Report

Critical Roles for Rac1 and Rac2 GTPases in B Cell Development and Signaling

See allHide authors and affiliations

Science  17 Oct 2003:
Vol. 302, Issue 5644, pp. 459-462
DOI: 10.1126/science.1089709

Abstract

The Rac1 guanosine triphosphatase (GTPase) has been implicated in multiple cellular functions, including actin dynamics, proliferation, apoptosis, adhesion, and migration resulting from signaling by multiple receptors, including the B cell antigen receptor (BCR). We used conditional gene targeting to generate mice with specific Rac1 deficiency in the B cell lineage. In the absence of both Rac1 and the highly related Rac2, B cell development was almost completely blocked. Both GTPases were required to transduce BCR signals leading to proliferation, survival and up-regulation of BAFF-R, a receptor for BAFF, a key survival molecule required for B cell development and maintenance.

B cell development in the bone marrow and spleen proceeds through a series of differentiation stages. Signals from the BCR and the pre-BCR play critical roles in this process (1). The Rac GTPases (Rac1, Rac2, and Rac3) have been implicated in a variety of signaling processes, downstream of multiple receptors (2). However most of these insights have come from studies using cell lines and overexpression of dominant active, or negative GTPases, with unclear specificity. A role for Rac2 in B cell development was demonstrated in targeted mice, which showed defects in marginal zone and peritoneal B1 cells (3). Although recent studies have reported activation of Rac1 after BCR stimulation (4), establishing the function of this ubiquitously expressed GTPase has been hampered by embryonic lethality in gene targeted mice (5). To evaluate the role of Rac1 in B cell development and activation, we generated mice bearing a conditional inactivation of the Rac1 gene in the B cell lineage.

Homologous recombination was used to introduce loxP sites flanking exons 4 and 5 of the Rac1 gene (Rac1flox) (fig. S1, A and B) (6). Mice bearing a deleted allele of Rac1 were also generated (Rac1Δ). Rac1flox/flox mice were crossed to CD19Cre/+ mice with B lineage–specific Cre recombinase expression (7). In the resultant Rac1flox/floxCD19Cre/+ (Rac1B) mice, the Rac1 gene was efficiently deleted in B cells (fig. S1C), resulting in little or no Rac1 protein (fig. S1D). Because Rac1 and Rac2 are highly homologous, a deficiency in one might be compensated by an increase in either the total amount or activation of the other. Analysis of Rac1-deficient B cells showed no change in either total or activated Rac2 (fig. S1, D and E). In contrast, despite unchanged levels of Rac1, Rac2-deficient B cells showed increased BCR-induced Rac1 activation (fig. S1F), suggesting hyperactivation of Rac1 may compensate for the lack of Rac2.

Numbers of B lineage cells in bone marrow, spleen, and lymph nodes of Rac1B mice were similar to those in control mice (Fig. 1, A to D; fig. S2, A to C). To circumvent potential effects of redundancy, Rac1B mice were crossed with Rac2-deficient mice (Rac2–/–) (8). Again, no significant change in pro-B, pre-B, or immature B cell numbers was seen in Rac1BRac2–/– mutant mice (Fig. 1, A and D; fig. S2A), suggesting that Rac1 and Rac2 are not required for the generation and expansion of early B lineage progenitors, though we have not directly determined Rac1 levels in mutant pro-B cells due to their limited availability.

Fig. 1.

Defective B cell development in the absence of Rac1 and Rac2. (A to C) Flow cytometric analysis of cells from Rac1+/+-Rac2+/+CD19Cre/+ (WT), Rac1flox/flox-Rac2+/+CD19Cre/+ (Rac1B), Rac1+/+-Rac2–/–CD19Cre/+ (Rac2–/–), Rac1flox/flox-Rac2–/–CD19Cre/+ (Rac1BRac2–/–) mice, all scatter-gated for lymphocytes. Numbers show percentages of cells falling into marked gates. (A) Expression of IgM and immunoglobulin D (IgD) on CD19+ bone marrow cells. Gates identify pro-B and pre-B cells (IgMIgD), immature B cells (IgM+IgD–/low), and mature recirculating B cells (IgM+IgD+). (B) Expression of CD21 and CD23 on B220+ splenocytes. Gates identify marginal zone cells (CD21+CD23) and all other B lineage cells (CD21CD23 and CD21+CD23+). B cells in Rac1BRac2–/– mice show slightly higher CD23 and lower CD21 levels. The reasons for this are unclear. (C) Expression of CD21 and heat stable antigen (HSA) on B220+ splenocytes gated as in (B) to exclude marginal zone cells. Gates identify transitional type 1 (T1; CD21HSAhigh), transitional type 2 (T2; CD21+HSAhigh) and mature recirculating B cells (CD21+HSAlow) (27). (D) Graph showing mean cell numbers in B cell compartments within bone marrow (two femurs and two tibia, n = 4), spleen, and peritoneal cavity (n = 6). Numbers were determined using the gates shown in (A) to (C) and in fig. S2. Error bars represent standard error of mean (SEM). Significant differences in cell numbers between Rac1BRac2–/– and WT or Rac2–/– mice are indicated: *, P < 0.01. (E) Graph showing mean (±SEM) percentage deletion of the Rac1flox allele from immature bone marrow cells (Imm), splenic T1, and mature recirculating B cells (Mat) sorted from Rac1B and Rac1BRac2–/– mice using gates shown in (A) and (C). Results shown are from five to nine independent measurements. Significant differences are indicated: **, P < 0.001.

In contrast, profound abnormalities were observed at later developmental stages. Rac1BRac2–/– mice had diminished numbers of splenic transitional type 1 and 2 (T1 and T2) and mature recirculating follicular B cells (Fig. 1, B to D; SOM Text 1). Similar reductions were also seen in mature marrow and lymph node B cells (Fig. 1A; fig. S2B). As previously described, splenic marginal zone B cells were reduced in Rac2–/– mice (3), and this was exacerbated in Rac1BRac2–/– mice (Fig. 1, B and D). Immunohistological analysis confirmed an absence of marginal zone B cells in both Rac2–/– and Rac1BRac2–/– mice as well as a severe reduction in mature follicular B cells in Rac1BRac2–/– mice (fig. S3). Finally, peritoneal CD5+ B cells (B1 cells) were reduced in Rac2–/– mice and absent in Rac1BRac2–/– mice (fig. 1D; fig. S2C).

To determine whether the few mature B cells remaining in Rac1BRac2–/– mice were Rac1-deficient, or whether these were a minor subset that had failed to delete Rac1, we measured the efficiency of Rac1 deletion in immature bone marrow B cells, splenic T1, and mature B cells. Rac1 was very efficiently deleted in Rac1B mature B cells, in contrast to Rac1BRac2–/– mice, which showed only 34% deletion (Fig. 1E), implying that few, if any, had a homozygous deletion of Rac1. In view of the efficient deletion of Rac1 in immature Rac1BRac2–/– B cells, these results demonstrate strong counterselection against the deletion of Rac1 in Rac1BRac2–/– but not Rac1B mice and confirms that, in the absence of both GTPases, B cell development is blocked at the T1 stage.

Signals from the BCR are critical for the differentiation of immature B cells to splenic T1 and mature B cells, termed positive selection (1). Thus B lineage cells unable to generate a BCR signal do not emigrate from the bone marrow to the spleen (911). Therefore, we determined whether BCR signaling defects might underlie the developmental block in Rac1BRac2–/– mice and whether Rac1 and Rac2 play a role in BCR-induced B cell activation. Because there are no mature B cells deficient in both GTPases (Fig. 1, D and E), we studied the responses of Rac1Δ/+Rac2–/– and Rac2–/– B cells. BCR-induced proliferation was impaired in Rac2–/– cells and was further reduced in Rac1Δ/+Rac2–/– cells (fig. S4A). Consistent with the role for Rac GTPases in tumor necrosis factor receptor (TNFR)- and toll-like receptor (TLR)–family receptor signaling (12, 13), Rac2–/– and Rac1Δ/+Rac2–/– B cells showed reduced proliferation to signals generated through CD40 and TLR4 (fig. S4B). Mutant cells proliferated normally to phorbol ester and ionomycin, demonstrating no inherent defect in proliferative ability (fig. S4B). Thus, both Rac1 and Rac2 transduce BCR signals, and Rac2 transduces mitogenic signals from CD40 and TLR4.

Because T1 cells do not proliferate in response to BCR stimulation (14), the reduced overall proliferation of Rac1Δ/+Rac2–/– cells could reflect an increased proportion of T1 cells. Therefore, our further studies used sorted mature B cells. Because Rac proteins may control either entry into cell cycle or survival (2), we measured both parameters. In the absence of stimuli, Rac-deficient B cells died at the same rate as wild-type cells (Fig. 2A). In contrast, Rac2–/– and Rac1Δ/+Rac2–/– B cells cultured with antibody to immunoglobulin M (IgM) showed decreased survival. Analysis of cell division showed that the fraction of cells triggered through the BCR to divide (progenitor frequency) as well as the mean number of divisions (burst size) was significantly reduced in both Rac2–/– and Rac1Δ/+Rac2–/– B cells (Fig. 2, B and C; fig. S5B) and that Rac2–/– B cells were arrested at the G0–G1 cell division stage transition (fig. S6, A and B). Together these data show that Rac1 and Rac2 transduce BCR signals required for both cell survival and efficient cell cycle entry.

Fig. 2.

Defective BCR-induced proliferation, survival, and signaling in Rac1- and Rac2-deficient B cells. (A to C) Mature splenic B cells of the indicated genotypes (sorted for B220+CD21+CD23+HSAlow, as in Fig. 1C) were labeled with CFSE (carboxyfluorescein diacetate succinimidyl ester), cultured for the indicated times ±antibody to mouse IgM (anti-IgM) F(ab')2, and stained with TO-PRO3 to identify viable cells. (A) Graphs showing mean (±SEM) percentage viability of mature splenic B cells, determined using a gate as in fig. S5A. Error bars are too small to see. (B) Graph showing mean (±SEM) progenitor frequency in cultures stimulated with anti-IgM F(ab')2. There was no detectable division at 24 hours. (C) Graph showing mean (±SEM) burst size among those viable cells that had divided at least once. (D and E) Immunoblots of total cell lysates from splenic B cells of the indicated genotypes stimulated with anti-IgM F(ab')2 for the indicated times. (D) Immunoblot probed with antiserum to pS473-Akt and reprobed with antiserum to Akt as a loading control. (E) Immunoblots probed with antibodies specific to cyclin D2 or Bcl-xL and reprobed with antiserum to ERK2 as a loading control.

Vav-family proteins are guanine nucleotide exchange factors for Rac1 and Rac2 activated by antigen receptor stimulation (15). In Vav1–/– thymocytes and in Vav3–/– chicken B cells, defects in antigen receptor-driven calcium flux were ascribed, in part, to defects in the activation of Rac GTPases leading to defective phosphoinositide 3-kinase (PI3K) activation (16, 17). In agreement with this, we found that Rac2–/– B cells showed reduced BCR-induced intracellular calcium flux (fig. S7) and a partial reduction in BCR-induced phosphorylation of the Akt kinase (a surrogate for PI3K activation), which was further reduced in Rac1Δ/+Rac2–/– cells (Fig. 2D). Akt has been reported to transduce survival signals, in part through the induction of genes such as Bcl-xL (18). We found decreased induction of Bcl-xL in Rac2 –/– and Rac1Δ/+ Rac2–/– cells, which may, in part, explain their defective survival (Fig. 2E). Finally, we found reduced induction of the cell cycle regulator cyclin D2 in Rac2–/– cells, which was further reduced in Rac1Δ/+Rac2–/– cells (Fig. 2E), consistent with their reduced cell cycle entry.

Defective BCR signaling in Rac1- and Rac2-deficient B cells suggests that Rac1 and Rac2 may transduce BCR signals that control positive selection. Another critical requirement for B cell development and survival results from BAFF binding to one of its receptors, BAFF-R (19). To determine whether defective B cell development in Rac1BRac2–/– mice might reflect compromised BAFF-R signaling, we examined BAFF-induced survival of immature B cells. Although BAFF allowed a considerable proportion of wild-type, Rac1B, and Rac2–/– cells to survive, almost all Rac1BRac2–/– immature B cells died (Fig. 3A; fig. S8A). This defective BAFF-induced survival could be due to compromised BAFF-R signaling, or to the absence of the receptor itself. To address this, we examined the ability of immature B cells to bind BAFF. Although wild-type and Rac1B immature B cells showed similar levels of BAFF binding, this was reduced in Rac2–/– cells and further reduced in Rac1BRac2–/– cells (Fig. 3B; fig. S8B). These studies cannot distinguish which of the three BAFF receptors is depleted, but it is most likely BAFF-R, because only mutations in this receptor cause decreased B cell survival (19) (SOM Text 2). Hence we conclude that the defective BAFF-induced survival of Rac1BRac2–/– immature B cells may be caused by decreased BAFF-R receptor expression.

Fig. 3.

Rac1 and Rac2 are required to transduce BCR signals to the up-regulation of BAFF-R mRNA. (A) Immature bone marrow B cells were cultured in the absence (Medium) or presence of BAFF. Graph shows mean (±SEM) percentage of live cells (identified as in fig. S8A) remaining after 72 hour culture. Each determination was carried out in triplicate. Results representative of three independent experiments. (B) Graph showing mean (±SEM) fluorescence intensity (MFI) of immature bone marrow B cells (gated as in Fig. 1A) stained without (–) and with biotinylated BAFF. Results show mean of five independent measurements. Significant differences in BAFF binding are indicated: *, P <0.01. (C and D) Mature splenic B cells (B220+CD21+HSAlow, Fig. 1C) were cultured for 24 hours in the absence (Medium) or presence of anti-IgM F(ab')2 antiserum and then stained with biotinylated BAFF, to evaluate cell surface levels of BAFF receptors or to determine mRNA levels for BAFF-R. (C) Graph showing MFI (±SEM) as a measure of BAFF binding. Significant differences in BAFF binding are indicated: **, P =0.029. (D) Graph showing mean (±SEM) fold induction of BAFF-R mRNA in mature splenic B cells after treatment with anti-IgM relative to cells cultured in medium alone. A fold induction of 1 indicates no induction.

Recent studies indicate that BCR stimulation up-regulates BAFF-R messenger RNA (mRNA) (20). Taken together with our results demonstrating defective BCR signaling in Rac-deficient mature B cells as well as reduced BAFF receptor expression and compromised positive selection in Rac1BRac2–/– immature B cells, we hypothesize that in immature B cells the BCR positive selection signal may induce BAFF-R up-regulation and that this occurs via Rac1 and Rac2. Consistent with this, Syk-deficient immature B cells, which are also blocked at the immature B cell stage due to an inability to transduce the BCR positive selection signal (11), showed a significant reduction in BAFF receptor levels (fig. S9, A and B). To directly examine whether Rac1 and Rac2 transduce BCR signals to the up-regulation of BAFF-R expression, we measured BAFF binding capacity and the induction of BAFF-R mRNA in BCR-stimulated B cells. Though wild-type cells showed a robust increase in both BAFF binding and BAFF-R mRNA, this was partially decreased in Rac2–/– cells and further reduced in Rac1Δ/+Rac2–/– cells (Fig. 3, C and D; fig. S10). In contrast, mRNA levels for TACI, a distinct BAFF receptor, remained unchanged in all genotypes (21). Thus we conclude that in mature B cells, Rac1 and Rac2 transduce BCR signals for selective up-regulation of BAFF-R mRNA.

Mice with a mutation in BAFF-R show a 40% reduction in the number of T1 cells and severely decreased numbers of T2, mature follicular B cells, and marginal zone B cells, but normal numbers of B1 cells (2225). Thus defective BAFF-R expression cannot fully account for the developmental block at the immature–T1 transition and the absence of B1 cells in Rac1BRac2–/– mice. It is more likely that these defects are due to compromised BCR signaling, because mutations that disrupt BCR signaling typically result in reduced B1 cell numbers (26) and failure to synthesize a BCR or to signal from it eliminates all splenic T1 cells (911). Finally, our results also show that there are notable differences in the phenotypes of Rac1BRac2–/– and Vav1–/–Vav2–/– mice (15), suggesting that in B cells these GTPases may be regulated by other exchange factors such as Vav3.

In conclusion, we have shown that Rac1 and Rac2 GTPases play critical overlapping roles in B cell development and BCR-induced activation. In mice deficient for both proteins, B cell development is arrested at the immature–T1 transition, consistent with defective BCR signaling. Furthermore, we have shown that Rac1 and Rac2 transduce BCR signals that control survival, cell cycle entry and up-regulation of BAFF-R mRNA.

Supporting Online Material

www.sciencemag.org/cgi/content/full/302/5644/459/DC1

Materials and Methods

SOM Text

Figs. S1 to S10

References

References and Notes

View Abstract

Navigate This Article