A novel Rac-dependent checkpoint in B cell development controls entry into the splenic white pulp and cell survival

Abstract

Rac1 and Rac2 GTPases transduce signals from multiple receptors leading to cell migration, adhesion, proliferation, and survival. In the absence of Rac1 and Rac2, B cell development is arrested at an IgD- transitional B cell stage that we term transitional type 0 (T0). We show that T0 cells cannot enter the white pulp of the spleen until they mature into the T1 and T2 stages, and that this entry into the white pulp requires integrin and chemokine receptor signaling and is required for cell survival. In the absence of Rac1 and Rac2, transitional B cells are unable to migrate in response to chemokines and cannot enter the splenic white pulp. We propose that loss of Rac1 and Rac2 causes arrest at the T0 stage at least in part because transitional B cells need to migrate into the white pulp to receive survival signals. Finally, we show that in the absence of Syk, a kinase that transduces B cell antigen receptor signals required for positive selection, development is arrested at the same T0 stage, with transitional B cells excluded from the white pulp. Thus, these studies identify a novel developmental checkpoint that coincides with B cell positive selection.

Figure 1.

Immature B cells are blocked in splenic development and accumulate in the blood in the absence of Rac1 and Rac2. (A) Contour plots of splenocytes from either WT or Rac1^BRac2^−/− mice showing separation of B220⁺ cells into immature (CD93⁺) and mature (CD93⁻) B cells. The immature cells were then further gated on IgM⁺ cells and separated according to the expression of IgD and CD23 into T0 (IgD⁻CD23⁻), T1 IgD⁺ (IgD⁺CD23⁻), and T2 (IgD⁺CD23⁺) subsets. Mature (CD93⁻) cells were separated according to expression of IgM and CD23 into MRF (IgM^+/−CD23⁺) and MZ (IgM⁺CD23⁻) subsets (not depicted). Numbers show the percentage of cells falling into gates or quadrants. (B) Mean (±SEM) numbers of T0, T1 IgD⁺, T2, MRF, and MZ splenic B cells in mice of the indicated genotypes (n = 6). (C) Contour plots showing expression of IgM and IgD on B220⁺ blood lymphocytes. Gates indicate immature (IgM⁺IgD⁻) and MRF (IgD⁺) B cells. Numbers show the percentage of cells falling into gates. (D) Mean (±SEM) numbers of immature and MRF B cells in the blood of mice of the indicated genotypes (n = 5). *, P < 0.05; **, P < 0.01.

Figure 2.

Analysis of population turnover by BrdU incorporation. (A and B) Turnover of immature T0-like B cells in the bone marrow and blood (CD93⁺IgM⁺IgD⁻CD23⁻) and of splenic T0, T1 IgD⁺, T2, and MRF B cells from mice of the indicated genotypes monitored by (A) continuous (n = 5–6) or (B) pulse-chase analysis of BrdU incorporation (n = 4–5) as a function of time after start of BrdU treatment. BrdU was administered to the mice starting at time 0. For pulse-chase analysis, BrdU was administered for the first 16 h only. Means ± SEM are shown. *, P < 0.05.

Figure 3.

Accumulation of Rac-deficient immature B cells in the blood may be caused by defective CXCR4 signaling. (A) Splenic T0 B cells from mice of the indicated genotypes (IgM^bLy5.2⁺) mixed with splenic T0 B cells from 129S8 (IgM^aLy5.2⁺) mice were injected into B6.SJL (IgM^bLy5.1⁺) mice. Graph shows the mean (±SEM) IgM^bLy5.2⁺ to IgM^aLy5.2⁺ ratio of B cells in the blood, spleen, and bone marrow 4 h after transfer, normalized to the input ratio (n = 5–17). (B) Mean (±SEM) number of IgM^b+ cells recovered from the blood, spleen, and bone marrow of B6.SJL mice into which WT splenic T0 B cells had been transferred 4 h earlier in the absence or presence of the CXCR4 inhibitor AMD3100 (n = 4). (C) Mean (±SEM) migration of immature bone marrow or splenic T0 cells from mice of the indicated genotypes in a Transwell assay in response to CXCL12. Wells were coated or not (−) with VCAM-1 (n = 6). *, P < 0.05; **, P < 0.01. BM, bone marrow.

Figure 4.

Transitional B cells deficient in both Rac1 and Rac2 fail to enter the white pulp of the spleen. (A and B) Contour plots show IgD and CD23 expression on B220⁺IgM^b+ cells from 129S8 (IgM^a) mice into which splenic T0 B cells from WT or Rac1^BRac2^−/− (IgM^b) mice had been transferred (A) 4 h or (B) 24 h earlier. The input cells before transfer are shown for comparison. Numbers indicate percentages of cells falling into quadrants. (C) Images showing immunofluorescence staining of sections from spleens of mice into which WT or Rac1^BRac2^−/− (IgM^b) T0 B cells had been transferred 4 or 24 h earlier. Staining for IgM^b (green) identifies transferred T0 B cells, and MadCAM-1 (red) defines the edges of the white pulp. (right) Transferred WT MRF B cells for comparison. Bar, 150 µm. (D) Mean (±SEM) ratio of transferred IgM^b+ T0 or MRF B cells ending up in white relative to red splenic pulp at 4 or 24 h after transfer, quantitated using sections such as those shown in C (n = 4–6). (E) Splenic T0 B cells from mice of the indicated genotypes (IgM^bLy5.2⁺) were injected into 129S8 (IgM^aLy5.2⁺) mice. Graph shows the mean (±SEM) percent recovery of transitional B cells in the spleen 24 h after transfer of T0 cells of the indicated genotype. Recovery of total transitional cells is shown, as well as subdivision of these into T0, T1 IgD⁺, and T2 cells (n = 5–8). *, P < 0.05; **, P < 0.01.

Figure 5.

Pertussis toxin blocks entry of T0 splenic B cells into the white pulp, as well as their survival. (A) Mean (±SEM) ratio of transferred IgM^b+ T0 B cells ending up in white relative to red splenic pulp at 24 h after transfers, as described in Fig. 4 (A–C). Transferred T0 cells were from WT or Rac1^BRac2^−/− mice. In some transfers of WT T0 B cells, the mice were pretreated with anti–LFA-1 and anti-α4 blocking antibodies (antiintegrin mAb), or the cells were treated with pertussis toxin (n = 4–6). (B and D) Contour plots show expression of IgD and CD23 on B220⁺IgM^b+ splenocytes from 129S8 (IgM^a+) mice into which (B) WT or (D) Rac1^BRac2^−/− (T0 B cells (both IgM^b+) had been transferred 24 h earlier and had been either pretreated with pertussis toxin or an oligomer of the B subunit of pertussis toxin (Oligo B). Oligo B controlled for any effects of pertussis toxin independent of the ADP-ribosylation activity of the A subunit, which inactivates Gαi-coupled GPCRs. Staining of input cells before transfer is shown for comparison. Numbers indicate percentages of cells falling into quadrants. (C and E) Graphs show the mean (±SEM) percent recovery of transitional B cells in the spleens of 129S8 mice 24 h after transfer of WT (C; n = 4–5) or Rac1^BRac2^−/− (E; n = 6–7) T0 B cells pretreated with Oligo B or pertussis toxin. Recovery of total transitional cells is shown as well as subdivision of these into T0, T1 IgD⁺, and T2 cells. *, P < 0.05; **, P < 0.01. PTX, pertussis toxin.

Figure 6.

Migration in response to lymphoid chemokines is defective in transitional B cells deficient in Rac1 and Rac2. (A) Mean (±SEM) migration in a Transwell assay of T0, T1 IgD⁺, and T2 splenic B cells from a WT mouse in response to CCL21, CXCL12, CXCL13, or no chemokine (n = 6). (B) Histograms showing cell-surface levels of α4, αL, β1, and β2 integrins, and CXCR4 and CXCR5 on T0, T1 IgD⁺, and T2 B cell subsets. (C) Graph as in A, with migration in response to different chemokines assayed in the presence of the integrin ligands ICAM-1, MadCAM-1, or VCAM-1, or with no ligand present (n = 6). (D) Graph as in A, comparing chemokine-induced migration of transitional B cells from WT and Rac1^BRac2^−/− mice (n = 6). *, P < 0.05; **, P < 0.01. No Chmk, no chemokine.

Figure 7.

Loss of Syk causes developmental arrest at the T0 transitional B cell stage. (A) Contour plots show IgD and CD23 expression on B220⁺CD93⁺IgM^hi splenocytes from irradiated B6.SJL-Rag2^−/− mice reconstituted 8 wk earlier with Syk^−/− or WT fetal liver. Numbers indicate percentages of cells falling into quadrants. (B) Mean (±SEM) numbers of pro–B (B220⁺IgM⁻IgD⁻CD2⁻), pre–B (B220⁺IgM⁻IgD⁻CD2⁺), and immature T0-like (Imm; B220⁺CD93⁺IgM⁺IgD⁻CD23⁻) cells in the bone marrow and splenic subsets (defined as in Fig. 1 B) of radiation chimeras described in A (n = 6). (C) Images showing immunofluorescence staining of sections from spleens of irradiated 129S8 (IgM^a) mice reconstituted 8 wk earlier with Syk^−/− or WT fetal liver. Staining for IgM^b (green) identifies donor B cells, and MadCAM-1 (red) defines the edges of the white pulp. Bar, 150 µm. **, P < 0.01.

Figure 8.

Expression of Bcl-xL in transitional B cells deficient in Rac1 and Rac2 rescues survival and development but not migration. Irradiated Rag1-deficient mice were reconstituted with bone marrow cells from WT or Rac1^BRac2^−/− mice that had been infected with MIGR1–Bcl-xL, a retroviral vector expressing GFP and Bcl-xL, or with a control vector (MIGR1) expressing only GFP. (A) Histograms show fluorescence of staining with an anti–Bcl-xL or isotype control antibody of GFP⁺ or GFP⁻ splenic T2 B cells from chimeras reconstituted with WT or Rac1^BRac2^−/− bone marrow cells infected with MIGR1–Bcl-xL. No increase in Bcl-xL expression was seen in T2 cells from chimeras reconstituted with cells infected with the MIGR1 vector control (not depicted; n = 3). (B) Contour plots show IgD and CD23 expression on B220⁺CD93⁺IgM^hiGFP⁺ splenocytes from chimeras reconstituted with WT or Rac1^BRac2^−/− bone marrow cells infected with MIGR1 (Control) or MIGR1–Bcl-xL (Bcl-xL) vectors. Numbers indicate percentages of cells falling into quadrants. (C) Mean (±SEM) percentage of B220⁺CD93⁺IgM^hiGFP⁺ transitional splenocytes from chimeras described in B that were T0 (IgD⁻CD23⁻), T1 IgD⁺ (IgD⁺CD23⁻), and T2 (IgD⁺CD23⁺), gated as in B (n > 4). (D) Mean (±SEM) percent migration in a Transwell assay of B220⁺IgM^hiGFP⁺ T0, T1 IgD⁺, and T2/MRF splenocytes from chimeras described in B. Migration was measured in response to no chemokine or the indicated chemokines. Colors of bars are as in C (n > 8). *, P < 0.05; **, P < 0.01. No Chmk, no chemokine.

Figure 9.

Vav family proteins are not required for entry into splenic white pulp. (A) Contour plots of splenocytes from either WT (B10.BR) or Vav1^−/−Vav2^−/−Vav3^−/− (VavTKO) mice showing separation of B220⁺ cells into immature (CD93⁺) and mature (CD93⁻) B cells. Note that the VavTKO mice are on a B10.BR background and are therefore compared with WT B10.BR mice. The immature cells were further gated on IgM⁺ cells and separated according to the expression of IgD and CD23 into T0 (IgD⁻CD23⁻), T1 IgD⁺ (IgD⁺CD23⁻), and T2 (IgD⁺CD23⁺) subsets. Mature (CD93⁻) cells were separated according to expression of IgM and CD23 into MRF (IgM^+/−CD23⁺) and MZ (IgM⁺CD23⁻) subsets (not depicted). Numbers show percentages of cells falling into gates or quadrants. (B) Mean (±SEM) number of splenic T0, T1 IgD⁺, T2, and MRF B cells in WT (n = 5), Rac1^BRac2^−/− (n = 7), B10.BR, and VavTKO mice (n = 7). (C) Mean (±SEM) migration in a Transwell assay of T0, T1 IgD⁺, and T2 splenic B cells from WT (B10.BR) or VavTKO mice in response to the indicated chemokines (n = 6). (D) Images showing immunofluorescence staining of sections from spleens of mice into which WT (B10.BR) or VavTKO (IgM^b) T0 B cells had been transferred 24 h earlier. Staining for IgM^b (green) identifies transferred T0 B cells, and MadCAM-1 (red) defines the edges of the white pulp. Bar, 150 µm. The graph shows the mean (±SEM) ratio of transferred IgM^b+ T0 B cells ending up in white relative to red splenic pulp 24 h after transfer (n = 4). *, P < 0.05; **, P < 0.01. No Chmk, no chemokine.

Figure 10.

Rac1 and Rac2 are required for chemokine-induced integrin adhesion. (A) Images showing immunofluorescence staining of sections from spleens of mice into which WT, Rac1^B, or Rac2^−/− CFSE-labeled MRF B cells had been transferred 4 h earlier. CFSE (green) identifies transferred B cells, and staining for MadCAM-1 (red) defines the edges of the white pulp. Bar, 300 µm. The graph shows the mean (±SEM) ratio of transferred MRF B cells ending up in white relative to red splenic pulp 4 h after transfer (n = 5). (B) Mean (±SEM) adhesion of splenic MRF B cells from mice of the indicated genotypes to either ICAM-1 or VCAM-1 incorporated into lipid bilayers. B cells were either unstimulated (Unst) or treated with phorbol 12,13-dibutyrate and ionomycin (P+I), CXCL12, CXCL13, or Mn²⁺. (C) Immunoblots showing levels of Rap1-GTP pulled down using a GST-RBD-RalGDS fusion protein, and of total Rap1 in cell lysates of splenic MRF B cells from mice of the indicated genotypes stimulated with CXCL13 for the indicated times. (D) Graph showing mean (±SEM) F-actin content in MRF B cells either unstimulated (−) or stimulated with the indicated chemokines, normalized to the F-actin content of WT unstimulated cells (100%; n = 5). Statistical significance in B and D was determined using an unpaired t test. *, P < 0.05; **, P < 0.01.