B cell homeostasis and plasma cell homing controlled by Krüppel-like factor 2

Abstract

Krüppel-like factor 2 (KLF2) controls T lymphocyte egress from lymphoid organs by regulating sphingosin-1 phosphate receptor 1 (S1Pr1). Here we show that this is not the case for B cells. Instead, KLF2 controls homeostasis of B cells in peripheral lymphatic organs and homing of plasma cells to the bone marrow, presumably by controlling the expression of β₇-integrin. In mice with a B cell-specific deletion of KLF2, S1Pr1 expression on B cells was only slightly affected. Accordingly, all splenic B cell subsets including B1 cells were present, but their numbers were increased with a clear bias for marginal zone (MZ) B cells. In contrast, fewer peyers patches harboring fewer B cells were found, and fewer B1 cells in the peritoneal cavity as well as recirculating B cells in the bone marrow were detected. Upon thymus-dependent immunization, IgG titers were diminished, and antigen-specific plasma cells were absent in the bone marrow, although numbers of antigen-specific splenic plasmablasts were normal. KLF2 plays also a role in determining the identity of follicular B cells, as KLF2-deficient follicular B cells showed calcium responses similar to those of MZ B cells and failed to down-regulate MZ B cell signature genes, such as CD21 and CXCR7.

Maturation of lymphocytes in primary lymphoid organs, controlled egress into the periphery, and proper positioning in lymphoid tissues are critical for efficient adaptive immune responses. Differential expression of Krüppel-like factor 2 (KLF2) promotes egress of T cells from lymphoid organs into the blood and T-cell migration to lymph nodes (1, 2). Accordingly, KLF2 is expressed in naive T cells, down-regulated upon activation, and reexpressed in memory T cells (1, 3–5). KLF2-deficient T cells inefficiently exit from the thymus and thus accumulate there; this is thought to be a consequence of a decrease in sphingosin-1 phosphate receptor 1 (S1Pr1) expression, which is regulated by KLF2 (2, 6). In addition, KLF2 increases the expression of β₇-integrin (gene symbol: Itgb7) and CD62L (l-selectin) on T cells (1, 2). However, other downstream effects of KLF2 expression are less clear, although they very likely contribute to the migratory behavior of T cells. For example, vav-cre- and lck-cre-mediated deletion of KLF2 in T cells resulted in the up-regulation of the inflammatory chemokine receptors CCR3 and CCR5 (7) on thymocytes, whereas CD4-cre–mediated deletion led to the up-regulation of only CXCR3 and spontaneous IL4 production in naive T cells (8).

Although the role of KLF2 in T-cell migration has been studied extensively, its function in B cells is not fully understood. KLF2 expression at the RNA and protein level is induced by pre-B cell receptor (pre-BCR) signals (9). Analogous to T cells, KLF2 transcripts are abundant in resting mature B cells, down-regulated upon mitogenic activation, and reexpressed in plasma and memory B cells (10–12). Because KLF2 controls the expression of β₇-integrin and CD62L, and because these adhesion molecules play also an important role in B cell trafficking (13–15), we thought that KLF2 controls homeostasis and trafficking of B lineage cells. To address this question, we investigated the function of KLF2 in a mouse model with a B cell-specific deletion of KLF2.

Results

KLF2 Expression Profiling in Mature B Cell Subsets.

Microarray analyses showed that KLF2 is expressed in pre-B cells, naive B cells, and plasma cells (9–12). We confirmed KLF2 mRNA expression in isolated plasma cells from bone marrow and resting B cells by Taqman RT-PCR (Fig. S1). In addition, we found a down-regulation of KLF2 transcripts upon stimulation with either LPS or a mixture of anti-CD40/IL4/anti-IgM (αBCR) (Fig. S1). Western blot analysis using rabbit serum raised against a KLF2 peptide (9) confirmed that KLF2 protein was expressed in freshly isolated CD43-negative splenic B cells and was down-regulated after treatment with an αBCR mixture (Fig. 1A). We also made two additional observations. In resting B cells, we found two forms of KLF2, suggesting that the protein is posttranslationally modified. Upon activation, the high molecular form disappeared faster than the low molecular form. More surprisingly, although abundant in isolated follicular (FO) B cells, KLF2 protein was barely detectable in marginal zone (MZ) B cells from WT mice (Fig. 1B).

Fig. 1.

Expression analysis of KLF2 in B cells. (A and B) Western blot analyses of KLF2 abundance in (A) CD43-depleted splenic B cells before and after mitogenic activation with αBCR mix and (B) purified splenic FO and MZ B cells from WT mice. Membranes were first stained with an anti-KLF2 serum and then with an antiactin serum. Actin abundance served as a control for the amount of loaded protein.

Normal B Cell Development in the Bone Marrow of KLF2-Deficient Animals.

One function of KLF2 is to keep cells in a quiescent state (16–18). Applied to B lymphoid cells, an increase of KLF2 expression upon pre-BCR induction should terminate the cell cycle of proliferating pre-B cells (9, 19). If so, KLF2 deficiency should result in hyperproliferation of pre-B cells, and thus, in an increase of the pre-B cell pool. To test this hypothesis, we established B cell-specific KLF2-knockout mice by crossing mice carrying a floxed KLF2 (KLF2^flox) allele (20) with the B cell-specific deleter strain mb1-cre (21) (Fig. S2 A and B). In these mice, deletion of KLF2 occurs in early B cell precursors, as demonstrated by Western blot analyses of CD19⁺ B lymphoid cells from bone marrow and spleen (Fig. S2C). In flow cytometric analyses, the frequencies and numbers of pro-B (CD19⁺/c-kit⁺), pre-B (CD19⁺/CD25⁺) and immature (B220^low/IgM⁺; IgM⁺/IgD⁻/CD93⁺) B cells (Fig. S3) did not differ between KLF2-deficient and KLF2-sufficient mice. This excludes KLF2 as the quiescence factor to terminate pre-BCR–mediated proliferative expansion of functional pre-B cells (19). However, frequencies and numbers of recirculating mature B cells (IgD⁺/IgM⁺ and B220^hi/IgM⁺) are decreased by half (Fig. 2A and Fig. S3C, respectively), indicating that KLF2-deficient mature B cells might have a defect in their homing capacity to the bone marrow.

Fig. 2.

Disturbed B cell homeostasis in peripheral lymphatic organs. Cell suspensions from different organs of WT (FACS dot plots to the left) and KLF2-deficient (FACS dot plots to the right) mice (6–10 wk old) were surface stained with indicated antibodies. Frequencies (Left and Center) and absolute cell numbers (Right) of (A) recirculating B cells (IgD⁺/IgM⁺) from bone marrow, (B) total B cells (B220⁺/IgM⁺), MZ B cells (B220⁺/CD23⁻/CD21^hi), FO B cells (B220⁺/CD23⁺/CD21^low), transitional B cells (T1: B220⁺/AA4.1⁺/IgM^hi/CD23⁻, T2: B220⁺/AA4.1⁺/IgM^hi/CD23⁺ and T3: B220⁺/AA4.1⁺/IgM^low/CD23⁺), as well as B1a (IgM⁺/CD5⁺/CD43⁺) and B1b cells (IgM⁺/CD5⁻/CD43⁺) from spleen are shown. (C) Immunofluorescence of acetone-fixed tissue cryosections (color code of antibodies shown to the right). Frequencies of total CD19⁺/IgM⁺ B cells in (D) inguinal lymph node, (E) blood, and (F) peyers patches as well as (G) frequencies of B1a (IgM⁺/CD5⁺/CD11b⁺), B1b (IgM⁺/CD5⁻/CD11b⁺), and B2 (IgM⁺/CD5⁻/CD11b⁻) cells in peritoneal lavages. Results of all analyzed littermates are summarized in the diagrams shown to the right, with one dot representing one mouse.

Splenomegaly and Increased Numbers of Splenic B Cell Subsets in KLF2-Deficient Mice.

Immature bone marrow B cells enter the blood stream and migrate to the spleen, where they develop into transitional B cells (22) and differentiate into mature FO or MZ B cells (23, 24). When we compared B cell subsets in the spleen from KLF2-deficient and WT mice by flow cytometry, we observed a clear effect of KLF2 deletion on B cell homeostasis (Fig. 2B). KLF2-deficient mice have enlarged spleens (Fig. S4A) with a two- to threefold increase in the total number of B220⁺/IgM⁺ splenic B cells, an increase in all transitional B cell stages (T1–T3), a two- to threefold increase in numbers of FO B (B220⁺/CD21^lo/CD23^hi), and a four- to fivefold increase of MZ B cells (B220⁺/CD21^hi/CD23^lo/neg) (Fig. 2B). In addition, slightly but not significantly increased numbers of B1a cells (IgM⁺/CD5⁺/CD43⁺) were detected in the spleen (Fig. 2B and Fig. S4B).

Next, we examined the splenic architecture. Lymphoid follicles in the spleen of WT mice consist of a core with IgD^hi/IgM^lo FO B cells, separated by a ring of MOMA⁺ metallophilic macrophages from the marginal zone harboring IgD^lo/IgM^hi MZ B cells (Fig. S5). In the spleen of KLF2-deficient mice, follicles with FO and MZ B cells are present. However, follicles are larger in size, and the separation between MZ and FO B cells is not as clear as in KLF2-deficient mice (Fig. 2C and Fig. S5).

B Cell Homeostasis Is Disturbed in Other Peripheral Lymphatic Compartments of KLF2-Deficient Mice.

Mature FO B cells leave the spleen, circulate in the bloodstream and lymphstream, and enter other peripheral lymphatic organs (25). To determine whether KLF2 controls migration and homing of mature peripheral B cell subsets, we analyzed other lymphoid compartments for the presence of B cells. By flow-cytometric analysis, we found no differences in the numbers and frequencies of CD19⁺/IgM⁺ B cells in inguinal lymph nodes (Fig. 2D). In contrast, in the blood of KLF2-deficient animals, frequencies of B cells were decreased two- to threefold (Fig. 2E). Similarly, the peyers patches in KLF2-deficient animals were smaller, their numbers were reduced (Fig. S6A) and they contained two- to threefold fewer CD19^+/IgM⁺ B cells (Fig. 2F). Consistent with the reduced number of peyers patches and B cells, serum titers of “natural” IgA were reduced as determined by ELISA (Fig. 3A, Right).

Fig. 3.

Antibody serum titers before and after thymus-independent immunization. Sera of WT mice (open blue triangles) and KLF2-deficient mice (filled red triangles) were collected and analyzed by ELISA for (A) total Ig as well as antigen-specific Ig titers upon i.v. immunization with (B) 50 μg TNP-LPS and (C) 25 μg TNP–ficoll.

The effect of KLF2 deletion on B1 cells in the peritoneal cavity was even more severe, with a clear reduction in B1a (IgM⁺/CD5⁺) and B1b (IgM⁺/CD5⁻/CD11b⁺) cells (Fig. 2G and Fig. S6B). Because B1 cells are the major source for production of “natural” IgM in the serum (26), we determined the amount of serum Ig titers by ELISA but could not detect any differences in “natural” IgM or IgG serum levels (Fig. 3A, Left and Center).

Reduced Numbers of Plasma Cells Observed in Bone Marrow After Boost Immunization with Thmyus-Dependent Antigen.

To investigate whether KLF2-deficient B cells could participate in humoral immune responses, we immunized WT as well as KLF2-deficient animals with either thymus-independent (TI) or thymus-dependent (TD) antigens (Figs. 3 and 4). Despite an increase in numbers of FO and MZ B cells in the spleen, antigen-specific antibody IgM and IgG titers in response to TI-1 (TNP-LPS, Fig. 3B) and TI-2 (TNP-ficoll, Fig. 3C) antigens were fairly normal in KLF2-deficient animals, with the exception of a slight but significant increase in IgG titers against TNP-LPS at day 14 after immunization (Fig. 3B). Similarly, IgM and IgG responses against the TD antigen TNP-coupled keyhole limbet hemocyanin (TNP-KLH) did not differ significantly between WT and KLF2-deficient mice 14 d after primary immunization (Fig. 4A). However, 35 d after the primary TNP-KLH immunization, antigen-specific total IgG (Fig. 4A) and 42 d after immunization also antigen-specific IgG1 titers were significantly reduced in KLF2-deficient mice (Fig. S7B).

Fig. 4.

Impaired thymus-dependent immune response in KLF2-deficient mice. Sera of WT mice (blue open triangles) and KLF2-deficient mice (red filled triangles) were collected after (A) primary and (B) boost TNP-KLH immunization and analyzed by ELISA for antigen-specific antibodies. (C) Flow-cytometric analyses of frequencies (Left and Center) and absolute numbers (Right) of CD138^hi/κ^hi/λ^hi plasma cells in spleen, blood, and bone marrow of WT and KLF2-deficient mice at indicated time points after boost TNP-KLH immunization. (D) ELISPOT assay of TNP-specific IgG-secreting cells in bone marrow and spleen 14 d after boost immunization with TNP-KLH. and D) Triplicate of one representative experiment (Left), and summary of results of all analyzed littermates, with one dot representing the mean value of ELISPOT triplicates of one mouse (Right).

The secondary immune response against TNP-KLH (Fig. 4B) mirrored the primary response. Over a period of 7 days, the kinetics of antigen-specific IgG and IgM titers did not differ between WT and knockout mice. However, total IgG and IgG1 titers, but not IgM titers, were significantly reduced 14 d after boost immunization (Fig. 4B and Fig. S7B). This was not due to a drop in the number of antibody-secreting (CD138^hi/κ^hiλ^hi) plasma cells in the blood and the spleen, as their numbers were similar 5 and 14 d after boost immunization in the blood and spleen, respectively (Fig. 4C). In contrast, the number and frequency of plasma cells in the bone marrow were two- to threefold decreased 14 d after boost immunization with TNP-KLH (Fig. 4C, Lower). To confirm this finding for antigen-specific plasma cells, we enumerated TNP-specific IgG-secreting cells by ELISPOT assays in spleen and bone marrow from mice 14 d after a secondary TNP-KLH immunization (Fig. 4D). We found similar numbers of antigen-specific IgG-secreting cells in the spleen of both WT and knockout mice. However, in contrast to WT mice, TNP-IgG-secreting cells were absent in the bone marrow of KLF2-deficient mice. Hence, KLF2 contributes to the homing of plasma cells to survival niches in the bone marrow.

KLF2 Controls Expression of CD62L and β₇-Integrin but Not S1Pr1.

Because KLF2-deficient mice have more B cells in the spleen and fewer B cells in most other peripheral organs, KLF2 could play a role in the egress, migration, and homing of peripheral B cell subsets by controlling the expression of genes encoding migration factors and adhesion molecules. For example, previous studies identified the chemoattractant receptor S1Pr1 and the adhesion molecules CD62L and β₇-integrin as KLF2 target genes (1, 2, 7, 27).

Egress of thymocytes and B cells from the thymus and the bone marrow, respectively, to the blood is controlled by expression of S1Pr1, a direct KLF2 target gene (2, 6, 28–30). In addition, S1Pr1 is required for efficient egress of plasma cells from the spleen into the blood (12). Because KLF2-deficient B cells egress from the bone marrow, and because KLF2-deficient plasma cells enter the bloodstream, we would expect that KLF2 deficiency, in contrast to T cells, would not significantly affect S1Pr1 expression in B cells. Indeed, this was the case. In microarray and flow-cytometric analyses, we detected, in splenic FO B and MZ B cells as well as in blood B cells, very similar levels of S1Pr1 (Fig. 5 A and B, Upper Row). We confirmed the presence of functional S1Pr1 on KLF2-deficient MZ B cells by injecting mice with the S1Pr1-agonist FTY720 (Fig. 5C). FTY720 decreases S1Pr1 on the cell surface, and, as a result, MZ B cells are attracted by the chemokine CXCL13 to the follicular zone (31). As documented in Fig. 5C, FTY720 abolishes the clear separation of FO B and MZ B cells.

Fig. 5.

KLF2 controls expression of CD62L and β₇-integrin but not S1Pr1. (A) Affymetrix microarray analysis of total RNA from FO B cells of WT and KLF2-deficient mice. RLU, relative light units for selected genes. (B) Flow-cytometric analyses for S1Pr1, α₄β₇-integrin, and CD62L of FO B (gated B220⁺/CD23⁺/CD21^low), MZ B (gated B220⁺/CD23⁻/CD21^hi) and blood B (gated CD19⁺, IgM⁺) cells of WT (gray histograms) and KLF2-deficient (open histograms) mice. (C) Immunofluorescence analyses of splenic acetone-fixed cryosections from DMSO-treated (Upper) and FTY720-treated (Lower) WT and KLF2-deficient mice. Color codes for antibodies are shown to the right.

There are conflicting reports about the role of KLF2 in controlling the expression of chemokine receptors in T cells (7, 8). In our microarray and TaqMan RT-PCR analyses, the expression of most chemokine receptors was not affected in KLF2-deficient FO and MZ B cells (Fig. S8). Only transcripts for CXCR7 are up-regulated in KLF2-deficient FO B cells on microarrays and TaqMan RT-PCR assays (Fig. 5A and Fig. S8). An almost identical profile of chemokine receptor expression in KLF2-deficient B cells is reported in the accompanying manuscript by Hart et al.

In contrast to S1Pr1, the abundance of the α₄β₇-integrin heterodimer was reduced in KLF2-deficient mice on recirculating bone marrow B (Fig. S9A), blood B (Fig. 5B), and FO B cells (Fig. 5 A and B and Fig. S9B). Here, we measured expression of β₇-integrin (gene symbol: Itgb7) mRNA by Affymetrix microarray and RT-PCR, and the amount of α₄β₇-integrin on protein level by flow cytometry. In contrast, α₄β₇-integrin is hardly expressed on WT and KLF2-deficient MZ B cells (Fig. 5B). This was not surprising, as WT MZ B cells barely produce KLF2 (Fig. 1B). CD62L, an adhesion molecule critical for lymphocytes to home to peripheral lymph nodes (13, 15), can still be detected at mRNA and protein levels on KLF2-deficient FO B and blood B cells. However, compared with the corresponding WT B cell subset, CD62L expression was reduced in KLF2-deficient B cells (Fig. 5B, Lower Row). Because MZ B cells from WT mice hardly express KLF2, it was not surprising that MZ B cells showed lower CD62L expression than WT FO B cells. In summary, KLF2-deficieny results in a strong reduction of β₇-integrin and a weaker down-regulation of CD62L on FO B cells.

KLF2 Deficiency Results in an MZ-Like Phenotype in FO B Cells.

Because KLF2 is barely detected in WT MZ B cells (Fig. 1B), we speculated that KLF2-deficient FO B cells might show some molecular and functional hallmarks of MZ B cells. This is indeed the case. As shown by flow cytometry (Figs. 2B and 6A) and on splenic cryosections (Fig. 6B), KLF2-deficient FO B cells fail to down-regulate complement receptors 1/2 (CD21/35) and express CD21/CD35 at levels similar to those on WT MZ B cells (Fig. 6A). In addition, CXCR7, a chemokine receptor expressed on MZ B cells (32), is up-regulated on KLF2-deficient FO B cells, as shown by microarray and TaqMan RT-PCR analyses (Fig. 5A and Fig. S8).

Fig. 6.

KLF2 deficiency results in a MZ-like phenotype in FO B cells. (A) Flow-cytometric analysis of CD21/CD35 expression on FO and MZ B cells of WT mice (shaded histograms) and KLF2-deficient mice (open histograms). (B) Immunofluorescence analysis of CD21/35 expression on splenic cryosections (color code of antibodies shown to the right). (C) Intra- and extracellular calcium signaling of FO and MZ B cells of WT (gray line) and KLF2-deficient mice (black line) upon anti-μHC F(ab′)₂ treatment. Results of one representative experiment (n = 3) are shown. Gating strategy for FO and MZ B cells is depicted in Fig. S10A.

Furthermore, compared with WT FO B cells, KLF2-deficient FO B cells showed a slight but statistically significant increase in the baseline of intracellular Ca²⁺, similar to that found in WT MZ B cells (Fig. 6C and Fig. S10). In contrast, KLF deficiency did not affect either intracellular or extracellular Ca²⁺ flux in MZ B cells (Fig. 6C and gating in Fig. S10A). Finally, MZ-like B cells with higher expression of CD21/35 could also be detected in the inguinal lymph nodes by flow cytometry (Fig. S11). These findings point to a function of KLF2 in determining the identity of FO B cells.

Discussion

In this study, we investigated the effect of KLF2 deficiency on plasma cell homing to the bone marrow and B cell homeostasis. In T cells, KLF2 controls the exit of thymocytes into the blood by regulating the expression of the chemoattractant receptor S1Pr1 (1, 2, 23). In contrast to T cells, we show here that KLF2 is not required to maintain surface expression of S1Pr1 on B lineage cells. Consistent with this, WT MZ B cells, which produce only small amounts of KLF2, show high levels of S1Pr1 (31). We also show here that S1Pr1 on KLF2-deficient B cells is functional. We conclude this from three findings: (i) KLF2-deficient B cells are displaced from the marginal zone upon FTY720 treatment in vivo; (ii) immature KLF2-deficient B cells egress from the bone marrow and are positioned to the correct anatomical location in splenic follicles; and (iii) KLF2-deficient plasma cells reach the bloodstream. The fact that all three events in B cell maturation depend on the expression of S1Pr1 (28, 30, 31) strongly supports our conclusion.

It is known that plasma cells require CXCR4 and CXCR3 to reach the bone marrow (33–35) but not how they enter the bone marrow and which survival signals are required there (36, 37). KLF2-deficient B cells respond to TD antigens and develop in the spleen into IgG-secreting plasma cells. However, these cells do not reach the bone marrow. Our finding that KLF2-deficient B cells down-regulate β₇-integrin on transcript level and the α₄β₇-heterodimer on protein level could add another player to the mechanism that controls the homing of plasma cells to the bone marrow. α₄β₇-Integrin is critical for IgA-secreting plasma cells to home to the gut (14, 38, 39), which could explain the decrease in natural serum IgA in KLF2-deficient mice. Furthermore, vascular cell adhesion molecule-1 (VCAM-1), a ligand for α₄β₇-integrin, is expressed in the bone marrow in combination with mucosal addressin cell adhesion molecule-1 (MAdCAM-1) (38). Therefore, it is tempting to speculate that α₄β₇-integrin is critical either for entry of plasma cells into the bone marrow or for their localization to survival niches in the bone marrow. However, some CD138^hi/IgL-chain^hi plasma cells were still present in the bone marrow of our KLF2-deficient mice. Therefore, KLF2 deficiency could either affect the trafficking of only a subset of plasma cells (e.g., from the spleen) or delay plasma cell homing to the bone marrow. Further characterization of KLF2-deficient plasma cells and transfer experiments with GFP-labeled plasma cells from spleen and other secondary organs will address this issue.

The down-regulation of α₄β₇-integrin in KLF2-deficient mice could also explain most of the defects in B cell homeostasis, as the phenotype is mirrored in mice with defective α₄β₇-integrin function. For example, treatment with blocking antibodies against α₄β₇-integrin (13), or germline deletion of β₇-integrin (15), results in an accumulation of B cells in the spleen and a reduced capacity of B cells to enter peyers patches and mesenteric lymph nodes but not peripheral lymph nodes. Unfortunately, none of these studies examined the effect of defective β₇-integrin function on B1 cell homeostasis in the peritoneal cavity or the presence of antigen-specific plasma cells in the bone marrow.

Surprisingly, the strong reduction in peritoneal B1 cells in our KLF2-KO mice was not accompanied by decreased serum IgM levels (26). The four- to fivefold increase in the MZ B cell compartment might compensate for the loss of the B1 cell compartment as a producer of “natural” IgM (24). Alternatively, splenic CD5-positive B1a cells, which were slightly increased in KLF2-KO mice, could be another source for most of the natural serum IgM (40).

During the preparation of our manuscript, Hoek et al. also reported the effect of conditional KLF2 deletion (via CD19-cre) on B cells (41). Although some of their findings, such as the increase in FO and MZ B cells and the reduction in peritoneal B1 cells, are consistent with our results, there are notable differences. For example, Hoek et al. did not analyze secondary immune responses and plasma cell homing. In addition, they propose that S1Pr1 and CXCR5 are strongly down-regulated in MZ B cells and up-regulated in FO B cells and conclude that this explains why FO B cells invade the marginal zone in the spleen. In contrast, we did not find significant changes in S1Pr1 expression at the RNA and protein level in KLF2-deficient MZ and FO B cells, and CXCR5 expression remained unchanged in both populations. The drastic down-regulation of S1Pr1 in KLF2-deficient MZ B cells in the Hoek et al. paper is surprising, as MZ B cells still disappear in that study after blockage of S1Pr1 function through FTY720 treatment. We do not know the reason for the differences between the results of Hoek et al. and our own. However, an accompanying paper by Hart et al. describes studies in a mouse with CD19-cre mediated KLF2 deletion with essentially the same results as reported in our paper.

In summary, KLF2 controls homeostasis of mature B cell subsets in peripheral lymphatic organs and homing of antibody-secreting plasma cells to the bone marrow, presumably by controlling the expression of the adhesion molecule β₇-integrin.

Materials and Methods

Procedures.

Standard procedures and methods such as animal handling, flow cytometry, isolation and activation of B cells, histology, immunizations and antigen-specific ELISA, ELISPOT assays, RNA and Western blot analyses, and statistical analyses are described in SI Appendix.

Analysis of Intracellular Calcium Flux.

Calcium flux measurements were performed as described (42, 43). Briefly, 5 × 10⁶ total spleen cells were incubated with the Ca²⁺ dye Indo-1 AM and fluorochrome-conjugated antibodies against CD21 and CD23 to distinguish between FO (CD23^hi/CD21^lo) and MZ (CD23^lo/neg/CD21^hi) B cells. To measure changes in intracellular Ca²⁺, stained cells were suspended in Ca²⁺-free Krebs–Ringer solution containing 0.5 mM EGTA and stimulated with 10 μg/mL anti-μHC F(ab')₂ (Jackson Laboratories). To measure extracellular Ca²⁺ influx, Ca²⁺ was restored after 240 s. Changes in Indo-1 AM fluorescence ratios were analyzed in gated FO and MZ B cell populations with a LSRII flow cytometer (BD Biosciences) using FlowJo software (Tree Star). Equal Indo-1 loading was assessed by ionomycin stimulation.

FTY720 Treatment.

For FTY treatment, 6- to 10-wk-old animals were injected i.p. with either 20 μg FTY720 (Cayman Chemicals) or an equivalent volume of saline. Four hours after injection, animals were killed, and spleens were removed and processed for cryosections.