B cell-activating factor (BAFF) and a proliferation-inducing ligand (APRIL) play key roles in peripheral B cell survival, maturation, and differentiation. BAFF and APRIL are produced by a variety of cell types such as macrophages/monocytes and dendritic cells. Our analysis shows that BAFF mRNA is also expressed in all B cell subsets isolated from bone marrow, spleen, and peritoneal cavity of BALB/c mice. APRIL expression is restricted to early stages of B cell development in the bone marrow and the peritoneal B1 subset. Stimulation of B2 and B1 cells with LPS or CpG-oligodeoxynucleotides induced MyD88-dependent plasma cell differentiation and intracellular expression of BAFF and APRIL. Furthermore, activation of B cells up-regulated membrane expression of BAFF. The finding that in vitro activation of B cells is inhibited by the antagonist transmembrane activator and calcium modulator ligand interactor Ig, indicates that BAFF and/or APRIL are released into the culture supernatants. It shows that B cell survival, proliferation, and differentiation are supported by an autocrine pathway. In vivo activation of B cells with a T-dependent Ag- induced BAFF expression in germinal center B cells. In (NZB × NZW)F1 mice with established autoimmune disease, marginal zone, germinal center B cells, as well as splenic plasma cells expressed high levels of BAFF. In (NZB × NZW)F1 mice, the continuous activation of B cells and thus overexpression of BAFF and APRIL may contribute to the development of autoimmune disease.
Two closely related cytokines of the TNF superfamily, B cell-activating factor (BAFF,3 also termed BLyS, TALL-1, zTNF-4, THANK, and TNF13B) and a proliferation-inducing ligand (APRIL), are central players in B cell development and homeostasis. BAFF-deficient mice have an almost complete loss of follicular (FO) and MZ B cells, although there is normal development of early B cells in the bone marrow (1, 2, 3, 4). Normal numbers of newly formed immature B cells leave the bone marrow and develop into the transitional T1 stage. In the absence of BAFF. B cells do not progress past the transitional T2 stage, resulting in impaired humoral immune responses (5). However, when BAFF is overexpressed, for example in BAFF- transgenic mice, self-reactive B cells may be rescued from peripheral deletion (6, 7). As a consequence, BAFF-transgenic mice develop a lupus-like autoimmune disease (8, 9, 10, 11). From these data, it is apparent that the level of BAFF has to be tightly regulated to ensure B cell survival on the one hand and to prevent autoimmunity on the other.
BAFF is produced by a number of different cell types. Expression of BAFF by follicular dendritic cells (FDC) may be essential for B cell homeostasis (12, 13, 14). However, the main sources of BAFF are monocytes, macrophages, dendritic cells, and neutrophils, although subpopulations of B and T cells have also been shown to express it (15, 16, 17, 18, 19, 20, 21). BAFF is found to be membrane associated and its expression is enhanced by cytokines, such as IFN-α, IFN-γ, and IL-10 or growth factors (17, 19, 20, 22).
Much less is known about APRIL which has no essential function in normal B cell development and plays only a minor role in B cell homeostasis (23). In immune responses APRIL acts as a costimulator for B and T cell proliferation and supports class switch (2, 4, 22).
BAFF binds to three separate receptors, the BAFF receptor (BAFF-R, BR3), the transmembrane activator and calcium modulator ligand interactor (TACI), and the B cell maturation Ag (24, 25, 26, 27). APRIL binds only to TACI and B cell maturation Ag. All three receptors are expressed on B cells, although their expression level changes with B cell maturation.
The analysis of human B cell tumor lines, such as B cell chronic lymphatic leukemia (B-CLL), multiple myeloma, and Hodgkin’s lymphoma cells, suggested that BAFF and APRIL support tumor survival by an autocrine pathway (21, 28, 29, 30). There is also evidence that normal nonmalignant human B cells up-regulate BAFF and APRIL upon activation (21, 28). However, it is thought that murine B cells do not express BAFF or APRIL (3), although an analysis of early murine B cell development suggested that B cell survival in the bone marrow may be supported by an autocrine pathway (31).
In this study, we show a detailed analysis of BAFF and APRIL expression in different B cell subsets isolated from the bone marrow, spleen, and peritoneal cavity. In vitro cultures demonstrate that upon stimulation with LPS or CpG-oligodeoxynucleotides (CpG-ODN), splenic B2 cells as well as peritoneal B1 cells up-regulate BAFF and APRIL expression. MyD88-deficient mice demonstrated that BAFF expression in B cells is regulated by the TLR signaling pathway. Immunization with the T-dependent Ag 2-phenyl-oxazolone (phOx) induced BAFF expression in germinal center (GC) B cells. In (NZB × NZW)F1 mice, expression of BAFF and APRIL mRNA in MZ and B1 cells was strongly up-regulated with increasing age and onset of disease. In addition, strong expression of BAFF and APRIL protein was found in plasma cells. Thus, expression of these cytokines in activated and differentiated B cells may support the development of autoimmune disease.
Materials and Methods
Experiments were performed with BALB/c, C57BL/6, MyD88−/− (C57BL/6 background), and (NZB × NZW)F1 mice. One to 2-mo-old BALB/c mice were immunized i.p. with a single injection of 100 μg of alum-precipitated phOx coupled to the carrier protein chicken serum albumin. (NZB × NZW)F1 mice were purchased from The Jackson Laboratory and bred in the Bundesinstitut für gesundheitlichen Verbraucherschutz und Veterinärmedizin, Berlin-Marienfelde. Animal experiments were approved by the institutional animal care and use committee.
Abs and reagents
Surface expression of BAFF on B cells was determined using a FITC-labeled anti-BAFF mAb (Buffy-2; Alexis). For isotype control, FITC-labeled rat IgM (Southern Biotechnology Associates) was used. To detect BAFF and APRIL expression, cytospins and frozen sections were stained with polyclonal rabbit anti-BAFF (Sigma-Aldrich) or anti-APRIL (Stressgen) Ab, respectively. As secondary Ab Alexa Fluor 546-conjugated goat anti-rabbit IgG (Molecular Probes) was used. To control for the specificity of the BAFF-specific mAb Buffy-2, purified B cells were incubated for 2 h with 200 ng of soluble BAFF and then stained with the mAb Buffy-2 or with polyclonal rabbit anti-BAFF Abs. The expression of the BAFF-R was controlled by staining with the biotinylated rat mAb 204406 (R&D Systems).
The antagonist TACI Ig was prepared by fusing the extracellular domain of human TACI (aa 1–154) to the Fc region of the human IgG1 H chain (32). The protein was expressed in the HEK 293T cell line and isolated from supernatants by protein G chromatography.
FITC-, PE-, Cy5-, PE-Cy7-, or biotin-conjugated anti-B220 (RA3-6B2), biotinylated anti-CD11b (MI/70.15.11), anti-CD11c (N418) and anti-CD90 (T24), FITC and Cy5 anti-CD19 (ID3) and anti-CD21 (7G6), Cy5 anti-IgM (M14), and biotinylated anti-κ L chain (187.1) Abs were provided by the Deutsches Rheuma ForschungsZentrum (DRFZ). FITC or biotinylated anti-CD43 (1B11) was obtained from Biolegend, PE or biotinylated anti-CD5 from eBioscience, FITC or PE anti-CD23 (B3B4) and PE anti-CD138 (281–2) from BD, Pharmingen, and biotinylated and FITC-peanut agglutinin (PNA) from Vector laboratories. To visualize biotinylated Ab Alexa Fluor 488-, PE-, or allophycocyanin-conjugated streptavidin was used (Molecular Probes and BD Biosciences).
Isolation of lymphocytes
Suspensions of bone marrow cells were flushed from tibias and femurs of 6–10-wk-old BALB/c mice and stained with B220-PE-, CD43-FITC-, and IgM-Cy5-specific Abs. To isolate pro-B cells, lymphocytes were gated on IgM− cells and B220+CD43+ cells were sorted as shown in Fig. 1,A. To isolate pre- and immature B cells, lymphocytes were gated on B220+CD43− and the IgM-negative (pre-B) and -positive (immature B cell) fractions sorted. After cell sorting, cells were controlled for purity (Fig. 1 A).
Spleen cells were stained with biotinylated Ab specific for CD43, CD11b, CD11c, and CD90 to remove myeloid, stromal, and T cells. After incubation with antibiotin microbeads (Miltenyi Biotec), B cells were enriched by MACS (Miltenyi Biotec) to >96% purity. To isolate total splenic B cells for setting up in vitro tissue cultures, the enriched fraction was stained with anti-B220-Cy5 and B cells sorted to >99% purity (Fig. 2 A).
For the isolation of T1/T2, FO, and MZ B cells, the MACS-enriched B cell fraction was stained with FITC-conjugated anti-B220, Cy5, anti-CD21/35, and PE anti-CD23 (Fig. 1 A). The different B cell subsets, MZ (CD21highCD23−), FO (CD21intCD23+), transitional T1 (CD21−CD23−), and T2 (CD21highCD23+) B cells were sorted by FACS (BD Biosciences).
B1 cells were isolated from the peritoneal cavity and sorted as CD5+B220lowIgM+ cells (Fig. 1 A). Purity of sorted cells was in the range of 98–99%.
Cell culture and stimulation
Spleen cells of three animals were pooled and duplicate cultures (106 sorted splenic B cells/ml) set up in RPMI 1640 supplemented with 10% FCS, 50 μM 2-ME, 100 U/ml penicillin, and 100 μg/ml streptomycin. B cells were stimulated with 25 μg/ml LPS (Sigma-Aldrich) or 10 μg/ml CpG-ODN 1826 (InvivoGen) for 2–4 days. Expression of mRNA for BAFF and APRIL was measured by semiquantitative RT-PCR or real-time PCR. Surface expression of BAFF on B cells was determined by FACS using the monoclonal anti-BAFF (Buffy-2) Ab.
To test whether B cells express biologically active BAFF and/or APRIL, sorted B cells were activated for 3 days with different concentrations of LPS (1 and 5 μg/ml). Cultures were set up in the presence or in the absence of the TACI Ig (20 μg/culture) fusion protein. As a control protein, human IgG was added. To monitor proliferation, cells were labeled with the CSFE using a standard protocol. The percentage of cells in S-G2-M phase was determined by FACS analysis. Briefly, after 3 days of activation, B cells were harvested, washed with PBS, and fixed overnight in 70% ethanol at 4°C. Fixed cells were incubated for 30 min in 50 μg/ml propidium iodide at 37°C before analysis.
Furthermore, cells were cultured for 3 days with LPS or CpG-ODN in the absence or presence of 50 μM furin-like convertase inhibitor decanoyl-Arg-Val-Lys-Arg-chloromethylketone (CMK; Alexis). For these experiments, cell cultures were set up using sorted splenic B2 or peritoneal B1 cells isolated from individual BALB/c mice.
Detection of BAFF and APRIL mRNA
Expression of mRNA for BAFF and APRIL was measured by semiquantitative RT-PCR or real-time PCR. Total RNA was prepared from 2 × 106 sorted cells using a RNeasy Mini Kit (Qiagen). RNA was treated with RNase-free DNase I (Qiagen) according to the manufacturer’s instruction. Concentration of total RNA was measured by NaNoDrop (Biotech International) and reverse transcribed into cDNA using an Ominiscript RT kit (Qiagen). cDNA was amplified with BAFF-, APRIL-, or CD11c-specific primers using 1.25 U/reaction AmpliTaq Gold polymerase (Applied Biosystems). The number of cycles is indicated for the different experiments. For the β-actin control, cDNA was amplified for 30 cycles. Each cycle consisted of 1 min at 94°C, 45 s at 63°C, and 20 s at 72°C. Amplified cDNA was visualized on agarose gels and specificity was checked by sequencing.
To control the purity of sorted cells, defined numbers of MZ, FO, and peritoneal B1 cells were prepared as described and sorted into PCR tubes containing 20 μl of reverse transcriptase buffer (Qiagen). In addition, a second spleen was gently homogenized and the suspension was digested in RPMI 1640/10% FCS containing 1 μg/ml collagenase for 45 min at 37°C. After three washes, cells were incubated with Cy5-conjugated anti-CD11c mAb (N418) and defined numbers of CD11c+ cells were sorted again into PCR tubes. For each of the cell populations, cDNA was directly transcribed using 10 μM oligo(dT) as primer and one-fifth of the reaction mixture was used to amplify BAFF, APRIL, CD11c, or β-actin cDNA in a seminested PCR (Qiagen). The first amplification consisted of 25 cycles of 1 min at 94°C, 1 min at 63°C, and 30 s at 72°C and the second round of 40 cycles for APRIL, 35 cycles for BAFF, 30 cycles for CD11c, and 25 cycles for β-actin.
Quantitative PCR to determine the relative amount of BAFF mRNA was performed with a LightCycler System (Roche Diagnostics) using the LighCycler FastStart DNA Master SYBR Green I (Roche Diagnostics). The real-time PCR products were analyzed on 4% agarose gels to check purify and specificity. Each sample from three independent experiments was run in triplicate. The unit number showing relative mRNA level in each sample was determined as a value of mRNA normalized against β-actin.
Cells were stained in FACS buffer with directly labeled or biotinylated Ab. Unspecific staining was inhibited by blocking for 10 min at 4°C with rat IgG (Sigma-Aldrich) and the mAb 2.4G2/75 specific for the FcγR (CD16/32). For intracellular staining, surface-labeled cells were washed twice and fixed in 2% (w/v) paraformaldehyde (Merck) for 20 min at room temperature. After washing, cells were permeabilized in PBS supplemented with 0.5% saponin (Sigma-Aldrich) and 0.5% BSA for 10 min at room temperature and stained with BAFF or κ-specific Abs for 30 min in the dark. Before analyses, cells were washed with 0.5% saponin buffer. Stained cells were analyzed using FACSCalibur or LSRII (BD Biosciences) and the CellQuest (BD Biosciences) or FlowJo software (Tree Star).
Cytospin, fluorescence staining, and confocal microscopy
B cells activated in vitro for 3 days with different stimuli were harvested. To remove receptor-bound BAFF and APRIL, cells were incubated with 0.1% sodium citrate/0.1% Triton X-100 (pH 5.2) overnight at 4°C. After three washes with PBS/0.5% BSA, cells were centrifuged onto glass slides (Menzel). Spleens were frozen in Tissue-Tec OCT compound and stored at −70°C. Frozen tissue sections of 7 μm were prepared, fixed in cold acetone for 10 min, and air dried.
To detect BAFF and APRIL expression, cytospins and sections were double stained with polyclonal rabbit anti-BAFF or anti-APRIL Ab and anti-κ Ab. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole. To analyze BAFF expression in GC, tissue sections were double stained with polyclonal rabbit anti-BAFF and biotinylated FDC-specific mAb M2 (Serotec) or biotinylated PNA (Vector Laboratories). Fluorescent microscopy images were captured with a SPOT RT camera (Diagnostic Instruments) or by confocal microscopy using Leica software.
ELISA for soluble BAFF
Soluble BAFF in supernatants was detected using a mouse anti-BAFF ELISA kit (APO-54N-019-KI01; Apotech) according to the manufacturer’s instruction The sensitivity of the kit is 0.2 ng/ml. ODs at 450 nm were measured with a microplate spectrophotometer (Spectrarax 190). The murine A20 cell line was used as a positive control for BAFF release. After 3 days in culture, without activation, a concentration of 1.6 ± 0.02 ng/ml soluble BAFF was measured (see Fig. 5).
All statistical analyses were performed using the Student t test; statistical significance of difference was determined as p < 0.05 or p < 0.01 or p < 0.001.
B cells express APRIL and BAFF mRNA
Both B cell lymphoma lines and normal human B cells have been reported to express BAFF and APRIL (8, 21, 29, 30, 31). This prompted us to analyze whether normal murine B cells also express these cytokines. We found that during early B cell development in the bone marrow, pro-B cells (B220+CD43+IgM−), pre-B cells (B220+CD43−IgM−), and immature B cells (B220+CD43−IgM+) express APRIL mRNA (Fig. 1,B). When immature B cells leave the bone marrow and differentiate into transitional T1/T2 B cells, APRIL mRNA is down-regulated and is no longer detectable in FO (CD21intCD23+) and MZ (CD21highCD23−) B cells. Although no APRIL mRNA was detected in mature B2 cells, B1 cells isolated from the peritoneal cavity expressed it at high levels (Fig. 1 B).
Surprisingly, all murine B cell subsets analyzed expressed BAFF mRNA (Fig. 1,C). A comparative analysis showed that pro-, pre-, and immature B cells isolated from the bone marrow express 4- to10-fold higher levels of BAFF mRNA than T1/T2, mature FO and MZ B cells (Fig. 1,C). With the differentiation of the immature B cell into the transitional T1/T2 stage, BAFF mRNA is down-regulated. However, this is different for peritoneal B1 cells that expressed high levels of BAFF mRNA. The level was comparable to that seen in the immature B cells isolated from the bone marrow (Fig. 1 C).
It is unlikely that BAFF and APRIL mRNA are derived from a contamination by monocytes/macrophages or dendritic cells since the purity of the sorted B cells was between 98 and 99%. To exclude the possibility of a few contaminating myeloid cells, defined numbers of splenic FO, MZ B cells, peritoneal B1 cells, and CD11c+ cells were sorted by FACS and tested for BAFF, APRIL and CD11c mRNA expression using seminested PCR. Fig. 1,D shows that as few as 10 MZ B cells expressed sufficient BAFF mRNA to give a positive signal. The intensity of the band was further increased when 100 MZ B cells were sorted. Ten FO B cells showed only a very weak band. However, the intensity increased when 100 FO B cells were sorted. As described, FO and MZ B cells were negative for APRIL mRNA expression. In contrast, high levels of BAFF and APRIL mRNA were seen in peritoneal B1 cells. Independent of whether 10 or 100 FO, MZ, or peritoneal B1 B cells were sorted, no PCR signal for CD11c was detectable (Fig. 1,D). Since already 10 CD11+ cells express high levels of CD11c (Fig. 1 D), it is unlikely that BAFF mRNA expression in FO, MZ, and peritoneal B1 B cells is due to a contamination with CD11c myeloid cells.
High levels of BAFF mRNA expression is associated with protein expression on the surface of B cells
To determine whether murine B cells express BAFF, bone marrow cells were triple stained with B220, anti-IgM, and the BAFF-specific mAb Buffy-2. The subset of IgM− B220low pre- and pro-B cells showed little BAFF expression. The signal was only slightly higher than in the isotype control. Comparably weak expression levels were seen for the subset of IgM+B220low immature B cells (data not shown).
To control for the specificity of the mAb Buffy-2, sorted splenic B cells were incubated for 2 h with soluble BAFF (Fig. 1,E, upper row of histograms). Before and after incubation, B cells were stained with BAFF-R (Fig. 1,E, left panel) or BAFF-specific Abs (Fig. 1,E, middle and right panels). Receptor bound BAFF was only detectable with BAFF specific polyclonal rabbit Abs (Fig. 1,E, right panel). In addition, pre-incubation with BAFF slightly reduced the signal obtained with anti-BAFF-R Abs (Fig. 1,E, left panel). These data show that the mAb Buffy-2 does not recognize BAFF when it is bound by its receptors. In contrast to the polyclonal rabbit anti-BAFF Abs, the mAb Buffy-2 recognizes only membrane-expressed BAFF. The B cell lymphoma lines WEHI-231 and A20 were used as positive controls and clearly showed membrane-associated BAFF expression (Fig. 1 E, second row of histograms).
Splenic B cells were enriched by MACS to a purity >96% and stained with the BAFF-specific mAb Buffy-2. FACS analysis showed that newly formed (B220low) and the majority of mature splenic B cells were negative for BAFF surface expression (Fig. 1,E, third row of histograms). Only a small fraction (1.38 ± 0.3%) of splenic B cells expressed membrane-bound BAFF (data not shown). The low frequency suggested that BAFF-positive cells may be splenic B1 cells. To further analyze whether splenic B1 cells express membrane-bound BAFF, a triple staining with Ab specific for BAFF, CD5, and B220 was performed (Fig. 1,E). Gating on CD5+B220low cells confirmed that B1 cells express low levels of BAFF in their membrane, whereas no expression was seen on splenic mature and newly formed B2 cells (Fig. 1,E, third row). When peritoneal B1 cells were analyzed for BAFF surface expression, the signal was again marginally higher than in the isotype control (Fig. 1 E, last row), supporting that B1 cells express membrane-bound BAFF.
In vitro activation of B cells induces APRIL and BAFF expression
To analyze whether activation of B cells induces APRIL and BAFF protein expression, splenic B cells were sorted and cultured in vitro with different stimuli. RNA was isolated and cytokine expression was analyzed. After 3 days in culture, the purity of B cells was rechecked by staining with CD19, CD3, CD11c, and CD11b. Because FACS analysis showed a purity of >99% B cells (Fig. 2,A) and because RT-PCR gave no signal for CD11c (Fig. 2 B), it is unlikely that cultured B cells are contaminated by myeloid cells.
RT-PCR showed an up-regulation of BAFF and APRIL mRNA when B cells were activated with CpG-ODN or with LPS (Fig. 2,B). The elevated levels of BAFF mRNA expression in LPS- or CpG-ODN-activated B cells correlated with the expression of BAFF on the surface of the cells. Splenic B cells expressed membrane-bound BAFF already 2 days after in vitro activation (Fig. 2 C). Surface expression of BAFF was confirmed by using biotinylated TACI Ig (data not shown).
A control experiment using MyD88-deficient mice demonstrated that the enhancement of BAFF expression is dependent on the TLR signaling pathway (Fig. 2 C). The negative result seen in the MyD88-deficient mice is not due to the C57BL/6 genetic background, since splenic B cells isolated from both BALB/c and C57BL/6 mice up-regulated BAFF expression when activated with CpG-ODN or LPS.
Activation of splenic B cells with LPS or CpG-ODN induces differentiation into plasma cells and expression of intracellular BAFF and APRIL. After 3 days in culture, B cells were harvested by cytospin and stained with BAFF- or APRIL-specific Ab. Costaining with κ-specific Ab demonstrated high levels of APRIL and BAFF expression in the cytoplasm of the newly generated plasma cells (Fig. 3, A and B). Comparable results were found when peritoneal B1 cells were activated with LPS or with CpG-ODN. B1 cells expressed low levels of BAFF and APRIL already before in vitro activation and after 3 days in culture with LPS or CpG-ODN a strong up-regulation of both cytokines was found. Again, double staining with κ-specific Ab showed BAFF and APRIL expression in the cytoplasm of plasma cells (data not shown).
CMK enhances the expression of membrane-bound BAFF on activated B cells
Sorted splenic B cells and also peritoneal B1 cells were activated with LPS or with CpG-ODN in the presence or absence of CMK, an inhibitor of furin-like ecto-peptidases which inhibits cleavage of BAFF from the cell surface and thus release of soluble BAFF (20). After 3 days in culture, B cells were harvested and stained with the BAFF-specific mAb Buffy-2. FACS analysis showed that activated B cells up-regulate membrane-bound BAFF expression (Fig. 4). In cultures with inhibitor, the mean fluorescence intensity (MFI) for BAFF expression was higher than in cultures without (Fig. 4). Activation of splenic B cells with LPS or CpG-ODN in the absence of CMK increased the MFI from 4.25 ± 0.24 (medium alone) to 12.86 ± 0.39 or 10.39 ± 1.36. In the presence of CMK, the MFI increased to 20.29 ± 1.92 or 17.8 ± 1.08, respectively. A similar up-regulation of BAFF expression was observed when peritoneal B1 cells were activated. When incubated with LPS, the MFI was 19.23 ± 1.48 in the absence of CMK and 26.64 ± 2.61 in its presence. Activation of B1 cells with CpG-ODN increased the MFI of BAFF expression to 24.94 ± 3.29 in the absence and 38.54 ± 3.59 in the presence of CMK.
Furthermore, when B2 cells were activated in the presence of CMK, a significant increase in the fraction of BAFFhigh B cells was seen (∼6% of total B cells; Fig. 4,A). A similar result was found for B1 cells, in particular when they were activated with CpG-ODN (Fig. 4 B). The finding of elevated levels of membrane-bound BAFF expression in the presence of CMK gives further evidence that activated B cells express BAFF. These findings indicate that upon activation B cells release soluble BAFF into the culture supernatant.
In vitro activation of B cells induces secretion of BAFF and/or APRIL
To look for the release of soluble BAFF, tissue culture supernatants of splenic B cells activated for 3 days with LPS or CpG-ODN were tested for the presence of soluble BAFF using an ELISA kit. The increase in the level of soluble BAFF was low, but detectable (p < 0.01). Activation with LPS or CpG-ODN yielded a concentration of 1.2 ± 0.1 or 0.9 ± 0.1 ng/ml soluble BAFF, respectively (Fig. 5 A).
The question arises whether the BAFF and APRIL expressed by B cells is of biological significance. To address this question, inhibition assays using the fusion protein TACI Ig were performed. Sorted B cells were incubated with different amounts of LPS in the presence or absence of TACI Ig. Fig. 5, B and C, shows that in the presence of TACI Ig there is a significant reduction in B cell activation. Although a low dose of 1 μg/ml LPS is sufficient to induce B cell proliferation (Fig. 5,C, upper graph), the number of B cells does not increase in the presence of TACI Ig. After 3 days of activation with 1 μg/ml LPS, approximately the same number of B cells were counted as at the start of cultures (Fig. 5,B). In the presence of TACI Ig, the frequency of cycling B cells was significantly reduced (Fig. 5,C, middle graph). Furthermore, the differentiation of activated B cells into CD138high plasma cells was inhibited (Fig. 5 C, lower graph). The finding that in the presence of the antagonist TACI Ig, B cell survival, proliferation, and differentiation are all inhibited indicates that in vitro activation of B cells induces secretion of biologically active BAFF and/or APRIL.
In vivo activation of B cells and up-regulation of BAFF expression
In primary follicles of BALB/c mice, FDC do not express BAFF (Fig. 3,C, C3). Staining with APRIL-specific Ab was also negative (data not shown). To induce a T-dependent immune response, BALB/c mice were immunized with phOx and 10 days after injection of Ag, GC formation was observed (Fig. 3 C).
Double staining with the FDC-specific M2 Ab and anti-BAFF showed that BAFF expression is up-regulated in FDC of the GC light zone (Fig. 3,C, C2). An analysis of BAFF expression at high magnification suggests that BAFF is also expressed in GC B cells, as B cells in the dark zone of the GC costained for PNA and BAFF (Fig. 3,C, C3). From staining of tissue sections, it is difficult to say whether BAFF is bound by its receptors or indeed expressed on the surface of B cells. However, FACS analysis of GC B cells showed up-regulation of membrane-bound BAFF and intracellular staining for BAFF confirmed endogenous production of BAFF by GC B cells (Fig. 6). In accordance with previous results, staining with APRIL-specific Ab showed no significant expression of APRIL in the FDC network and in GC B cells (14).
Enhanced levels of BAFF and APRIL mRNA expression in B cells of (NZB × NZW)F1 mice
(NZB × NZW)F1 mice, which at the age of 3–4 mo spontaneously develop a lupus-like syndrome, were used as a disease model for chronic autoimmunity. An increase in MZ and CD5+ B cell populations was found with increasing age and onset of disease (33). Furthermore, the frequency and the absolute numbers of both plasma cells and GC B cells increased with age (Fig. 7, A and B). The finding of both an expansion of the MZ and the spontaneous development of GC suggests that B cells are continuously activated in (NZB × NZW)F1 mice.
To test whether the activation of the immune system affects APRIL and BAFF expression in B cells, spleens were prepared from 4- to 6-wk-old (NZB × NZW)F1 mice before the onset of autoimmune disease, from 3- to 4-mo- old animals at the time point when the disease becomes apparent and from 6-mo-old mice with established autoimmune disease. In a first step, MZ and FO cells were isolated and the level of BAFF and APRIL mRNA was compared in young and old (NZB × NZW)F1 mice (Fig. 8). Both, female and male mice were analyzed. In contrast to BALB/c mice, where no APRIL expression was seen in FO and MZ B cells, an up-regulation was evident in (NZB × NZW)F1 mice even before the onset of autoimmune disease (Fig. 8,A). The level of APRIL mRNA increased further with the development of disease (Fig. 8 A).
Enhanced levels of BAFF mRNA expression were mainly seen in MZ B cells. A significant increase was found in female animals at the age of 2–4 mo and in male animals at the age of 6- mo (Fig. 8,B). For FO B cells, no significant up-regulation in BAFF mRNA was found. In B1 B cells, an increase in the level of BAFF mRNA expression with age was found. However, in 6- to 7-mo-old (NZB × NZW)F1 mice, the level of BAFF expression was not significantly different from that in BALB/c mice (Fig. 8 B).
BAFF expression is up-regulated in GC of (NZB × NZW)F1 mice
In 3- to 4-mo-old (NZB × NZW)F1 female mice, staining of splenic tissue sections showed large GC with strong BAFF expression in their dark and light zones (Fig. 3,D). A comparison of fully developed GC from immunized BALB/c mice with those from (NZB × NZW)F1 mice suggested enhanced BAFF expression in GC (Fig. 3, C and D). However, when BAFF expression in GC B cells was analyzed by FACS, no significant difference was found (Fig. 6). GC B cells from BALB/c, old and young (NZB × NZW)F1 mice showed a 3- to 4-fold increase in their MFI when compared with naive B cells (Fig. 6). Intracellular staining confirmed that GC B cells express high levels of BAFF (Fig. 6).
Staining of splenic tissue sections of 3- to 4-mo-old (NZB × NZW)F1 mice showed APRIL expression in the light zone of GC. Using confocal microscopy, colocalization of the FDC-M2 signal with APRIL expression was found (data not shown). Whether APRIL expression is also enhanced in GC B cells could not be determined.
In the spleen of (NZB × NZW)F1 mice plasma cells express BAFF and APRIL
In vitro activation of B cells with LPS or CpG-ODN induced BAFF and APRIL expression. The newly developing plasma cells showed high cytoplasmic expression of these cytokines (Fig. 3, A and B). We therefore analyzed whether in vivo-generated plasma cells also express BAFF and APRIL. Splenic tissue sections of 6- to 7-mo-old (NZB × NZW)F1 mice were prepared and stained for BAFF or APRIL expression. Costaining with κ- specific Ab demonstrated that the majority of plasma cells in the spleen of (NZB × NZW)F1 mice express both BAFF and APRIL (Fig. 3,E). FACS analysis confirmed that in (NZB × NZW)F1 mice with established autoimmune disease practically all splenic plasma cells (CD19lowCD138+) express high levels of both membrane-bound and also intracellular BAFF (Fig. 9).
BAFF is a fundamental survival factor for B cells and plays an essential role in the homeostatic regulation of the naive peripheral B cell pools (2, 3, 4). In the absence of BAFF or BAFF-R, only a few transitional B cells will differentiate into FO and MZ B cells (5, 24, 25, 26). APRIL has no essential function in B cell development. However, there is evidence that it supports B cell proliferation and acts as cofactor in class switch (2). In this study, we show for the first time that BAFF and/or APRIL produced by murine B cells themselves support B cell development and survival.
Our analysis shows that B cells at all stages of differentiation express BAFF mRNA, while APRIL mRNA was restricted to early B cell development in the bone marrow and to peritoneal B1 B cells. As described for human peripheral B cells, no membrane-bound expression of BAFF was detectable on resting mature B cells (21, 28, 29), but both in vitro and in vivo activation resulted in up-regulation of BAFF and APRIL. Using MyD88-deficient mice, we show that the up-regulation of BAFF and APRIL expression following treatment with LPS or CpG-ODN depends on MyD88-TLR4 and TLR9 complex signaling (Fig. 2).
One might argue that BAFF may be released by contaminating myeloid cells and that the increased level of BAFF surface expression which we see is due to up-regulation of BAFF-R. Were that to be the case, no increase in the level of BAFF should be observed when B cells are activated in the presence of the protease inhibitor CMK (Fig. 4). However the opposite was found. When B cells were activated with LPS or CpG-ODN in the presence of CMK, the MFI of BAFF expression increased significantly (p < 0.01). These results suggest that B cells release soluble BAFF which may then support an autocrine survival pathway; however, the mechanisms are unclear.
In general, cytokines of the TNF superfamily are functional both in their membrane-bound and in soluble form when shed by enzymatic cleavage from the surface through a furin-like convertase (34). An alternative direct secretion pathway for BAFF has been reported, but it is unique to neutrophils (19). Also, for APRIL it was found that it is secreted following intracellular processing in the Golgi apparatus (35). Our data demonstrate membrane-bound and intracellular cytoplasmic BAFF and APRIL expression in B cells. It is therefore possible that the survival of B cells and, in particular, of plasma cells may be supported by autocrine cytokines signaling both through receptors in the outer membrane and in the cytosolic compartments (8). Our finding that in the presence of the antagonist TACI Ig B cell activation is inhibited (Fig. 5) indicates that B cells are supported by an autocrine pathway. Upon stimulation, B cells themselves release biologically active BAFF and/or APRIL into the culture supernatants.
The analysis of the GC reaction suggested that BAFF and APRIL are required for both the development of the mature FDC network in the GC and for the maintenance of the GC (26, 36, 37). Our data show that immunization of BALB/c mice with a T-dependent Ag up-regulates BAFF expression both in FDC and in GC B cells (Fig. 3). High-affinity interaction between the Ag and the BCR may up-regulate BAFF expression and help the Ag-selected B cell to survive and to differentiate into an effector cell. The differentiation of GC B cells seems to be independent of APRIL because we see no significant expression of it in GC. This is in accordance with a previous report (14).
In T-independent responses, the simultaneous up-regulation of both BAFF and APRIL in TLR-activated B cells may help to compete with Ag-inexperienced B cells for short-term survival niches. The autocrine pathway may thus provide a mechanism that promotes survival and differentiation of Ag-activated B cells in defined microenvironments within the peripheral lymphoid organs.
Our analysis of (NZB × NZW)F1 mice shows that the chronic activation of the immune system in these animals enhances BAFF and APRIL expression in B cells. In particular, plasma cells expressed high levels of BAFF and APRIL and this may enhance their life span in the spleen. Furthermore, cytokine secretion by plasma cells may also contribute to the increased levels of BAFF in sera of the (NZB × NZW)F1 mouse.
Similarly, in patients with autoimmune diseases like lupus erythematosus, Sjögren’s syndrome, and rheumatoid arthritis, elevated levels of BAFF may be generated by activated B cells (38, 39, 40). Autocrine BAFF will then become part of a vicious circle in which enhanced serum levels of BAFF lower the threshold of affinity-based selection and result in an increase in the frequency of autoreactive B cells.
We find that the majority of plasma cells in autoimmune (NZB × NZW)F1 mice express high levels of BAFF and APRIL. This observation is potentially of clinical importance, particularly with respect to ongoing clinical trials of BAFF and APRIL antagonists. These trials aim to deplete pathogenic autoreactive B and plasma cells. However, if these cells themselves produce their own survival factors, they may turn out to be resistant to this type of therapy.
We are thankful for technical support by members of the DRFZ, in particular S. Schürer and G. Steinhauser for their help in preparing TACI Ig. We also thank S. Fillatreau, R. S. Jack, and A. Radbruch for critical discussion.
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 work was supported by a grant from the Government of Vietnam and the DAAD (to V.T.C.), the Bundesministerium für Bildung und Forschung Grant NGFN2 and the SFB 650. The DRFZ is supported by the Berlin Senate of Research and Education.
Abbreviations used in this paper: BAFF, B cell-activating factor; APRIL, a proliferation-inducing ligand; MZ, marginal zone; FO, follicular; GC, germinal center; FDC, follicular dendritic cells; CMK, chloromethylketone; phOx, 2-phenyl-oxazolone; PNA, peanut agglutinin; MFI, mean fluorescence intensity; TACI, transmembrane activator and calcium modulator ligand interactor; ODN, oligodeoxynucleotide; BAFF-R, BAFF receptor.