by 10 M U0126 at the 4-and 18-h time points (Fig.  1C). Treatment with 10 M PD98059 also led to a reduced accumulation of cyclin D2 mRNA in response to anti-Ig stimulation at 4 and 18 h (Fig. 1D); however, PD98059 did not completely suppress cyclin D2 mRNA induction, which may reflect its inability to completely suppress BCR-induced ERK activation (Fig. 1A). These data suggest that MEK1/2-ERK activity is required for the accumulation of cyclin D2 mRNA following BCR cross-linking in splenic B cells.
BCR Induction of Cyclin D2 Protein but Not CDK4 Requires Active MEK1/2-ERK-By having observed that the accumulation of cyclin D2 mRNA in response to BCR cross-linking is dependent upon an active MEK1/2-ERK pathway, we next determined whether these effects were mirrored at the level of cyclin D2 protein. Quiescent splenic B cells were pretreated with U0126 or PD98059 as above and then stimulated with 10 g/ml anti-Ig for 18 h. The presence of 10 M U0126 led to a nearly complete inhibition of anti-Ig-induced cyclin D2 protein accumulation ( Fig. 2A). PD98059 also reduced cyclin D2 protein accumulation in response to anti-Ig stimulation, although its inhibitory effect was not as great in comparison with U0126 (Fig. 2B). The reduction in cyclin D2 protein in the presence of the MEK1/2 inhibitors did not arise from changes in the total amount of cellular protein being compared across conditions, as evidenced by equal levels of ␤-actin expression (Fig. 2, ␤-actin).
CDK4 represents the primary G 1 -CDK target of cyclin D2 in mature splenic B cells, and the initial phosphorylation of endogenous pRb is mediated by cyclin D2-CDK4 holoenzyme complexes (5,13). Therefore, we sought to determine if the levels of CDK4 protein and pRb phosphorylation were affected in B cells treated with MEK1/2 inhibitors. Western blot analysis of whole cell lysates from quiescent B cells stimulated with 10 g/ml anti-Ig revealed that the accumulation of CDK4 protein was not affected by treatment with 10 M U0126 at 18 h (Fig. 2C). Although CDK4 protein levels are not affected by U0126 or PD98059 (data not shown), the lack of sustained cyclin D2 accumulation is probably sufficient to prevent activation of endogenous CDK4. This is consistent with the finding that treatment with 10 M U0126 is sufficient to inhibit anti-Iginduced hyperphosphorylation of endogenous pRb (Fig. 2D). We also monitored activity of in vivo D-type cyclin-associated

hsp90-dependent Pathway Couples BCR to Cyclin D2 Accumulation
CDK activity in splenic B cells using a highly specific Ab that recognizes phosphorylation of pRb on Ser 807 /Ser 811 by CDK4/6 (29,30). Splenic B cells treated with 10 M U0126 exhibited a nearly complete inhibition of BCR-induced pRb phosphorylation on serine residues 807/811, suggesting that the MEK1/2-ERK pathway is required for endogenous CDK4/6 activity (Fig.  2E).
Geldanamycin Impairs Signaling through MEK1/2-ERK in Response to BCR Cross-linking-To gain further support for the involvement of the MEK1/2-ERK pathway in BCR signaling to cyclin D2 accumulation, we sought to determine if the benzoquinone ansamycin anti-tumor drug geldanamycin (GA) inhibits activation of the MEK1/2-ERK pathway via depletion of cellular Raf-1. hsp90 is a selective target of GA, and upon binding it leads to destabilization of hsp90 client proteins (31,32). In the case of Raf-1, GA promotes its degradation via a proteasome-mediated pathway (33)(34)(35). Treatment of quiescent splenic B cells with 2 M GA for 4 h led to a nearly complete depletion of cellular Raf-1, whereas MEK1/2, p44/p42 ERK, and hsp90 levels were not decreased (Fig. 3A). We observed protection of Raf-1 depletion from GA treated B cells by the proteasome inhibitor, MG-132 (Fig. 3B). Moreover, GA treatment led to a rapid disruption of Raf-1-hsp90 complexes, which preceded Raf-1 depletion, as evidenced by a reduced amount of coprecipitated hsp90 in Raf-1 immune complexes (Fig. 3C).
We next tested whether GA blocks downstream activation of MEK1/2 and ERK1/2 in response to BCR cross-linking. BCR signals phosphorylation of MEK1/2 within 1-5 min, with phosphorylation decreasing to control levels by 20 min (Fig. 4A). Treatment with 2 or 0.2 M GA blocked BCR-induced phosphorylation of MEK1/2 on activation residues Ser 217 /Ser 221 (Fig.  4B). BCR-induced phosphorylation of ERK1/2 on activation residues Thr 201 /Tyr 204 was also abrogated (Fig. 4B). To test whether the inhibitory action of GA was specific for the Raf-1-MEK1/2-ERK pathway, phosphorylation of Akt on Ser 473 and CREB on Ser 133 was monitored as downstream readouts for phosphatidylinositol 3-kinase and PLC␥2-protein kinase C pathway activation, respectively (36,37). As shown in Fig. 4C, GA had no measurable inhibitory effect on anti-Ig-induced phosphorylation of CREB on Ser 133 . Similarly, the rapid phosphorylation of Akt in response to BCR cross-linking was not inhibited by GA at the time points examined (Fig. 4C). Collectively, these results suggest that GA does not impair general BCR-mediated signaling but rather disrupts Raf-1-dependent signaling through MEK1/2 and ERK.
Cyclin D2 Accumulation and Proliferation in Response to Anti-Ig Stimulation of Splenic B Lymphocytes Is Blocked by GA-We observed that GA treatment of splenic B cells led to a nearly complete block in cyclin D2 mRNA accumulation at 4 h, whereas cyclin D2 mRNA accumulation was completely blocked at 18 h following stimulation with 10 g/ml anti-Ig (Fig. 5A). This was accompanied by a suppression of cyclin D2 protein accumulation in response to BCR cross-linking at several time points examined (Fig. 5B). By contrast, the accumulation of CDK4 protein in response to BCR cross-linking was reduced ϳ40% by GA, as determined by analyzing the resulting autoradiogram by densitometry (Fig. 5B). Since cyclin D2 accumulation is rate-limiting for G 1 -to-S phase progression in splenic B cells stimulated with 10 g/ml anti-Ig, we tested the functional consequences of GA-dependent inhibition of cyclin D2 accumulation on proliferation. As shown in Fig. 5C, splenic B cells undergo cellular proliferation upon BCR cross-linking with 10 g/ml anti-Ig. In comparison, treatment with 2 or 0.2 M GA completely blocked anti-Ig-induced proliferation of splenic B cells as monitored by [ 3 H]thymidine incorporation. DISCUSSION Mature B cells activate a Raf-1-dependent MEK1/2 and ERK2 signaling module upon BCR cross-linking (2,36). A recent report by DeFranco and co-workers (20) provided evidence that BCR-induced proliferation of mature splenic B cells requires signaling through MEK1/2 and ERK; however, the mechanism by which this pathway mediates BCR-induced proliferation remains unknown. Given that the principal initial determinant to G 1 -to-S phase progression lies in the sustained accumulation of cyclin D2, we sought herein to examine whether components of the MEK1/2-ERK pathway are required for induction of cyclin D2 expression by the BCR (13,15,28). We show that the MEK1/2 inhibitors, PD98059 and U0126, inhibit BCR-mediated activation of ERK. Although certain caveats exist with regard to the specificity of such reagents in vivo, neither reagent measurably inhibited BCR-induced activation of SAPK/JNK or p38 MAPK. Importantly, treatment with U0126 and, to a lesser degree, PD98059 inhibits BCRinduced expression of cyclin D2 at the mRNA and protein levels. The block in cyclin D2 expression is accompanied by impaired phosphorylation of endogenous pRb on CDK4/6-targeted Ser 807 /Ser 811 residues. These findings provide a molecular explanation for the block in BCR-induced proliferation of splenic B cells in which MEK1/2 activity has been abrogated, as recently reported by DeFranco and co-workers (20).
Experiments in which hsp90 function was disrupted provide additional support for the role of MEK1/2 and ERK in BCRinduced cyclin D2 mRNA expression. hsp90 functions as a general protein chaperone, maintaining the functional state of a select group of protein kinases (e.g. Raf-1 and CDK4) involved in signal transduction and cell cycle control (34,35,38). The ansamycin antibiotic GA disrupts the association of hsp90 with its client proteins (33)(34)(35). Only recently was it discovered that GA binds with high specificity within the ADP/ATP binding pocket of the hsp90, thereby inhibiting the function of hsp90 (31). GA also targets the hsp90 homolog, GP96, localized in the endoplasmic reticulum (32). Interestingly, ansamycins cause a pRb-dependent G 1 phase arrest in tumor cells (39). We find that treatment of normal splenic B cells with GA inhibits BCR-induced activation of MEK1/2 and ERK and that this occurs via a mechanism involving proteasome-mediated depletion of endogenous Raf-1 but not MEK1/2 or ERK. Perhaps most importantly, GA-mediated disruption of BCR-coupled signaling through Raf-1-MEK1/2-ERK prevents downstream induction of cyclin D2 mRNA expression. Consistent with the role of cyclin D2 in mediating G 1 -to-S phase progression, we find that GA blocks anti-Ig-stimulated splenic B cell proliferation.
The findings herein also indicate that cyclin D2 protein accumulation is blocked following inhibition of BCR-coupled Raf-1-MEK1/2-ERK signaling. This is probably secondary to the block in cyclin D2 mRNA expression, in part because up-regulation of cyclin D2 protein in quiescent B cells stimulated with anti-Ig requires de novo mRNA and cyclin D2 protein synthesis (13,40). We cannot rule out the possibility that components of the Raf-1-MEK1/2-ERK pathway may act concurrently to control the magnitude and duration of cyclin D2 protein availability via post-transcriptional mechanisms (4,18,19,41). In this view, it may be significant that a recent report by Lam and co-workers (42) suggests that inhibition of phosphatidylinositol 3-kinase with LY294002 down-regulates cyclin D2 steady-state levels in immature WEHI-231 B-cell lymphomas.
Our finding that the MEK1/2-ERK pathway is necessary for BCR-mediated cyclin D2 gene expression in normal splenic B cells is supported, albeit indirectly, by studies in mice containing mutant Btk. B cells from Btk-null and xid mutant mice exhibit defective cyclin D2 expression and fail to proliferate in response to BCR cross-linking (43,44). BCR-mediated ERK activity in splenic B cells from xid mutant mice is severely attenuated in comparison with wild-type littermates, suggesting that Btk is important for BCR-mediated ERK activation in murine splenic B cells (45). These findings are also supported by studies in DT40 chicken B cells expressing a conditional Btk-estrogen receptor (Btk:ER) fusion protein and Btk-null DT40 B cells, for which BCR-induced ERK2 activation is de- were incubated for 4 and 18 h with 10 g/ml anti-Ig (aIg) in the presence or absence of a 60-min pretreatment with 2 M GA. RNA was isolated, and RT-PCR was carried out as described under "Experimental Procedures." Cyclin D2 expression is reported as a proportion (-fold induction) of expression present in unstimulated (time 0) B cell samples. B, quiescent splenic B lymphocytes (M) were stimulated with 10 g/ml anti-Ig (aIg) for 9 or 12 h, following a 60-min pretreatment with 2 M GA. Whole cell extracts were prepared and immunoblotted with either anti-cyclin D2, anti-CDK4, or anti-␤-actin Abs. C, splenic B cells were cultured in media alone (M) or stimulated with 10 g/ml anti-Ig (aIg), following a 60-min preincubation with GA at the indicated concentrations. After 42 h, [ 3 H]thymidine was added, and proliferation was assayed after an additional 6 h as described (13). These data are representative of three independent experiments. hsp90-dependent Pathway Couples BCR to Cyclin D2 Accumulation pendent on Btk (46,47).
xid B cells also exhibit impaired PLC␥2 activation and attenuated Ca 2ϩ mobilization (48). The significance of this signaling phenotype to the defect in BCR-induced cyclin D2 expression is not known for the xid mouse; however, a similar mutant phenotype has been demonstrated in BLNK-deficient mice (21). BLNK is an Src homology 2 and 3 domain-containing adaptor protein that is phosphorylated upon BCR ligation and links Syk activation to Vav, Nck, Grb2, and PLC␥2 (3). Anti-Ig-stimulated B cells from BLNK-deficient mice do not proliferate due to an inability to induce cyclin D2 (and CDK4) expression (21). These cells exhibit impaired PLC␥2 activation and downstream activation of NF-B in response to BCR crosslinking; however, in contrast to xid mutant mice, BCR-induced activation of ERK, SAPK/JNK, and p38 MAPK appears normal. The phenotype of splenic B cells from mice null for the vav proto-oncogene, a Rho family guanine-nucleotide exchange factor, resembles the BLNK Ϫ/Ϫ mice (22). Vav-deficient splenic B cells are capable of activating ERK and SAPK/JNK in response to BCR cross-linking, yet these cells exhibit a defective [Ca 2ϩ ] response, impaired cyclin D2 expression, and fail to proliferate (22). Taken together, these observations suggest that cyclin D2 expression depends upon [Ca 2ϩ ] mobilization triggered by PLC␥2 activity. Moreover, the BLNK Ϫ/Ϫ and Vav Ϫ/Ϫ mutant phenotypes imply that expression of cyclin D2 is impaired despite the presence of active ERK.
The above findings seem to contradict the studies herein with normal splenic B cells. The discrepancy between these data may reflect the readouts for cyclin D2 expression used, namely that the BLNK-and Vav-deficient studies examined cyclin D2 protein levels, whereas our study has identified a role for MEK1/2-ERK in cyclin D2 gene expression. With this in mind, it can be envisaged that cyclin D2 protein availability may be controlled at the translational and/or post-translational levels through PLC␥2 and/or Ca 2ϩ -dependent pathways, whereas cyclin D2 gene expression is regulated by the MEK1/ 2-ERK pathway. Equal is the consideration that full induction of cyclin D2 gene expression is perhaps achieved by several distinct signaling pathways (i.e. Raf-1-MEK1/2-ERK, PLC␥2, and/or Ca 2ϩ mobilization). We base this conclusion on a recent study involving the regulation of cyclin D2 gene transcription by the colony-stimulating factor-1 receptor. Dey et al. (49) found that in BAC1.2F5 macrophages colony-stimulating factor-1 receptor signaling to cyclin D2 gene promoter activation required three distinct pathways, a rottlerin-sensitive protein kinase C␦-c-myc pathway, ERK, and a Src-dependent pathway. Additional work is needed to determine whether a similar scenario exists in mature B lymphocytes and specifically whether PLC␥2-and/or Ca 2ϩ -dependent pathways regulate cyclin D2 gene expression.
The requirement for c-myc in colony-stimulating factor-1 receptor induced cyclin D2 gene expression may be significant to BCR signaling, given a recent report demonstrating induction of cyclin D2 mRNA after expression of a conditional Mycestrogen receptor (Myc-ER) fusion protein (50). Furthermore, Myc, as part of a heterodimeric complex with Max, has been shown to activate the murine cyclin D2 gene promoter via a high affinity E-box in Rat1 and NIH-3T3 cells (51,52). In this view, the established role for NF-B/Rel in regulating c-myc expression in murine B cells may provide a mechanism whereby PLC␥2, and hence Ca 2ϩ mobilization could regulate cyclin D2 gene expression (53). By contrast, the mechanism by which the Raf-1-MEK1/2-ERK pathway couples BCR to cyclin D2 mRNA accumulation in mature B cells is not yet defined. The 5Ј-flanking region of the murine cyclin D2 gene promoter contains several transcriptional factor binding sites that are targets for ERK; among these, PEA3, AP-1, CCAAT/enhancerbinding protein, CREB, and Sp1 have been shown to be directly phosphorylated by ERK or regulated via an intermediate downstream ERK-dependent kinase (54). Clearly, more work is warranted to identify the nuclear trans-acting factors that contribute to the full induction of cyclin D2 gene expression in response to BCR ligation. A schematic representation of the signaling pathways by which the BCR may induce cyclin D2 gene expression is shown in Fig. 6. It should be noted that additional signaling pathway(s) may also be involved in the induction of cyclin D2 gene expression.
In summary, this is the first study to provide direct evidence for a role for active Raf-1-MEK1/2-ERK in BCR-induced cyclin D2 expression in normal ex vivo splenic B cells. These findings also provide a molecular explanation for the requirement of MEK1/2 in BCR-induced proliferation of splenic B cells.