B cell maturation Ag (BCMA), a member of the TNFR superfamily expressed on B cells, binds to a proliferation-inducing ligand (APRIL) and B cell-activating factor of the TNF family (BAFF) but the specific B cell responses regulated by BCMA remain unclear. This study demonstrates that ligation of A20 B cells transfected with BCMA induces the expression of CD40, CD80/B7-1, CD86/B7-2, MHC class II, and CD54/ICAM-1, which subsequently enhances the presentation of OVA peptide Ag to DO11.10 T cells. BCMA expression in murine splenic B cells can be induced with IL-4 and IL-6, allowing subsequent treatment with APRIL or agonist anti-BCMA to similarly induce Ag presentation. A comparative analysis of hybrid receptors of TNFR2 fused to the cytoplasmic domains of APRIL/BAFF receptors found that only BCMA, but not transmembrane activator and calcium-modulator and cyclophilin ligand interactor or BAFF-R, is capable of activating Ag presentation. Although all three receptors can trigger NF-κB signaling, only BCMA activates the JNK pathway conferring on BCMA the specific ability to activate this Ag presentation response.
Membersof the TNF and TNFR superfamilies are critical for the development and response of the immune system (1). Two recent additions to the TNF superfamily, B cell-activating factor (BAFF),3 also known as BLYS, TALL-1, zTNF4, or THANK, and a proliferation-inducing ligand (APRIL), also known as TALL-2 or TRDL-1, have been shown to have a role in regulating B cells (2). BAFF was initially cloned by sequence homology to the TNF family and was shown to stimulate B cell growth and Ig production in vitro (3, 4). BAFF was subsequently shown to be required for maintaining the survival of peripheral B cell as baff−/− mice lack follicular and marginal zone B cells (5, 6). The B cell developmental blockade in the BAFF-deficient animals appears to be in the T1 to T2 transition step during the maturation of B cells and as a result, these animals have impaired humoral responses to both T cell-dependent and T cell-independent Ags. In contrast, transgenic overexpression of BAFF in mice resulted in elevated number of B cells and these animals subsequently developed a systemic lupus erythematosus (SLE)-like autoimmune disorder (7, 8, 9). Furthermore, increased BAFF levels were observed in the NZW × BF1 mouse model of SLE as the disease progresses and antagonist soluble BAFF-R inhibited the development of lupus in these animals (9, 10). These observations in animal models have led to the suggestion that elevated BAFF levels may play a role in human autoimmune disorders. Consistent with this hypothesis, patients with SLE were reported to have elevated levels of BAFF when compared with normal individuals (11, 12).
In contrast to the well-characterized physiological role of BAFF, that of APRIL remains less clear. APRIL is the closest relative of BAFF in the TNF family, sharing 33% sequence identity, and was initially reported to stimulate the proliferation of tumor cells (13). Subsequent in vitro experiments demonstrated that APRIL could also stimulate proliferation, Ig secretion, and class switching in B cells (14, 15). Two groups recently generated april−/− mice with slightly different observations. One group reported that these mice did not have any overt immune defects (16), whereas a second group observed defects in isotype class switching to IgA (17). The reason for the discrepancy between the two reports is unclear at this point. Thus in contrast to BAFF, APRIL appears to have a role in the immune system that can be largely compensated by other cytokines. In addition to effects in the immune system, there is a growing body of evidence that BAFF and APRIL are likely to be involved in lymphoid malignancies by conferring antiapoptotic signals to tumor cells (18, 19, 20, 21, 22). BAFF and APRIL can be produced by the tumor cells and thus can act in an autocrine manner, or they can be produced by accessory cells to act in a paracrine manner.
Both BAFF and APRIL can bind to multiple receptors of the TNFR superfamily expressed on B cells. BAFF binds to BCMA, transmembrane activator and calcium-modulator and cyclophilin ligand interactor (TACI), and BAFF-R, whereas APRIL binds only to BCMA and TACI (2). The prosurvival effects of BAFF on B cells is most likely mediated by BAFF-R in that the A/WySnJ mouse strain that has a loss-of-function mutation in BAFF-R lacks peripheral B cells, a phenotype that is most similar to baff−/− mice (23, 24). On the contrary, taci−/− mice have elevated B cell numbers (25, 26), and older taci−/− mice subsequently develop autoimmunity (27), suggesting that TACI may play a negative regulatory role in B cells. To date, the function of BCMA has remained the most enigmatic of the three receptors. BCMA was initially cloned as a translocation product from a human T cell lymphoma and its expression profile was shown to correlate with the maturation of B cells, with the highest level observed in plasma cell lines (28). Although both BAFF and APRIL can bind to BCMA, APRIL binds with much higher affinity (29, 30) suggesting perhaps that the APRIL-BCMA interaction may be more physiologically relevant. However, the functional consequences of BCMA ligation on B cells have remained unclear and the generation of bcma−/− mice has not provided further clues to this question. Initial examination of the BCMA-deficient mice revealed no discernible phenotype (5, 31) but a more recent analysis indicated that BCMA has a role in maintaining the survival of long-lived plasma cells (32). Whether BCMA can regulate other functions of B cells has heretofore remained unknown.
In this study, we report our cloning of BCMA using an expression cloning strategy designed to identify molecules that activate the NF-κB signaling pathway. Our subsequent analysis of BCMA revealed a novel function for this receptor in triggering Ag presentation in B cells that is dependent on NF-κB and JNK signaling. This ability to induce Ag presentation is specific to BCMA and cannot be triggered by signals generated by either TACI or BAFF-R.
Materials and Methods
Expression cloning of NF-κB-activating molecules
The NF-κB reporter cell line is a derivative of the EBV-encoded nuclear Ag 293EBNA cell line (Invitrogen Life Technologies) that was cotransfected with a NF-κB-GFP reporter plasmid and an expression construct encoding two antiapoptotic molecules CrmA and Bcl-xL. Stable selection of the plasmids was achieved by resistance against hygromycin B, which was encoded on the antiapoptotic plasmid. The activated human T cell cDNA library in the Ef1α promoter-driven vector pEAK8 was obtained from Edge BioSystems. The library was divided into pools each of ∼500 clones, cultured in Terrific broth (TB) media, and miniprep DNA generated from the cultures using Qiagen. Reporter cells were seeded onto 24-well plates (75,000 cells/500 μl/well) and transfected the next day with 0.4 μg of the miniprep DNA using calcium phosphate. At 48 h after transfection, plates were analyzed by fluorescent microscopy for GFP-positive wells. The initial TB cultures from the pools that gave rise to GFP-positive wells were titered and divided into subpools of ∼50 clones. Transfections and GFP screens were repeated with the miniprep DNA from the subpool cultures. Cultures from positive subpools were plated on Luria-Bertani agar and subpools of five clones each screened in the same manner. In the last step, individual clones from the final GFP-positive subpools were tested and DNA sequencing conducted on the positive clones.
Full-length murine BCMA and TACI were amplified by Pfu Turbo DNA polymerase (Stratagene) from IMAGE (clone nos. 1349977 and 3411532, respectively; American Type Culture Collection (ATCC)). The murine BCMA-ΔC83 mutant was also generated by PCR amplification using the full-length BCMA as the template and replacing codon 104 with a stop codon. All three receptors were cloned in-frame downstream of a CD5 leader and hemagglutinin (HA) epitope. To generate chimeric receptors, the extracellular and transmembrane domain of human TNFR2 (residues 1–287) was amplified from a TNFR2 plasmid (33) and cloned in-frame to the intracellular domains of murine BCMA (residues 74–185), murine TACI (residues 150–249), and murine BAFFR (residues 98–175) using standard PCR methodology. The primers used are: mBCMA, forward GGAATTCACGCGTAAGATGAACCCCGAGGCCCTG, reverse GGGATCCGCGGCCGCTTATCTAG-TGTGAGTTGGCTT; mTACI, forward GGAATTCACGCGTAGGCGTAGAGGAGAGCCACTA, reverse GGGATCCGCGGCCGCTCAAGTTGCCGGACGAGCGTC; and mBAFFR, forward GGAATTCACGCGTAGGTGGCGTCAACAGCTCAGG, reverse GGGATCCGCGGCCGCCTATTGCTCTGGGCCAGCTGT. PCR products are flanked with Mlu1 (5′ end) and NotI (3′ end) sites for subcloning downstream of human TNFR2. For the negative control, the intracellular domain was replaced by an in-frame stop codon. The intracellular domain of murine BAFF-R was amplified from IMAGE (clone no. 5547292; ATCC). All constructs were sequenced and verified by automated sequencing (PE Applied Biosystems). All receptors were cloned into a previously described Moloney murine leukemia virus-based retroviral vector upstream of an internal ribosome entry site (IRES)-puromycin resistance cassette (34). The dominant negative JNK kinase (JNKK)2 or MAPK kinase (MKK)7 (JNKK2/MKK7) is a lysine to methionine substitution at residue 149 of human JNKK2, which has been previously described to behave in a dominant negative manner (35). Mutagenesis was performed using the Quikchange site-directed mutagenesis kit (Stratagene) and verified by sequencing. The I-κB super-repressor (I-κBSR) with the serine to alanine substitutions at residues 32 and 36 of I-κBα was subcloned from a previously described plasmid pEAK8-IκBα double negative (36). The JNKK2 (K149M) mutant, I-κBSR, and the GST irrelevant control were cloned into the retroviral vector upstream of an IRES-GFP cassette. Infection of A20 B cells with vesicular stomatitis virus G protein-pseudotyped retrovirus was conducted as described (34). All experiments were conducted with bulk populations of puromycin-resistant or GFP-positive A20 cells to avoid clonal variability.
The murine A20 B cell line was obtained from ATCC and maintained in DMEM (Invitrogen Life Technologies) supplemented with 10% FBS (HyClone Laboratories), 50 μM 2-ME, 2 mM l-glutamine, and 15 μg/ml gentamicin. Stable transfectants were kept in the presence of 1 μg/ml puromycin. 293EBNA and its derivatives were cultured in DMEM with 10% bovine calf serum and 15 μg/ml gentamcin. The DO11.10 T cell hybridoma was a generous gift of Dr. J. Zhang (Kimmel Cancer Center, Philadelphia, PA) and was cultured in RPMI 1640 (Cellgro) supplemented with 10% FBS, 50 μM 2-ME, and penicillin-streptomycin. Mouse B cells were magnetically isolated from spleen using the B cell isolation kit purchased from Miltenyi Biotec by depletion of CD43+, CD4+, and Ter-119+ cells. The purity of the isolated B cells was ∼98% B220-positive as analyzed by flow cytometry. Mouse spleens were obtained from 4- to 6-wk-old female BALB/c mice purchased from Charles River Breeding Laboratories. The mice were maintained in pathogen-free facilities in accordance with the guidelines of the Institutional Animal Care and Use Committee of Mount Sinai School of Medicine.
The OVA 323–339 peptide was synthesized by Peptides International. Murine IL-4, IL-6, APRIL, and human BAFF were purchased from PeproTech. Abs used were from the following sources: TNFR2 (AB-226-PB) and BCMA (AF593 and MAB593; R&D Systems); I-κBα mAb (Imgenex); HA mAb (clone 3F10; Roche Applied Science); phospho-JNK, JNK, phospho-p38, p38, phospho-ERK, and ERK (Cell Signaling Technology).
Cells were gently lysed in ice-cold buffer containing 1% Nonidet P-40, 20 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 30 mM NaF supplemented with protease inhibitors. Protein concentrations in the extract were determined using the bicinchoninic acid protein assay kit (Pierce) and equivalent amounts of protein were resolved on SDS-PAGE gels. Western blotting was performed using standard methodologies.
A total of 5 × 105 A20 B cells were stained with 1 μg of FITC-conjugated Ab in 100 μl of complete medium in the presence of Fc block in accordance with manufacturer’s instructions (BD Pharmingen). Stained cells were analyzed on a FACScan or FACSCalibur flow cytometer using the CellQuest software (BD Biosciences). FITC- and biotin-conjugated anti-CD40, anti-CD80/B7-1, anti-CD86/B7-2, anti-MHC class II, anti-CD54/ICAM-1 Abs, and streptavidin-allophycocyanin were purchased from eBioscience. Allophycocyanin-conjugated goat anti-rat IgG was obtained from BD Pharmingen.
Real-time PCR analysis
Primary B cells were treated as indicated and harvested for total RNA isolation by using RNA STAT-60 (Iso-Tex). First-strand cDNA was synthesized by using ThermoScript RT-PCR System (Invitrogen Life Technologies). The primers used for real-time PCR analysis of the APRIL/BAFF receptors are the same as those described earlier for cloning of their intracellular domains. Primers for GAPDH controls are forward, CTCATGACCACAGTCCATGCCATC and reverse, CTGCTTCACCACCTTCTTGAT-GTC. Quantitative real-time PCR was performed using the LightCycler instrument and FastStart DNA Master SYBR Green I kit according to manufacturer’s instructions (Roche). Data analysis was performed using the LightCycler Software version 3.5.
A total of 5 × 106 A20 B cells were stimulated with APRIL (100 ng/ml) or anti-human TNFR2 (10 μg/ml) for varying lengths of time. Nuclear extracts were prepared, and NF-κB EMSA was performed as previously described (37).
IL-2 secretion assays
A total of 2 × 105 A20 B cells in 0.4 ml were treated with APRIL (0–100 ng/ml) or BAFF (0–100 ng/ml) or anti-human TNFR2 (0–10 μg/ml) for 24 h. The next day, 1 × 106 DO11.10 T cells in 0.1 ml were added to the B cells with varying concentrations of the OVA peptide for another 24 h. For primary B cells, 5 × 106 B cells in 0.4 ml were stimulated with the different combinations of cytokines for 48 h before the addition of 1 × 106 DO11.10 T cells and OVA peptide for another 24 h. Culture supernatants were harvested and assayed in triplicate using IL-2 ELISA kits (eBioscience).
Identification of BCMA as an NF-κB-activating molecule by functional cloning
As the NF-κB signaling pathway controls numerous aspects of the immune system, we reasoned that an unbiased search for molecules that activate this signaling pathway would likely lead to the identification of molecules that critically regulate biochemical signaling events and downstream immunological responses. To do so, we established an expression cloning approach that allowed us to rapidly identify molecules with such characteristics. This approach was based on the observation that high-level expression of receptors and pathway-specific signaling molecules is often sufficient to activate the NF-κB pathway in a ligand-independent manner. We first generated a NF-κB reporter cell line by stable transfection of 293EBNA cells with a construct consisting of multimerized NF-κB promoter elements upstream of GFP, together with a plasmid that expresses CrmA and Bcl-xL. The latter antiapoptotic genes served to avoid the loss of cDNAs that encode genes with direct proapoptotic effects themselves, or nonapoptotic cDNAs that are accompanied by cDNAs with proapoptotic effects. An activated human T cell cDNA expression was then screened by sib selection using the NF-κB reporter cell line. The complexity of the cDNA library was first reduced by division into multiple pools of ∼500 cDNA clones per pool (Fig. 1,A). DNA obtained from these pools were transfected into the NF-κB reporter cells and examined by fluorescence microscopy for GFP-positive cells. Positive pools were subdivided into ever-smaller size pools and the screening process repeated until eventually a single cDNA clone capable of activating NF-κB was isolated (Fig. 1,A). A number of molecules previously known to activate NF-κB, as well as novel genes, were isolated in the screen. Among the former were several receptors including BCMA, FAS, DR3, lymphotoxin β receptor, and TLR4. The different stages during the sib selection process that led to the cloning of BCMA were shown (Fig. 1 B). The initial pool where BCMA was present at a frequency of 1/500 produced a fluorescence pattern in which only a few cells were green. The intensity of the fluorescent signal increased as the pool size decreased, reflecting the increased frequency of the target cDNA. The cloning of BCMA as a NF-κB-activating protein is consistent with previous reports showing that BCMA induces the NF-κB signaling pathway (38, 39). Nonetheless, the biological responses regulated by BCMA have remained unclear and we therefore focused our subsequent analyses on the characterization of this receptor.
BCMA induces Ag presentation in A20 B cells
Because BCMA is expressed predominantly on B cells (28), we explored the function of BCMA by testing the effects of expressing and activating this receptor in various B cell models. Our initial RT-PCR analysis of the murine A20 B cell line indicated that this cell line has no detectable expression of endogenous BCMA and BAFFR, but TACI expression was detected (Fig. 2,A). The lack of BAFF-R transcripts in our A20 cell line is a consistent finding and is somewhat surprising because many other B cell lines have been shown to express BAFF-R (23). The fact that the RT-PCR analysis conducted in the same experiment detected BAFF-R expression in the positive control splenic B cells but not in the A20 cell lines indicated that this observation is accurate and is not due to an experimental error. We subsequently stably expressed full-length murine BCMA in A20 B cells using retroviruses. Bulk populations of puromycin-resistant cells were used in all our experiments to avoid the effects of clonal variation. As controls, we also expressed a signaling-defective BCMA mutant with a truncation of the C-terminal 83 residues and a full-length murine TACI. All constructs were epitope-tagged with HA and Western blotting indicated that these receptors are expressed, albeit full-length BCMA and TACI are expressed at lower levels than the signaling-defective BCMA-ΔC83 mutant (Fig. 2,B). This difference in expression level is likely due to the fact that high level expression of wild-type BCMA and TACI leads to cell death (our unpublished observation) and is not tolerated by the cells. We also tested whether these receptors were functional by stimulating with their ligand APRIL and examining the activation of the canonical NF-κB pathway. APRIL (100 ng/ml) did not induce detectable I-κBα degradation in untransfected A20 B cells or in cells expressing the signaling defective BCMA-ΔC83 mutant. In contrast, APRIL stimulation of cells expressing either full-length BCMA or TACI led to the time-dependent degradation of I-κBα followed by its re-expression (Fig. 2,C), indicating that the two full-length APRIL receptors are functional. The functionality of the full-length BCMA and TACI receptors were further confirmed by examining I-κBα phosphorylation using a phospho-I-κBα mAb (our unpublished observation), as well as by examining nuclear translocation of NF-κB transcription factors using EMSA with a NF-κB-binding probe previously described (Fig. 2 D) (37).
To obtain a better understanding of the role of BCMA, we examined the effects of APRIL stimulation on the expression of several immunologically relevant surface proteins on B cells by flow cytometry. Overnight stimulation of BCMA-expressing cells with 100 ng/ml APRIL significantly increased the expression of CD40, CD80/B7-1, CD86/B7-2, MHC class II, and CD54/ICAM-1 (Fig. 3,A). This APRIL-mediated increase in expression was not observed in untransfected A20 cells (our unpublished observation) or in A20 cells expressing the signaling-defective BCMA-ΔC83 mutant (Fig. 3,A). The surface molecules that were up-regulated are known to be critical for Ag presentation and we therefore hypothesized that this immunologically important process is regulated by BCMA in B cells. To test this hypothesis further, we investigated whether BCMA ligation has any effect on IL-2 production by T cells in the presence of Ag. A20 cells (H-2d haplotype) expressing the different receptors were stimulated with APRIL for 24 h and then cocultured with DO11.10 T cell hybridomas in the presence of varying concentrations of the OVA 323–339 peptide. Culture supernatants were harvested after an additional 24 h and IL-2 levels quantified by ELISA. APRIL stimulation of BCMA-expressing B cells dramatically induces IL-2 production by T cells that is dose-dependent on Ag concentration (Fig. 3,B). Little IL-2 production occurred if the BCMA-expressing cells were not stimulated with APRIL. In contrast, APRIL stimulation of untransfected A20 cells, or those expressing the BCMA-ΔC83 mutant, did not have any effect on IL-2 production by the T cells (Fig. 3,B). This last observation is consistent with the fact that APRIL failed to induce the expression of Ag presenting molecules in those two cell lines (Fig. 3 A). The absence of any IL-2 production by the T cells when cocultured with the untransfected A20 cells and similarly treated with APRIL also shows that APRIL itself does not have a direct effect on IL-2 production by the T cells. Furthermore, direct treatment of the DO11.10 T cells with APRIL also does not induce IL-2 secretion (our unpublished observation).
In addition to APRIL, BAFF also binds BCMA albeit at a lower affinity (29, 30). Hence we examined whether BAFF has the same effect as APRIL. Stimulation of BCMA-expressing A20 cells with BAFF also induced Ag presentation to T cells in a dose-dependent manner (Fig. 3,C). A comparison of the dose responses to BAFF and APRIL revealed that APRIL is more potent at activating Ag presentation (Fig. 3 C). On a molar basis, the effect of APRIL is even greater as the molecular mass of APRIL (22 kD) is larger than BAFF (17 kD). This observation is consistent with the reported higher affinity of APRIL for BCMA (29, 30). The results thus far demonstrate that BCMA ligation induces B cells to become more efficient APCs.
Aside from BCMA, both APRIL and BAFF can bind to TACI, another member of the TNFR superfamily expressed on B cells (2). We therefore asked whether TACI ligation has any effect on Ag presentation by B cells. A20 cells stably transfected with either the full-length BCMA or TACI were stimulated with 100 ng/ml APRIL and cocultured with the DO11.10 T cells in the presence of varying concentrations of the OVA peptide (Fig. 3,D). Although APRIL stimulation of the BCMA-expressing cells led to IL-2 production by the T cells, identical stimulation of the TACI-expressing cells did not result in IL-2 production. Likewise, APRIL stimulation of TACI-expressing cells also did not induce the surface expression of CD40, CD80, MHC class II and ICAM-1 (Fig. 3,A). This differential effect is not due to differences in BCMA and TACI expression. Both receptors are equivalently expressed as assayed by anti-HA Western blotting and both can induce signaling when stimulated by ligand (Fig. 2). Hence, TACI does not have the same effect as BCMA on Ag presentation in B cells.
BCMA ligation induces Ag presentation in splenic B cells
We next examined whether BCMA ligation has any effect on Ag presentation in primary B cells. The expression profile of BCMA, TACI, and BAFF-R in resting B cells isolated by MACS beads from BALB/c splenocytes was examined by PCR analysis. Little BCMA expression was detectable in resting B cells, whereas both TACI and BAFF-R were detected (Fig. 2,A). The lack of BCMA expression in peripheral naive murine B cells observed in this study is similar to previous analyses in both mouse and human B cells (21, 40, 41). Because BCMA expression appeared to increase as B cells differentiate to plasma cells (28), we evaluated the effects of B cell differentiation cytokines on BCMA expression by real-time PCR analysis. Treatment of splenic B cells with IL-6 alone had no effect on BCMA expression, whereas IL-4 alone resulted in a modest increase in BCMA expression (Fig. 4,A). However, IL-4 and IL-6 synergize to significantly induce BCMA expression. To confirm the results from the PCR analysis, we examined the expression of BCMA on splenic B cells by flow cytometry using anti-BCMA mAb. Consistent with the PCR analysis, stimulation of splenic B cells with a combination of IL-4 and IL-6 resulted in increased surface expression of BCMA (Fig. 4 B). In contrast to BCMA, parallel samples analyzed for TACI and BAFF-R expression indicated that these two receptors remained unchanged by these cytokines.
To investigate whether BCMA expression on splenic B cells induced by IL-4 and IL-6 can lead to responsiveness to APRIL and enhanced Ag presentation, the surface expression of costimulatory molecules treated with a combination of IL-4, IL-6, and APRIL was compared with cells treated with IL-4 and IL-6 only (Fig. 4,C). Similar to that observed previously in the A20 cell line, APRIL stimulation increased the expression of CD40, CD80, CD86, and MHC class II on splenic B cells. However, CD54 was not induced by APRIL in splenic B cells, indicating a minor difference between splenic B cells and the A20 B cell line. To examine Ag presentation to T cells, B cells stimulated with the various combinations of IL-4, IL-6, and APRIL were incubated with DO11.10 T cells in the presence of OVA peptide. APRIL stimulation alone had very little effect on Ag presentation in splenic B cells (our unpublished observation) most likely because BCMA expression in resting B cells is very low. Although IL-4 and IL-6 stimulation resulted in increased BCMA expression (Fig. 4,A), this activity had only minor effects on IL-2 production by T cells in the absence of APRIL (Fig. 4,D). However, upon addition of APRIL to IL-4 and IL-6, significant enhancement of IL-2 production by T cells was observed at two different Ag concentrations tested (Fig. 4 D). Therefore under conditions in which BCMA expression is induced, i.e., after IL-4 and IL-6 treatment, APRIL is now able to stimulate B cells to present Ag.
The previous experiment strongly suggests that stimulation of BCMA by APRIL induces splenic B cells to present Ag to T cells. Because APRIL can bind to both BCMA and TACI, the possibility remains that the effect of APRIL on Ag presentation is not mediated through its interaction with BCMA but rather via binding to TACI, even though the results from transfected A20 cells indicated this effect to be unlikely. To formally demonstrate that BCMA expressed on splenic B cells is indeed able to induce Ag presentation, we tested the effect of stimulating BCMA with agonist Abs. Polyclonal goat Abs directed against murine BCMA were first tested for specificity and agonistic activity using BCMA- and TACI-transfected A20 B cell lines. The BCMA Abs were able to activate NF-κB signaling in A20 cells expressing BCMA, but not in cells expressing TACI (our unpublished observation). Similarly, these Abs were also able to stimulate A20 B cells expressing BCMA, but not those expressing TACI, to present Ag to T cells (Fig. 4,E). Thus, these polyclonal Abs are specific for BCMA and function as agonist. Splenic B cells stimulated with IL-4 and IL-6 in the presence of either APRIL or anti-BCMA were cocultured with DO11.10 T cells in the presence of OVA peptide and IL-2 production assayed. Similar to the experiment shown in Fig. 4,D, B cells stimulated with IL-4 and IL-6 did not present Ag to T cells, whereas those stimulated with IL-4, IL-6, and APRIL induced T cells to produce IL-2 (Fig. 4,F). More importantly, anti-BCMA had the same effect as APRIL and was able to activate Ag presentation by B cells in the presence of IL-4 and IL-6 (Fig. 4 F). These observations demonstrate that specific ligation of BCMA on primary B cells by agonist Abs activates Ag presentation by B cells. Thus, BCMA ligation induces Ag presentation in both the A20 cell line and primary B cells.
Differential effect of APRIL/BAFF receptors on Ag presentation
The experiments shown earlier in Fig. 3 indicated that BCMA but not TACI induces Ag presentation in B cells. To investigate this differential effect more thoroughly and to also examine the third TNFR-related receptor BAFF-R, a chimeric receptor approach was used. A20 B cells were stably transfected with chimeric receptors consisting of the ectodomain of human TNFR2 fused to the intracellular signaling domains of murine BCMA, TACI, and BAFF-R. As a negative control, the TNFR2 ectodomain was fused in frame to a stop codon. The chimeric receptor approach allowed us to normalize for receptor expression level and “ligand-receptor” interaction as all the receptors can be stimulated by the same ligand, i.e., agonist Abs specific for human TNFR2. Hence a fair comparison of the effects of triggering signaling from the three APRIL/BAFF receptors on downstream biological responses can be conducted. All receptors are equivalently expressed on A20 cells (our unpublished observation) and those containing the intracellular do-main of BCMA, TACI, and BAFF-R are functional as cross-link-ing with anti-TNFR2 led to the activation of the NF-κB pathway (see Fig. 6 A), similar to that previously described for ectodysplasin anhidrotic receptor chimeric receptors (10).
The four A20 cell lines expressing the different chimeric receptors were stimulated with anti-TNFR2 for 24 h and the surface expression of CD40, CD80, CD86, MHC class II, and ICAM-1 was analyzed by flow cytometry (Fig. 5,A). Similar to our earlier observation with full-length BCMA, cross-linking of the TNFR2-BCMA chimeric receptor induced the expression of the Ag presentation molecules. This effect of receptor cross-linking was not observed in A20 cells expressing the negative control TNFR2-EC, TNFR2-TACI, or TNFR2-BAFFR chimeras (Fig. 5,A). This observation suggests that only BCMA is capable of activating signals leading to Ag presentation in B cells. To test this further, we next stimulated the different A20 cell lines with different amounts of anti-TNFR2 for 24 h, followed by the addition of DO11.10 T cells and 80 ng/ml OVA peptide for another 24 h. In line with the observations in Fig. 5,A, anti-TNFR2 cross-linking of the B cells expressing TNFR2-BCMA resulted in a dose-dependent production of IL-2 by the T cells (Fig. 5 B). This receptor-dependent presentation of Ag to T cells was not observed in B cells expressing the negative control TNFR2-EC, or TNFR2-TACI or TNFR2-BAFFR. These chimeric receptor studies demonstrate that among the three APRIL/BAFF receptors, only BCMA is capable of generating the appropriate signals to trigger Ag presentation in B cells.
BCMA-mediated Ag presentation is NF-κB and JNK-dependent
The specific ability of BCMA to trigger Ag presentation suggests that the signaling pathways triggered by BCMA are likely to be different from that of TACI and BAFF-R. Although cross-linking of TNFR2-BCMA, TNFR2-TACI, and TNFR2-BAFFR all activated the canonical NF-κB signaling pathway as assayed by I-κBα degradation and NF-κB gel shift analysis (Fig. 6,A), only the TNFR2-BCMA chimera induced Ag presentation in B cells (Fig. 5). Therefore, activation of the NF-κB pathway does not appear to be sufficient for Ag presentation and it is very likely that BCMA triggers an additional signaling pathway required for this response that is not shared with TACI or BAFF-R. To identify this BCMA-specific signaling pathway, we examine the ability of the different TNFR2 chimeric receptors to activate MAPK signaling pathways. Among the three chimeric BAFF/APRIL receptors, only cross-linking of the TNFR2-BCMA receptor resulted in activation of the JNK signaling pathway as evidenced by increased phosphorylation of JNK (Fig. 6,B). In contrast to the differential effect on JNK signaling, all three chimeric BAFF/APRIL receptors weakly induced the phosphorylation of p38 (Fig. 6,B). ERK phosphorylation was induced by ligation of TNFR2-TACI but was not significantly induced by ligation of TNFR2-BCMA or TNFR2-BAFFR (Fig. 6 B). Thus, the MAPK signaling pathway that correlated with BCMA-mediated Ag presentation is the JNK signaling pathway.
To confirm that the effect of the TNFR2-BCMA receptor on JNK signaling is also reflected in wild-type BCMA, A20 B cells expressing full-length BCMA were stimulated with APRIL for different lengths of time and analyzed by Western blotting with Abs specific for phospho-JNK. Similar to the TNFR2 chimeric receptor, ligation of full-length BCMA also resulted in JNK activation whereas analogous stimulation of untransfected A20 B cells, or cells expressing full-length TACI or BCMA-ΔC83 did not result in JNK signaling (Fig. 6 C). These observations are consistent with earlier studies showing that overexpression of BCMA in HEK 293 cells induced JNK activation (38) and strongly suggests that the ability to activate JNK signaling confers on BCMA the specific capability to induce Ag presentation.
The expression of CD40, CD80, CD86, MHC class II, and ICAM-1 are known to be regulated by NF-κB (42, 43, 44, 45). Furthermore, multiple AP-1 enhancer sites are present in the promoter regions of CD40 and CD80 (46), suggesting that the JNK signaling pathway is also critical for this process. We therefore examined the requirement for NF-κB and JNK signaling in BCMA-mediated Ag presentation. The BCMA-expressing A20 cells were transduced with retroviruses encoding either an irrelevant control protein or the I-κBSR mutant to inhibit NF-κB signaling. Both coding sequences are located upstream of an IRES-GFP cassette and transduced cells that were GFP-positive were sorted by flow cytometry. Sorted cells were left untreated or treated with APRIL for 24 h followed by staining with biotinylated Abs and streptavidin-allophycocyanin. As shown in Fig. 6,D, transfection of the I-κBSR mutant but not the control protein inhibited APRIL-induced up-regulation of CD40, CD80, CD86, MHC class II, and ICAM-1. These B cell lines were also stimulated with varying concentrations of APRIL for 24 h before coculturing with DO11.10 T cells in the presence of 80 ng/ml OVA peptide. Culture supernatants were subsequently harvested for IL-2 ELISA. Consistent with the effect on the surface molecules, the I-κBSR also inhibited the ability of APRIL to induce BCMA-expressing B cells to present Ag to T cells (Fig. 6 E). These observations demonstrate that BCMA-induced Ag presentation in B cells is dependent on the NF-κB signaling pathway.
To examine whether BCMA-induced JNK signaling has a role in Ag presentation in B cells, the effect of blocking JNK activation using a dominant negative JNKK2/MKK7 was tested using the strategy described earlier for the I-κBSR mutant. BCMA-expressing A20 cells were infected with retroviruses encoding a kinase inactive JNKK2 (K149M) mutant upstream of an IRES-GFP cassette and infected cells were sorted based on GFP expression. Expression of the JNKK2KM dominant negative effectively inhibited APRIL-induced expression of CD40, CD80, CD86, and ICAM-1 in the BCMA-expressing B cells (Fig. 6,D) indicating that the JNK signaling pathway is critical for the induction of these surface molecules. The induction of MHC class II expression did not appear to be significantly affected by the blockade of the JNK pathway. The blockade of BCMA-mediated up-regulation of CD40, CD80, CD86, and ICAM-1 by the JNKK2KM dominant negative also inhibited the ability of BCMA to induce B cells to present Ag to T cells (Fig. 6 E). These observations strongly suggest that the JNK signaling pathway is critical for Ag presentation induced by BCMA in B cells and is the likely explanation for the specific capability of BCMA, but not TACI or BAFF-R, to trigger this response.
The study on BCMA reported here initiated from efforts to identify signaling molecules that activate the NF-κB pathway using an expression cloning strategy. This strategy based on the high-level expression of signaling molecules to autonomously activate the NF-κB pathway led to the identification of both uncharacterized NF-κB-activating molecules, as well as those previously known to do so. The latter observation validated the use of this strategy to identify novel molecules involved in the regulation of the NF-κB pathway. Use of the NF-κB-GFP reporter cell line greatly simplified the screening process allowing rapid identification of molecules that activate NF-κB. This approach provides an advantage over similar strategies that required the use of enzymatic assays such as luciferase activity (47). Coupled with the use of a powerful expression vector driven by the EF1α promoter, this enabled the size of the initial pools to be in the range of 500-1000 cDNA per pool. A number of TNFR superfamily members were identified in the screen conducted including BCMA, a receptor for APRIL and BAFF. The two other known receptors for BAFF, TACI and BAFFR, were not identified in the initial screen. It is possible that the mRNAs encoding TACI and BAFFR may be present at a low frequency in the T cell library and to date, only a small fraction of this library has been screened. It is likely that cDNA libraries generated from different cell types will give rise to a different array of NF-κB-activating genes and one may be more likely to find clones encoding TACI and BAFF-R in a B cell library. Although the focus to date has been on the NF-κB pathway, the principles on which this expression cloning technology is based can be adapted to other signaling pathways in which pathway-specific reporters can be developed.
Although it is known that BCMA binds to APRIL and BAFF, its function has remained obscure. This present study has revealed a novel function for BCMA in activating Ag presentation in B cells. Ligation of BCMA resulted in the up-regulation of surface molecules critical for Ag presentation such as CD40, CD80, CD86, MHC class II, and ICAM-1/CD54. This process is dependent on the NF-κB signaling pathway triggered by BCMA. One of the most intriguing aspects of this study is the observation that among the three APRIL/BAFF receptors, only BCMA is capable of generating signals that lead to Ag presentation. This finding is despite the observation that the intracellular domains of BCMA, TACI, and BAFFR are all capable of activating the canonical NF-κB pathway. Therefore, there must exist significant differences in signaling pathways triggered by the three receptors, which then account for the different biological responses elicited. Because members of the TNFR superfamily including BCMA are known to activate MAPK signaling pathways (38), a comparison of the three APRIL/BAFF receptors in their ability to activate MAPK signaling was conducted. The JNK signaling pathway was found to be specifically activated by BCMA but not by TACI or BAFF-R, suggesting that the ability to trigger the JNK pathway confers on BCMA the specificity to activate Ag presentation. Indeed, blockade of the JNK pathway inhibited BCMA-mediated Ag presentation indicating that this pathway is critical for this biological response. Both the NF-κB and JNK signaling pathways are thus essential for BCMA to induce Ag presentation in B cells.
The effect of BCMA ligation on Ag presentation was also observed in primary B cells. Resting murine splenic B cells do not appear to express significant level of BCMA. Upon stimulation with IL-4 and IL-6, BCMA expression on B cells is up-regulated rendering these cells responsive to APRIL and subsequently to APRIL-mediated Ag presentation to T cells. We attribute this effect of APRIL to its activation of BCMA based on several observations. First, because APRIL does not bind BAFF-R (23), this effect is not mediated by BAFF-R. Second, APRIL stimulation alone has little effect on Ag presentation by resting B cells even though TACI is expressed on resting B cells. Third, APRIL was only able to elicit a response when B cells were also stimulated with cytokines (i.e., IL-4 and IL-6) that induce BCMA expression. Fourth, APRIL stimulation of TACI-transfected A20 B cells had no effect whereas it had an effect in BCMA-transfected cells. Fifth, agonistic Abs against BCMA were able to induce primary B cells to present Ag to T cells. Taken together, these results show that ligation of BCMA on B cells leads to enhanced Ag presentation and indirectly to T cell activation. A previous report showed that in vivo administration of APRIL resulted in T cell activation, most likely as a result of direct APRIL binding to TACI expressed on T cells (14). It is also possible that APRIL could activate T cells indirectly through BCMA-mediated induction of stimulatory molecules on B cells. The observation that primary B cells that are stimulated by cytokines to express BCMA can present Ag to T cells in response to APRIL raises the question whether cells that constitutively express BCMA, such as plasma cells and transitional T1 B cells (28, 32, 41), are able to do so in response to APRIL. Furthermore, it is not known whether all B cells can respond to APRIL in this manner or whether certain subsets of B cells such as marginal zone or B1 cells may respond differently. These questions will be the subject of future studies.
The functional characterization of BCMA adds this receptor to the list of signals that are known to induce Ag presentation in B cells, including TLRs and CD40. Because CD40 and BCMA are both members of the TNFR superfamily, it is likely that the two receptors share similar mechanisms to activate Ag presentation. For instance, both receptors interact with TNFR-associated factor adapters and can activate the NF-κB and JNK signaling pathway (38, 39, 48, 49, 50, 51, 52). Future studies will be aimed at dissecting the proximal signaling mechanism that is used specifically by BCMA and that may be shared with CD40 but not with TACI or BAFFR to activate Ag presentation. It is noteworthy that BCMA ligation also increases the expression of CD40. Because CD40 ligation also activates NF-κB and JNK signaling, which in turn can induce CD80, CD86, MHC class II, and ICAM-1 expression, BCMA and CD40 may synergize to increase Ag presentation to T cells. Both APRIL and BAFF are up-regulated in activated T cells (53) and thus during B cell-T cell interactions, APRIL or BAFF can synergize with CD40L to enhance Ag presentation by B cells. In addition, APRIL and BAFF expression can be induced on B cells following CD40 stimulation (18), raising the possibility that CD40 may also provide a feed-forward effect to BCMA further enhancing Ag presentation. Because both BCMA and CD40 can induce Ag presentation, this function of BCMA is likely to be dispensable in B cells and can be compensated for by CD40. This function may partly explain why the bcma−/− mice displayed no overt defect in immune responses (5, 31). Another reason why this Ag presentation function of BCMA is dispensable in the context of the organism is the redundancy provided by the other APCs, i.e., dendritic cells and macrophages.
The discovery that BCMA can induce Ag presentation may have wider clinical implications for APRIL and BAFF. The present study suggests that APRIL and BAFF could potentially be used as an adjuvant to activate BCMA to increase the immunogenicity of vaccines, perhaps by enhancing the humoral response. In a similar vein, BCMA may play a role in the pathogenesis of autoimmune disorders such as SLE, in which elevated levels of BAFF have been reported (11, 12). One mechanism by which elevated BAFF can contribute to the development of autoimmunity is believed to be due to BAFF-R-mediated prolonged survival of autoreactive B cells (10, 54). The observation that BCMA can trigger Ag presentation suggests the additional possibility that elevated BAFF levels, through the activation of BCMA on B cells with autoreactive specificities, may enhance the presentation of self-Ags to autoreactive T cells. This enhancement is likely to contribute to the breakdown of tolerance to self-Ags.
We thank Lloyd Mayer for the OVA 323–339 peptide, Tom Moran and Sergio Lira for providing mice, and Jianke Zhang for the DO11.10 T cell hybridoma.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This study was supported in part by an Extended Fellowship (to M.Y.) and by a research grant (to A.T.T.) from the Systemic Lupus Erythematosus Foundation.
Abbreviations used in this paper: BAFF, B cell-activating factor of the TNF family; BCMA, B cell maturation Ag; TACI, transmembrane activator and calcium-modulator and cyclophilin ligand interactor; SLE, systemic lupus erythematosus; APRIL, a proliferation-inducing ligand; HA, hemagglutinin; I-κBSR, I-κB super-repressor; IRES, internal ribosome entry site; JNKK, JNK kinase; MKK, MAPK kinase.