A Rasa3-Gαi signaling axis orchestrates B lymphocyte trafficking into and through lymphoid organs

SUMMARY

By modulating integrin affinity, the GTP/GDP status of Rap1 affects the arrest, transendothelial migration, and the trafficking of lymphocytes, yet how chemoattractant receptors control Rap1 remains incompletely resolved. Rasa3, a Rap1 GTPase-activating protein, limits the duration that Rap1 remains GTP bound. Here, we investigated how Rasa3 deficiency impacted chemoattractant receptor signaling and B lymphocyte trafficking in mice. The loss of Rasa3 disrupted the usual dynamic regulation of the GDP/GTP status of Rap1 causing a striking phenotype characterized by a severe maldistribution of B cells within lymphoid organs and major reductions in mucosal and blood B cells. Rasa3 loss raised basal Rap1-GTP levels, disrupted integrin binding, and unexpectedly caused defects in chemoattractant receptor signaling. At the plasma membrane GTP-bound but not GDP-bound Gαi resided within 10 angstroms of Rasa3. Thus, Rasa3 couples Gαi signaling to Rap1-GTP levels tuning chemokine signaling and integrin affinity to promote B lymphocyte trafficking and function.

In brief

B lymphocytes must traverse multiple body sites to generate plasma cells and make protective antibodies. Park et al. identify a reciprocal relationship between Rasa3 and G_αi that controls Rap1 activity, thereby promoting B lymphocyte entrance into and migration through lymphoid organs. These three proteins collaborate to allow normal humoral immunity.

Graphical Abstract

INTRODUCTION

B lymphocyte adhesion, migration, and tissue compartmentalization facilitates an efficient humoral immune response by ensuring B cell exposure to antigens and vaccines.¹ Chemoattractant receptors that activate the heterotrimeric G protein Gi play a central role in both homeostatic B cell trafficking and in directing B cells to inflammatory sites.² Ligand-occupied chemoattractant receptors trigger receptor/heterotrimeric Gi protein coupling, G_αi subunit guanosine diphosphate (GDP)-guanosine 5′-triphosphate (GTP) exchange, functional G_αi dissociation from G_βγ subunits, and subsequent G_αi-GTP and G_βγ effector activation.^3–6 In B cells, G_βγ effectors include phospholipase C-β (PLC-β), adenylyl cyclases, and phosphatidylinositol 3-kinase-γ (PI3K-γ). Known direct G_αi-GTP effectors include certain adenylyl cyclase isoforms. G_αi-GTP and G_βγ primary effectors rapidly engage intracellular molecules including protein kinases and guanine nucleotide exchange factors (GEFs) that trigger small GTPases such as Ras, Rap, and Rac to undergo nucleotide exchange.^7–9 Much like heterotrimeric Gα subunits, these small GTPases are molecular switches that cycle between active and inactive states depending upon whether they are GTP bound or not. However, unlike heterotrimer Gα subunits, the small GTPases lack an intrinsic GTPase activity making them completely reliant on GTPase-activating proteins (GAPs) for their deactivation and re-cycling.¹⁰

In B lymphocytes the small GTPase Rap1 plays a pivotal role in regulating adhesion, migration, and immune synapse formation.¹¹ Murine B cells express two highly related Rap1 isoforms, Rap1a and Rap1b. Mice lacking both isoforms have impaired integrin activation and a severe B cell trafficking defect. B cells enter lymph nodes (LNs) poorly, the peritoneum lacks B-1a cells, blood B cells numbers increase, splenic marginal zone B cells are missing, and chemokines cannot elicit directional cell migration. Complicating the study of Rap1 signaling in B cells and other cell types are the numerous GEFs and GAPs that affect its GTP/GDP status (KEGG PATHWAY database: mmu04015).¹² Among the RAP1 GAPs, Rasa3 and Sipa1 are the most highly expressed in murine B cells.¹³ Mice lacking Rasa3 are not viable largely because of defective platelet function,¹⁴ while mice lacking Sip1a develop a spectrum of myeloproliferative stem cell disorders and a lupus-like autoimmune disease with an accumulation of self-reactive B-1a cells in the peritoneum.¹⁵ T lymphocytes express Rasa3 and T cell-specific deletion increases Rap1-GTP levels, causes T cell lymphopenia, impairs LN entry and egress, and results in poor T cell-dependent humoral immunity.¹⁶ Rasa3-deficient T cells show an increase in both integrin-mediated adhesion and T cell migration defects. Loss of both Rasa3 and Sip1a in T cells exacerbates these phenotypes.¹⁷

Two findings directly link Rasa3 to Gi signaling. First, Rasa3 preferentially co-immunoprecipitates with GTPase-deficient G_αi2 (largely GTP bound) versus wild-type Gα_i2.¹⁸ While the importance of this interaction is unknown, the related protein Rasa2 also binds GTPase-deficient Gα_i2, which enhances Ras signaling by rerouting Rasa2 away from Ras-GTP.¹⁹ Second, G_αi signaling affects the intracellular localization of Rasa3, which enhances Rap1 signaling and integrin activation.²⁰ To investigate the role of Rasa3 in B lymphocyte biology and to explore whether the loss of Rasa3 reciprocally affects G_αi signaling, we generated mice with a B cell-specific deletion of Rasa3. A severe B cell trafficking defect occurred with marked B cell lymphopenia, loss of B cells from LNs, a dramatic reduction in Peyer’s patch (PP) B cells, and an expansion of splenic B cells. Unexpectedly proximal chemokine receptor signaling was also impaired in its absence. In HeLa cells, RASA3 localized at the plasma membrane closely associated with G_αi-GTP but not G_αi-GDP. Finally, in a human B cell line RASA3 localized at the plasma membrane enriched at sites of active RAP1 nucleotide exchange. Collectively, these findings suggest that Rasa3 directly intersects with Gi and Rap1 signaling pathways to control B lymphocyte trafficking and compartmentalization.

RESULTS

B cell lymphopenia with normal bone marrow B cell development but an expanded splenic B cell population

To assess the role of Rasa3 in B cell biology, we crossed mb1^cre/+ mice to Rasa3^fl/fl mice to generate mice that lost Rasa3 expression during B cell development, hereafter referred to as Rasa3 BKO mice. Based on past studies using mb1^cre/+ mice,² the lack of Rasa3 expression should become evident in fraction B cells in the Hardy scheme of bone marrow B cell development, in which fraction A–C cells correspond to pro-B cell stages, fraction D cells to pre-B cells, fraction E cells to immature B cells, and fraction F cells to mature recirculating cells.²¹ We found that mice lacking B cell Rasa3 expression were born at the expected Mendelian ratios. As T cell lymphopenia occurred in Rasa3 T cell-specific knockout mice,¹⁶ we checked the cellular composition of the blood in the Rasa3 BKO mice. We found very low blood cell counts with an overall 90% decrease in B cells (Figure 1A). However, defective bone marrow B cell development did not account for the low blood B cell numbers. While we noted fewer recirculating bone marrow B cells (perhaps a consequence the B cell lymphopenia) and a minor increase in fraction D cells, the expected numbers of newly formed bone marrow B cells were present (Figure 1B). Thus, poor bone marrow development of B cells or bone marrow entrapment did not account for the B cell lymphopenia.

Figure 1. Immune phenotyping of Rasa3 BKO mice reveals a severe maldistribution of B cells.

(A) Blood cell number and cell distribution, unpaired t test.

(B) Bone marrow B cells and B cell development, unpaired t test.

(C) Spleen cell numbers, unpaired t test.

(D) Splenic B cell development, unpaired t test.

(E) Immunohistochemistry of representative spleen sections and assessment of B/T cell zones. Scale bars represent 500 μm, unpaired t test.

(F) pLN and mLN B and T cell numbers, unpaired t test.

(G) Immunohistochemistry of representative iLNs. Scale bars represent 100 μm.

(H) PPs cell numbers and germinal center B cells in mLN and PPs, unpaired t test.

(I) Immunohistochemistry of representative PP sections. Scale bars represent 100 μm.

(J) Peritoneal cavity lymphocytes, unpaired t test.

(K) Thymocytes and thymus cell populations, unpaired t test.

(L) Immunohistochemistry of representative thymus sections. Scale bars represent 200 μm.

(M) B220⁺ and CD138⁺ cells located in the outlined boxes overlying the thymus cortex, unpaired t test. *p < 0.05, **p < 0.005, ***p < 0.0005.

Immature murine B cells leave the bone marrow and complete their differentiation to follicular and marginal zone B cells in the spleen.²² Upon initial inspection, the Rasa3 BKO mice had visibly enlarged spleens, which contained 1.7-fold more CD45⁺ cells and 2-fold more B cells than did the spleens from control mice (Figure 1C). The increase in splenic B cells was accompanied by an absolute increase in splenic CD4 and CD8 T cells although they declined on a percentage basis. Further analysis of splenic B cell differentiation revealed a slight increase in the numbers of T2 follicular B cells and a slight reduction in follicular B cells. While Rasa3 BKO mice spleens had an increase in absolute numbers of marginal zone B cells on a percentage basis, they did not differ from the controls (Figure 1D). Similarly, Rasa3 BKO mice spleens had an absolute increase in germinal center B cells but no difference on a percentage basis. To assess where the B cells reside within the spleen, we used confocal microscopy to image control and Rasa3 BKO spleen sections immunostained for B220, CD21, and Ki67. Imaging confirmed the expanded spleen size, with no obvious disruption of the white pulp (Figure 1E). Consistent with the expanded B cell compartment in the spleen, the ratio between the sizes of the white pulp B and T cell zones was increased (Figure 1E).

To exit the spleen, B cells must leave the white pulp, a transition that requires S1pr1 signaling.²³ Whether B cells directly cross the marginal sinus or access the bridging channels to access the red pulp is unclear. However, once in the red pulp, B cells encounter splenic sinusoids, where sphingosine 1-phosphate (S1P) levels are high. Again, S1pr1 signaling helps promote their movement across the ICAM-1 rich endothelium into the venous blood to re-enter the systemic circulation. In the Rasa3 BKO mice spleens, numerous B cells accumulated near the venous sinusoids (Figure S1). These findings suggest that splenic B cell entrapment due to difficulties in exiting from the white and red pulp contributed to the low blood B cell numbers in the Rasa3 BKO mice.

B cell deficit in LNs and PPs from the Rasa3 BKO mice

In contrast to the spleen, both the inguinal and mesenteric LNs contained a lower percentage of B cells than did those from control mice (Figure 1F). Combined with the reduction in LN lymphocyte recovery, the Rasa3 BKO mice peripheral and mesenteric LNs had approximately half as many B cells as did control LNs (1.5 × 10⁷/pLN versus 0.7 × 10⁷/pLN; 2.5 × 10⁷/mLN versus 1.1 × 10⁷/mLN). Images of the Rasa3^fl/fl and Rasa3 BKO inguinal lymph node are shown (Figure 1G). Examination of the small intestine for PPs revealed a reduction in visible PPs in the Rasa3 BKO mice. Found along the antimesenteric side of the small intestine, primarily in the distal portion ileum, 6–12 PPs are typically visible in C57BL/6 mice. In this study the control mice had an average of 8 PPs while the Rasa3 BKO mice had only 3. The residual PPs in the Rasa3 BKO mice had fewer cells with a decreased percentage of B cells (6.9 × 10⁴/PP versus 1.8 × 10⁴/PP) (Figure 1H). Accounting for the reduction in PP numbers, this resulted in a 90% reduction of PP B cells in the Rasa3 BKO mice. Despite the reduction of B cells in mesenteric LNs and PPs, the percentage of B cells that possessed a germinal center phenotype did not differ from the control mice (Figure 1H). To assess where the B cells reside within LNs and PPs, we again used confocal microscopy to image sections prepared from control and Rasa3 BKO LNs and PPs and immunostained for CD21, CD4, Lyve-1 (lymphatics), PNAd (high endothelial venules, HEVs), and Ki67 (proliferating cells, i.e., germinal center B cells). While this loss of Rasa3 in B cells reduced the B cell numbers in the LN and PPs, this loss did not disrupt overall B and T cell compartmentalization or the normal organ architectures (Figure 1I).

Expanded peritoneal and thymic B cell populations in the Rasa3^fl/flmb1^cre/+ mice

Having documented B cell lymphopenia and a reduction in LN and PP B cells in the Rasa3 BKO mice, we evaluated other sites where B cells reside. The mouse peritoneal and pleural cavities contain three different types of B cells: B-1a, B-1b, and conventional B cells or B-2 cells.²⁴ B-1 cells predominately reside in body cavities being retained by CXCL13 signaling and integrin engagement.²⁵ In contrast, B-2 cells tend to circulate through the body cavities, exiting via the lymphatics into adjacent LNs. To measure peritoneal B cell numbers, we injected cold phosphate-buffered saline into the peritoneal cavity and 5 min later recovered the fluid to determine its cellular composition. The peritoneal fluid recovered from the Rasa3 BKO mice had more cells with more lymphocytes than recovered from the controls (Figure 1J). The lymphocyte composition also differed as the peritoneal fluid from the Rasa3 BKO mice had significantly more B-2 cells. The absolute numbers of B-1a cells were slightly reduced while B-1b cell numbers were slightly increased (Figure 1J). Thus, despite the B cell lymphopenia, the peritoneal cavity contained more B-2 cells suggesting a misbalance between B-2 cell ingress and egress in the Rasa3 BKO mice. In contrast, the loss of Rasa3 in B-1a and B-1b cells only marginally affected their counts in the peritoneum.

B cells also reside in the thymus.²⁶ Bone marrow-derived progenitor cells traffic to the thymus where they encounter a unique microenvironment that influences their differentiation, phenotype, and function. Committed thymic B cells progenitors are B220^low/CD43⁺ surface Ig⁻ cells, which develop into B220^high/CD43⁻ surface Ig⁺ B cells. Phenotypically they resemble activated memory B cells.²⁷ The thymuses from the Rasa3 BKO mice contained slightly more cells than the control mice (Figure 1K), but a normal distribution of thymocyte subsets. On a percentage basis of the lymphocyte gate, we found more B220⁺CD19⁺ cells than in the control mice. We also found that an expansion of B220^high cells accounted for the excess percentage of B220⁺/CD19⁺ cells. When normalized to the number of CD4 single-positive cells, we again noted an expansion of B cells as well as plasma cells (Figure 1K). To determine whether the loss of Rasa3 affected the localization of thymic B cells we used confocal microscopy to image thymic sections from control and Rasa3 BKO mice immunostained for CD19, UEA-1, and CD4 (Figure 1L). Typically, B220^low cells localize at the cortical/medullary junction, while B220^high and plasma cells accumulate in the medullary region.²⁸ The thymic sections prepared from the Rasa3 BKO mice contained more B220⁺ cells than those prepared from the Rasa3^fl/fl mice. Furthermore, the Rasa3 BKO B220⁺ B cells were in the cortex distal to the cortico-medullary junction (Figure 1M). Some of the CD138⁺ cells were displaced residing superficially in the thymic cortex. Thus, the Rasa3-deficient B cells showed increased and abnormal localization within the thymus.

Rasa3 facilitates B cells trafficking through the spleen, peripheral LNs, and lung

To assess the trafficking of the Rasa3 BKO B cells, we purified and differentially labeled B cells from Rasa3^fl/fl and Rasa3 BKO mice and adoptively transferred them into C57BL/6 mice. In some instances, we treated the recipient mouse 2 h after cell transfer with an antibody to CD62L, thereby blocking further LN entrance of blood-borne B cells. At 2 h following transfer we found little difference in the recovered control and BKO B cells at the various sites except for a 20% decrease in BKO B cells recovered from mesenteric LNs. However, by 21 h post transfer, comparatively fewer Rasa3 BKO B cells resided in the blood, bone marrow, and inguinal LN, and relatively more in the spleen (Figure 2A). Adding a CD62L antibody 2 h post cell transfer raised the relative recovery of Rasa3 BKO B cells from peripheral LNs returning the ratio to nearly 1:1 (Figure 2A). In contrast, the CD62L antibody did not affect the relative recovery of WT and KO B cells from mesenteric LNs. These data suggested that poor entrance rather than enhanced egress accounted for the reduced recovery of KO B cells from peripheral LNs. However, while the flow cytometry assay provides a measure of the lymphocyte flux through LN and spleen, it does not assess the location of the transferred cells. Therefore, we used confocal microscopy to examine the relative numbers and location of the adoptively transferred cells. At 24 h post adoptive cell transfer, we found fewer KO B cells in the splenic follicles compared to control cells but similar numbers at the B:T border (Figures 2B and S2). This argued that the Rasa3 BKO B cells had difficulty in entering the splenic B cell follicles. Nonetheless, the enlarged splenic white pulp in the Rasa3 BKO mice indicates that the Rasa3-deficient B cells eventually enter and accumulate over time.

Figure 2. Adoptive transfer of Rasa3 BKO B cells confirms their impaired splenic and LN trafficking.

(A) Schematic of the experimental procedure used in in vivo homing and egress assays. Graph shows the ratio of adoptively transferred B cells in the specified tissues. Ordinary one-way ANOVA multiple comparison.

(B) Image of the splenic white pulp 24 h after adoptive cell transfer (left). Boxes indicate the analyzed regions, CD21 signal removed (right). Scale bars represent 50 μm. Graph shows percent of adoptively transferred cells in the follicle or at the border. Ordinary one-way ANOVA multiple comparison.

(C) The graph shows the number of adoptively transferred cells at each indicated time point localized on the HEVs, unpaired t test.

(D) Schematic of the stepwise process of B cell transendothelial migration (TEM) through a high endothelial venule (HEV), depicting key stages.

(E) Graphs show the time duration for indicated TEM process, unpaired t test.

(F) Images (left) iLN HEVs 1 h post adoptive cell transfer. Boxes a and b highlight the HEV-associated cortical ridge; scale bars represent 200 μm and 50 μm. Images (right) of iLN 24 h after adoptive cell transfer. Boxes (below) indicate analyzed regions; yellow, subcapsular; green, follicular; white, cortical ridge. Scale bars represent 100 μm.

(G) Graphs showing localization of B cells 1 h and 24 h after transfer, unpaired t test.

(H) Graphs showing B cell motility parameters. Rasa3^fl/fl and Rasa3 BKO B cells transferred 24 h prior to iLN imaging (2:1 ratio), unpaired t test. *p < 0.05, **p < 0.005, ***p < 0.0005.

To examine the entrance of B cells into LNs, we used intravital microscopy to visualize differentially labeled B cells from Rasa3^fl/fl and Rasa3 BKO mice as they appeared on the LN HEVs. During the first 40 min of imaging, we found that approximately twice as many Rasa3 BKO B cells accumulated on the HEVs as control B cells (Figure 2C; Video S1). Once firmly adhered, B cells search for a site to penetrate the endothelial barrier. Subsequently they cross the endothelial basement membrane (1^st TEM) to enter the perivascular space where they remain until exit (2^nd TEM) into the LN cortical ridge (Figure 2D). Comparing the wild-type and Rasa3 BKO B cells revealed that that the wild-type B cells required less time to accomplish each of these steps. On average the control B cells took approximately 16 min to enter the LN cortical ridge while the Rasa3 BKO B cells required 40 min (Figure 2E). Next, we examined fixed LN sections 1 h and 24 h following adoptive cell transfer (Figures 2F and 2G). At 1 h post transfer, we focused on the HEV areas. Although WT and Rasa3 BKO B cells showed similar numbers of cells, more Rasa3 BKO B cells remained within the confines of the HEVs. At 24 h post adoptive cell transfer, as expected we found fewer Rasa3 BKO B cells in the LNs and particularly fewer within the B cells follicles. To have sufficient Rasa3 BKO B cells to image within B cell follicles we adoptively transferred twice as many Rasa3 BKO B cells as WT cells. Tracking the follicular cells revealed that the Rasa3 BKO B cells moved faster, straighter, and with less speed variability than did the control B cells (Figure 2H; Video S2). These measurements are consistent with a model in which the loss of Rasa3 causes altered integrin/Rap1 control. Faster, straighter, and more persistent cell tracks result from a reduction in adhesive pausing and turning.

Next, we imaged the Rasa3 BKO and control B cells near the cortical lymphatics, a major site for B lymphocyte egress from LNs.²⁹ At 2 h after cell transfer, we administrated a CD62L antibody to block further entrance. At 8 h both control and Rasa3 BKO B cells were located near the cortical lymphatic (Figure S3A). Although relatively few cells had entered the lymphatic sinuses, most that had were KO B cells. By 24 h post transfer, more cells had localized within the lymphatic sinuses, but the Rasa3 BKO B cells still predominated. While this result suggested that the Rasa3 BKO B cells egressed more rapidly than did the control B cells, we found that many of the Rasa3 BKO B cells in the sinus crossed back into the LN parenchyma. In contrast, once control B cells crossed into the sinus, they rarely returned (Figures S3B and S3C). Finally, we checked the motility parameters of the B cells located near the lymphatics (Video S3). Despite the differences between the control and Rasa3 BKO B cells in the follicle, near the lymphatics their motility parameters did not differ (Figure S3D). These data indicate that (1) Rasa3 BKO B cells enter LNs poorly and those that do so tend to remain near the HEVs in the cortical ridge area, (2) Rasa3 BKO B cells that gain access to the LN follicle moved faster along straighter paths, and (3) Rasa3 BKO B cells that enter the lymphatic sinus can inappropriately return to the LN parenchyma.

To assess B cell trafficking through the lung we adoptively transferred fluorescently labeled Rasa3^fl/fl and Rasa3 BKO B cells and examined lung sections 1 and 4 h after transfer. We found at both time points a nearly 2-fold excess of Rasa3 BKO B cells compared to controls (Figure S4). Thus, the trapping of the Rasa3 BKO B cells within the lungs also likely contributed to the low number of B cells in the blood.

Rasa3 enhances B cell entrance into and retention within PPs

Consistent with the reduction in PP B cells in the Rasa3 BKO mice, we also recovered fewer adoptively transferred Rasa3 BKO B cells compared to control B cells in the PPs of recipient mice with a small reduction at 2 h post transfer and a more severe reduction at 21 h. However, administering CD62L antibody at 2 h did not change the recovery ratio (Figure 3A). We then used intravital microscopy to further evaluate the adoptively transferred B cells. First, we generated composite images of an entire PP. To outline the B cell-rich regions, we transferred labeled control B cells; to help identify the HEVs, we injected a PNAd antibody; and to outline the blood vessels and lymphatics, we intravenously administered Evans blue dye. Shown are three composite images acquired at different depths from the capsule (Figure 3B; Video S4). Visualized by their exclusion of Evans dye within the extensive lymphatic network, unlabeled endogenous cells jammed several sites, which we termed lymphatic portals as the cells directly exited from them into nearby, rapidly flowing lymph channels. A few transferred B cells resided in the germinal centers while most localized in the follicular mantle region adjacent to the germinal centers. Next, we evaluated intravenously adoptively transferred control and Rasa3 BKO B cells in the HEVs of recipient wild-type mice. We imaged the transferred cells for 60 min beginning 5 min after tail vein injection (Figures 3C and 3D, Video S5). Twice as many control B cells had arrived at the 5 min time point as did Rasa3 BKO B cells. The number of Rasa3^fl/fl B cells accumulated over the first 30 min and then declined while the Rasa3 BKO B cells failed to accumulate. Tracking the B cell on the HEVs revealed that both the control and the Rasa3 BKO B cells resided for many minutes on the HEVs before attempting to cross. In the imaging experiment shown, 7 Rasa3 BKO B cells transmigrated while 31 Rasa3^fl/fl B cells crossed. Because of the low transmigration frequency, we did not evaluate the different stages of transmigration.

Figure 3. Poor PP entrance and enhanced egress of Rasa3 BKO B cells.

(A) Graph shows the ratio of adoptively transferred Rasa3^fl/fl and Rasa3 BKO B cells in PPs. Ordinary one-way ANOVA multiple comparison. Like Figure 2A.

(B) Micrographs of a PP from a C57BL/6 mouse. Wild-type B cells adoptively transferred 24 h (pink) and 48 h (cyan) before imaging; Evans blue and CD31 antibody injected 30 min prior to imaging. Scale bars represent 200 μm. GC, germinal center; HEV, high endothelial venule; LP, lymphatic portal.

(C) Micrograph of a PP HEV 5 min post B cell adoptive transfer. CD31 antibody injected 25 min prior B cell adoptive transfer. Scale bar represents 50 μm.

(D) 3D reconstruction of PP HEVs overlaid with tracks of Rasa3 BKO (pink) and control B cells (green) Cell motility tracks were generated from 5 to 66 min post-transfer. The graph shows the number of cells at each indicated time point. Unpaired t test.

(E) Confocal image of PP section from wild-type mouse 8 h post Rasa3^fl/fl and Rasa3 BKO B cell transfer immunostained with Lyve-1 antibody (green).

(F) Graph showing percent Rasa3^fl/fl and Rasa3 BKO B cells 8 h post adoptive transfer. Data from boxes shown in (E). Unpaired t test.

(G) Micrograph of PP section focused on efferent lymphatic (green) 8 h after adoptive B cell transfer Lymphatic endothelial cells (Lyve-1, green). 3D reconstruction (right) of lymphatic and transferred B cells visualized. Scale bars represent 50 μm, 20 μm, and 25 μm.

(H) Schematic showing FTY720 experimental protocol. Transferred cells were analyzed 8 h after cell transfer. Graph shows the ratio of adoptively transferred B cells in PPs. Unpaired t test. Micrograph from intravital images of a PP verifying efficacy of FTY720 (right). Adoptively transferred C57BL/6 B cells at indicated time before imaging. Arrows show direction of lymph flow. FTY720 administered or not 4 h before imaging. Scale bars represent 50 μm. *p < 0.05, **p < 0.005, ***p < 0.0005.

Eight hours post adoptive transfer, the recipient PPs had more than twice as many control B cells (Figures 3E and 3F). Many of the control and Rasa3 BKO B cells were localized near the lymphatic portals although within the portals we predominately found the BKO B cells suggesting that they had prematurely egressed (Figure 3G). To provide evidence that the Rasa3 BKO B cells egressed more rapidly, we compared the ratio of adoptively transferred Rasa3^fl/fl B cells to Rasa3 BKO B cells in the PPs of recipient mice pretreated, or not, with FTY720 to limit egress. The FTY720 pretreatment partially normalized the ratio between the control and KO B cells recovered from PPs arguing that in its absence the Rasa3 BKO B cells egressed more rapidly (Figure 3H). We verified that the FTY720 blocked B cell egress from PPs by imaging the lymphatics of the non-treated and FTY720-treated mice. The images showed that the FTY720 treatment emptied the PP lymphatics of both transferred B cells and endogenous cell shadows (Figure 3H; Video S6). These results indicate that the loss of Rasa3 in B cells affects PPs by decreasing the attachment of B cells to HEVs, reducing their entry, and facilitating their egress. Consequently, PPs are small with few B cells or even visibly absent. Together the adoptive transfer experiments confirmed a major disruption in B cell homing and homeostasis in the Rasa3 BKO mice.

Rasa3 limits integrin activation in B cells

The phenotyping of the Rasa3 BKO mice, the abnormal trafficking of the Rasa3 BKO B we have documented, and the known role of Rasa3 as a Rap1 GAP implicated Rasa3 in the basal and chemoattractant-mediated integrin activation. First, we checked whether the loss of Rasa3 affected Cxcr5, Cxcr4, or Ccr7 expression or that of several adhesion molecules. We found a small increase in the expression of CD44 on the Rasa3 BKO B cells; otherwise there were no other significant differences (Figure 4A). To assess CXCL13-induced LFA-1 activation, we measured the time-dependent binding of three different concentrations of soluble-complexed ICAM-1 Fc fusion protein (sICAM-1) to wild-type splenic B cells using a standard Ca²⁺/Mg²⁺ buffer. The two highest concentrations of sICAM-1 showed a rapid time-dependent binding to the CXCL13-stimulated B cells (Figure 4B). Using the middle concentration of sICAM-1, we bypassed the proximal signaling pathways by treating the Rasa3^fl/fl and the Rasa3 BKO B cells with phorbol 12-myristate 13-acetate (PMA). This resulted in a similar level of sICAM-1 binding (Figure 4C). However, when we assessed basal LFA-1 activity across time, we found that the unstimulated Rasa3 BKO B cells bound more sICAM-1 than did the control B cells. A similar assessment using a soluble MAdCAM-1 Fc fusion protein (sMAdCAM-1) also demonstrated an elevated basal binding to the Rasa3 BKO B cells (Figure 4D). The higher percentage of cells binding sMAdCAM-1 versus sICAM-1 indicates that at baseline α₄β₇ is partially activated, while LFA-1 is mostly inactive. Next, we stimulated the B cells with CXCL13 and assessed sICAM-1 and sMAdCAM-1 binding over time. Subtracting out the baseline background revealed that the Rasa3 BKO B cells had a poor response to the CXCL13 stimulation. (Figure 4E). These data indicate that Rasa3 actively limits Rap1-GTP signaling in splenic B cells.

Figure 4. Rasa3 maintains low basal LFA-1 and α4β7 activation.

(A) Flow cytometry analysis of chemokine and adhesion related protein expression on splenic B cells. Results shown as mean fluorescent intensity (MFI).

(B) Time-dependent binding of sICAM-1 to CXCL13-stimulated splenic B cells cultured at 37°C.

(C) Time-dependent binding of sICAM-1 to PMA activated (30 min) splenic B cells cultured at 37°C.

(D) Time-dependent binding of sICAM-1 or sMAdCAM-1 to unstimulated splenic B cells cultured at 37°C. Unpaired t test for indicated time points.

(E) Time-dependent binding of sICAM-1 or sMAdCAM-1 to splenic CXCL13-stimulated B cells cultured at 37°C. Basal binding in absence of CXLC13 subtracted. Unpaired t test for indicated time points.

(F) Time-dependent binding sICAM-1 or sMAdCAM-1 to unstimulated splenic B cells on ice. Unpaired t test for indicated time points.

(G) Time-dependent binding sMAdCAM-1 or sICAM-1 to B cells on ice or at 37°C in Mn²⁺ binding buffer. Unpaired t test for indicated time points. *p < 0.05, **p < 0.005, ***p < 0.0005.

Next, to assess input from inside-out signaling, we compared basal integrin binding at 4°C, which largely limits intracellular signaling. In contrast to the results at 37°C, at 4°C equal percentages of Rasa3 BKO and Rasa3^fl/fl B cells bound sMAdCAM-1 and sICAM-1. The differences noted at 37°C disappeared (Figure 4F). We repeated the experiment but substituted a Mn²⁺ buffer for the Ca²⁺/Mg²⁺ buffer to stabilize the high-affinity conformations of α₄β₇ and LFA-1. At 37°C and 4°C, the Mn²⁺ buffer led to nearly all the Rasa3^fl/fl and Rasa3 BKO B cells binding sMAdCAM-1 (Figure 4G), suggesting that Mn²⁺ strongly activates α₄β₇ on B cells. At 4°C using the Mn²⁺ buffer the Rasa3^fl/fl B cells and Rasa3 BKO B cells exhibited a similar binding profile to sICAM-1. However, at 37°C a much higher percentage of the Rasa3 BKO B cells bound the sICAM-1 (Figure 4G). This suggests that at 37°C the Rasa3 BKO B cells Rap1 GEF activity is not opposed by Rasa3, and Rap1-GTP enhances the direct activating effect of Mn²⁺ on ICAM-1 binding, whereas at 4°C those temperature-dependent processes are frozen. Wild-type cells have intact Rasa3, which balances the Rap1 GEF activity and Rap1-GDP predominates, subsequently only a small additional gain in ICAM-1 binding occurs at 37°C.

Loss of Rasa3 impairs CXCL13-induced increases in intracellular calcium, F-actin, Rap1-GTP, and directed cell migration

Although the basal F-actin levels in the Rasa3 BKO B cells slightly exceeded that of control B cells (Figure 5A), when stimulated with CXCL13 the Rasa3 BKO B cells did not raise their F-actin content like the control B cells (Figure 5A). Loss of Rasa3 in B cells also reduced the intracellular calcium response to all the test chemokines and to S1P but did not impact the calcium response to anti-IgM stimulation (Figure 5B). In contrast, the phosphorylation of AKT, ERK, p70S6, and Pyk2 proceeded much like control B cells (Figure 5C). However, loss of Rasa3 B cells elevated the basal level of Rap1-GTP and thereby blunted the fold increase after exposure to CXCL13 (Figure 5D). These results indicate that Rasa3 helps maintains low basal levels of RAP1-GTP and suggests that G_αi signaling results in its rapid deactivation, a step needed to increase Rap1-GTP levels following chemoattractant exposure.

Figure 5. The dynamic regulation of Rap1-GTP by Rasa3 facilitates chemokine receptor signaling.

(A) Basal and CXCL13 induced increased in F-actin as assessed by phalloidin staining and flow cytometry, unpaired t tests.

(B) Time-resolved intracellular calcium responses to indicated chemoattractants and anti-IgM using splenic B cells from Rasa3^fl/fl and Rasa3 BKO mice, unpaired t test of peak responses.

(C) Cell lysates prepared Rasa3^fl/fl and Rasa3 BKO splenic B cells stimulated for various durations with CXCL13 (1 μg/mL) immunoblotted for indicated proteins.

(D) Rap1-GTP and Ras-GTP pull down assay using Rasa3^fl/fl and Rasa3 BKO splenic B cells stimulated for indicated durations with CXCL13 (1 μg/mL). Representative immunoblots shown, fold changes from 3 independent experiments for Rap1-GTP and 2 for Ras-GTP. Results normalized to basal result with Rasa3^fl/fl or Rasa3 BKO B cells.

(E) Basal and specific migration of Rasa3^fl/fl or Rasa3 BKO splenic B cells to indicated chemoattractants. Each result from 12 B cell preparations analyzed in duplicate with chemokines and 5 in duplicate for S1P, unpaired t test.

(F) Total migration of Rasa3^fl/fl or Rasa3 BKO splenic B cells to CXCL13 using chemotaxis chambers coated with indicated concentrations of recombinant ICAM-1 or MAdCAM-1. Each result from 6 B cell preparations analyzed in duplicate, unpaired t test. *p < 0.05, **p < 0.005, ***p < 0.0005.

Examining the migration of Rasa3^fl/fl and Rasa3 BKO B cells to various chemoattractants revealed that Rasa3 BKO B cells basal migration exceeded that of control B cells (Figure 5E). However, when corrected for the different rates of spontaneous migration, Rasa3 BKO B cells migrated less to the high concentrations of chemokines but surprisingly better to the optimal concentration of S1P (Figure 5E). Examining different splenic B cells subsets showed that all subsets exhibited an increase in basal migration. The loss of Rasa3 most affected the specific migration of the T1 B cells subset and least affected the T2-marginal zone precursor B cells (Figure S5). Coating the wells of the migration chamber with either ICAM-1 or MAdCAM-1 slightly exacerbated the difference between the Rasa3^fl/fl and Rasa3 BKO B cells (Figure 5F).

Mixed bone marrow chimeras confirm trafficking defects noted in the Rasa3 BKO mice

To assess how Rasa3 BKO B cells competed versus wild-type cells, we made mixed bone marrow chimera using bone marrow from B6.SJL-Ptprc^a Pepc^b/BoyJ (CD45.1) and either Rasa3^fl/fl C57BL/6J or Rasa3 BKO C57BL/6J (CD45.2). Because we found considerable difference in the trafficking of the CD45.1 and CD45.2 B cells, we compared the recovery and function of the CD45.2 Rasa3 wild-type and Rasa3 BKO B cells. In contrast to the Rasa3 BKO mice, the chimeric mice at the time of assessment had normal appearing spleens, LNs, and PPs. However, Rasa3 BKO B cells underpopulated blood, LNs, and PPs, while overpopulating the spleen. We noted a loss of bone marrow fraction F cells, a skewing of splenic B cell differentiation, elevated numbers of splenic germinal center B cells, but reduced germinal center B cells in LNs and PPs (Figure 6A). Among the various anatomic sites, PPs were the most severely affected with major losses of all the B cell subsets (Figure 6A). Likely as a consequence, the Rasa3 BKO B cells under-populated the lamina propria (Figure 6A). Like the B cells from the Rasa3 BKO mice, the Rasa3 BKO B cells from the chimeric mice had enhanced basal migration and reduced specific migration to CXCL12 and CXCL13 but enhanced migration to S1P (Figure 6B). We also took advantage of the differential markers to compare the localization of the wild-type and Rasa3 BKO B cells in the lamina propria and in PPs. Fewer Rasa3 BKO CD19⁺ cells and CD19⁻ IgA⁺ cells localized in the lamina propria in the mixed chimera mice, and fewer Rasa3 BKO cells localized in PP germinal centers (Figures 6C and 6D). While the mb-1 Cre mice have been extensively used to achieve B lineage deletion, the possibility is not excluded that loss of one allele of mb-1 affected the germinal center response under competitive conditions. Overall, the chimeric mouse studies confirmed the findings in the Rasa3 BKO mice.

Figure 6. Lack of B cells in the lamina propria and low IgA and IgG3 levels in Rasa3 BKO mice.

(A) Flow cytometry results from analyses of lymphoid organs from mixed bone marrow chimeras of Rasa3^fl/fl and Rasa3 BKO (CD45.2), gated on CD45.2⁺ cells, unpaired t test.

(B) Non-specific and specific migration to the indicated chemoattractants using splenic B cells from mixed bone marrow chimeric mice. Representative of 3 experiments, unpaired t test.

(C) Micrographs of IgA⁺ plasma cells in lamina propria of bone marrow chimeras. Reconstituted Rasa3 BKO cells were visualized using a CD45.2 antibody (green), and IgA expression was detected with an IgA antibody (pink). Epithelial cells were stained with Ep-CAM antibody (gray). Stars mark IgA⁺ plasma cells that are also positive for CD45.2. Scale bars represent 25 μm. Right graph depicts the percentage of reconstituted CD19⁺ B cells in the lamina propria, and left graph shows the percentage of reconstituted CD19⁻TCRβ⁻ IgA⁺ plasma cells in the lamina propria of bone marrow chimera mouse, unpaired t test.

(D) Micrographs of germinal centers (GCs) in PPs from bone marrow chimera mice. Reconstituted Rasa3 BKO were visualized using a CD45.2 antibody (green), while CD45.1 wild-type cells were detected with a CD45.1 antibody (pink). The GC within the B cell follicle is indicated. Scale bars represent 100 μm. Graph depicting the ratio of reconstituted Rasa3 BKO and control B cells in the PP GC.

(E) Serum immunoglobulin concentrations in 9- to 12-week-old Rasa3^fl/fl and Rasa3 BKO mice (n = 12 vs. 16). Data are presented as mean ± SEM. Statistical significance was determined using an unpaired t test. *p < 0.05, **p < 0.005, ***p < 0.0005.

Next, we checked whether the maldistribution of B cells had affected the normal levels of serum immunoglobulins. While the Rasa3 BKO mice had normal amounts of IgM, IgG2b, IgG2c, and IgE they had reduced levels of IgA and IgG3, and a mild decrease in IgG1 (Figure 6E). This profile points to a problem in non-follicular and mucosal B cell compartments. The low IgA is consistent with the loss of PPs and reduced number of IgA plasma cells in the lamina propria, whereas the IgG3 defect suggests impaired marginal zone and or B1 B cell function. However, the normal IgM, IgG2a/2b, and IgE indicate that basic B cell activation and Th1/Th2 cytokine switching are preserved.

Both G_αi2-GDP and G_αi2-GTP colocalize at the plasma membrane with Rasa3, but only G_αi2-GTP is in close physical proximity

To examine whether G_αi2 affected RASA3 localization, we expressed fluorescently tagged human RASA3 in HeLa cells in the presence or absence of G_αi2-YFP or G_αi2Q205L-YFP, which lacks GTPase activity and hence remains GTP bound (Figure 7A). In HeLa cells both G_αi2-YFP and G_αi2Q205L-YFP localized both at the plasma membrane and in the cytoplasm. RASA3 was predominately localized at the plasma membrane, in contrast to the more diffuse RASA2 expression pattern. Co-expression of the G_αi2 constructs or RAP1A did not alter the RASA3 membrane expression. In addition to its plasma membrane expression, much of the Rap1A localized in the cytosol. TIRF imaging revealed RASA3 and G_αi2Q205L enrichment in membrane filopodia and microspikes (Figure 7A). To assess the localization of RASA3 in human B cells, we used RPMI 8866 cells and derived permanently transfected cells lines that expressed varying levels of RASA3-mScarlet (RASA3 transfectants [TFs]). Immunoblotting cell lysates showed varying expression levels of the RASA3 construct. As we detected two bands with the RASA3 antibody, we re-blotted with a red fluorescent protein-specific antibody, which identified a single band suggesting a partial loss of the fluorescent tag. The RPMI 8866 RASA3 cells had higher CCR7 and slightly decreased CXCR4 and integrin β7 expression. The RASA3-mScarlet expression pattern in the RPMI 8866 cells mirrored that we had observed in HeLa cells (Figure 7B). Less sICAM-1 and sMAdCAM-1 bound to the RASA3 TFs compared to the parent cell line. However, in contrast to the primary murine B cells, the addition of CXCL12 or CCL19 did not increase the basal binding observed in the absence of chemokine (Figure S6). Transiently transfecting the RASA3-mScarlet B cell transfectants with G_αi2Q205L YFP allowed assessment of the coordinate expression of RASA3 and G_αi2-GTP during live cell imaging. G_αi2Q205L and RASA3-mScarlet colocalized at the plasma membrane and were both enriched in filopodia (Figure 7C; Video S7). A co-expressed RAP1-GTP sensor colocalized with RASA3 in membrane ruffles and lamellipodia, known sites enriched for Rap1-GTP (Figure 7D; Video S8).

Figure 7. Plasma membrane expression of RASA3 and close association with G_αi2-GTP.

(A) Expression of fluorescently tagged constructs in HeLa cells. Single confocal slices and co-localization shown. Top row images inverted. Bottom row images from live cell TIRF microscopy. Scale bars represent 3 μm.

(B) RASA3-mScarlet expression in RPMI 8866, a human B cell line. Immunoblot, flow cytometry, and single confocal slice and brightfield overlay shown. Scale bar represents 3 μm. Unpaired t test.

(C) Expression of RASA3-mScarlet and G_αi2Q205L-YFP in RPMI 8866 cells seeded on ICAM-1/CXCL12-coated plates. A time-lapse image sequence of a 3-μm z-projection was acquired. Individual and colocalization channels shown. Scale bar represents 5 μm.

(D) Expression of RASA3-mScarlet and RalGDS-GFP in RPMI 8866 cells seated on ICAM-1/CXCL12-coated places. A time-lapse image sequence of a 3-μm z-projection was acquired. Individual and colocalization channels shown. Scale bar represents 5 μm.

(E) FLIM using RASA3-GFP or GFP as the donor with indicated acceptors. Quantification of RASA3-GFP or GFP lifetimes. Unpaired t test. *p < 0.05, **p < 0.005, ***p < 0.0005.

To assess whether G_αi2 and RASA3 resided in the same plasma membrane nanodomains, we performed Förster resonance energy transfer (FRET)-based fluorescence lifetime imaging (Figure 7E). The fluorescence lifetime of the RASA3-GFP donor was partially quenched by the presence of the G_αi2-mScarlet acceptor, but pre-treating the HeLa cells with pertussis toxin eliminated the shortened lifetime indicating that RASA3 was closely apposed to GTP bound (within 10 angstroms), not GDP-bound G_αi2. Furthermore, expression of the G_αi2Q205L-mScarlet fluorescent acceptor decreased the lifetime of RASA3-GFP donor significantly more than did wild-type G_αi2-mScarlet. The G_αi2Q205L quenching was mainly observed at discrete sites along the plasma membrane. Thus, in B cells G_αi, RAP1, and RASA3 colocalize at the plasma membrane but only when G_αi is GTP bound can it impact RASA3. This underscores the importance of G_αi nucleotide exchange as a mechanism to rapidly upregulate RAP1 signaling by releasing the RASA3 brake.

DISCUSSION

In the absence of an active GEF Rap1 is GDP bound. Due to its high affinity for the GTPase active site, GDP release is very slow. To promote GTP binding, a Rap1 GEF catalyzes the dissociation of GDP, the typically rate-limiting step in the GTPase cycle.³⁰ In contrast, Rasa3 rapidly triggers GTP hydrolysis by introducing a catalytic residue into the GTPase active site. In quiescent B cells, basal Rap1 GEF activity exchanges GDP for GTP, but because Rasa3 is active it ensures that Rap1-bound GTP is rapidly hydrolyzed to GDP, thereby favoring the GDP-bound or off status of Rap1. By limiting the duration that Rap1 remains GTP bound, Rasa3 prevents spurious Rap1-GTP signaling that would be detrimental to normal B cell function and homeostasis. Supporting constitutive Rap1 GEF activity in B cells shifting the Rasa3 BKO B cells to 4°C, which prevents constitutive signaling, normalized their binding to sICAM-1 and sMAdCAM-1. Simultaneous Rap1 GEF and GAP activity provides a mechanism by which normal B cells can very rapidly upregulate Rap1-GTP levels by turning off Rasa3, thereby circumventing the need for the slower Rap1 GEF activation. This mechanism is much like pressing the accelerator and brake of a car simultaneously and then removing the brake to quickly accelerate or in this case rapidly increase active Rap1-GTP.

This loss of a brake in the Rasa3 BKO B cells led to high basal levels of Rap1-GTP, which precluded the usual dynamic regulation of the GDP/GTP-bound status of Rap1. In turn, this caused a striking phenotype characterized by a severe maldistribution of B cells within lymphoid organs and a major reduction in blood B cells. The blood and PPs had approximately 90% fewer B cells than expected and the inguinal and mesenteric LNs had 50% less. Despite the lymphopenia and reduced LN and PP B cells, the overall B cell population in the Rasa3 BKO mice was not reduced as the spleen had a 2-fold increase, and the thymus and peritoneal cavity each had a 2.5-fold increase. Furthermore, serum immunoglobulin levels except IgA and IgG3 were normal. The B cell lymphopenia likely arose from the retention of B cells within the spleen, lungs, and perhaps the liver. Our data further suggest that the splenic B cell retention resulted from defects in the Rasa3 BKO B cell egress from both the white pulp and the red pulp as we noted an expansion of B cell follicles and an accumulation of B cells around red pulp sinuses. Thus, the loss of Rasa3 and dysregulated regulation of Rap1 in B cells interfered with the signals that control B cell splenic retention and egress. The expansion of splenic B cells contrasts with the 50% reduction of splenic T cells noted in mice generated by crossing Rasa3^fl/fl to Cd4^cre/+ mice,¹⁶ suggesting a fundamental difference in how Rasa3 affects B and T cell trafficking through the spleen.

While the B cell lymphopenia partially explained the lack of Rasa3 BKO B cells in LNs, our data indicate that trafficking defects also contributed. Adoptively transferred Rasa3 BKO B cells were retained on HEVs due to poor transmigration, and they did not properly egress from LNs. Furthermore, the loss of Rasa3 more severely compromised PPs. Adoptively transferred Rasa3 BKO B cells adhered poorly to the PP HEVs, failed to traffic properly, and rapidly egressed. The differential roles of LFA-1/ICAM-1 and α4β7/MAdCAM-1 interactions in LN and PP entry likely explained the contrasting behavior of the Rasa3 BKO B cells in the LN and PP HEVs. In ex vivo flow chamber studies, LFA-1/ICAM-1 interactions mediate slow T cell rolling and T cell firm adhesion. Both processes depend upon Rap1-GTP. In contrast, α4β7/MAdCAM-1 interactions mediate T cell tethering, rolling, and arrest, but only T cell arrest depends upon increases in Rap1-GTP.³¹ Intriguingly, dual Rasa3/Sip1a KO T cells strongly adhere to solid phased MAdCAM-1 under static conditions but fail to adhere under flow.³¹ An analogous situation followed the adoptive transfer of Rasa3 BKO B cells into wild-type mice where they accumulated on the inguinal LN HEVs but failed to do so on the PP HEVs. While this aligns with the in vitro T cell experiments, surprisingly the Rasa3-deficient T cells normally populated mLNs and PPs, while the Rasa3-deficient B cells clearly did not.

Another mouse model with a B cell trafficking defect resembles the phenotype described here. Mice with a single amino acid change in G_αi2 at position 184, which changes a glycine residue to a serine (G184S), express a G_αi2 protein no longer subject to the GAP activity of RGS proteins.³² The loss of RGS protein interactions extends the half-life of G_αi2-GTP, disrupting normal GTP/GDP cycling. Similarities between the models include the overall distribution of B cells; increases in non-specific but decreases in specific B cell migration to chemokines; and decreased intracellular calcium responses. A possible connection between the two models is the interaction between G_αi2-GTP and Rasa3.¹⁸ While in many cell types G_αi-GTP targets adenylyl cyclase isoforms AC1, AC5, and AC6, B cells express very low levels of those isoforms; rather, they express AC7, which does not bind G_αi-GTP.^13,33 The lack of G_αi binding adenylyl cyclase isoforms may increase the likelihood that G_αi2-GTP targets Rasa2 and Rasa3. Supporting this contention are the previously mentioned human G_αi2 mutations that impair its intrinsic GTPase activity and which sequester RASA2, limiting its GAP activity for RAS.^19,34 These patients exhibit a complex phenotype, some of which result from the sequestration of RASA2, but also likely RASA3. Furthermore, the sequestration of RASA2 by G_αi2-GTP may provide an explanation for the increased germinal centers and heightened B cell proliferation noted previously in the G_αi2 G184S mice.³²

Like Rasa2, the interaction of Rasa3 with G_αi-GTP should rapidly segregate Rasa3 from Rap1-GTP promoting Rap1-GTP signaling. No structural information informs how Rasa3 binds to G_αi, although it presumably binds the G_αi switch regions as it interacts with G_αi-GTP, not G_αi-GDP. Rasa3 lacks any structural similarity to RGS proteins, which also bind G_αi-GTP. G_βγ subunits also bind G_αi-GTP, directly competing with RGS proteins.³⁵ Thus, the relative balance between RGS proteins, freed G_βγ subunits, and Rasa3 likely fine tunes Gi signaling in B cells and likely other cell types. Supporting a role for Rasa3 in facilitating Gi signaling the early intracellular calcium response and the rapid rise in F-actin following CXCL13 exposure only reach 50%–60% of that observed in control B cells. While further studies are needed to understand the underlying mechanism, one possibility is that the G_αi-GTP/Rasa3 interaction promotes RGS binding/function accelerating GTP hydrolysis. A precedent exists for such a possibility as the Gq-effector GRK2 enhances the activity of RGS4, a relatively poor Gq GAP.³⁶ Thus, the absence of RASA3 would impair the termination/resetting of Gi signaling, which fits the observed chemokine signaling defects.

In conclusion, the loss of Rasa3 in B cells revealed a complex phenotype that overlapped but differs in several substantial ways from that observed with loss of Rasa3 in T cells.¹⁶ Severe lymphopenia occurred in both models, yet the Rasa3-deficient T and B cells variably populated the mesenteric LNs, PPs, and spleen. The loss of Rasa3 in both models led to an increase in basal level of Rap1-GTP and a loss in the dynamic regulation of GDP/GTP-bound forms of Rap1. However, chemokines elicited an exaggerated increase in sICAM-1 binding to Rasa3 KO T but a subnormal response using the Rasa3 BKO B cells. Both the KO T and B cells exhibited defects in LN entry and egress although the Rasa3 BKO rapidly exited LNs while Rasa3 BKO B cells often reverse migrated from the efferent lymph to the LN parenchyma. Further studies directly comparing Rasa3 BKO T and B cells may provide insights into their respective lymphoid organ trafficking patterns. Our study provides insights into the interconnections between the signaling pathways that engage both large and small G-proteins and in particular how Rasa3 and G_αi collaborate to modulate both G_αi and Rap1 signaling to maintain B lymphocyte homing and homeostasis. Finally, pharmaceutically targeting Rasa3 offers a means to indirectly manipulate integrin adhesiveness, cell migration, and immune cell trafficking.

Limitations of the study

Structural data demonstrating the G_αi2-GTP/Rasa3 interaction is lacking. An AlphaFold-predicted structure for Rasa3 is available, which may serve to dock Rasa3 with the known structure of G_αi-GTP. This may help unravel how this interaction impacts G_αi and Rap1 signaling and reveal whether Rasa3/G_αi-GTP interactions impacts the binding of RGS proteins and G_βγ subunits. Only indirect evidence supports the contention that G_αi2-GTP limits the Rap1 GAP activity of Rasa3.

RESOURCE AVAILABILITY

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, John H. Kehrl (jkehrl@niaid.nih.gov).

Materials availability

All unique/stable reagents generated in this study are available from the lead contact with a completed materials transfer agreement.

STAR★METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

C57BL/6 (MGI:2159736, IMSR_JAX:000664) mice and the CD45.1 (MGI:4819849, IMSR_JAX:002014) mice were obtained from the Jackson Laboratory. The C57/BL6 mb1^cre/+ mice were kindly provided by Dr. M. Reth and were bred with the Rasa3^fl/fl (MGI:5756029) mice³⁷ to generate the Rasa3^fl/flmb1^cre/+ mice. For the bone marrow reconstitutions, 6-wk-old CD45.1 mice were twice irradiated with 550 rad for total of 1100 rad. Mixed chimeric mice were made by reconstituting the irradiated CD45.1 mice with a 1:1 mixture of bone marrow prepared from C57BL/6 CD45.1 wild-type mice and either CD45.2 Rasa3^fl/fl mice or CD45.2 Rasa3^fl/flmb1^cre/+ mice. The engraftment was monitored by sampling the blood 28 d later. The mice were used 6–8 weeks after reconstitution. All mice used in this study were 6–14 weeks of age. Mice were housed under specific pathogen–free conditions. All the animal experiments and protocols used in the study were approved by the National Institute of Allergy and Infectious Diseases’ Animal Care and Use Committee at the National Institutes of Health. Ethical approval number is LIRID 15.

RPMI 8866 cells were kindly provided by Dr. Claudia Cicala, NIAID, NIH. The cell lines were obtained from Millipore Sigma (95041316-1VL, Sigma-Aldrich). Authentication was not independently performed in our laboratory. Mycoplasma contamination testing was not performed during this study. The cells were transfected with RASA3-mScarlet using the Nucleofector transfection protocol (Amaxa biosystems nucleofector I, X-01 program). For stable cell line generation, cells were cultured in RPMI with 10% FBS for 40 h, then selected with G418 (400 μg/mL). RASA3-mScarlet-expressing cells were FACS-sorted after 3 weeks initially into 3 different populations based on mScarlet fluorescent intensity. The pool of cells expressing the highest level of RASA3-mScarlet were maintained in G418 (200 μg/mL).

METHOD DETAILS

Cells

Spleens, LNs, and PPs were removed and gently dissociated into single-cell suspensions. Bone marrow cells were collected by flushing isolated femurs with PBS. Peripheral blood samples were collected by mandibular vein bleeding. After removing RBCs with Tris–NH₄Cl, the cells were resuspended in PBS containing 1% BSA at 4°C. B cells were isolated by negative depletion using biotinylated Abs to CD4, CD8, CD3, CD117, CD11b, CD11c, Gr-1 (Ly-6C and Ly-6G), and TCRgd and Dynabeads M-280 Streptavidin (Thermo Fisher Scientific). The B cell purities were >95%. When needed, splenic, lymph, bone marrow, or B cells were cultured in RPMI 1640 containing 10% heat-inactivated fetal calf serum, 2 mM L-glutamine, antibiotics (100 IU/mL penicillin and 100 μg/mL streptomycin), 1 mM sodium pyruvate, and 50 μM 2-ME. For ICAM-1 and MAdCAM-1 binding assays, purified B cells were resuspended in Ca²⁺/Mg²⁺, or Mn²⁺ binding buffer (PBS supplemented with 0.1% fatty-acid–free BSA, 0.5 mM MgCl₂, and 0.9 mM CaCl₂, or 1mM MnCl₂) and incubated for at least 30 min at 37°C. For F-actin polymerization assays, cells were resuspended in Opti-MEM Medium and incubated for at least 30 min at 37°C.

Flow cytometry

Single cells were resuspended in PBS, 2% FCS, and strained with fluorochrome-conjugated or biotinylated Abs against B220 (RA3-6B2), CD19 (1D3), CD23 (B3B4), CD21/35 (4E3), CD93 (AA4.1), CD43 (S7), IgD (11-26c-2a), IgM (R6-60.2), CD24 (M1/69), BP-1 (6C3), CD3 (145-2C11), CD4 (GK1.5 or RM4-5), CD8 (53–6.7), CD11c (HL3), CD11b (M1/70), CD184 (CXCR4, 2B11), CCR7 (4B12), CXCR5 (2G8), CD11a (M17/4), CD49d (9C10, MFR4.B), CD54 (3E2), CD62L (MEL-16), NK1.1 (PK136), TCRγδ (GL3), Ly-6G (1A8), Ly-6C (AL-21), CD45.1 (A20), or CD45.2. Biotin-labeled Abs were visualized with fluorochrome-conjugated streptavidin. A LIVE/DEAD Fixable Aqua Dead Cell Stain Kit was used in all experiments to exclude dead cells. Compensation was performed using CompBeads and ArC Amine Reactive Compensation Beads individually stained with each fluorochrome. Compensation matrices were calculated with FACSDiva software. Data acquisition was done on FACSCelesta SORP or FACSymphony A3 flow cytometer and analyzed with FlowJo software version 10.10.0.

Chemotaxis and ICAM-1/MAdCAM-1–coated plate migration assays

Chemotaxis assays were performed using 5-μm HTS Transwell 96-well supports. Splenic B cells were immunostained for B cell subsets with fluorochrome-conjugated antibodies against B220, CD21, CD23, and CD24, washed twice, resuspended in complete RPMI 1640 medium containing charcoal-stripped FCS, and incubated for 1 h at 37°C. Cells were then added to the upper wells of the 96-well Transwell plate. The lower wells contained various concentrations of CCL19, CXCL12, CXCL13, or S1P in 240 μL of medium. After 3 h of incubation at 37°C, cells that had migrated to the lower wells were counted using a FACSCelesta SORP or FACSymphony A3 flow cytometer. Nonspecific migration was defined as the percentage of cells migrating in the absence of chemoattractant, and specific migration was calculated by subtracting nonspecific migration from total migration. Percent migration was calculated as the number of cells in the lower well divided by the total number of cells in the starting suspension, multiplied by 100. For the ICAM-1– or MAdCAM-1–coated plate migration assay, 96-well plates were coated with recombinant ICAM-1/Fc or MAdCAM-1/Fc (1, 5, or 10 μg/mL) overnight at 4°C. Plates were then washed twice with PBS before adding cells. Migration was assessed using the same procedure described for the chemotaxis assays.

Immunoblotting

Splenic B cells were lysed in RIPA buffer with 1 mM NaF, 1 mM PMSF, 1 mM DTT, and 1 mM Na₃VO₄ with cOmplete protease inhibitor cocktail and PhosStop phosphatase inhibitor tablets. Cell lysates were separated by SDS-PAGE and transferred to nitrocellulose membranes by the iBLOT gel transfer system. Membranes were incubated with 5% nonfat milk (w/v) in TBS buffer (25 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.1% Tween 20) for 1 h and incubated with the primary antibody in TBS buffer with 2.5% nonfat milk or 5% BSA (w/v) overnight by shaking at 4°C. The appropriate second antibodies conjugated to HRP were used to detect the protein of interest with enhanced chemiluminescence (ECL). Primary and secondary antibodies used, their source, and dilutions are in Resource Table.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Biotin anti-mouse CD117 (clone 1B8)	BD Biosciences	Cat# 553353; RRID: AB_394804
Biotin anti-mouse CD11b (clone M1/70)	BD Biosciences	Cat#553309; RRID: AB_394773
Biotin anti-mouse CD11c (clone HL3)	BD Biosciences	Cat#553800; RRID: AB_395059
Biotin anti-mouse CD3e (clone 145-2C11)	BD Biosciences	Cat#553060; RRID: AB_394593
Biotin anti-mouse CD4 (clone GK1.5)	BioLegend	Cat#100404; RRID: AB_312689
Biotin anti-mouse CD8 (clone 53–6.7)	BioLegend	Cat#100704; RRID: AB_312743
Biotin anti-mouse Gr-1 (clone RB6-8C5)	BD Biosciences	Cat#1088430; RRID: AB_394641
Biotin anti-mouse NK1.1 (clone PK136)	BD Biosciences	Cat# 553163; RRID: AB_394675
Biotin anti-mouse TCRγδ (clone GL3)	BD Biosciences	Cat# 553176; RRID: AB_394687
Biotin anti-mouse TER-119 (clone TER-119)	BD Biosciences	Cat#553672; RRID: AB_394985
PE/Cyanine7 anti-mouse CD196 (CCR6) Antibody	BioLegend	Cat#129816; RRID: AB_2072798
PE anti-mouse CD197 (CCR7) Antibody	BioLegend	Cat# 120106; RRID: AB_ 389358
PE anti-mouse CD11a Antibody	BioLegend	Cat# 101107; RRID: AB_ 312780
Brilliant Violet 650™ anti-mouse/human CD11b	BioLegend	Cat# 101259; RRID: AB_2566568
Brilliant Violet 785™ anti-mouse CD11c	BioLegend	Cat# 117336; RRID: AB_2565268
PE/Cyanine7 anti-mouse CD11c	BD Biosciences	Cat# 558079; RRID: AB_647251
PE anti-mouse CD138	BD Biosciences	Cat#553714; RRID: AB_395000
PE/Dazzle™ 594 anti-mouse CD19	BioLegend	Cat#115554; RRID: AB_2564001
Brilliant Violet 421™ anti-mouse CD21/CD35 (CR2/CR1) (clone 7E9)	BioLegend	Cat#123422; RRID: AB_2650891
FITC anti-mouse CD21/CD35 (CR2/CR1) Antibody	BioLegend	Cat#123408; RRID: AB_940405
Brilliant Violet 605™ anti-mouse CD23	BioLegend	Cat# 101637; RRID: AB_2832279
PE anti-mouse CD23 (clone B3B4)	BioLegend	Cat# 101608; RRID: AB_312833
Pacific Blue™ anti-mouse CD24 (clone M1/69)	BioLegend	Cat#101820; RRID: AB_572011
Brilliant Violet 421™ anti-mouse CD24 (clone M1/69)	BioLegend	Cat#101820; RRID: AB_572011
Brilliant Violet 605™ anti-mouse CD24 Antibody	BioLegend	Cat#101827; RRID: AB_2563464
PE/Cyanine7 anti-mouse CD24 (clone M1/69)	BioLegend	Cat#101822; RRID: AB_756048
PE anti-mouse Ly-51 Antibody	BioLegend	Cat#108308; RRID: AB_313365
PE anti-Mouse CD25	BD Biosciences	Cat#561065; RRID: AB_395101
FITC anti-Mouse CD3 Molecular Complex	BD Biosciences	Cat#555274; RRID: AB_395698
Pacific Blue™ anti-mouse CD3	BioLegend	Cat#100214; RRID: AB_100214
Brilliant Violet 421™ anti-mouse CD3	BioLegend	Cat#100228; RRID: AB_2562553
Pacific Blue™ anti-mouse CD38	Invitrogen	Cat#A15393; RRID: AB_2534407
Alexa Fluor® 700 anti-mouse CD4	BioLegend	Cat#100430; RRID: AB_493699
APC/Cyanine7 anti-mouse CD4	BioLegend	Cat#100414; RRID: AB_312699
Brilliant Violet 785™ anti-mouse CD4	BioLegend	Cat#100453; RRID: AB_2565843
APC anti-mouse CD43	BD Biosciences	Cat#560663; RRID: AB_1727479
FITC anti-mouse CD44	BD Biosciences	Cat#553133; RRID: AB_2076224
PE anti-mouse CD44	BD Biosciences	Cat#553134; RRID: AB_394649
Brilliant Violet 711™ anti-mouse CD45	BioLegend	Cat#103147; RRID: AB_2564383
PE anti-mouse CD45.1	BioLegend	Cat# 110708; RRID: AB_313497
Brilliant Violet 421™ anti-mouse CD45.1	BioLegend	Cat#110732; RRID: AB_2562563
Alexa Fluor® 488 anti-mouse CD45.2	BioLegend	Cat# 109816; RRID: AB_492868
APC anti-mouse CD45.2	BioLegend	Cat#109814; RRID: AB_389211
APC anti-mouse/human CD45R/B220 (clone RA3-6B2)	BioLegend	Cat#103212; RRID: AB_312997
Brilliant Violet 605™ anti-mouse/human CD45R/B220 (clone RA3-6B2)	BioLegend	Cat#103243; RRID: AB_2563312
Brilliant Violet 650™ anti-mouse/human CD45R/B220 (clone RA3-6B2)	BioLegend	Cat#103241; RRID: AB_11204069
Brilliant Violet 750™ anti-mouse/human CD45R/B220 (clone RA3-6B2)	BioLegend	Cat#103261; RRID: AB_2734157
PerCP/Cy5.5 anti-mouse B220 (clone RA3-6B2)	BioLegend	Cat#103236; RRID: AB_893354
PE anti-mouse CD49d	BioLegend	Cat#103705; RRID: AB_313047
Brilliant Violet 421™ anti-mouse CD5	BioLegend	Cat#100629; RRID: AB_2565430
PerCP/Cyanine5.5 anti-mouse CD5	BioLegend	Cat#100624; RRID: AB_2563433
PE anti-mouse CD54	BioLegend	Cat#116108; RRID: AB_313699
PE anti-mouse CD62L	BioLegend	Cat#104408; RRID: AB_313095
APC anti-mouse CD62L	BioLegend	Cat#104412; RRID: AB_313099
PE/Cyanine7 anti-mouse CD69	BD Biosciences	Cat#561930; RRID: AB_394508
PE/Cyanine5 anti-mouse CD8a	BioLegend	Cat#100710; RRID: AB_312749
Alexa Fluor® 700 anti-mouse CD8a	BioLegend	Cat#100730; RRID: AB_493703
Brilliant Violet 605™ anti-mouse CD95	BD Biosciences	Cat# 740367; RRID: AB_ 2740099
PE anti-mouse CD184 (CXCR4)	Invitrogen	Cat#12-9991-82; RRID: AB_891391
PE anti-mouse CD185 (CXCR5)	BD Biosciences	Cat# 551959; RRID: AB_394300
APC anti-mouse CD185 (CXCR5)	BD Biosciences	Cat# 560615; RRID: AB_1727519
APC anti-mouse F4/80	BioLegend	Cat#123116; RRID: AB_893481
FITC anti-mouse/human GL7 Antigen	BioLegend	Cat# 144604; RRID: AB_2561697
PE anti-mouse IgA	Invitrogen	Cat#12-4204-82; RRID: AB_465917
APC/Cyanine7 anti-mouse IgD	BioLegend	Cat#405716; RRID: AB_10662544
PerCP/Cyanine5.5 anti-mouse IgD	BioLegend	Cat#405710; RRID: AB_1575113
APC anti-mouse IgG1	BD Biosciences	Cat#550874; RRID: AB_398470
PE Mouse IgG1, κ Isotype Ctrl	BioLegend	Cat# 981804; RRID: AB_3076354
PE Rat IgG2a, κ Isotype Ctrl	BioLegend	Cat# 400508; RRID: AB_326530
PE Mouse IgG2b, κ Isotype Ctrl	BioLegend	Cat# 402204; RRID: AB_3097041
PE Armenian Hamster IgG Isotype Ctrl	BioLegend	Cat# 400908; RRID: AB_326593
PE/Cyanine7 anti-mouse IgM	BD Biosciences	Cat#552867; RRID: AB_394500
FITC anti-mouse Ly-6C	BD Biosciences	Cat#553104; RRID: AB_394628
APC anti-mouse Ly6G	BD Biosciences	Cat#560599; RRID: AB_1727560
PerCP/Cyanine5.5 anti-mouse Ly6G	BD Biosciences	Cat#560602; RRID: AB_1727563
PE/Cyanine7 anti-mouse Ly6G	BD Biosciences	Cat#560601; RRID: AB_1727562
PE anti-mouse I-A/I-E	BioLegend	Cat# 107608; RRID: AB_313323
PE/Cyanine7 anti-mouse NK-1.1	BioLegend	Cat# 108714; RRID: AB_389364
FITC anti-mouse NK-1.1	BioLegend	Cat#108706; RRID: AB_313393
PE/Cyanine7 anti-mouse CD279 (PD-1)	Invitrogen	Cat#25-9985-82; RRID: AB_10853805
APC/Cyanine7 anti-mouse Siglec F	BD Biosciences	Cat#565527; RRID: AB_2732831
PerCP/Cy5.5 anti-mouse TCRb	Invitrogen	Cat#45-5961-82; RRID: AB_925763
FITC anti-mouse TCRgd	Invitrogen	Cat# 11-5711-82; RRID: AB_465238
PE/Cyanine7 anti-mouse TCRgd	Invitrogen	Cat# 25-5711-82; RRID: AB_2573464
Alexa Fluor® 488 anti-mouse CD31 (clone 390)	BioLegend	Cat# 102414; RRID: AB_493408
Alexa Fluor® 647 anti-mouse/human PNAd	BioLegend	Cat# 120808; RRID: AB_2783060
Alexa Fluor® 488 anti-mouse LYVE-1	R&D Systems	Cat# FAB2125G; RRID: AB_10892330
PE anti-mouse Ki67	Invitrogen	Cat# 12-5698-82; RRID: AB_11150954
PE anti-mouse CD326 (EpCAM)	Invitrogen	Cat# 17-5791-82; RRID: AB_2716944
Alexa Fluor® 647 anti-mouse/human CD324 (E-Cadherin)	BioLegend	Cat# 147308; RRID: AB_2563955
FITC anti-mouse CD3	BioLegend	Cat# 100204; RRID: AB_312661
CD21/35 imaging	BioLegend	Cat# 123424 RRID: AB_2629578
Purified NA/LE anti-mouse CD62L	BD Biosciences	Cat# 553147; RRID: AB_394662
Alexa Fluor® 594 anti-mouse CD31 (clone 390)	BioLegend	Cat# 102432; RRID: AB_2617017
AffiniPure® F(ab’)2 Fragment Goat Anti-Mouse IgM, μ chain specific (min X Hu, Bov, Hrs Sr Prot)	Jackson ImmunoResearch	Cat# 115-006-075; RRID: AB_2338474
Alexa Fluor® 647 AffiniPure® F(ab’)2 Fragment Goat Anti-Human IgG, Fcγ fragment specific (min X Bov, Ms, Rb Sr Prot)	Jackson ImmunoResearch	Cat# 115-006-075; RRID: AB_2337902
Alexa Fluor™ 647 Phalloidin	Thermo Fisher Scientific	Cat#A22287; RRID: AB_2620155
Phospho-Pyk2 (Tyr402)	Cell Signaling	Cat# 3291S; RRID: AB_2300530
Pyk2	Cell Signaling	Cat# 3292S; RRID: AB_ 2174097
Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) (E10)	Cell Signaling	Cat# 9106S; RRID: AB_331768
p44/42 MAPK (Erk1/2)	Cell Signaling	Cat# 9102S; RRID: AB_330744
Phospho-p70 S6 Kinase (Thr389) (D5U1O)	Cell Signaling	Cat# 97596S; RRID: AB_2800283
p70 S6 Kinase (49D7)	Cell Signaling	Cat# 2708S; RRID: AB_390722
GAP1-InsP4 BP (RASA3) Antibody	Santa Cruz	Cat# sc-398283
Akt (pan) (11E7)	Cell Signaling	Cat# 2920S; RRID: AB_1147620
Phospho-Akt (Ser473) (D9E)	Cell Signaling	Cat# 4060L; RRID: AB_2315049
Anti-Rap1 (Kit component)	Thermo Fisher Scientific	Cat# 1862344
Active Rap1 Pull-Down and Detection Kit	Thermo Fisher Scientific	Cat# 16120
Anti-Ras	Cell Signaling	Cat# 3965S; AB_2180216
Active Ras Pull-Down and Detection Kit	Thermo Fisher Scientific	Cat# 16117
Anti-β-Actin– Peroxidase-linked	Sigma-Aldrich	Cat# A-3854; RRID: AB_262011
Anti-rabbit IgG, HRP-linked	Cell Signaling	Cat# 7074; RRID: AB_2099233
Anti-mouse IgG, HRP-linked	Cell Signaling	Cat# 7076; RRID: AB_330924
LIVE/DEAD™ Fixable aqua dead cell stain kit	Thermo Fisher Scientific	Cat#L34966
LIVE/DEAD™ Fixable yellow dead cell stain kit	Thermo Fisher Scientific	Cat#L34968
LIVE/DEAD™ Fixable Near-IR dead stain kit	Thermo Fisher Scientific	Cat#L34976
LEGENDplex™ Mouse Immunoglobulin Isotyping Panel (6-plex) with V-bottom Plate	BioLegend	Cat# 740493
LEGENDplex™ Mouse IgE Assay with V-bottom Plate	BioLegend	Cat# 740770
Chemicals, Peptides, and Recombinant Proteins
Evans Blue	Sigma-Aldrich	Cat#E2129-10G
7α,25-dihydroxy Cholesterol (7a, 25-OHC)	Cayman Chemical	Cat# 11032
Sphingosine 1-phosphate, ≥95%	Sigma-Aldrich	Cat# S9666-1MG
Sphingosine-1-phosphate (d17:1, S1P)	Cayman Chemical	Cat# 22498
Phorbol 12-myristate 13-acetate (PMA)	Cayman Chemical	Cat# 10008014
Recombinant murine CXCL12	PeproTech	Cat#250-20A
Recombinant murine CCL19	PeproTech	Cat# 250-27B
Recombinant murine CXCL13	PeproTech	Cat# 250-24
Recombinant Mouse ICAM-1/CD54 Fc Chimera Protein, CF	R&D Systems	Cat# 796-IC-050
Recombinant Mouse MAdCAM Fc Chimera Protein, CF	R&D Systems	Cat# 993-MC-050
Fixation Buffer	BioLegend	Cat# 420801
Intracellular Staining Permeabilization Wash Buffer (10x)	BioLegend	Cat# 421002
FLIPR Calcium 4 Assay kit	Molecular Devices	Cat# R8142
Bovine Serum albumin fatty acid free (low endotoxin)	Sigma-Aldrich	Cat# 8806
2,2,2-Tribromoethanol (Avertin)	Sigma-Aldrich	Cat# T48402-5G
2-methyl-2-butanol	Sigma-Aldrich	Cat# 152463
Alexa Fluor 647 protein labeling kit	Thermo Fisher Scientific	Cat# A20173
CellTracker™ Green CMFDA Dye (5-chloromethylfluorescein diacetate)	Thermo Fisher Scientific	Cat# C2925
FLIPR Calcium 4 assay kit	Molecular Devices	Cat# R8142
Pertussis toxin	Sigma-Aldrich	Cat# 516560
Software and Algorithms
SOFT max Pro 7.x	Molecular Devices	https://www.moleculardevices.com/
FlowJo v10.8.x	Tree Star	https://www.flowjo.com/
ImageJ 1.53a	NIH	https://imagej.nih.gov/ij/
Imaris v10.0.1	Oxford Instruments	http://www.bitplane.com/imaris
Prism v9.3.1	GraphPad	https://www.graphpad.com/scientific-software/prism/
Adobe Photoshop	Adobe	https://www.adobe.com
LASX-Single molecule detection software	Leica	https://www.leica-microsystems.com/products/microscope-software/p/leica-las-x-ls/
Other
Flow cytometers		See Table S1
RASA3-GFP		This study
RASA3-mScarlet		This study
YFP-RASA2		Gift Helen Su, NIAID, NIH
Gαi2-YFP		Gift Al Gilman, UTSW
Gαi2Q205L YFP		Gift Al Gilman, UTSW
Gαi2-mScarlet		This study
Gαi2Q205L mScarlet		This study
RalGDS-GFP		Gift Tatsuo Kinashi, Kansai Medical University
Rap1A-OFP	Sinobiological	Cat# MG52137-ACR
mScarlet	Addgene	Cat#189754

Ras and Rap1 activation assay

Ras-GTP and Rap1-GTP pulldown assays were performed according to the manufacturer’s protocol. Briefly, single-cell suspensions of splenic B cells isolated from Rasa3^fl/fl and Rasa3^fl/flmb1^cre/+ mice were prepared. Cells were washed twice and resuspended in Opti-MEM. For each time-point, 2.5 × 10⁷ cells were preincubated for 10 min at 37°C and then stimulated with CXCL13 (1 μg/mL). Activation was stopped by adding ice-cold PBS, followed by centrifugation and lysis in ice-cold 1× Lysis/Binding/Washing Buffer supplemented with protease and kinase inhibitors. Pulldowns were performed using 20 μg GST–RalGDS-RBD for Rap1-GTP and 40 μg GST–Raf-1-RBD for Ras-GTP, incubated with 500 μg of cell lysate and 100 μL glutathione resin for 1 h on ice. Samples were washed twice with Lysis/Binding/Washing Buffer and eluted by boiling in 1× sample buffer. Pulldown samples and total cell lysates (input) were separated on 4–12% Bis-Tris polyacrylamide gels and transferred to PVDF membranes. Membranes were probed with anti-pan Ras or anti-Rap1 antibodies.

ICAM-1 and MAdCAM-1 binding assays

Flow cytometry–based ICAM-1 binding was performed with minor modifications to a previously published method.¹⁶ Briefly, purified B cells were immunostained with CD19, CD23, CD24, CD21 to allow the B cell subsets to be distinguished. To prepare sICAM-1, recombinant mouse ICAM-1–Fc (50 μg/mL, R&D Systems) was mixed with Alexa Fluor 647 AffiniPure F(ab’)₂ Fragment Goat Anti-Human IgG, Fcγ fragment specific (1:6.25 dilution, Jackson ImmunoResearch) in PBS and incubated on ice for at least 30 min. For on-ice binding assays, 2 × 10⁶ B cells per sample were washed and resuspended in 46 μL ice-cold binding buffer (PBS supplemented with 0.1% fatty-acid–free BSA and either 0.9 mM CaCl₂/0.5 mM MgCl₂ or 1 mM MnCl₂). Ice-cold sICAM-1 (7.9 μL) was then added to each sample. Two parallel conditions were prepared: (i) cells + sICAM-1 + buffer, and (ii) cells + sICAM-1 pre-incubated with CXCL13 (final concentration 1 μg/mL). The 4°C incubations were performed on ice. For 37°C binding assays, identical sample mixes were incubated at 37°C in a Thermomixer for the indicated times. For inside-out signaling induction, cells were pretreated with 50 ng/mL phorbol 12-myristate 13-acetate (PMA) for 30 min at 37°C before performing the sICAM-1 binding assay. At each designated time point, 10 μL of the reaction (≈1.5 × 10⁵ cells) was removed and immediately fixed in 500 μL 4% paraformaldehyde (PFA) for 5 min at room temperature. Samples were washed once in FACS buffer and analyzed by flow cytometry. Flow cytometric data were analyzed in FlowJo. ICAM-1⁺ and ICAM-1⁻ populations were gated based on the 0-min time point. The same protocol was used to assess sICAM-1 and sMAdCAM-1 binding to the RPMI 8866 cells.

Serum antibody measurements

Mouse serum immunoglobulin levels were measured using the LEGENDplex Mouse Immunoglobulin Isotyping Panel (6-plex) for IgG1, IgG2a, IgG2b, IgG3, IgA, and IgM, and the LEGENDplex Mouse IgE Assay for IgE, which was analyzed separately. Blood samples were collected and allowed to clot for at least 30 min at room temperature, centrifuged at 1,000 × g for 20 min, and serum aliquots were stored at −20°C until use. Beads, detection antibodies, and streptavidin-PE were added to each sample and standard according to the manufacturer’s instructions, followed by incubation on a shaker (~800 rpm) at room temperature in the dark. Samples were washed and analyzed by flow cytometry, and data were processed using LEGENDplex Data Analysis Software.

F-actin polymerization

Actin polymerization was evaluated on purified splenic B cells as described.³⁸ Purified B cells were immunostained with CD19, CD23, CD24, CD21 to allow the B cell subsets to be distinguished. The cells were rested at 3 × 10⁶ cells/ml in Opti-MEM medium for 30 min prior to the F-actin Polymerization assay. The cells were stimulated with CXCL13 (1 μg/ml) for various time points. At each indicated time point (0s–300s), a 100-μL aliquot was taken from the cell suspension, fixed in 4% paraformaldehyde for 10 min on ice, and then permeabilized and stained with Intracellular Staining Permeabilization wash buffer protocol using the Alexa Fluor 647 phalloidin. The F-actin polymerization was measured using flow cytometry by gating on the B220⁺ cells. F-actin polymerization percentage was calculated as follows: 100 × [geometric mean fluorescence intensity (MFI) (each time point, incubation with CXCL13) – geometric MFI (0 time point)]/geometric MFI (0 time point).

Intracellular calcium measurements

Splenic B220⁺ cells were seeded at 3 × 10⁵ cells per 100-μL loading medium (RPMI 1640, 10% charcoal-stripped FCS) into poly-D-lysine–coated 96-well black wall, clear-bottom microtiter plates. An equal volume of assay loading buffer (FLIPR Calcium 4 Assay Kit) in HBSS supplemented with 20 mM HEPES and 2 mM probenecid was added. Cells were incubated for 1 h at 37°C before adding chemokine, anti-mouse IgM antibody (aIgM), 7α,25-dihydroxy Cholesterol (7a, 25-OHC), or S1P, and then the calcium flux peak was measured using a FlexStation 3. The data were analyzed with SoftMax Pro 5.2. Data are shown as fluorescent counts, and the y axis is labeled as iLm1.

Immunohistochemistry

Immunohistochemistry was performed using a modified method of a previously published protocol.³⁹ Briefly, freshly isolated LNs, thymuses, PPs, or spleens were fixed in newly prepared 4% PFA overnight at 4°C on an agitation stage. Lymphoid organs were embedded in 4% low melting point agarose in PBS and sectioned with a vibratome at a thickness of 30 μm. Thick sections were blocked in PBS containing 10% fetal calf serum, anti-Fcγ receptor (1 mg/mL), and 0.1% Triton X-100 for 30 min at room temperature. Sections were stained overnight at 4°C on an agitation stage with the indicated antibodies, which included anti-peripheral node addressin (PNAd) (MECA-79), anti–lymphatic vessel endothelial hyaluronan receptor 1 (LYVE-1) (clone 223322), B220 (RA3-6B2), CD21/35 (7E9), Ki67 (SilA15), CD45.2 (104), CD138 (281-2), IgA (mA-6E1), Ep-CAM (G8.8), CD3 (17A2), and E-cadherin (DECMA-1). The lectin UEA-I (Ulex europaeus agglutinin) was used to stain thymuses. Stained thick sections were analyzed using a Leica SP8 confocal microscope equipped with an HC PL APO CS2 40× (NA, 1.30) oil objective and images were processed with Leica LAS AF software and Imaris software 10.0.1 64×.

Homing and egress assays

Purified splenic B cells from Rasa3^fl/fl and Rasa3 BKO mice were labeled with 1 μM eFluor 450 or 2.5 μM CMTMR for 15 min at 37°C, and equal numbers of viable cells (8–10 million) were injected i.v. into recipient mice. After 2 h, spleen, inguinal LNs, mesenteric LNs, and PPs were removed and gently dissociated into single-cell suspensions. Data acquisition was done on a FACSCelesta SORP flow cytometer and analyzed with FlowJo software. The egress assay used a similar initial approach, but 2 h after the cell transfer, mice were injected i.v. with CD62L Ab (5 mg/kg body weight). After 18 h, spleen, iLNs, mLNs, and PPs were removed and gently dissociated into single-cell suspensions. Peripheral blood was collected by retro-orbital eye bleeding. After removing RBCs with Tris–NH₄Cl, the cells were resuspended in PBS containing 1% BSA at 4°C. The ratio between the numbers of cells in the various compartments 2 h after transfer versus after CD62L Ab injection were calculated. In some experiments FTY720 (2 mg/kg body weight) was injected into the peritoneum of recipient mice 2 h prior to B cell adoptive transfer. PPs were collected 8 h later.

Intravital two-photon laser-scanning microscopy

Inguinal LNs were prepared for intravital microscopy as described previously.²⁹ The microanatomy of the organs was delineated by tail vein injection of labeled antibodies before imaging. Cell populations were labeled for 10 min at 37°C with 2.5–5 μM CellTracker Red CMTMR or 2 μM of eFluor 450. Five to ten million labeled cells of each population in 200 μL of PBS were adoptively transferred by tail vein injection into recipient mice. After anesthesia, the skin and fatty tissue over the il LN were removed. The mouse was placed in a prewarmed coverglass chamber slide. The chamber slide was then placed into the temperature control chamber on the microscope. PPs were prepared for intravital imaging as follows. After initial anesthesia (Avertin 300 mg/kg, i.p.) a small loop of intestine with PPs was exposed and glued with n-butyl cyanoacrylate to a custom-made metal holder. After attachment, the mouse was placed over a pre-warmed cover glass window on universal mounting frame AK-Set. The exposed organs were kept moist with saline wetted gauze. The mounting frame was placed into the temperature control chamber on the microscope and maintained at 37.0 ± 0.5°C. Once stabilized onto imaging stage/insert the mice received isoflurane (2% for induction of anesthesia, and 1–1.5% for maintenance, vaporized in an 80:20 mixture of oxygen and air). The exposed organs were intravitally imaged from the capsule over a range of depths (0–220 μm). All two-photon imaging was performed with a Leica SP8 inverted five-channel confocal microscope equipped with 25× water-dipping objective, 0.95 numerical aperture (immersion medium used distilled water). Two-photon excitation was provided by a Mai Tai Ti:Sapphire laser with a 10-W pump, tuned wavelength ranges from 820 to 920 nm. Emitted fluorescence was collected using a four-channel non-descanned detector. Wavelength separation was through a dichroic mirror at 560 nm and then separated again through a dichroic mirror at 495 nm followed by 525/50-nm emission filter for Alexa Fluor 488; the eFluor 450, or second harmonic signal, was collected by 460/50-nm emission filter, a dichroic mirror at 650 nm, followed by 610/60-nm emission filter for CMTMR or Alexa Fluor 594, and the Evans blue. For four-dimensional analysis of cell behavior, stacks of various numbers of section (z step = 2, 3, 4, 5, or 6 μm) were acquired every 2.5, 5, or 30 s to provide an imaging volume of 20–120 μm in depth. Sequences of image stacks were transformed into volume-rendered four-dimensional videos using Imaris software v.9.0.1 x64 and the tracks analysis was used for semiautomated tracking of cell motility in three dimensions by using the following parameters: autoregressive motion algorithm, estimated diameter 10 μm, background subtraction true, maximum distance 20 μm, and maximum gap size 3. Tracks acquired that could be tracked for at least 20% of total imaging duration were used for analysis. Some tracks were manually examined and verified. Calculations of the cell motility parameters (speed, track length, displacement, straightness, and speed variability) were performed using the Imaris software 10.0.1 64×. Statistical analysis was performed using Prism software. Annotations on videos and video editing were performed using Adobe Premiere Pro 2050.

HeLa cell transfection and imaging fluorescent proteins

HeLa cells were plated on 8-well micro-Slides and transiently transfected with X-tremeGENE HP DNA transfection reagent to express G_αi2-YFP, G_αi2Q205L-YFP, RASA3-mScarlet, and other fluorescently tagged proteins as indicated. After 40 h, live-cell confocal or TIRF imaging was performed. Confocal images were acquired with a Leica SP8 or Leica Stellaris confocal microscope equipped with appropriate lasers, an HC PL APO CS2 63X/1.4NA oil-immersion objective, HyD hybrid detectors, Adaptive focus control, and an environmental chamber. Images were captured with a pixel size of 0.12 × 0.12 × 0.85μm and a frame rate of approximately 30 s. Processed colocalization signals were generated using the Coloc function in Imaris software. TIRF images were acquired using a Leica DMI8 equipped with an HCX PL APO 100X/1.46NA TIRF objective, 488nm and 561nm lasers, appropriate filter cubes, Adaptive focus control, and an Andor iXon Ultra EMCCD camera. Images were captured with a pixel size of 0.13μm x 0.13μm. TIRF images were processed on the fly by using the THUNDER Computational Clearing routine to reduce diffuse background fluorescence and improve feature visibility.

Imaging RPMI 8866 RASA3 transfectants

To assess the localization of RASA3 and Rap1-GTP in RPMI 8866 human B cells, we used the RASA3-stable RPMI 8866 cells and localized Rap1-GTP by expression of the RalGDS-GFP fusion protein. The RalGDS-GFP plasmid was kindly provided by Dr. Tatsuo Kinashi, Kansai Medical University, Japan. Two days post-nucleofection, cells were plated in 8-well chambers pre-coated with ICAM-1/CXCL12 (200 μL of 5 μg/mL ICAM-1 and 0.2 μg/mL CXCL12 in dH₂O per well, incubated at 4°C overnight) for 5 min and then directly imaged at 10-s intervals using a Leica SP8 inverted five-channel confocal microscope. Imaging was performed with a 63× oil-immersion objective (1.40 NA). Live imaging was conducted in an incubation chamber, maintaining an air temperature of 37.0 ± 0.5°C with 5% CO2. Images were acquired with Leica LAS AF software and processed using Imaris software v.10.0.1 64×. Same method used to image the RPMI 8866 RASA3 stable transfectants transiently expressing G_αi2Q205L-YFP.

Confocal and fluorescence lifetime imaging microscopy

Confocal images were acquired on a DMI8-SP8-FALCON confocal microscope. Images were acquired with fixed acquisition settings and co-localizations were done using Imaris Image analysis software. For fluorescence lifetime imaging microscopy (FLIM), corresponding samples were imaged on a Leica SP8WLL falcon inverted confocal microscope with a 63× oil immersion objective. Donor Lifetime measurements of GFP were measured by exciting at 488nm wavelength by using tunable White Light Laser (WLL) system tuned at 80 MHz frequency settings. Acquired fluorescent transients in the range of 1000 or more photons per pixel, were analyzed using LASX single molecule detection analysis software (LASX-Single molecule detection software) to determine fluoresce lifetime measurements. Mean fluorescent lifetime and FRET efficiency calculations were observed and measured as described previously.⁴⁰

QUANTIFICATION AND STATISTICAL ANALYSIS

All experiments were performed using at least six mice per experimental group. Data are presented as mean ± SEM. Statistical analyses for comparisons between two groups were performed using unpaired two-tailed t-tests, while analyses involving multiple groups or repeated measurements were performed using ANOVA, with experiment-to-experiment variability accounted for. Individual mice were treated as biological replicates, and technical triplicates were used to ensure measurement accuracy. A p-value of <0.05 was considered statistically significant, unless otherwise stated (*p < 0.05; **p < 0.005; ***p < 0.0005). In vivo results represent samples from 2 to 10 mice per experimental group. Results represent mean values of at least triplicate samples. SEM and p values were calculated with a t test or ANOVA using GraphPad Prism. Representative images were placed in figures. Primary image data which was analyzed and calculated using Leica LAS AF software or Imaris software 10.0.1 64×. Final data was acquired and processed with Microsoft Excel software.

Supplementary Material

Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2026.117243.

Highlights.

• Blood and tissue B cells actively suppress integrin activation

• Loss of Rasa3 in murine B cells severely disrupts B lymphocyte trafficking

• Rasa3 couples G_αi signaling to Rap1-GTP levels

• Rasa3 amplifies G_αi signaling

Data and code availability

• Data reported in this paper will be shared by the lead contact upon request.

• This paper does not include or rely on any original code.

• All processed video files are provided in the supplemental information.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.