CD40 Ligand-mediated Activation of the de Novo RelB NF-κB Synthesis Pathway in Transformed B Cells Promotes Rescue from Apoptosis*

CD40, a tumor necrosis factor receptor family member, is expressed on B lymphocytes. Interaction between CD40 and its ligand (CD40L), expressed on activated T lymphocytes, is critical for B cell survival. Here, we demonstrate that CD40 signals B cell survival in part via transcriptional activation of the RelB NF-κB subunit. CD40L treatment of chronic lymphocytic leukemia cells induced levels of relB mRNA. Similarly, CD40L-mediated rescue of WEHI 231 B lymphoma cells from apoptosis induced upon B cell receptor (surface IgM) engagement led to increased relB mRNA levels. Recently, we characterized a new de novo synthesis pathway for the RelB NF-κB subunit, induced by the cytomegalovirus IE1 protein, in which binding of p50/p65 NF-κB and c-Jun/Fra-2 AP-1 complexes to the relB promoter works in synergy to potently activate transcription (Wang, X., and Sonenshein, G. E. (2005) J. Virol. 79, 95–105). CD40L treatment of WEHI 231 cells caused induction of AP-1 family members Fra-2, c-Jun, JunD, and JunB. Cotransfection of Fra-2 with the Jun AP-1 subunits and p50/c-Rel NF-κB led to synergistic activation of the relB promoter. Ectopic expression of relB or RelB knockdown using small interfering RNA demonstrated the important role of this subunit in control of WEHI 231 cell survival and implicated activation of the anti-apoptotic factors Survivin and manganese superoxide dismutase. Thus, CD40 engagement of transformed B cells activates relB gene transcription via a process we have termed the de novo RelB synthesis pathway, which protects these cells from apoptosis.

CD40, a 48-kDa transmembrane protein belonging to the tumor necrosis factor receptor superfamily, is expressed on B lymphocytes as well as dendritic cells, macrophages, epithelial cells, and hematopoietic progenitor cells (1). Its ligand (CD40L) 2 is found on activated T lymphocytes; interaction between the receptor and ligand induces B cell activation and proliferation, Ig secretion, memory cell formation, and isotype switching (2). The critical role that CD40 plays in B cell function is made evident by the absence of germinal centers and secondary immune responses in CD40-deficient mice and the association of mutations in CD40 and CD40L with the human disease X-linked hyper-IgM syndrome (3). In addition, CD40 signaling is also involved in neoplastic cell proliferation. The interaction between CD40 and its ligand can inhibit apoptosis of normal and transformed B cells induced by engagement of the B cell receptor (BCR), serum deprivation, or treatment with a chemotherapeutic agent (4). For example, triggering of CD40 by the anti-CD40 monoclonal antibody G28-5 inhibits apoptosis induced by the chemotherapeutic agent fludarabine in B cells from patients with chronic lymphocytic leukemia (CLL) (5). A variety of second messengers are activated after treatment of B cells with CD40L, including NF-B (6). In WEHI 231 murine B lymphoma cells, a model often used to study NF-B and its effects on B cell survival, CD40L stimulation following BCR engagement rescues cells from apoptosis (7). BCR engagement of WEHI 231 cells normally leads to an initial transient increase in p50/c-Rel NF-B factor activity at 1 h, which is followed by a rapid decline in levels (8), and then to apoptosis (9). Studies from our laboratories have shown that CD40L-mediated rescue of WEHI 231 B cells is due in part to maintenance of NF-B factor binding (10), although the nature of the subunits responsible was not determined.
The mammalian NF-B family members are p65 (RelA), RelB, c-Rel, p105/p50, and p100/p52. These subunits contain a 300-amino acid long region termed the Rel homology domain, which is involved in subunit dimerization and binding to DNA (11). For the most part, the different members can form heteroand homodimers, which vary significantly in their transactivation potential (11,12). As we first showed by antisense and ectopic c-Rel expression studies in splenic B lymphocytes and WEHI 231 B cells (9), NF-B transcription factors play critical roles in proliferation control and B cell survival (13,14). NF-B, which regulates the c-myc, and c-myb genes (8,15,16), has also been implicated in promoting neoplastic transformation (13,14). In most non-B cells, NF-B is sequestered in the cytoplasm bound to specific inhibitory proteins (IB (inhibitor of B protein)) of which IB-␣ is the prototype. In B cells, NF-B activity is constitutive, but can be further induced or modulated (17)(18)(19). Activation of classical NF-B (p50/p65) through the canonical pathway is mediated via the IB kinase complex, con-taining two kinases, IB kinase-␣ and -␤, and multiple copies of a structural protein IB kinase-␥ or NEMO (20,21). The IB kinase complex phosphorylates IB-␣, leading to its subsequent ubiquitination and degradation. In the alternative NF-B activation pathway, p100/RelB complexes are sequestered in the cytoplasm via the C-terminal ankyrin repeats of p100 (20). Following activation of IB kinase-␣, p100 is phosphorylated, ubiquitinated, and clipped by the proteasome into p52, and the resulting p52/RelB complexes can translocate to the nucleus. For example, CD40L treatment of murine B lymphocytes induces p52/RelB complexes via the alternative pathway (22,23). In addition to post-translational regulation, RelB can also be controlled at a transcriptional level. The relB promoter has two proximal B elements and is an NF-B target (24). Our laboratory characterized a novel de novo RelB synthesis pathway induced by the cytomegalovirus IE1 (immediate early-1) protein that occurs via binding of both classical NF-B and AP-1 complexes containing c-Jun with either Fra-2 or Fra-1 to a distal AP-1 element, leading to synergistic transcriptional activation of the relB promoter (25).
Recently, Gricks et al. (26) used microarray analysis to begin to assess the effects of CD40L treatment on patients with CLL and noted an apparent activation of relB mRNA levels, which was absent in cells from healthy individuals. The physiological relevance of this activation was not further elucidated. Here, we confirm that CD40L induces relB mRNA levels in CLL cells as well as in WEHI 231 B lymphoma cells. Using WEHI 231 cells as a model, we show for the first time that CD40L engagement causes induction of the de novo RelB synthesis pathway, involving c-Jun, JunD, JunB, and Fra-2 AP-1 factors that work in synergy with p50/c-Rel NF-B. Notably, RelB plays a critical role in rescue from apoptosis, indicating that transcriptional regulation of the RelB NF-B subunit is a new mechanism whereby cell survival is controlled.

EXPERIMENTAL PROCEDURES
Cell Isolation, Culture, and Treatment Conditions-Peripheral blood was drawn from five CLL patients with flow cytometry-confirmed B-CLL with an Institutional Review Board approved consent. Patient samples were arbitrarily labeled JM37, JM40, JM42, JM61, and JM62. Leukemic cells were isolated by centrifugation using Histopaque 1077 (Sigma). Cultures contained Ͼ90% leukemic cells, which were maintained in RPMI 1640 medium (BioWhittaker, Walkersville, MD) supplemented with 10% fetal calf serum, 2 mmol/liter L-glutamine, 100 units/ml penicillin, and 100 units/ml streptomycin (Sigma) at 37°C and 5% CO 2 . After isolation, cells were allowed to rest for 24 h in fresh medium. For CD40L treatment, 20 -30 ϫ 10 6 cells were treated with supernatants containing CD40L and anti-CD8 antibody at the optimal concentrations of 1:5 and 1:40, respectively, as described previously (10). WEHI 231 murine B lymphoma cells, which express abundant constitutive levels of p50/c-Rel and p50 homodimers, were grown as described previously (9). Sixteen hours prior to treatment, WEHI 231 cells were plated in fresh medium at a density of 2.2 ϫ 10 5 cells/ml. Goat anti-mouse IgM (Jackson Immuno-Research Laboratories) was added to a final concentration of 15.0 g/ml for the specified times. Where indicated, cells were treated with supernatants containing CD40L and anti-CD8 antibody at the optimal concentrations of 1:5 and 1:40, respectively, as described previously (10). NIH 3T3 fibroblasts were grown and transfected as published previously (25).
RelB and c-Rel WEHI 231 Stable Transfectants-WEHI 231 cells were electroporated as described previously (33) with either the murine RelB or c-Rel expression vector or their corresponding EV DNAs. Mixed populations of stable transfectants were selected with 1400 g/ml Geneticin (G418, Sigma) for 2 weeks. Single clones were obtained by serial dilution of the mixed population of transfectants.
Retroviral Gene Delivery-Retrovirus stock was generated using the ecotropic packaging cell line Bosc23 (34). The pCLEco vector was cotransfected with either pRelB-sense or pRelB-siRNA into Bosc23 cells using FuGene 6 (Roche Applied Science). Two days later, supernatants containing viral particles were harvested and used to infect WEHI 231 cells in the presence of 6 g/ml Polybrene (Sigma). After 48 h, infected cells were selected with 0.4 g/ml puromycin (Sigma). Single clones were obtained by serial dilution of the mixed population of RelB siRNA transfectants.
Electrophoretic Mobility Shift Assay (EMSA)-Nuclear extracts were prepared and subjected to EMSA as described previously (36). The relB AP-1 element sequence is 5Ј-GATC-CTGCAATAGCGTCATCACAG-3Ј (25), where the underlined region indicates the core binding element. The sequence of the Sp-1 oligonucleotide (used as a control for equal extract loading) is 5Ј-ATTCGATCGGGGCGGGGCGACC-3Ј.
Cell Viability, Cell Cycle, and Apoptosis Assays-For quantification of viable cells, cells were incubated with 0.2% trypan blue for 10 -20 min, and the percentage of dead cells was determined as described previously (9). Treatments were done in triplicate, and values shown are each the mean Ϯ S.D. of three counts. For evaluation of cellular DNA content, cells stained with propidium iodide were analyzed by flow cytometry as described previously (37). Data (5000 events) were collected on linear and log scales to assess cell cycle progression and apoptosis, respectively, and analyzed using WinMDI software.

RESULTS
CD40L Induces relB mRNA Levels in CLL Cells-To begin to measure the effects of CD40L on relB mRNA expression, B-CLL cells were isolated from three patients (coded JM37, JM40, and JM42) and left untreated or treated with CD40L for 24 h. RNA was isolated and subjected to RT-PCR analysis, and values for relB were normalized to GAPDH, which served as a loading control (Fig. 1A). Stimulation of CLL cells from patients JM37, JM40, and JM42 with CD40L resulted in increases in the relB mRNA levels of 7.1-, 6.6-, and 4.6-fold, respectively. These findings are consistent with the microarray analysis performed by Gricks et al. (26), who found that CLL cells responded to CD40L treatment via activation of relB mRNA, whereas repression of relB was seen in B cells from healthy individuals.
CD40L Increases relB mRNA Levels in WEHI 231 Cells-We next measured whether the effects of CD40L treatment on relB mRNA levels can be extended to other transformed B cells and selected WEHI 231 B lymphoma cells, in which CD40L promotes cell survival from anti-IgM-induced apoptosis (7). WEHI 231 cells were treated for 6 h with CD40L or anti-IgM alone or in combination, and cytoplasmic RNA was subjected to Northern blot analysis (Fig. 1B). Treatment of WEHI 231 cells with CD40L either alone or in combination with anti-IgM increased relB mRNA levels, whereas anti-IgM treatment alone decreased the levels. In this and a duplicate experiment, treatment with CD40L or CD40L plus anti-IgM resulted in 3.1-and 2.9-fold increases in relB mRNA levels, respectively, whereas anti-IgM treatment resulted in a 0.6-fold decrease (Fig. 1C). Thus, CD40L treatment induces relB mRNA in WEHI 231 B lymphoma cells, as seen in CLL cells.
CD40L Causes an Increase in Nuclear Levels of RelB-To test for the expected activation of RelB protein, B-CLL cells isolated from three patients (coded JM42, JM61, and JM62) were left untreated or treated with CD40L for 24 h. Nuclear proteins were extracted and subjected to immunoblot analysis for RelB and Oct-1 (used as a loading control) (Fig. 1D). CD40L stimulation of CLL cells resulted in a dramatic induction of RelB protein, ranging from 8.7-to 11.4-fold compared with untreated samples. On average, the induction of RelB protein levels appeared to be higher than that of relB mRNA levels upon CD40L stimulation (Fig. 1A). CD40L treatment has been shown to activate p52/RelB complexes via the alternative pathway, which could explain this difference in protein and mRNA levels.
To test for this pathway, the nuclear levels of p52 were assessed (Fig. 1D). In untreated CLL cells, p52 levels were not detectable (Fig. 1D). Upon CD40L stimulation, nuclear p52 levels increased in all three CLL patient samples. The data indicate that the alternative pathway is activated in CLL cells upon CD40L stimulation and could account for the observed disparity between RelB protein and relB mRNA levels. Quantitative assessment of the contribution of this pathway to the RelB increase was not possible, however, as we were unable to detect p52 in the untreated samples.
We next studied the kinetics of changes in RelB protein levels in WEHI 231 cells induced by CD40 engagement. Nuclear proteins were isolated from cells in exponential growth or following treatment for 1, 4, or 10 h with CD40L or anti-IgM alone or in combination and analyzed by immunoblotting. An increase in the nuclear levels of RelB was detectable by 1 h of treatment with either CD40L alone or plus anti-IgM and was more substantial by 4 h and increased further by 10 h (Fig. 2A). Analysis of Oct-1 levels confirmed essentially equal loading. The immunoblots were subjected to densitometry, and the data for RelB normalized to Oct-1 are presented in Fig. 2C. By 10 h, CD40L alone or with anti-IgM increased RelB levels by 10.4-and 9.1-fold, respectively. The data at the 4-h time point from this and a duplicate experiment were plotted. Similarly to what was seen after CD40L stimulation of CLL cells, treatment of WEHI 231 cells with CD40L alone or in combination with anti-IgM resulted in 8.8-and 6.6-fold increases in RelB (Fig. 2D), respectively, which were substantially higher than the ϳ3-fold increase in mRNA levels observed at 6 h (Fig. 1C). , and JM42 were either left untreated or stimulated with CD40L for 24 h. Total RNA was extracted and analyzed by RT-PCR for the RNA levels of relB and GAPDH (used as a control for equal loading). The expression levels of relB were normalized to those of GAPDH and are presented as the -fold change relative to untreated samples (set at 1). B, cytoplasmic mRNA was isolated from asynchronously growing WEHI 231 cells (0) or following treatment with anti-IgM (␣IgM) or CD40L alone or in combination for 6 h, and samples (20 g) were subjected to Northern blot analysis using as a DNA probe either murine relB or Gapdh (used as a loading control). C, the relB blot in B and that from a duplicate experiment were subjected to densitometry and normalized to Gapdh. The average -fold change Ϯ S.E. in relB mRNA for the different treatments relative to levels in control untreated cells (set at 1) is presented. D, B-CLL cells isolated from patients JM42, JM61, and JM62 were either left untreated or stimulated with CD40L for 24 h. Nuclear extracts were prepared and subjected to immunoblot analysis for the levels of RelB, p52, and Oct-1 (used as a control for equal loading). Next, to test for alternative pathway activation in WEHI 231 cells, the nuclear levels of p52 were assessed ( Fig. 2A). CD40L alone or with anti-IgM caused a detectable increase in p52 at 1 h, which continued to 4 h and then either remained elevated or dropped back slightly by the 10-h time point, such that levels were increased by 3.6-or 2.3-fold, respectively, at the 10-h time point (Fig. 2C) and by 4.2-or 2.3-fold at 4 h (Fig. 2D). Anti-IgM treatment alone led to decreased p52 levels (Fig. 2), suggesting that the lower extent of increase in p52 upon the co-treatment probably results from anti-IgM-induced signaling. Taken together, these data suggest that CD40L leads to activation of the de novo RelB synthesis pathway, which leads to a sustained increase in relB mRNA and RelB subunit expression, and to the alternative pathway, which leads to an increase in the levels of p52/RelB complexes, which appears more transient.
Treatment with anti-IgM alone led to a slight increase in RelB levels and to substantial decreases in c-Rel and p65 levels between 1 and 4 h, which persisted to 10 h (Fig. 2B); we showed previously that this leads to apoptosis (8,9). Whereas CD40L treatment alone caused substantial increases in the levels of c-Rel, it was insufficient to overcome anti-IgMmediated shutoff of this subunit (Fig. 2B). CD40L treatment resulted in large increases in p65 and p50 and led to maintenance of levels equivalent to those in untreated cells upon co-treatment. Thus, CD40L-mediated rescue from anti-IgM-induced cell death also overrides the shutoff of these two NF-B subunits.
CD40L Induces AP-1 Family Members-In the cytomegalovirus IE1-induced de novo RelB synthesis pathway, synergistic transcriptional activation of the relB promoter by binding of both classical NF-B and AP-1 complexes containing c-Jun with either Fra-2 or Fra-1 increases relB mRNA levels (25,38). Thus, we next tested for activation of AP-1 components that mediate de novo relB synthesis in WEHI 231 cells. Nuclear extracts were prepared from untreated WEHI 231 cells or following treatment with CD40L, anti-IgM, or anti-IgM plus CD40L for 4 h and subjected to EMSA for AP-1 binding using the AP-1 element upstream of the relB promoter as a probe. As seen in Fig. 3A, extracts from control WEHI 231 cells displayed two bands of AP-1 binding (bands 1 and 2). Whereas CD40L or anti-IgM treatment alone resulted in a slight increase in binding, a much more robust increase was seen upon anti-IgM/CD40L co-treatment. In addition, the appearance of a new faster migrating complex (band 3) was evident. Sp-1 binding confirmed essentially equal loading. The increases in AP-1 factor binding seen with either CD40L alone or the combined anti-IgM/CD40L treatment were maintained up to 10 h (Fig. 3B). Thus, CD40L-mediated rescue from anti-IgM-induced death of WEHI 231 cells leads to enhanced AP-1 factor binding. . CD40L treatment of WEHI 231 cells activates AP-1 family members. A, nuclear extracts were isolated from asynchronously growing WEHI 231 cells (0) or following treatment with anti-IgM (␣IgM) or CD40L alone or in combination for 4 h. Samples (2.5 g) were used in EMSA with the relB AP-1 site as a probe. Binding to an Sp-1 oligonucleotide was used as a control for equal loading. B, AP-1 binding to the relB AP-1 site was examined (as described for A) using extracts isolated following 10 h of treatment. C, nuclear extracts (50 -70 g) isolated from asynchronously growing WEHI 231 cells or from cells treated with anti-IgM or CD40L alone or in combination for 4 h were subjected to immunoblot analysis using antibodies against the indicated AP-1 family members. Oct-1 protein levels confirmed equal loading. ns, a nonspecific band previously reported in the literature (65); Jun D FL , full-length JunD isoform; Jun D ⌬ , truncated isoform of JunD generated through utilization of an alternative translation start site. The bar to the right indicates the position of a more slowly migrating Fra-2 complex likely the result of phosphorylation (39). D, nuclear extracts (2.5 g/sample) isolated from WEHI 231 cells exposed to the combined anti-IgM and CD40L treatment for 4 h were used in an EMSA in the absence (Ϫ) or presence of antibodies against c-Jun, JunD, JunB, Fra-2, Fra-1, and c-Fos. The relB AP-1 element was used as a probe. The middle and upper panels are darker exposures of the upper part of the lower panel, indicated by a bracket (7 and 60 days versus 2 days, respectively). The asterisks indicate a clearing in an AP-1 band, and the bars indicate a supershifting band.
To identify the nature of the individual AP-1 subunits activated by CD40L, anti-IgM, or anti-IgM plus CD40L co-treatment, nuclear extracts were subjected to immunoblotting for c-Jun, JunD, JunB, Fra-1, Fra-2, and c-Fos (Fig. 3C). The levels of c-Jun and JunD increased upon treatment with CD40L or anti-IgM alone, and treatment with both CD40L and anti-IgM showed essentially an additive effect. JunB levels were seen to increase substantially only in cells exposed to anti-IgM/CD40L co-treatment. In the case of Fra-2, a new slower migrating band was visible upon CD40L treatment alone or in combination with anti-IgM, consistent with the appearance of a phosphorylated, more active form (39). Fra-1 levels were barely detectable in untreated cells and increased modestly with anti-IgM and somewhat more substantially with the anti-IgM/CD40L co-treatment. Very low levels of c-Fos were detected in WEHI 231 cells and increased upon treatment only with anti-IgM, but appeared unaffected by CD40L (Fig. 3C). Thus, c-Jun, JunD, JunB, and Fra-2 levels or activities increase upon treatment with CD40L alone or in combination with anti-IgM and could participate in the de novo RelB synthesis pathway.
CD40L Induces Binding of c-Jun, JunD, JunB, and Fra-2 to the relB AP-1 Site-We next tested for binding of AP-1 family members to the relB AP-1 site. Nuclear extracts isolated from WEHI 231 cells co-treated with CD40L and anti-IgM for 4 h were used in antibody gel shift analysis. Addition of antibody against c-Jun, JunD, JunB, or Fra-2 resulted in a clearing of AP-1 bands (Fig. 3D, middle and lower panels) and the appearance of supershifted bands, which were more clearly visible in the darker exposures (Fig. 3D, upper and middle panels). No changes were observed in the samples with anti-Fra-1 and antic-Fos antibodies. Thus, the data confirm binding of the AP-1 family members c-Jun, JunD, JunB, and Fra-2 induced by anti-IgM/CD40L co-treatment.

p50/c-Rel NF-B Transactivates the relB Promoter in Synergy with AP-1 Complexes Containing Fra-2 with c-Jun, JunD, or
JunB-Previously, p50/p65 NF-B was shown to activate the relB promoter in synergy with c-Jun/Fra-2 AP-1 complexes (25). Although p50 and p65 subunit expression is induced by CD40L treatment, the predominant NF-B complex in WEHI 231 B lymphoma cells is p50/c-Rel (8,40). To determine whether p50/c-Rel complexes work in synergy with AP-1, in particular, c-Jun, JunB, JunD, and Fra-2, NIH 3T3 cells were selected for their high transfection efficiency and low basal levels of NF-B and AP-1 (25,32). NIH 3T3 cells were cotransfected with vectors expressing c-Jun, Fra-2, p50, or c-Rel alone or in combination and the p1.7 relB promoter-luciferase reporter construct (Fig. 4A). Transfection of c-Jun or Fra-2 alone or p50/c-Rel had little effect on relB promoter activity (Ͻ2-fold increase). Whereas the combination of c-Jun and Fra-2 increased relB promoter activity by 16.2 Ϯ 1.5fold, a striking 133.9 Ϯ 15.4-fold induction was seen upon cotransfection of vectors expressing c-Jun/Fra-2 and p50/c-Rel (Fig. 4A), indicating that these subunits have potent synergistic effects on relB promoter activity.
The ability of JunD and JunB to transactivate the relB promoter alone or in combination with Fra-2 and p50/c-Rel was similarly tested (Fig. 4, B and C). JunD alone was found to increase relB promoter activity by 4.6 Ϯ 0.1-fold, whereas Fra-2 alone again showed little effect (Fig. 4B). A combination of JunD and Fra-2 induced promoter activity by 45.3 Ϯ 2.8-fold, whereas a combination of JunD/Fra-2 and p50/c-Rel complexes caused an even larger increase in relB promoter activity of 115.4 Ϯ 12.9-fold. Finally, expression of JunB alone or in combination with Fra-2 resulted in a 10.9 Ϯ 1.2-or 45.7 Ϯ 2.3-fold induction, respectively (Fig. 4C). The combination of JunB/Fra-2 and p50/c-Rel complexes resulted in a more potent 80.4 Ϯ 3.8-fold increase in relB promoter activity. Taken together, these data show that Fra-2 plus any of the three Jun family members that are induced by CD40L (c-Jun, JunD, or JunB) can synergize with p50/ c-Rel NF-B to potently transactivate the relB promoter.
Knockdown of RelB Promotes Anti-IgM-induced Death of WEHI 231 Cells-To test whether endogenous RelB expression promotes WEHI 231 cell survival, an siRNA strategy and anti-IgM treatment were used. WEHI 231 cells were infected with pRelB-siRNA and control pRelB-sense retroviral vectors (expressing RelB siRNA and RelB control sense RNA, respectively) as described previously (29). Stable mixed populations of transfectants were selected using puromycin, and two individual clones were isolated by limiting dilution (WEHI-siRelB clones 8 and 10). Both WEHI-siRelB clones 8 and 10 had lower basal levels of nuclear RelB compared with the control WEHI-sense RelB population (Fig. 5A). After anti-IgM treatment for 24 h, RelB levels were seen to increase in all cells, as observed above ( Fig. 2A), but remained substantially lower in the two WEHI-siRelB clones compared with untreated control cells. An analysis of cell cycle progression by flow cytometry revealed considerably more growth arrest in the RelB siRNA-expressing cells upon anti-IgM treatment compared with control cells (62.6 and 63.4% of WEHI-siRelB clone 8 and 10 cells, respectively, were in G 0 /G 1 versus 45.1% of WEHI-sense RelB cells) (Fig. 5B). Previously, anti-IgMtreated WEHI 231 cells were shown to die via apoptosis using DNA laddering and flow cytometry analysis of sub-G 1 DNA content and dead cells quantified by trypan blue exclusion (9). When anti-IgM-induced cell death was assessed by trypan blue exclusion, WEHI-siRelB clones 8 and 10 displayed substantially increased levels of cell death (40.6 and 39.1%, respectively) compared with the 18.1% seen with the control WEHIsense RelB population after 48 h of anti-IgM treatment (Fig. 5C). These results were confirmed by flow cytometry (Fig. 5D). WEHI-siRelB clones 8 and 10 displayed a greater percentage of apoptotic cells (40.3 and 30.1%, respectively) compared with 18.0% apoptosis seen in the sense RelB population (Fig. 5D). These findings implicate RelB in the control of WEHI 231 B lymphoma cell viability.
Genes Encoding Anti-apoptotic Factors MnSOD and Survivin Are RelB Targets in WEHI 231 Cells-To identify possible targets of RelB that mediate survival, we examined the expression of MnSOD, Survivin, Bcl-2, and Bcl-x L , known NF-B targets with anti-apoptotic functions (41)(42)(43)(44), as well as of Bax, the pro-apoptotic Bcl-2 family member. Whole cell extracts were isolated from untreated WEHI-sense RelB cells and WEHI-siRelB clones 8 and 10 and subjected to immunoblot analysis (Fig. 6A). Knockdown of RelB was paralleled by a decrease in MnSOD and Survivin protein levels, whereas expression of Bcl-2, Bcl-x L , and Bax remained unchanged. To determine whether the decrease in MnSOD and Survivin could be related to a decline in mRNA levels, RT-PCR was performed (Fig. 6B). Expression of the RelB siRNA vector resulted in a 50 -70% repression of relB mRNA in the two clones, consistent with the RelB knockdown seen above. Of note, MnSOD and survivin RNA levels were substantially reduced in WEHI-siRelB clones 8 and 10: 56 -63% for MnSOD and 35-49% for survivin. Thus, these data implicate RelB in the control of expression of the two known anti-apoptotic factors MnSOD and Survivin in WEHI 231 cells.

Ectopic RelB Expression Induces Prosurvival Gene Expression and Rescues WEHI 231 Cells from Anti-IgM-induced Death-
To test whether ectopic RelB expression is sufficient to rescue cells from anti-IgM-mediated apoptosis, WEHI 231 cells were transfected with pMexNeo EV DNA or a murine relB expression vector DNA. Mixed populations of stable transfectants (WEHI-EV MP and WEHI-RelB MP cells) were selected and two clones (WEHI-RelB clones 1 and 7) were isolated from the WEHI-RelB MP cells by limiting dilution. Nuclear extracts were analyzed by immunoblotting for RelB expression. Basal levels were found to be similar in the untreated WEHI-EV MP and WEHI-RelB MP cells, but were significantly higher in WEHI-RelB clone 1, whereas intermediate levels were seen in WEHI-RelB clone 7 (Fig. 7A). Treatment with anti-IgM for 24 h resulted in a slight increase in RelB levels in WEHI-EV MP cells, which was more pronounced in the WEHI-RelB MP cells; the higher basal levels of nuclear RelB in WEHI-RelB clones 1 and 7 remained essentially constant (Fig. 7A). Notably, trypan blue staining demonstrated that ectopic RelB expression promoted survival of WEHI 231 cells following BCR engagement. Although 36.7% of control WEHI-EV MP cells died, only 20.2, 2.5, and 13.7% cell death was observed with WEHI-RelB MP cells and WEHI-RelB clones 1 and 7, respectively (Fig. 7B), consistent with the relative levels of RelB expression. Next, we examined the effects of RelB expression on MnSOD, Survivin, Bcl-2, Bcl-x L , and Bax protein levels (Fig. 7C). Ectopic RelB expression increased the levels of MnSOD and Survivin, whereas no changes in the levels of Bcl-2, Bcl-x L , and Bax were seen. Thus, RelB regulates MnSOD and Survivin. Consistent with these findings, CD40L treatment of B-CLL cells increased the mRNA levels of MNSOD and survivin (Fig. 7D). Taken together, these data indicate that ectopic RelB induces expression of prosurvival genes and that these proteins can protect WEHI 231 B lymphoma cells from anti-IgM-induced death.

Ectopic c-Rel Expression Induces Prosurvival Proteins and Rescues WEHI 231 Cells from Anti-IgM-induced Death-Previ-
ously, we demonstrated that ectopic c-Rel expression protects WEHI 231 B cells from apoptosis (9). To compare the prosurvival genes controlled by these two NF-B subunits, WEHI 231 cells were stably transfected with either the empty backbone vector (WEHI-EV) or a c-Rel expression vector, and single WEHI-c-Rel clones 3 and 4 were isolated. Overexpression of c-Rel in the two clonal lines, which was confirmed by immunoblotting, led to increased levels of MnSOD, Survivin, Bcl-x L , and Bcl-2 (Fig. 8A) and enhanced survival upon anti-IgM treatment (Fig. 8B), consistent with our earlier findings (9). Thus, c-Rel induces a wider spectrum of prosurvival proteins compared with RelB.

DISCUSSION
CD40L stimulation of both patient B-CLL cells and WEHI 231 B lymphoma cells induced mRNA levels of the relB NF-B subunit. In the WEHI 231 cell model system, CD40L-mediated RelB activation occurred via the newly identified de novo RelB synthesis pathway, and RelB was found to be essential for the protection of these cells from apoptosis induced by BCR engagement. Recently, we reported that cytomegalovirus IE1 protein mediates activation of p50/p65 NF-B and c-Jun/Fra-2, which work in synergy to potently induce relB gene transcription (25). Of note, CD40 activation was found to significantly increase relB mRNA levels in B cells isolated from CLL patients in microarray analysis (26). Using RT-PCR to quantify mRNA levels, we have confirmed that CD40L treatment of B-CLL cells results in increased relB mRNA levels. This mRNA increase could be reproduced in the WEHI 231 B lymphoma cell model. Furthermore, CD40L treatment of WEHI 231 cells was found to increase the nuclear levels of c-Jun, JunD, and Fra-2 phospho-FIGURE 6. MnSOD and Survivin are RelB targets in WEHI 231 cells. A, whole cell extracts were isolated from untreated WEHI-sense RelB or WEHI-siRelB clone 8 and 10 cells in asynchronous growth, and samples (20 -50 g) were subjected to immunoblot analysis for the levels of Survivin, MnSOD, Bax, Bclx L , Bcl-2, and ␤-actin (used as a loading control). B, cytoplasmic mRNA (5 g) isolated from asynchronously growing WEHI-sense RelB or WEHI-siRelB clone 8 and 10 cells was analyzed by RT-PCR for the RNA levels of relB, MnSOD, survivin, and Gapdh (used as a control for equal loading). Expression levels relative to control sense RelB (set at 100%) are given below each lane. rylation, whereas co-treatment with anti-IgM increased the levels of JunB as well. Notably, complexes of c-Jun, JunD, and JunB plus Fra-2 potently synergized with p50/c-Rel, the major NF-B complex in WEHI 231 cells, strongly inducing relB promoter activity. Thus, we have identified p50/c-Rel as a potential partner to potently activate relB promoter activity, and we have extended the AP-1 complexes that can work in synergy with Fra-2 to JunD and JunB. Previously, c-Jun/c-Fos AP-1 complexes were found to be unable to synergize with p50/p65 to activate the relB promoter (25); we have now found they are also unable to function with p50/c-Rel (data not shown). Overall, our findings identify a new pathway activated by CD40 receptor engagement: the de novo RelB synthesis pathway, which plays a critical role in survival of transformed B cells.
The RelB NF-B factor was originally reported to be associated with regulation of adaptive immune responses, specifically differentiation of dendritic cells (45,46). RelB-deficient mice show a variety of phenotypic abnormalities (45,(47)(48)(49). These include thymic atrophy due to a reduced population of dendritic and medullary epithelial cells, lack of clearly developed lymph nodes, myeloid hyperplasia, and abnormal microarchitecture of the spleen that does not support proper germinal center formation. More recently, RelB has been implicated in tumorigenesis and mammary gland development. For example, aberrantly elevated levels of nuclear RelB were observed in mammary tumors induced by dimethyl[a]benzanthracene treatment (50) or by enforced expression of c-Rel under the control of the murine mammary tumor virus in transgenic mice (31). Furthermore, delayed activation in the mammary gland of p52/RelB by the alternative pathway rescues gland development in a murine mammary tumor virus super-repressor IB-␣ mouse model (31). RelB is the most frequently detected nuclear NF-B subunit in prostate cancer, with levels that directly correlate with the Gleason score, indicating an association with cancer progression (51). Stoffel et al. (52) showed that p50/RelB complexes mediate inhibition of apoptosis induced by DNA damage in mucosa-associated lymphoid tissue lymphoma. Very recently, we observed RelB in estrogen receptor-␣-negative breast cancer cells and showed that it promotes survival and an invasive phenotype of these highly malignant cancers (53).
In normal B cells, the interaction between CD40 and its ligand provides critical signals for survival, proliferation, and eventual differentiation into antibody-producing plasma cells (54). Similarly, in low grade B cell malignancies, CD40 signaling protects from apoptosis and promotes proliferation (55). In some B-CLL, Burkitt lymphoma and non-Hodgkin lymphoma cells, this anti-apoptotic effect is achieved in the absence of a T cell stimulus, as concurrent expression of the CD40 receptor and ligand on the affected B cells provides constitutive activation of the CD40 signaling pathway and, more specifically, increased NF-B activity (56 -58). In addition, patients with B-CLL exhibit high levels of soluble CD40L in their peripheral RelB expression vector (WEHI-EV MP or WEHI-RelB MP cells, respectively) or single WEHI-RelB clones 1 and 7 were either left untreated (0) or treated with anti-IgM (␣IgM) for 24 h, and nuclear extracts were prepared. Samples (20 g) were subjected to immunoblot analysis for the levels of RelB and Oct-1 (used as a loading control). B, the stable lines in A were treated with anti-IgM for 48 h, and trypan blue analysis was performed to determine the percent dead cells in the population. C, whole cell extracts were isolated from untreated WEHI-EV MP, WEHI-RelB MP, and WEHI-RelB clone 1 and 7 cells, and samples (20 -50 g) were subjected to immunoblot analysis for RelB, MnSOD, Bcl-2, Bax, Survivin, Bcl-x L , and ␤-actin (used as a loading control). D, RNAs from patient JM37, JM40, and JM42 B-CLL cells either left untreated or stimulated with CD40L for 24 h (as described for Fig. 1A) were analyzed by RT-PCR for the levels of MNSOD, survivin, and GAPDH. blood associated with both tumor survival and resistance to chemotherapy (59). Interestingly, activation of RelB by CD40L was so potent that it was able to overwhelm the reduction by RelB siRNA (data not shown). Of note, Zarnegar et al. (60) have previously reported that CD40 ligation is able to maintain significant cell survival of B cells isolated from p50 Ϫ/Ϫ c-Rel Ϫ/Ϫ p65 ϩ/Ϫ mice compared with wild-type cells, suggesting a role for RelB. Our studies provide direct evidence for the effects of RelB on survival and proliferation of WEHI 231, extending the observations made previously (and confirmed here) showing that ectopic c-Rel expression can protect WEHI 231 cells from anti-IgM-induced death (9). RelB has been shown to regulate several oncogenes and genes that promote cell growth and survival, including c-myc, c-myb, cyclin D 1 , MnSOD, and bcl-x L (38,41,50,61,62). In our study, the levels of c-Myc were not altered upon overexpression or knockdown of RelB (data not shown), possibly because of the maintained expression of other NF-B subunits, which also target this factor. Although Gelinas and co-workers (63) identified the bcl-x L gene as a direct target of c-Rel and RelA, the role of RelB has been somewhat controversial. Jacque et al. (61) found that knockdown of RelB using siRNA technology strongly represses bcl-x L mRNA. In our studies, p50/RelB complexes failed to bind to the B element of the bcl-x L promoter or to induce its transcription (38). Consistent with our previous data, in this study, overexpression or knockdown of RelB in WEHI 231 cells did not change the levels of Bcl-x L , whereas ectopic c-Rel induced its expression. These differences suggest that RelB may be involved in bcl-x L regulation indirectly through a still unknown factor. Bcl-2 levels in WEHI 231 cells were also unchanged by RelB, whereas they were induced with c-Rel. Interestingly, our data implicate RelB in the regulation of the antioxidant and anti-apoptotic factor MnSOD in WEHI 231 cells. This is consistent with a recent report by Josson et al. (41) indicating that MnSOD, regulated through RelB, increases the resistance of prostate cancer cells to radiation. Finally, CD40L stimulation has been shown to improve the viability of B cells isolated from patients with CLL, leading to the specific up-regulation of Survivin (64). Here, we have shown that CD40L stimulation of patient B-CLL cells increases both survivin and MNSOD mRNA levels. Consistently, our data indicate that RelB regulates the levels of Survivin and MnSOD in WEHI 231 cells. Overall, our data identify a transcriptional mechanism regulating RelB expression that plays an important role in the control of B cell survival and proliferation.