• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of hhmipaAbout Author manuscriptsSubmit a manuscriptHHMI Howard Hughes Medical Institute; Author Manuscript; Accepted for publication in peer reviewed journal
Nat Immunol. Author manuscript; available in PMC Jul 29, 2008.
Published in final edited form as:
Published online Nov 25, 2007. doi:  10.1038/ni1542
PMCID: PMC2488964

Follicular shuttling of marginal zone B cells facilitates antigen transport


The splenic marginal zone is a site of blood flow and the specialized B cell population that inhabits this compartment has been implicated in the capture and follicular delivery of blood-borne antigens. However, the mechanism accounting for this antigen transport has been unknown. Here we show that marginal zone B cells were not confined to the marginal zone but continuously shuttled between the marginal zone and follicular areas such that many of the cells visited a follicle every few hours. Migration to the follicle required CXCR5 whereas return to the marginal zone was promoted by S1P1 and S1P3. Treatment with an S1P1 antagonist caused marginal zone B cell displacement from the marginal zone. Marginal zone-follicular shuttling of marginal zone B cells provides an efficient mechanism for systemic antigen capture and delivery to follicular dendritic cells.

The spleen is a major site for the induction of antibody responses against blood-borne antigens. Many B cells in the spleen are situated in follicles within the white pulp, where they migrate over the processes of follicular dendritic cells (FDCs) in search of antigen. FDCs are also a source of B lymphocyte chemoattractant (BLC or CXCL13), a chemokine that attracts B cells to follicles by engaging the receptor CXCR51. Marginal zone B cells constitute a second major B cell population in the spleen and are so-named because of their location in the marginal zone between the white and red pulp areas2,3. The marginal zone is separated from the white pulp by the marginal sinuses, sites where terminal arterioles open and release blood4. The outer borders of the sinuses are porous and blood can pass through the marginal zone before reaching the red pulp and returning to circulation via venous sinuses. These features ensure that cells situated in the marginal zone are readily exposed to blood-borne antigens2,3. Marginal zone B cells have a unique surface phenotype, expressing high amounts of the complement receptors CD21 (CR2) and CD35 (CR1) and the non-classical major histocompatibility complex (MHC) molecule, CD1d3. In rodents, marginal zone B cells are restricted to the spleen and do not recirculate2.

The pathways by which blood-borne antigens are delivered to splenic follicles have been under investigation for more than 30 years5-7. Several studies have shown that marginal zone B cells rapidly capture complement-opsonized antigens, such as Ficoll, via CD21 and CD35 (refs.5-7). Treatment with lipopolysaccharide (LPS) or pertussis toxin (PTX) causes displacement of marginal zone B cells from the marginal zone and this finding was correlated with an immediate deficit in the ability of injected antigens to become deposited on FDCs8-10. These studies implicated marginal zone B cells in antigen transport but did not exclude the alternative possibility that other LPS or PTX-sensitive cell types were required. The uncertainty about the contribution of marginal zone B cells to antigen transport has been amplified by the inability to detect reductions in marginal zone B cell numbers within the marginal zone following exposure to opsonized-antigens8,10.

Marginal zone B cell positioning in the marginal zone depends on the sphingosine-1-phosphate (S1P) receptor, S1P1 (ref.11) encoded by endothelial differentiation gene-1 (Edg1). Marginal zone B cells that lack S1P1 or have been exposed to the S1P1 modulating drug, FTY720, fail to position in the marginal zone and are instead found inside the follicles11,12. S1P is abundant in blood and is thought to be present in much lower amounts in the follicles and T zones of lymphoid tissues13,14. However, some studies have questioned the view that the S1P1 requirement directly implicates a role for S1P in cell positioning and have instead suggested that the role(s) of this ligand-receptor pair in lymphocyte function may be developmental15,16. Also complicating the picture are experiments with T cells showing that continual exposure to blood S1P causes down-modulation of S1P1 (refs.14,17), making it unclear if marginal zone B cells could continually respond to plasma S1P in the marginal zone.

Here we provide evidence, using a small molecule antagonist, that engagement of S1P1 by ligand is continually required for marginal zone B cell positioning in the marginal zone. Using mice treated with FTY720 or lacking S1P1 in B cells we then show that marginal zone B cell positioning in the marginal zone was necessary for capture of blood-borne antigens and for their efficient deposition on FDCs. However, if these cells lacked CXCR5 and were not able to migrate into follicles, antigen was no longer transported to FDCs. Using an in vivo antibody labeling procedure, we find that in the resting state up to half of the marginal zone B cells were located in the follicle. By comparing the amount of antibody labeling after 5 and 20 min we obtained evidence that marginal zone B cells constitutively shuttled between marginal zone and follicle even in the absence of immunization and independent of B cell receptor (BCR) specificity or complement receptor expression. Follicle to marginal zone shuttling was regulated by the balance of CXCR5 and S1P1 abundance. Finally, we demonstrate that S1P3 contributed to the efficient positioning of marginal zone B cells in the marginal zone.


S1P1 antagonists displace marginal zone B cells

To test whether S1P1 engagement by ligand was a constitutive requirement for marginal zone B cell positioning we treated mice with the S1P1 antagonist, VPC44116 (ref.18). Mice were treated with three doses of VPC44116 or equivalent volumes of carrier over 3 h and tissues were then isolated for histological sectioning or flow cytometric analysis. T cells from blood, spleen and lymph nodes of treated animals had upregulated S1P1, establishing that sufficient antagonist was injected to reduce ligand engagement of the receptor (Fig. 1a).

Figure 1
Treatment with S1P1 antagonist VPC44116 for 3 h causes displacement of marginal zone B cells into follicles. (a) FACS histograms of S1P1 on naive CD4+ T cells (CD4+ CD62Lhi) in blood, spleen, and peripheral lymph node (PLN) after treatment with VPC44116 ...

The less complete upregulation of S1P1 on cells in blood compared to lymphoid tissues most likely reflects incomplete antagonism of receptor engagement by the high amounts of S1P in blood13. Although the S1P1 antibody preparation specifically stains wild-type but not S1P1-deficient T cells19, extensive efforts to stain B cells were unsuccessful due to high non-specific background staining on this cell type (data not shown). Analysis of tissue sections from mice treated for 3 h with VPC44116 showed that marginal zone B cells had relocalized into the follicles (Fig. 1b). These data suggest that marginal zone B cell localization in the marginal zone is promoted by S1P signaling through S1P1 in these cells.

Marginal zone positioning facilitates antigen capture

As an approach to test the importance of marginal zone positioning in capture of blood-borne antigen, mice were pretreated with FTY720 to cause displacement of marginal zone B cells into follicles11 and then injected i.v. with 5 or 50 μg of trinitrophenol (TNP)-Ficoll and analyzed after 30 min. In saline treated control mice, marginal zone B cells bound substantially greater amounts of TNP-Ficoll than follicular B cells and this binding was CD21-CD35 dependent (Fig. 2a and data not shown) as expected20. By contrast, in mice pretreated with FTY720, the amount of TNP-Ficoll bound to the majority of marginal zone B cells was minimal, similar to the amounts seen on follicular B cells (Fig. 2a). The inclusion of allelically marked ex vivo control B cells established that there was minimal in vitro capture of TNP-Ficoll complexes when mice had been injected with 5 μg of TNP-Ficoll and this amount was used in all subsequent experiments. To further test if the reduced TNP-Ficoll binding to marginal zone B cells was due to the redistribution of cells into follicles rather than an indirect effect of FTY720 we tested the impact of FTY720 on Ficoll binding in CXCL13-deficient mice. In these animals, FTY720 treatment does not cause displacement of marginal zone B cells from the marginal zone11 and it had minimal effect on Ficoll binding (Fig. 2b). Similarly, FTY720 did not affect TNP-Ficoll binding by circulating B cells in the blood (Fig. 2c).

Figure 2
Marginal zone B cell positioning in the marginal zone is required for efficient binding of blood-borne antigen. Mice were injected i.v. with TNP-Ficoll and 30 min later tissues were analyzed. (a-e) Flow cytometric analysis of TNP-Ficoll binding to splenic ...

A concern with the use of FTY720 to study access to a blood-borne antigen was the known activity of FTY720 in reducing vascular permeability21. We therefore performed further experiments using S1P1-deficient fetal liver chimeras. As expected11, the majority of marginal zone B cells in S1P1-deficient animals failed to locate in the marginal zone but instead could be found within follicles (data not shown). Similarly to the results obtained by FTY720 treatment, 30 min after intravenous TNP-Ficoll injection, marginal zone B cells in S1P1-deficient animals bound little TNP-Ficoll (Fig. 2d). To confirm that the effect of S1P1-deficiency was due to loss of S1P1 in B cells, we performed experiments in Edg1flox/- mice22 that had been crossed with Cd19-cre mice23. Quantitative PCR analysis on sorted cells indicated that the Cre-mediated ablation of Edg1 in marginal zone B cells was efficient and immunohistochemical staining showed that the lodgment of marginal zone B cells in the marginal zone was reduced to an extent similar to the reduction in Edg1-/- fetal liver chimeras (data not shown). Again, the marginal zone B cells bound little TNP-Ficoll (Fig. 2e). Immunohistochemical analysis of tissue sections showed substantial amounts of TNP-Ficoll overlapping with the B cell distribution in the splenic marginal zone of control mice (Fig. 2f). In the S1P1-deficient animals there was still accumulation of TNP-Ficoll at the marginal zone, but in a more punctate pattern, perhaps reflecting association with marginal zone macrophages (Fig. 2f). Taken together, these observations suggest that marginal zone B cell localization at the marginal zone is essential for their ability to rapidly capture blood-borne antigen.

Marginal zone B cells transport antigen to follicles

Within 3 h of intravenous injection, Ficoll is found in large amounts on splenic FDCs. To test whether marginal zone B cells must be present in the marginal zone for this transport to occur, we assessed the extent of TNP-Ficoll deposition in follicles of mice pretreated with FTY720 or in S1P1-deficient fetal liver chimeras. Both FTY720 pretreatment and S1P1 deficiency were found to reduce, though not completely block, TNP-Ficoll deposition on FDCs, compared to the deposition seen in matched controls (Fig. 3a,b). TNP-Ficoll deposition was similarly reduced in Cd19-cre Edg1flox/- mice (Fig. 3c), confirming that B cell expression of S1P1 is required for efficient deposition of opsonized-Ficoll on FDCs.

Figure 3
Marginal zone B cell positioning in the marginal zone is required for efficient delivery of blood-borne immune complexes to FDCs. (a-c) IHC micrographs of spleens from mice that were injected with 5 μg TNP-Ficoll for 3 h. Sections were stained ...

While these experiments supported the conclusion that marginal zone B cells need to locate in the marginal zone to facilitate transport of a blood-borne antigen to FDCs, it remained possible that their role was solely in capture of the blood-borne antigen rather than in its transport. To further test this issue we used a mixed bone marrow chimera approach to generate mice in which follicles could develop normally but where marginal zone B cells would be unable to migrate from the marginal zone into follicles. In mice lacking CD19, follicles form, but marginal zone B cells fail to develop whereas in CXCR5-deficient mice, B cells develop normally but are unable to migrate into follicles and the B cell areas lack FDCs24-26. When mice that had been reconstituted with a mixture of CD19-deficient and CXCR5-deficient bone marrow were injected with TNP-Ficoll, very effective antigen-capture by CXCR5-deficient marginal zone B cells was observed but there was little or no deposition of this antigen on FDCs (Fig. 3d). Normal Ficoll deposition occurred in mixed bone marrow chimeric mice containing CXCR5-wildtype B cells (Fig. 3d). Staining with antibodies to CD35 confirmed the presence of FDCs in both types of mixed bone marrow chimeric mice (Fig. 3d). Taken together, these experiments provide strong evidence that marginal zone B cells not only capture systemic antigens in the marginal zone, but they also migrate in a CXCR5-dependent manner to follicles to facilitate antigen deposition on FDCs.

Folliclular shuttling of marginal zone B cells

Despite the efficiency of the TNP-Ficoll transport detected in wild-type mice, we were not able to observe a change in the number of B cells in the marginal zone following TNP-Ficoll injection (Fig. 4a and data not shown for earlier time points). However, we had repeatedly noted in CD1d-stained sections of control spleens that there were cells staining with the same intensity as marginal zone B cells located inside splenic follicles (Fig. 4a and data not shown). These combined observations led us to consider the possibility that there was constitutive migration of marginal zone B cells into follicles occurring in a manner independent of complement receptor CR1/2-mediated antigen capture. To examine this possibility further we needed a more quantitative method of assessing the number of B cells present in the marginal zone. Since this compartment is rapidly exposed to blood-borne antigen we asked whether an antibody pulse-labeling procedure developed to label cells in the vascular compartment (J. Pereira and J.G.C., manuscript in preparation) could be applied and tested whether a brief in vivo exposure to a CD21 antibody would selectively label marginal zone B cells amongst splenic B cells. Flow cytometric analysis of spleen cells from mice that had been treated with anti-CD21-phycoerythrin (PE) for 5 min revealed strong staining on a fraction of the marginal zone B cells (Fig. 4b). Microscopy analysis revealed intense staining on B cells in the MZ (Fig. 4c). Importantly, the flow cytometric analysis showed that the marginal zone B cell staining was bimodal with only ~50% of the cells becoming brightly labeled (Fig. 4b). We speculated that the poorly labeled cells corresponded to cells that had been shielded from the injected antibody because they were present within follicles during the 5 min treatment. Like lymph nodes, the white pulp cords of spleen are surrounded by stromal cells and associated extracellular matrix that together form a barrier restricting free diffusion of molecules the size of antibodies27-29. Follicular B cell staining was lower than on marginal zone B cells as expected though there was a tendency for some of the cells to show positive staining (Fig. 4b), consistent with the presence of recirculating follicular B cells in the splenic marginal zone and red pulp30. To test whether follicular localization shielded marginal zone B cells from exposure to the injected antibody, we treated S1P1-deficient mice with anti-CD21-PE. In these animals, the majority of marginal zone B cells were poorly stained following the 5 min antibody injection (Fig. 4d,e). A similarly low fraction of cells became CD21-PE labeled in mice that had been pretreated for 3 h with FTY720 (Fig. 4e). We then injected CXCR5-deficient mice with anti-CD21-PE and found that all the marginal zone B cells became labeled in 5 min (Fig. 4f). The staining intensity on these cells was lower than on the brightly stained subset in wild-type controls, perhaps because there are many more B cells located in the marginal zone of mice deficient in the CXCL13-CXCR5 axis compared to control mice26. To address the concern that white pulp organization is disturbed in CXCR5-deficient mice we also treated bone marrow chimeras that had been reconstituted with a mixture of CXCR5-deficient and S1P1-deficient bone marrow. These mice have well-developed lymphoid follicles but the CXCR5-deficient B cells are unable to access these areas. In these animals, all the CXCR5-deficient marginal zone B cells became rapidly labeled whereas only few S1P1-deficient marginal zone B cells became labeled (Fig. 4g) consistent with the conclusion that cells located within the marginal zone become heavily labeled whereas cells present within the follicle are shielded from immediate exposure to the CD21-PE.

Figure 4
In vivo antibody labeling provides evidence for marginal zone-follicular shuttling of marginal zone B cells in the absence of immunization. (a) IHC micrographs of spleen sections from non-treated mice, or mice that were immunized with TNP-Ficoll for 3 ...

We then asked whether there was a change in the fraction of brightly labeled cells after 20 min and found at that time point ~70% of marginal zone B cells were highly labeled in wild-type mice (Fig. 4e, h). After longer time periods, all marginal zone B cells became antibody-labeled, consistent with a study that examined labeling 1 h after anti-CD21-biotin injection29. Importantly, in Cd19-cre Edg1flox/- mice the majority of marginal zone B cells remained poorly labeled at 20 and even at 40 min, indicating that the increased labeling was not a result of free antibody access to the follicle (Fig. 4e, i). The brightly stained cells that could be detected in these mice at 20 min appeared to correspond to a small number of cells that were residing within the marginal zone (Fig. 4c). Reciprocally, in CXCR5-deficient mice, all the marginal zone B cells were brightly labeled, consistent with all of these cells being in the marginal zone and accessible to the antibody (Fig. 4c,j). Together, these observations are consistent with the view that as many as 40% of the marginal zone B cells that were in the follicle of wild-type mice during the initial 5 min of labeling travel back into the marginal zone during the next 15 min. If this were the case, it would be anticipated that these cells would be replaced by CD21-PE labeled cells migrating into the follicle from the marginal zone. Microscopy analysis of splenic tissue sections taken 20 min after antibody injection showed that in addition to staining throughout the marginal zone there was now also labeling of a population of B cells within the follicle (Fig. 4c). By contrast, there were few CD21-brightly stained cells in splenic sections from 20 min treated S1P1-deficient mice (Fig. 4c) ruling out the possibility that the brightly staining cells arose due to diffusion of the antibody. To exclude the possibility that the brightly staining cells appearing within follicles after 20 min were all recirculating B cells, we also treated Cd19-/- mice that lack marginal zone B cells but have normal numbers of follicular B cells23,24. This analysis indicated a lack of brightly stained cells (Fig. 4c). In wild-type mice, similar findings regarding marginal zone B cell labeling were made after injection of anti-CD1d-PE, B220-PE, CD35-biotin, CD21-fluorescein isothiocyanate (FITC), CD19-PE, and CD45.2-PE (data not shown) indicating that the shuttling was not dependent on engagement of CR1/2 by the CD21 antibody, nor were these observations notably affected by the nature of the antibody or the type of conjugation. The marginal zone B cell shuttling did not depend on expression of CR1/2 as similar labeling profiles were observed in Cr2-/- mice at 5 and 20 min after anti-CD45.2-PE injection (Fig. 4k). In addition, marginal zone B cell shuttling seems to be independent of their BCR specificity. In Ig-heavy chain knockin animals in which about half of the B cells are hen egg lysozyme (HEL) specific (VDJ mice), marginal zone B cells were similarly distributed between the marginal zone and follicles whether they were HEL specific or not (Fig. 4l). As a further approach to test for rapid migration of marginal zone B cells into the follicle we asked whether any marginal zone B cells that labeled with antibody during an initial pulse became protected from exposure to a subsequently injected antibody. Indeed, a fraction of the marginal zone B cells that had labeled with anti-CD35 during a 15-min pulse were protected from labeling by anti-CD21 in the ensuing 5 min (Fig. 4m, upper left quadrant in right panel). The lower CD35 antibody staining on the protected (CD21-PE negative) cells was observed in repeated experiments and might indicate that after migration into follicles there is rapid loss of some antibody to surrounding B cells or FDCs. In control mice treated with both antibodies simultaneously, the labeled marginal zone B cells bound both antibodies equally, as expected (Fig. 4m, left panel). Taken together, these observations indicate that many marginal zone B cells are rapidly shuttling between marginal zone and follicle in a homeostatic manner that is independent of complement receptor or BCR engagement.

Actived C4 on marginal zone B cells

The antibody, FDC-M2, has been widely used as an FDC marker and is now known to bind an activated form of complement C4 (ref.31). In tumor necrosis factor (TNF)-deficient mice, animals that have poorly developed follicles and lack FDCs, there is an accumulation of FDC-M2 staining cells in the marginal zone32. Similar findings have been made for CXCR5-deficient mice33. We confirmed the marginal zone FDC-M2 staining pattern in CXCR5-deficient mice (Fig. 5a) and observed similar staining in CXCL13-deficient mice (not shown). Although it was suggested that these cells may be mislocalized FDCs or FDC precursors32,33 we speculated that in the absence of marginal zone-follicular shuttling, marginal zone B cells may be unable to off-load complement-coated antigens on FDCs and they might therefore accumulate C4-containing antigens over time. Indeed, by flow cytometric analysis we found that the marginal zone B cells in CXCR5-deficient mice had significantly higher amounts of active C4 on their surface (Fig. 5b,c). Interestingly, CXCR5-deficient marginal zone B cells also had slightly elevated surface CD21 expression (Fig. 5c). However, the fold difference in FDC-M2 staining was at least 3-fold greater than the difference in CD21 expression suggesting that the marginal zone B cells in unimmunized mice have a propensity to accumulate C4-containing immune complexes. Therefore, marginal zone B cell shuttling appears to have an homeostatic role in depositing complement opsonized antigens to FDCs.

Figure 5
FDC-M2 accumulation on marginal zone B in CXCR5-deficient mice. (a) IHC micrographs of FDC-M2 staining in CXCR5 wild-type (WT) and CXCR5-deficient (KO) mice. Spleen sections were stained for FDC-M2 (dark blue) and MAdCAM-1 (brown). Scale bar, 50 μm. ...

S1P1 and CXCR5 balance regulates shuttling

To test whether the balance of S1P1 and CXCR5 activity was important in regulating the shuttling behavior observed above, we performed flow cytometric analysis of B cells from antibody injected S1P1 or CXCR5 heterozygous mice. In Edg1+/- mice treated for 5 and 20 min with CD21-PE, fewer cells were brightly labeled than in controls (Fig. 6a,c) indicating a shift in the balance in favor of follicular positioning. Although we have not been able to quantitate S1P1 surface protein amounts on B cells due to high background staining, Edg1+/- T cells have about half the normal amounts of surface S1P1 of wild-type T cells17. Reciprocally, in mice heterozygous for the gene encoding CXCR5, Blr1, a greater fraction of the cells were CD21-PE labeled in 5 min than in wild-type controls and this was also the case at 20 min (Fig. 6b,c) indicating that a 50% reduction in CXCR5 protein abundance led to a shift in the balance of responsiveness in favor of S1P signaling and positioning in the marginal zone. These observations reveal a delicate balance of signaling by S1P1 and CXCR5 and provide support for a model where the principal function of each receptor in marginal zone B cells is to attract or retain the cells in the reciprocal compartments.

Figure 6
Balance of S1P1 and CXCR5 responsiveness, and S1P3, control marginal zone B cell shuttling. (a, b) Flow cytometric analysis of marginal zone B cell labeling by injected antibodies. (a) Chimeras reconstituted with Edg1+/+ and Edg1+/- BM, and (b) CXCR5 ...

S1P3 is required for marginal zone B cell positioning

Although S1P1 deficiency and FTY720 treatment led to a marked depletion of marginal zone B cells from the marginal zone and accumulation in the follicle, the in vivo CD21-PE labeling approach revealed that there was a residual population of B cells located in the marginal zone that were detectable by flow cytometric and immunofluorescence analysis (Fig. 4c,d). Previous work had shown that S1P3, encoded by endothelial differentiation gene-3 (Edg3), is abundant in marginal zone B cells and supported strong in vitro S1P chemotactic responsiveness11 though the distribution of marginal zone B cells in S1P3-deficient animals appeared little affected11,34. Because the above findings indicated that many marginal zone B cells are rapidly shuttling between follicle and marginal zone, it seemed likely that the chemotactic function of S1P3 may be involved in the return of marginal zone B cells from the follicles to the marginal zone. For these studies we used S1P3-deficient bone marrow chimeras to exclude possible roles of S1P3 in stromal cells34. When we quantified the S1P3-deficient marginal zone B cells that became rapidly labeled following CD21-PE or CD35-biotin antibody treatment, the fraction was lower than in control animals (Fig. 6d,e). At 20 min the extent of labeling of S1P3-deficient and control cells was similar (Fig. 6e) perhaps reflecting a less accurate quantitation of cell distribution at this time point due to the elevated staining on all B cells. These observations suggest that at homeostasis, a greater fraction of the S1P3-deficient marginal zone B cells are located inside the follicle than in the marginal zone compared to wild-type mice. Moreover, when S1P3-deficient bone marrow chimeras were treated with FTY720 and then tested for marginal zone B cells remaining in the marginal zone, a more complete displacement of the cells occurred (Fig. 6d,e). Interestingly, the treatment of S1P3-deficient mice with FTY720 did not fully block marginal zone B cell shuttling; compared to the minimal labeling with the injected antibody at 5 min, more marginal zone B cells became labeled at 20 min (Fig. 6d). In summary, both S1P1 and S1P3 promote marginal zone B cell positioning in the marginal zone and an additional ligand-receptor system may exist that makes a similar contribution to marginal zone B cell positioning as S1P3.


Marginal zone B cells have long been thought to be a sessile cell population that localizes selectively in the splenic marginal zone. The above findings provide strong evidence that this is not the case and that instead the cells are shuttling continuously back-and-forth between marginal zone and follicle. The propensity to shuttle into the follicle is promoted by CXCR5 whereas shuttling to the marginal zone is mainly controlled by S1P1 and S1P3. The relative amount of these receptors is important in determining the propensity for the cells to dwell in one compartment versus the other. Our findings indicate that marginal zone B cell shuttling is important for the efficient capture and transport of complement-bound antigen from the blood to splenic FDCs.

Taking our observations together with findings in several other studies, we propose the following model to account for marginal zone-follicular shuttling of marginal zone B cells. The high abundance of S1P in blood and the blood-rich nature of the marginal zone make it likely that the marginal zone is an S1P-high environment. Our demonstration that antagonism of S1P1 ligand binding rapidly displaces marginal zone B cells is consistent with this conclusion. By contrast, the lymphoid regions of lymphoid organs are S1P-low environments13. CXCL13 is abundant in follicles and transcript expression extends to the marginal sinus but not into the marginal zone35,36. Marginal zone B cells express 2-3 fold higher amounts of S1P1 transcripts and similar amounts of CXCR5 transcripts to follicular B cells (11 and data not shown). Over-expression of S1P1 in follicular B cells is sufficient to exclude them from migrating into splenic follicles in short term transfer experiments17. We propose that upon commitment to a marginal zone B cell fate and upregulation of S1P1 and S1P3, marginal zone B cells are attracted to the marginal zone where they can adhere and resist the shear forces associated with local blood flow due to their high expression of αLβ2 and α4β1 (refs.37,38). However, we speculate that after a short period they are exposed to a sufficient amount of S1P to down-modulate S1P1 and no longer be efficiently retained in the marginal zone. Although reagents are not yet available to stain mouse B cells for S1P1, studies with T cells and transfected cell lines have shown that S1P1 is highly sensitive to down-modulation after S1P exposure13,39-41. Following S1P1 down-modulation the marginal zone B cell response to CXCL13 dominates and the cells migrate to the follicle. Within the follicle, FDCs are thought to successfully compete for binding of opsonized antigens carried by the migrating marginal zone B cells due to their exceptionally high expression of CD21 and CD35. After a short period of migration within this CXCL13-high but S1P-low environment we suggest that marginal zone B cells regain sufficient S1P1 surface expression to respond again to marginal zone S1P and now migrate back to this compartment. When T cells migrate from blood into lymph nodes they upregulate S1P1 within 30-60 min17. The sensitive relationship between S1P1 and CXCR5 surface abundance and propensity of marginal zone B cells to locate in the marginal zone or follicle suggests that even partial decreases or increases in receptor function might be sufficient to promote movement of the cell from one compartment to the other. Repeated cycles of this process would lead to constitutive marginal zone B cell shuttling between marginal zone and follicle. In future work it will be important to further test this model by disrupting desensitization motifs in S1P1 and by developing approaches to image marginal zone B cell migration in real-time within the intact spleen.

The potential involvement of marginal zone B cells in the transport of opsonized antigens into follicles has long been under investigation5-7. Early electron microscopy studies identified rare examples of splenic (possibly marginal zone) B cells with labeled antigen at their uropod that may have been migrating in the direction of the follicle42. Correlative evidence that marginal zone B cells were needed for antigen transport was obtained by performing treatments that displace the cells and showing that this reduced subsequent deposition of injected antigen on FDCs8-10. An anti-CD21 injection study also observed a correlation between initial labeling of marginal zone B cells and subsequent antibody appearance on FDCs29. However, these studies could not exclude the possibility of indirect effects of the treatments and they did not establish how following immune complex or antibody capture the marginal zone B cells underwent repositioning into the follicle. Another experiment showed that CD19-deficient mice were defective in transport of immune complexes to splenic FDCs, consistent with a role for marginal zone B cells but not excluding the possibility that these cells performed only the capture and other cells performed the transport10. Our work builds from these findings to directly demonstrate that marginal zone B cells must be located in the marginal zone to efficiently capture opsonized antigen and that their transport of the antigen may then be a default feature of their shuttling behavior rather than being induced by complement receptor engagement.

Our finding that the fraction of marginal zone B cells labeled by anti-CD21 antibody injection increases from ~50% to ~70% in 15 min suggests that as many as 40% of the marginal zone B cells that were within follicles during the first 5 min redistribute back to the marginal zone in 15 min (corresponding to ~20% of total marginal zone B cells). At this rate of flux, the majority of marginal zone B cells would shuttle in and out of the follicle over a period of hours. Consistent with this, we find that compared to the heavy loading of the entire marginal zone B cell population with TNP-Ficoll 30 min after injection, the cells have all largely been cleared of surface complexes by 2-3 h. This clearance appears to depend on shuttling as marginal zone B cells in CXCR5- and CXCL13-deficient mice fail to become cleared of surface labeling over this time frame. Indeed, it would appear that constitutive shuttling contributes in an ongoing fashion to removal of opsonized antigens from the surface of marginal zone B cells since marginal zone B cells in CXCR5- and CXCL13-deficient mice have elevated amounts of activated C4 on their surface even in the absence of intentional immunization. It will be of interest in future studies to determine what fraction of these constitutive complexes contain autoantigens as this might implicate marginal zone-follicular shuttling and FDC loading with blood-borne antigen as a mechanism promoting self-tolerance of follicular B cells.

Although we found that there was a marked defect in antigen deposition on FDCs when marginal zone B cells were defective in shuttling to the marginal zone due to S1P1-deficiency (or after FTY720 treatment), the antigen deposition was not completely prevented. This was also the case in FTY720 treated S1P3-deficient mice, the conditions where we observed the most severe defect in marginal zone B cell shuttling. We speculate that the residual transport and deposition of opsonized antigen on splenic FDCs is mediated by recirculating follicular B cells. Blood B cells were able to rapidly capture low amounts of injected Ficoll and recent work has shown that follicular B cells in lymph nodes capture and transport immune complexes in a CD21-dependent manner to FDCs43.

The strong chemotactic activity of the S1P3 receptor in marginal zone B cells made it surprising that S1P3-deficient mice and bone marrow chimeras showed little evidence of impairment in B cell positioning in the marginal zone when examined non-quantitatively by immunohistochemistry11,34. However, using quantitative flow cytometric assessment we have now been able to reveal a contribution of this receptor to marginal zone B cell positioning in the marginal zone. It remains puzzling that in chemotaxis assays S1P3 is more active than S1P1 yet in vivo, S1P1 plays a greater role in promoting cell positioning in the marginal zone. One possible explanation for this discrepancy is that in vitro assays are poor at reporting the in vivo promigratory activity of S1P1, perhaps because of the high sensitivity of S1P1 to down-modulation by ligand. In this regard, in preliminary experiments with transfected cells we find that S1P3 is more resistant to ligand-mediated down-modulation than S1P1 (data not shown). In contrast to the activity of FTY720 in down-modulating and promoting the degradation of S1P1 (refs.40,41), cells taken from mice injected with FTY720 retained S1P3 function11. Thus the more complete displacement of marginal zone B cells by FTY720 in S1P3-deficient mice than in wild-type mice is consistent with FTY720 not inactivating S1P3. It is notable that even S1P3-deficient mice retain small numbers of marginal zone B cells in the marginal zone after FTY720 treatment, suggesting that at least one further mechanism functions to promote localization of marginal zone B cells in this zone. Finally, it was recently shown that shear forces are important for integrin activation of T cells by chemokines38. As the marginal zone is constantly flushed by blood and integrins are vital for marginal zone B cell lodgement at the marginal zone37 it will be important to study the possible role of shear forces in promoting marginal zone B cell adhesion as the cells shuttle into this zone.


Mice, fetal liver and bone marrow chimeras

Wild-type C57BL/6 (B6) and congenic B6 CD45.1 mice 2-4 months old were from the National Cancer Institutes. Edg1+/- and Edg1-/- fetal liver (FL) chimeras were generated as described11. Edg1+/- mice44 were crossed for 11 generations or more to B6. Cd19-cre mice23 on a B6 background were crossed to Edg1+/- mice. Insertion of cre disrupts the Cd19 coding sequence, leading to CD19 deficiency in the homozygous situation (Cd19-cre/cre, also referred to as Cd19-/-). Cd19-cre/cre Edg1+/- mice were crossed to mice in which the Edg1 alleles are flanked by loxP (Edg1flox/flox)22. This resulted in Cd19-cre Edg1flox/+ or Cd19-cre Edg1flox/- mice in which the B cell lineage has either one or no functional Edg1 alleles, respectively. To generate bone marrow (BM) chimeras, lethally irradiated B6 CD45.1 mice were reconstituted with: BM from Edg1+/- FL chimeras that was mixed 95:5 with BM from CXCR5-deficient (Blr1-/-) CD45.1 mice; BM from Cd19-/- mice that was mixed 90:10 with BM from B6 CD45.1 or Blr1-/- CD45.1 mice; BM from Edg1+/- B6 mice that was mixed 65:35 with BM from B6 CD45.1 mice; BM from S1P3-/- mice that was mixed 50:50 with BM from B6 CD45.1 mice or the two BM types were used alone. Cr2-/-45, Cxcl13-/-46, Blr1-/-47, Edg3-/-48, and VDJ9-κ5 mice49 were previously described. Mice were housed in a specific pathogen-free environment in the Laboratory Animal Research Center at the University of California, San Francisco, and all animal procedures were approved by the Institutional Animal Care and Use Committee at the University of California, San Francisco.

Treatments and TNP-Ficoll injection

For VPC44116 treatment18, the drug was first dissolved at 20 mM in a mixture of dimethylsulfoxide (DMSO) and 1N HCl at 95:5 ratio and immediately diluted to a final concentration of 1 mM in aqueous solution containing 2% 2-hydroxypropyl-β-cyclodextrin (Sigma). A carrier control solution was similarly prepared in the absence of VPC44116. Mice were injected i.p. with 10 mg/Kg of VPC44116, or the equivalent volume of carrier, or saline solution. Mice were injected at time -3, -2 and -1 h and were sacrificed at 0 h.

For FTY720 pretreatment, mice were injected i.p. with 1mg/Kg of the drug or an equivalent volume of saline for 3 h. For TNP-Ficoll deposition studies, mice were injected i.v. with different doses (see text) of TNP(89)- or TNP(65)-Ficoll (Biosearch Technologies Inc.) and were sacrificed 30 min or 3 h thereafter. Tissues were processed on ice to minimize further binding of the Ficoll in vitro.

In vivo labeling of B cells

For in vivo labeling of B cells, a method adapted from J. Pereira (manuscript in preparation) was used. Mice were i.v. injected for the times indicated in the text with the indicated amounts of the following antibodies: 1 μg anti-CD21-PE (7G6, recognizing both CD21 and CD35), 0.5 μg CD21-FITC (7G6), 1 μg CD35-biotin (8C12, recognizing selectively CD35), 1 μg CD19-PE (1D3) all from BD Bioscience, 5 μg CD1d-PE (1B1), 2 μg CD45.2-PE (clone 104) both from eBioscience, 1.5 μg B220-PE (RA-3-6B2) from Invitrogen, Caltag. Tissues were processed on ice to minimize further binding of the antibodies in vitro.

Flow cytometric and immunohistochemical analysis

Flow cytometric analysis was on a FACSCalibur or LSRII (both from Becton Dickinson). S1P1 staining was previously described17. All other antibodies were from BD Bioscience. Marginal zone and follicular B cells were gated on CD19 or B220 positive cells. Marginal zone B cells were gated as CD23lo CD21hi or CD23lo CD1dhi and follicular B cells were gated as CD23hi CD21med or CD23hi CD1dmed. Since B cells may bind injected TNP-Ficoll or injected antibodies upon processing of the tissue in vitro, tissues from syngeneic non-treated mice were mixed with tissues from treated animals before further processing, and these cells were used as background controls. Immunohistochemical (IHC) and immunofluorescence (IF) analysis of spleen sections were performed as described11.

Statistical analysis

Data was analyzed using a two-sample student’s t-test.


We are grateful to J. Pereira for sharing the in vivo labeling technique, J. An and Y. Xu for excellent technical help, R. Proia (NIDDKD, Bethesda, MD) for S1P1- and S1P3-deficient mice, M. Lipp (Max Delbrück Center of Molecular Medicine, Berlin, Germany) and R. Forster (Hannover Medical School, Hannover, Germany) for CXCR5-deficient mice, X. Wu and J. Atkinson (Washington University School of Medicine, St. Louis, Missouri) for Cr2-/- mice, K. Rajewsky (The CBR Institute for Biomedical Research, Harvard Medical School, Boston, MA) for CD19-cre mice, K. Lynch and T. MacDonald (University of Virginia Medical Center, Chartlottesville, VA) for providing VPC44116, C. Allen for help with the LSRII, L. Shiow for helpful discussion and T. Phan for comments on the manuscript. G.C. is an Associate and J.G.C. is an Investigator of the Howard Hughes Medical Institute. This work was supported in part by NIH grant AI40098.


The authors declare no competing financial interests.


1. Cyster JG. Chemokines, sphingosine-1-phosphate, and cell migration in secondary lymphoid organs. Annu. Rev. Immunol. 2005;23:127–159. [PubMed]
2. MacLennan ICM, Gray D, Kumararatne DS, Bazin H. The lymphocytes of splenic marginal zones: a distinct B-cell lineage. Immunol. Today. 1982;3:305–307.
3. Martin F, Kearney JF. Marginal-zone B cells. Nat. Rev. Immunol. 2002;2:323–335. [PubMed]
4. Kraal G. Cells in the marginal zone of the spleen. Int. Rev. Cytology. 1992;132:31–73. [PubMed]
5. Nossal GJ, Austin CM, Pye J, Mitchell J. Antigens in immunity. XII. Antigen trapping in the spleen. Int. Arch. Allergy Appl. Immunol. 1966;29:368–383. [PubMed]
6. Mitchell J, Abbot A. Antigens in immunity. XVI. A light and electron microscope study of antigen localization in the rat spleen. Immunology. 1971;21:207–224. [PMC free article] [PubMed]
7. Brown JC, Harris G, Papamichail M, Sljivic VS, Holborow EJ. The localization of aggregated human -globulin in the spleens of normal mice. Immunology. 1973;24:955–968. [PMC free article] [PubMed]
8. Gray D, Kumararatne DS, Lortan J, Khan M, MacLennan IC. Relation of intra-splenic migration of marginal zone B cells to antigen localization on follicular dendritic cells. Immunology. 1984;52:659–669. [PMC free article] [PubMed]
9. Groeneveld PH, Erich T, Kraal G. The differential effects of bacterial lipopolysaccharide (LPS) on splenic non-lymphoid cells demonstrated by monoclonal antibodies. Immunology. 1986;58:285–290. [PMC free article] [PubMed]
10. Ferguson AR, Youd ME, Corley RB. Marginal zone B cells transport and deposit IgM-containing immune complexes onto follicular dendritic cells. Int. Immunol. 2004;16:1411–1422. [PubMed]
11. Cinamon G, et al. Sphingosine 1-phosphate receptor 1 promotes B cell localization in the splenic marginal zone. Nat. Immunol. 2004;5:713–720. [PubMed]
12. Vora KA, et al. Sphingosine 1-phosphate receptor agonist FTY720-phosphate causes marginal zone B cell displacement. J. Leukoc. Biol. 2005;78:471–480. [PubMed]
13. Schwab SR, et al. Lymphocyte sequestration through S1P lyase inhibition and disruption of S1P gradients. Science. 2005;309:1735–1739. [PubMed]
14. Pappu R, et al. Promotion of Lymphocyte Egress into Blood and Lymph by Distinct Sources of Sphingosine-1-Phosphate. Science. 2007;316:295–298. [PubMed]
15. Rosen H, Goetzl EJ. Sphingosine 1-phosphate and its receptors: an autocrine and paracrine network. Nat. Rev. Immunol. 2005;5:560–570. [PubMed]
16. Rosen H, Sanna MG, Cahalan SM, Gonzalez-Cabrera PJ. Tipping the gatekeeper: S1P regulation of endothelial barrier function. Trends Immunol. 2007;28:102–107. [PubMed]
17. Lo CG, Xu Y, Proia RL, Cyster JG. Cyclical modulation of sphingosine-1-phosphate receptor 1 surface expression during lymphocyte recirculation and relationship to lymphoid organ transit. J. Exp. Med. 2005;201:291–301. [PMC free article] [PubMed]
18. Foss FW, Jr., et al. Synthesis and biological evaluation of gamma-aminophosphonates as potent, subtype-selective sphingosine 1-phosphate receptor agonists and antagonists. Bioorg. Med. Chem. 2006 [PMC free article] [PubMed]
19. Lo CG, Lu TT, Cyster JG. Integrin-dependence of lymphocyte entry into the splenic white pulp. J. Exp. Med. 2003;197:353–361. [PMC free article] [PubMed]
20. Guinamard R, Okigaki M, Schlessinger J, Ravetch JV. Absence of marginal zone B cells in Pyk-2 deficient mice define their role in the humoral response. Nat. Immunol. 2000;1:31–36. [PubMed]
21. Sanchez T, et al. Phosphorylation and action of the immunomodulator FTY720 inhibits vascular endothelial cell growth factor-induced vascular permeability. J. Biol. Chem. 2003;278:47281–47290. [PubMed]
22. Allende ML, Dreier JL, Mandala S, Proia RL. Expression of the sphingosine-1-phosphate receptor, S1P1, on T-cells controls thymic emigration. J. Biol. Chem. 2004;279:15396–15401. [PubMed]
23. Rickert RC, Rajewsky K, Roes J. Impairment of T-cell-dependent B-cell responses and B-1 cell development in CD19-deficient mice. Nature. 1995;376:352–355. [PubMed]
24. Martin F, Kearney JF. Positive selection from newly formed to marginal zone B cells depends on the rate of clonal production, CD19, and btk. Immunity. 2000;12:39–49. [PubMed]
25. Forster R, et al. A putative chemokine receptor, BLR1, directs B cell migration to defined lymphoid organs and specific anatomic compartments of the spleen. Cell. 1996;87:1037–1047. [PubMed]
26. Ansel KM, et al. A chemokine driven positive feedback loop organizes lymphoid follicles. Nature. 2000;406:309–314. [PubMed]
27. Gretz JE, Norbury CC, Anderson AO, Proudfoot AE, Shaw S. Lymph-borne chemokines and other low molecular weight molecules reach high endothelial venules via specialized conduits while a functional barrier limits access to the lymphocyte microenvironments in lymph node cortex. J. Exp. Med. 2000;192:1425–1440. [PMC free article] [PubMed]
28. Nolte MA, et al. A conduit system distributes chemokines and small blood-borne molecules through the splenic white pulp. J. Exp. Med. 2003;198:505–512. [PMC free article] [PubMed]
29. Whipple EC, Shanahan RS, Ditto AH, Taylor RP, Lindorfer MA. Analyses of the in vivo trafficking of stoichiometric doses of an anti-complement receptor 1/2 monoclonal antibody infused intravenously in mice. J. Immunol. 2004;173:2297–2306. [PubMed]
30. Nolte MA, Hoen EN, van Stijn A, Kraal G, Mebius RE. Isolation of the intact white pulp. Quantitative and qualitative analysis of the cellular composition of the splenic compartments. Eur. J. Immunol. 2000;30:626–634. [PubMed]
31. Taylor PR, et al. The follicular dendritic cell restricted epitope, FDC-M2, is complement C4; localization of immune complexes in mouse tissues. Eur. J. Immunol. 2002;32:1888–1896. [PubMed]
32. Pasparakis M, Kousteni S, Peschon J, Kollias G. Tumor necrosis factor and the p55TNF receptor are required for optimal development of the marginal sinus and for migration of follicular dendritic cell precursors into splenic follicles. Cell. Immunol. 2000;201:33–41. [PubMed]
33. Voigt I, et al. CXCR5-deficient mice develop functional germinal centers in the splenic T cell zone. Eur. J. Immunol. 2000;30:560–567. [PubMed]
34. Girkontaite I, et al. The sphingosine-1-phosphate (S1P) lysophospholipid receptor S1P3 regulates MAdCAM-1+ endothelial cells in splenic marginal sinus organization. J. Exp. Med. 2004;200:1491–1501. [PMC free article] [PubMed]
35. Gunn MD, et al. A B-cell-homing chemokine made in lymphoid follicles activates Burkitt’s lymphoma receptor-1. Nature. 1998;391:799–803. [PubMed]
36. Ngo VN, et al. Lymphotoxin alpha/beta and tumor necrosis factor are required for stromal cell expression of homing chemokines in B and T cell areas of the spleen. J. Exp. Med. 1999;189:403–412. [PMC free article] [PubMed]
37. Lu TT, Cyster JG. Integrin-mediated long-term B cell retention in the splenic marginal zone. Science. 2002;297:409–412. [PubMed]
38. Woolf E, et al. Lymph node chemokines promote sustained T lymphocyte motility without triggering stable integrin adhesiveness in the absence of shear forces. Nat. Immunol. 2007;8:1076–1085. [PubMed]
39. Liu CH, et al. Ligand-induced trafficking of the sphingosine-1-phosphate receptor EDG-1. Mol. Biol. Cell. 1999;10:1179–1190. [PMC free article] [PubMed]
40. Oo ML, et al. Immunosuppressive and Anti-angiogenic Sphingosine 1-Phosphate Receptor-1 Agonists Induce Ubiquitinylation and Proteasomal Degradation of the Receptor. J. Biol. Chem. 2007;282:9082–9089. [PubMed]
41. Gonzalez-Cabrera PJ, Hla T, Rosen H. Mapping pathways downstream of sphingosine 1-phosphate subtype 1 by differential chemical perturbation and proteomics. J. Biol. Chem. 2007;282:7254–7264. [PubMed]
42. Veerman AJ, van Rooijen N. Lymphocyte capping and lymphocyte migration as associated events in the in vivo antigen trapping process. An electron-microscopic autoradiographic study in the spleen of mice. Cell Tissue Res. 1975;161:211–217. [PubMed]
43. Phan TG, Grigorova I, Okada T, Cyster JG. Subcapsular encounter and complement-dependent transport of immune complexes by lymph node B cells. Nat. Immunol. 2007;8:992–1000. [PubMed]
44. Liu Y, et al. Edg-1, the G protein-coupled receptor for sphingosine-1-phosphate, is essential for vascular maturation. J. Clin. Invest. 2000;106:951–961. [PMC free article] [PubMed]
45. Molina H, et al. Markedly impaired humoral immune response in mice deficient in complement receptors 1 and 2. Proc. Natl. Acad. Sci. U. S. A. 1996;93:3357–3361. [PMC free article] [PubMed]
46. Ansel KM, Harris RB, Cyster JG. CXCL13 Is Required for B1 Cell Homing, Natural Antibody Production, and Body Cavity Immunity. Immunity. 2002;16:67–76. [PubMed]
47. Förster R, et al. A putative chemokine receptor, BLR1, directs B cell migration to defined lymphoid organs and specific anatomic compartments of the spleen. Cell. 1996;87:1037–1047. [PubMed]
48. Kono M, et al. The sphingosine-1-phosphate receptors S1P1, S1P2, and S1P3 function coordinately during embryonic angiogenesis. J. Biol. Chem. 2004;279:29367–29373. [PubMed]
49. Allen CD, Okada T, Tang HL, Cyster JG. Imaging of germinal center selection events during affinity maturation. Science. 2007;315:528–531. [PubMed]
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...