The BCR-triggered responses of mature and transitional immature B cells differ at both the biochemical and functional level. In this study, we show that in mature B cells, BCR signaling triggers Vav phosphorylation and Rac1 activation. Furthermore, we demonstrate that although downstream actin-dependent BCR capping is independent of Rac1 activation, actin-dependent membrane ruffling and cell spreading are Rac1-dependent processes. In contrast, BCR-induced Vav phosphorylation and Rac1 activation is impaired in transitional immature B cells, resulting in defects in actin polymerization-dependent spreading and membrane ruffling while Rac1-independent BCR capping remains intact. Because transitional immature murine B cells maintain lower steady-state levels of plasma membrane cholesterol, we augmented their levels to that of mature B cells and found that BCR-induced Rac1 activation and Rac1-dependent membrane ruffling and cell spreading were restored. These studies provide a direct link between B cell cholesterol levels and downstream cellular signaling processes.
Antigen- or anti-BCR-mediated aggregation of the BCR on mature B cells leads to its organization into cholesterol-enriched membrane microdomains termed lipid rafts, actin-dependent polarization of the engaged BCRs into concentrated “caps”, and dynamic changes in plasma membrane morphology termed “ruffling” (1, 2, 3, 4). Each of these processes occurs coincident with generation of sustained signals necessary for B cell activation. These signals are initiated by Src kinase-mediated tyrosine phosphorylation of ITAMs contained within BCR Igα and Igβ, followed quickly by recruitment and activation of spleen tyrosine kinase (Syk) (5, 6). Src and Syk kinases phosphorylate adaptor proteins such as B-cell linker protein which in turn organizes a macromolecular signaling complex containing phospholipase Cγ2 (PLCγ2),3 Bruton’s tyrosine kinase (Btk) PI3K, Vav as well as other proteins (7, 8). PLCγ2 activation leads to hydrolysis of phosphatidylinositol 4,5 bisphosphate and ultimately to calcium flux and NF-κB activation (9, 10). Signals downstream of Vav activate members of the Rho GTPase family such as Rac, Cdc42, and RhoA (11) and recruitment of Vav to lipid rafts appears to be necessary for optimal Rac1 activation (12). Rac1 activation correlates with the ability of cells to undergo stimulation-induced cytoskeleton reorganization, possibly involving ezrin dephosphorylation-dependent dissociation of the actin cytoskeleton from the plasma membrane (13).
Mature splenic B cells develop from immature B cells that emigrate from the bone marrow. These emigrants and those that first enter the spleen are known as transitional immature B cells. Like bone marrow immature stage B cells, transitional immature B cells pass through developmental checkpoints where self-reactive B cells can be identified and eliminated. Signaling through the BCR at these immature B cell stages triggers tolerogenic responses such as apoptosis, anergy, and the reprogramming of BCR specificity by a process referred to as receptor editing (14, 15, 16). B cells that escape negative selection develop into follicular immunocompetent mature B cells that have the potential to differentiate into effector cells in response to BCR signaling (17). Thus, the responses of immature B cell subsets and mature B cells are different. Although previous studies have indicated that this difference is linked to unique BCR-signaling phenotypes (9), the molecular basis for this difference remains undefined.
Proteins such as Vav and Rac that are involved in actin cytoskeleton remodeling play an essential role in normal B cell development and survival. Indeed, B cells present in Vav1/Vav2-null mice and Rac1/Rac2-null mice have similar phenotypes (18, 19, 20), characterized by a developmental block that prevents transitional immature B cells from differentiating into follicular mature B cells. In vitro, anti-BCR-induced proliferation and survival are compromised in these B cells. In addition, B cells lacking the adaptor protein Bam32 are deficient in Rac1 activation, do not undergo BCR-induced membrane ruffling, and have a transient signaling phenotype characterized by susceptibility to anti-BCR-induced cell death (21, 22). Together, these studies indicate that the ability to effectively signal through the Vav/Rac1 pathway is required for the induction of BCR-induced cellular responses normally observed in mature B cells and suggests that BCR-induced responses in transitional immature B cells may more closely recapitulate what is observed in Vav1/Vav2-, Rac1/Rac2-, or Bam32-null B cells.
Biochemical characterization of BCR-induced signaling in follicular mature B cells and transitional immature B cell has revealed many differences. For example, transitional immature B cells are relatively impaired in their ability to activate protein kinase C (PKC) (9) and exhibit an associated inability to sustain signaling through the PKCβ/NF-κB/c-myc pathway (23). Furthermore, although mature B cells readily translocate their BCR to the lipid raft compartment following stimulation as determined by both biochemical means and immunofluorescent microscopy, transitional immature B cells do not (1, 23). We have previously reported that the plasma membrane of both early transitional immature B cells (T1) and the more mature T2 subset contain less unesterified free cholesterol than do mature B cells (23). Furthermore, augmenting membrane cholesterol levels in transitional immature B cells to levels comparable to mature B cells rescues inducible BCR association with lipid rafts and sustained signaling through the PLCγ2/NF-κB/c-myc pathway. Therefore, membrane cholesterol content and lipid raft compartmentalization are necessary for the sustained BCR signaling missing in the immature B cell subsets (23). As a direct link between sustained AgR signaling and actin cytoskeleton reorganization is supported by a number of studies (24, 25, 26, 27), our current studies investigate whether the inability of transitional immature B cells to copolarize their BCR with lipid rafts following BCR stimulation prevents efficient cytoskeletal reorganization, thereby precluding sustained BCR-induced signaling. Understanding the mechanisms involved in developmental regulation of BCR signaling would be a significant step for understanding B cell-negative selection and for designing strategies to manipulate Ag-specific B cell responses.
Materials and Methods
Isolation and definition of mature and transitional immature B cell subsets
Transitional immature B cells were isolated from spleens of 6- to 10-wk-old autoreconstituting BALB/c mice 14 days after sublethal irradiation (500 rad) by depletion of CD43+ cells using MACS beads. Mature B cells were similarly enriched from nonirradiated mice. Enriched populations ranged from 95 to 98% B cells as assessed by B220 expression. Nonsorted mature populations generally contained 10–15% transitional immature B cells as judged by CD93 (AA4.1+) expression. Mice were maintained under pathogen-free conditions at the University of Pennsylvania, and experimental procedures in these animals were performed according to local and National Institutes of Health guidelines.
Immunofluorescence and microscopy for BCR polarization/capping
Purified B cells (2 × 106 in 200 μl of PBS) were stimulated with a goat IgG F(ab′)2 anti-BCR Ab (Jackson ImmunoResearch Laboratories) at 37°C for 5 min, and the reaction was stopped by addition of cold PBS/0.1% BSA/0.02% azide. Cells were fixed in 4% paraformaldehyde/0.1% glutaraldehyde, followed by washing and resuspension in 0.5 mg/ml sodium borohydride to stop fixation. Cells were then incubated with a Cy3-conjugated secondary Ab directed to the anti-BCR primary Ab at 4°C, mounted onto glass dishes, and overlaid with Molecular Probes Prolong Gold anti-fade with 4′,6-diamidino-2-phenylindole (DAPI). Cells were visualized using a Zeiss Axiovert 200M inverted epifluorescence microscope equipped with a Sensicam QE high-performance camera. Images were captured and analyzed using Slidebook image analysis software (Intelligent Imaging Innovations). The No Neighbors deconvolution module of the software was used to remove out-of-focus light. Cells were scored for capping according to the following criteria: 1) cells were only counted if not in contact with other cells. 2) The BCR was visually scored as capped when the BCR polarized on less than half the circumference of the cell. If it was not clear that the majority of the BCR polarized visually, then a mask was drawn on opposite poles of the cell and the pixel intensity was determined using the Slidebook software. 3) At least 50 cells were scored for each condition per experiment.
Immunofluorescence for live cell imaging and membrane ruffling
Purified B cells (20 × 106 in 1 ml of PBS) were incubated with a monovalent Fab directed against the BCR H chain (μ-chain specific) coupled to Cy3 (Jackson ImmunoResearch Laboratories) that does not cross-link the BCR. Following washing, the B cells (2 × 106 cell in 200 μl) were mounted onto a glass plate and placed on the inverted epifluorescent microscope at room temperature. The AgR was then cross-linked with an anti-L chain Ab, and B cells were imaged every 30 s for 10 min.
Immunofluorescence for B cell/bead conjugates
Polystyrene 4.5-μm beads were incubated with BSA or anti-BCR L chain Ab at 4°C overnight. Purified B cells (10 × 106 in 500 μl of PBS) were centrifuged with prewashed beads (1:1 ratio of beads to cells) for 5 s to facilitate conjugate formation and then incubated at 37°C for 10 min. The reaction was stopped and cells were fixed as described above. Following plating, the BCR was detected with a monovalent Fab directed against the BCR μH chain (described above). Cells were then permeabilized with 0.1% saponin and F-actin was detected with Alexa Fluor 488 phalloidin (Molecular Probes). Cells were visualized as described above.
Immunofluorescence for B cell spreading
For spreading assays, purified B cells (20 × 106 in 1 ml of PBS) were incubated with a monovalent Fab directed against the BCR μH chain and washed as described above. The B cells (2 × 106 cell in 200 μl) were mounted onto either poly-l-lysine coated- or anti-BCR L chain-coated glass plates and heated at 37°C for 10 min. The reaction was stopped and cells were fixed as described above. Cells were scored for surface area according to the following criteria: 1) cells were only scored if the complete outline of the cell was visible. 2) At least 50 cells were scored for each condition per experiment. Slidebook software reported the relative surface area after a mask was created around the circumference of the cell. Statistical analysis was performed using the Student t test. A p value ≤0.001 was considered significant while a p value ≥0.05 was considered insignificant.
Treatment of B cells with Tat-control or Tat-Rac1 dominant-negative peptides
Tat-control and Tat-Rac1 dominant-negative peptides were purchased from SynPep. Purified B cells (5 × 106 cells in 500 μl of PBS) were incubated with 0.5 mg/ml Tat-control peptide or Tat-Rac1 dominant-negative peptide for 5 min at 37°C. Cells were washed and immunofluorescence assays for B cell spreading and B cell capping were performed as described above.
Addition of cholesterol to cells and filipin staining
Cholesterol was added to transitional immature B cells as previously described (23). Briefly, transitional immature B cells (1 × 107 cells/ml) were incubated in RPMI 1640 containing 300 μg/ml cholesterol (Sigma-Aldrich) plus 5 mM methyl-β-cyclodextrin (MβCD) (Sigma-Aldrich) at 37°C for 120 min. To determine the amount of membrane-associated cholesterol in each cell type, 1 × 106 cells were fixed in 120 μl of 4% paraformaldehyde at room temperature for 12 min, washed three times in PBS, and incubated with filipin (1:100 of 25 mg/ml stock; Sigma-Aldrich) for 1 h at 4°C. Cells were then subject to flow cytometric analysis and the filipin-binding levels were determined.
Purified B cells were (5–10 × 106 cells/sample) were stimulated with a goat IgG F(ab′)2 anti-BCR as described above. Cells were lysed in 2× Mg-containing lysis buffer (Upstate Biotechnology). A total of 10 μg of PAK assay reagent (Upstate Biotechnology) was added to the samples and incubated for 60 min at 4°C. PAK beads were pelleted at full speed for 30 s and washed in Mg-containing lysis buffer three times and run on SDS-PAGE. Total Rac1 was determined on Western blots using anti-Rac1 Ab (Upstate Biotechnology). Western blots were treated with ECL (Amersham) and exposed to x-ray film. Relative band intensities were measured using NIH ImageJ software.
Biochemical analysis and Abs used for Western blotting
Purified B cells (5–10 × 106 cells/sample) were stimulated with a goat IgG F(ab′)2 anti-BCR as described above and lysed in ice cold radioimmunoprecipitation assay lysis buffer. Abs used in Western blotting were 4G10 (Upstate Biotechnology), anti-Vav (Santa Cruz Biotechnology), anti-phospho-JNK and anti-JNK (Cell Signaling Technology). Blots were developed and quantified as described above.
Transitional immature B cells exhibit selective defects in BCR-induced actin-dependent plasma membrane processes
Capping of the BCR following anti-BCR- or Ag-induced aggregation is dependent on both tyrosine kinase signaling and actin polymerization (28). As we and others have previously shown that BCR-induced signals in transitional immature B cells differ from those observed in mature B cells (23, 29), we assessed whether transitional immature B cells were able to cap their BCR after stimulation with soluble anti-BCR Ab. Fig. 1,a illustrates the uniform distribution of the BCR on both unstimulated transitional immature and mature B cells. Aggregation of the BCR on mature B cells with soluble anti-BCR resulted in a typical capping pattern and the capping response observed in transitional immature B cells was not remarkably different. Treatment with the actin inhibitor cytochalasin D decreased the capping response in both mature and transitional immature B cells (Fig. 1 b), verifying that this response was dependent on actin polymerization. Although cytochalasin D-treated B cells clearly did not organize BCR into single caps, they nevertheless exhibited a punctate pattern that is likely due to actin-independent aggregation of BCR complexes as a consequence of passive clustering by cross-linking receptors with anti-BCR.
The capping studies described above used B cells that had been subject to fixation after anti-BCR stimulation. However, actin-dependent membrane changes such as membrane ruffling and cell spreading are dynamic processes that are difficult to capture unless visualized in real time on live cells. Real-time imaging of live mature and transitional immature B cells revealed that BCR capping began ∼2–3 min after AgR cross-linking and, consistent with studies using fixed cells, occurred in both mature and transitional immature B cell populations (Fig. 1,c). In contrast, while membrane ruffling was clearly observed in the live differential interference contrast images of the BCR-stimulated mature B cells (Fig. 1,c, left, top and bottom panels), transitional immature B cells redistributed their BCR, but did not demonstrate the membrane ruffling response (Fig. 1 c, right, top and bottom panels). Cytochalasin D treatment of mature B cells blocked both anti-BCR-induced receptor capping and membrane ruffling, indicating that membrane ruffling was also an actin-dependent process (data not shown). Because transitional immature were capable of actin-dependent BCR capping but not actin-dependent BCR-induced membrane ruffling, these results suggest that capping and ruffling are regulated by distinct mechanisms.
Transitional immature and mature B cells differ in their ability to spread in response to substrate-bound anti-BCR
Mature B cells that contact immobilized Ag polarize their BCR and reorganize their actin cytoskeleton toward the contact site and increase their area of contact (spreading) to engage more BCR (30, 31). Because transitional immature and mature B cells differed in their ability to undergo actin-dependent membrane ruffling in response to soluble anti-BCR stimulation, we next assessed whether BCR polarization, actin cytoskeleton reorganization, and BCR-induced spreading were similarly impaired in transitional immature B cells following exposure to immobilized anti-BCR. For these assays, beads coated with an Ab directed against the BCR L chain were used to model associations with insoluble Ag and BSA-coated beads served as a negative control. Both transitional immature and mature B cells in contact with BSA-coated beads exhibited a rounded morphology and uniform distribution of BCR and F-actin (Fig. 2,a). In comparison, mature B cells polarized both BCR and F-actin toward the anti-BCR coated beads resulting in an accumulation of F-actin and BCR toward the contact point. In contrast, transitional immature B cells in contact with anti-BCR-coated beads maintained a rounded morphology and uniform distribution of both BCR and F-actin (Fig. 2 b). Therefore, similar to their inability to undergo membrane ruffling in response to soluble anti-BCR stimulation, transitional immature B cells fail to reorganize either the BCR or F-actin toward beads coated with immobilized anti-BCR Ab.
We next assessed the ability of transitional immature and mature B cells to spread on anti-BCR L chain-coated glass dishes. The extent of spreading by mature B cells plated on anti-L chain-coated glass dishes was noticeably and significantly (p < 0.0001) different from transitional immature B cells (Fig. 3,a). Live images of mature and transitional immature B cells were taken every 30 s for 10 min (Fig. 3 c). It was noted that not all of the mature B cells spread, a finding that has previously been reported in both B and T cells (32, 33).
Finally, to verify that the differences seen in anti-BCR-induced spreading between transitional immature and mature B cells were not caused by anomalous behavior of transitional immature B cells generated in the autoreconstitution system, we simultaneously visualized mature and transitional immature B cell splenocytes from nonirradiated mice in which ∼10–15% are transitional immature (T1 and T2) B cells (23) that can be identified by their expression of CD93 (recognized by the AA4.1 Ab). Consistent with the data obtained in the reconstituting system, transitional immature B cells from nonirradiated mice did not detectably spread on anti-BCR-coated plates whereas most mature B cells evidenced a statistically significant degree of morphological change and spreading (Fig. 3, d and e). Thus, the use of the autoreconstitution model does not introduce artifacts that account for the different responses of the two developmental subsets of B cells.
Anti-BCR-induced spreading, but not BCR capping, is Rac1 dependent
As Rac has been implicated in regulating cytoskeletal reorganization and B cells from mice deficient in the genes encoding Rac1/Rac2 display signaling and functional responses similar to transitional immature B cells (20), we next wished to ascertain whether development regulation of Rac activation might be responsible for the differential ability of mature and transitional immature B cells to undergo dynamic actin-dependent membrane ruffling and/or cell spreading. As primary murine B cells are not amenable to transfection studies, we used a Tat-fused Rac1 dominant-negative peptide containing the hypervariable C-terminal membrane-binding domain of Rac1 that has been shown to block membrane ruffling, cell-cell adhesion, and migration in several nonlymphoid cell lines and in T cells by inhibiting the membrane localization of endogenous Rac (34, 35) for these experiments. Mature B cells were treated with 0.5 mg/ml of the Tat-fused Rac1 dominant-negative peptide or control Tat peptide alone under conditions previously shown to facilitate 100% inclusion of the peptide (34). This concentration of peptide did not alter surface BCR expression (data not shown). Mature B cells treated with either the Tat control peptide or the Tat-fused Rac1 dominant-negative peptide were able to cap BCR after stimulation with a soluble anti-BCR (Fig. 4 a), indicating that actin-dependent BCR capping in response to soluble anti-BCR is not dependent on Rac1.
We next wanted to determine whether Rac1 was involved in BCR-induced actin-dependent cell spreading. Although mature B cells treated with either the control Tat peptide or the Tat-Rac1DN peptide exhibited similar morphology on poly-l-lysine-coated plates, clear differences were observed in the spreading responses of BCR stimulated mature B cells (Fig. 4, c and d). Specifically, although increased contact area (spreading) was evident in the BCR immunofluorescence and DIC images of control peptide-treated B cells on the anti-BCR-coated plates, this response was significantly decreased (p < 0.0001) using Tat-Rac dominant-negative peptide-treated B cells. Together, these results suggest that capping of the BCR in mature B cells in response to BCR stimulation is Rac1 independent whereas BCR-induced cell spreading is dependent on Rac1.
Elevation of transitional immature B cell membrane cholesterol levels converts their spreading and ruffling responses to a mature stage phenotype
We have previously shown that both T1 and T2 transitional immature B cell subsets maintain lower steady-state levels of membrane-associated unesterified free cholesterol than do mature B cells. Augmentation of transitional immature B cell plasma membrane cholesterol to a level approximating that of mature B cells resulted in compartmentalization of the BCR into lipid rafts and a sustained signaling response resembling that observed in mature B cells (23). As cytoskeletal reorganization in T cells appears to be lipid raft dependent (36), we next assessed whether restoration of BCR compartmentalization into lipid rafts following augmentation of membrane cholesterol levels in transitional immature B cells influences their ability to promote membrane ruffling and cell spreading.
Filipin is a polyene antibiotic that binds selectively to 3β-hydroxysterols (37) and as we have previously shown that all of the transitional immature and mature B cell cholesterol exists in the unesterified free, membrane-associated form (23), the level of filipin binding represents a valid means to compare relative levels of membrane cholesterol in these populations. Similar to our previous studies (23), the flow cytometric analysis of filipin binding presented in Fig. 5 a reveals that plasma membrane cholesterol of transitional immature B cells can be augmented to levels that approximate those of mature B cells following MβCD/cholesterol treatment. Cholesterol augmentation did not alter the surface BCR expression on transitional immature B cells (data not shown). Using these cholesterol-augmented cells, we assessed the cell spreading and membrane ruffling responses following BCR stimulation with substrate-bound and soluble anti-BCR, respectively.
Consistent with data presented above, mature B cells underwent a BCR-induced spreading response whereas control transitional immature B cells did not following stimulation with immobilized anti-BCR. In contrast, transitional immature B cells in which cholesterol levels were augmented underwent a BCR-induced spreading response in a manner virtually indistinguishable from mature B cells (Fig. 5, b and c) and which was significantly different from control transitional immature B cells (p < 0.001). Finally, as shown in Fig. 5,d, mature B cells once again showed BCR capping and membrane ruffling in response to soluble anti-BCR, while the control transitional immature B cells were capable of capping, but did not undergo membrane ruffling (Fig. 5,d). As observed in the spreading response, augmentation of cholesterol levels in transitional immature B cells resulted in BCR-induced membrane ruffling (Fig. 5 d). Together, these studies indicate that elevation of membrane cholesterol levels in transitional B cells promotes BCR-induced cell spreading and membrane ruffling, thereby converting their functional response to one more closely resembling that of a mature B cell.
The Vav/Rac1/JNK-signaling pathway is impaired in transitional immature B cells and rescued by elevation of plasma membrane cholesterol levels
Because we found that anti-BCR-induced spreading was a Rac1-dependent event, and spreading was developmentally regulated and influenced by membrane cholesterol levels, we next wanted to assess Rac1 activation in mature, transitional immature, and cholesterol-augmented transitional immature B cells. As the Rac1 GTPase is active in its GTP-bound state, we assessed Rac activation by determining the amount of GTP-loaded Rac1 after BCR stimulation with soluble anti-BCR. Mature B cells typically exhibited a 2-fold increase in GTP-loaded Rac1 5 min after BCR stimulation and maintained increased GTP-loaded Rac1 through 10 min (Fig. 5,a). By comparison, the levels of GTP-Rac1 did not increase after BCR stimulation in control transitional immature B cells, even though they expressed Rac1 and were capable of increasing GTP-loaded Rac1 (Fig. 6 a, last lane). Cholesterol augmentation of transitional immature B cells rescued the ability of these cells to GTP-load Rac1 after BCR stimulation. Although the kinetics between mature and cholesterol-augmented transitional immature B cells were not identical, they both inducibly increased GTP-loaded Rac1.
To confirm and extend the Rac1 GTP-loading results, we next looked at signaling pathways both upstream and downstream of Rac1. Vav is an early substrate for tyrosine phosphorylation after AgR stimulation (12, 38) and phosphorylated Vav serves as a guanine nucleotide exchange factor for Rac1 activation (39). Vav phosphorylation was assessed in mature, transitional immature, and cholesterol-augmented transitional immature B cells by immunoprecipitation and Western blotting (Fig. 6 b). Consistent with the Rac1 GTP-loading results, BCR stimulation of mature B cells resulted in an increase of Vav phosphorylation as early as 2 min that peaked at 15 min poststimulation. Likewise, the apparent inability of transitional immature B cells to activate Rac1 was associated with reduced Vav phosphorylation in both resting and anti-BCR-stimulated cells. Finally, cholesterol augmentation of transitional immature B cells resulted in phosphorylation of Vav at levels similar to that seen in mature B cells, indicating that restoration of Vav phosphorylation may promote the increased Rac1 activation observed in these cells.
GTP loading of Rac1 has previously been linked to phosphorylation and activation of JNK (40, 41). Therefore, we assayed JNK phosphorylation in B cells as a downstream indicator of Rac1 activation (Fig. 6,c). In mature B cells, inducible JNK phosphorylation was detected as early as 2 min and peaked at 15 min. Levels of phospho-JNK in transitional immature B cells were by comparison to mature B cells reproducibly lower. Cholesterol augmentation of transitional immature B cells resulted in phosphorylation of JNK similar to that seen in mature B cells. Taken together, these results suggest that transitional immature B cells are relatively impaired in their activation of the Vav/Rac1/JNK-signaling pathway but that augmentation of their membrane cholesterol levels allows them to engage this pathway to levels that recapitulate those observed in mature B cells. Thus, developmentally regulated membrane cholesterol levels appear to play a direct role in regulating BCR-induced responses during B cell development (Fig. 6).
A number of BCR proximal events distinguish BCR signaling in follicular mature and immature B cell subsets. To a large extent, the common theme is an inability to effectively initiate and/or sustain BCR-induced signals through the phosphatidylinositol 4,5 bisphosphate hydrolysis/PKCβ/NF-κB/c-myc pathway (23, 29, 42, 43). The strong link between this pathway and the regulation of Ag-induced mature B cell proliferation and Ag-dependent and -independent survival (9, 23, 29, 44, 45, 46, 47, 48) argues that processes associated with this signaling pathway determine the responsiveness of mature B cells to BCR signals. Together, these studies indicate that the ability to sustain signaling through this pathway determines B cell fate decisions in response to Ag. Furthermore, they suggest that the regulation of cellular processes associated with transient vs sustained signaling might determine whether a B cell response will be directed toward tolerance as opposed to activation.
We have previously shown BCR-stimulated transitional immature B cells fail to translocate the BCR into rafts after BCR cross-linking. We document here that compared with mature B cells, transitional immature B cells are relatively deficient in their ability to phosphorylate Vav and GTP-load Rac1. Kurosaki and colleagues (12) have previously shown in DT40 cell lines that Vav must be recruited into lipid raft microdomains by growth factor receptor bound protein-2 and/or B cell linker protein for anti-BCR-induced Vav phosphorylation and GTP loading of Rac1 to occur. Given this finding, we argue that the impaired Vav/Rac1 signaling phenotype of the transitional immature B cells is due to the demonstrated inability of the transitional immature BCR to stably localize to lipid rafts as a consequence of developmentally regulated lower levels of unesterified, membrane-associated cholesterol. This failure to localize to rafts contributes to impaired Vav/Rac activation, defective cytoskeletal reorganization and inability to sustain signaling pathways necessary for survival and activation and is consistent with the finding that the BCR-induced functional responses of B cells from either Vav1/Vav2- or Rac1/Rac2-null B cells resemble those of transitional immature B cells (18, 19, 20). However, we readily acknowledge that membrane cholesterol levels may influence the differential responsiveness of transitional immature and mature B cells by affecting the composition and/or endocytosis of the BCR-signaling complex (49) or altering gene expression through sterol-regulated transcription pathways (50) and future studies will address these possibilities.
Although we are keenly interested in the relative importance of developmentally regulated membrane cholesterol levels in maintaining B cell tolerance, the MβCD/cholesterol add back method used above is not suitable for long-term functional assays as continuous exposure is detrimental to the cells and removal of MβCD/cholesterol results in the gradual reduction of membrane cholesterol to baseline levels. Although we have attempted apoptosis assays following washout of MβCD/cholesterol, the results were inconclusive. We believe that the apparent inconsistency of these assays is related to the rate at which membrane cholesterol is lost from the transitional immature B cells in an individual experiment. However, we are encouraged by our finding that transitional immature B cells externalize phosphatidylserine to a much greater extent than do either mature B cells or cholesterol-augmented transitional immature B cells following BCR stimulation in vitro. Although these data are consistent with the idea that membrane cholesterol levels influence B cell tolerance, phosphatidylserine externalization has been associated with both apoptosis and positive selection in B cells, complicating the interpretation of such a finding. Currently, we are developing a more stable genetic approach to test the tolerance sensitivity of cholesterol-augmented transitional immature B cells in vivo and which will provide a more definitive assessment of the role of developmentally regulated membrane cholesterol levels in the initiation and/or maintenance of B cell tolerance.
The ability of both B and T lymphocytes to undergo anti-AgR-induced spreading has received much attention lately (31, 32). Batista and colleagues (31) recently identified an actin-dependent spreading and contraction response in B cells that allows B cells to increase surface area to gather Ag during the spreading phase and internalize Ag during the contraction phase. Using live cell imaging, we observed BCR-induced spreading of mature B cells but not of transitional immature B cells. We have previously shown that transitional immature B cells do not completely activate CD4 T cells (51) and it is tempting to speculate that the relative inability of transitional immature B cells to undergo BCR-induced spreading correlates with their inability to completely activate CD4 T cells.
A recent study by Blery et al. (49) also addressed a role for membrane cholesterol in B cell responses to BCR signaling. In these studies, MβCD-mediated lowering of membrane cholesterol levels in anergic B cells resulted in decreased Ag-induced internalization and sustained BCR signaling. Their results, suggesting that lower cholesterol levels lead to sustained BCR signaling, appear at first inconsistent with our studies. However, in their studies steady-state cholesterol levels were indistinguishable between naive, Ag-responsive and anergic, Ag-unresponsive B cells, arguing that normal steady-state cholesterol levels are not responsible for the differential responsiveness of normal and anergic B cells. In the present study, we compared populations of B cells that differ not only in their responsiveness to BCR signaling but also in their steady-state membrane cholesterol levels. Although it is difficult to directly compare the two studies, the differences observed in cholesterol dependence for sustaining BCR signaling in Ag-exposed anergic B cells and naive but unresponsive transitional immature B cells warrants further study.
We believe that the significance of our current studies extends well beyond the characterization of Vav and Rac1 activation and their role in cell spreading in developing B cells. Our results reveal multiple levels of actin cytoskeleton reorganization during the initial phases of BCR signaling that are each regulated by and linked to different cellular processes. Both transitional immature and mature B cells have the ability to undergo actin-dependent, Rac1-independent capping of the BCR following BCR stimulation. In contrast, transitional immature B cells are relatively defective in the Rac1-dependent cytoskeletal reorganization leading to membrane ruffling and spreading after BCR stimulation as compared with mature B cells, but this impairment can be reversed by augmenting membrane cholesterol levels. It was unexpected that the induction of membrane ruffling/cell spreading would be so tightly correlated with the signaling and cell-fate decisions that have been reported by us and others to characterize the differential responsiveness of immature and mature stage B cells (23, 52). Further studies will determine whether these specific differences in proximal signaling and their apparent tight link to membrane cholesterol levels play a determining role in the differential sensitivity of immature and mature B cells to Ag-specific tolerance induction.
We thank Xinyu Zhao at the Penn Imaging Core facility for help with confocal microscopy and Fredrick G. Karnell with FACS analysis. We also thank Drs. Leslie King, Richard Siegel, David Allman, Jan Burkhardt, and Shannon Grande for scientific insights and critical reading of the manuscript.
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 National Institutes of Health Grants CA93615 and AI32592 (to J.G.M.). R.J.B. was supported by funding from the National Cancer Institute.
Abbreviations used in this paper: PLC, phospholipase C; PKC, protein kinase C; MβCD, methyl-β-cyclodextrin; DAPI, 4′,6′-diamidino-2-phenylindole; DIC, differential interference contrast.