Rap1b Regulates B Cell Development, Homing, and T Cell-Dependent Humoral Immunity¹

The small GTPase, Rap1, belongs to the Ras superfamily and cycles between an inactive GDP-bound and an active GTP-bound conformation (1, 2). This ubiquitously expressed molecular switch regulates cell proliferation, differentiation, and adhesion by distinct mechanisms through integrin activation and MAPK cascades (3, 4). Multiple extracellular stimuli can activate Rap1. This activation is mediated by various guanine nucleotide exchange factors (GEFs),³ such as C3G, CalDAG-GEF, and EPAC (5, 6, 7). C3G is activated by receptor-associated protein tyrosine kinases (5). CalDAG-GEF responds to calcium and diacylglycerol, downstream of phospholipase C signaling (8). EPAC is a cAMP-dependent GEF (7). Active Rap1-GTPase recruits multiple effectors (e.g., RapL/Nore1B and RIAM) to execute their cellular functions (9, 10). RapL, a Rap1-binding molecule, is a major effector of Rap1-GTPase in immune cells and associates with α_Lβ₂ integrin (9). RIAM contains Ras-associated and pleckstrin homology domains and interacts with profilin and Ena/VASP proteins that regulate actin dynamics (10). Rap1 signaling is terminated by GTPase-activating proteins (GAPs), such as SPA-1 (signal-induced proliferation-associated protein 1) and RapGAPs, which convert active GTP-bound to inactive GDP-bound form (11, 12).

Integrins mediate cellular contacts to extracellular matrix or to their counter receptors, thereby regulating cell motility, polarity, growth, and survival (13). Rap1 is expressed in multiple hematopoietic cells and one of its key functions is to regulate integrin-mediated cell adhesion (4). This amplifies a series of hematopoietic cellular processes including aggregation, migration, extravasation, and homing (4, 14, 15). Mice lacking CalDAG-GEFI show impaired platelet adhesion and aggregation leading to impaired thrombus formation (16). Deficiency of C3G, one of the major Rap-specific GEFs, resulted in severe embryonic lethality, partly due to aberrant cell adhesion and spreading (17). Overexpression of Rap1, RapGEFs, GAPs, or their mutant proteins altered the activation of Rap1 and helped to uncover its signaling functions in different cell types (18, 19, 20, 21). Rap1 can be activated through Ag or chemokine receptors (18, 22, 23). Active Rap1 plays an important role in immunological synapse formation between T cells and APCs (22).

Recent studies have well documented that interference in Rap1 functions can result in severe pathological conditions. In human, leukocyte adhesion deficiency syndrome III is associated with defective Rap1 activation and impaired stabilization of integrin bonds (24). Additionally, knockout mice for SPA-1 or RapL have provided critical insights into in vivo functions of Rap1-GTPase (25, 26). SPA-1 is the major Rap1GAP in the hematopoietic progenitors in bone marrow (26). SPA-1^−/− mice show an increase in the hematopoietic stem cells and eventually develop leukemia (26). SPA-1 deficiency also results in aberrant BCR repertoire and autoimmune diseases (27). RapL deficiency causes defective chemokine-triggered lymphocyte adhesion and migration (25). However, many of these RapGEFs, GAPs, and effectors are promiscuous and can associate with multiple Rap1 GTPases; therefore the direct evaluation of Rap1 functions is limited in the described knockout mouse models.

Rap1 is closely related to the Rap2 family (Rap2a/2b/2c) and shares 60% homology (1, 2, 28). Rap1 exists in two isoforms, Rap1a and Rap1b, which are encoded by different genes. Rap1 isoforms share more than 95% amino acid homology and differ only in eight amino acids with five located at their C termini. Due to the high homology of Rap1a and Rap1b, their functions have been studied undistinguished in some studies (22, 23, 29). However, the different subcellular localization and tissue expression patterns of these two isoforms have been described and indicate their distinct functions (30, 31, 32, 33, 34, 35). Studies have indicated the role of Rap1 isoforms in neuron, T cells, and platelets (19, 34, 35). Rap1a has been shown to positively regulate T cell functions in both transgenic and knockout mouse models (19, 36). Recently, we have generated Rap1b gene knockout mice (34). Approximately 40% of the Rap1b^−/− fetuses die during embryogenesis. Surviving Rap1b^−/− mice show bleeding disorder due to reduced platelet aggregation that is similar to the phenotype in CalDAG-GEFI^−/− mice (16, 34). Irrespective of these findings, the differential expression of Rap1 isoforms and their specific roles in B cell have yet to be determined. In this study, using Rap1b^−/− mice, we show that Rap1b is the dominant Rap1 isoform in B cells. Rap1b is crucial for early B cell development, marginal zone (MZ) B cell maturation, and T-dependent humoral responses. Additionally, absence of Rap1b leads to the reduced B cell adhesion and chemotaxis and in vivo homing. However, Rap1b is dispensable for BCR-mediated splenic B cell phosphorylation of Erk1/2, p38, and AKT.

Rap1b^−/− mice were previously described (34). The Rap1b^+/− mice were crossed with Rap1b^+/− or Rap1b^−/− mice to obtain Rap1b^+/+ and Rap1b^−/− mice. The following Abs and reagents were used in this study: recombinant proteins Rap1a-GST and Rap1b-GST (85–184 aa) (Novus Biologicals); rabbit anti-Rap1 polyclonal Ab, goat anti-rabbit mouse IgG (H+L) HRP (Santa Cruz Biotechnology); rabbit anti-Rap1b mAb (36E1), rabbit anti-Erk1/2 polyclonal Ab, rabbit anti-phospho-Erk1/2 mAb (197G2), rabbit anti-p38 polyclonal Ab, rabbit anti-phospho-p38 mAb (3D7), rabbit anti-AKT polyclonal Ab, rabbit anti-phospho-AKT (Ser⁴⁷³) mAb (193H12), and anti-GST mAb (Cell Signaling Technology); rabbit anti-β-actin Ab (Abcam); rat anti-mouse B220 (RA3-6B2), rat anti-mouse IgM (R6-60.2), rat anti-mouse IgD (11-26c.2a), rat anti-mouse CD19 (1D3), rat anti-mouse CD1d (1B1), rat anti-mouse CD21/CD35 (4E3), rat-anti-mouse CD23 (B3B4), Armenian hamster anti-moue CD3ε (145-2C11), rat anti-mouse CD16/CD32(93), rat anti-mouse CD43, rat anti-mouse CD25 and IL-4 (eBioscience); rat anti-mouse CD4 (H129.19), rat anti-mouse c-Kit, and anti-mouse CD8 (53-6.2) (BD Pharmingen); MOMA-1-FITC (Serotec); goat anti-mouse IgM F(ab′)₂ (Jackson ImmunoResearch Laboratories); mouse anti-GAPDH LPS (Escherichia coli O111:B4; Sigma-Aldrich); CFSE (C-1157) and TAMRA (C-300) (Molecular Probes); and stromal cell-derived factor (SDF)-1 and CXCL13 (R&D Systems).

Single-cell suspensions were prepared by gently mincing the dissected organs through 70-μm cell strainers. T cells and resting B cells were purified by anti-CD3-PE mAb/anti-PE Microbeads (BD Pharmingen) and B cell isolation kit (Miltenyi Biotec). Separated cells were stained with anti-CD19 and anti-CD3ε mAbs, and their purity was >95% as confirmed by flow cytometry. Flow cytometry analysis was conducted in LSR-II and analyzed with FACSDiva software (BD Biosciences). For isolating newly forming, follicular (FO), and MZ B cell populations, single-cell suspensions from spleens were stained with anti-B220-Pacific Blue, anti-CD21-FITC, and anti-CD23-PE-Cy7 mAbs and sorted using FACSAria (BD Biosciences).

Spleens were embedded in Tissue-Tek (Sakura Finetek), frozen with dry ice-isopropanol mixture, and kept at −80°C until sectioning. The 10-μm cryostat sections were generated, air-dried for 1 h at room temperature, and fixed in acetone for 10 min. The sections were blocked with 0.1% BSA/PBS containing anti-CD16/CD32 mAb for 45 min, incubated for 1 h at room temperature in dark with the rat anti-mouse CD3ε-PE, anti-B220-Cy5, MOMA-1-FITC, anti-mouse IgM-PE, or anti-mouse IgD-FITC mAbs. Sections were mounted in aqueous mounting medium and visualized on inverted Nikon Eclipse TE200 fluorescence microscope. Images are at a magnification of ×40.

Bone marrow cells from tibias and femurs of wild-type (WT) and Rap1^−/− mice were prepared and resuspended in 2% FBS/RPMI 1640 (Invitrogen Life Technologies). A total of 5 × 10⁶ bone marrow cells (200 μl) were injected i.v. to reconstitute 6- to 8-wk-old, sublethally irradiated (300 rad) JAK3^−/− and lethally irradiated (1100 rad) Rap1b^−/− mice 4 h after the irradiation. The recipient mice were sacrificed 8 wk later and the spleens were analyzed for reconstitution of B cells by flow cytometry.

B cells were purified from spleens and resuspended in HEPES buffer (10 mM HEPES/0.5% FCS/RPMI 1640). A total of 15 × 10⁶ B cells were used for each lane. Cells were stimulated with 50 μg/ml goat anti-mouse IgM F(ab′)₂ or 100 ng/ml SDF-1 for 2 min at 37°C or left unstimulated. Stimulation was terminated by adding one volume of 2X ice-cold lysis buffer (100 mM Tris (pH 7.4), 150 mM NaCl, 2% Nonidet P-40, 1% deoxycholic acid, 0.2% SDS, 2 mM sodium orthovanadate, with Phosphatase Inhibitor Cocktail (PhosSTOP; Roche Diagnostics)) and Proteinase Inhibitor Cocktail (Sigma-Aldrich). Lysates were incubated for 30 min on ice and centrifuged at 15,000 × g for 15 min at 4°C. An aliquot of cell lysates (1 × 10⁶ cells) was taken out for evaluating the total Rap1 loading. The remaining cell lysates were incubated with glutathione-agarose beads that precoupled with GST-RalGDS-Ras binding domain at 4°C (37). This assay is based on the fact that RalGDS binds with high affinity to Rap1-GTPase but not to Rap1-GDP (37). The beads were washed, and 30 μl of 1X Laemmli buffer was added and heated at 95°C to elute the pull-down proteins. The active Rap1 was analyzed by an Ab that recognizes both Rap1a and Rap1b. For Western blotting, cell lysate aliquot (corresponding to 1 × 10⁶ cells) or pull-down protein pellets were separated by SDS-PAGE, transferred to polyvinylidene difluoride membrane, and probed with primary and secondary Abs conjugated with HRP. The signal was detected by ECL (GE Healthcare Life Sciences).

Primary pre-B cell cultures were previously described (38). Briefly, erythrocyte-depleted total bone marrow cells were cultured with 2 ng/ml recombinant murine IL-7 (R&D Systems) at 2 × 10⁶/ml for 5–7 days. The medium was replaced with fresh IL-7 every second day. For BrdU labeling, on day 5, bone marrow cells treated with IL-7 were seeded into 24 wells and cultured for another 3 days with IL-7. These cells were pulsed for 3 h with 10 μM BrdU at the end of day 3. Cells were stained with anti-B220 and anti-IgM Abs, resuspended in cold PBS, fixed with ice-cold ethanol (95%), and incubated on ice for 30 min. Cells were fixed or permeabilized with 1% paraformaldehyde/0.01% Tween 20 at 4°C overnight. Cells were resuspended with 50 Kunitz DNaseI/ml in 0.15 M NaCl, 4.2 mM MgCl₂ (pH 5) for 10 min at 37°C, washed, stained with 1/50 anti-BrdU-FITC (BD Biosciences) for 30 min, and analyzed.

Purified B cells were plated into 96-well plates in triplicates (2 × 10⁴ cells/well) and stimulated with 10 μg/ml goat anti-mouse IgM F(ab′)₂, 10 μg/ml goat anti-mouse IgM F(ab′)₂ plus 10 ng/ml IL-4, or 1 μg/ml LPS in 10% FBS for 72 h. A total of 1 μCi of [³H]thymidine (GE Healthcare Life Sciences) was added 16 h before harvest. Labeled DNA from cells was collected on GSC filters (Whatman), and the radioactivity was measured in a microplate scintillation counter.

For the T-independent immune response, 6- to 8-wk-old mice were i.p. immunized with 100 μg of trinitrophenol (TNP)-Ficoll (Biosearch Technologies) in PBS. A total of 50 μl of blood was collected from the retro-orbital venous plexus with capillary tubes on days 0 and 7 after immunization. For the T-dependent response, mice were i.p. injected with 100 μg of TNP-keyhole limpet hemocyanin (KLH) in 200 μl of aluminum hydroxide adjuvant and Imject Alum (Pierce) and boosted with 10 μg of TNP-KLH without adjuvant on day 21. Sera were collected on days 0, 7, and 28. Sera were analyzed by ELISA for basal and TNP-specific Ig isotypes using Clonotyping System (Southern Biotechnology Associates).

Chemotaxis assay was performed in triplicates in 5.0-μm Transwell (5.0-μm pore, 6.5 mm; Costar). The lower compartment contained 600 μl of 150 ng/ml SDF-1 or 500 ng/ml CXCL13 in HEPES buffer (RPMI 1640, 10 mM HEPES, 0.5% FBS). Purified B cells were resuspended to 5 × 10⁶/ml in the same buffer. The 100 μl of B cell suspension was added to the upper chamber and allowed to migrate for 3 h. The number of viable cells that migrated into the lower compartment was determined by flow cytometry and expressed as a percentage of the input cells that were directly added to the lower compartment. For sphingosine 1-phosphate (S1P)-induced migration assay, different concentrations of S1P (Biomol) in HEPES buffer were added to the lower compartments and the RBC-depleted splenocytes from WT and Rap1b^−/− were added to the upper chambers. At 3 h later, the cells that migrated into the lower compartment were stained with anti-B220, anti-CD21, and anti-CD23 Abs.

S1P-induced adhesion of B cells to ICAM-1 was described (39). Briefly, Immulon 4 HBX 96-well flat-bottom plates (ISC BioExpress) were coated with 1% fat-free BSA/PBS, recombinant mouse ICAM-1/Fc 10 μg/ml (R&D Systems), triplicates, keep at 4°C overnight. Splenocytes were incubated in 10% FBS/RPMI 1640 for 30 min at 37°C in a tissue culture flask to eliminate adherent macrophages. Nonadherent floating cells were collected and RBCs were lysed. Cells were resensitized in RPMI 1640 containing 0.5% fatty acid-free BSA and 10 mM HEPES at 37°C for 1 h. The plates were gently washed with PBS twice and blocked with 1% fatty acid-free BSA/PBS at 37°C for 1 h. A total of 3 × 10⁵ cells in 100 μl of buffer with or without 100 nM S1P and incubated at 37°C for 60 min nonadherent cells wells were washed away and adherent cells were released by keeping the plates on ice for 20 min with RPMI 1640/5 mM EDTA. Collected cells were stained for B220, CD23, and CD21 markers. Adhesion was calculated as a percentage of input cells.

To quantify the adhesion of B cell progenitors from bone marrow, IL-7-driven bone marrow cells were labeled on day 7 with 2 μM CFSE and 10 μM TAMRA for 10 min at room temperature, and the reaction was terminated by adding one volume of FBS. Cells were washed and resuspended in 10% FBS medium. A mixture of 0.5 × 10⁶ CFSE- and TAMRA-labeled cells was added onto M2-10B4 confluent stromal cell layer (American Type Culture Collection) and incubated for 5 h. Unbound cells were washed away with warm plain medium and adherent cells were released with 5 mM EDTA/plain medium. Stromal cell layers were removed by filtering before flow cytometry.

RBC-depleted splenocytes from WT and Rap1b^−/− were resuspended to 20 × 10⁶ cells/ml in PBS and labeled with 2 μM CFSE and 10 μM TAMRA as bone marrow pro-B cells and resuspended in 2% FBS medium. A mixture of 20 × 10⁶ CFSE- and TAMRA-labeled cells was i.v. injected into recipient WT mice. At 3 h later, single-cell preparations from spleen, peripheral lymph nodes (LNs) (inguinal and auxiliary), and mesenteric LN of recipient mice and a portion of labeled cells were stained with anti-B220-allophycocyanin or anti-CD3-PE-Cy7 mAbs. The homing efficiency of Rap1b^−/−/WT B and T cells were calculated compared with the input ratio, which was normalized to 1. A dye swap was also performed.

Data were presented as mean ± SE, and statistical significance was calculated using the Student t test. A value of p < 0.05 was considered significant.

Because Rap1 exists in two isoforms, we first analyzed the relative levels of Rap1a and Rap1b in B cells. Expression of total Rap1 protein was assayed using an affinity purified polyclonal IgG that recognizes an identical epitope in the C terminus of both Rap1a and Rap1b isoforms. Fig. 1,A shows that the total Rap1 was readily detectable in the WT-derived B cells. However, only a residual level of Rap1 protein was present in B cells from Rap1b^−/− mice. With recombinant GST-Rap1A full-length protein and GST-Rap1B (85–184 aa), we show that anti-Rap1 has similar affinities to both Rap1a and Rap1b proteins compared with anti-GST Ab (Fig. 1,B). Therefore, the residual protein band in Rap1b^−/− B cells represents Rap1a. Similar Rap1 expression pattern was also observed in T cells (Fig. 1,A). This finding demonstrates that the Rap1b is the major isoform in B and T cells. Rap1b-specific mAb confirmed the absence of Rap1b in B and T cells from Rap1b^−/− mice (Fig. 1 A). The dominant expression of Rap1b suggested its potential role in B cell development and function.

As Rap1b is highly expressed in B cells, we analyzed its role in B cell development. The total cellularity of bone marrow in Rap1b^−/− mice was significantly less compared with that of WT (WT: 45.4 ± 8.8 × 10⁶ cells vs Rap1b^−/−: 31.0 ± 6.4 × 10⁶ cells; n = 6+6 mice; and p = 0.014). The percentage of B220⁺ B cells were comparable between Rap1b^−/− and WT mice. However, because of reduced cellularity, the absolute number of B220⁺ B cells was significantly reduced in Rap1b^−/− mice (Fig. 1,C). Based on B220 and IgM expression, B cells in the bone marrow can be divided into different maturation stages: pro/pre-B cells (B220⁺IgM⁻), immature B cells (B220⁺IgM⁺), and recirculating mature B cells (B220^high). The percentages of pro/pre-B cells were reduced in Rap1b^−/− mice compared with WT. Likewise, the absolute number of both pro/pre-B cells and immature B cells were significantly reduced in Rap1b^−/− mice (Fig. 1,C). In contrast, the number of recirculating mature B cell B220^high population was comparable between Rap1b^−/− and WT mice (Fig. 1,C). B cell development in bone marrow can be further characterized through the expression of CD43 (leukosialin), CD117 (c-Kit), and CD25 (IL-2Rα). Early pro-B cells are marked by the expression of B220⁺CD43⁺, which develop into B220⁺CD43⁻ B cells. As pro/pre-B cells mature, c-Kit expression decreases while CD25 level increases. The relative (WT: 2.6 ± 0.7% and Rap1b^−/−: 2.8 ± 0.5%) and the absolute (WT: 0.41 ± 0.167 × 10⁶ cells and Rap1b^−/−: 0.264 ± 0.076 × 10⁶ cells) B220⁺CD43⁺ cell numbers did not differ significantly between Rap1b^−/− and WT bone marrow (n = 6 + 6 mice; p = 0.0855) (Fig. 1,D). However, B220^mediumCD43⁻ B cells were much lower in Rap1b^−/− in both relative percentages (WT: 35.2 ± 8.4% and Rap1b^−/−: 25.9 ± 10.3%) and absolute number (WT: 6.5 ± 2.09 × 10⁶ cells and Rap1b^−/−: 2.58 ± 1.27 × 10⁶ cells, n = 6 + 6 mice; p = 0.003), which is consistent with the reduced B220⁺IgM⁻ populations (Fig. 1,D, upper). The increased B220⁺c-kit⁺ population in Rap1b^−/− and with relatively decreased B220⁺CD25⁺ population (Fig. 1,D, middle and lower) indicate a delay in B cell maturation in Rap1b^−/− mice. In peripheral blood, there is an increased IgM⁺IgD⁻ B cell population in Rap1b^−/− mice (Fig. 1 E). Thus, the lack of Rap1b resulted in a developmental impairment of B cell progenitor maturing into pro/pre-B cell and immature B cell stages.

To further gain the insight of B development in Rap1b^−/− mice, we conducted in vitro proliferation of pro/pre-B cells derived from IL-7-driven bone marrow culture. The reduced pro/pre-B cells in the bone marrow of Rap1b^−/− mice may be due to either reduced proliferation or adhesion to stromal cells. Bone marrow cells cultured with IL-7 contained similar purity of B220⁺ and B220⁺IgM⁻ B cells in WT and Rap1b^−/− mice (Fig. 1,F, top). In addition, the Rap1b^−/− pro/pre-B cells showed similar BrdU incorporation as WT cells (Fig. 1,F, bottom), thereby excluding a defect in the proliferation as a probable cause for our observations. However, consistent to the role of Rap1b in regulating integrin activation, we observed a significantly decreased adhesion of Rap1b^−/− pro/pre-B cells to the mouse stromal cell line M2-10B4, reduced to 60% of WT (Fig. 1 G).

Cells from auxiliary and inguinal LNs were stained with anti-B220, anti-IgM, and anti-IgD Abs. The percentage of B220⁺, IgM⁺IgD⁺, and IgM⁻IgD⁺ cells was comparable between Rap1b^−/− and WT mice (Fig. 2 D). Additionally, we examined the T cell development in Rap1b^−/− mice. The cellularity of thymus, spleen, and LN from Rap1b^−/− mice was comparable to that of WT (thymus, WT: 120.7 ± 8.9 × 10⁶ cells; Rap1b^−/−: 119.0 ± 40.8 × 10⁶ cells, spleen, WT: 107.6 ± 24.6 × 10⁶ cells; Rap1b^−/−: 99.7 ± 16.2 × 10⁶ cells; and LN: 4.8 ± 1.1 × 10⁶ cells; Rap1b^−/−: 4.8 ± 2.4 × 10⁶ cells; n = 6–8 mice). T cell development appears normal in thymus, spleen, and LN based on CD4 and CD8 expression in the absence of Rap1b (data not shown). These results indicate Rap1b plays a role in B but not in T cell development.

Next, we analyzed the B cell maturation in the spleen. The percentage B220⁺ B cells were comparable between Rap1b^−/− and WT mice (Fig. 2,A). However, when we stained splenocytes with anti-IgM and anti-IgD Abs, Rap1b^−/− mice showed a reduction in IgM⁺IgD^low/− population (Fig. 2,B). B220⁺ splenic B cells can be further separated into newly forming B cells (CD21^low/−CD23^low/−), FO B cells (CD23^highCD21^low/−), and MZ B cells (CD21^highCD23^low/−). The percentages of FO B cells were comparable between Rap1b^−/− and WT mice. However, the MZ B cells were significantly reduced (∼55%), whereas the newly forming B cells were relatively increased in Rap1b^−/− mice (Fig. 2,C, upper panels). MZ B cells can also be defined by their exclusive high-level expression of IgM, CD21, and CD1d (40). Staining for these markers confirmed the reduction in IgM^highCD21^high and IgM^highCD1d^high B cell population (Fig. 2, D, middle and lower panels).

MZ B cells reside in a unique location around the follicles in the spleen. To examine the lymphocyte organization, we performed H&E staining of spleen cryosections. Rap1b^−/− mice displayed normal spleen architecture of red and white pulps (Fig. 3,A). Staining with anti-CD3ε and anti-B220 mAbs revealed a normal B and T cell organization within the white pulps of the spleens (Fig. 3,B). To view the MZ, we used MOMA-1 mAb, which stains the metallophillic macrophages. The location of metallophillic macrophages defines the boundary between the FO and MZ. The spleen sections from the WT mice revealed a clearly defined band of B220⁺ MZ area outside of the MOMA-1⁺ metallophillic macrophages. However, similar regions were greatly reduced and less contiguous in Rap1b^−/− mice (Fig. 3,C). Staining splenocytes with anti-IgM and anti-IgD mAbs also demonstrated a reduction in IgM^highIgD⁻ MZ B cells in the Rap1b^−/− mice (Fig. 3 D). The immunostaining results are consistent with the flow cytometry data. The significant alteration in the newly forming B cell to MZ B cell ratio indicates that the Rap1b plays a critical role in the homeostasis of MZ B cells.

Rap1b is expressed in multiple cell lineages (4). Therefore, the defects we observed in MZ B cells can be cell intrinsic or due to an abnormality in the stromal microenvironment. To distinguish between these two possibilities, we performed bone marrow transplantation experiments. As expected, the transfer of WT bone marrow cells into JAK3^−/− mice led to the reconstitution of newly forming, FO, and MZ B cells (Fig. 4,A). However, similar reconstitutions using Rap1b^−/− bone marrow cells into JAK3^−/− mice gave rise to normal FO B cells, but significantly reduced MZ B cells (p = 0.007; n = 4 mice) and slightly increased newly forming B cells, which mimicked the phenotype in Rap1b^−/− mice (Fig. 4,B). Additionally, transfer of WT bone marrow cells to sublethally irradiated Rap1b1b^−/− mice showed they were able to reconstitute the MZ population in the Rap1b^−/− host (Fig. 4 C). Taken together, our results demonstrate that the B cell defects in the spleen of Rap1b^−/− mice were cell intrinsic, and the splenic environment in Rap1b^−/− mice was able to support the development of WT B cells.

Earlier study indicated a role for Rap1 signaling in lymphocyte proliferation (41, 42). Therefore, we hypothesized that the reduction in MZ B cells in Rap1b^−/− mice could be due to defective cell proliferation. We purified B cells from spleens and stimulated them with goat anti-IgM F(ab′)₂, goat anti-IgM F(ab′)₂ plus IL-4, or LPS. [³H]Thymidine incorporation assay showed comparable proliferation of Rap1b^−/− B cells as that of WT (Fig. 5,A). To further investigate the proliferation of B cell subsets, FO B cells and MZ B cells were sorted, CFSE-labeled, and stimulated with LPS. Percentages of dividing cells between Rap1b^−/− and WT mice were compared and found to be equivalent (Fig. 5 B). Therefore, it seems that Rap1b does not regulate splenic B cell proliferation under current stimulation conditions.

Earlier studies have shown that BCR engagement and chemokine stimulation lead to increased GTP-bound active Rap1 in B cells (18, 23, 42). However, the specific isoform activated is not known. Therefore, we purified resting B cells from spleens and stimulated them with goat anti-IgM F(ab′)₂ and SDF-1. The active Rap1-GTP was analyzed by a pull-down assay. In WT B cells, Rap1-GTP was increased to more than 2-fold compared with the basal level (Fig. 5,C, upper). Overexposure led to the diminished difference between control and activated samples (Fig. 5,C, middle). As the Rap1a level was extremely low compared with Rap1b in B cells and no good commercial Rap1a specific Ab are available, we used anti-Rap1 Ab that recognizes both the isoforms. The basal level of Rap1-GTP in Rap1b^−/− resting B cells was hardly detectable even after a prolonged exposure (Fig. 5,D, lane 4), which led to the saturation of Rap1-GTP in WT sample. Upon stimulation, Rap1-GTP was also increased in Rap1b^−/− B cells. However, it was much lower than that in WT cells (Fig. 5 D, lanes 5 and 6). These results indicate that although both Rap1a and Rap1b can be converted to active GTP-bound forms through BCR or chemokine receptor-mediated activation, the major isoform is Rap1b.

Activation of Erk1/2, p38, and AKT are critical for lymphocytes proliferation and survival (43, 44). Yet the role of Rap1b in the activation of these molecules is not clear. Overexpression of RapGAPII in B cells, which inhibited the activation of both Rap1 and Rap2, did not affect Erk1/2 and p38 phosphorylation but led to increased AKT activation (45). To test the role of Rap1b in these events, we examined the BCR or SDF-1-stimulated splenic B cells. Fig. 5,E shows that the levels of Erk1/2 and p38 phosphorylation were comparable between Rap1b^−/− and WT. It is important to note that the BCR- or SDF-1-mediated activation resulted in only a minimal increase of Rap1-GTP in Rap1b^−/− B cells (Fig. 5,D). Therefore, Rap1b does not seem to regulate the phosphorylation of Erk1/2 and p38. Furthermore, we stimulated splenic B cells with goat anti-IgM F(ab′)₂ Ab and found similar levels of AKT phosphorylation at its Ser⁴⁷³ residue in Rap1b^−/− as that of WT. This finding indicates that Rap1b-GTPase does not regulate AKT phosphorylation during BCR signaling in splenic B cells (Fig. 5 F).

The present study indicates that Rap1b is the dominant isoform in B cells and regulates B cell development. To test whether Rap1b plays a role in B cell-mediated immunity, we first analyzed the basal Ig levels in the sera of Rap1b^−/− mice. Fig. 6,A shows that the preimmune Ig IgM, IgG1, IgG2a, IgG2b, IgG3, and IgA levels were comparable between WT and Rap1b^−/− mice. Because MZ B cells have been shown to play an important role in T-independent Ab generation (46), we challenged the mice with T-independent Ag TNP-Ficoll and collected sera to analyze the presence of Ag-specific Abs. Rap1b^−/− mice were able to generate similar amounts of TNP-specific IgM and IgG3, the two major Ig isotypes generated during T-independent responses, as generated by WT mice (Fig. 6,B). Optimal Ab response in Rap1b^−/− mice indicates that the remaining MZ B cells were sufficient for T-independent Ab generation. Next, we examined the T-dependent immune responses by i.p. injection of TNP-KLH. In the primary response, the generation of TNP-specific IgM, IgG1, and IgG3 was comparable between Rap1b^−/− and WT mice (Fig. 6,C). However, after Ag boost, the TNP-specific IgG1 was significantly lower in the Rap1b^−/− mice compared with that of WT (Fig. 6,C). Although not significant, the TNP-specific IgG3 were also lower in Rap1b^−/− mice during the secondary responses (Fig. 6 C). Therefore, we conclude that Rap1b is required for optimal T-dependent but not T-independent humoral responses.

Rap1 regulates cell adhesion and migration (30). To better understand the functional role of Rap1b in B cell migration, we first examined the chemokines SDF-1 and CXCL13 and lipid S1P-induced chemotaxis in Transwell assays. Rap1b^−/− splenic B cells displayed significantly reduced migration toward chemokine gradient (Fig. 7,A). S1P is important for lymphocyte migration (47, 48). Similar to SDF-1 and CXCL13, S1P has been shown to activate Rap1 and integrin activation is required for the correct localization of MZ B cells (39, 49). Consistent with published data (50), MZ B cells showed robust response to S1P and showed higher migration toward S1P. In contrast, Rap1b^−/− B cells showed dramatically reduced migration compared with WT (Fig. 7,B). MZ B cells have higher α_L and β₂ integrin expression and show higher adhesion to ICAM-1 compared with FO B cells (39). Similarly, S1P-induced adhesion to ICAM-1 was also reduced in Rap1b^−/− B cells (Fig. 7,C). To further investigate these defects, we performed an in vivo homing assay. Rap1b^−/− and WT-derived splenocytes were separately labeled with TAMRA and CFSE-DA, respectively, premixed and i.v. injected into the recipient mice. To provide a Rap1b-sufficient microenvironment, we used the WT mice as the recipients. Three hours later, cells from spleen, auxiliary/inguinal and mesenteric LN were stained with anti-B220 and anti-CD3ε mAbs, and the ratio of Rap1b^−/− to WT-derived lymphocytes were quantified (Fig. 7,D). The homing of B and T cells from Rap1b^−/− and WT donors into the recipient spleens were comparable. However, the ability of Rap1b^−/− B cells to migrate into the auxiliary or inguinal LN was significantly reduced (∼60%). Similarly, homing of Rap1b^−/− B cells into mesenteric LN was also affected (Fig. 7,E, left). We also observed reduced ability of Rap1b^−/− T cells to migrate into the auxiliary or inguinal LN (Fig. 7 E, right). Together, these results demonstrate that Rap1b plays a critical role in chemokine-mediated migration and in the in vivo homing of B cells.

In this study, we demonstrate that Rap1b is the dominant isoform in B cells. Using gene knockout mice, we show that Rap1b is essential for pro/pre-B cell development in the bone marrow and for the homeostasis of MZ B cells in the spleen. Furthermore, Rap1b is required for optimal T-dependent humoral immunity, chemokine- and S1P-stimulated migration, S1P-induced adhesion to ICAM-1, and in vivo homing of B cells. Rap1b is also important for B cell progenitor adhesion to stromal cell lines. Rap1b is not required for splenic B cell proliferation and BCR-mediated activation of Erk1/2, p38, or AKT.

One of the key findings in this study is the role of Rap1b in B cell development in the bone marrow. A reduction of pro/pre-B cells and immature B cells indicates that the requirement of Rap1b cannot be compensated by Rap1a or Rap2 isoforms. Several gene knockout mice with altered Rap1 signaling have been recently generated (25, 26, 36). Deletion of SPA-1, a principle Rap1GAP in lineage negative (Lin⁻) bone marrow cells, resulted in an accumulation of Rap1-GTP and lead to an increase of the hematopoietic progenitors (27). Expression of constitutively active Rap1, RapE63, resulted in the Lin⁻ bone marrow cell expansion, whereas SPA-1 led to the suppression (27). However, in the absence of Rap1b, bone marrow-derived B220⁺ B cell lineage showed similar proliferation rates as that of WT. Interestingly, lack of Rap1b resulted in reduced adhesion to bone marrow stromal cell lines. The reduced bone marrow pro/pre-B cells in Rap1b^−/− may be due to reduced adhesion rather than a decreased expansion of B cell lineage. In addition, although lack of Rap1b did not adversely affect the number of B220⁺CD43⁺ B progenitors, there was a delayed maturation as indicated by more c-kit⁺ and fewer CD25⁺ cells in Rap1b^−/− bone marrow. In peripheral blood, a higher percentage of B cells displayed IgM⁺IgD⁻ immature phenotype compared with that of WT. Reduced adhesion of pro/pre-B cell to stromal cells could also cause a corresponding decrease of B progenitors in the bone marrow of Rap1b^−/− mice. Taken together, these results indicate a possible differential regulation of Lin⁻ and Lin⁺ bone marrow cells by Rap1b. The reduced bone marrow pro/pre-B cells in Rap1b^−/− may be due to reduced adhesion rather than an augmented expansion of B cell lineage.

Spleen has a unique architecture composed of stromal and hematopoietic cells that are vital for B cell maturation and function. A strong reduction of MZ B cells in Rap1b^−/− mice may be due to the deficiency in B or splenic stromal cells. Rap1b^−/− mice showed normal organization of spleen and the defect of MZ B cells was recapitulated in our bone marrow transplantation experiment. Together, these indicate the MZ B defect is cell intrinsic. A MZ B cell defect is also observed in RapL^−/− mice validating the role of Rap1-RapL signaling in MZ B cell development (25). However, the MZ B cells in RapL^−/− mice are nearly absent, which is a more severe phenotype compared with Rap1b^−/− mice. The multiple upstream RapL activators, Rap1 and Rap2 isoforms, could account for the phenotypic differences in these two mouse models (51). Expression of Rap2 in B cells has been previously established (18). Immature B cells emigrate from the bone marrow and home to the secondary lymphoidal organs such as spleen (40). Once migrated into the spleen, newly formed B cells have the potential to differentiate into either FO or MZ B cells (40). Reduced adhesion of pro/pre-B cell to stromal cells may lead to an augmented egress of immature B cells from bone marrow, which may account for the normal FO B cells in Rap1b^−/− mice. Rap1 signaling seems not essential for egressed immature B cells to differentiate into FO B cells in Rap1b^−/− spleen. This is partly supported by bone marrow progenitors transfected with constitutively active RapE63 that showed similar differential potential as vector-transfected control cells (26). An increase in the newly forming B cells in Rap1b^−/− mice could indicate the specific role of Rap1 in the maturation for MZ but not FO B cells. The role of integrin in the retention of MZ B cells has been demonstrated (39). MZ B cells have higher expression level of α_Lβ₂ and α₄β₁ integrins than FO B cells. As S1P induced activation of Rap1 (49), the dramatically reduced migration to S1P and S1P-induced adhesion to ICAM-1 of Rap1b^−/− splenic MZ B cells may contribute to the defect in MZ B cells. In contrast, in the preleukemic SPA^−/− mice, the development of FO and MZ B cells is normal (27). This discrepancy between Rap1b^−/− and SPA^−/− mice is due to the differential expression of these two proteins in bone marrow and spleen B cells. SPA-1 expression level is considerably higher in Lin⁻ than Lin⁺ cells, which suggests Rap1 signaling may regulate lineage development differentially at different stages. The reduced MZ B cells in Rap1^−/− mice is not due to defective cell proliferation as we did not observe a role of Rap1b in LPS- or BCR-induced proliferation of total spleen B cells or sorted FO and MZ B cell, which is supported by the normal proliferation of SPA-1^−/− splenic B cells (27). Thus, the role of Rap1 signaling in cell proliferation seems to be cell type-dependent as it has been shown to positively regulate the proliferation of hematopoietic (26) and peritoneal B1 cells (27). The relatively normal development of T cells in the absence of Rap1b could be due to a compensatory effect from a different member of the Rap family such as Rap2.

Cell proliferation, differentiation and survival can be regulated by Erk1/2, p38, and AKT-mediated signaling pathways (52). Ras is one of the major activator of Erk1/2 and p38 (53). Rap1-GTPase has been shown to negatively regulate Ras-mediated Erk and p38 activation in some studies; however, it is controversial (1, 45, 54, 55). Our present study shows that phosphorylation of Erk1/2 and p38 was not changed after BCR and SDF-1 stimulation in Rap1b^−/− B cells, although Rap1-GTP was nearly absent. This finding is consistent with earlier studies in which expression of constitutively active V₁₂Rap1a in thymocytes, activation of Rap1 by cAMP in NIH3T3 cells, or overexpression of RapGAPII in B cells exhibited an unaltered Erk1/2 phosphorylation (19, 45, 56). Regulation of Erk1/2 activation by Rap1 has been shown to be cell context-dependent. In SPA-1^−/− mice, Rap1 signaling has been shown to regulate Erk phosphorylation positively in Lin⁻ bone marrow cells and negatively in anergic T cells (26, 57). Earlier studies have also that Rap1 signaling AKT phosphorylation (41, 45). In contrast, the similar BCR-mediated activation of AKT (Ser⁴⁷³) phosphorylation was observed in Rap1b^−/− B cells as that in WT. This discrepancy may be due to the dual GAP activity of RapGAPII directed to both Rap1 and Rap2. It is most likely that the Rap2 but not Rap1 that regulates the AKT phosphorylation as Rap2 has its unique effectors (58, 59, 60).

Cell migration is controlled by multiple coordinated processes including chemoattraction, adhesion, and transmigration through the blood vessels. Chemokine-directed migration of B cells is essential for B cell development, differentiation, and Ab generation (61). The defect of Rap1b^−/− B cells in vitro chemotaxis assay and in vivo short-term homing to LN is consistent with the established role of Rap1b in cell migration. Interestingly, the LN cellularity of Rap1b^−/− mice was normal. The defect in the short-term homing assay may reflect only a slower cell migration. During the long developmental period, Rap1b^−/− B and T cells may still be able to migrate to LN and display normal cellularity in LN. In RapL^−/− mice, the defective migration of lymphocytes leads to atrophic thymus, spleen and LNs (25), which could be due to that RapL relays multiple upstream Rap GTPase signaling (51). The migration of T cells in short-term homing assay was far less severe in Rap1b^−/− mice compare with B cells, which may reflect a differential expression pattern of Rap family proteins in B and T cells.

The defect in T-dependent but not T-independent humoral response was impaired in the Rap1b^−/− mice. The defects in short-term homing of B and T cells in the absence of Rap1b could contribute to this defect. During the T2-independent immune response, MZ and B1 are two predominant B cell types that are recruited for Ab generation. Although, the MZ B cells are reduced to 40% of WT in our experiments, the residual MZ B cells might be enough for T2-independent Ag-induced Ab production. However, the defects in short-term homing of B and T cells in the absence of Rap1b could have contributed to the suboptimal T-dependent Ab production. In addition to cell migration, several other important steps are required for the successful T-dependent immune response. The presentation of antigenic epitope/MHC class II on APCs (dendritic cells or B cells) to Th cells is crucial, which is dependent on successful stable formation of immunological synapses that contain LFA-1, ICAM-1, CD40, and other molecules (62). Rap1 regulates T cell polarization and thereby the interaction with APCs (22). The dominant negative form of Rap1, Rap1N17, has been shown to block TCR-dependent increase in adhesion to ICAM-1. Furthermore, Rap1a^−/− T cells failed to properly polarize LFA-1 upon CD3 stimulation (36). Thus, lack of Rap1b could impair the efficiency of T cell-APC interactions leading to a reduction in the secondary T-dependent Ab generation. Our present observations were supported by the findings in the SPA-1^−/− mice (27). T cells from SPA-1^−/− mice had selective impairment in memory T cell responses against a T-dependent Ag with normal primary Ab response. Lack of SPA-1 also resulted in an augmented levels anti-dsDNA Abs from B1a B cells leading to lupus-like autoimmune disease (27). Therefore, a regulated Rap1-signaling is critical for optimal immune responses. Of note, during publication of our results, Chen et al. published similar results (63).

However, the precise role of Rap1b in B cell development and memory T-dependent humoral response are still not complete elucidated. Induced deletion of Rap1b in B lineage could address this question more specifically. A conditional ablation of Rap1b will help to dissect the distinct role of Rap1b in B cells, T cells, and dendritic cells, the major cell types that involve in T-dependent immunity. Rap1a and Rap1b have different C-terminal sequences leading to distinct posttranslational modifications, which may contribute to isoform-specific function (31, 32, 34). The incomplete loss of MZ B cells and intact T-independent immunity in Rap1b^−/− mice may indicate the compensatory role of Rap1a or Rap2 proteins. Continued research using Rap1a or Rap2-deficient mouse models will greatly help us to understand the function of Rap protein family.

We thank Jillian Dargatz for helping with mouse care, Barbara A. Fleming for cryosectioning, and Asanga Samarakoon for technical support.

The authors have no financial conflict of interest.