Research ArticleHUMORAL IMMUNITY

R-Ras2 is required for germinal center formation to aid B cells during energetically demanding processes

See allHide authors and affiliations

Science Signaling  29 May 2018:
Vol. 11, Issue 532, eaal1506
DOI: 10.1126/scisignal.aal1506

R-Ras2 revs replication

Germinal centers are specialized lymphoid structures that facilitate the development of high-affinity antibody responses by B cells, a process that requires the guanosine triphosphate hydrolase activity of Ras family members. Mendoza et al. characterized germinal center formation in mice lacking R-Ras2 or other related family members and found that Ras proteins are functionally specialized. Only B cell–intrinsic R-Ras2 activity was necessary for cellular proliferation and germinal center development. R-Ras2 deficiency specifically reduced the expression of mitochondrial tRNAs necessary for mitochondrial replication and glucose metabolism in B cells. These data suggest that R-Ras2 signaling is essential for the metabolic reprograming of germinal center B cells to supply the energy required to generate effective antibody responses.

Abstract

Upon antigen recognition within peripheral lymphoid organs, B cells interact with T cells and other immune cells to transiently form morphological structures called germinal centers (GCs), which are required for B cell clonal expansion, immunoglobulin class switching, and affinity maturation. This process, known as the GC response, is an energetically demanding process that requires the metabolic reprogramming of B cells. We showed that the Ras-related guanosine triphosphate hydrolase (GTPase) R-Ras2 (also known as TC21) plays an essential, nonredundant, and B cell–intrinsic role in the GC response. Both the conversion of B cells into GC B cells and their expansion were impaired in mice lacking R-Ras2, but not in those lacking a highly related R-Ras subfamily member or both the classic H-Ras and N-Ras GTPases. In the absence of R-Ras2, activated B cells did not exhibit increased oxidative phosphorylation or aerobic glycolysis. We showed that R-Ras2 was an effector of both the B cell receptor (BCR) and CD40 and that, in its absence, B cells exhibited impaired activation of the PI3K-Akt-mTORC1 pathway, reduced mitochondrial DNA replication, and decreased expression of genes involved in glucose metabolism. Because most human B cell lymphomas originate from GC B cells or B cells that have undergone the GC response, our data suggest that R-Ras2 may also regulate metabolism in B cell malignancies.

INTRODUCTION

Germinal centers (GCs) are temporary structures formed in secondary lymphoid organs where B cells undergo clonal expansion, immunoglobulin (Ig) class switching, and affinity maturation. Among other cells types, GCs contain B cells, a specialized subset of dendritic cells (DCs) termed follicular DCs (FDCs), and a specialized subset of helper T cells termed T follicular helper (TFH) cells. GCs are organized into two compartments, namely, the dark zone (DZ) and the light zone (LZ). In the DZ, B cells termed centroblasts rapidly proliferate and undergo somatic hypermutation of their Ig-encoding variable genes. B cells of the LZ, termed centrocytes, undergo Ig class switching. During the GC response, those B cell clones expressing the highest-affinity antibodies are selected on the basis of their competitive advantage over others expressing low-affinity antibodies to receive signals from the B cell receptor (BCR) and the receptor CD40. These signals come from FDCs, which activate the BCRs, and TFH cells, which activate CD40 (1, 2). B cells present in GCs express a panel of characteristic markers and are referred to as GC B cells.

The rapid proliferation of B cells during the GC response requires metabolic reprogramming to generate building blocks for the synthesis of new macromolecules, as well as to meet increasing energy demands. Unlike antigen-stimulated T cells, which disproportionately increase lactate production over oxygen consumption, B cell activation results in a proportional increase of both glycolysis and oxidative phosphorylation (35). The control of metabolism in activated B cells during the GC reaction is largely unknown beyond a few pioneering reports pointing to the nuclear factor κB (NF-κB) subunit c-Rel and the regulator of mRNA splicing and stability HuR as key modulators of metabolism (6, 7). The phosphoinositide 3-kinase (PI3K) and Akt (PI3K-Akt) pathway plays a fundamental role in the control of cell growth and metabolism (8) and is crucial for GC B cells to cycle between the LZ and the DZ (9). However, the relevance of this pathway to B cell metabolism in the GC and how it is controlled by plasma membrane receptors have not yet been explored.

Ras proteins are upstream regulators of the PI3K-Akt pathway; when activated, they recruit and activate the catalytic subunit of type Ia PI3Ks. In addition, Ras proteins regulate both glycolytic and mitochondrial metabolism (10), making these proteins candidates for the transduction of signals that regulate metabolism in B cells in the GC. The Ras family of small guanosine triphosphatases (GTPases) is made up of 39 genes in the human genome (11). The classic Ras subfamily (Hras, Kras, and Nras) has been investigated intensively, because its members are frequently mutated to constitutively active (oncogenic) forms in human cancer. The lesser-known Ras-related (R-Ras) subfamily proteins show a 55 to 60% overall amino acid identity with the classical Ras members (12, 13). This subfamily is also composed of three genes: Rras (also known as and hereafter referred to as Rras1), Rras2 (also known as TC21), and Rras3. Rras2 displays transforming activities similar to those exhibited by the classic Ras GTPases (1419). In addition, R-Ras proteins share guanine nucleotide exchange factors (GEFs) (20) and GTPase-activating proteins (GAPs) (21) with classic Ras proteins, suggesting that R-Ras and classic Ras proteins are activated by similar mechanisms and signals. Thus, R-Ras proteins could perform many of the functions initially ascribed to classical Ras proteins in the immune system.

Most pioneering studies on Ras protein function involved expressing dominant-negative or constitutively active mutants of classical Ras proteins. These types of experiments indicated that the classical Ras proteins play a fundamental role in T and B cell development or activity (22, 23). However, the study of Ras-deficient mice has produced different results, and neither single- nor double-mutant mice deficient in Hras and Nras display an immune-related phenotype (24). These results support the idea of functional specialization of the classic Ras proteins in different immunological processes, suggesting a differential regulation of signaling pathways, which is also reflected in mouse fibroblasts (25). Consistent with the idea of functional specialization, studies on mice deficient in R-Ras proteins implicate these GTPases in different processes. Thus, Rras1−/− mice mount inefficient immune responses to tumors because of an impaired ability of DCs to prime allogenic and antigen-specific T cell responses (26). In addition, T cells lacking R-Ras1 show defects in trafficking in high endothelial venules during an effective immune response (27). R-Ras2 also plays important physiological roles in tonic T cell receptor (TCR) and BCR signaling, and this protein is constitutively associated with resting antigen receptors (28). In addition, R-Ras2 is essential for T cell trogocytosis, a phagocytic-like mechanism that promotes the internalization of TCR bound to antigen presented on major histocompatibility complex (MHC) from the immunological synapse (29).

Immunization of Rras2−/− mice with T-dependent antigens, which require T helper cells to stimulate antibody production by B cells, induces a poor humoral response, whereas the response to type I and type II T-independent antigens, which do not require T helper cells to stimulate antibody production, was not inhibited (28). The impaired T-dependent humoral response correlates with the formation of fewer and smaller GCs, as detected by immunohistochemistry using peanut agglutinin lectin staining. Here, we performed in-depth studies regarding the requirement for mechanisms regulating R-Ras2 and its redundancy with other Ras GTPases in the GC response. We provide evidence that R-Ras2, and not R-Ras1, H-Ras, or N-Ras, played an essential role in the GC response. This role was B cell–intrinsic, and the compromised activation of B cells in GCs correlated with deficient PI3K-Akt pathway activation by two receptors on B cells: the BCR and CD40. We further showed that R-Ras2 deficiency impaired energy production by affecting both mitochondrial function and aerobic glycolysis. Thus, R-Ras2 enables a connection between key membrane receptors and metabolism in activated and GC B cells.

RESULTS

R-Ras2 plays an essential, nonredundant role in the formation of GCs

The homology between R-Ras and classic Ras small GTPases is particularly evident in the nucleotide binding loops (G1–G5) and in the switch I and switch II effector loops, which are essentially indistinguishable (fig. S1A). This sequence identity could reflect a functional redundancy among the six members of both these subfamilies. Although Rras2 is the most highly expressed GTPase at the mRNA level in marginal zone B cells, follicular B cells, and GC B cells, other members of both subfamilies are also expressed (fig. S1, B and C). Therefore, we wondered whether the GC role of R-Ras2 was specific to this GTPase or whether it was also fulfilled by other members of the R-Ras or classic Ras subfamilies. Therefore, we generated a germ-line knockout mouse of Rras (hereafter referred to as Rras1), which is the most closely related to Rras2 of the R-ras and classic ras genes. Exons 2 to 6 of Rras1, which encode most of the protein, were deleted, and we failed to detect Rras1 expression in lymphoid organs from these mice by quantitative polymerase chain reaction (qPCR) and immunoblotting (fig. S1, D and E). Analysis of B and T cell populations in the spleen and lymph nodes showed that Rras1−/− mice had a 30 to 40% reduction in the total number of T cells compared to their wild-type (WT) Rras1+/+ counterparts (fig. S1, F to H). However, unlike Rras2−/− mice (28), Rras1−/− mice had normal numbers of B cells in these tissues (Fig. 1A and fig. S1F). Double-mutant mice (Rras1−/−Rras2−/−) reflected a similar phenotype to the Rras2−/− mice, indicating that these GTPases play independent roles in controlling the size of lymphoid cell populations (Fig. 1A and fig. S1H). Because we did not detect Rras3 expression in lymphoid cells (fig. S1B), our data indicated that R-Ras2 is the only Ras-related GTPase with a specific nonredundant role in B cell homeostasis.

Fig. 1 Nonredundant role of R-Ras2 in GC formation.

(A) Flow cytometry analysis of the frequency of marginal zone (MZ) phenotype (CD21highCD23low) or follicular (CD21lowCD23high) B cells within B220+ splenic B cells from mice of the indicated genotype. Representative two-color contour plots (top) and means of the total number of each B cell population ± SEM (bottom) are from five mice per group. (B) Flow cytometry analysis of the frequency of GL7+CD95+ GC B cells within B220+ splenic B cells from mice of the indicated genotype at 7 days after SRBC immunization. Representative two-color contour plots (left) and means of the GC B cell percentage ± SEM (right) are from at least four mice per group. (C) Flow cytometry analysis of the percentage of GC B cells in Peyer’s patches from nonimmunized mice. Representative two-color contour plots (left) and means of the GC B cell percentage ± SEM (right) are from at least two mice per group. (D) Flow cytometry analysis of the frequency of PD-1+CXCR5+ TFH within CD4+B220 splenic T cells from mice of the indicated genotype at 7 days after SRBC immunization. Representative two-color contour plots (left) and means of the TFH percentage ± SEM (right) are from at least two mice per group. All data are representative of at least three independent experiments. One-way analysis of variance (ANOVA) was applied to compare the different groups. Dunnett’s multiple comparison test was used to derive P values. *P < 0.05, **P < 0.005 by two-tailed Student’s t test. ns, not significant; p.i., postinjection.

To confirm previous results showing that the absence of R-Ras2 gives rise to defects in the GC response (28) and to determine whether these defects are specific to this GTPase or whether similar defects are caused by deficiencies in R-Ras or classical Ras proteins, we studied the formation of GCs and B cells in these structures in Rras1−/−, Rras2−/−, and Rras1−/−Rras2−/− mice. We also analyzed mice that lacked two classic Ras proteins, Hras−/−Nras−/− (fig. S1, B and C). All mice were immunized with sheep red blood cells (SRBCs) to induce GC formation. We stained peripheral lymphoid tissues, spleen, and Peyer’s patches with anti-CD95 and anti-GL7 to identify the GC B cells within the general B220+ B cell population, and quantified the number of these B cells by flow cytometry. We detected a 2.5-fold reduction in the number of GC B cells in mice deficient in R-Ras2 (Fig. 1B and fig. S1I). This defect was specific to R-Ras2, because the numbers of GC B cells in R-Ras1–deficient and HRas-NRas double-mutant mice were not significantly different from those in WT mice. In addition, the loss of GC B cells in mice deficient in both R-Ras1 and R-Ras2 was similar to that of the Rras2−/− mice (Fig. 1B and fig. S1I), indicating that R-Ras1 is not functionally redundant with R-Ras2. We also investigated GC B cells in Peyer’s patches, which are lymphoid organs associated with the small intestine and normally contain abundant GCs, probably due to repeated exposure to intestinal antigens (30). There were significantly fewer GC B cells in R-Ras2–deficient and R-Ras1–R-Ras2 double–deficient mice than in WT or R-Ras1–deficient mice (Fig. 1C), confirming that R-Ras2 plays a nonredundant and essential role in the formation of GC B cells.

A subset of CD4+ T cells, referred to as TFH cells, that are positive for the transcription factor Bcl6 and the surface markers CXCR5 and programmed cell death protein 1 (PD-1) provide survival signals to GC B cells. TFH cells secrete cytokines that modulate Ig class switching and produce the ligand CD40L, which engages CD40 on B cells (31). However, the generation of TFH cells was comparable in Rras2−/− and WT mice (Fig. 1D and fig. S1I). Likewise, other Ras deficiencies (R-Ras1, H-Ras, and N-Ras) did not significantly affect TFH cell numbers. Furthermore, R-Ras2–deficient TFH cells had normal amounts of TFH proteins, including CD40L and inducible T cell costimulator (ICOS) (fig. S2A). Similarly, the abundance of the surface proteins CD40, ICOS ligand (ICOSL), and interleukin-21 receptor (IL-21R) on R-Ras2–deficient GC B cells was indistinguishable from that in WT and R-Ras1–deficient mice (fig. S2B). Together, these data suggested that the deficient GC B cell response of Rras2−/− mice is not due to anomalies in the number or phenotype of TFH cells, but rather that it may derive from an intrinsic B cell defect.

The role of R-Ras2 in GC B cell formation is B cell–intrinsic

To determine whether R-Ras2 is required for the formation of antigen-specific GC B cells, we used B1-8hi knockin mice (32), which have a recombinant heavy chain antibody variable region derived from a 4-hidroxy-3-nitrophenylacetyl (NP)–binding antibody (B1-8). These mice were immunized with an NP-modified chicken gamma globulin (NP-CGG), through which we tracked the formation of antigen-specific B cells. During activation in the GCs, B cells undergo class switching in which the antibody heavy chain switches from IgM to other classes (IgG1, IgG2a, IgG2b, IgG3, IgA, and IgE in mice). Compared to WT controls at day 7 after immunization, R-Ras2 deficiency impaired the formation of NP-binding class-switched IgG1+ B cells with GC markers (GL7 and CD95) both in spleen and lymph nodes (Fig. 2A). Serum collected from those mice at days 7 and 14 revealed a deficiency in the production of all class-switched high-affinity IgGs reacting with a low valence–substituted NP-ovalbumin (Fig. 2B). To confirm the defect of Rras2−/− mice in the generation of antigen-specific GC B cells, we also immunized nontransgenic mice with NP-CGG and determined the generation of NP antigen-specific GC and memory B cells according to the presence of IgG1 as a mature class-switched Ig and the abundance of CD38, which is down-regulated in GC B cells (33). The formation of GC and memory B cells was impaired in Rras2−/− mice when compared to WT or Rras1−/− mice (Fig. 2C). The total number of antigen-specific class-switched B cells [B cells (NIP+IgG1+); Fig. 2C] was significantly lower in these Rras2−/− mice than in the other two genotypes. Because these cells come from GC B cells, the reduction in NIP+IgG1+ B cells further supported the hypothesis that R-Ras2 is selectively required for GC B cell formation. The analysis of antibodies in the serum showed that R-Ras2 deficiency did not impair the generation of NP-specific IgMs of low affinity but inhibited the generation of class-switched IgGs of low and high affinity (Fig. 2D), confirming that the reduction in antigen-specific class-switched B cells resulted in a functional deficiency in the antigen response.

Fig. 2 B cell–intrinsic role of R-Ras2 in the GC response.

(A) Flow cytometry analysis of the percentage of GC B cells (GL7+CD95+) within CD19+NP+ B cells in Rras2−/− or WT transgenic B1-8hi mice at 7 days after immunization with NP-CGG. Representative two-color contour plots (left) and means of the TFH cells percentage ± SEM (right) are from at least three mice per group. (B) Enzyme-linked immunosorbent assay (ELISA) analysis of NP-ovalbumin–reactive antibodies in sera taken from Rras2−/− or WT B1-8hi mice at 7 and 14 days p.i. with NP-CGG. Data are means ± SEM from at least two mice per group. (C) Flow cytometry analysis of the frequency of NP antigen-specific IgG1+ GC (B220+NIP+IgG1+CD38) and memory (B220+NIP+IgG1+CD38+) B cells at 12 days after immunization with NP-CGG. Representative two-color contour plots (left) and means of TFH cells percentage ± SEM (right) are from at least two mice per group. (D) ELISA analysis of NP-reactive antibodies IgM, IgG, and IgG1 in sera from mice of the indicated genotypes at different time points after immunization with NP-CGG. Data are means ± SEM from at least two mice per group. (E) Flow cytometry analysis of the frequency of donor CD45.2+ Rras2−/− or WT B cells in host CD45.1+ mice at the indicated time points after immunization with SRBC. Representative two-color dot plots (left) and means of the total numbers of donor cells ± SEM (right) are from at least four mice per group. (F) Flow cytometry analysis of the frequency of CD95+GL7+ GC B cells within donor CD45.2+ Rras2−/− or WT B cells in host CD45.1+ mice at the indicated time points after immunization with SRBC. Representative two-color dot plots (left) and means of total numbers of donor (CD45.2+) and host (CD45.2) cells ± SEM (right) are from at least four mice per group. All data are representative of at least three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by two-tailed Student’s t test.

Although the absence of anomalies in TFH cells (Fig. 1D) suggested that GC B cell deficiency in Rras2−/− mice is not secondary to a T cell defect, we could not yet rule out an influence of other cell types affected by R-Ras2 deficiency. To determine whether R-Ras2 is required for GC B cell formation in a B cell–intrinsic manner, we performed adoptive transfer experiments in which B cells bearing the marker CD45.2 were purified from either Rras2−/− or WT mice and transferred into CD45.1+ WT recipient mice. These mice were subsequently immunized with NP-CGG, and the differentiation of the transferred B cells into GC B cells was analyzed by tracking the CD45.2+ population. Because we previously encountered a defect in the survival of R-Ras2–deficient B cells 21 days after transfer into WT mice (28), we first examined whether B cell survival was affected by R-Ras2 deficiency in the shorter adoptive transfer experiments (4 to 10 days). We found no significant effect of R-Ras2 deficiency on the number of recovered CD45.2+ B cells through day 10 (Fig. 2E). In contrast, there was a threefold reduction in the number of GC B cells (GL7+CD95+) in the absence of R-Ras2 both at 7 and 10 days after stimulation (Fig. 2F). We monitored the endogenous CD45.2 GC B cell population as an internal immunization control and found no differences regardless of the genotype of transferred B cells (Fig. 2F). Because the inoculated CD45.2+ Rras2−/− B cells were in a WT context, these results suggested that the role of R-Ras2 in GC B cell formation is B cell–intrinsic.

R-Ras2–deficient GC B cells are predominantly DZ B cells

Exploration of the temporal aspect of the defect in the formation of GC B cells in response to R-Ras2 deficiency could provide insight into whether the defect was specific to the LZ- or DZ-mediated events in the GC response. Therefore, we immunized WT and Rras2−/− mice with SRBC and quantified the absolute number of total B cells and GC B cells at days 7 and 14 after immunization. We found that the number of R-Ras2–deficient total B cells increased between days 7 and 14, which is similar to the expansion that occurred with cells in WT mice (Fig. 3A). However, by day 7, the number of GC B cells was already reduced in Rras2−/− compared to WT mice, and these cells did not expand from days 7 to 14 as the GC B cells in the WT mice did (Fig. 3A). These data suggested that R-Ras2 is required for proliferation of B cells during the GC response and their expansion in vivo (34). To address whether there is a defect in proliferation, we performed cell cycle analysis of GC B cells (B220+GL7+CD38) at days 7 and 14 after immunization. At day 7, the total GC B cell population was partially arrested at the G2-M stage in Rras2−/− mice, concomitant with a reduction in the percentage of cells in the S phase (Fig. 3B).

Fig. 3 R-Ras2–deficient GC B cells predominantly present a DZ B cell phenotype.

(A) The total number of splenic total B (B220+) and GC B cells (B220+GL7+CD95+) from SRBC-immunized WT and Rras2−/− mice at the indicated time points was determined by flow cytometry. Data are means ± SEM and are from at least three mice per group. (B) Flow cytometry analysis of cell cycle phase in GC B cells (B220+GL7+CD38) from WT and Rras2−/− mice at the indicated times after immunization with SRBC. Representative two-color dot plots (left) and means of the frequency of GC cells in apoptosis (7AAD), G0-G1 (7AADloBrdU), G2-M (7AAD+BrdUlo), and S phase (BrdU+) ± SEM (right) are from at least four mice per group. (C) Flow cytometry analysis of the frequency of centroblasts (CXCR4+CD86) and centrocytes (CXCR4CD86+) within GC B cells in WT and Rras2−/− mice at the indicated times after SRBC immunization. Two-color dot plots (left) are representative of at least seven mice. Quantification of the frequency of centroblasts, centrocytes, and their relative ratio are means ± SEM (right) from three independent experiments. (D) Flow cytometry analysis of cell cycle phase in CXCR4+CD86 centroblasts (left) and CXCR4CD86+ centrocytes (right) within B220+GL7+ CD38 GC B cells from WT and Rras2−/− mice at the indicated times after immunization with SRBC. Quantification of the frequency of cells in apoptosis (7AAD), G0-G1 (7AADloBrdU), G2-M (7AAD+BrdUlo), and S phase (BrdU+) are means ± SEM (right) from at least three mice per group. Data in (A), (B), and (D) are representative of at least three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by two-tailed Student’s t test.

GC B cell proliferation is believed to start at the LZ stage (centrocytes) when GC B cells receive stimuli through their BCR and CD40 receptors, whereas proliferation at the DZ stage (centroblasts) is independent of these receptors (35). To determine whether R-Ras2 deficiency altered the distribution of centroblasts and centrocytes in GC B cells, we used the abundance of CXCR4 and CD86 to define these two populations: Centroblasts are CXCR4+CD86, and centrocytes are CXCR4CD86+. We analyzed centroblasts and centrocytes in the population of B cells from the spleens of SRBC-immunized WT and Rras2−/− mice. The overall numbers of both populations of GC B cells were reduced in the absence of R-Ras2 at both day 7 and day 14 after immunization, but R-Ras2 deficiency was associated with a higher centroblast/centrocyte ratio than in the WT mice (Fig. 3C). Hence, GC B cells were mainly found as centroblasts in the absence of R-Ras2 at day 7 after SRBC immunization.

We performed cell cycle analyses and detected apoptotic cells in the GC centroblast and centrocyte populations at day 7 and day 14 after immunization of WT and Rras2−/− mice. At day 7, we found fewer GC B cells in the S phase in Rras2−/− mice than in WT mice, although this was most pronounced in the centrocyte population (Fig. 3D). In addition, in Rras2−/− mice, both centrocytes and centroblasts accumulated in the G0-G1 phase at day 7 after immunization. GC B cells receive antigen and CD40 signals that make them enter the S phase at the centrocyte stage, and from this stage, they enter the centroblast stage in which they engage several cycles of cell division independent of stimuli (36). Thus, our results suggest that Rras2−/− GC B cells may receive insufficient proliferative signals at the centrocyte stage. Because the number of GC B cells was already reduced by day 14 after SRBC immunization in Rras2−/− mice (Fig. 3A and fig. S2C), we noted fewer differences in centroblast and centrocyte populations between WT and Rras2−/− mice (Fig. 3C). However, the proportion of apoptotic centroblasts was increased in the absence of R-Ras2. (Fig. 3D). Together, these results reinforce the role of R-Ras2 in the GC reaction and more precisely at the centrocyte stage, when cells receive stimuli through the BCR and CD40.

R-Ras2 is required for GC B cell proliferation and PI3K signaling in response to BCR and CD40

Because GC B cells receive stimuli through both the BCR and CD40 in the LZ, we determined whether B cells proliferated in response to these stimuli in the absence of R-Ras2. Purified naïve B cells proliferated in response to either anti-IgM or with anti-CD40. In the absence of R-Ras2, but not R-Ras1, B cell proliferation in response to either of these stimuli was reduced (Fig. 4A), suggesting that R-Ras2 functions downstream of both these receptors. R-Ras2 is required for tonic PI3K activation but not for extracellular signal–regulated kinase (ERK) activation downstream of BCR activation (28). Activation of PI3K is required for BCR-triggered B cell proliferation (37), and R-Ras2 recruits and activates the catalytic p110δ subunit of PI3K, which plays an essential role in GC formation (38, 39). Therefore, we examined whether deficient activation of PI3K was responsible for the impaired proliferative response of naïve B cells to BCR or CD40 stimulation. Akt is a PI3K-dependent effector that is recruited to the plasma membrane by binding to phosphatidylinositol 3,4,5-trisphosphate (PIP3). At the membrane, Akt1 is phosphorylated at Thr308 by PDK1, which is a PIP3-dependent kinase, and at Ser473 by the kinase in the complex mTORC2. In turn, Akt activates the kinase complex mTORC1, which phosphorylates the p70S6 kinase, and this protein further phosphorylates the ribosomal protein S6. Using phosphorylation-specific antibodies, we found that R-Ras2 deficiency decreased BCR-induced phosphorylation of PDK1 at an activating site (Ser241), of Akt at Thr308 and Ser473, of p70S6K at Thr389, and of S6 at Ser240/244, but not ERK phosphorylation by the upstream kinase mitogen-activated protein kinase (MEK) at Thr202 and Tyr204 (Fig. 4B and fig. S3A). R-Ras2 deficiency impaired the phosphorylation not only of Akt, PDK1, and S6 but also of ERK in response to stimulation with antibody against CD40 (Fig. 4C and fig. S3A), suggesting that R-Ras2 plays a role in the activation of both the PI3K/Akt and ERK pathways by CD40 in naïve B cells.

Fig. 4 Defective proliferative response and activation of the PI3K-Akt pathway by BCR or CD40 in the absence of R-Ras2.

(A) Proliferation of purified naïve B cells of the indicated genotype stimulated for 3 days with anti-IgM or anti-CD40 was measured by 3H-thymidine incorporation. Data are means ± SEM of at least three mice and representative of three independent experiments. cpm, counts per minute. (B and C) Western blot analysis of phosphorylated PDK1, Akt, p70S6K, and S6 at the indicated time points on lysates from purified naïve B cells stimulated with either anti-IgM (B) or anti-CD40 (C). Representative blots (left) and mean quantified densitometry data ± SEM (right) are from three independent experiments. (D) Coimmunoprecipitation of Rras2 with anti-CD40 from lysates of purified total splenic B cells from WT mice and treated with or without anti-CD40 for 5 min. Blots are representative of at least three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by two-tailed Student’s t test.

R-Ras2 is recruited to the antigen receptors of both B cells and T cells, the BCR and TCR, respectively, in a constitutive manner and functions as a direct effector of these receptors (28). Because our data indicated that R-Ras2 was also downstream of CD40 in B cells, we immunoprecipitated CD40 from B cells with and without CD40 stimulation and evaluated whether R-Ras2 coprecipitated with this receptor was present. R-Ras2 coprecipitated with CD40 independent of triggering (Fig. 4D). This result suggested that a direct or indirect association of R-Ras2 with membrane receptors might be a common theme in the mechanism of action of this GTPase. To determine whether similar signaling defects occurred in GC B cells, we evaluated phosphorylation of Akt and ERK by flow cytometry (fig. S3B) after B cell stimulation with either anti-IgM or anti-CD40 (Fig. 5A). Similar to naïve B cells, R-Ras2 but not R-Ras1 was required for BCR-triggered phosphorylation of Akt and for CD40-triggered phosphorylation of Akt and ERK in GC B cells. Together, these data indicated that, downstream of both the BCR and CD40, R-Ras2 stimulates PI3K/Akt activation in B cells, including GC B cells, and that B cells from mice deficient in R-Ras2 have reduced PI3K activity compared to cells from WT and Rras1−/− mice. We tested whether PI3K activity was also defective in both centrocyte and centroblast cell populations in the absence of R-Ras2. Phosphorylation-specific flow cytometry analysis within these two populations showed that both centrocytes and centroblasts had less p70S6K activity when R-Ras2 expression was lost (Fig. 5B).

Fig. 5 Defective activation of the PI3K-Akt pathway in GC B cells in the absence of R-Ras2.

(A) Flow cytometry analysis of Akt and ERK phosphorylation after either anti-IgM or anti-CD40 stimulation of GL7+CD95+ GC B cells from WT and Rras2−/− mice at 7 days after SRBC immunization. Representative histogram plots (left) and mean fluorescence intensity (MFI) ± SEM are from three mice per group. (B) Flow cytometry analysis of p70S6K phosphorylation after either anti-IgM stimulation of total GC B cells (B220+CD95+CD38), centroblasts (B220+CD95+CD38CXCR4+CD86), or centrocytes (B220+CD95+CD38CXCR4CD86+) from WT and Rras2−/− mice at 7 days after immunization with SRBC. Representative histogram plots (left) and MFI ± SEM are from three mice per group. (C) Flow cytometry analysis of FoxO1 phosphorylation after either anti-IgM or anti-CD40 stimulation of total GC B cells (B220+CD95+GL7+) from WT and Rras2−/− mice at 7 days after immunization with SRBC. Representative histogram plots (left) and MFI ± SEM are from three mice per group. (D) Flow cytometry analysis of total FoxO1 abundance in total GC B cells (B220+CD95+CD38) from WT and Rras2−/− mice at 7 days after immunization with SRBC. Shaded histograms show FoxO1 levels in nonstimulated cells. Dashed histograms show FoxO1 levels in IgM-stimulated cells. Representative histogram plots (left) and normalized MFI ± SEM are from three mice per group. (E) Reverse transcription (RT)–qPCR for the expression of the indicated FoxO1 target genes relative to hypoxanthine phosphoribosyltransferase control in FACS (fluorescence-activated cell sorting)–sorted GC B cells (B220+CD95+GL7+) from WT and Rras2−/− mice at day 7 after SRBC immunization. All data are representative of at least three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by two-tailed paired Student’s t test.

FoxO1 is a target of Akt that is important in B cells because this transcription factor promotes the expression of target genes involved in cell cycle arrest, apoptosis, and gluconeogenesis (40). Phosphorylation of FoxO1 at Thr24 by Akt promotes its proteasomal degradation, resulting in blockade of FOXO1-mediated transcription of target genes involved in cell cycle arrest and apoptosis. After BCR or CD40 stimulation, we found that, concomitant with decreased PI3K/Akt activity, FoxO1 phosphorylation (Fig. 5C) and subsequent degradation (Fig. 5D) were significantly decreased in R-Ras2–deficient GC B cells, compared to cells from WT controls.

PI3K and FoxO1 have reciprocal roles in the maintenance of the LZ and DZ, respectively (9, 41); FoxO1 is required for B cell proliferation in the DZ, whereas PI3K down-regulates FoxO1 activity in the LZ. Because PI3K activity opposes FoxO1 stability and R-Ras2 activates PI3K, our results appear consistent with a model in which R-Ras2 deficiency may lead to increased FoxO1 activity due to insufficient PI3K activity after BCR and CD40 stimulation. Because the FoxO1 gene promoter has a site for regulation by FoxO transcription factors, including FoxO1 itself (42), we predicted that increased FoxO1 stability could increase FoxO1 transcriptional activity at its own promoter in the absence of R-Ras2. When we examined B cells from immunized WT mice, we found that the expression of FoxO1 mRNA was greater in endogenous GC B cells than in follicular B cells of mice at 7 days after SRBC injection, and this difference in transcript expression correlated with greater FoxO1 protein abundance in GC B cells (fig. S3C, D and E). Furthermore, although FoxO1 mRNA expression in R-Ras2–deficient GC cells was decreased when compared to WT cells (fig. S3D), the FoxO1 protein abundance was similar (fig. S3E). Thus, we found that R-Ras2 deficiency only affected a subset of the FoxO1 target genes in endogenous GC B cells (Fig. 5E), with the gene encoding IL-1β showing the greatest increase in expression in the absence of R-Ras2. Because it seemed unlikely that this limited effect on FoxO1 target genes was responsible for the GC phenotypes found in R-Ras2–deficient mice, we analyzed global transcriptional differences between WT and Rras2-deficient GC B cells and metabolic characteristics of GC B cells.

R-Ras2 controls mitochondrial function in B cells

To examine the requirement for R-Ras2 on gene transcription in GC B cells in a global manner, we performed whole-genome microarray analysis on sorted GC B cells from WT and Rras2−/− mice. Although analysis of gene expression data by significance analysis of microarrays (SAM) showed that the expression of only nine transcripts was altered by the loss of R-Ras2 [Fig. 6A and Gene Expression Omnibus (GEO) GSE113599], gene set enrichment analysis (GSEA) suggested an increased expression of genes involved in the differentiation of GC B cells (Fig. 6B) and the G1-to-S cell cycle transition in WT when compared to R-Ras2–deficient GC B cells (fig. S4). This enrichment matched our results from the earlier phenotypic studies (Fig. 3B). Other gene signatures affected by R-Ras2 deficiency related to the tricarboxylic acid (TCA) cycle, glucose metabolism, and cellular respiration (fig. S4), suggesting a metabolic defect in R-Ras2–deficient GC B cells. Of the genes altered by the absence of R-Ras2 (Fig. 6C and data file S1), transcripts for four mitochondrial DNA (mtDNA)–encoded tRNAs were down-regulated compared to WT cells (Fig. 6C). Mitochondrial tRNA down-regulation was confirmed by qPCR (Fig. 6D). These results suggested that R-Ras2 deficiency affected mtDNA replication or mitochondrial transcription in B cells and likely affects cell metabolism.

Fig. 6 Impaired mitochondrial function in GC B cells in the absence of R-Ras2.

(A to C) Microarray analysis of the transcriptomes from WT and R-Ras2–deficient GC B cells isolated at day 7 after SRBC immunization. Differentially expressed transcripts are depicted as green circles on a SAM plot (A). GSEA identified transcripts that correlate with the “GC vs. naïve B cell up-regulation” signature (B). Nine differentially expressed genes were identified and mitochondria-encoded tRNAs are highlighted in red (C). Data are representative of four mice per group. p0, the estimated prior probability that a gene is not differentially expressed; FDR, false discovery rate; d(i), statistics “d” for each gene i. (D) RT-qPCR for the differentially expressed mitochondria-encoded tRNAs in GC B cells from WT and RRas2−/− mice at day 7 after SRBC immunization. Data are means ± SEM from three mice per group. (E) qPCR quantification of mtDNA content in FACS-sorted GC B cells (CD19+CD95+GL7+) from WT and RRas2−/− mice at day 7 after SRBC immunization (pool of four mice per group). Data are means of the ratio of CO1 mitochondrial-encoded DNA to 18S nuclear-encoded DNA ± SEM from three mice per group. (F) RT-qPCR for the indicated mitochondrial genes in FACS-sorted GC B cells (CD19+CD95+CD38) from WT and RRas2−/− mice at day 7 after SRBC immunization. Data are means ± SEM from three mice per group. (G) Representative transmission electron micrographs of WT and R-Ras2–deficient GC B cells (CD19+CD95+GL7+) indicate mitochondria morphology (red arrowheads). Representative micrographs (left) and mean number of number of mitochondria with normal or atypical morphology per cell per section ± SEM (right) are from 13 cells per group. (H) Mitochondrial health and ROS activity in WT and R-Ras2–deficient GC B cells (B220+CD95+CD38) were measured by flow cytometry. The ratio of MMP to MM (left) and fluorescence of ROS-reactive CM-H2DCFDA (right) are means ± SEM from at least three mice per group. All data are representative of at least three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 by two-tailed Student’s t test.

Mammalian cells contain 1000 to 10,000 copies of mtDNA organized in clusters, known as nucleoids, that contain 2 to 10 copies each (43). To determine whether R-Ras2 deficiency affected mtDNA replication, we measured the relative abundance of mtDNA relative to a control of nuclear DNA by qPCR. We found no differences in mtDNA abundance in resting B cells, whereas activated B cells from Rras2−/− mice showed a 1.5-fold reduction in mtDNA compared to cells from WT mice (fig. S5A). Similarly, sorted R-Ras2–deficient GC B cells exhibited a threefold decrease in mtDNA abundance when compared to WT cells (Fig. 6E), indicating that R-Ras2 is required for mtDNA replication in activated B cells, particularly during the GC reaction. We also noted diminished mRNA expression from genes encoding mitochondrial proteins of the respiratory chain complexes (mt-ATP6, mt-CO1, mt-CO2, and mt-ND1) in R-Ras2–deficient GC B cells (Fig. 6F). Mitochondria fuse and divide, changing morphology in response to stress conditions. The number and extension of cristae formed by the inner membrane reflect functionality, because the respiratory chain proteins are located in this membrane (44). Transmission electron microscopy micrographs showed that GC B cells from Rras2−/− mice contained abnormally swollen mitochondria, with few and poorly differentiated cristae (Fig. 6G). These data suggested that mitochondrial function may be impaired in the absence of R-Ras2.

To examine mitochondrial functionality, we evaluated three different parameters: total mitochondrial mass (MM), mitochondrial membrane potential (MMP), and production of reactive oxygen species (ROS), using MitoTracker Green, MitoTracker Red, and CM-H2DCFDA, respectively (45). The MMP/MM ratio in B cells lacking R-Ras2 was lower than that in WT cells, due to a defect in MMP (fig. S5B). This difference was also evident in GC B cells lacking R-Ras2 (Fig. 6H). In vitro stimulation induces ROS production by both WT and R-Ras2–deficient B cells (fig. S5C). However, GC B cells isolated from immunized Rras2−/− mice had a 1.5-fold increase in ROS production compared to their WT GC B cell counterparts (Fig. 6H). The increase in ROS production in R-Ras2–deficient GC B cells correlated with lower mitochondrial activity. These results suggested that R-Ras2 is required to maintain MMP and, therefore, mitochondrial function after B cell stimulation.

To determine whether R-Ras2 deficiency affected mitochondrial respiration, we used two methods: We measured oxygen consumption upon addition of inhibitors of the respiratory chain with an extracellular flux analyzer, and we estimated how much oxygen was used in the production of adenosine 5′-triphosphate (ATP) versus other cellular processes by adding a proton ionophore to stop mitochondrial ATP production. We found that, in resting conditions, basal respiration was equivalent in WT and R-Ras2–deficient B cells but that the latter could not increase basal respiration in response to B cell stimulation (Fig. 7A). We used the respiratory chain complex V inhibitor oligomycin and the proton ionophore carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP) to estimate oxygen consumption during production of ATP by oxidative phosphorylation and to measure the maximum respiratory capacity of mitochondria (reserve capacity), respectively. Unlike WT B cells, R-Ras2–deficient B cells did not increase oxygen consumption (Fig. 7A, basal respiration) to cover ATP demands in activated B cells, and their reserve capacity was diminished in activated conditions (Fig. 7A, reserve capacity). These results showed that mitochondrial function is impaired in B cells lacking R-Ras2 and indicated that these cells are likely incapable of meeting the energy demands incurred during B cell activation.

Fig. 7 R-Ras2 deficiency prevents increased mitochondrial respiration upon activation.

(A) The Oxygen consumption rate (OCR) of WT and R-Ras2–deficient purified B cells stimulated with anti-IgM (left) or anti-CD40 (right) for 24 hours. Sequential addition of the indicated inhibitors was used to define the indicated respiratory parameters. Representative OCR measurements (top) and quantified values as means ± SEM are from an experiment performed in quintuplicate. (B) ATP cell content of WT and Rras2−/− purified B cells in basal conditions or stimulated with anti-IgM for 24 hours. Data are means ± SEM from at least four mice per group. (C) The fluorescence of ATP-sensitive quinacrine was measured by flow cytometry in quinacrine-treated GC (B220+CD95+GL7+) B cells from WT and Rras2−/− mice at day 7 after SRBC immunization. Data are means ± SEM from at least four mice per group. (D) Our proposed model describing the role of R-Ras2 downstream of BCR and CD40 and controlling mitochondrial activity through the PI3K-Akt-mTORC1 pathway. All data are representative of at least three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by two-tailed Student’s t test.

A failure in ATP production by oxidative phosphorylation could be compensated by increasing aerobic glycolysis. However, the acidification of the extracellular medium by lactate production, a by-product of glycolysis, was reduced in R-Ras2–deficient versus WT control B cells in activated conditions (fig. S5D). Furthermore, the glycolytic reserve, manifested as increased lactate production after blocking mitochondrial complex V with oligomycin, was significantly reduced in the absence of R-Ras2 (fig. S5D). These results suggested that Rras2−/− B cells cannot cope with increasing ATP demands during B cell activation by either increasing oxidative phosphorylation or glycolysis. The cytoplasmic concentrations of ATP were reduced in both resting and activated B cells in the absence of R-Ras2 (Fig. 7B). Intracellular ATP, measured as a function of the ATP-sensitive dye quinacrine, was significantly reduced in GC B cells lacking R-Ras2 (Fig. 7C).

Inhibition of mTORC1 impairs mitochondrial activity and biogenesis (46). mTORC1 phosphorylates the serine/threonine kinase p70S6K at Thr389, enhancing its activity (47, 48). We found that p70S6K phosphorylation induced upon stimulation of the BCR and phosphorylation of its substrate S6 were impaired in R-Ras2–deficient B cells compared to WT controls (Figs. 4, B and C, and 5B), thus pointing to defective activity of mTORC1 in the absence of R-Ras2. Because mTORC1 is positively regulated by Akt, our data link R-Ras2 activity downstream of membrane receptors (BCR and CD40) with the control of cell metabolism and mitochondrial activity by the PI3K-Akt-mTORC1 pathway (Fig. 7D). Thus, R-Ras2 helps activated B cells meet the increased energy demands by transducing signals from the BCR and CD40. Because the GC reaction involves several energy-demanding processes, impairment in increasing metabolic capacity could explain why Rras2−/− mice have such a strong GC phenotype.

DISCUSSION

Here, we provide evidence that R-Ras2, a member of the Ras-related subfamily of GTPases, plays a fundamental, B cell–intrinsic role in the GC response by mediating the proliferative and differentiation signals provided by two key receptors: the BCR and CD40. This role is specific to R-Ras2, because mice deficient in the closely related molecule R-Ras1 or in the classical molecules H-Ras and N-Ras do not develop a GC phenotype. We previously showed that R-Ras2 is essential for the survival and homeostatic proliferation of mature B cells (28). Here, we identify a previously uncharacterized specific function for this GTPase in controlling the development of the GC. Ras proteins are thought to be highly redundant because of their high degree of homology and the conservation of the effector loops. Such functional redundancy has been described for Rac1 and Rac2 during T cell and B cell development (49, 50). However, evidence that reflects the specific functional roles fulfilled by individual members of this family is now accumulating (51).

Previous studies of transgenic mice expressing a dominant-negative mutant H-Ras suggested that classical Ras proteins are required for the memory B cell response by providing survival signals that oppose apoptosis stimulated by the BCR (52). However, the use of dominant-negative H-Ras diminishes not only the activity of the three classical Ras proteins (53) but also that of R-Ras2, because all of these proteins share effectors and regulatory proteins and they have identical switch sequences (54). Considering our previous results regarding B cell homeostasis (28) and our current data showing that R-Ras2, but not classical Ras proteins, controls the GC reaction, we hypothesize that some, if not most, of the roles initially ascribed to classical Ras in B cells are in fact functions fulfilled by R-Ras2. Furthermore, the fact that classical Ras members do not compensate for the loss of R-Ras2 activity supports this view. Although the Raf-ERK and PI3K-Akt pathways are the two major pathways activated by classical Ras proteins, there is compelling in vitro and in vivo evidence that R-Ras2 plays a nonredundant role in the activation of PI3K. This functional specialization of R-Ras2 is especially critical for hematopoietic cells, because R-Ras2, but not classical Ras proteins, interacts with and activates the p110δ catalytic subunit of PI3K (55). Note that p110δ is only expressed by hematopoietic cells (56), and it seems to have a catalytic activity that is independent of regulatory subunits (57). Furthermore, genetic ablation of this subunit results in a marked reduction in marginal zone B cells, impaired homeostatic proliferation, and defective GC formation (58, 59).

We previously demonstrated that R-Ras2 is responsible for the direct recruitment of p110δ to the BCR, placing this GTPase in one of the mechanisms responsible for PI3K activation by the BCR, together with B cell adapter for phosphoinositide 3-kinase (BCAP) and noncatalytic region of tyrosine kinase (Nck) adaptor proteins (28, 59). Here, we showed that R-Ras2 mediated the activation of PI3K downstream of the BCR in GC B cells, as well as downstream of CD40. BCR- and CD40-dependent signals are essential for GC formation, providing essential proliferative and survival signals to LZ B cells (centrocytes) that are driven to commence or reinitiate rapid cell proliferation as DZ B cells (centroblasts). In contrast, proliferation of DZ B cells is independent of further BCR and CD40 signals. The few GC B cells detected in Rras2−/− mice had a DZ GC B cell phenotype. Therefore, these results suggest that R-Ras2 is likely critical for LZ GC B cell survival and proliferation, akin to the functional positioning of this GTPase immediately downstream of the BCR and CD40. Furthermore, two studies showed that either FoxO1 ablation or induction of PI3K activity results in loss of the DZ and that PI3K activity is enhanced in LZ GC B cells, suggesting that PI3K and FoxO1 activities oppose each other during GC B cell cycling between the LZ and the DZ (9, 41). These reports are consistent with the detected effects of R-Ras2 deficiency, resulting in reduced PI3K activity, reduced FoxO1 phosphorylation by Akt, FoxO1 protein degradation, and an increased proportion of DZ versus LZ GC B cells. We therefore propose that R-Ras2 is crucial in mediating LZ-to-DZ GC B cell cycling through its role in the activation of PI3K signaling by BCR and CD40. The specific defect in the LZ population of Rras2−/− mice also supports the present model of the GC reaction in which the DZ is the site of clonal B cell expansion and antigen receptor editing, whereas the LZ is the site of selection by antigen binding and assistance from T cells. The collapse of the LZ in the S phase of the cell cycle offers support to the “cellular timer” model regarding the GC. In this model, DZ-to-LZ GC B cell conversion progresses independently of any signals derived from the DZ, whereas active BCR and CD40 signaling in the LZ is required to reset the timer and reenter the DZ stage (35).

What are the immediate consequences of R-Ras2 deficiency at the cellular level? It was initially surprising to find such small differences in gene transcription profiles between R-Ras2–deficient and WT GC B cells, suggesting that the effect of R-Ras2 deficiency was not transcriptional. A similar finding in a breast cancer cell line depleted of R-Ras2 through short hairpin RNA (shRNA)–mediated approaches hinted at a role for R-Ras2 in the control of protein synthesis (60). However, although the SAM analysis showed few substantial transcriptional differences gene by gene between R-Ras2–deficient and WT GC B cells, GSEA showed statistically significant differences in groups of genes encoding enzymes involved in glucose metabolism, cellular respiration, and the TCA cycle. Together, these results suggest that R-Ras2 deficiency affects the transcription of genes dedicated to metabolism. This transcriptional effect on metabolism is consistent with the effects of FoxO1 regulation, because this transcription factor is also involved in the regulation of cell metabolism (61). However, not all of the roles of R-Ras2 are transcriptional. We found an effect of R-Ras2 deficiency on mtDNA replication and function, which are outcomes that can also be linked to the PI3K-Akt pathway through mTORC1. Thus, R-Ras2 seems to be immediately downstream of membrane receptors that regulate metabolism through a dual effect of the PI3K-Akt pathway on the activities of FoxO1 and mTORC1. Unlike T cells, which seem to meet the growing energy demands incurred upon their activation by enhancing aerobic glycolysis (62), B cells seem to increase both aerobic glycolysis and oxidative phosphorylation during activation (3). Both pathways of ATP production are markedly limited in the absence of R-Ras2, suggesting that R-Ras2 is primarily required for B cells to cope with the increasing energy demands encountered upon B cell activation and for GC B cells to reenter the cell cycle at the LZ stage after stimulation of the BCR and CD40.

The relevance and specificity shown by R-Ras2 in B cell homeostasis and GC formation warrant further experiments to pinpoint its involvement in other processes, such as B cell development. Similarly, the positioning of R-Ras2 downstream not only of antigen receptors (TCR and BCR) but also of CD40 paves the way for future studies on its relevance downstream of other key receptors for lymphoid cell biology, such as CD19 and CD28, particularly regarding their control of PI3K activity and cell metabolism. In addition, the profound effect of R-Ras2 deficiency on GC B cell energy supply points to metabolic regulation as one of the key cellular processes controlling the GC reaction. Finally, the fact that most human B cell lymphomas originate from GC or post-GC B cells (63) and that BCR signaling and PI3Kδ have a prominent signaling role in chronic lymphocytic leukemia and diffuse large B cell lymphoma (64) place R-Ras2 in the spotlight as a possible important target in the treatment of B cell malignancies.

MATERIALS AND METHODS

Mice and immunizations

Rras2−/− and Hras−/−Nras−/− mice were generated as described previously (24, 28), and all of the mice were maintained under specific pathogen–free conditions at the animal facility of the Centro de Biología Molecular Severo Ochoa in accordance with national and European guidelines. All procedures were approved by the ethical committee of the Centro de Biología Molecular Severo Ochoa. Where indicated, 6- to 12-week-old mice were immunized intraperitoneally with 2 × 109 SRBC or NP-CGG to generate a GC response, as described previously (65). Spleens were harvested 7 days postinjection, except where specified.

Generation of Rras1−/− mice

Knockout mice lacking RRAS expression were generated by genOway. The BAL1-HR targeting vector contained a neo cassette flanked by FLP recombinase target sequences inserted in the intron 1 and LoxP sites flanking exons 2 and 6. The construct was used in the electroporation of 129Sv/Pas embryonic stem (ES) cells, which were then selected with G418. Primary screening for 3′ homologous recombination was performed by PCR and was verified by 5′ Southern blots and 3′ Southern blots. Three independent ES clones of 48 were positive for homologous recombination and were injected into the blastocysts of C57BL/6J mice. One of the ES clones gave rise to 10 pups, of which 7 were identified as chimeras, including 1 male chimera with 99% chimerism. This male mouse was crossed with a C57BL/6 Flp deleter female mouse, and 14 mice with germ-line transmission were tested by PCR for the presence of the targeted allele. Two highly chimeric mice (98 and 99% chimerism) were each mated with C57BL/6J Cre deleter females to excise the floxed region (exons 2 to 6) and generate the constitutive knockout allele. Two of the six agouti F1 mice tested by PCR for excision were positive for both the excised and the non-excised alleles, indicating that they were mosaics. These mice were crossed with C57BL/6 mice, and the 14 pups generated were screened by PCR for the presence of the WT, excised, and non-excised alleles. Four were heterozygous for the WT and excised alleles, indicating that they had the knockout allele in heterozygosis, as confirmed by Southern blots. One pup identified as heterozygous for the knockin allele was crossed with C57BL/6J mice to generate the Rras−/− colony at the Centro de Biología Molecular Severo Ochoa.

Flow cytometry

Cells were incubated with anti-CD16/32 antibody in phosphate-buffered saline (PBS), 1% bovine serum albumin (BSA), and 0.02% sodium azide before labeling with saturating amounts of the fluorochrome-labeled or biotinylated monoclonal antibodies indicated in the figure legends, and fluorochrome-labeled streptavidin where applicable (listed in table S1). For intracellular labeling of pAkt and pERK, cells were first stimulated at the times indicated in the figure legends, and stimulation was stopped by adding paraformaldehyde (PFA) to a final concentration of 2% for 10 min at room temperature. Fixed cells were first stained with extracellular antibodies and then permeabilized for 30 min with 90% methanol on ice. Cells were washed and labeled intracellularly overnight with phosphor-specific antibodies. For intracellular staining of pp70S6K, splenic cells were stained on the surface as described earlier. After surface staining, the cells were treated with fixation/permeabilization buffer (BD Cytofix/Cytoperm kit, BD Biosciences) for 30 min. Anti-ppS6K antibody was then diluted in permeabilization buffer and incubated with the samples overnight at 4°C, and the samples were then washed. Labeled cells were analyzed on a FACSCalibur or FACSCanto II flow cytometer (Becton Dickinson), and the data were analyzed with a FlowJo software (Tree Star). Counting of total cells was performed with CountBright beads.

Anti-NP antibody production

Soluble Igs were quantified by ELISA for isotype determination (SouthernBiotech) with a 1:100 dilution of the sera from the immunized mice. For determination of high-affinity and low-affinity nitrophenol (NP)–reactive antibodies, Costar p96 flat-bottom plates were incubated overnight at 4°C with 100 μl of NP(7)-BSA or NP(41)-BSA (5 μg/ml) in PBS. BSA with a low grade of derivatization with nitrophenol [NP(7), that is, an average of 7 NP molecules per BSA molecule] was used to estimate the relative abundance of high-affinity antibodies because binding will be essentially monovalent, whereas BSA with a high grade of substitution [NP(41), that is, an average of 41 NP molecules per BSA molecule] was used to study total NP-specific binding because low-affinity antibodies also bind in a multivalent fashion to highly derivatized BSA due to an avidity effect. After overnight incubation with NP-derivatized BSA, plates were washed twice, blocked with 1% BSA, and ELISA was developed by standard procedures.

B cell proliferation

Splenic B cells were purified by negative selection using a combination of biotinylated anti-CD43 and anti-CD11b antibodies followed by incubation with a 3:1 ratio of streptavidin-coated magnetic beads (Invitrogen). CD43+ and CD11b+ cells were removed with a magnet (DynaMag, Invitrogen). The proliferation of purified B cells was analyzed by 3H-thymidine incorporation after 48 hours in culture with RPMI 1640, 10% fetal bovine serum, supplemented with 20 μM 2-mercaptoethanol and 10 mM sodium pyruvate after stimulation with IL-4 (5 ng/ml) and anti-IgM or anti-CD40 at the concentrations indicated in the figure legends.

BrdU proliferation assay

WT or Rras2−/− mice were injected intraperitoneally with 1 mg of BrdU 7 or 14 days after they had been immunized with SRBCs. After 90 min, the mice were sacrificed. Splenocytes (2 × 106 cells) were prepared and stained with fluorescent antibodies against the appropriate cell surface markers. Intranuclear BrdU staining was performed with a BrdU Flow kit (eBioscience) according to the manufacturer’s instructions. At the flow cytometer, the samples were acquired at 400 events/s to optimize resolution.

Measurement of MM and potential

ROS, MM, and MMP were measured with 5 μM CM-H2DCFDA, 100 nM MitoTracker Green, and 100 nM MitoTracker Red probes, respectively (all from Invitrogen), as previously described (45). Briefly, naïve primary cells or cells that had been stimulated in vitro with IgM and CD40 were incubated for 30 to 45 min at 37°C with the different probes in RPMI medium. The cells were then washed and labeled with antibodies against cell surface markers for flow cytometry analysis.

Measurement of cellular oxygen consumption

OCR and extracellular acidification rate were measured with an XF24 extracellular flux analyzer (Seahorse Bioscience). Two to 2.5 × 106 B cells per well (five wells per treatment for each independent experiment) were plated the day before analysis with anti–IgM-F(ab)2 antibody (10 μg/ml) or anti-CD40 antibody (10 μg/ml), where indicated in the figure legends. Cells were preincubated with nonbuffered Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 25 mM glucose, 1 mM pyruvate, and 2 mM glutamine for 1 hour at 37°C in a CO2-free incubator. OCR measurements were programmed with successive injections of nonbuffered DMEM, oligomycin (5 μg/ml; a mitochondrial respiratory complex V inhibitor), 300 nM FCCP (a proton ionophore that dissipates the proton gradient across the inner mitochondrial membrane), and 1 μM rotenone (a mitochondrial respiratory complex I inhibitor) together with 1 μM antimycin A (a mitochondrial respiratory complex III inhibitor). Calculations were performed according to the manufacturer’s instructions. Briefly, OCR was measured before and after the addition of inhibitors to derive several parameters of mitochondrial respiration: (i) Basal respiration was derived by subtracting nonmitochondrial respiration (the OCR after a mixture of rotenone and antimycin A was added) to the initial baseline cellular OCR; (ii) ATP turnover, also known as ATP-linked respiration, was calculated by subtracting the OCR after oligomycin addition from baseline cellular OCR; (iii) reserve capacity was calculated by subtracting basal respiration from maximal respiratory capacity measured by the OCR after FCCP addition; and (iv) proton leak respiration is calculated by subtracting nonmitochondrial respiration from the OCR value generated after addition of oligomycin.

Gene expression analysis

Splenic GC B cells from pooled samples of four to six SRBC-immunized mice were sorted in a FACSVantage sorter (Becton Dickinson), and RNA was extracted with an RNeasy kit (QIAGEN, 74134) according to the manufacturer’s instructions. RNA integrity was assessed using the Agilent 2100 Bioanalyzer (Agilent). Labeling and hybridizations were performed according to protocols from Affymetrix. Briefly, 100 ng of total RNA was amplified and labeled with the WT Plus Reagent Kit (Affymetrix) and then hybridized to Mouse Gene 2.0 ST Array (Affymetrix). Washing and scanning were performed using an Affymetrix GeneChip System (GeneChip Hybridization Oven 645, GeneChip Fluidics Station 450, and GeneChip Scanner 7G). GSEAs were performed using GSEA version 2.2.2.

Reverse transcription quantitative polymerase chain reaction

Total RNA was extracted from purified B or T cells using an RNeasy kit (QIAGEN, 74134) according to the manufacturer’s instructions. This RNA was reverse-transcribed using random hexamers and SuperScript II reverse transcriptase (Invitrogen), and the reverse-transcribed RNA was amplified with the appropriate primers (listed in table S1) using the SYBR Green PCR mix (Applied Biosciences). All primers were designed to span at least one intron. Expression of the gene of interest was normalized to that of a plasmid (GeneArt) containing the sequences of the six fragments amplified by the primers used for RT-qPCR. The plasmid was linearized with Asc I before amplification and quantified in a NanoDrop spectrophotometer (Thermo Scientific). This enabled measurement of the number of copies of each mRNA per nanogram of total RNA, and finally, all values were normalized to those for RRas1. To perform RT-qPCR analysis of the differentially expressed mitochondrial-encoded tRNAs, RNA extraction and RT-qPCR analysis of cells from WT and Rras2−/− mice were performed with the miRNeasy Mini Kit (QIAGEN) according to the manufacturer’s instructions. The cDNA was generated from 50 ng of RNA using ExiLERATE LNA qPCR system (Exiqon). A sample of the cDNA reaction (3 μl of a 1:10 dilution) was used as template for RT-qPCR analysis. The abundances of the mt-Tr, mt-TS2, mt-Tl2, and mt-Tf mitochondrial tRNAs and the 18S ribosomal RNA (rRNA) gene product were determined by RT-qPCR analysis using the GoTaq qPCR Master Mix (Promega) and specific oligonucleotides (ExiLERATE LNA qPCR custom primer set; table S1) in the CFX384 Touch Real-Time PCR Detection System Lightcycler (Bio-Rad). Each value was normalized by the 18S rRNA gene and expressed as the relative RNA abundance to that of cells obtained from WT mice. Experiments were performed in triplicate.

Measurement of mtDNA copy number

mtDNA copy number was assessed by qPCR analysis of total DNA samples extracted from sorted GC B cells or purified B cells that were then stimulated with anti–IgM-F(ab)2 (10 μg/ml), where indicated in the figure legends. mtDNA content was analyzed by calculating the ratio between mitochondrial complex 1 (CO1) gene content and nuclear encoded ARBP [Mus musculus ribosomal protein, large, P0 (Rplp0)] gene content (sequences are listed in table S1).

Transmission electron microscopy analysis

Isolated B cells or cell-sorted GC B cells from WT and Rras2−/− mice were fixed with 4% PFA and 2% glutaraldehyde in Sörensen phosphate buffer (pH 7.4) for 90 min at room temperature. Cell samples were then processed for embedding in epoxy TAAB 812 Resin (TAAB Laboratories) according to standard procedures. Ultrathin sections (80 nm thick) were stained with saturated uranyl acetate and lead citrate by standard procedures. Samples were examined at 80 kV in a Jeol JEM-1010 electron microscope. Images were captured with a TemCam-F416 (4 K × 4 K) digital camera Tietz Video and Image Processing Systems.

Quantification of intracellular ATP

Purified splenic B cells from WT and Rras2−/− mice were incubated for 24 hours at 37°C with or without anti-IgM (10 μg/ml). The amount of ATP content was measured by bioluminescence with the ATP Determination Kit (Molecular Probes) according to the manufacturer’s instructions, and the concentrations of ATP were calculated by interpolation on a standard curve generated from known concentrations of ATP (provided in the kit). Alternatively, ATP content was measured in GC B cells by flow cytometry after treatment with quinacrine (66). Briefly, splenic cells from SRBC-immunized WT and Rras2−/− mice were collected on day 7 after immunization and were stained with the appropriate antibodies against cell surface markers together with quinacrine (Sigma) at a final concentration of 5 μM for 30 min at room temperature in the dark. Cells were then washed and analyzed by flow cytometry.

Immunoprecipitation and Western blotting

For Western blotting of whole-cell lysates, purified splenic B cells (1 × 106 cells per time point) were activated for different times with anti-IgM or anti-CD40 (10 μg/ml) at 37°C. Cells were then lysed in Brij96 lysis buffer containing protease and phosphatase inhibitors [0.33% Brij96, 140 mM NaCl, 20 mM tris-HCl (pH 7.8), 10 mM iodoacetamide, 1 mM PMSF (phenylmethylsulfonyl fluoride), leupeptin (1 mg/ml), aprotinin (1 mg/ml), 1 mM sodium orthovanadate, and 20 mM sodium fluoride]. The samples were then resolved by SDS–polyacrylamide gel electrophoresis (PAGE) and analyzed by Western blotting with the antibodies listed in table S1, as described previously (29). Blots were quantified with ImageJ Software. For immunoprecipitation, a total of 3 × 107 purified splenic B cells were stimulated for different times with anti-CD40 (10 μg/ml) at 37°C, and the cells were then lysed in Brij96 lysis buffer for 30 min on ice. A postnuclear supernatant was prepared by centrifugation at 10,000g for 20 min at 4°C, and immunoprecipitation was carried out by incubation of the postnuclear supernatant with 10 μg of anti-CD40 for 1 hour at 4°C under continuous rotation followed by a 30-min incubation with 20 μl of protein-A Sepharose beads (Sigma). Beads were collected by a brief (seconds) centrifugation in an Eppendorf centrifuge, and the supernatant was discarded. The bead pellet was washed five times with 1 ml of Brij96 lysis buffer and finally resuspended in 20 ml of Laemmli buffer for SDS-PAGE. Western blotting was carried out after transfer to a nitrocellulose membrane with anti–R-Ras2 monoclonal antibody (Exbio).

Statistical analysis

Quantitative data are shown as means ± SEM. Differences between data sets were analyzed with the two-tailed Mann-Whitney test or by one-way ANOVA (when more than two groups of mice were compared) with a GraphPad Prism software.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/11/532/eaal1506/DC1

Fig. S1. Nonredundant role of R-Ras2 in B cell function.

Fig. S2. R-Ras2–deficient mice show a defective GC reaction, but GC B cells and TFH cells have normal amounts of surface receptors.

Fig. S3. Defective activation of the PI3K-Akt pathway in GC B cells in the absence of R-Ras2.

Fig. S4. GSEA data.

Fig. S5. R-Ras2 deficiency causes dysfunctional mitochondria.

Table S1. Antibodies and other materials.

REFERENCES AND NOTES

Acknowledgments: We thank M. Mittelbrunn for helpful advice and H. van Santen and M. Sefton for the critical reading of the manuscript. We also thank F. Bertaux and K. Thiam of genOway for the generation of the Rras1−/− mice. We are indebted to C. Prieto, V. Blanco, M. Rejas, and T. Gómez for their expert technical assistance. Funding: This work was supported by grants from the European Research Council ERC 2013–Advanced Grant 334763 NOVARIPP (to B.A.), the Ministerio de Economía y Competitividad [SAF2016-76394-R (to B.A.) and SAF2015-64556-R (to X.R.B.)], the Fondo de Investigaciones Sanitarias [FIS PI16/02137 (to E.S.)], the Instituto de Salud Carlos III RTICC [RD12/0036/0001 (to E.S.)], the Castilla-León Government [SA043U16 (to E.S.) and BIO/SA01/15, CSI049U16 (to X.R.B.)], the Worldwide Cancer Research [14-1248 (to X.R.B.)], the Fundación Ramón Areces, and the Fundación Científica de la Asociación Española Contra el Cáncer [GC16173472GARC (to B.A. and X.R.B.)]. Author contributions: P.M., N.M.-M., P.H.-A., P.D., I.F.-P., E.R.B., D.R.-G., and E.C. performed the experiments. A.M.-R., E.S., X.R.B., T.K., and B.A. designed and supervised the research. P.M. and N.M.-M. phenotyped the mice. I.F.-P. and E.C. generated classic Ras-deficient mice and were supervised by E.S. and X.R.B. M.D.D.-M. performed the GSEA analysis. P.M. and P.H.-A. performed the Seahorse experiments, which were supervised by A.M.-R. T.K. supervised the GC experiments. C.L.O. and B.A. prepared the manuscript, which was revised by P.M., N.M.-M., T.K., and X.R.B. Statistical analysis was performed by B.A. and P.M. D.A.-L. analyzed microarray data. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The microarray data have been deposited at the GEO repository with the data set identifier GSE113599. All other data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.
View Abstract

Navigate This Article