B cells-activating factor
Human B cell-activating factor (BAFF) induces mouse surface IgM+ B cells of the immature type from bone marrow and of the immature types 1 and 2 from spleen, as well as of the mature type from spleen to increased longevity in tissue culture. BAFF does so polyclonally and without inducing proliferation in any of these B cell subpopulations. BAFF induces phenotypic and functional maturation of immature to mature B cells so that all immature cells loose C1qRp (AA4.1, 493) expression and type 1 immature cells up-regulate IgD, CD21 and CD23. Immature B cells of types 1 and 2, upon pre-incubation with BAFF, change their reactiveness to Ig-specific antibodies so that they no longer enter apoptosis but now proliferate. However, BAFF does not seem to overcome negative selection of developing immature B cells in vitro.
In the adult, B lymphocytes develop in bone marrow in a series of steps which finally leads to the surface deposition of IgM (sIgM) on so-called immature B cells (for reviews see 1, 2). These immature B cells are short-lived with life-spans estimated to be between 2–4 days 3–11. They leave the bone marrow to enter the spleen, still as short-lived immature B cells. Between the exit from the bone marrow and the entry into the spleen 80% of the immature B cells are lost, possibly due to negative selection by autoantigens 3–13. In the spleen, immature B cells can be distinguished from their mature counterparts by the expression of the antigen recognized by the mAb 493 developed in our laboratory 2–4. Rather recently we discovered that the antigen recognized by the 493 mAb is the complement component C1q like receptor C1qRp (14, 15, and Rolink et al., unpublished observation). This receptor is also recognized by the mAb AA4.1 14. At least two types of immature B cells can be distinguished in spleen, called type 1 as sIgMhigh sIgDlow or IgD– CD21– CD23– C1qRp+ cells, and type 2 as sIgMint sIgDhigh CD21+ CD23+ C1qRp+ cells 16. Both types of cells undergo apoptosis upon stimulation by sIg-specific antibodies 3, 11. Immature B cells then further develop in the spleen into mature B cells. Mature B cells have life-spans of over six weeks and react to stimulation via sIg by proliferation 2–11.
A new molecule belonging to the TNF family of ligands, called B cells-activating factor (BAFF) (also termed TALL-1, THANK, BlyS and zTNF4 17–21) was discovered and found to stimulate splenic immature B cells to survival in vitro, specifically the immature B cells of type 2 22. In costimulation with μH chain-specific antibodies it was found to change immature B cells to a mature phenotype 22. BAFF has been found to be the exclusive ligand, which finds three receptors, TACI, BCMA and BAFF-R for itsactions on B cells 17, 21, 23–31. Mice transgenic for BAFF have been seen to develop SLE-like autoimmune disease, which has ledto the speculation that BAFF inhibits negative selection by autoantigens via the Ig antigen receptor in allowing immature B cells to transit from an apoptosis-inducible state to a proliferation-inducible state 32–34.
In this study we extended the in vitro observations of the actions of BAFF on B lineage cells. We show that three stages of immature B cells, one in bone marrow and the two subsequentones in spleen, as well as mature splenic B cells are all induced to longevity in tissue culture in a polyclonal fashion and without entering proliferation. Moreover, we show that immature B cells,upon stimulation with BAFF, acquire the phenotype of a mature B cell and proliferate instead of undergoing apoptosis when stimulated via their B cell receptor. However, BAFF does not seem to rescuedeveloping immature B cells from negative selection in vitro.
2.1 BAFF is a survival factor for immature and mature B cells
Immature B cells from the bone marrow of C57BL/6 mice were FACS-purified as B220int IgM+ cells. Immature B cells from the spleen of the same strain of mice were FACS-purified into two subpopulations, called transitional 1 as B220+ C1qRp+ CD23– and transitional 2 as B220+ C1qRp+, CD23+ (Fig. 1). Mature B cells from the spleen were FACS-purified as B220+ C1qRp– CD23+ cells (Fig. 1). These four cell populations were then cultured in the presence or absence of BAFF and the numbers of live cells remaining in culture were monitored each day for 5 days. Results in Fig. 2A show that the approximate life-spans of immature B cells from the bone marrow, and those of type 1 from the spleen, was 2 days, while that of immature B cells of type 2 and of mature B cells was around 2.5 days (Fig. 2B). In the presence of BAFF a dramatic increase in the life expectancy of all four B cell types was observed, so that after 4 days the number of cells in culture had not decreased significantly (Fig. 2A, B).
2.2 BAFF does not induce B cell proliferation
To test whether the observed increase in survival was due to BAFF-induced proliferation, total splenic immature (B220+ C1qRp+) and mature B cells (B220+ C1qRp–) were FACS-purified. As seen by the lack of [3H]thymidine uptake of these cells cultured for 3 days, BAFF does not induce proliferation (Fig. 3A, B). On the other hand, both immature and mature B cells show a strong proliferation response in the presence of the mitogen LPS, while addition of IgM-specific antibody induces proliferation of mature and apoptosis in immature B cells (3, and Fig. 3A, B).
Previously it was shown that BAFF can act as a potent costimulator together with anti-IgM for proliferation of B cells in vitro17, 20, 22, 33. As shown in Fig. 3B, the [3H]thymidine uptake in mature B cells stimulated with anti-IgM in the presence of BAFF is approximately double that obtained when the same cells were stimulated with anti-IgM alone. Immature B cells on the other hand do not proliferate when stimulated with anti-IgM in the presence of BAFF (Fig. 3A).
2.3 BAFF is a maturation factor for immature B cells
It has been shown that immature B cells undergo apoptosis upon sIg cross-linking, while mature B cells, under the same conditions, are induced to proliferate 1–3, 11. Results in Fig. 4A confirm and extend these findings by showing that in both splenic immature B cell populations anti-IgM antibody induces apoptosis while in mature B cells proliferation is induced. Costimulation of immature B cells with anti-IgM and BAFF does not significantly reduce the induced apoptosis (data not shown). However, pre-incubation of these immature B cells with BAFF for 3 days changes their reactivity, so that they no longer undergo apoptosis but instead proliferate when later exposed to anti-IgM (Fig. 4B).
We conclude from these experiments that BAFF is a survival factor for immature and mature B cells, and moreover, induces the differentiation of immature B cells into mature B cells. This latter point is further substantiated by the finding that C1qRp+ IgMhigh IgDlow CD21– CD23– immature type 1 splenic B cells, upon incubation for 3 days with BAFF, become to a large extent C1qRp– IgMint IgDhigh CD21+ CD23+, i.e. acquire a mature B cell phenotype (Fig. 5).
2.4 BAFF does not inhibit negative selection of B cells in vitro
Mice transgenic for BAFF have been shown to develop SLE-like autoimmune disease 32–34. This has led to the speculation that BAFF might inhibit negative selection of autoreactive immature B cells produced in the bone marrow. Here, using an in vitro differentiation system we tested the potential influence of BAFF on this process of negative selection. Previously, we showed that ex vivo isolated pre-B I cells proliferate and differentiate in vitro, in a pre-BCR-dependent manner, into sIgM+ immature B cells of which 90% express a κ-L chain and 10% a λ-L chain 35 (Fig. 6A). When pre-B I cells are induced to differentiate in the presence of anti-κ-L chain antibody, cell recovery at the end of the culture period is identical to that of control cultures, indicating that no apoptosis is induced by the anti-κ-L chain antibody. However, the percentage of sIgM+ immature B cells that appear in the presence of anti-κ-L chain antibody is dramatically decreased and all immature B cells formed use λ-L chains (Fig. 6B). Thus, under these culture conditions, the formation of κ-L chain-bearing immature B cells is completely inhibited, most likely due to negative selection as normally operating in the bone marrow 12, 13.
The in vitro differentiation of pre-B I cells into immature B cells is neither qualitatively nor quantitatively influenced by the presence of BAFF (Fig. 6C). Moreover, the inhibition of the formation of κ-L chain-bearing immature B cells, as observed in the presence of anti-κ-L chain antibody, cannot be overcome by BAFF, indicating that BAFF does not interfere with negative selection in this in vitro system.
In B cell development of adult mice, the first cells expressing IgM on their surface are immature B cells, which are short-lived with life-spans of 2–4 days. These immature B cells exit the bone marrow to first enter the spleen, still as immature sIgM+ B cells, where they then become mature.
At the transition from immature B cells in bone marrow to mature B cells in spleen, four out of five cells are lost, presumably due to negative selection, i.e. by induction of apoptosis due to an interaction of IgM on their surface with autoantigens. Moreover, it has been shown that peripheral B cells are deleted when conditional targeting of sIg expression on them is abolished 36, 37. This sIgM expression and possibly constant signaling via this receptor is an absolute necessity for the maintenance of the peripheral B cell pool.
Immature B cells in the spleen have life-spans of 2–4 days, while mature B cells have a life-span of over 6 weeks. Recent experiments 23, 24, 26, 34, 38 have indicated that immature B cells next to sIgM require signaling via the receptor BAFF-R, induced by its ligand BAFF, for their transition into themature B cell compartment. It remains to be studied in more detail, whether the mere surface deposition of IgM, or its interaction with antigens is needed in a cooperative action with BAFF to accomplish the transition from immature to the mature stage of a B cell.
Batten et al. 22 conducted in vitro experiments with mouse splenic B cells that suggested that type 2 immature B cells were the major targets for BAFF to induce increased survival in vitro, and that with costimulation by μH chain-specific-antibody these cells could develop to a mature phenotype. It was shown very recently that BAFF-deficient mice have a severe B cell developmental defect 23, 24, 34. In these mice, B cell development up to immature type 1 in the spleen seems to be normal. However, immature type 2, follicular and marginal zone B cells are practically absent. A similar defect in B cell development was previously observed in the A/WySnJ mouse 39–41. Recently it was shown that these mice carry a mutation in the gene encoding BAFF-R 26. Due to this mutation A/WySnJ mice express a BAFF-R lacking the cytoplasmic tail. Thus the BAFF-induced B cell maturation appears to be mediated via BAFF-R.
Mice lacking the BAFF receptor BCMA 42 show no abnormalities in B cell development, indicating that this receptor is at least redundant in this maturation process. On the other hand, mice lacking the BAFF receptor TACI 43, 44, have increased numbers of peripheral B cells, suggesting that the interaction of BAFF with TACI results in asignal that partly inhibits B cell development.
In this study we have confirmed the in vitro survival results with type 2 immature cells 22; however, we also show that splenic immature B cells of type 1, immatureB cells of bone marrow and even mature B cells from spleen are all induced to longevity in vitro by BAFF. The finding that mature B cells require BAFF to survive in vitro may suggest that the interaction of BAFF with its receptor is also required for the longevity of mature B cells in vivo. A mouse in which BAFF-R is specifically deleted in the mature B cell compartment would be needed to test this hypothesis.
Our in vitro studies also show that the action of BAFF on all these B cell subpopulations is polyclonal, i.e. independent of the specificity of sIgM. Furthermore, the findingthat immature B cells upon pre-incubation with BAFF become resistant to anti-IgM-induced apoptosis, and in fact like mature B cells proliferate under these conditions, together with the finding that immature B cells in the presence of BAFF acquire a mature phenotype, shows that BAFF plays a crucial role in the transition from immature to mature B cells. Whether this maturation is directly induced by BAFF or whether it is an intrinsic property of immature B cells but requires prolonged survival is still an open question.
BAFF-transgenic mice have increased numbers of type 2 immature B cells and extended marginal zones with B cells in vivo. They also develop SLE-like autoimmune disease 32–34. This had led to the speculation that BAFF-induced maturation competes with negative selection of the emerging B cell repertoire by autoantigens and in fact, allows the survival of autoantigen-reactive B cells and their transition into the mature, peripheral pool of B cells. Our in vitro data suggest that BAFF does not interfere with central B cell toleranceinduction normally occurring in the bone marrow. However, our data also indicate that immature autoreactive B cells might be rescued from negative selection when exposed to the actions of BAFF prior to contact with their autoantigen. Whether such a sequence of reactions with immature B cells can also occur in normal physiological or in other pathophysiological states resulting in SLE-type autoimmune disease remains to be elucidated in detail.
4 Materials and methods
C57BL/6 (B6) mice between 6 and 8 weeks of age were purchased from IFFA CREDO, France.
4.2 Flow cytometric analysis and cell sorting
Cell surface staining was performed as previously described 3, 4. The following homemade mAb were used: M41 (anti-IgM), 1.19 (anti-IgD), 493 (anti-C1qRp), RA3–6B2 (anti-B220), ACK-4 (anti-c-kit) and 187.1 (anti-κ-L chain ). mAb against CD21, CD23 and λ1+2-L chain were purchased from BD/PharMingen (San Diego, CA). For sorting of stained cells the MoFlo (Cytomation Inc., Fort Collins, CO) was used.
4.3 Cell cultures
Immature B cells from bone marrow were sorted as B220int IgM+ cells, while immature B cells of type 1 and type 2 from the spleen were FACS-purified by being B220+ C1qRp+ CD23– and B220+ C1qRp+ CD23+, respectively (Fig. 1). Total immature B cells from spleen were sorted as B220+ C1qRp+. Mature B cells were purified as B220+ C1qRp– CD23+ (Fig. 1). The sorted B cell populations were cultured in IMDM containing 5×10–5 M 2-ME, 0.03% primatone (Quest International, Naarden, The Netherlands) and 2% FCS at 5×105 cells/ml in the presence or absence of 200 ng/ml BAFF (Apotech Biochemicals, Epalinges,Switzerland). At various time points viable cells were counted in a light microscope using a Neubauer hemocytometer and the trypan blue exclusion test. In some experiments cells were harvested on day 3 and subjected to flow cytometric analysis, or recultured for 36 h in the presence of 20 μg/ml anti-IgM. Thereafter, the cells were spun down and resuspended in 0.5 ml hypotonic fluorochrome solutioncontaining 50 μg/ml propidium iodide, 0.1% sodium citrate and 0.1% Triton X-100 45. After overnight incubation at 4 °C in the dark, samples were analyzed on a FACScalibur.
To determine the capacity of BAFF to induce proliferation, sorted immature and mature splenic B cells were cultured at 5×105 cells/ml for 3 days in medium containing either 200 ng/ml BAFF, 200 ng/ml BAFF plus 20 μg/ml anti-IgM, 20 μg/ml anti-IgM or 10 μg/ml LPS. Cultures were pulsed with 1 μCi [3H]thymidine during the last 6 h of culture.
B220+ c-kit+ bone marrow pre-B I cells were sorted and cultured at 4×105 cells/ml in the presence or absence of 20 μg/ml monoclonal 187.1 anti-κ-L chain antibody with or without 200 ng/ml BAFF. At day 4 of culture the appearance of sIg+ B cells was determined by FACS.
We thank Jan Andersson for critical reading of the manuscript. A. G. R. is the holder of the chair in Immunology endowed by F. Hoffmann-La Roche Ltd., Basel, Switzerland. This work was supported by the Swiss National Science Foundation (No. 3100-066692.01).