Abstract
Resistance to inhibitors of cholinesterase 8A (Ric-8A) is a highly evolutionarily conserved cytosolic protein initially identified in Caenorhabditis elegans, where it was assigned a regulatory role in asymmetric cell divisions. It functions as a guanine nucleotide exchange factor for Gαi, Gαq, and Gα12/13 and as a molecular chaperone required for the initial association of nascent Gα subunits with cellular membranes in embryonic stem cell lines. To test its role in hematopoiesis and B lymphocytes specifically, we generated ric8fl/flvav1-cre and ric8fl/flmb1-cre mice. The major hematopoietic cell lineages developed in the ric8fl/flvav1-cre mice, notwithstanding severe reduction in Gαi2/3, Gαq, and Gα13 proteins. B lymphocyte–specific loss of Ric-8A did not compromise bone marrow B lymphopoiesis, but splenic marginal zone B cell development failed, and B cells underpopulated lymphoid organs. The ric8fl/flmb1-cre B cells exhibited poor responses to chemokines, abnormal trafficking, improper in situ positioning, and loss of polarity components during B cell differentiation. The ric8fl/flmb1-cre mice had a severely disrupted lymphoid architecture and poor primary and secondary Ab responses. In B lymphocytes, Ric-8A is essential for normal Gα protein levels and is required for B cell differentiation, trafficking, and Ab responses.
Introduction
In canonical G-protein signaling, an agonist binds a G-protein–coupled receptor, which adopts a conformation that triggers the Gα subunit of a heterotrimeric G-protein to exchange GDP for GTP, resulting in the functional dissociation of Gα from its associated Gβγ. This leads to the activation of downstream intracellular effector enzymes that mediate cellular responses. For example, most chemokine receptors signal by triggering Gαi nucleotide exchange, resulting in the activation of small GTPases that communicate with actin regulatory proteins to drive the cell motility needed for chemotaxis (1). In noncanonical G-protein signaling, the guanine nucleotide exchange factor (GEF) activity exerted by the G-protein–coupled receptor is replaced by the action of intracellular GEFs. One such intracellular GEF is resistance to inhibitors of cholinesterase 8A (Ric-8A). Ric-8A acts on Gαi, Gαq, and Gα12/13, whereas a related protein, Ric-8B, acts on Gαs (2).
A highly evolutionarily conserved cytosolic protein Ric-8 was initially identified in Caenorhabditis elegans, in which its functions include a regulatory role in asymmetric cell divisions (3–5). In human cells, Ric-8A recruits to the cell cortex a signaling complex that helps orient the mitotic spindle in response to spatial clues (6). In noncanonical signaling pathways, Gα subunits are often paired with proteins containing one or more conserved Gαi/o–Loco interaction (GoLoco) motifs, also known as G-protein regulatory motifs, which act as a guanine nucleotide dissociation inhibitor, much as Gβγ does in the canonical pathway (7). In Drosophila, a GoLoco protein Pins (Partner of Inscuteable) forms an apical protein complex essential for neuroblast asymmetric cell division (8). In humans, mutations in the GoLoco protein Leu-Gly-Asn–enriched protein (LGN) cause brain malformations and hearing loss in the Chudley–McCullough syndrome (9). Also in human cells, the noncanonical G-protein signaling proteins Ric-8A and LGN localize at the midbody during cytokinesis along with Gαi and RGS14 (regulator of G-protein signaling 14), a GoLoco motif–containing protein (10, 11). Interference with Gαi expression or with Gαi nucleotide exchange prolongs cytokinesis, the final step of the cell cycle (12). Gene targeting the GoLoco protein Ags3 in mice leads to phenotypes in the kidney, as well as the nervous and immune systems (13–15). Arguing for a role of activator of G-protein signaling 3 (AGS3) in Gαi regulation in immune cells, ags3−/− dendritic cells and lymphocytes exhibit suboptimal responses to chemokines in chemotaxis, calcium mobilization, and effector protein activation assays (16). In addition, during neutrophil chemotaxis GDP-bound Gαi accumulates at the leading edge, where it recruits the adaptor molecule Inscuteable (Insc), LGN/AGS3, and the Par3–atypical protein kinase C (aPKC) polarity complex. Insc−/− neutrophils poorly stabilize leading edge pseudopods, and that stability can be restored by the addition of wild-type (WT) Insc protein, but not by a mutant protein that does not bind LGN/AGS3 (17). These studies and others implicate Ric-8A and noncanonical G-protein signaling in an array of biologic processes likely to affect immune cells (18).
Ablation of ric8 in mice results in early embryonic lethality, as embryos died at embryonic days 6.5–8.5. The mice die shortly after initiation of gastrulation with a disorganized epiblast (19). Derived ric8−/− embryonic cell lines exhibited pleiotropic G-protein signaling defects and ∼85% loss of Gαi1/2, Gαq, and Gα13 proteins. An accelerated rate of protein degradation accounted for the reduced levels of proteins. These data indicated that Ric-8A had an additional function: that of a molecular chaperone helping to target newly translated Gαi, Gαq, and Gα12/13 proteins to cellular membranes (20, 21). Conditional gene targeting using a floxed ric8 allele and an hGFAP-cre that targets Ric-8A expression in neural progenitors and astroglia resulted in mice with a disorganized Bergmann glial scaffolding, defective granule cell migration, and disrupted Purkinje cell positioning (22). A synapsin I promoter–driven Cre ablated Ric-8A function in most differentiated neuron populations and resulted in early postnatal death owing to a severe neuromuscular phenotype (23). However, whether the phenotypes that arose in these conditionally ric8-targeted mice resulted from Gα protein deficiency or were due to a loss of Ric-8A function in noncanonical G-protein signaling was unexplored in these studies.
Despite increasing evidence that asymmetric localization of proteins during lymphocyte cell division contributes to differential cell fates and the known role of Gα proteins and their partners in model organism asymmetric cell divisions, relatively little attention has been paid to whether they participate in asymmetric cell divisions in lymphocytes. One study did note that interference with the Pins (LGN)/G-protein module reduced the number of dividing T cells with a mitotic axis, compatible with asymmetric cell division (24). We sought to determine whether Ric-8A had chaperone-like activity for Gα subunits in hematopoietic cells, to investigate the consequences of a specific loss of Ric-8A in B cells, and to determine whether the loss of Ric-8A affected B lymphocyte symmetric and asymmetric cell divisions. We found that Ric-8A has chaperone-like activity for Gαi2, Gαi3, and Gαq, whereas steady state levels of Gαs and Gα12 were unaffected in spleen cells and bone marrow–derived macrophages. A loss of Ric-8A in B cells led to a severe B cell immunodeficiency likely because of the Gαi proteins. In response to mitotic signals, the Ric-8A–deficient and WT B cells divided symmetrically with an equal frequency, although on occasion the final abscission step was delayed in the absence of Ric-8A. In contrast, activated ric8fl/flmb1-cre B cells and germinal center (GC) B cells from immunized ric8fl/flmb1-cre mice underwent fewer asymmetric cell divisions when compared with control cells. The implications of our results are discussed.
Materials and Methods
Animals
C57BL/6 and B6.SJL-Ptprca Pepcb/BoyJ mice were obtained from The Jackson Laboratory. The previously characterized Ric-8Afl/fl mice (22) on a mixed background were backcrossed 10 times on to C57BL/6. The C57/BL6 mb1-cre mice were kindly provided by Dr. Michael Reth (25). The C57/BL6 vav1-cre mice were obtained from The Jackson Laboratory and previously characterized (26). For bone marrow reconstitution, 7-wk-old B6.SJL-Ptprca Pepcb/BoyJ (CD45.1) mice were irradiated twice with 550 rads, for a total of 1100 rads, and received bone marrow from C57BL/6 CD45.2 control or mutant mice. The engraftment was monitored by sampling the blood 28 d later. The mice were used 6–8 wk after reconstitution. All mice were used in this study were 6–14 wk 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.
Cells
Splenic B cells were isolated by negative depletion, using biotinylated Abs to CD4, CD8, Gr-1 (Ly-6C and Ly 6G), and CD11c and Dynabeads M-280 Streptavidin (Invitrogen). The B cell purity was >95%. When needed, B cells were cultured in RPMI 1640 containing 10% FCS (Life Technologies), 2 mM l-glutamine, antibiotics (100 IU/ml penicillin and 100 μg/ml streptomycin), 1 mM sodium pyruvate, and 50 μM 2-ME. When very high purity B cells were needed, they were isolated by cell sorting following immunostaining for CD19 and B220.
Flow cytometry and Abs
Single cells were resuspended in PBS and 2% FBS, and stained with fluorochrome-conjugated or biotinylated Abs against B220 (RA3-6B2), IgD (11-26c-2a), IgM (R6-60.2), CD24 (M1/69), CD3 (145-2C11), CD4 (GK1.5), CD5 (53-7.3), CD8 (53-6.7), CD11c (HL3), CD11b (M1/70), CD138 (281-2), CD19 (1D3), CD38 (90), IgG1 (×56), CD93 (AA4.1), BP-1 (6C3), GL-7 (GL-7, Ly-77), CD95 (Jo2), CD21 (4E3), CD23 (B3B4), CD43 (S7), CD184 (CXCR4, 2B11), CXCR5 (2G8), CCR7 (4B12), CD11a (M17/4), PD-1 (CD279, RMP1-30), CD45.1 (A20), or CD45.2 (104) (all from eBioscience, BioLegend, or BD Pharmingen). Biotin-labeled Abs were visualized with fluorochrome-conjugated streptavidin (eBioscience). The LIVE/DEAD Fixable Aqua Dead Cell Stain Kit (Molecular Probes) was used in all experiments to exclude dead cells. Data acquisition was done on a FACSCanto II (BD) flow cytometer and analyzed with FlowJo software (TreeStar). The following flow cytometry gates were done to identify bone marrow B cell subsets: fraction (Fr.) A: B220+CD43+CD24−BP1−; Fr. B: B220+CD43+CD24+BP1−; Fr. C: B220+CD43+CD24+BP1+; Fr. D: B220+CD43−IgM−IgD−; Fr. E: B220+CD43−IgM+IgD−; and Fr. F: B220+CD43−IgM+IgD+. Gating for splenic subsets was as follows: T0: B220+CD93+IgM+IgD−CD23−; transitional type 1 (T1): B220+CD93+IgM+IgD+CD23−CD21−; follicular (FO): B220+CD23+CD21−CD24+; transitional type 2 (T2)–FO: B220+CD23+CD21−CD24++; marginal zone precursor (MZP): B220+CD23+CD21+CD24++; marginal zone (MZ): B220+CD23−CD21+CD24+; GC: B220+CD19+CD38−CD95+GL-7+; and FO helper T cell: CD4+CXCR5+PD-1+. The following gates were used to identify thymocyte subsets (B220−CD11c−Gr1−): double positive: CD4+CD8+; double negative (DN)1: CD4−CD8−CD44+CD25−; DN2: CD4−CD8−CD44+CD25+; DN3: CD4−CD8−CD44−CD25+; DN4: CD4−CD8−CD44−CD25−; mature CD4 single positive (SP): CD4+CD8−CD69−CD62L+CD24−; immature CD4SP: CD4+CD8−CD69+CD62L−CD24+; mature CD8SP: CD4−CD8+CD69−CD62L+CD24−; immature CD8SP: CD4−CD8+CD69+CD62L−CD24+.
Cell proliferation
The cell proliferation studies were performed using CFSE (Molecular Probes) in a standard dye dilution assay. Purified B cells were stimulated for 96 h with various combinations of the following reagents: 1 μg/ml CD40 (HM40-3), 1 μg/ml LPS (055:B5, Sigma-Aldrich), recombinant mouse IL-4 (10 ng/ml), or 10 μg/ml AffiniPure F(ab′)2 fragment goat anti-mouse IgM (Jackson ImmunoResearch Laboratories). Data acquisition was done on a FACSCanto II flow cytometer. The proliferation index is the average number of cell divisions that a cell in the original population undergoes; the division index is the average number of cell divisions of the responding cells; and the percent cell division is defined as the proliferation index divided by the division index, and multiplying the results by 100, assuming no cell death, was done using FlowJo software.
Chemotaxis assays
Chemotaxis assays were performed using a Transwell chamber (Costar) as previously described (27). Splenic B cells were immunostained for B cell subsets with fluorochrome-conjugated Abs against B220, CD21, CD23, CD24, CD45.1, and CD45.2, washed twice, resuspended in complete RPMI 1640 medium, and added in a volume of 100 μl to the upper wells of a 24-well Transwell plate with a 5-μm insert. Lower wells contained various doses of chemokines in 600 μl complete RPMI 1640 medium. The numbers of cells that migrated to the lower well after 2 h incubation were counted using a MACSQuant flow cytometer (Miltenyi Biotec). The percent migration was calculated by the numbers of cells of a given subset that migrated into the bottom chamber divided by the total number of cells of that subset in the starting cell suspension, and multiplying the results by 100. CXCL13, CCL19, and CXCL12 were purchased (R&D Systems). Fatty acid–free BSA was purchased (Sigma-Aldrich).
Intracellular calcium measurements
Cells were seeded at 105 cells per 100 μl loading medium (RPMI 1640, 10% FBS) into poly-d-lysine–coated 96-well black wall, clear-bottom microtiter plates (Nalgene Nunc). An equal volume of assay loading buffer (FLIPR Calcium 4 assay kit; Molecular Devices) 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 or unconjugated AffiniPure F(ab′)2 fragment goat anti-mouse IgM (Jackson ImmunoResearch Laboratories), and then the calcium flux peak was measured using a FlexStation 3 (Molecular Devices). The data were analyzed with SOFT max Pro 5.2 (Molecular Devices). Data are shown as fluorescent counts, and the y-axis is labeled as Lm1.
ImageStream: labeling, acquisition, and analysis
The splenic B cells were stained with B220 PE-Cy7 conjugated for 10 min, then fixed in 2% PFA. The PKCζ distribution was visualized by indirect labeling when Ab was diluted (Santa Cruz Biotechnology; 1:100) in permeabilization wash buffer consisting of 3% BSA and 0.1% Triton X-100 in PBS. Samples were incubated for 2 h at room temperature. Primary Ab was removed, and 1:200 dilution of secondary Alexa 488–conjugated Ab (Jackson ImmunoResearch Laboratories) was added and incubated at room temperature in the dark for 1 h. Secondary Ab was removed, and cells were resuspended in 100 μL PBS. Just prior to running on the ImageStream, all samples had Draq5 (Cell Signaling) added (20 nM final concentration) to visualize the nucleus. At least 100,000 events were collected for all samples on an ImageStream MarkII, using 488 nm, 405 nm, and 642 nm laser excitations. Cell populations were hierarchically gated for single cells that were in focus and were positive for both PKCζ and B220. Then, based on a Draq5 intensity histogram, we gated on cells that were in the G2/M phase. An additional gating for GC was applied for those experiments in which we focused on B220+Fas+GL7+ cells. Following data acquisition, the spatial relationship between B220 and Draq5 was measured using a feature (δ centroid intensity weighted) calculated in the IDEAS software package that eliminates nonspecific asymmetric distributions. The same feature was applied between PKCζ and Draq5, and asymmetry was defined when the δ centroid value was greater than half the median of the B cell radius.
Immunohistochemistry
Immunohistochemistry was performed using a modified method of a previously published protocol (28). Briefly, freshly isolated lymph nodes and spleens were fixed in freshly prepared 4% paraformaldehyde (Electron Microscopy Science) overnight at 4°C on the agitation stage. Lymph nodes and spleens were embedded in 4% low melting agarose (Invitrogen) in PBS and sectioned with a vibratome (Leica VT-1000 S) at a 30-μm thickness. Thick sections were blocked in PBS containing 10% FCS, 1 mg/ml anti-Fcγ receptor (BD Biosciences), and 0.1% Triton X-100 (Sigma-Aldrich) for 30 min at room temperature. Sections were stained overnight at 4°C on the agitation stage with the following Abs: anti-B220, anti-CD3e, anti-CD4, anti-Ki67 (all from eBioscience), anti-CD169 (R&D Systems), and anti-CD21/35 (BioLegend). Stained thick sections were microscopically analyzed using a Leica SP5 confocal microscope (Leica Microsystem), and images were processed with Leica LAS AF software (Leica Microsystem) and Imaris v.7.6.1 (Bitplane).
Live cell time-lapse confocal microscopy
Splenic B cells were prepared from ric8wt/wt Lifeact and ric8fl/flmb1-cre Lifeact mice. The isolated B cells were cultured at an initial concentration of 1 × 106 cells per milliliter in complete lymphocyte medium in the presence of 2 μg/ml LPS (from E. coli, serotype R515 [Re], TLR grade; ENZO Life Sciences), for 48 h. For live cell imaging, B cells were allowed to adhere to ICAM-1 + VCAM-1 (Recombinant Mouse ICAM-1/CD54 Fc Chimera, CF; Recombinant Mouse VCAM-1/CD106 Fc Chimera; R&D Systems) on coated glass-bottom dishes (no. 1.5 coverglass; MatTek). Confocal imaging was performed using a Leica SP8 inverted five-channel confocal microscope (Leica Microsystems) equipped with incubation chamber (CO2, 37°C) for live cell imaging (Pecon). The argon laser was tuned to 488-nm excitation wavelength, using laser power between 0.2 and 1%. Z stacks were acquired every 10–12 s over a time period of 1 h, and single-plane images were used to generate video. Images were processed using Imaris (Bitplane) software.
Intravital and spleen section two-photon laser scanning microscopy
Inguinal lymph nodes were prepared for intravital microscopy as described (29). Cell populations were labeled for 15 min at 37°C with 1 μM green cell tracker CMFDA, 2.5- to 5-μM red cell tracker CMTMR (Molecular Probes). A total of 5–10 million labeled cells of each population in 200 ml PBS were adoptively transferred by tail vein injection into 6- to 10-wk-old recipient mice. After anesthetizing the mice by i.p. injection of Avertin (tribromoethanol, 300 mg/kg; Sigma-Aldrich), the skin and fatty tissue over inguinal lymph nodes were removed. The mouse was placed in a prewarmed coverglass chamber slide (Nalgene, Nunc). The chamber slide was then placed into the temperature control chamber on the Leica SP5 microscope. The temperature of the air was monitored and maintained at 37.0 ± 0.5°C. The inguinal lymph node was intravitally imaged from the capsule over a range of depths (10–220 μm). All two-photon imaging was performed with a Leica SP5 inverted five-channel confocal microscope (Leica Microsystems) equipped with 0.95 numerical aperture (immersion medium used distilled water). Two-photon excitation was provided by a Mai Tai Ti:Sapphire laser (Spectra Physics) with a 10-W pump, tuned to 810 or 910 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 a 525/50 emission filter for CMFDA (Molecular Probes); a dichroic mirror at 650 nm, followed by 610/60-nm emission filter for CMTMR; and the Evans blue signal was collected by 680/50-nm emission filter. Sequences of image stacks were transformed into volume-rendered four-dimensional videos using Imaris software, and tracking analysis was transformed using autoregressive motion algorithm of Imaris software v.7.6.1 (Bitplane). Polarity measurements were performed with Imaris software.
Immunizations and ELISA
WT and mutant 6- to 8 wk-old mice were immunized with SRBCs, 4-hydroxy-3-nitrophenylacetyl (NP)35–keyhole limpet hemocyanin (KLH), or NP40-Ficoll. For the SRBC immunizations, 200 μl 10% solution of SRBCs (Lonza Walkersville) was given by i.p. injection. NP35-KLH (Biosearch Technologies) was mixed with Imject Alum (Thermo Scientific) and introduced into mice (100 μg) via i.p. injection. Mice were boosted with the same dose of Ag at the indicated days, along with Alum. Mice were also immunized with 25 μg NP40-Ficoll (Biosearch Technologies) via i.p. injection. Serum NP-specific Ig levels in these mice were analyzed by ELISA. Briefly, 96-well ELISA plates (Nunc) were coated with NP30-BSA (Biosearch Technologies) overnight at 4°C, washed, and blocked with 1% BSA fraction V (Sigma-Aldrich); serum titers were then added to the plates and incubated 4 h at 4°C. After washing, alkaline phosphatase–labeled goat anti-mouse Ig isotype–specific Abs were added for 2 h at room temperature (SouthernBiotech). After washing, pNPP One Component Substrate (SouthernBiotech) was used to detect the amount of secondary Ab bound.
Western blotting
Purified splenic B cells were lysed on ice for 10 min in 50 mM HEPES (pH 7.4), 250 mM NaCl, 2 mM EDTA, 0.2% Nonidet P-40, containing cOmplete Protease Inhibitors (Roche) and supplemented with 100 μM orthovanadate sodium, 10 μM NaF, and 10 μM PMSF. After centrifugation at 13000 × g for 10 min, the supernatant was collected and protein concentration determined (ODλ = 280 nm measurement, Nanodrop; ThermoScientific). A total of 50 μg lysates was resolved on 4–20% Tris-Glycine gels. After transfer onto nitrocellulose, blots were probed with anti-actin–HRP-conjugated (Sigma-Aldrich; 1:20,000), anti-Gαi2 mouse monoclonal (Santa Cruz Biotechnology; 1:500), anti-Gαi3 rabbit polyclonal (Santa Cruz Biotechnology; 1:500), anti-Gα12 rabbit polyclonal (Santa Cruz Biotechnology; S20, 1:50), anti-Gαs rabbit polyclonal (EMD Millipore; 1:100), anti-Gαq rabbit polyclonal (EMD Millipore; 1:100), anti-Gα13 mouse monoclonal (New East Biosciences), anti–Ric-8A rabbit polyclonal (kind gift from Dr. G. Tall, University of Rochester), anti-LGN rabbit polyclonal (kind gift from Dr. J. Blumer, University of South Carolina) and were revealed using HRP-conjugated TrueBlot (eBioscience) secondary Ab (1:10,000). Immunoblots were revealed by luminescence (ECL; Amersham Bioscience).
Statistics
In vivo results represent samples from three to nine mice per experimental group. Results represent mean values of at least triplicate samples. SEM and p values were calculated with the Student t test or two-way ANOVA using GraphPad Prism (GraphPad software). The p values were as follows: *p < 0.05, **p < 0.005, ***p < 0.0005.
Results
Disrupting Ric-8A expression in mouse hematopoietic cells results in a loss of Gαi2, Gαi3, and Gαq; anemia; leukocytosis; and a loss of B lymphocytes
To assess the role of Ric-8A in mouse hematopoietic cells, we crossed ric8fl/fl mice to vav1-cre transgenic mice to generate ric8fl/flvav1-cre mice (both strains on a C57/BL6 background). Ric-8A and Vav1 are expressed in hematopoietic stem cells, progenitors, and cells of the hematopoietic lineage (18, 26, 30). The ric8fl/flvav1-cre mice were born with the expected mendelian frequency. However, we noted that they had an increased perinatal mortality, diminished vigor, and reduced longevity, with many mice dying within 4 mo of birth; yet their body weights at 6 wk were similar to that of controls (C. Boularan, unpublished observations). Pathological examination did not reveal any consistent cause of their early demise. The loss of a single allele of ric8 in hematopoietic cells was without evident consequence, as the ric8fl/wtvav1-cre mice thrived similarly to WT mice. We verified the loss of Ric-8A in the conditionally deleted hematopoietic cells by immunoblotting cell lysates prepared from splenocytes and bone marrow–derived macrophages for Ric-8A expression (Fig. 1A). Next, we checked the expression levels of Gα proteins and that of LGN. We found that the loss of Ric-8A had led to a severe reduction in Gαi2, Gαi3, Gα13, and Gαq, whereas the levels of Gα12, Gαs, and LGN were similar to those of WT mice (Fig. 1A). Assessment of blood obtained from the ric8fl/flvav1-cre mice revealed modest anemia, an increase in white cell numbers, and normal platelet numbers. A lymphocytosis accounted for the noted leukocytosis (Fig. 1B, 1C). Assessment of B and T cells in the spleen and thymus did not reveal any striking abnormalities, whereas lymph nodes, Peyer’s patches, and bone marrow all had a reduction in B cell numbers (Fig. 1D, 1E). An analysis of B cell development in the bone marrow and spleen revealed a significant loss of T2 transitional and MZ B cells in the spleen (Fig. 1F). These results indicate that Ric-8A is needed for normal levels of Gαi2, Gαi3, Gα13, and Gαq in hematopoietic cells, but not for Gα12. Its loss caused modest anemia, decreased longevity, and loss of B lymphocytes, particularly from lymph nodes and Peyer’s patches, and interfered with B cell development in the spleen. Surprisingly, T cell development proceeded relatively normally, and secondary lymphoid organs were populated with normal numbers of T cells (Fig. 1D, 1E, 1G). We did observe an increase in the number of mature SP CD4 and CD8 T cells, a result consistent with a mild thymus egress defect likely owing to the reductions in Gαi2 and Gαi3.
Targeting ric8 in hematopoietic cells causes the loss of Gα proteins in spleen cells, anemia, impaired MZ B cell development, increased mature CD4+ T thymocytes, and reduced peripheral B cell numbers. (A) Immunoblot analysis of cell lysates prepared from spleen cells and bone marrow–derived macrophages (BMDMs) from ric8fl/flvav1-cre mice, ric8fl/+vav1-cre mice, and ric8+/+vav1-cre mice for the indicated proteins. Similar results in three separate experiments. (B and C) Complete blood counts using blood obtained from vav1-cre (n = 7) and ric8fl/flvav1-cre (n = 6) mice. (D and E) Flow cytometry characterization of cells prepared from the lymphoid organs of the mice used in parts (B) and (C). Single-cell suspensions were prepared from spleen, thymus, inguinal lymph node, Peyer’s patches, and bone marrow. The numbers of lymphocytes, CD4+, CD8+, and B220+ cells in each preparation are shown. (F) Flow cytometry analysis of bone marrow and splenic B cell development in vav1-cre and ric8fl/flvav1-cre mice. The bone marrow development is shown as the percentage of cells in Fr. A–F based on the B220+ cell gate. Spleen B cell development is shown as the percentage of cells in T1, T2, FO, and MZ B cell subsets. (G) Flow cytometry analysis of thymocytes prepared from in vav1-cre and ric8fl/flvav1-cre mice. The percentages of thymocytes in DN subsets 1–4, double positive (DP), CD4 SP, and CD8 SP are shown. The CD4 and CD8 SP cells are divided into immature and mature based on expression of CD69 and CD62L. *p < 0.05, **p < 0.005. Baso, basophils; Eosin, eosinophils; Lymph, lymphocytes; Mono, monocytes; PMN, neutrophils; RBC, RBC count; WBC, WBC count.
Targeting ric8 in hematopoietic cells causes the loss of Gα proteins in spleen cells, anemia, impaired MZ B cell development, increased mature CD4+ T thymocytes, and reduced peripheral B cell numbers. (A) Immunoblot analysis of cell lysates prepared from spleen cells and bone marrow–derived macrophages (BMDMs) from ric8fl/flvav1-cre mice, ric8fl/+vav1-cre mice, and ric8+/+vav1-cre mice for the indicated proteins. Similar results in three separate experiments. (B and C) Complete blood counts using blood obtained from vav1-cre (n = 7) and ric8fl/flvav1-cre (n = 6) mice. (D and E) Flow cytometry characterization of cells prepared from the lymphoid organs of the mice used in parts (B) and (C). Single-cell suspensions were prepared from spleen, thymus, inguinal lymph node, Peyer’s patches, and bone marrow. The numbers of lymphocytes, CD4+, CD8+, and B220+ cells in each preparation are shown. (F) Flow cytometry analysis of bone marrow and splenic B cell development in vav1-cre and ric8fl/flvav1-cre mice. The bone marrow development is shown as the percentage of cells in Fr. A–F based on the B220+ cell gate. Spleen B cell development is shown as the percentage of cells in T1, T2, FO, and MZ B cell subsets. (G) Flow cytometry analysis of thymocytes prepared from in vav1-cre and ric8fl/flvav1-cre mice. The percentages of thymocytes in DN subsets 1–4, double positive (DP), CD4 SP, and CD8 SP are shown. The CD4 and CD8 SP cells are divided into immature and mature based on expression of CD69 and CD62L. *p < 0.05, **p < 0.005. Baso, basophils; Eosin, eosinophils; Lymph, lymphocytes; Mono, monocytes; PMN, neutrophils; RBC, RBC count; WBC, WBC count.
Disrupting Ric-8A expression in mouse B cells leads to a loss of Gαi2, Gαi3, and Gαq; a reduction in B lymphocytes; and major decreases in GC B cells and early switched memory B cells
Because the initial assessment of the loss of Ric-8A in hematopoietic cells had suggested that B cells had been the most severely affected, and to remove the complicating issue of the loss of Ric-8A in multiple cell types, we crossed the ric8fl/fl mice to mb1-cre mice and generated ric8fl/fl mb1-cre mice. The mb1-cre mice have a single normal mb1 (CD79a) allele that is sufficient for normal B cell development and function, whereas the other allele expresses the Cre recombinase at the pro-B stage of B cell bone marrow development (25). Because both ric-8 and mb1 are located on chromosome 7, we had to screen the progeny for a cross-over event that resulted in a floxed ric-8 allele and an mb-1 Cre allele on the same chromosome. Having established that mouse line on a C57BL/6 background, we assessed Ric-8A and Gα protein expression in B cells from WT and mice that lacked B cell–specific expression of ric8. Similar to the vav1-cre mice, the loss of ric8 expression in B cells led to a marked decrease in their Gαi2, Gαi3, and Gαq protein levels, although Gαi3 was less affected (Fig. 2A). Attempts to discern whether the loss of Ric-8A affected Gα13 levels in B cells were stymied by our failure to detect Gα13 in WT B cell lysates. In the blood, the mice had more lymphocytes and normal numbers of other leukocytes, with the exception of a modest increase in the number of neutrophils (Fig. 2B). Like the ric8fl/flvav1-cre mice, the ric8fl/flmb1-cre mice had reduced numbers of B cells in lymph nodes, Peyer’s patches, and the spleen (Fig. 2C). B cell development in the bone marrow was largely intact, and in contrast to the ric8fl/flvav1-cre mice, no reduction of B cells in the bone marrow was noted (Fig. 2D, data not shown). Similar to the ric8fl/flvav1-cre mice, there was marked loss of MZPs and MZ B cells in the spleen (Fig. 2E). The Ric-8A expression levels did not differ appreciably between the splenic B cell subsets in the WT mice (Fig. 2E). Immunizing the ric8fl/flmb1-cre mice led to a very poor GC response in the spleen, lymph nodes, and Peyer’s patches; reduced plasma cell response in lymph nodes and in Peyer’s patches; and a very poor induction of switched memory cells at all of the tested sites (Fig. 2F). These results indicate that the loss of Ric-8A in B cells would likely lead to a severe impairment in humoral immunity.
B cell–specific loss of Ric-8A confirms its role as a chaperone for Gαi2/3, Gα13, and Gαq and reveals a B cell–intrinsic role for Ric-8A in B cell development and function. (A) Ric-8A, Gα proteins, and actin expressions were detected by immunoblot of lysates prepared from sorted B220+CD19+ splenic cells from ric8wt/wtmb1-cre and ric8fl/flmb1-cre mice. (B) Data shown are from complete blood counts from 14 ric8wt/wtmb1-cre and 12 ric8fl/flmb1-cre mice. (C) Flow cytometry characterization of lymphoid organs from control and ric8fl/flmb1-cre mice used in (B). Absolute numbers of cells are shown from the spleen, from mesenteric and inguinal lymph nodes (mLN and iLN, respectively), and from Peyer’s patches (PP). (D) Flow cytometry was used to characterize B cell development in the bone marrow of ric8wt/wtmb1-cre and ric8fl/flmb1-cre mice. Shown are the percentages of B220+ cells in bone marrow, Fr. A–F. (E) Flow cytometry was used to characterize B cell development in the spleen of ric8wt/wtmb1-cre and ric8fl/flmb1-cre mice. Shown are the percentages of B220+ cells in the T1, T2-FO, T2-MZP, FO, and MZ subsets. Below is an immunoblot of Ric-8A expression in the different subsets. (F) Flow cytometry was used to analyze cells isolated from the spleen, iLN, and PP 10 d following immunization of ric8wt/wtmb1-cre and ric8fl/flmb1-cre mice with SRBCs. Absolute numbers of B220+, CD4+, CD8+, GC, T FO helper cells (TFH), early plasma cell (B220+CD138+), and switched memory cells (B220+, CD38+/−, CD138−, IgG1+) are indicated. Results are from the analysis of four experimental and four control mice. Experiment was repeated three times with similar results. *p < 0.05, **p < 0.005.
B cell–specific loss of Ric-8A confirms its role as a chaperone for Gαi2/3, Gα13, and Gαq and reveals a B cell–intrinsic role for Ric-8A in B cell development and function. (A) Ric-8A, Gα proteins, and actin expressions were detected by immunoblot of lysates prepared from sorted B220+CD19+ splenic cells from ric8wt/wtmb1-cre and ric8fl/flmb1-cre mice. (B) Data shown are from complete blood counts from 14 ric8wt/wtmb1-cre and 12 ric8fl/flmb1-cre mice. (C) Flow cytometry characterization of lymphoid organs from control and ric8fl/flmb1-cre mice used in (B). Absolute numbers of cells are shown from the spleen, from mesenteric and inguinal lymph nodes (mLN and iLN, respectively), and from Peyer’s patches (PP). (D) Flow cytometry was used to characterize B cell development in the bone marrow of ric8wt/wtmb1-cre and ric8fl/flmb1-cre mice. Shown are the percentages of B220+ cells in bone marrow, Fr. A–F. (E) Flow cytometry was used to characterize B cell development in the spleen of ric8wt/wtmb1-cre and ric8fl/flmb1-cre mice. Shown are the percentages of B220+ cells in the T1, T2-FO, T2-MZP, FO, and MZ subsets. Below is an immunoblot of Ric-8A expression in the different subsets. (F) Flow cytometry was used to analyze cells isolated from the spleen, iLN, and PP 10 d following immunization of ric8wt/wtmb1-cre and ric8fl/flmb1-cre mice with SRBCs. Absolute numbers of B220+, CD4+, CD8+, GC, T FO helper cells (TFH), early plasma cell (B220+CD138+), and switched memory cells (B220+, CD38+/−, CD138−, IgG1+) are indicated. Results are from the analysis of four experimental and four control mice. Experiment was repeated three times with similar results. *p < 0.05, **p < 0.005.
The loss of Ric-8A reduces B cell chemotaxis and profoundly impairs chemokine-induced intracellular calcium responses
We used splenic B cells isolated from mixed chimeric mice reconstituted with 1:1 mix of bone marrow from ric8fl/flmb1-cre and WT mice in standard chemotaxis assays using CXCL12, CCL19, and CXCL13 as chemoattractants. We could distinguish the WT and the Ric-8A–deficient B cells by flow cytometry, using the differential expression of CD45.1 versus CD45.2. This allowed a direct comparison between the two sources of B cells in the same assay. We examined the response of the entire population of splenic B cells as well as the splenic B cell subsets (Fig. 3A). We did not compare the MZ B cells, as the ric8fl/flmb1-cre mice essentially lacked them. The ric8fl/flmb1-cre B cells responded only at the highest ligand concentration tested. Although the loss of Ric-8A had markedly reduced the responsiveness of the cells, sufficient Gαi remained to allow the cells to respond to the higher dose of chemoattractant. In contrast, mice lacking Gαi2 and Gαi3 exhibit little or no chemotaxis, even at high ligand concentrations (27). The absence of Ric-8A in B cells did not affect their expression of CXCR4, CCR7, or CXCR5 or that of several different homing and adhesion receptors (Fig. 3B). This finding differs from the modest reductions in CXCR4, CCR7, and CXCR5 expression noted in Gαi2/3-deficient B cells (27). The B cells from the ric8fl/flmb1-cre mice exhibited a profoundly impaired intracellular calcium response to the three chemokines. Even at the highest ligand concentration, little intracellular calcium was elicited (Fig. 3C). Although lymphocyte chemoattractant receptor signaling, including increases in intracellular calcium, is absolutely dependent on Gαi, it is possible that Gαq also contributes to the calcium flux. In the context of reduced levels of Gαi proteins, the loss of Gαq in the ric8fl/flmb1-cre B cells may have resulted in a more severe loss in the mobilization of intracellular calcium than otherwise.
B cell–specific loss of Ric-8A impairs responses to chemokines. (A) Chemotaxis assays using splenic B cells from mixed bone marrow chimera mice. C57BL/6 CD45.1 mice were reconstituted with bone marrow from ric8fl/flmb1-cre CD45.2 mice and C57BL/6 CD45.1 mice. At 8 wk later, purified B220+ cells were immunostained for B cell subset markers, CD45.1, and CD45.2 and were subjected to chemotaxis assays with the indicated concentrations of chemokine. The percentages of migratory cells for each subset and each genotype are shown. The data are from four reconstituted mice, performed in duplicate. The indicated chemokine concentrations are nanograms per milliliter. Data shown as mean ±SEM. (B) Flow cytometry analysis of CXCR4, ICAM1, CXCR5, CCR7, CD11a, CD62L, and α4β7 expression on B220+ cells from the same mice used in (A). WT and ric8fl/flmb1-cre cells were distinguished by CD45.1 versus CD45.2 immunostaining. (C) Intracellular calcium response to chemokines using B cells from mb1-cre and ric8fl/flmb1-cre mice. The cells were stimulated with the indicated amount of CXCL12, CCL19, or CXCL13 and the induced change in intracellular calcium monitored over 3 min. Data represent the maximal calcium change plotted against the chemokine concentration (in nanograms per milliliter). Data are from three experiments.
B cell–specific loss of Ric-8A impairs responses to chemokines. (A) Chemotaxis assays using splenic B cells from mixed bone marrow chimera mice. C57BL/6 CD45.1 mice were reconstituted with bone marrow from ric8fl/flmb1-cre CD45.2 mice and C57BL/6 CD45.1 mice. At 8 wk later, purified B220+ cells were immunostained for B cell subset markers, CD45.1, and CD45.2 and were subjected to chemotaxis assays with the indicated concentrations of chemokine. The percentages of migratory cells for each subset and each genotype are shown. The data are from four reconstituted mice, performed in duplicate. The indicated chemokine concentrations are nanograms per milliliter. Data shown as mean ±SEM. (B) Flow cytometry analysis of CXCR4, ICAM1, CXCR5, CCR7, CD11a, CD62L, and α4β7 expression on B220+ cells from the same mice used in (A). WT and ric8fl/flmb1-cre cells were distinguished by CD45.1 versus CD45.2 immunostaining. (C) Intracellular calcium response to chemokines using B cells from mb1-cre and ric8fl/flmb1-cre mice. The cells were stimulated with the indicated amount of CXCL12, CCL19, or CXCL13 and the induced change in intracellular calcium monitored over 3 min. Data represent the maximal calcium change plotted against the chemokine concentration (in nanograms per milliliter). Data are from three experiments.
Lymphoid organ architecture is disrupted in the ric8fl/flmb1-cre mice
To assess the status of the lymphoid architecture in the ric8fl/flmb1-cre mice, we performed high-resolution multicolor confocal microscopy with spleens and lymph nodes from nonimmunized and immunized mice (Fig. 4A, 4B). Thick sections were immunostained for B220 (B cell marker), CD21 (enriched on FO dendritic cells and MZ B cells), CD4 (marker for CD4+ T cells), Ki67 (expressed by activated and proliferating cells), and CD169 (marker for MZ macrophages). The loss of Ric-8A in B cells severely disturbed the organization of the spleen. Shown are images of the spleens from WT (Fig. 4A) and ric8fl/flmb1-cre mice (Fig. 4B). Nearly an 80% reduction in the number of visible primary B cell follicles in the ric8fl/flmb1-cre mice spleens was observed. There was little or no discernible MZ surrounding the B cell follicles and no spontaneous GC formation. Numerous ric8fl/flmb1-cre B cells inappropriately localized in T cell zones, and an excessive number of B cells resided in the red pulp. The residual B cell follicles were poorly delineated from the T cell zone, as the normal sharp B/T borders were missing in the mutant mice. Following immunization, those GCs that formed were small and disorganized. No discernible light and dark zones were identifiable, and Ki67+ cells were distributed throughout the GC-like structure. The lymph node architecture in the ric8fl/flmb1-cre mice largely mirrored that of the spleen. The nonimmunized mice had small, poorly delineated primary follicles, and immunization led to a poor GC response, with small GCs present only occasionally in the primary follicles (Fig. 4A, 4B). These results show that the loss of Ric-8A in B cells profoundly disturbs the normal lymphoid architecture in secondary lymphoid organs; however, the abnormalities are not as extreme as those noted in mice whose B cells totally lacked both Gαi2 and Gαi3 (27).
Ric8fl/flmb1-cre mice have disorganized lymphoid organs and poorly structured GCs. (A and B) Confocal microscopy imaging of the spleen and lymph nodes from nonimmunized mice and mice 8 d post SRBC immunization. Spleen and lymph node sections from ric8fl/fl (A) and ric8fl/flmb1-cre (B) mice were immunostained for B220 (green), CD4 (blue), CD21/35 (red), Ki67 (white), and CD169 (pink). The top panel images are tiled images of the spleens. Scale bar, 300 μm. The two images in the middle panels are ×4 electronic zooms of the spleens from a nonimmunized and an immunized mouse. Scale bar, 80 μm. The two images in the bottom panels are from the inguinal lymph node of a nonimmunized and an immunized mouse. Scale bar, 200 μm. GC, B cell zone (B), and T cell zone (TCZ) are indicated.
Ric8fl/flmb1-cre mice have disorganized lymphoid organs and poorly structured GCs. (A and B) Confocal microscopy imaging of the spleen and lymph nodes from nonimmunized mice and mice 8 d post SRBC immunization. Spleen and lymph node sections from ric8fl/fl (A) and ric8fl/flmb1-cre (B) mice were immunostained for B220 (green), CD4 (blue), CD21/35 (red), Ki67 (white), and CD169 (pink). The top panel images are tiled images of the spleens. Scale bar, 300 μm. The two images in the middle panels are ×4 electronic zooms of the spleens from a nonimmunized and an immunized mouse. Scale bar, 80 μm. The two images in the bottom panels are from the inguinal lymph node of a nonimmunized and an immunized mouse. Scale bar, 200 μm. GC, B cell zone (B), and T cell zone (TCZ) are indicated.
B cells from ric8fl/flmb1-cre mice enter the splenic white pulp and lymph nodes poorly, tend to reside near lymph node high endothelial venules or in the red pulp of the spleen, and have reduced in vivo motility
Because the Ric-8A–deficient B cells can enter into lymph nodes, we could assess their localization and motility by intravital two-photon microscopy. B cells from ric8fl/flmb1-cre and WT mice were differentially fluorescently labeled and adoptively transferred to WT mice at a 1:1 and 3:1 ratio, respectively. The following day, the inguinal lymph node of the recipient mouse was prepared for intravital microscopy. The blood vessels were outlined by i.v. injection of Evans blue. As expected, fewer Ric-8A–deficient B cells resided in the lymph node 24 h after transfer. Those cells that had entered tended to reside near the high endothelial venules, many failing to enter into the lymph node follicle (Fig. 5A, Supplemental Video 1). Following the imaging procedure, the animals were sacrificed, and localization of the transferred B cells in the spleen was assessed. Fewer Ric-8A–deficient B cells than WT entered the splenic white pulp, as most of the transferred mutant B cells resided in the red pulp (Fig. 5B). Analysis of the imaging data from the inguinal lymph node allowed us to track the movement of the transferred B cells over time and to derive motility parameters. We focused only on the B cells that had entered into the follicle. The B cells from the ric8fl/flmb1-cre mice moved slower and more erratically than did the WT B cells. They were also less polarized (Fig. 5C). These results are consistent with an impaired input from chemoattractant receptors, which help support much of the spontaneous motility of B cells in lymph nodes.
Intravital microscopy of the inguinal lymph node and imaging of spleen sections reveal that loss of Ric-8A leads to B cells with decreased motility that access B cell niches poorly. (A) Images acquired during intravital microscopy of the inguinal lymph node of a recipient mouse that had received an i.v. transfer of WT (green) and ric8Afl/flmb1-cre (red) B cells (1:1 ratio) 18 h prior to imaging. Evans blue was infused i.v. immediately before the imaging to outline blood vessels (gray). (B) Standard confocal microscopy of a thick spleen section from the same mouse used for intravital microscopy focused on a splenic primary follicle. The T cell zone (TCZ) and B cell zone are shown. The section was immunostained for CD4 (green) and CD169 (red). ric8fl/flmb1-cre B cells (pink) and ric8fl/fl (blue) B cells are shown. The location of the transferred B cells in the surrounding red pulp is indicated with arrowheads of the corresponding color. (C) Motility parameters of the ric8fl/flmb1-cre B cells and ric8fl/fl (control) B cells in the lymph node follicle of the recipient mice. Also shown are polarity measurements (long axis/short axis) of the control and ric-8fl/flmb1-cre B cells. Results are based on the analysis of two separate imaging experiments. **p < 0.005, ***p < 0.0005.
Intravital microscopy of the inguinal lymph node and imaging of spleen sections reveal that loss of Ric-8A leads to B cells with decreased motility that access B cell niches poorly. (A) Images acquired during intravital microscopy of the inguinal lymph node of a recipient mouse that had received an i.v. transfer of WT (green) and ric8Afl/flmb1-cre (red) B cells (1:1 ratio) 18 h prior to imaging. Evans blue was infused i.v. immediately before the imaging to outline blood vessels (gray). (B) Standard confocal microscopy of a thick spleen section from the same mouse used for intravital microscopy focused on a splenic primary follicle. The T cell zone (TCZ) and B cell zone are shown. The section was immunostained for CD4 (green) and CD169 (red). ric8fl/flmb1-cre B cells (pink) and ric8fl/fl (blue) B cells are shown. The location of the transferred B cells in the surrounding red pulp is indicated with arrowheads of the corresponding color. (C) Motility parameters of the ric8fl/flmb1-cre B cells and ric8fl/fl (control) B cells in the lymph node follicle of the recipient mice. Also shown are polarity measurements (long axis/short axis) of the control and ric-8fl/flmb1-cre B cells. Results are based on the analysis of two separate imaging experiments. **p < 0.005, ***p < 0.0005.
Loss of Ric-8A reduced anti-IgM–induced increases in intracellular calcium, did not affect B cell proliferation, but enhanced B cell survival
As B cells that lack Gαi2 or both Gαi2/3 have a reduced anti-IgM–induced intracellular calcium response (27), we checked the response of the B cells from the ric8fl/flmb1-cre mice. We found that like Gαi-deficient B cells, the Ric-8A–deficient B cells also demonstrated a suboptimal peak intracellular calcium response to anti-Ig stimulation. We verified these results using B cells purified by the ric8fl/flvav1-cre mice. We noted a decrease following stimulation with either a low or a high concentration of anti-IgM (Fig. 6A). Using a CFSE dye dilution assay, we checked the proliferative response of B cells from the ric8fl/flmb1-cre and WT mice. Although we found little difference in the proliferation index (the average number of cell divisions that a cell in the original population undergoes) or the division index (the average number of cell divisions of the responding cells), we did note a modest reduction in the perentage of Ric-8A–deficient B cells that divided when Il-4 was present in the culture. We also found that at the end of the culture period we recovered more Ric-8A–deficient B cells following CD40 or CD40 and IgM stimulation, suggesting that the loss of Ric-8A had led to an improved B cell survival rate (Fig. 6B). B cells from mice lacking Gαq are known to have an intrinsic survival advantage over normal B cells (31), suggesting that the loss of Gαq in Ric-8A–deficient B cells may have accounted for their enhanced survival.
Whereas Ric-8A–deficient B cells respond relatively normally to in vitro B cell proliferative signals, Ric8fl/flmb1-cre mice have low serum Ig levels and generate poor Ab responses following immunization. (A) Control (mb1-cre), ric8fl/flmb1-cre, and ric8fl/flvav1-cre purified B cells were stimulated with anti-IgM, and the induced changes in intracellular calcium were monitored over 3 min. The graph (top) shows results from ric8fl/fl versus ric8fl/flmb1-cre B cells. The data (bottom) are shown as a percentage of the maximal response of control cells. The experimental value is the mean of three determinations. Experiments were repeated two times with similar results. (B) Analysis of flow cytometry results from CFSE dye dilution assays using B cells purified from 3 WT (ric8fl/fl) and 3 ric8fl/flmb1-cre mice and stimulated for 96 h as indicated. The division index, percentage of cells divided, and the proliferation index were calculated using FlowJo. The numbers of viable B cells recovered at the end of the experiment are shown. Each assay was done in duplicate. Similar results in two other experiments. (C) ELISA assay to measure Ig isotypes in the sera of control (ric8Afl/fl, vav1-cre, and mb1-cre) and ric8Afl/flmb1-cre mice. Sera analyzed were from six mice for each genotype. (D) ELISA assay measuring specific Ab in the serum of NP-Ficoll–immunized control and ric8fl/flmb1-cre mice. Four mice of each genotype were immunized and sera collected at the indicated time points. ELISA results from individual time points were compared. (E) ELISA assay measuring specific Ab in the serum of NP-KLH–immunized control and ric8fl/flmb1-cre mice. Four mice of each genotype were immunized, and serum was collected at the indicated time points. Mice were boosted at day 60. Results from individual time points were compared. *p < 0.05, **p < 0.005.
Whereas Ric-8A–deficient B cells respond relatively normally to in vitro B cell proliferative signals, Ric8fl/flmb1-cre mice have low serum Ig levels and generate poor Ab responses following immunization. (A) Control (mb1-cre), ric8fl/flmb1-cre, and ric8fl/flvav1-cre purified B cells were stimulated with anti-IgM, and the induced changes in intracellular calcium were monitored over 3 min. The graph (top) shows results from ric8fl/fl versus ric8fl/flmb1-cre B cells. The data (bottom) are shown as a percentage of the maximal response of control cells. The experimental value is the mean of three determinations. Experiments were repeated two times with similar results. (B) Analysis of flow cytometry results from CFSE dye dilution assays using B cells purified from 3 WT (ric8fl/fl) and 3 ric8fl/flmb1-cre mice and stimulated for 96 h as indicated. The division index, percentage of cells divided, and the proliferation index were calculated using FlowJo. The numbers of viable B cells recovered at the end of the experiment are shown. Each assay was done in duplicate. Similar results in two other experiments. (C) ELISA assay to measure Ig isotypes in the sera of control (ric8Afl/fl, vav1-cre, and mb1-cre) and ric8Afl/flmb1-cre mice. Sera analyzed were from six mice for each genotype. (D) ELISA assay measuring specific Ab in the serum of NP-Ficoll–immunized control and ric8fl/flmb1-cre mice. Four mice of each genotype were immunized and sera collected at the indicated time points. ELISA results from individual time points were compared. (E) ELISA assay measuring specific Ab in the serum of NP-KLH–immunized control and ric8fl/flmb1-cre mice. Four mice of each genotype were immunized, and serum was collected at the indicated time points. Mice were boosted at day 60. Results from individual time points were compared. *p < 0.05, **p < 0.005.
The lack of Ric-8A in B cells reduces serum Ig levels and significantly impairs the humoral immune response to a thymus-dependent and an independent Ag
The reduction in MZ B cells and the reduced number of peripheral B cells portended lower levels of serum Igs in the ric8fl/flmb1-cre mice. As anticipated, we found decreased serum levels of IgA, IgM, and IgG isotypes. In contrast, the mb1-cre mice had a serum Ig profile similar to that of the vav1-cre mice, arguing that a loss of single mb-1 allele does not significantly compromise humoral immunity (Fig. 6C). Next, we check specific Ab responses following immunization with a thymus-dependent or a thymus-independent Ag. The ric8fl/flmb1-cre mice had little or no increase in serum Ig specific for NP following immunization with NP-Ficoll, which will elicit a thymus-independent B cell response (Fig. 6D). This finding is consistent with the marked loss of MZ B cells in these mice. Immunization with NP-KLH, which triggers a thymus-dependent response, led to a relatively normal primary and secondary NP-specific IgM response. In contrast, both the primary and secondary NP IgG responses were very poor in the immunized ric8fl/flmb1-cre mice (Fig. 6E). This result is consistent with the poor GC response and the reduction in IgG1+ B cells noted following immunization of the ric8fl/flmb1-cre mice with SRBCs. These results show a marked impairment in humoral immunity, particularly in eliciting responses to new Ags.
The ric8fl/flmb1-cre B cells exhibit an occasional delay in cytokinesis abscission, decreased F-actin levels, and a reduced incidence of asymmetric cell divisions
We had previously noted that in human cell lines Gαi proteins, RGS14, LGN, and Ric-8A all accumulated at the midbody during cytokinesis (10, 11). We had also shown that lowering Ric-8A expression delayed the abscission time of dividing human cells, which correlated with increased intercellular bridge length and multinucleation (12). To visualize dividing ric8fl/flmb1-cre B cells, we generated ric8fl/flmb1-cre EGFP-Lifeact mice. Lifeact accurately labels F-actin and is useful for visualizing dynamic changes in F-actin networks (32). B cells from these mice and control mice were cultured on ICAM-1–coated plates and imaged following LPS stimulation. The stimulated B cells from ric8fl/flmb1-cre mice underwent symmetric mitosis like that of B cells from WT EGFP-Lifeact mice. However, as we had observed with the human cell lines, we noted that some of the ric8fl/flmb1-cre EGFP-Lifeact B cells had a delayed abscission time following mitosis (Fig. 7A, 7B, Supplemental Video 2, Supplemental Video 3). The B cells from the ric8fl/flmb1-cre mice also exhibited a reduced motility, had difficulty in maintaining a polarized phenotype, and had lower overall levels of F-actin (Fig. 7C, data not shown).
Occasional cytokinesis failure and reduced asymmetric cell division in B cells from ric8fl/flmb1-cre mice. (A) Confocal microscopy of division of a ric8wt/wt Lifeact B cell and ric8Afl/flmb1-cre Lifeact B cells. Splenic B cells purified from ric8wt/wt Lifeact and ric8Afl/flmb1-cre Lifeact mice were placed on ICAM-1–coated plates, stimulated with LPS, and imaged using time-lapse live cell confocal microscopy. Image stacks (three to five slices) were acquired every 15–20 s for 1 h, and dividing B cells were identified by morphology changes and varying Lifeact expression. Single-slice confocal images of WT (top panel) and ric8fl/flmb1-cre B cells (bottom three panels) are shown. Scale bar, 5 μm. (B) Delayed cytokinesis in the ric8fl/flmb1-cre B cells. LPS-activated B cells plated as in (A) were imaged, and the number of minutes required from initial cleavage furrow formation to completion of cytokinesis was determined. WT Lifeact and ric8fl/flmb1-cre Lifeact B cells were compared. The results are from three separate imaging experiments. (C) Confocal microscopy images and F-actin levels in LPS-stimulated WT (Lifeact) and ric8fl/flmb1-cre B cells plated on ICAM-1–coated plates. The level of Lifeact immunofluorescence in single cells was quantitated using Imaris, and results are shown in the graph on the right. Scale bar, 10 μm. (D) Imagestream analysis to measure the number of asymmetrically dividing ric8Afl/fl and ric8Afl/flmb1-cre B cells. Prior to the analysis, purified B cells were stimulated, or not, with CD40 + IgM or LPS for 72 h. The B cells were immunostained for B220-PE-Cy7, fixed, immunostained for PKCζ, and counterstained with DRAQ5 to outline the nuclei. A representative graph showing the number of cells harboring an asymmetric PKCζ distribution is shown (left panel), and beneath are images from a symmetric (above) and an asymmetric (below) dividing cell. The numbers of cells with asymmetric PKCζ are shown (middle panel). The populations’ proliferative status is shown (right panel). (E) Imagestream analysis to determine the percentage of GC B cells showing an asymmetric PKCζ distribution among the GC B cells analyzed from the spleens from WT or ric8fl/flmb1-cre mice immunized 10 d previously with SRBCs. The B cells were immunostained for B220-PE-Cy7, fixed, immunostained for aPKC, and counterstained with DRAQ5 to outline the nuclei. *p < 0.05, **p < 0.005.
Occasional cytokinesis failure and reduced asymmetric cell division in B cells from ric8fl/flmb1-cre mice. (A) Confocal microscopy of division of a ric8wt/wt Lifeact B cell and ric8Afl/flmb1-cre Lifeact B cells. Splenic B cells purified from ric8wt/wt Lifeact and ric8Afl/flmb1-cre Lifeact mice were placed on ICAM-1–coated plates, stimulated with LPS, and imaged using time-lapse live cell confocal microscopy. Image stacks (three to five slices) were acquired every 15–20 s for 1 h, and dividing B cells were identified by morphology changes and varying Lifeact expression. Single-slice confocal images of WT (top panel) and ric8fl/flmb1-cre B cells (bottom three panels) are shown. Scale bar, 5 μm. (B) Delayed cytokinesis in the ric8fl/flmb1-cre B cells. LPS-activated B cells plated as in (A) were imaged, and the number of minutes required from initial cleavage furrow formation to completion of cytokinesis was determined. WT Lifeact and ric8fl/flmb1-cre Lifeact B cells were compared. The results are from three separate imaging experiments. (C) Confocal microscopy images and F-actin levels in LPS-stimulated WT (Lifeact) and ric8fl/flmb1-cre B cells plated on ICAM-1–coated plates. The level of Lifeact immunofluorescence in single cells was quantitated using Imaris, and results are shown in the graph on the right. Scale bar, 10 μm. (D) Imagestream analysis to measure the number of asymmetrically dividing ric8Afl/fl and ric8Afl/flmb1-cre B cells. Prior to the analysis, purified B cells were stimulated, or not, with CD40 + IgM or LPS for 72 h. The B cells were immunostained for B220-PE-Cy7, fixed, immunostained for PKCζ, and counterstained with DRAQ5 to outline the nuclei. A representative graph showing the number of cells harboring an asymmetric PKCζ distribution is shown (left panel), and beneath are images from a symmetric (above) and an asymmetric (below) dividing cell. The numbers of cells with asymmetric PKCζ are shown (middle panel). The populations’ proliferative status is shown (right panel). (E) Imagestream analysis to determine the percentage of GC B cells showing an asymmetric PKCζ distribution among the GC B cells analyzed from the spleens from WT or ric8fl/flmb1-cre mice immunized 10 d previously with SRBCs. The B cells were immunostained for B220-PE-Cy7, fixed, immunostained for aPKC, and counterstained with DRAQ5 to outline the nuclei. *p < 0.05, **p < 0.005.
As heterotrimeric G-protein signaling proteins can shape and regulate the orientation of the spindle body with the axis of polarity during asymmetric cell division (33), we assessed the role of the B cell–specific ric8 deletion in asymmetric cell division. Polarity is regulated by complexes of evolutionarily conserved polarity proteins known as the Scribble [including Scribble, lethal giant larvae, and Dlg (discs large)] and Par [including Par3, Par6, and aPKC (PKCζ)] complexes, which antagonize each other to define molecularly distinct regions of the cell (34). We used a fast and unbiased methodology to quantify the frequency of dividing B cells with an asymmetric PKCζ distribution (Fig. 7D, left panel). This approach showed that the loss of Ric-8A expression decreases the propensity of dividing B cells to undergo asymmetric polarization of PKCζ following in vitro stimulation with LPS or CD40 plus IgM (Fig. 7D, middle panel). The decrease in asymmetric localization of PKCζ cannot be accounted for by a reduction in the dividing cells, as we found no significant decrease in the percentage of dividing cells in ric8fl/flmb1-cre B cells at 72 h post stimulation, using this assay (Fig. 7D, right panel). Although there is a severe reduction in GC B cells in the immunized ric8fl/flmb1-cre mice, we could use the same high-throughput methodology to analyze splenic B cells from immunized mice to determine the percentage of GC cells with an asymmetric localization of PKCζ. We found that 10 d following SRBC immunization, ∼6% of the phenotypic GC B cells (B220+GL7+Fas+) from the WT mice had an asymmetric distribution of PKCζ, whereas only 2% of the phenotypically similar cells from the ric8fl/flmb1-cre B cells had an asymmetric distribution (Fig. 7E). Together these results point to the importance of heterotrimeric G-protein signaling proteins in the regulation of polarity components during B cell activation and differentiation.
Discussion
Ric-8A is a GEF for a subset of Gα subunits and is needed for normal levels of Gαi and Gαq and likely Gα13 proteins in lymphocytes and macrophages. Mice with hematopoietic cell–specific loss of Ric-8A had lymphocytosis, anemia, and a shortened lifespan. Yet blood neutrophil, eosinophil, basophil, and platelet numbers in the blood were relatively normal. This finding indicates that hematopoietic stem cell function and bone marrow egress were largely intact. Hematopoietic cell differentiation proceeded despite the lack of Ric-8A and decreased levels of Gαi, Gαq, and Gα13 proteins in hematopoietic progenitors. Nevertheless, functional defects in mature hematopoietic cells can be expected. The causes of anemia and decreased longevity in these mice are unknown, but one possibility is bleeding resulting from impaired platelet function. The absence of Gαq, Gαi, or Gα13 is known to impair platelet function and can cause a bleeding diathesis (35–37). The lymphocytosis was accompanied by a reduction in B cells in secondary lymphoid organs, which suggested a B lymphocyte trafficking defect. Surprisingly, T cell trafficking was less affected and thymocyte development proceeded with evidence of only a mild egress defect. Nevertheless, further studies using a T cell–specific Cre are warranted. In this study, we focused on the role of Ric-8A in B cells and B cell function by generating mice with a B lymphocyte–specific loss of Ric-8A. Study of these mice confirmed the importance of Ric-8A for B cell trafficking; chemokine receptor and BCR signaling; GC and memory B cell responses; asymmetric cell division; and humoral responses to novel Ags.
As indicated in the introduction, the steady state levels of Gαi, Gαq, and Gα13 were reduced by 85% in the ric8−/− embryonic stem cells compared with similar cells from WT mice. Our study shows that absence of Ric-8A in lymphocytes and bone marrow–derived macrophages causes a similar reduction in Gαi2/3, Gαq, and Gα13, although Gα12 levels were unaffected. Presumably, other hematopoietic cells in the ric8fl/flvav1-cre mice are similarly affected, although we have not directly examined them. All hematopoietic cells have ric8 mRNA transcripts, although neutrophils conspicuously have the highest levels (18). Consistent with a major loss of Gαi proteins in neutrophils, the ric8fl/flvav1-cre neutrophils were very poorly recruited to inflammatory sites in vivo (O. Kamenyeva, unpublished observations).
By a variety of in vitro and in vivo assays, the responses of the ric8fl/flmb1-cre B cells to chemoattractants are severely degraded. Essentially, none of the ric8fl/flmb1-cre B cells migrated to near-optimal concentrations of CXCL12 and CCL19, and only at the highest concentration of CXCL13 did we observe some migration. This finding is likely accounted for by residual expression of Gαi2 and Gαi3 in the ric8fl/flmb1-cre B cells, as the Gαi2/3-deficient B cells are refractory (27). The loss of MZ B cells in the ric8fl/flmb1-cre mice also argues for a physiologically significant loss in Gαi signaling. A reduction in MZ B cells was noted in Gαi2-deficient mice, and the sphingosine 1-phosphate receptor 1 and cannabinoid receptor 2 are needed for the proper positioning of MZ B cells (38–40).
Engagement of chemoattractant receptors also elicits a rise in intracellular Ca2+, which is lost in the absence of Gαi signaling. As previously mentioned, it is possible that the marked reduction in intracellular Ca2+ following exposure to chemokines may have arisen from the reduction of Gαq in the setting of a reduced, but not absent, Gαi in these mice. The disruption of the lymphoid organ architecture is also consistent with a major defect in B cell trafficking. In the spleen, Gαi nucleotide exchange is needed to enter into the white pulp (41), and loss of CXCR5 leads to a failure to populate B cell follicles (42). In the ric8fl/flmb1-cre mice, the splenic white pulp is poorly populated with B cells, the B/T borders are indistinct, MZ B cells are rare, and many B cells resided in the red pulp. Lymph nodes also had small B cell zones with indistinct B/T borders. Although not shown, Peyer’s patches were small and poorly developed in the ric8fl/flmb1-cre mice. This was also observed in mice with Gαi2/Gαi3-deficient B cells and, as previously discussed, may be secondary to a reduction in B cell recruitment into the developing Peyer’s patches (27). The adoptive transfer studies supported the lymphoid architecture changes, as many ric8fl/flmb1-cre B cells resided near high endothelial venules, failing to enter into the lymph node follicle and the splenic white pulp. Those B cells that did enter the follicle exhibited a motility pattern consistent with diminished Gαi signaling (43). Thus, the loss of Ric-8A in B cells reduces Gαi2 and Gαi3 levels sufficiently to severely impair chemoattractant receptor signaling. This loss was not accompanied by any significant change in chemoattractant receptor expression. This finding was in contrast to the Gαi2/3-deficient B cells, in which a modest decrease in CXCR4, CCR7, and CXCR5 expression was noted (27).
The reduction of BCR-mediated increases in intracellular calcium found with the Ric-8A–deficient B cells was also seen with the Gαi2- and Gαi2/3-deficient B cells (27). In contrast, Gαq−/− B cells are hyperresponsive to BCR crosslinking, exhibiting higher amounts of phosphorylated AKT, ERK, and PLCγ2 following stimulation. They also had an augmented proliferative response to LPS and enhanced B cell viability (31). Although the ric8fl/flmb1-cre B cells had major reductions in both Gαq and Gαi, the ric8fl/flmb1-cre B cell phenotype reflected the loss of Gαi more than the loss of Gαq. Like the Gnai3−/−Gnai2fl/flmb1-cre mice, the ric8fl/flmb1-cre mice had a reduced number of splenic B cells and a severe reduction in MZ B cells, whereas the Gnaq−/− mice had an expansion of splenic B cells, including MZ B cells. We did note that with some inductive signals in in vitro cultures we recovered more ric8fl/flmb1-cre B cells than control B cells, a result consistent with the improved B cell survival noted with the Gαq−/− B cells. We saw no evidence of autoimmunity in the ric8fl/flmb1-cre mice, as had been noted in the Gnaq−/− mice.
Of interest, the serum Ig profiles and Ab responses to immunization in the ric8fl/flmb1-cre mice differed from those of both the Gnai3−/−Gnai2fl/flmb1-cre and the Gnai2fl/flmb1-cre mice. The Gnai3−/−Gnai2fl/flmb1-cre mice had a more severe phenotype, exhibiting a hyper IgM–like syndrome with elevated serum IgM and severely depressed other Ig isotypes (27). The Gnai2fl/flmb1-cre mice had a more modest phenotype, with reductions only in serum IgG1 and IgG2b. When immunized with a thymus-independent Ag, these mice had a slightly decreased IgM and IgG3 response and paradoxically an elevated IgG2C response. For unclear reasons, the ric8fl/flmb1-cre mice had a more severe phenotype with essentially no specific response to the same thymus-independent Ag. Similarly, a thymus-dependent Ag elicited a much poorer response in the ric8fl/flmb1-cre mice than had been observed with the Gnai2fl/flmb1-cre mice. The ric8fl/flmb1-cre mice had a poor primary and essentially no secondary response. This finding is likely a consequence of the much poorer induction of GC cells in the ric8fl/flmb1-cre mice (∼10-fold decrease) than had been observed in either the Gnai2fl/flmb1-cre or the Gnai3−/− mice (∼2-fold reduction in each strain). The GC response in the ric8fl/flmb1-cre mice much more resembled that previously observed with the Gnai3−/−Gnai2fl/flmb1-cre mice. Whereas B cell chemotaxis is highly dependent on Gαi2, and less so on Gαi3, evidently the GC response depends on the presence of both Gαi2 and Gαi3 (27). CXCR4, CXCR5, and Ebi2 signaling help orchestrate GC B cell dynamics (44–46), and their combined loss would likely recapitulate the severe GC phenotype noted in the ric8fl/flmb1-cre and Gnai3−/−Gnai2fl/flmb1-cre mice.
In recent years, a conserved general mechanism for asymmetric cell division has been discovered: asymmetric localization of Par proteins polarizes the cell cortex, orients the mitotic spindle through heterotrimeric G-proteins, and directs the segregation of determinants into only one of the two daughter cells. The polarity axis can be established by either cell-intrinsic or -extrinsic environmental signal (47). These observations have been extended to hematopoiesis and lymphocyte differentiation, although not without controversies (48). In dividing T cells, an unequal partitioning of membrane and cytoplasmic proteins can lead to the generation of two phenotypically distinct populations of daughter cells (24, 49). In B cells, the polarized secretion of lysosomes at the B cell synapse couples Ag extraction to presentation, and the asymmetric segregation of Ag during cell division shapes the subsequent capacity of B cells to present Ag (50, 51). Furthermore, GC B cells asymmetrically segregate the transcriptional regulator Bcl6, the IL-21R, and an aPKC to one side of the plane of division by a mechanism that depends upon extrinsic signals from the microenvironment. However, another study indicated that Scribble-mediated asymmetric cell division is not required for humoral immunity (52). In the current study, we addressed whether the loss of Ric-8A affected asymmetric partitioning of PKCζ. Our results showed that loss of Ric-8A expression and function switched off the molecular mechanisms in activated and GC B cells that had led to the asymmetric distribution of PKCζ. This finding is likely secondary to the loss of the activities of Ric-8A as an exchange factor and a chaperone, both of which have been implicated in mitotic spindle orientation. Despite this observation and those from others, many questions remain unanswered. For example, it remains unclear whether asymmetric cell divisions affect the decisions of germinal B cells to re-enter the GC, to adopt one of two alternative fates: a plasma cell or a memory B cell. More detailed studies combining fate mapping techniques (53) with sorting of cells harboring an asymmetric distribution (54) should help reveal the role that asymmetric cell division plays in lymphocyte fate decisions. Other interesting questions arise: how do G-proteins act on microtubules and how is this process regulated? A growing body of evidence links a polarized Cdc42 activation to the initiation of asymmetric cell division (51, 55). One model that could explain an asymmetric activation of Cdc42 is a spatial organization of GoLoco proteins that could free Gα protein. Their subsequent activation could be Ric-8A dependent to mediate Cdc42 activation (56).
This study shows that hematopoiesis, bone marrow B cell development, and thymopoiesis occur in the absence of Ric-8A. The loss of Ric-8A from splenocytes, B cells, and likely other hematopoietic cell types leads to major decreases in the cellular content of Gαi, Gαq, and likely Gα13 proteins. The B lymphocyte–specific loss of Ric-8A impairs B cell development in the spleen and severely impairs B lymphocyte trafficking and differentiation. B cells respond poorly to chemoattractants and fail to appropriately populate B cell niches. Humoral immunity is adversely affected with a major loss in GC formation, a marked reduction in Ab responses to novel Ags, and a nearly absent IgG memory response. The loss of Ric-8A in B cells interferes with asymmetric cell divisions, although their importance in B cell differentiation remains unclear. The ric8fl/fl mice will be a valuable resource for studying the roles of Ric-8A, Gαi, Gαq, and Gα13 proteins in immune and hematopoietic cells.
Acknowledgements
We thank Dr. Anthony Fauci for continued support.
Footnotes
This work was supported by the intramural program of the National Institute of Allergy and Infectious Diseases.
The online version of this article contains supplemental material.
Abbreviations used in this article:
- AGS3
activator of G-protein signaling 3
- aPKC
atypical protein kinase C
- DN
double negative
- FO
follicular
- Fr.
fraction
- GC
germinal center
- GEF
guanine nucleotide exchange factor
- GoLoco
Gαi/o–Loco interaction
- Insc
Inscuteable
- KLH
keyhole limpet hemocyanin
- LGN
Leu-Gly-Asn–enriched protein
- MZ
marginal zone
- MZP
marginal zone precursor
- NP
4-hydroxy-3-nitrophenylacetyl
- Ric-8A
resistance to inhibitors of cholinesterase 8A
- SP
single positive
- T1
transitional type 1
- T2
transitional type 2.
References
Disclosures
The authors have no financial conflicts of interest.