Inositol 1,3,4,5-tetrakisphosphate controls proapoptotic Bim gene expression and survival in B cells

Abstract

The contribution of the B isoform of inositol 1,4,5-trisphosphate [Ins(1,4,5)P₃] 3-kinase (or Itpkb) and inositol 1,3,4,5-tetrakisphosphate [Ins(1,3,4,5)P₄], its reaction product, to B cell function and development remains unknown. Here, we show that mice deficient in Itpkb have defects in B cell survival leading to specific and intrinsic developmental alterations in the B cell lineage and antigen unresponsiveness in vivo. The decreased B cell survival is associated with a decreased phosphorylation of Erk1/2 and increased Bim gene expression. B cell survival, development, and antigen responsiveness are normalized in parallel to reduced expression of Bim in Itpkb^−/− Bim^+/− mice. Analysis of the signaling pathway downstream of Itpkb revealed that Ins(1,3,4,5)P₄ regulates subcellular distribution of Rasa3, a Ras GTPase-activating protein acting as an Ins(1,3,4,5)P₄ receptor. Together, our results indicate that Itpkb and Ins(1,3,4,5)P₄ mediate a survival signal in B cells via a Rasa3–Erk signaling pathway controlling proapoptotic Bim gene expression.

Cross-linking of the antigen receptor on T and B lymphocytes by antigen is followed by the rapid activation of phospholipase Cγ, resulting in the hydrolysis of phosphatidylinositol 4,5-bisphosphate and the production of inositol 1,4,5-trisphosphate [Ins(1,4,5)P₃] and diacylglycerol. Ins(1,4,5)P₃ causes release of calcium from the endoplasmic reticulum and is the substrate of two main metabolic pathways. Dephosphorylation by type I Ins(1,4,5)P₃ 5-phosphatase generates the inactive metabolite inositol 1,4-bisphosphate (1, 2). Phosphorylation by isoforms A, B, and C of Ins(1,4,5)P₃ 3-kinase (Itpkb) or inositol polyphosphate multikinase results in the production of inositol 1,3,4,5-tetrakisphosphate [Ins(1,3,4,5)P₄] (3). In various cellular systems, it has been demonstrated that Ins(1,3,4,5)P₄ is a modulator of Ins(1,4,5)P₃ levels and calcium mobilization (3–5). However, Ins(1,4,5)P₃ and calcium concentrations in response to T cell antigen receptor stimulation were found normal in double-positive thymocytes from Itpkb-deficient mice, despite an important decrease in Ins(1,3,4,5)P₄ production resulting in profound defects in the final maturation of these cells (6, 7). Alternative pathways to mediate the cellular effects of Ins(1,3,4,5)P₄ include the binding to pleckstrin homology (PH) domains of specific proteins acting as Ins(1,3,4,5)P₄ receptor, and the synthesis of higher inositol phosphates (8).

To assess the potential contribution of the Ins(1,4,5)P₃-Itpkb-Ins(1,3,4,5)P₄ signaling pathway to B cell function and development, we have analyzed the B cell lineage of Itpkb^−/− mice. Our results indicate that Itpkb and Ins(1,3,4,5)P₄ mediate a survival signal in B cells by regulating Erk signaling pathway and proapoptotic Bim gene expression.

Results

Impaired B Cell Development in Itpkb^−/− Mice.

In the bone marrow of Itpkb^−/− mice, absolute numbers of pro-B/large pre-B (CD43⁺ B220⁺ IgM⁻) and small pre-B (CD43⁻ B220⁺ IgM⁻) cells were significantly reduced in knockout mice as compared with controls [Table 1 and supporting information (SI) Fig. 4]. Interestingly, an ≈3-fold decreased number of mature CD43⁻ B220⁺⁺ IgM⁺ recirculating B cells was also observed in the mutant mice, compared with Itpkb^+/+ mice (Table 1 and SI Fig. 4). Finally, no difference was observed in the number of CD19⁻ CD138⁺ plasma cells between both groups (Table 1).

Table 1.

Absolute numbers of B cell subsets in bone marrow and spleen of adult mice

Subset	Cell population	Itpkb+/+, mean ± SD (×106), n = 10	Itpkb−/−, mean ± SD (×106), n = 7	P
Bone marrow	Total	52.5 ± 13.5	37.3 ± 14.0	0.0632
	Pro-B/large pre-B(CD43+ B220+ IgM−)	0.82 ± 0.14	0.48 ± 0.14	0.0007
	Small pre-B(CD43− B220+ IgM−)	4.00 ± 1.42	2.05 ± 0.70	0.0095
	Immature(CD43− B220+ IgM+)	0.96 ± 0.46	0.49 ± 0.31	0.0541
	Mature recirculating(CD43− B220++ IgM+)	1.28 ± 0.53	0.39 ± 0.19	0.0018
	Plasma cells(CD19− CD138+)	0.76 ± 0.67	0.41 ± 0.36	0.2
Spleen	Total	59.60 ± 14.77	24.29 ± 8.52	<0.0001
	B220+	32.28 ± 10.29	13.61 ± 5.20	0.0003
	T1 (IgM++ CD21− CD23−)	3.99 ± 2.65	2.74 ± 0.98	0.200
	T2 (IgM++ CD21+ CD23+)	2.57 ± 1.64	0.86 ± 0.74	0.012
	FoM (IgM+ CD21+ CD23+)	16.00 ± 4.08	4.11 ± 2.89	<0.0001
	MZM (IgM++ CD21+ CD23−)	1.42 ± 0.54	1.43 ± 0.61	0.96
	B1* (B220+ CD5+)	1.19 ± 0.23	0.78 ± 0.53	0.25

The bone marrow B cell subsets were determined on the basis of CD43, B220, and IgM expression. Total cell numbers are for both femurs. Splenic B cell subsets were determined on the basis of CD21, CD23, CD5, B220, and IgM expression. The P value is determined by unpaired t test for α = 0.05 (Student or Welch corrected when appropriate).

*n = 3 for each group.

In the spleen of Itpkb^−/− mice, flow cytometry revealed a decreased number of total B cells and specifically a 3-fold and a 4-fold reduction in the number of T2 (IgM⁺⁺ CD21⁺ CD23⁺) and FoM (IgM⁺ CD21⁺ CD23⁺) cells, respectively (Table 1). No difference was detected between Itpkb^+/+ and Itpkb^−/− mice in the number of T1 (IgM⁺⁺ CD21⁻ CD23⁻), MZM (IgM⁺⁺ CD21⁺ CD23⁻), and B1 (B220⁺ CD5⁺) cells (Table 1). These results were confirmed by staining for CD21 and CD24 after gating on B220⁺ cells or by staining with IgM and IgD antibodies (Table 2 and SI Fig. 5). Immunohistochemistry on splenic sections confirmed the significant reduction in the number of CD21⁺ CD23⁺ T2 and FoM cells and the presence of normal numbers of CD21⁺ CD23⁻ MZM cells at the periphery of the follicle (SI Fig. 6). Finally, no difference in the percentage of B1 cells was detected in the peritoneal cavity from mutant and control mice (data not shown).

Table 2.

Splenic B cell populations in adult mice based on IgM/IgD staining (mean ± SD, ×10⁶)

Mice	n	T1-MZ (IgM++ IgD−)	T2 (IgM++ IgD+)	FoM (IgM+ IgD+)
Itpkb+/+*	4	7.97 ± 4.02	7.67 ± 4.06	17.4 ± 7.78
Itpkb−/−*	4	4.89 ± 2.53	2.05 ± 1.37†	7.08 ± 3.15†
T+Itpkb−/−*	4	4.00 ± 1.86	2.00 ± 1.42†	6.03 ± 2.37†
Scid Itpkb+/+‡	4	7.37 ± 2.49	7.44 ± 2.84	22.4 ± 9.66
Scid Itpkb−/−‡	4	8.35 ± 5.36	2.35 ± 1.43§	7.85 ± 3.94§

*Experiments where Itpkb is specifically expressed in T cells of Itpkb^−/− mice; statistical analysis is based on the one-way ANOVA and subgroup post test on Bonferroni's multiple comparison test if significant one-way ANOVA for α = 0.05.

^†For T2 and FoM populations, the Itpkb^+/+ group is statistically different from the Itpkb^−/− and the T⁺ Itpkb^−/− groups (P < 0.05), even though these last two groups are not different.

^‡Reconstitution experiments; the P value is determined by unpaired t test (Student or Welch corrected when appropriate).

^§P < 0.05.

Altogether, these data indicate that Itpkb deficiency leads to B cell defects at specific developmental stages in the bone marrow and the spleen of adult mice, and that Itpkb is particularly important for the maturation of FoM B cells, but not MZ mature and B1 B cells.

The Developmental Defects Are Intrinsic to B Cells.

To explore whether the B cell defects observed in Itpkb^−/− mice were intrinsic to the B cell lineage, we reconstituted SCID mice with fetal liver cells from either Itpkb^+/+ or Itpkb^−/− mice. Significantly decreased numbers of pro-B/large pre-B, small pre-B, and mature recirculating B cells and follicular B cells were found in SCID mice reconstituted with Itpkb^−/− cells, as in Itpkb^−/− mice (Table 2, SI Fig. 7, and data not shown). We next investigated the potential effect of the absence of T cells in Itpkb^−/− mice on B cell development. Itpkb^−/− mice were genetically modified to express an Itpkb transgene specifically in the T cell compartment (T⁺ Itpkb^−/− mice). These T⁺ Itpkb^−/− mice have mature thymocytes and peripheral T cells (SI Fig. 8 and data not shown). The presence of mature T cells in T⁺ Itpkb^−/− mice did not significantly modify the B cell defects found in Itpkb^−/− mice (Table 2 and SI Fig. 7).

Altogether, these data show that Itpkb^−/− stromal cells and the absence of T cells are not necessary for the development of the B cell phenotype in Itpkb^−/− mice and indicate that the observed defects are intrinsic to B cells.

Impaired Function of Itpkb^−/− B Cells.

To explore the function of B cells persisting in Itpkb^−/− mice, the response to a classic T cell-independent type 2 (TI-2) antigen, 4-hydroxy-3-nitrophenylacetyl (NP)-Ficoll, was analyzed in vivo (Fig. 1a). NP-specific IgG3 antibodies were detected in control mice, but not in Itpkb^−/− and T⁺ Itpkb^−/− mice. Similar results were obtained for NP-specific IgM antibodies (data not shown). These results indicate that B cells persisting in the mutant mice are unresponsive to TI-2 antigen in vivo, and that this defect is not the consequence of the absence of T cells in these mice.

Fig. 1.

In vivo and in vitro function and survival of Itpkb^−/− B cells. (a) Levels of anti-NP IgG3 antibodies in response to NP-Ficoll antigen immunization; data are mean ± SD for seven Itpkb^+/+, four Itpkb^−/−, and four T⁺ Itpkb^−/− mice. *, P = 0.0027 by one-way ANOVA. The Itpkb^+/+ group is statistically different from the Itpkb^−/− and the T⁺ Itpkb^−/− groups by Bonferroni's multiple comparison test (P < 0.01). (b and c) Splenic B cell survival in culture with or without goat F(ab′)₂ anti-mouse IgM (10 μg/ml) (b) and with recombinant BAFF (rBAFF) (100 ng/ml) and with or without goat F(ab′)₂ anti-mouse IgM (10 μg/ml) or with LPS (100 ng/ml) (c), based on propidium iodine staining. Results are expressed as a ratio between living cells at t = 24 or 48 h and living cells at t = 0 h. The data represent mean ± SEM of five independent experiments, except for the rBAFF experiment (mean ± SEM of two experiments). Statistical analysis was realized by using Student's unpaired t test or Welch corrected when appropriate: *, P = 0.0013; **, P = 0.0036; ***, P = 0.031.

Decreased Survival of Itpkb^−/− B Cells in Vitro.

The survival of purified Itpkb^+/+ and Itpkb^−/− splenic B cells was investigated in culture. The percentage of living B cells, as assessed by propidium iodide staining, was significantly reduced in Itpkb^−/− mice in the absence of any stimulus (Fig. 1b). F(ab′)₂ anti-IgM stimulation, which is known to increase the apoptosis of B cells in culture, further decreased the viability of the mutant cells, as compared with Itpkb^+/+ B cells (Fig. 1b). The presence of the prosurvival B cell-activating factor (BAFF) resulted in a markedly increased survival in both mutant and control B cells (Fig. 1c). When F(ab′)₂ anti-IgM and BAFF were added together, the survival of Itpkb^+/+ and Itpkb^−/− B cells was significantly increased compared with anti-IgM alone (Fig. 1 b and c). In these conditions, a significant difference was still observed between the survival of Itpkb^+/+ and Itpkb^−/− B cells after stimulation (Fig. 1c). Anti-CD40 stimulation, another prosurvival factor, reproduced the results obtained with BAFF (data not shown). Finally, mutant and control B cells survived equally in the presence of LPS (Fig. 1c).

Together, our results indicate that Itpkb mediates a survival signal downstream of the BCR, and that survival signals downstream of the BAFF receptor, CD40, and Toll-like receptor are not affected by Itpkb deficiency.

Role of the Proapoptotic Bim Gene in the Survival of Itpkb^−/− B Cells.

Specific members of the Bcl-2 family, including Bax, Bim, Bcl-2, Bcl-xL, Mcl-1, and Bcla1a, control B cell survival and function (9). Quantitative real-time RT-PCR analysis revealed a slight, but very significant, increase in Bim messenger RNA level in splenic resting follicular B cells persisting in Itpkb^−/− mice, as compared with Itpkb^+/+ mice (Fig. 2a). In Itpkb^−/− Eμ-2–22 BCL-2 mice, which express a human BCL-2 transgene specifically in the B cell population (10), the numbers of splenic follicular B cells are not decreased anymore (Table 3). However, a significant increase in Bim mRNA level was still detected in these cells (Fig. 2a), indicating that the deregulated Bim expression is not simply the consequence of the reduced FoM B cell number in Itpkb-deficient mice. Flow cytometry confirmed that follicular B cells present in Itpkb^−/− and Itpkb^−/− Eμ-2–22 BCL-2 mice expressed a higher level of Bim protein, as compared with control mice (Fig. 2b). On the contrary, no increase in Bim protein level was detected in T1-MZ B cells (data not shown).

Fig. 2.

Bim expression and role in Itpkb^−/− B cells. (a) Quantitative real-time RT-PCR on RNA isolated from FACS-sorted splenic follicular Itpkb^+/+, Itpkb^−/−, and Itpkb^−/− Eμ-2–22 BCL-2 B cells (SI Table 4). Data are mean ± SEM of three independent experiments after normalization to Pbgd, Hprt, and Rpl13a mRNA. *, P = 0.0027; **, P = 0.0001 by one sample t test comparing the expression of the gene of interest to theoretical mean 1.00 (as no modulation of expression). (b) Analysis of bim expression by flow cytometry on splenic follicular B cells. Splenic follicular B cells were gated on IgM/IgD expression, and mean fluorescence intensity (MFI) of Bim expression ± SEM is presented for five Itpkb^+/+, seven Itpkb^−/−, and two Itpkb^−/− Eμ-2–22 BCL-2 mice after subtraction of nonspecific MFI observed in splenic follicular B cells of Bim^−/− mice. *, P = 0.028 using Student's unpaired t test for Itpkb^+/+ and Itpkb^−/− groups. (c) Survival of Itpkb^−/−Bim^+/+, Itpkb^−/−Bim^+/−, Itpkb^+/+Bim^+/+, and Itpkb^−/− Eμ-2–22 BCL-2 B cells in culture analyzed as in Fig. 1b. Data represent mean ± SEM of two independent experiments. (d) Levels of anti-NP IgG3 antibodies in response to NP-Ficoll antigen immunization; data are mean ± SEM for six Itpkb^+/+ and three Itpkb^−/− Bim^+/− mice.

Table 3.

B cell subsets in spleen of Itpkb^−/− Eμ-2–22 BCL-2 and Itpkb^−/−Bim^+/− mice, based on IgM/IgD staining (mean ± SD, ×10⁶)

Mice	n	T1-MZ (IgM++ IgD−)	T2 (IgM++ IgD+)	FoM (IgM+ IgD+)
Itpkb+/+	5	7.97 ± 4.02	7.67 ± 4.06	17.4 ± 7.78
Itpkb−/−	3	4.89 ± 2.53	2.05 ± 1.37	7.08 ± 3.15
Itpkb+/+Bim+/−	4	10.6 ± 3.5	14.9 ± 4.3	40.5 ± 7.2
Itpkb−/−Bim+/−	7	12.0 ± 3.6*	6.5 ± 1.9*	26.9 ± 7.1†
Itpkb−/− Eμ-2–22 BCL-2	2	6.1 ± 2.1	5.7 ± 0.8	40.4 ± 7.0

Statistical analysis is based on one-way ANOVA and Bonferroni's multiple comparison test for Itpkb^+/+ vs Itpkb^−/−Bim^+/− and Itpkb^−/− vs. Itpkb^−/−Bim^+/− if significant one-way ANOVA for α=0.05. Itpkb^+/+ and Itpkb^−/−Bim^+/− B cell populations are statistically the same; Itpkb^−/− and Itpkb^−/−Bim^+/− B cell population are statistically different.

*, P < 0.05;

†, P < 0.01.

These results indicate that Itpkb deficiency is associated with a specific overexpression of the proapoptotic Bim protein in follicular B cells and that overexpression of the antiapoptotic Bcl-2 protein, if sufficient to restore a normal number of FoM B cells, is not sufficient to normalize Bim expression.

The complete absence of Bim in mice results in a selective increase in the numbers of T2 and follicular mature B cells in the spleen, two populations affected by Itpkb inactivation (11). To test whether the increased Bim expression detected in follicular Itpkb^−/− B cells is sufficient to reduce their survival and number, Itpkb^+/− mice were crossed with Bim^+/− mice initially described by Bouillet et al. (12). FoM B cell analysis in Itpkb^−/−Bim^+/− mice demonstrated that haploinsufficiency at the Bim locus was sufficient to restore a normal Bim expression level (data not shown). As a consequence, no B cell developmental defects were detected in these mice. Indeed, there were normal numbers of pro-B/large pre-B [mean ± SD for Itpkb^+/+Bim^+/+ and Itpkb^−/−Bim^+/−, respectively (×10⁵): 4.6 ± 2.6; 5.0 ± 2.8], small pre-B [mean ± SD for Itpkb^+/+Bim^+/+ and Itpkb^−/−Bim^+/−, respectively (×10⁶): 4.1 ± 2.6; 4.2 ± 1.4], and mature recirculating B cells in the bone marrow (SI Fig. 9a), as well as follicular Itpkb^−/− Bim^+/− B cells in the spleen (Table 3 and SI Fig. 9b). When cultured with or without F(ab′)₂ anti-IgM, splenic B cells from Itpkb^−/−Bim^+/− and Itpkb^+/+Bim^+/+ mice survived similarly, indicating that Itpkb^−/− B cells recovered a normal viability when one allele of the Bim gene was inactivated (Fig. 2c). Finally, after NP-Ficoll immunization, Itpkb^−/−Bim^+/− and Itpkb^+/+Bim^+/+ mice mounted a similar anti-NP IgG3 antiboby response (Fig. 2d).

Together, these data show that Bim is overexpressed in B cells from Itpkb^−/− mice and underline its important role in the reduced survival and developmental defects observed in these mice.

Itpkb and Ins(1,3,4,5)P₄ Control Erk Activation and Bim Level in B Cells.

The concentrations of Ins(1,4,5)P₃ were not significantly different in splenic B cells from Itpkb^+/+ and Itpkb^−/− mice before and after stimulation with F(ab′)₂ anti-mouse IgM (SI Fig. 10). Interestingly, calcium mobilization in response to anti-IgM was decreased in Itpkb^−/− follicular B cells (SI Fig. 11). However, calcium mobilization in response to ionomycin was also systematically lower in the mutant follicular B cells as compared with control cells, precluding any conclusion on the role of Itpkb in the control of calcium levels in these cells.

To circumvent the problems caused by the decreased number and survival of follicular B cells in Itpkb^−/− mice when analyzing the signaling pathway leading from Itpkb deficiency to Bim overexpression, Itpkb^−/− Eμ-2–22 BCL-2 mice, which overexpressed Bim similarly to Itpkb^−/− mice (Fig. 2 a and b), were compared with Itpkb^−/− Eμ-2–22 BCL-2 mice in these experiments. Before and up to 10 min after B cell receptor (BCR) stimulation, Bim was overexpressed in Itpkb^−/− Eμ-2–22-BCL-2 splenic B cells, as compared with Itpkb^+/+ Eμ-2–22-BCL-2 B cells (Fig. 3a). The pattern of Bim migration was also different in Itpkb^+/+ and Itpkb^−/− Eμ-2–22-BCL-2 mice 5 and 10 min after BCR stimulation, suggesting that Bim phosphorylation defects also occurred in the mutant mice (Fig. 3a).

Fig. 3.

Analysis of the signaling pathway leading to Bim increased expression. (a) Western blot for Bim expression in splenic B cells from Itpkb^+/+ Eμ-2–22 BCL-2 and Itpkb^−/− Eμ-2–22 BCL-2 stimulated with goat F(ab′)2 anti-mouse IgM (10 μg/ml) for different times. β-Actin was used as control. The blot is representative of three independent experiments. (b) Analysis of Rasa3 expression and Erk1/2 and Bim_EL phosphorylation after splenic B cells activation with goat F(ab′)₂ anti-mouse IgM (10 μg/ml) for different times or ionomycin (50 ng/ml) plus phorbol 12-myristate 13-acetate (10 ng/ml) for 5 min (I+P). Erk1/2 and β-actin were used as controls. The blot represents one of three independent experiments. (c) Rasa3–GFP subcellular distribution in Cos-7 cells transfected with Rasa3-GFP alone (Upper) or with Rasa3-GFP and Itpkb (Lower). Pictures of representative cells are presented 0, 3, and 5 min after the addition of 100 μM ATP. The profile of fluorescence intensity from membrane to membrane (as depicted by axis presented in small pictures at t = 0) is given below each cell. The fluorescence intensity is represented as arbitrary units. The ratio (R) between the mean fluorescence intensities of the membranes (A1 and A2) and the mean fluorescence intensity of cytosol (B) is given. The pictures are representative of three independent transfection experiments. (d) Ins(1,3,4,5)P₄ decreases membrane localization of Rasa3–GFP in HeLa cells. Rasa3–GFP-transfected HeLa cells were cultured for 30 min with DMSO-containing solvent, 2,6-di-O-butyryl-Ins(1,2,4,5)P₄/PM (50 μg/ml) or 2,6-di-O-butyryl-Ins(1,3,4,5)P₄/AM (50 μg/ml). Pictures of representative cells are presented 0 and 30 min after addition of each compound. Data are representative of three independent experiments.

The Ras–Erk signaling pathway regulates expression and phosphorylation of Bim, targeting this protein for ubiquitination and proteasomal degradation in various cells, including B cells (11, 13–16). Erk activation has also been shown defective in thymocytes from Itpkb^−/− mice (7). Therefore, its phosphorylation was analyzed in B cells from Itpkb^+/+ and Itpkb^−/− Eμ-2-22-BCL2 mice. Erk1 and Erk2 were found to be much less phosphorylated in Itpkb^−/− Eμ-2–22-BCL-2 B cells than in control B cells after BCR activation (Fig. 3b). Together, these results indicate that Itpkb and Ins(1,3,4,5)P₄ are crucial for the optimal activation of the Ras/Erk pathway in B cells.

Ins(1,3,4,5)P₄ Controls the Subcellular Distribution of Rasa3.

Rasa3 is a GTPase-activating protein that binds to the GTP-bound form of Ras and catalyzes its inactivation, thereby modulating the Ras/Erk signaling pathway (17). Both Rasa3 mRNA and protein are highly expressed in T and B cells (ref. 18 and http://symatlas.gnf.org/SymAtlas). The protein is essentially membrane-associated through the interaction of its PH domain with phosphatidylinositol 4,5-bisphosphate (19). Because Rasa3 PH domain has also a high affinity for the water-soluble Ins(1,3,4,5)P₄ (20, 21), it has been suggested that Ins(1,3,4,5)P₄ production in stimulated thymocytes could result in a substantial redistribution of Rasa3 from the plasma membrane to the cytosol, allowing unhindered Ras activation (7). Accordingly, Itpkb deficiency and decreased Ins(1,3,4,5)P₄ production should prevent Rasa3 sequestration in the cytosol, suppress optimal Ras and Erk activation, and lead to abnormal thymocyte maturation (7). The potential role of Rasa3 in the Erk signaling defects observed in Itpkb^−/− B cells was analyzed. First, Rasa3 protein expression was similar in mutant and control B cells, excluding the possibility that an abnormal concentration of this protein is responsible for the defects (Fig. 3b). Second, Cos-7 cells were used to test whether the subcellular distribution of Rasa3 could be modulated by Itpkb activation and/or Ins(1,3,4,5)P₄ production, because the poor survival of purified B cells when cultured for extended period and their very low transfection efficiency made these cells unsuitable for Rasa3–GFP translocation analysis. In nonstimulated Cos-7 cells transfected with the Rasa3–GFP construct and an Itpkb expression vector or the Rasa3–GFP construct alone, an intense green fluorescence signal was localized at the membrane (Fig. 3c). When cells cotransfected with Rasa3–GFP and Itpkb were stimulated with ATP to activate phospholipase C (4), the intensity of the membrane signal significantly decreased in ≈20% of the fluorescent cells within 3–5 min after ATP addition (Fig. 3c). By contrast, in Rasa3–GFP transfected cells, the intensity of the membrane fluorescent signal never decreased during the same time periods, indicating that Rasa3–GFP translocation only occurs when Itpkb is coexpressed in the cells (Fig. 3c). Indeed, Itpkb transfection in Cos-7 cells resulted in a 16-fold increase in Itpkb activity (from 527 ± 124 nmol/min per g of proteins in nontransfected cells to 8,475 ± 2,519 nmol/min per g of proteins in Itpkb-transfected cells), the basal activity in nontransfected cells probably reflecting the presence of other Itpks in these cells (SI Fig. 12). A substantial increase in the concentrations of Ins(1,3,4,5)P₄ and inositol 1,3,4-trisphosphate, a specific metabolite of Ins(1,3,4,5)P₄, was also detected in Itpkb-transfected cells 3 min after ATP stimulation, as compared with nontransfected cells (SI Fig. 12).

To test the direct effect of Ins(1,3,4,5)P₄ on Rasa3–GFP subcellular distribution, HeLa cells were transfected with the Rasa3–GFP construct and treated with a membrane-permeant derivative of either Ins(1,3,4,5)P₄ [d-2,6-di-O-butyryl-myo-inositol 1,3,4,5-tetrakisphosphate octakisacetoxymethyl ester (Bt₂Ins(1,3,4,5)P₄/AM)], or Ins(1,2,4,5)P₄ [d-2,6-di-O-butyryl-myo-inositol 1,2,4,5-tetrakisphosphate octakispropionyloxy methyl ester (Bt₂Ins(1,2,4,5)P₄/PM)] as a control (SI Fig. 13). These membrane-permeant derivatives diffuse into cells and are subsequently hydrolyzed by nonspecific cellular esterases to release the active compounds. Translocation of Rasa3–GFP from the plasma membrane to the cytosol occurred slowly in the presence of Bt₂Ins(1,3,4,5)P₄/AM, probably reflecting the slow penetration and metabolism of the compound and the need to reach a threshold concentration in HeLa cells to be effective (Fig. 3d and SI Movie 1). By contrast, no redistribution of Rasa3–GFP was detected in the presence of Bt₂Ins(1,2,4,5)P₄/PM, which has a ≈100-fold lower affinity for Rasa3 PH domain than Ins(1,3,4,5)P₄ (19) or in the presence of DMSO, both used as negative controls (Fig. 3d and SI Movies 2 and 3).

These results suggest that Itpkb and Ins(1,3,4,5)P₄ control the subcellular distribution of Rasa3 and provide a good candidate for the link between Ins(1,3,4,5)P₄ and the Ras/Erk signaling pathway to modulate proapoptotic Bim gene expression.

Discussion

Mice deficient for Itpkb present, besides the previously reported T cell differentiation defects, specific and intrinsic alterations in the B cell compartment. These developmental and functional alterations are the consequence of a decreased B cell survival caused by a selective increase in the proapoptotic Bim mRNA and protein expression, associated with a decreased Bim phosphorylation pattern. The essential role of Bim overexpression in the setting of the B cell phenotype in Itpkb^−/− mice was clearly demonstrated when analyzing Itpkb^−/−Bim^+/− mice. Indeed, the partial inactivation of Bim in these mice is sufficient to normalize Bim expression in B cells and B cell survival, numbers, and functions. A similar effect of the loss of a single allele of Bim has also been reported in Bcl-2^−/− mice, where the abnormal lymphocyte survival and the kidney disease were prevented (22). Interestingly, total inactivation of Bim in Itpkb^−/− Bim^−/− double knockout mice resulted in the presence of peripheral T cells, reaching ≈60% of the normal numbers of CD4⁺ and CD8⁺ T cells in the spleen (Y.M. and S.S., unpublished results). Together, these data indicate that a tight control of Bim protein expression is crucial for B and T cell survival, and that Itpkb and Ins(1,3,4,5)P₄ constitute one of the signaling pathways that subtly tune Bim expression at the transcriptional/phosphorylation levels in lymphoid cells.

The exact mechanism(s) that mediate Ins(1,3,4,5)P₄ actions in cells are still controversial. We report here that in the few follicular B cells persisting in Itpkb^−/− mice Ins(1,4,5)P₃ production was normal, but calcium mobilization in response to BCR stimulation or ionomycin was significantly decreased (SI Fig. 11). The reason for these defects is still unclear, and we cannot exclude a direct role for Ins(1,3,4,5)P₄ in the control of calcium mobilization in B cells. Interestingly, reports have been published on the effects of Bcl-2 family members, including Bim, on Ins(1,4,5)P₃ receptor activation (23–25). They indicate that the ratio of proapoptotic versus antiapoptotic Bcl-2 family members, which is altered in our present study, can control calcium release from the endoplasmic reticulum.

In a recent study by Miller et al. (26) on another strain of Itpkb-deficient mice (called Ms. T-less mice), calcium concentrations in response to BCR stimulation were found significantly increased in mutant B cells. Based on these results, Miller et al. suggested that Ins(1,3,4,5)P₄ negatively modulates store-operated calcium channels and calcium entry in B cells. They reported a similar calcium-entry defect when thymocytes were incubated with cell permeable Ins(1,3,4,5)P₄ (see supporting figure 6 in ref. 26), although no alteration in calcium mobilization has been observed in previous studies on both mutant mice (6, 7). At the present time, we have no clear explanation for the calcium mobilization discrepancies between the two strains of mutant mice. Interestingly, other parameters of the B cell phenotype are different between Itpkb^−/− and Ms. T-less mice, including the numbers of T1 and MZ B splenic cells (which are significantly reduced in Ms. T-less mice), the number of peritoneal B-1 B cells (which is significantly increased in their study), and the presence of alterations in the bone marrow (our study). We propose that these differences may reflect the different mutation introduced in the Itpkb gene and/or result from the different housing and intestinal bacterial flora between the two strains of mice. We suggest that the role of Itpkb in B cell development, survival and function, including calcium responses, should also be carefully evaluated in BCR transgenic models.

In both strains of mice, mutant B cells failed to proliferate in response to BCR stimulation, even when they expressed the Bcl-2 transgene, indicating that Itpkb and Ins(1,3,4,5)P₄ are also important for B cell proliferation, independently of their action on Bim (ref. 26 and Y.M. and S.S., unpublished results).

Defects in Erk phosphorylation have been previously reported in T cell antigen receptor-stimulated thymocytes from Ms. T-less mice (7, 27). We report here a similar defect in BCR-stimulated Itpkb^−/− B cells and propose that it participates in Bim overexpression and the decreased survival detected in the mutant B cells. Indeed, in other cells, it has been demonstrated that Erk-mediated phosphorylation can target Bim for ubiquitination and proteasomal degradation (16, 28, 29). It is noteworthy that in Ms. T-less B cells, Erk1 phosphorylation seemed significantly decreased 2, 5, and 15 min after BCR stimulation (see figure 5 in ref. 26). Unfortunately, B cell survival was not investigated in this study.

PH domains bind not only membrane phosphoinositides, but also their cognate head groups, which are produced as cytosolic inositol phosphates (reviewed in ref. 30). In this case, the subcellular distribution of the PH domain-containing protein is determined by the relative concentrations of and affinity toward membrane-bound phosphoinositide and water-soluble inositol phosphate (31, 32). Our results indicate that Itpkb activity and membrane-permeant Ins(1,3,4,5)P₄ cause a similar subcellular redistribution of the Ins(1,3,4,5)P₄ receptor Rasa3 from the plasma membrane, where its PH domain interacts with phosphatidylinositol 4,5-bisphosphate, to the cytosol. Because Rasa3 is a GTPase-activating protein acting on Ras, its persistent or prolonged membrane localization, close to its substrate, could explain the defects in Erk activation and proapoptotic Bim gene expression/phosphorylation observed in Itpkb^−/− B cells. Interestingly, recent experiments in thymocytes from Ms. T-less mice indicate that Ins(1,3,4,5)P₄ controls subcellular localization of other PH domain-containing proteins besides Rasa3 (27).

In conclusion, these data in B cells and thymocytes from Itpkb^−/− and Ms. T-less mice indicate that one of the physiological mechanisms of Ins(1,3,4,5)P₄ action is the control of subcellular distribution of specific PH domain-containing proteins, resulting in the fine regulation of important signaling cascade like the Erk pathway. What exactly determines the specificity of the proteins targeted by Ins(1,3,4,5)P₄ in different cells or during different physiological processes remains to be investigated.

Experimental Procedures

Mice, immunization, antibodies, flow cytometry analysis, immunohistochemistry, ELISA, Western blot analysis, determination of inositol phosphates in intact cells, follicular B cell sorting, relative real-time RT-PCR, and statistics were performed as described (6). Further information can be found in SI Text.

Production of T⁺ Itpkb^−/− Mice.

To express an HA-tagged murine Itpkb protein specifically in the T cell lineage of mice, we used a combination of the LTH1 (lck-tTA) mouse line, which produces a tet-regulatable transcriptional activator (tTA) specifically in the T cell lineage (33, 34) and the TW2 mouse line. The TW2 mouse line was produced by Nucleis (Lyon, France). The TW2 transgene contains sequences encoding the TetO–hCMV minimal promoter, which requires tTA for its activity, and the mouse HA-tagged Itpkb cDNA (V.P. and S.S., unpublished results). Mating LTH1 Itpkb^+/− mice with TW2 Itpkb^+/− mice generated double transgenic LTH1 TW2 Itpkb^−/− mice, or T⁺ Itpkb^−/− mice, expressing the Itpkb protein specifically in the T cell lineage in the absence of doxycyclin administration (data not shown). Unlike Itpkb^−/− mice (6, 7), T⁺ Itpkb^−/− mice have mature peripheral T cells and survive in our conventional animal room for >12 months (SI Fig. 8 and data not shown).

B Lymphocyte Purification and Stimulation.

B cells were isolated from spleen by depletion of non-B cells and cultured as described (35). Activation conditions were as follows: 10 μg/ml of F(ab′)₂ anti-IgM (Jackson Immunoresearch, West Grove, PA) and/or with 6 μg/ml of anti-CD40 (eBioscience, San Diego, CA) or 100 ng/ml of recombinant BAFF (R & D Systems, Minneapolis, MN) or with 100 ng/ml of LPS. Cells were stained with propidium iodide for survival analysis at 24 and 48 h and passed on a FC 500 flow cytometry system (Beckman Coulter, Fullerton, CA).

Rasa3–GFP Translocation Analysis.

Cos-7 cells were transiently transfected with Rasa3–GFP (20) and pcDNA3-Itpkb constructs (4) or Rasa3–GFP construct alone by using FuGENE 6.0 (Roche Diagnostics, Basel, Switzerland). Forty-eight hours later, cells were stimulated with ATP (100 μM), and fluorescence was analyzed on a confocal microscope (TCS SP2 AOBS; Leica, Vienna) for the indicated times. HeLa cells transfected with a Rasa3–GFP construct were cultured in glass-bottom dishes (MatTek, Ashland, MA). Cell membrane-permeant derivative of Ins(1,3,4,5)P₄, Bt₂Ins(1,3,4,5)P₄/AM (50 μg/ml) (Echelon Bioscience, Salt Lake City, UT), was added onto the transfected cells and incubated for 30 min. A time-lapse video at the green fluorescence channel was taken every 1 min for 30 min on an IX71 fluorescent microscope (Olympus, Melville, NY) to monitor the translocation of Rasa3–GFP. The isoform Ins(1,2,4,5)P₄, Bt₂Ins(1,2,4,5)P₄/PM (Echelon Bioscience) and DMSO solvant were used as controls.

Abbreviations

Ins(1,3,4,5)P₄

inositol 1,3,4,5-tetrakisphosphate

Ins(1,4,5)P₃

inositol 1,4,5-trisphosphate

Itpkb

Ins(1,4,5)P₃ 3-kinase

PH

pleckstrin homology

NP

4-hydroxy-3-nitrophenylacetyl

BAFF

B cell-activating factor

BCR

B cell receptor.