|Home | About | Journals | Submit | Contact Us | Français|
BAFF-R (B cell-activating factor belonging to the tumor necrosis factor family-receptor) regulates B lymphocyte survival, maturation, homeostasis, and self tolerance through signaling mechanisms that are not completely understood. A spontaneous BAFF-R mutation, Bcmd-1, disrupts BAFF-R signaling. However, it is not clear why the Bcmd-1-encoded BAFF-R fails to adequately support B cell survival, optimal CD21/35 expression, and B-cell tolerance to dsDNA, since it is 95% identical to the wild-type (wt) BAFF-R and retains the only known signaling motif. A retrotransposon insertion in A/WySnJ strain mice generated the Bcmd-1 allele, replacing the eight C-terminal BAFF-R residues with 21 retrotransposon-encoded residues. New data reported here show that the displaced residues, previously thought to have no signaling role, are essential for optimal CD21/35 expression but contribute little to B cell survival signaling. Analysis of wt Baffr or Bcmd-1 homozygous (A/WySnJ X B6.BCL2)F2 mice confirmed that BCL2 complemented Bcmd-1 for B cell survival but not CD21/35 expression. Through in vivo retroviral transduction experiments, we show that Baffr complemented Bcmd-1 for B cell survival but not CD21/35 expression, whereas the BaffrΔ103-175 deletion mutant lacking the BAFF-R cytoplasmic domain failed to support these functions. Importantly, we show that the BaffrΔ168-175 deletion mutant lacking the retrotransposon-displaced residues, and a BaffrT170A mutant lacking a critical threonine, supported B cell survival but failed to support optimal CD21/35 expression. These data provide the first evidence for a possible bifurcation at the receptor level in the BAFF-R signaling pathway. We suggest that discrete BAFF-R cytoplasmic domains may interact with distinct downstream pathways to provide fine control over B cell survival, maturation, and tolerance induction.
Peripheral B lymphocytes are consumed in the defense against microbial invasion, and replenished through bone marrow B cell hematopoiesis. Although much is known about peripheral B lymphocyte survival, homeostasis, maturation, and self-tolerance induction, questions remain as to the precise signaling mechanisms that control these integrated processes. The cytokine BAFF (B-cell-activating factor belonging to the tumor necrosis factor family), through BAFF-receptor (BAFF-R) signaling, provides indispensable support for transitional B lymphocytes as they migrate from the bone marrow to the spleen and mature into fully immuno-competent peripheral B lymphocytes (Mackay et al., 2007). The evidence supporting this conclusion derives in part from the severe B lymphopenia observed in mice with targeted mutations in either BAFF (Gorelik et al., 2004; Schiemann et al., 2001) or BAFF-R (Sasaki et al., 2004; Schiemann et al., 2001; Shulga-Morskaya et al., 2004; Thompson et al., 2001; von Bulow et al., 2001; Yan et al., 2001). Signals transduced by BAFF-R also enhance B cell expression of the Cr2 gene encoding CD21/35, completing antigen-independent B cell maturation (Amanna et al., 2003; Rahman and Manser, 2004; Sasaki et al., 2004; Shulga-Morskaya et al., 2004; Tardivel et al., 2004). This is a significant step in B cell maturation, because proper CD19/CD21 complex expression is required for B cell selection into the marginal zone (MZ) pool (Cariappa et al., 2001), and for coupling the innate immune recognition of microbial antigens to B cell activation through the complement system (Fearon and Carroll, 2000).
BAFF binding to BAFF-R propagates an intracellular signal primarily via non-canonical nuclear factor-kappaB2 (NF-κB2) pathway activation (Claudio et al., 2002; Kayagaki et al., 2002; Siebenlist et al., 2005), leading to enhanced expression of the Bcl-2 and Bcl-xL survival genes (Batten et al., 2000). The BAFF-R cytoplasmic domain motif PVPAT (murine aa 154-158, human aa 162-166) and murine Thr167 (human Thr175) are required for binding of tumor necrosis factor receptor-associated factor 3 (TRAF3) to BAFF-R (Ni et al., 2004; Xu and Shu, 2002). TRAF3 may act as a negative regulator of non-canonical NF-κB2 by targeting the vital kinase of the pathway, NF-κB-inducing kinase (NIK), for degradation. Binding of TRAF3 to BAFF-R is thought to sequester TRAF3 so NIK can accumulate and activate the NF-κB2 pathway (Gardam et al., 2008). Thus, BAFF is required for peripheral B cell homeostasis because it activates the NF-κB2 pathway leading to survival gene expression.
In contrast to survival, the molecular details that explain how BAFF enhances CD21/35 expression are unknown. Dependence of Cr2 gene expression on NF-κB2 pathway activation has not been reported (Martin, 2007). BAFF might regulate CD21/35 expression and survival independently, since a brief blockade of BAFF in vivo reduced CD21/35 expression without reducing B cell survival (Gorelik et al., 2004). Long-term blockade of BAFF through TACI-Ig-transgene expression reduced B cell survival, CD21/35 expression, and MZ B cell development (Tardivel et al., 2004). However, introducing the BCL2 transgene corrected the survival defect, but not the CD21/35 or MZ B cell defects (Tardivel et al., 2004). These BAFF blockade studies do not address which of the three BAFF receptors, BAFF-R, transmembrane activator and CAML interactor (TACI), or B cell maturation antigen (BCMA), were involved in controlling survival and CD21/35 (Mackay et al., 2007). Our data suggested that the BAFF-R might control both traits, because BAFF-R-mutant A/WySnJ mice had suboptimal B cell survival, CD21/35 expression, and MZ B cell development (Amanna et al., 2003). Enforced bcl-xL gene expression in A/WySnJ B cells restored B cell longevity and follicular B cell development, but did not correct the CD21/35 and MZ B cell defects (Amanna et al., 2003). Similar findings were reported for BCL2 transgene expression in A/WySnJ mice (Rahman and Manser, 2004). Together, these data support a role for the BAFF-R in B cell survival and CD21/35 control.
The interpretation of the A/WySnJ data is confounded by the complex nature of the BAFF-R mutation. A spontaneous retrotransposon insertion into the Baffr gene generated the Bcmd-1 (B cell maturation defect-1) allele (Amanna et al., 2003). The insertion replaced the eight C-terminal amino acyl (aa) residues 168-175 of the BAFF-R with 21 retrotransposon-encoded residues (Thompson et al., 2001; Yan et al., 2001). Although these eight residues are part of a 25-aa region that is completely conserved between human and mouse (Thompson et al., 2001), they were disordered in electron density maps and are considered to have no signaling role (Ni et al., 2004). The Bcmd-1-encoded BAFF-R is 95% identical to wt BAFF-R at the amino acid level, is correctly expressed in B cell development, and binds BAFF (Shulga-Morskaya et al., 2004; Thompson et al., 2001).
The Bcmd-1 mutation decreased peripheral B cell survival causing B lymphopenia (Amanna et al., 2001; Miller and Hayes, 1991), diminished the relative maturity of the surviving B cells (Amanna et al., 2003; Rahman and Manser, 2004), and reduced the T-dependent IgG responses (Miller et al., 1992) through an adverse effect on the germinal center reaction (Rahman et al., 2003). Although Bcmd-1 was widely considered to be a complete loss-of-function allele (Kayagaki et al., 2002; Mackay et al., 2005), we documented phenotypic and functional differences between the Bcmd-1-mutant A/WySnJ mice and congenic AW.Baffr−/−mice that must be due to retention of partial function and an absence of dominant-negative effects (Mayne et al., 2009). The most striking difference was development of a systemic lupus erythematosus-like syndrome in the Bcmd-1-mutant A/WySnJ mice, but not the congenic AW.Baffr−/− mice (Mayne et al., 2008).
The present research sought to define the signaling capabilities of the BAFF-R cytoplasmic domain, particularly the C-terminal 8 aa residues, to better understand why the Bcmd-1-expressed mutant receptor failed to support CD21/35 expression (Amanna et al., 2003), and drove a lupus-like autoimmune syndrome in A/WySnJ mice (Mayne et al., 2008). Since B lymphopenic and CD21/35-deficient A/WySnJ mice lack aa 168-175, we hypothesized that the C-terminal 8 aa residues control CD21/35 expression and B cell survival. We tested this hypothesis by performing complementation studies in A/WySnJ mice. Wild-type and mutated Baffr genes were retrovirally transduced into A/WySnJ HSC, and the transduced HSC-derived B lymphocytes were compared to Bcmd-1 mutant B lymphocytes within A/WySnJ host mice. Our results show unequivocally that CD21/35 control resides within the C-terminal 8 aa domain, whereas this region is largely dispensible for B cell survival.
Male and female A/J, A/WySnJ, and (A/WySnJ X BCL2)F2 mice were produced in our specific pathogen-free mouse colony in the Department of Biochemistry, University of Wisconsin-Madison. The mice were maintained at 23°C with 40–60% humidity and 12 h light-dark cycles. Mice were age 7–10 wks. The protocols were approved by the Institutional Animal Care and Use Committee. A/WySnJ mice were bred to C57BL/6-TgN(BCL2)22Wehi mice (The Jackson Laboratory, Bar Harbor, ME), which carry a BCL2 transgene expressed in B lineage cells (Strasser et al., 1991). F1 mice were intercrossed to produce (A/WySnJ X BCL2)F2 mice. DNA isolated from tail sections was genotyped for the presence of the BCL2 transgene and Baffr genotype through PCR as previously described (Mayne et al., 2008; Strasser et al., 1991).
Flow cytometric analysis was performed as previously reported (Mayne et al., 2008). In short, duplicate samples (106 cells/sample) of RBC-depleted splenocytes were stained with titered amounts of mAb conjugates. Stained samples were analyzed on a FACScalibur™ using CELLQuest™ software (BD Biosciences, Franklin Lakes, NJ). The streptavidin-APC and streptavidin-PercCp-Cy-5.5 secondary labels, the PE- and biotin-coupled rat mAb to mouse CD23 (Clone B3B4), and the FITC-coupled rat mAb to mouse CD21/35 (Clone 7G6) were obtained from BD Biosciences. The APC-goat polyclonal Ab to mouse IgM, and the FITC-, PE-, APC- and biotin-rat IgG2B and IgG2A isotype control mAb were obtained from Caltag Laboratories (Burlingame, CA). PE-coupled rat mAb to mouse CD21/35 was a generous gift from Dr. John Kearney (University of Alabama, Birmingham, AL).
The wild-type Baffr gene was PCR amplified from an A/J mouse splenic cDNA library. Deep Vent DNA polymerase (New England Biolabs, Beverly, MA) was used to amplify basepairs 8 to 589 of the BAFF-R mRNA sequence (GenBank ID 15208476; forward primer 5′-GCCCAGACTCGGAACTGT-3′ and reverse primer 5′-AGCCTCCACTGCTGCTATTG-3′). This fragment was cloned into the pCRII-TOPO vector using the TOPO TA PCR Cloning Kit for Sequencing (Invitrogen, Carlsbad, CA). Plasmid DNA was isolated and sequenced at the UW-Madison Biotechnology Center’s DNA Sequence Laboratory. Inserts from error-free clones were excised with EcoRI and subcloned into the EcoRI site of the MigRI vector’s multiple cloning site (Pear et al., 1998). Site-directed mutations of Baffr were created using PCR and a Quick-Change PCR kit (Stratagene).
The various MigRI retroviral particles were produced and A/WySnJ hematopoietic stem cells (HSC) were transduced as previously described (Amanna et al., 2003; Cariappa et al., 2001). Briefly, purified vector DNA (Wizard Plus Maxipreps; Promega, Madison, WI) was CaPO4 precipitated onto 80–90% confluent monolayers of the Bosc23 or Phoenix-Eco packaging cells grown in IMDM supplemented with 10% heat-inactivated FBS (Hyclone, Logan, UT), 100 U/ml penicillin/streptomycin, and 2 mmol/L L-glutamine. Retroviral supernatants were collected 48 to 72 hrs later, stored at −70 °C, and titered on NIH/3T3 cells.
To produce chimeric mice, the A/WySnJ donors were injected intravenously with 5-fluorouracil (5 mg in 400 μl PBS; Sigma-Aldrich, St. Louis, MO). Five days later, bone marrow cells were collected and cultured (10–20×106 cells) in IMDM (4 mL) supplemented with 15% heat-inactivated FBS, 5% WEHI-3B conditioned medium, 100 U/mL penicillin/streptomycin, 2 mmol/L L-glutamine, recombinant mouse IL-3 (6 ng/mL), recombinant mouse IL-6 (10 ng/mL), and recombinant mouse stem cell factor (100ng/mL; all from Peprotech, Rocky Hill, NJ). After 24 h of culture, the medium was removed, and two sequential retroviral transductions separated by 24 h were performed by inoculation and centrifugation as described (Pear et al., 1998). The transduced cells were collected 4 h after the second spinoculation, resuspended in PBS, and 0.5×106 to 1×106 cells were injected i.v. into each lethally irradiated A/WySnJ mouse. Lethal irradiation (1100 Rads in two doses, separated by 4 h) was done just prior to bone marrow cell transfer. Splenocytes were collected and analyzed 11–12 wks later.
The Bcmd-1 mutation consists of a truncation and an addition of 21 retrotransposon-encoded residues as illustrated in Fig. 1A. The mutant BAFF-R signaling defects might be due to the added residues, including a stretch of Arg, His, and Lys residues that could disrupt protein interactions through charge-charge repulsion. Alternatively, the 8 truncated residues might harbor an unidentified signaling motif.
As a first step toward determining why the Bcmd-1 mutant receptor failed to support CD21/35 expression, we confirmed that BAFF-R-mutant A/WySnJ B cells expressed less CD21/35 than B cells from the closely-related A/J strain. IgM-stained splenic B cells were gated on the FO B cell population and evaluated for CD21/35 expression (Fig. 2). As expected, the mature wt B cells had a CD21/35 mean fluorescence index (MFI) that was 118±38 units higher than the Bcmd-1 B cells. We also confirmed that complementing Bcmd-1 with a B cell survival signal was sufficient to restore B cell development, but not sufficient to restore optimal CD21/35 expression (Amanna et al., 2003; Rahman and Manser, 2004; Tardivel et al., 2004). Mice from an (A/WySnJ X BCL2)F2 intercross were genotyped and analyzed for CD21/35 expression. In the absence of the BCL2 transgene, the B cell subset proportions were similar between homozygous Baffr mice and A/J mice, and between homozygous Bcmd-1 mice and A/WySnJ mice, as expected (Fig. 3A, Table I). In the presence of the BCL2 transgene, the homozygous Bcmd-1 mice had increased FO B cells, decreased NF B cells, and decreased MZ B cells, as reported (Rahman and Manser, 2004). Also as reported, the BCL2 transgene decreased the MZ B cells, did not change the FO or NF B cells in the homozygous Baffr mice. The homozygous Baffr B cells had a 2-fold higher CD21/35 MFI than homozygous Bcmd-1 B cells (Fig. 3B, Table I).
We next extended our analysis of FO Bcmd-1 B cells expressing the bcl-xL survival gene from a MigRI retroviral vector in AW.bcl-xL mice (Amanna et al., 2003). We previously reported that the retrovirally transduced bcl-xL gene complemented Bcmd-1 with respect to FO B cell development. Here, we found that bcl-xL failed to complement Bcmd-1 with respect to optimal CD21/35 expression, since the FO B cells from AW.bcl-xL mice had significantly lower CD21/35 expression than the FO B cells from A/J mice, whether they were transduced with the MigRI virus carrying the bcl-xL survival gene or the empty vector MigRI control virus (data not shown). These data show that restoring Bcmd-1 B cell development through BCL2 or bcl-xL survival gene expression was not sufficient to enable optimal CD21/35 expression.
To define the signaling capabilities of BAFF-R cytoplasmic domains more precisely in vivo, we used the MigRI retrovirus system to transduce Baffr genes with defined mutations into A/WySnJ HSC, transplanted the HSC into lethally-irradiated A/WySnJ recipients, and evaluated B lymphocyte development and CD21/35 expression. Using A/WySnJ mice as recipients for these experiments raises a concern about interference from the endogenous mutant BAFF-R proteins in the retrovirally-transduced B cells. We attempted to use B6.Baffr-ko mice lacking endogenous BAFF-R proteins as recipients (Sasaki et al., 2004), but the combination of extremely poor MigRI transduction efficiencies, coupled with the virtual absence of B lymphocytes in the mice, made this approach untenable. Since we had previously shown that Bcmd-1 is a hypomorphic partial loss-of-function allele and does not act in dominant-negative fashion (Mayne et al., 2009), we used A/WySnJ recipients where retroviral transduction efficiency was typically ~5–20% and each recipient had some B cells for analysis. This approach allowed us to evaluate each mutant Baffr construct for complementation of Bcmd-1, because recipient mice had B cells that were a mixture cells expressing only Bcmd-1 and cells expressing Bcmd-1 plus the desired Baffr mutation.
The MigRI retroviral vector used in these experiments expresses the gene of interest on a bicistronic mRNA containing an internal ribosomal entry site (IRES) and encoding green fluorescent protein (GFP) as a marker of successful gene transfer (Fig. 1B) (Pear et al., 1998). We first created retroviral vectors representing the full-length transcript (MigRI.Baffr), and a naturally occurring mRNA splice variant (MigRI.BaffrΔ118-128), a full cytoplasmic domain truncation mutant (MigRI.BaffrΔ103-175), and an empty MigRI vector to use as a negative control (Fig 1C). Each of these vectors was used to transduce A/WySnJ HSC, which were then transferred into irradiated recipient A/WySnJ mice. After 11–12 wk of hematopoietic cell reconstitution, the splenocytes were immunostained and analyzed by flow cytometery for IgM, GFP, and BAFF-R protein expression. The AW.Baffr, AW.BaffrΔ118-128, and AW.BaffrΔ103-175 splenic B cells all showed a robust correlation between increasing GFP and increasing BAFF-R immunofluorescence, indicating that these vectors encoded a BAFF-R protein that was correctly expressed on the B-cell surface (data not shown).
The CD21/23 gating scheme was then used to analyze the IgM+GFP+ and IgM+GFP− B cell populations (Oliver et al., 1999). For all retroviral constructs tested, B cell phenotypes were confirmed using the IgM/IgD (Cariappa et al., 2001), and IgM/C1qRp/CD23 (Allman et al., 2001) gating schemes (data not shown). The data from the CD21/23 gating scheme are shown to highlight changes in CD21/35 expression (note that a CD21-PE antibody was used for Table II in contrast to a CD21-FITC antibody for Table I, thus absolute MFI values between the tables differ). There were no significant differences in the B cell subset percentages or the FO B cell CD21/35 MFI between GFP+ and GFP− B cells from the recipients of MigRI-transduced HSC, indicating that expression of the empty retroviral vector had no impact on B cell development (Fig. 4, Table II). However, there was a 58% decline in the NF B cell percentage, a 1.5-fold increase in the FO B cell percentage, and a 1.8-fold increase in FO B cell CD21/35 MFI between GFP+ and GFP− B cells in the recipients of HSC transduced with the MigRI.Baffr vector (Fig. 4, Table II). Similar results were obtained with the MigRI.BaffrΔ118-128 vector (Table II). These changes were not observed with the MigRI.BaffrΔ103-175 vector (Fig. 5, Table II), which encoded a mutant protein lacking a cytoplasmic domain, even though this protein attained cell surface expression and retained the regions necessary to bind BAFF. Collectively these data establish that retroviral-mediated transfer of the Baffr gene (and the splice variant) complemented Bcmd-1 for B cell development and CD21/35 expression, and further, that the BAFF-R residues 103-175 were essential for these survival and maturation signals.
To test the hypothesis that the 8 truncated aa residues might have an essential role in BAFF-R signaling, we produced AW.BaffrΔ168-175 mice using retroviral gene transduction as before (Fig. 1C), and analyzed their splenocytes by flow cytometery. The close correlation between increasing GFP expression and increasing BAFF-R protein expression in B cells from these mice again verified that vector-encoded BAFF-R proteins were expressed on the B-cell surface (data not shown). Surprisingly, transduction of the BaffrΔ168-175 mutant yielded a 35% decline in the NF B cell percentage and a 1.2-fold increase in the FO B cell percentage, but no change whatsoever in FO B cell CD21/35 MFI when GFP+ B cells were compared to GFP− B cells (Fig 5, Table II). This result was unexpected, because a previous report did not implicate this region in TRAF3 binding (Ni et al., 2004). We conclude that aa 168-175 have an essential role in BAFF-R-mediated signaling for optimal CD21/35 expression, although these residues appear to contribute little to B cell survival signaling.
Because previous experiments did not implicate BAFF-R residues 168-175 in TRAF3 binding (Ni et al., 2004; Xu and Shu, 2002), and TRAF3 was the only adapter protein involved in BAFF-R survival signaling (Claudio et al., 2002; Kayagaki et al., 2002), we reasoned that the displaced BAFF-R residues 168-175 might harbor a previously unidentified signaling motif for CD21/35 upregulation. We noted that this conserved BAFF-R region shares some homology with the CD40 receptor (Amanna et al., 2003), another member of the tumor necrosis factor receptor (TNFR) family that is critical for B cell antigen responses. Among the homologous residues were several Thr, known to be common phosphorylation targets, and a Pro, known to be important in protein folding. Accordingly, to look for an unidentified signaling motif, we elected to mutate Thr170 and Pro173, and evaluate the mutants for possible disruption of BAFF-R-mediated signaling. We hypothesized that one or both of these point mutants would fail to complement Bcmd-1 for optimal CD21/35 expression.
The MigRI.BaffrT170A and MigRI.BaffrP173A retroviral vectors were generated through site-directed mutagenesis, AW.BaffrT170A and AW.BaffrP173A mice were created, and flow cytometry was performed on GFP+ and GFP− splenic B cells as described above. The robust correlation between increasing GFP fluorescence and increasing BAFF-R protein expression indicated proper surface expression of these mutant BAFF-R proteins on splenic B cells (data not shown). The BaffrT170A and BaffrP173A mutants yielded 31% and 45% declines, respectively, in the NF B cell percentages, and 1.3-fold increases in the FO B cell percentages (Fig. 6, Table II). The BaffrP173A mutant also supported a highly significant 1.7-fold increase in FO B cell CD21/35 intensity, similar to the increase seen with wt Baffr transduction. Contrastingly, the BaffrT170A mutant supported a very weak 1.2-fold increase in CD21/35 intensity (Fig. 6, Table II). Thus, both point mutants supported B cell survival, but only the BaffrP173A mutant fully supported optimal CD21/35 expression. We interpreted these data as suggesting that the displaced BAFF-R residues 168-175 harbor a previously unidentified signaling motif for CD21/35 upregulation, and suggest that Thr170 is part of the motif.
The effects of all the expressed mutations on B cell survival and CD21/35 expression are summarized in Fig. 7. For this summary, we first calculated the difference between the GFP+ and GFP− cell populations within each mouse, and subsequently calculated the mean values and S.D. shown in Fig. 7. This allows a direct comparison to be made between transduced and untransduced cells within the same experimental animal. Overall, our data suggest that the BAFF-R signal bifurcation occurs at the level of the receptor, with aa 103-167 signaling primarily B cell survival and aa 168-175, particularly Thr170, signaling CD21/35 upregulation.
The new data presented here provide the first direct evidence that the BAFF-R signal transduction pathway may be bifurcated at the receptor level to regulate B cell survival and CD21/35 expression independently. Analysis of B cell development and CD21/35 expression in Baffr or Bcmd-1 homozygous (A/WySnJ X B6.BCL2)F2 mice and in AW.bcl-xL mice confirmed that the survival gene expression complemented the signaling-impaired Bcmd-1 allele with respect to B cell survival, but not CD21/35 expression (Amanna et al., 2003; Rahman and Manser, 2004; Tardivel et al., 2004). These results are consistent with a bifurcation in the BAFF-R pathway, but the complexity of the Bcmd-1 mutation and the fact that BCL2 transgene expression altered FO and MZ B cell development in normal mice confounds this interpretation.
We extended the analysis by performing retroviral gene transduction studies in Bcmd-1 mice. This approach allowed us to test specific mutations in vivo, and to compare transduced with untransduced cells within individual animals. Our in vivo experiments established that the survival and CD21/35 enhancement signals emanated from the BAFF-R cytoplasmic domain, since the wt Baffr gene and the natural BaffrΔ118-128 splice variant supported B cell development and optimal CD21/35 expression, whereas the BaffrΔ103-175 deletion mutant did not. The BaffrΔ168-175 variant, which lacks the residues that were displaced by the retrotransposon in the Bcmd-1 allele, failed to support optimal CD21/35 expression. This was surprising since the deleted residues were previously thought to have no signaling role (Ni et al., 2004). The BaffrT170A mutant similarly failed to support optimal CD21/35 expression. Thus, our data establish that the C-terminal 8 residues of the BAFF-R, particularly Thr170, contribute significantly to optimal CD21/35 expression, although they appear to be less important for survival signaling.
Survival signals emanating from the BAFF-R are transmitted by the non-canonical NF-κB2 pathway (Claudio et al., 2002; Kayagaki et al., 2002). This pathway begins with NF-κB-inducing kinase (NIK) and leads ultimately to enhanced transcription of the Bcl-2 and Bcl-xL pro-survival genes (Mackay et al., 2007). BAFF binding causes BAFF-R monomers to undergo oligomerization and bind TRAF3 (Claudio et al., 2002; Kayagaki et al., 2002). TRAF3 binding to the BAFF-R cytoplasmic domain prevents TRAF3-mediated NIK degradation, allowing NIK to activate non-canonical NF-κB2 (Liao et al., 2004; Sasaki et al., 2008). Previous studies reported TRAF3 binding to BAFF-R via a PVPAT motif (murine aa 154-158, human aa 162-166) (Ni et al., 2004), so it was not surprising that the BaffrΔ168-175 variant supported B cell survival. However, our data present a paradox. The Bcmd-1- and BaffrΔ168-175-encoded receptors retain the PVPAT motif, but only the BaffrΔ168-175-encoded receptor signaled B cell survival. The explanation for this paradox may relate to the fact that the partial loss-of-function protein expressed by Bcmd-1 contains a region of 8 retrotransposon-encoded amino acids (RPHIRRHK) with a high positive charge density at the C-terminus (Amanna et al., 2003). We speculate that the Bcmd-1-encoded protein signals survival poorly because this positively-charged region of the retrotransposon-encoded tail impedes receptor oligomerization or TRAF3 binding.
An important question remaining to be answered is how the C-terminal 8 residues of the BAFF-R, particularly murine Thr 170, contribute to optimal CD21/35 expression. Our retroviral Baffr gene transfer data show unequivocally that optimal CD21/35 expression is under BAFF-R-mediated control, confirming Baffr gene ablation data from others (Sasaki et al., 2008; Sasaki et al., 2004). Recent research has shown that TRAF2 and TRAF3 cooperate in BAFF-R-mediated signaling, with TRAF2 signaling degradation of TRAF3, eliminating TRAF3-mediated suppression of NIK (Gardam et al., 2008). In those studies, B cell-specific deletion of either TRAF2 or TRAF3 resulted in BAFF-R-independent survival and maturation of B lymphocytes, including enhancement of CD21/35 expression. Thus, from the perspective of non-canonical NF-κB2 signaling, the B cell survival and CD21/35 maturation phenotypes appear linked. It is possible that BAFF-R residues 168-175 are required for full NF-κB2 pathway activation, and that suboptimal levels of NF-κB2 signaling in the BaffrΔ168-175 mutant support survival but not optimal CD21/35 expression. This single pathway hypothesis is consistent with our data comparing the wt Baffr to the BaffrΔ168-175 mutant with respect to B cell development.
An alternative possibility is that the BAFF-R-derived signals for optimal CD21/35 expression depend on a second pathway that utilizes aa 168-175. The finding that NF-κB proteins are located at the CD21/35 (Cr2) gene promoter of the CD21-deficient A/WySnJ mouse suggests that non-NF-κB signals may be necessary for full CD21/35 expression (Debnath et al., 2007). In this context, it is intriguing that a motif search of the BAFF-R cytoplasmic domain identified the TTKT sequence (murine aa 167-170, human aa 175-178) within the deleted region as a potential target site for protein kinase C (PKC)-mediated phosphorylation. It is possible that phosphorylation of the TTKT sequence, possibly by PKC, may be needed for optimal CD21/35 protein expression. Interestingly, PKC-beta interacts with components of the NF-κB pathway, influencing NF-κB signaling (Guo et al., 2004). It is also noteworthy that B cell receptor signal transduction activates PKC-delta, and that PKC-delta is specifically required for transitional B cell tolerance induction (Guo et al., 2004), because the Bcmd-1-encoded BAFF-R lacks the TTKT motif and drives the development of a lupus-like systemic autoimmune syndrome in A/WySnJ mice (Mayne et al., 2008; Mayne et al., 2009). The TTKT motif may also play a role in BAFF-R binding to TRAF3, since murine BAFF-R Thr167 (human Thr175) was necessary for this binding (Ni et al., 2004). We suggest that the BAFF-R cytoplasmic TTKT sequence may be a target site for PKC-delta or PKC-beta-mediated phosphorylation, and further, that such phosphorylation may be part of the cross-talk between the B cell receptor and the BAFF-R receptor that determines transitional B cell tolerance induction via fine control of B cell survival.
BAFF-R provides the predominant survival signal to B cells in the periphery and also drives expression of the B cell co-receptor CD21/35 on these cells. This receptor thus promotes B cell survival for maintenance of the B lymphocyte pool, and also provides essential signals for coupling the complement system of the innate immune response to B cell activation (Cariappa et al., 2001; Fearon and Carroll, 2000). Our new data show that the survival signals and CD21/35 upregulation signals clearly derive from the cytoplasmic domain of the BAFF-R, but may emanate from distinct regions. Moreover, our new data raise the possibility that a previously unrecognized signaling motif resides in the C-terminal 8 residues that were previously thought to be structurally and functionally insignificant (Ni et al., 2004). We hypothesize that murine residues 167-170, TTKT, constitute a potential PKC phosphorylation site that is linked to BAFF-R signaling for CD21/35 maturation, and may also be linked to the systemic autoimmune phenotypes we observed in A/WySnJ mice (Mayne et al., 2008; Mayne et al., 2009). Further experiments are underway to evaluate the phosphorylation status of the TTKT motif and the TRAF binding capabilities of wild-type and mutant BAFF-R proteins in the expectation that the results will provide insights into how B cell signaling pathways are integrated to exert fine control over B cell survival and tolerance induction.
We thank Drs. John Kearney, Antonius Rolink, and Warren Pear for reagents, and Dr. Gail Bishop for providing critical comments on the manuscript.
1This work was supported by an NIH Predoctoral Training Grant in Genetics (5 T32 GM07133) and the Lupus Foundation of America Gina Finzi Memorial Student Fellowship to C.G.M and National Institutes of Health Grant T32GM07215 and the University of Wisconsin, Madison, Peterson Predoctoral Fellowship to IJA.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.