It is well established that Peyer’s patches (PPs) are sites for the differentiation of IgA plasma cell precursors, but molecular and cellular mechanisms in their trafficking remain to be elucidated. In this study, we show that alterations in type 1 sphingosine 1-phosphate (S1P) receptor expression during B cell differentiation in the PPs control the emigration of IgA plasma cell precursors. Type 1 S1P receptor expression decreased during the differentiation of IgM+B220+ B cells to IgA+B220+ B cells, but recovered on IgA+B220− plasmablasts for their emigration from the PPs. Thus, IgA+B220− plasmablasts migrated in response to S1P in vitro. Additionally, IgA+ plasmablasts selectively accumulated in lymphatic regions of PPs when S1P-mediated signaling was disrupted by FTY720 treatment. This accumulation of IgA+ plasmablasts in the PPs led to their reduction in the intestinal lamina propria and simultaneous impairment of Ag-specific intestinal IgA production against orally administered Ag. These findings suggest that S1P regulates the retention and emigration of PP B cells and plays key roles in the induction of intestinal IgA production.
Secretory IgA (S-IgA)4 in the intestinal lumen acts as the gateway controller against pathogenic and commensal microorganisms (1, 2, 3). S-IgA production is achieved by two distinct subsets of B cells, termed B1-B and B2-B cells (3). B1-B cells can be distinguished from B2-B cells by cell surface molecules (e.g., B220, IgM, IgD, CD5, and Mac-1), origin, and growth properties (4). B1-B cells migrate from nonorganized tissues such as the peritoneal cavity to the intestinal lamina propria (iLP), where they further differentiate to IgA-secreting plasma cells (PCs) (3). In addition to the peritoneal B cells, the common mucosal immune system (CMIS) acts as an alternative pathway for S-IgA production (1, 2). In the CMIS-dependent pathway, Peyer’s patches (PPs) act as inductive tissues, where B cells are primed and switched from the μ- to α-chain by immunological interactions with dendritic cells (DCs) and T cells, and differentiate to the precursors of IgA+ PCs with the expression of gut-tropic chemokine receptors (e.g., CCR9) and adhesion molecules (e.g., α4β7 integrin) (5, 6, 7). These changes allow the precursors of IgA+ PCs to traffic specifically to the iLP, where they further differentiate to IgA-secreting PCs.
The tissue-specific homing of Ag-primed lymphocytes is tightly regulated by a combination of adhesion molecules and chemokines (8, 9). In addition to these molecules, sphingosine 1-phosphate (S1P) also regulates lymphocyte trafficking, especially emigration from the organized lymphoid tissues, such as the thymus and secondary lymphoid organs (9, 10). Five types of S1P receptors have been identified, with the type 1 S1P receptor (S1P1) primarily expressed on lymphocytes (10, 11). S1P1 expression on T cells is cyclically modulated during their circulation, which determines their retention in lymph nodes or exit into the blood and lymph in response to the S1P gradient (12, 13, 14). T cells also control S1P1 expression during their development and activation (15, 16, 17). Thus, treatment with FTY720, a S1P1 modulator, decreases the number of circulating T cells in both blood and lymph by inhibiting their emigration from the secondary lymphoid organs and thymus (18).
In addition to T cells, S1P is also involved in the regulation of B cell trafficking. FTY720 impairs plasma Ab production, especially against T-dependent Ag due to the abolishment of germinal center (GC) formation (19, 20, 21). S1P is also important in the determination of PC tropism into bone marrow (22). Furthermore, we have recently demonstrated that S1P contributes to the regulation of peritoneal B cell trafficking into the intestine and subsequent intestinal S-IgA production (23), as well as naive αβTCR+ intraepithelial T lymphocytes in naive mice (24) and pathogenic T and mast cell trafficking into the intestine under intestinal allergic conditions (25). These findings provide strong evidence that S1P plays an essential role in the regulation of lymphocyte trafficking in both systemic and mucosal humoral immunity. However, it is still unclear whether S1P is involved in the regulation of intestinal S-IgA production via the CMIS, a core pathway for the development of Ag-specific IgA B cells and subsequent Ag-specific mucosal Ab production. Additionally, although the details of S1P1 expression during T cell differentiation and activation are well studied (9, 10, 11, 15, 16, 17), the contribution of S1P1 expression during B cell differentiation still remains to be elucidated.
PPs could potentially be used to investigate these two unresolved, but immunologically important questions simultaneously because PP is the main inductive tissue for CMIS-mediated intestinal humoral immunity and shows spontaneous class switching recombination (CSR) from μ- to α-chain to supply IgA-committed B cells to distant effector sites such as the iLP (26). This feature of PPs allowed us to investigate the involvement of S1P in intestinal IgA production via the CMIS, as well as to elucidate S1P1 expression during B cell differentiation from IgM+ to IgA+ B cells. Thus, we first investigated S1P1 expression during B cell differentiation in the PPs, and then assessed the involvement of S1P in Ag-specific intestinal S-IgA production. These findings provide new evidence that the level of S1P1 expression correlates with distinct stages of IgA-committed B cell development in the PPs, and hence, S1P regulates the retention and emigration of IgA-committed B cells for the subsequent S-IgA production via the CMIS-mediated pathway.
Materials and Methods
Mice and experimental treatment
Female BALB/c mice (7–9 wk) were purchased from Japan CLEA or Japan SLC and provided with sterile food and water ad libitum. For treatment with FTY720 (Novartis Pharmaceuticals), mice were injected i.p. with FTY720 (1 mg/kg/time) (23, 24, 25). For oral immunization, mice were deprived of food for 15 h and then given a solution of sodium bicarbonate to neutralize stomach acid before oral immunization (27). Thirty minutes later, mice were orally immunized with 1 mg of OVA (Sigma-Aldrich) with 10 μg of cholera toxin (List Biological Laboratories). This oral immunization procedure was conducted on days 0, 7, and 14. All animals were maintained in the experimental animal facility at the University of Tokyo, and experiments were conducted in accordance with the guidelines provided by the Animal Care and Use Committee of the University of Tokyo.
Lymphocytes were isolated from the spleen, PPs, and iLP, as previously described (23, 24, 25). Briefly, single-cell suspensions were prepared from the spleen by passing them through a 70-μm mesh filter. To isolate lymphocytes from the PPs, an enzymatic dissociation protocol with collagenase (Nitta gelatin) was used. To isolate lymphocytes from the iLP, after removing PPs and isolated lymphoid follicles, small intestines were cut into 2-cm pieces and stirred in RPMI 1640 containing 1 mM EDTA and 2% FCS. The tissues were then stirred in 0.5 mg/ml collagenase before undergoing a discontinuous Percoll gradient centrifugation. Lymphocytes were isolated at the interface between the 40 and 75% layers.
Flow cytometry and cell sorting
Flow cytometry and cell sorting were performed, as previously described (23, 24, 25). Cells were preincubated with 10 μg/ml anti-CD16/32 Ab (BD Biosciences) and then stained with fluorescent Abs specific for B220, CD138, α4β7 integrin, IgA, IgM (BD Biosciences), peanut agglutinin (PNA; Vector Laboratories), and CCR9 (R&D Systems). A Viaprobe (BD Biosciences) was used to discriminate between dead and living cells. Cytofix and Cytoperm kit (BD Biosciences) and ethidium monoazide bromide (Invitrogen Life Technologies) were used for intracellular IgA staining and for discriminating between dead and live cells, respectively. Stained cells were then subjected to flow cytometric analysis using FACSCalibur (BD Biosciences). For B cell purification, T cells were depleted using biotin-conjugated Abs specific for CD4 and CD8α together with streptavidin-conjugated magnetic-activated cell sorter beads (Miltenyi Biotec). These T cell-depleted PP cells were subjected to cell sorting using FACSAria (BD Biosciences).
In vitro migration assay
In vitro migration assays using PP B cells were performed according to a previously established method (16, 22). Briefly, PP B cells were isolated from mock- or FTY720-treated mice. These cells were stained with appropriate fluorescence-conjugated anti-IgA, IgM, and B220 Abs; washed three times with RPMI 1640 medium containing 0.5% fatty acid-free BSA (Calbiochem); and applied to the upper chambers of Transwell plates (pore diameter, 5 μm; Corning-Costar). Various concentrations of S1P (0, 1, 10, or 100 nM; Sigma-Aldrich) were added to the lower wells. After a 3-h incubation, the B cells that had migrated into the lower wells were enumerated by flow cytometry.
Analysis of OVA-specific Ab responses and total Ab levels by ELISA and ELISPOT
One week after the last immunization, fecal samples were collected and lymphocytes were isolated for the enumeration of OVA-specific Ab responses by ELISA and ELISPOT, respectively. Standard OVA-specific ELISA and ELISPOT were performed, as previously described (28, 29). In separate experiments, the total numbers of blood-circulating IgA-forming cells and the amounts of serum IgA in nonimmunized mice were determined by ELISPOT and ELISA, as described previously (23).
Immunohistochemical analysis was performed, as previously described (24). Briefly, PPs were fixed in 4% paraformaldehyde for 15 h at 4°C, washed, and treated in 20% sucrose for 12 h at 4°C. The tissues were embedded in OCT compound (Sakura Finetechnical). Cryostat sections (7 μm) were preblocked with an anti-CD16/CD32 Ab for 15 min at room temperature and stained with fluorescent-conjugated PNA lectin or Abs specific for IgA and B220 for 15 h at 4°C. Counterstaining was performed using 4′,6′-diamidino-2-phenylindole (Sigma-Aldrich). Podoplanin, CD4, and CD11c were stained using the TSA-Direct kit (PerkinElmer), according to the manufacturer’s instructions (30). Briefly, cryostat sections (7 μm) were treated with 3% H2O2 for 10 min to quench endogenous peroxidase activity. Sections were blocked with anti-CD16/CD32 Ab in TNT buffer (0.1 M Tris-HCl (pH 7.5), 0.15 M NaCl, 0.05% Tween 20) for 15 min at room temperature. Next, sections were stained with purified anti-podoplanin Ab (Acris Antibodies) plus biotin-conjugated anti-hamster IgG mixture (BD Biosciences), biotin-conjugated anti-CD4 Ab (BD Biosciences), or biotin-conjugated anti-CD11c Ab (BD Biosciences) overnight, followed by incubation with HRP-conjugated streptavidin (Pierce) for 30 min at 4°C and amplification of the fluorescent signal with Cy5-tyramide. The specimens were analyzed using a confocal laser-scanning microscope (TCS SP2; Leica Microsystems).
In vitro culture of PP B cells
To measure IgA production by IgA+B220− B cells, purified PP IgA+B220− B cells (104 cells/well) were cultured in U-bottom, 96-well microtiter plates with 500 pg/ml murine IL-6 (R&D Systems) for 72 h (31). The amount of IgA in the culture supernatant was determined by ELISA, as described previously (23, 31). To induce differentiation of PP B cells, stromal cells were isolated from iLP and cocultured (4000 cells/well) with purified IgM+ B220+ PP B cells (105 cells/well) in the presence of 1 μg/ml anti-CD40 Ab (BD Biosciences), 100 ng/ml murine IL-5 (R&D Systems), and 1 ng/ml human TGF-β (PeproTech) for 6 days with a protocol previously described (32). These differentiated B cells were purified by FACSAria for subsequent RT-PCR analysis.
To measure mRNA expression for S1P1, quantitative RT-PCR using LightCycler (Roche Diagnostic Systems) was performed, as previously described (23, 24, 25). Total RNA was prepared using TRIzol reagent (Invitrogen Life Technologies), and cDNA was synthesized using Powerscript reverse transcriptase (BD Biosciences). The oligonucleotide primers and probes specific for S1P1 (forward primer, 5′-TACACTCTGACCAACAAGGA-3′; reverse primer, 5′-ATAATGGTCTCTGGGTTGTC-3′; FITC probe, 5′-TGCTGGCAATTCAAGAGGCCCATCATC-3′; LCRed 640 probe, 5′-CAGGCATGGAATTTAGCCGCAGCAAATC-3′) and GAPDH (forward primer, 5′-TGAACGGGAAGCTCACTGG-3′; reverse primer, 5′-TCCACCACCCTGTTGCTGTA-3′; FITC probe, 5′-CTGAGGACCAGGTTGTCTCCTGCGA-3′; LCRed 640 probe, 5′-TTCAACAGCAACTCCCACTCTTCCACC-3′) were designed and produced by Nihon Gene Research Laboratory.
The results for treatment vs control groups were compared using Student’s t test or Welch’s t test. Statistical significance was established at p < 0.05.
Alteration of S1P1 expression during B cell differentiation in the PPs
We first investigated S1P1 expression at different B cell developmental stages based on the expression pattern of Ig H chain in the PPs. B cells isolated from the PPs of naive mice consisted of three distinct Ig expression patterns, as follows: IgM+IgA− (70%), IgM+IgA+ (1%), and IgM−IgA+ (4%; Fig. 1,A, left panels). The first two populations exclusively expressed B220, whereas IgM−IgA+ B cells were composed of IgA+B220+ B cells and IgA+B220− B cells (Fig. 1,A, right panel). Staining of intracellular IgA showed that no expression of intracellular IgA was noted in IgA+B220+ cells, whereas high levels of intracellular IgA were expressed in IgA+B220− cells (Fig. 1,B). These intracellular IgA expressions in PP IgA+B220− cells were similar to those in IgA+B220− B cells in the iLP, which are predominant population of iLP B cells (Fig. 1,A, right panel, and Fig. 1 B). These findings suggest that naive IgM+B220+ B cells switch to IgM−IgA+B220+ cells through IgM+IgA+B220+ B cells under the control of CSR and then IgM−IgA+B220+ cells further differentiated to IgM−IgA+B220− PC precursors in the PPs (6).
To determine the S1P1 expression at different stages of B cell differentiation in the PPs, we performed S1P1-specific quantitative RT-PCR using purified PP B cells. Naive IgM+IgA−B220+ B cells expressed high levels of S1P1, and the S1P1 expression was markedly decreased (∼40-fold) in PP B cells that underwent IgA commitment, such as IgM+IgA+B220+ B cells (Fig. 1,C). Whereas the IgM−IgA+B220+ B cells retained a low level of S1P1 expression, the expression on IgM−IgA+B220− B cells increased to a level similar to that of IgM−IgA+B220− B cells located in the iLP (Fig. 1 C). These results demonstrate that S1P1 expression was altered during B cell differentiation to IgA+ cells in the PPs.
Regulation of S1P1 expression in in vitro differentiated PP B cells
We next investigated whether S1P expression is similarly regulated in in vitro differentiated PP B cells. To address this issue, we used in vitro B cell differentiation model using purified PP IgM+IgA−B220+ B cells (32). Six days coculture of IgM+IgA−B220+ B cells with iLP stromal cells plus IL-5, TGF-β, and anti-CD40 Ab induced IgM−IgA+B220+ B cells (3.6%) and further differentiated IgM−IgA+B220− plasmablasts (3.5%) (Fig. 2,A). Like in vivo differentiated B cells (Fig. 1), in vitro differentiated IgM−IgA+B220+ B cells showed decreased levels of S1P1 (Fig. 2,B). However, unlike in vivo differentiated IgM−IgA+B220− B cells (Fig. 1), the S1P1 expression was not recovered on in vitro differentiated IgM−IgA+B220− cells (Fig. 2 B). These data suggested that the recovery of S1P1 expression was not simply coincided with the differentiation to IgM−IgA+B220− cells and required additional unknown factor(s).
FTY720 treatment causes accumulation of IgA+B220− plasmablasts in PPs
We next investigated whether the alteration of S1P1 expression during B cell differentiation in the PPs was truly associated with B cell commitment to CMIS-mediated cell trafficking, especially emigration from the PPs. To address this issue, we used FTY720, an immunomodulator of S1P receptors (18). When we analyzed IgM+ and/or IgA+ B cells in the PPs of mice treated i.p. with FTY720 for 5 days, no significant difference was noted among IgM+IgA−, IgM+IgA+, and IgM−IgAlow B cells, whereas the relative abundance and total number of IgM−IgAhigh B cells were increased (Fig. 3,A). These increased IgM−IgAhigh B cells were B220− (Fig. 3,A), which was consistent with the high expression of S1P1 on IgA+B220− B cells (Fig. 1,C). The accumulation of IgAhighB220− B cells in the PPs was coincident with the reduction of the same population in the iLP and IgA Ab-forming cells (AFCs) in the blood (Fig. 3, B and C). The 5-day treatment with FTY720 did not affect the amount of serum IgA, probably due to the t1/2 of IgA and IgA-forming cell (our unpublished data), whereas the 4-wk treatment with FTY720 resulted in decreased amounts of serum IgA (Fig. 3,C). Because IgA+B220− cells include both plasmablasts and PCs, we next examined which of these cell types FTY720 targeted. FTY720 did not change the frequency of IgA+ PCs (CD138highB220−) in the PPs, but did cause accumulation of IgA+ plasmablasts (CD138int/−B220−) in the PPs (Fig. 3,D). In agreement with this observation, the number of CD138int/−IgA+ plasmablasts was reduced in the iLP of FTY720-treated mice without affecting the number of IgA+ PCs (Fig. 3 D).
We next analyzed the kinetics of cell recovery after FTY720 treatment. Five treatments with FTY720 resulted in the accumulation of IgA+B220− B cells in the PPs and simultaneous reduction in the iLP (Fig. 3,E). Partial recuperation occurred by day 4, with full recovery observed 8 days after the termination of FTY720 treatment (Fig. 3 E). These data suggest that the inhibition of S1P-mediated signaling by FTY720 reversibly hampered the migration of B cells from the PPs into the iLP.
To directly investigate whether IgA+ plasmablasts migrated in response to S1P and whether this response was inhibited by FTY720 treatment, we performed an in vitro Transwell migration assay. Consistent with S1P1 expression (Fig. 1,C), S1P1+IgM+B220+ B cells and S1P1+IgA+B220− plasmablasts, but not S1P1−IgA+B220+ B cells isolated from mock-treated mice migrated toward S1P (Fig. 4). In addition, both S1P1+IgM+B220+ B cells and S1P1+IgA+B220− plasmablasts from FTY720-treated mice failed to migrate in response to S1P (Fig. 4), which is in accord with previous data that FTY720 treatment abolished the reactivity to S1P in lymphocytes (16, 22). These data suggested that the altered S1P1 expression on B cells in the PPs could be a key biological determinant of whether PP B cells stay or emigrate from the PPs, and that S1P plays a key role in regulating the emigration of S1P1+IgA+ plasmablasts from the PPs.
FTY720 does not influence the expression of gut-homing molecules on and IgA production by IgA+ plasmablasts
An obvious explanation for the effect of FTY720 on the emigration of IgA+ plasmablasts would be the influence of S1P on the expression of gut-homing-associated adhesion molecules and chemokine receptor. Therefore, we examined the expression of the adhesion molecule, α4β7 integrin, and the chemokine receptor, CCR9, which determine gut tropism (33, 34). Flow cytometric analysis demonstrated that accumulated IgA+B220− plasmablasts expressed high levels of α4β7 integrins and CCR9 (Fig. 5,A), indicating that accumulation of IgA+B220− plasmablasts in the PPs and their simultaneous decrease in the iLP were not due to a lack of migration ability through these gut-homing molecules. In addition, these accumulated IgA+B220− plasmablasts were capable of producing substantial amounts of IgA. Indeed, IgA+B220− plasmablasts isolated from the PPs of mock- or FTY720-treated mice showed comparable levels of IgA production induced by IL-6 treatment (Fig. 5 B). Consistent with these results, we also found that FTY720 did not affect the distribution of B220+ B cells, CD4+ T cells, or CD11c+ DCs in the PPs (our unpublished data), which are all involved in the appropriate induction of IgA-committed B cells, including CSR and expression of gut-homing molecules (1, 2, 7). These findings suggested that FTY720 did not influence the immunological nature of IgA-committed B cells, including their gut-homing molecules, and differentiation to Ab production ability; FTY720 solely affected their egress from the PPs.
FTY720 inhibits emigration of IgA+ plasmablasts from the lymphatic area of the basal side of PPs
We next performed confocal microscopic analysis to determine the sites in which IgA+ plasmablasts accumulated after FTY720 treatment. In mock-treated mice, IgA+ B cells were found in the GCs, follicle-associated epithelium, and lymphatic area of the basal side of PPs (Fig. 6,A). In FTY720-treated mice, however, IgA+ B cells accumulated only on the basal side of PPs (Fig. 6,B, arrows). Some IgA+ B cells bound to podoplanin+ lymph (Fig. 6 B, bottom, arrowheads). These data clearly indicate that S1P regulates the emigration of IgA+ plasmablasts from the lymph around the basal side of PPs without affecting other immunological functions, including expression of gut-homing molecules, class switching to IgA+ B cells, and the ability to differentiate to IgA-producing cells.
FTY720-mediated inhibition of IgA+ plasmablast emigration from the PPs abolishes the subsequent induction of Ag-specific intestinal IgA production
We next examined whether the S1P-mediated regulation of IgA+ plasmablast emigration from the PPs is crucial for the induction of efficient Ag-specific Ab responses against orally administered Ag. To address this issue, mice were orally immunized with OVA plus cholera toxin, a mucosal adjuvant. An ELISPOT assay revealed that OVA-specific IgA AFCs were induced in the iLP after oral immunization in control mice (Fig. 7,A). In contrast, daily treatment with FTY720 during immunization resulted in a decreased number of OVA-specific IgA AFCs in the iLP, which was associated with the accumulation of OVA-specific IgA AFCs in the PPs (Fig. 7,A). In nonimmunized control mice, negligible levels of OVA-specific IgA AFCs were detected (mock PP, 0.5 ± 0.14; FTY720 PP, 1.4 ± 0.22; mock LP, 3.5 ± 0.29; FTY720 LP, 3.5 ± 0.28 cells/106 cells). Consistent with the results obtained from the analysis of nonimmunized mice treated with FTY720 (Fig. 3,D), the increase of OVA-specific IgA AFCs in the PPs was coincident with the accumulation of IgA+B220−CD138−/low plasmablasts in the PPs (Fig. 7,B). We also measured the levels of OVA-specific IgA in feces to examine whether the altered trafficking of IgA+ plasmablasts affected actual Ab production in the intestinal lumen. We found that OVA-specific fecal IgA was markedly decreased in the FTY720-treated mice (Fig. 7 C). These data clearly indicate that the migration of Ag-specific IgA+ plasmablasts from the PPs into the iLP is a prerequisite for the efficient production of Ag-specific S-IgA Abs in the intestinal lumen.
Previous studies reported that FTY720 treatment inhibited Ab responses against systemically immunized T-dependent Ag by abolishing GC formation in the systemic lymph nodes (19, 20, 21). Thus, we next examined GC formation in the PPs and spleen. In naive mice, there were few PNAhigh B220+ GC B cells in the spleen, and 4% of lymphocytes were PNAhigh B220+ GC B cells in the PPs, and FTY720 barely affected these populations in naive condition (our unpublished data). When we examined mice orally immunized with OVA plus cholera toxin, the numbers of PNAhigh B220+ GC B cells were increased in the spleen, but not significantly changed in the PPs compared with naive mice (Fig. 8,A). Consistent with the previous findings (19, 20, 21), reduced numbers of PNAhigh B220+ GC B cells were seen in the spleen of mice receiving FTY720 during the oral immunization period (Fig. 8,A). In contrast, comparable numbers of PNAhigh B220+ GC B cells were detected in the PPs of FTY720-treated mice (Fig. 8,A). Histological analysis confirmed that GC formation in the PPs was not impaired by FTY720 treatment. Thus, comparable GC formation containing B220+ PNA+ cells was noted in the FTY720-treated mice (Fig. 8 B). These results indicate that, unlike systemic immune compartments (19, 20, 21), FTY720 reduced intestinal S-IgA production against orally inoculated Ag by inhibiting IgA+ plasmablast emigration from the PPs without affecting GC formation in the PPs.
More than 30 years ago, PPs were discovered to be the induction sites of precursor of IgA-producing PCs (5). The current study indicates that PPs are the main sites for the induction of IgA+ plasmablasts, which selectively use S1P for emigration from the PPs by changing their S1P1 expression during differentiation to IgA+ B cells in the PPs. Because a previous study showed that activation of B cells by LPS or BCR stimulation induced the reduction of S1P1 expression (19, 20, 21), it is likely that PP milieu containing abundant IL-4- or TGF-β-producing cells allows activated B cells to differentiate to IgA+ B cells with simultaneous reduction of S1P1. The reduction of S1P1 expression enables IgA+ B cells to stay at the PPs for efficient differentiation to IgA+ plasmablasts. The IgA+ plasmablasts then recover S1P1 expression together with the expression of gut-homing molecules (e.g., α4β7 integrin and CCR9), allowing them to exit from the PPs and migrate into the iLP for the final differentiation to IgA-producing PCs. This S1P-mediated regulation system observed in this study is in agreement with a previous data on the systemic immunization model (22). It was shown that differential expression of S1P1 and Kruppel-like factor 2 (KLF2), a transcription factor that increases S1P1 expression (35), in differentiating PCs is a key factor to determine their trafficking from spleen to the bone marrow. Indeed, high levels of S1P1 and KLF2 were noted in Blimp1int CXCR4int IgG AFCs in the spleen and blood (22).
In addition, our current results provide further evidence that the differentiation of IgM+ B cells to IgA+ plasmablasts in the PPs is not sufficient for effective intestinal IgA production against intestinal Ags, but that the cells require appropriate egress from PPs to the iLP for the final differentiation to IgA synthesis. This idea was previously proposed based on studies showing that intestinal S-IgA production was impaired in mice deficient in gut-homing molecules, such as α4β7 integrin and CCR9 (34, 36). Our results convincingly demonstrate the role of lipid mediator for the egress of IgA-committed B cells from the inductive tissue by showing for the first time that intestinal S-IgA production against ingested Ag was impaired by inhibiting the S1P-mediated emigration of IgA+ plasmablasts from the PPs, without affecting their expression of gut-homing molecules and their ability to produce IgA (Figs. 5 and 7). Our current findings do not exclude the possibility that FTY720 inhibits the emigration of IgA+ plasmablasts via prevention of endothelial cell function, which was previously reported to express S1P receptors (37). However, our current in vitro data on S1P1 expression on PP B cells (Fig. 1) and their migration in response to S1P (Fig. 4) and in vivo data on FTY720-treated mice (Fig. 3) collectively suggest that S1P1 on PP B cells itself is responsible for the FTY720-sensitive emigration from the PPs.
Although the recovery of S1P1 expression was noted in IgM−IgA+B220− plasmablasts in the PPs in vivo (Fig. 1), in vitro differentiated IgM−IgA+B220− cells did not show the S1P1 recovery (Fig. 2 B). It was possible that continuous stimulation of differentiated IgM−IgA+B220− cells with the cytokines and anti-CD40 Ab might inhibit the recovery of S1P1 expression because previous studies indicated that several hours were required for the full recovery of S1P1 (38, 39). To test this possibility, we removed the CSR-related molecules after their differentiation to IgM−IgA+B220− cells, but we found that S1P1 expression was still low after removing the CSR-related molecules (our unpublished data). Additionally, coculture with whole PP cells in the presence of the CSR-related molecules resulted in the induction of the IgM−IgA+B220− cells, but they still did not recover the S1P1 expression (our unpublished data). Thus, it seems that well-organized structure of PP and/or some unknown factors are required for the S1P1 recovery during differentiation to IgM−IgA+B220− plasmablasts in the PPs. Related to the imprinting system in the gut-associated lymphocyte trafficking, accumulating evidence has shown that the expression of gut-homing molecules (e.g., α4β7 integrin and CCR9) is induced by the interaction with gut-associated DCs producing retinoic acid (7). In contrast to the retinoic acid-mediated expression of α4β7 integrin and CCR9, retinoic acid did not affect the S1P1 expression of PP B cells in vitro (our unpublished data). Another possibility is that cytokines produced in the intestinal compartment may cooperatively regulate S1P1 expression. In this context, the expression of KLF2, a transcription factor for the increase of S1P1 expression (22), is regulated by several cytokines, including IL-7, which is abundantly produced by intestinal epithelial cells (40, 41). Thus, it is possible that gut-associated cytokines (e.g., IL-7) may enhance the re-expression of S1P1 in PP IgA+ B cells, but this circumstance is not established in vitro because freshly isolated epithelial cells cannot retain the ability in vitro due to their low viability. These points, including the effect of organized structure of PPs, represent challenges for future studies.
Our current findings suggest that S1P is not involved in cellular distribution in the PPs because FTY720 treatment did not affect the distribution of CD4+ T cells, B220+ B cells, and CD11c+ DCs in the PPs, with the exception of the IgA+ plasmablast emigration from the lymphatic area of the basal side of PPs (Fig. 6 and our unpublished data). In contrast, B cell distribution in the spleen was affected by FTY720 (12, 19). Moreover, unlike the inhibitory effects of FTY720 on GC formation in the spleen after systemic immunization (19, 20, 21), we show that FTY720 did not affect GC formation in the PPs (Fig. 8). A major difference between the spleen and PPs is that spleen is located in germfree condition, whereas PPs are exposed to continuous stimulation by environmental Ags (e.g., microbial and food Ags), which may account for these different effects of FTY720. B cells in the systemic immune compartments (e.g., spleen) are normally in a quiescent state, and thus, no GC formation is detected in intact mice. In contrast, PPs contain GCs in intact mice that are induced by stimulation from intestinal microbiota (26). Thus, FTY720 may inhibit the formation of new GCs, such as GC formation in the spleen induced by immunization, but does not interfere with established GCs in the PPs.
Another unresolved question is why IgA+ plasmablasts selectively accumulated by FTY720 treatment, yet IgM+ B cells, which express high levels of S1P1 and show comparable reactivity to S1P in vitro, did not change (Figs. 1, 3, and 4). Similar selective effects of FTY720 on IgA+ B cells were observed in our previous study on peritoneal B cell trafficking into the iLP (23). In that study, we found that IgA+ B cells, but not IgM+ B cells, were inhibited by FTY720 treatment from migrating from the peritoneal cavity into the iLP (23). There is a mutual interaction between S1P and chemokines in lymphocyte trafficking, with some levels of hierarchy (19, 22, 42, 43, 44, 45), so a cooperative pathway mediated by both S1P and chemokines may determine the selective effects of FTY720 on IgA+ B cells. Indeed, it was previously reported that CCR10 expression was prevalent on IgA+ B cells with plasmablast and PC phenotypes in the blood and the intestine, but expression was negligible on IgA− B cells (46). By contrast, IgM+ B cells, but not IgA+ B cells, predominantly expressed CCR7 (our unpublished data), which was reported to regulate S1P-mediated T cell trafficking together with additional Gαi-coupled receptors (45). Therefore, the expression of identified and/or unidentified chemokine receptor(s) on IgM+ or IgA+ B cells may determine their dependency (or lack thereof) on S1P for emigration from the PPs. Our group is currently conducting research on the involvement of gut-associated cytokines and chemokines in the regulation of S1P-mediated intestinal B cell trafficking.
In addition to the IgA+ plasmablasts, the few cells showing IgA+ PC phenotypes (e.g., CD138high) in the PPs were not affected by FTY720 treatment (Fig. 3,D), suggesting the presence of an alternative, S1P-independent differentiation pathway to IgA+ PCs in the PPs. In this context, a recent study demonstrated that PP-DCs produce retinoic acid, IL-5, and IL-6, which provide a milieu for class switching from μ- to α-chain, as well as IgA production (7). IgAhigh cells in the follicle-associated epithelium (FAE) were barely affected by FTY720 treatment (Fig. 5), and DCs are abundant in the FAE (30). Thus, the DCs may induce IgA+ PCs in the FAE in a S1P-independent manner. This idea is supported by a previous report that B cells near M cells generally situated in FAE region showed a memory cell phenotype (47). Hence, S1P-independent IgA+ cells in the FAE might provide a rapid response against newly arriving Ags for creating local immunity in the PPs.
Our previous work indicated the pivotal role of S1P in the regulation of peritoneal B cell trafficking for intestinal S-IgA production (23), and our current findings show that B cells can alter S1P1 expression to regulate their retention on and emigration from PPs. Together, these findings show that S1P is a key molecule in the versatile S-IgA production pathways in the intestinal immune system.
We thank Dr. K. Kabashima (University of Occupational and Environmental Health, Japan) for helpful comments on this article. We thank Novartis Pharmaceuticals for providing FTY720.
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 grants from Core Research for Evolutional Science and Technology of Japan Science and Technology; Ministry of Education, Science, Sports, and Culture; Ministry of Health and Welfare in Japan; Waksman Foundation; Yakult Bio-Science Foundation; and Mochida Memorial Foundation for Medical and Pharmaceutical Research.
Abbreviations used in this paper: S-IgA, secretory IgA; AFC, Ab-forming cell; CMIS, common mucosal immune system; CSR, class switch recombination; DC, dendritic cell; GC, germinal center; iLP, intestinal lamina propria; int, intermediate; KLF2, Kruppel-like factor 2; PC, plasma cell; PNA, peanut agglutinin; PP, Peyer’s patch; S1P, sphingosine 1-phosphate; S1P1, type 1 S1P receptor; FAE, follicle-associated epithelium.