Inositol 1,3,4,5-tetrakisphosphate is essential for T lymphocyte development

© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

ARTICLES

Inositol 1,3,4,5-tetrakisphosphate is essential for T lymphocyte development Valérie Pouillon1, Romana Hascakova-Bartova2, Bernard Pajak3, Emmanuelle Adam4, Françoise Bex5, Valérie Dewaste2, Carine Van Lint4, Oberdan Leo3, Christophe Erneux2 & Stéphane Schurmans1 Inositol 1,4,5-trisphosphate (Ins(1,4,5)P3) is phosphorylated by Ins(1,4,5)P3 3-kinase, generating inositol 1,3,4,5tetrakisphosphate (Ins(1,3,4,5)P4). The physiological function of Ins(1,3,4,5)P4 is still unclear, but it has been reported to be a potential modulator of calcium mobilization. Disruption of the gene encoding the ubiquitously expressed Ins(1,4,5)P3 3-kinase isoform B (Itpkb) in mice caused a severe T cell deficiency due to major alterations in thymocyte responsiveness and selection. However, we were unable to detect substantial defects in Ins(1,4,5)P3 amounts or calcium mobilization in Itpkb–/– thymocytes. These data indicate that Itpkb and Ins(1,3,4,5)P4 define an essential signaling pathway for T cell precursor responsiveness and development.

Ins(1,4,5)P3 is a calcium-mobilizing second messenger produced in response to cell stimulation. Two main pathways for Ins(1,4,5)P3 metabolism have been described so far, including dephosphorylation by Ins(1,4,5)P3 5-phosphatase and phosphorylation by Ins(1,4,5)P3 3-kinase, generating inositol 1,4-bisphosphate (Ins(1,4)P2) and Ins(1,3,4,5)P4, respectively1. The physiological function of these metabolites is still unclear, but Ins(1,3,4,5)P4 is reported to be a potential modulator of calcium mobilization1,2. In immune cells, activation of the Ins(1,4,5)P3 calcium-signaling pathway is essential for various effector functions, including proliferation, cytokine production and cytotoxicity. In thymocytes, the metabolism of inositol phosphates downstream of Ins(1,4,5)P3 through phosphorylation by Ins(1,4,5)P3 3-kinase is particularly well developed. Large amounts of Ins(1,3,4,5)P4 are generated in these cells after stimulation with concanavalin A (Con A)3, and the metabolism of inositol 1,3,4,5,6-pentakisphosphate (Ins(1,3,4,5,6)P5) and inositol hexakisphosphate (InsP6), two inositol phosphates resulting from specific phosphorylation and dephosphorylation of the obligatory Ins(1,3,4,5)P4 precursor, is very dynamic in rat thymocytes during cell cycle progression4. Three isoforms of the Ins(1,4,5)P3 3-kinase have been cloned and analyzed for their stimulation, tissue expression and subcellular localization5–10. These isoenzymes are characterized by the presence of a conserved catalytic unit in their C-terminal sequence, but can be distinguished by their N-terminal sequence and tissue distribution, indicating nonredundant functions in vivo6,7,9. The gene encoding isoform A of Ins(1,4,5)P3 3-kinase (Itpka) is expressed exclusively in specific neuronal subpopulations in the central nervous system and in testis9,11. Mice deficient in Itpka are viable and fertile, but develop an enhanced hippocampal CA1 long-term potentiation after

high-frequency neuronal stimulation12. Itpkb and Itpkc are expressed fairly ubiquitously. As no specific function has been linked to these isoforms in vivo, we generated and analyzed Itpkb- and Itpkc-deficient mice. RESULTS Generation of Itpkb-deficient mice We disrupted Itpkb in embryonic stem cells by homologous recombination (Fig. 1a). We inserted a neomycin-resistance cassette into the first coding exon in reverse orientation to the targeted gene, replacing a 0.3-kilobase (kb) fragment of Itpkb. We identified homologous recombinants by Southern blot analysis (Fig. 1b). Chimeric mice were obtained from a recombinant clone and the mutation was transmitted to the progeny. We verified the null mutation in Itpkb–/– mice by the absence of wild-type mRNA transcripts, at the expected size, in RNA hybridization analyses of the lung, where Itpkb has its highest expression (Fig. 1c). However, a higher-molecular-weight transcript was present in Itpkb–/– and Itpkb+/– mice. This transcript encompassed Itpkb as well as the neomycin-resistance cassette; we confirmed this result by RT-PCR analysis (Fig. 1d). This altered transcript could lead to the production of a truncated protein comprising the first 266 amino acids of Itpkb fused to 16 amino acids provided by the neomycin-resistance cassette. However, this truncated protein is most likely nonfunctional, as it lacks the entire catalytic unit. To confirm this, we measured Ins(1,4,5)P3 3-kinase activity in Itpkb+/+, Itpkb+/– and Itpkb–/– mice. We found an allele dosage–dependent decrease in Ins(1,4,5)P3 3-kinase activity in thymi from mice homozygous or heterozygous for a null allele in Itpkb (Fig. 1e). These results indicate that the mutated Itpkb allele generated is a null allele.

1IRIBHM,

IBMM, rue des Professeurs Jeener et Brachet 12, 6041 Gosselies, Belgium. 2IRIBHM, Campus Erasme, route de Lennik 808, 1070 Brussels, Belgium. de Physiologie Animale, IBMM, rue des Professeurs Jeener et Brachet 12, 6041 Gosselies, Belgium. 4Laboratoire de Chimie Biologique, IBMM, rue des Professeurs Jeener et Brachet 12, 6041 Gosselies, Belgium. 5Laboratoire de Microbiologie, CERIA, Avenue E. Gryson 1, 1070 Brussels, Belgium. Correspondence should be addressed to S.S. ([email protected]). 3Laboratoire

Published online 28 September 2003; doi:10.1038/ni980

1136

VOLUME 4 NUMBER 10 OCTOBER 2003 NATURE IMMUNOLOGY

ARTICLES

© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

a

b

c

d

e

Figure 1 Targeted disruption of mouse Itpkb. (a) Wild-type allele, targeting vector, predicted structure of a mutated allele, probe used in Southern blot analysis, DNA fragments generated after digestion with HindIII and EcoRV, and primers used in RT-PCR. Filled boxes, exons. E, EcoRI; H, HindIII; K, KpnI; S, SalI. (b) Southern blots of a recombinant embryonic stem clone. Left and right margins, length of genomic DNA fragments generated after digestion with restriction enzymes (above blot). (c) RNA hybridization analysis of total lung RNA isolated from Itpkb+/+, Itpkb+/– and Itpkb–/– mice. The probe was a 1.4-kb genomic DNA fragment localized in the first exon. (d) RT-PCR analysis of thymic mRNA with primers located in the 0.3-kb deleted region and in the third exon (top), or in the neor cassette and the third exon (bottom). (e) Ins(1,4,5)P3 3-kinase activity expressed as the amount of Ins(1,3,4,5)P4 produced by thymic cells from Itpkb+/+, Itpkb+/– and Itpkb–/– mice. One representative experiment of two is shown. Data are expressed as means ± s.e.m. of triplicate points *, P < 0.0001.

Itpkb–/– mice were born at the expected mendelian frequency and were indistinguishable from control mice at birth. However, these mice progressively developed reduced mobility, abdominal swelling, diarrhea and partial hair loss and became stunted for growth. Within the first 3 months of age, 50% of Itpkb–/– mice died, and nearly all were

a

b

dead before reaching 6 months of age (Fig. 2a). Macroscopic and microscopic examination showed that most of them had dilated stomachs and intestines containing many giardia parasites. In contrast, we found no giardia in the intestinal lumens of Itpkb+/– and Itpkb+/+ mice. As the control of acute giardia infection in mice is dependent on T cells13, we suspected a defect in T cell development and/or function. Thus, we did histological, immunohistochemical, biochemical and flow cytometric analyses of lymph nodes, spleens and thymi from Itpkb mutant mice. Impaired T cell development in Itpkb–/– mice Lymph nodes and spleen were characterized by a nearly complete absence of mature CD4+ and CD8+ T cells by immunohistochemistry and flow cytometry (Figs. 2b and 3a and data not shown). In contrast, B220+ B cells in secondary lymphoid organs and CD5+ B cells in the peritoneal cavity were still present in Itpkb–/– mice (data not shown). In the thymi of Itpkb–/– mice, there was a nearly complete absence of medulla, as shown by a reduced immunodetection of UEA-1, a lectin specific for a determinant on medullary epithelial cells14 (Fig. 2b). Total thymic cellularity in Itpkb–/– mice was increased at 4 weeks, compared with that of Itpkb+/– and Itpkb+/+ mice (Itpkb+/+, 2.3 ± 0.45 × 108, n = 4; Itpkb+/–, 1.8 ± 0.35 × 108, n = 3; Itpkb–/–, 3.7 ± 0.47 × 108, n = 5; P < 0.01). However, in older Itpkb–/– mice, the thymus seemed to involute earlier, possibly as a result of the stress caused by the infection. We noted a developmental block between double-positive and mature single-positive thymocytes in Itpkb–/– mice. Increased numbers and percentages of double-positive thymocytes were associated Figure 2 Characterization of Itpkb-deficient mice. (a) Survival curves of Itpkb mutant mice. (b) Immunohistochemical staining of the spleen and the thymus. Splenic sections were stained with anti-B220 (red) and anti-Thy1.2 (blue); thymic sections were stained with UEA-1 (blue).

NATURE IMMUNOLOGY VOLUME 4 NUMBER 10 OCTOBER 2003

1137

ARTICLES

© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

a

+/+

+/–

–/–

b

+/+ 18.9% +/– 10.5% –/– 1.8%**

+/+ 18.0% +/– 13.3% –/– 3.5%**

with considerably reduced numbers and percentages of CD4+ and CD8+ single-positive thymocytes in Itpkb–/– mice, compared with those of Itpkb+/+ mice (Fig. 3a and Supplementary Fig. 1 online). Itpkb+/– mice had intermediate numbers and percentages of singlepositive thymocytes, indicating that the efficiency of the maturation of double-positive thymocytes correlated with Itpkb expression. Cell size and expression of CD3int and T cell receptor (TCR)αβint as well as CD4 and CD8 were indistinguishable between wild-type and Itpkb–/– double-positive thymocytes. In contrast, CD3hi and TCRαβhi cells, which correspond to single-positive thymocytes, were almost absent in Itpkb–/– mice, and present in intermediate amounts in heterozygous mice (Fig. 3b). Finally, we found no additional block during the transition from double-negative to double-positive thymocytes in the mutant mice (Fig. 3a). Decreased Ins(1,3,4,5)P4 production by Itpkb–/– mice Stimulation of intact mouse thymocytes with Con A results in the rapid production of Ins(1,4,5)P3 and of large amounts of Ins(1,3,4,5)P4 (Fig. 4a; ref. 3). To define the effects of Itpkb deficiency on the concentration of Ins(1,4,5)P3 and Ins(1,3,4,5)P4, we stimulated thymocytes from Itpkb+/+ and Itpkb–/– mice with Con A, or left them unstimulated. Ins(1,4,5)P3 and Ins(1,3,4,5)P4 were undetectable in unstimulated thymocytes (Fig. 4a). After stimulation, Ins(1,4,5)P3 concentrations were not significantly different in Itpkb mutant and control mice, whereas Ins(1,3,4,5)P4 concentrations were

a

Figure 3 Impaired T cell development in Itpkb mutant mice. (a) Flow cytometry of thymocytes, splenocytes and lymph node cells. Numbers in quadrants indicate percentages of each population. Results are representative of a minimum of four mice per group. *, P = 0.001; **, P < 0.0001. (b) Flow cytometry of thymocytes. Top, comparison of size (FSC), CD4 and CD8 expression on gated double-positive thymocytes from Itpkb+/+ (thick lines), Itpkb+/– (thin lines) and Itpkb–/– (shaded lines) mice. Bottom, CD3 and TCRαβ expression on total thymocytes from Itpkb+/+ (thick lines), Itpkb+/– (thin lines) and Itpkb–/– (shaded lines) mice. Numbers indicate percentages of CD3hi and TCRαβhi cells for each genotype. Results are representative of a minimum of two mice per group. *, P < 0.05; **, P < 0.01. All fluorescence values are in log scale; FSC is in linear scale.

significantly lower in Itpkb–/– mice (Fig. 4a,b). We obtained similar results after stimulating cells with a CD3 monoclonal antibody (mAb 7D6; data not shown). These data indicate that deficiency in Itpkb does not alter Ins(1,4,5)P3 generation, but leads to decreased Ins(1,4,5)P3 3-kinase activity and Ins(1,3,4,5)P4 concentration after thymocyte stimulation. This deficiency results in abnormal T cell development characterized by a block during the transition from double-positive to single-positive thymocytes. The block is specific for T cells, as peripheral lymphoid organs still contained normal numbers of non-T cells, and as only concentrations of T cell–dependent immunoglobulin (Ig) isotypes (IgG1, IgG2a and IgG2b) were notably decreased in the sera of 8-week-old Itpkb–/– mice (data not shown). The thymic defect is intrinsic to hematopoietic cells We next investigated the involvement of thymic stromal cells in the phenotype of Itpkb–/– mice. We used severe combined immunodeficient (SCID) mice, which are characterized by the presence of a developmental block at the transition from double-negative to doublepositive thymocytes. We reconstituted these mice with fetal liver cells from either Itpkb+/+ or Itpkb–/– embryos, then 4 weeks later isolated reconstituted SCID thymus and peripheral lymphoid organs and analyzed them by flow cytometry. As in Itpkb–/– mice, a block still occurred at the transition from double-positive to single-positive thymocytes, and we found no mature T cells in the periphery of SCID mice injected

b

Figure 4 Production of inositol phosphates in thymocytes. (a) HPLC profiles from unstimulated (NS) Itpkb+/+ thymocytes, and from Itpkb+/+ and Itpkb–/– thymocytes stimulated for 5 min with Con A. (b) Relative amounts of Ins(1,4,5)P3 and Ins(1,3,4,5)P4 in Itpkb–/– thymocytes (filled bars) and Itpkb+/+ thymocytes (open bars) after stimulation. Data are expressed as means ± s.e.m. of duplicate points of one representative experiment of two. WT, wild-type. *, P < 0.01.

1138

VOLUME 4 NUMBER 10 OCTOBER 2003 NATURE IMMUNOLOGY

Figure 5 Reconstitution of SCID mice. Flow cytometry of thymocytes and splenocytes from SCID mice reconstituted with Itpkb+/+ or Itpkb–/– fetal liver cells or not reconstituted (NR). The thymic cellularity of SCID mice reconstituted with Itpkb+/+ or Itpkb–/– fetal liver cells was 57.0 ± 17.2 × 106 (n = 5) or 35.8 ± 24.2 × 106 (n = 5), respectively. Numbers in quadrants indicate percentages of each population.

with Itpkb–/– fetal cells (Fig. 5). Thus, mutated thymic stromal cells are not necessary for the development of the phenotype, and support the idea of the involvement of an intrinsic hematopoietic cell defect. Impaired thymocyte selection in Itpkb–/– thymi Thymocytes at the double-positive stage are confronted with positive and negative selection to ensure that only T cells bearing useful and harmless antigen receptors mature for export to the periphery15. To investigate the selection efficiency in Itpkb–/– mice, we crossed Itpkbdeficient mice with mice transgenic for the anti–H-Y TCR, which binds the Smcy-encoded H-Y peptide KCSRNRQYL (amino acids 738–746) presented by H-2Db major histocompatibility complex (MHC) class I antigen on cells derived from male mice16. Female Itpkb+/+ anti–H-Y TCR mice, lacking the male-specific H-Y antigen, provided the appropriate environment for positive selection of thymocytes expressing the anti–H-Y TCR (Fig. 6a,b). In contrast, female Itpkb+/– and Itpkb–/– +/+

+/–

–/–

b

anti–H-Y TCR mice partially or completely failed to positively select double-positive thymocytes and to develop CD8+ single-positive anti–H-Y TCR cells (Fig. 6a). We obtained similar results after crossing our Itpkb mutant mice with transgenic mice expressing the OT1 TCR specific for ovalbumin peptide SIINFEKL (amino acids 257–264) presented by H-2Kb MHC class I antigen17 (Fig. 6c). A hallmark of doublepositive thymocytes undergoing positive selection is the induction of CD69 expression18. Consistent with a substantial failure of positive selection, the Itpkb–/– double-positive thymocyte population contained fewer CD69+ cells than did Itpkb+/+ (Fig. 6d). We investigated the efficiency of negative selection in Itpkb–/– anti–H-Y TCR male mice. Negative selection seemed to proceed normally in mutant mice, as we found no difference in thymocytes subsets between control and Itpkb–/– mice (Fig. 6a,b). To analyze the efficiency of negative selection at low concentrations of self antigen, we cultured Itpkb-deficient anti–H-Y TCR thymocytes from female mice with H-2Db adherent splenic cells pulsed with decreasing concentrations of the specific H-Y deleting peptide. After 2 d, we collected thymocytes and estimated the percentage of surviving double-positive cells by flow cytometry for each peptide concentration (Fig. 6e). In mutant and control mice, the survival of double-positive thymocytes decreased with increasing concentrations of selecting peptide, from 10–10 to 10–3 M. However, at specific peptide concentrations, the percentage of surviving double-positive thymocytes was significantly higher in Itpkb+/– and Itpkb–/– mice than in Itpkb+/+ mice. Our results indicate that Itpkb is required for positive selection of double-positive thymocytes, and that the partial or complete Itpkb deficiency reduces the efficiency of negative selection, resulting in a resistance to deletion at low and intermediate concentrations of selecting peptide. Thus, the phenotype in Itpkb+/– and Itpkb–/– thymocytes is most likely the consequence of a lower response of thymocytes to antigens, or a higher threshold of activation, rather than an increased sensitivity to apoptotic stimuli. To investigate whether overexpression of the antiapoptotic Bcl-2 protein could overcome the thymocyte maturation defect, we crossed our Itpkb mutant mice with transgenic LckPr-Bcl2 mice. These mice express

c

+/+

–/–

d

CD4

CD4

Females

a

CD8

Males

© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

ARTICLES

CD8

e

Figure 6 Impaired positive and negative selection in Itpkb–/– thymocytes. (a) Flow cytometry of thymocytes from wild-type (+/+), Itpkb+/– and Itpkb–/– anti–H-Y TCR-transgenic mice in the context of H-2Db. Results are representative of a minimum of two mice per group. (b) Number of thymocytes in anti–H-Y TCR-transgenic mice. A minimum of two mice were analyzed per group. Data are expressed as means ± s.e.m. (c) Flow cytometry of thymocytes from Itpkb+/+ and Itpkb–/– OT1 TCR-transgenic mice. (d) Percentages of CD69+ double-positive (DP) thymocytes from wild-type (+/+), Itpkb+/– and Itpkb–/– mice. Four mice were analyzed per group. Data are expressed as means ± s.e.m. (e) Deletion of anti–H-Y TCR-transgenic thymocytes in vitro. Thymocytes from female Itpkb+/+ (n = 9), Itpkb+/– (n = 8) and Itpkb–/– (n = 3) mice were cocultured with H-2Db adherent splenic cells pulsed with the H-Y peptide or not pulsed (100% survival). For all panels, *, P < 0.05; **, P < 0.01; ***, P < 0.001. Percentages of living double-positive thymocytes after 2 d in culture with splenic cells in the absence of antigen were: Itpkb+/+, 39.7 ± 4.3%; Itpkb+/–, 65.1 ± 4.0%; Itpkb–/–, 88.4 ± 2.6%. In a and c, numbers in quadrants indicate percentages of each population.

NATURE IMMUNOLOGY VOLUME 4 NUMBER 10 OCTOBER 2003

1139

ARTICLES

© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

a

b

c

d

e

Figure 7 Calcium mobilization in thymocytes. Mean fluorescence emitted by the calcium-binding dye Fluo-3 was calculated for time intervals and is presented as ± s.e.m. (a) Stimulation of double-positive thymocytes from Itpkb+/+ (n = 3), Itpkb–/– (n = 2) and Itpkb–/– Itpkc–/– (n = 2) mice with CD3 mAb 7D6 (concentrations, above graph). (b) Stimulation of total thymocytes from female anti–H-Y TCR-transgenic Itpkb+/+ (n = 5) and Itpkb–/– (n = 2) mice with antigen-presenting cells pulsed with H-Y peptide (concentrations, above graph). Curve breaks indicate the time needed for mixing, pelleting and resuspending the cells. (c) Stimulation of double-positive thymocytes with ionomycin (300 ng/ml). (d) Stimulation of total thymocytes from Itpkb+/+ (n = 3) and Itpkb–/– (n = 3) mice with CD3 mAb 7D6 (2 µg/ml) in the presence of EGTA (1 mM). (e) Stimulation of Itpkb+/+ (n = 4) and Itpkb–/– (n = 5) total thymocytes with thapsigargin (1 µM).

Bcl2 under the control of the proximal promoter of the tyrosine kinase–encoding gene Lck, which drives expression of the transgene in immature cortical double-positive thymocytes. As a result, these thymocytes are protected from a wide variety of apoptotic stimuli, including glucocorticoids, radiation and treatment with antibody to CD3 (antiCD3). However, clonal deletion of self-reactive T cells still occurs. T cell maturation is altered, resulting in increased percentages of mature thymocytes skewed toward the CD8+ subset and increased numbers of T cells in spleen and lymph nodes19. Our mice crossed with these transgenic LckPr-Bcl2 mice also showed increased numbers of single-positive thymocytes and peripheral T lymphocytes (Supplementary Fig. 1 online). We also noted increased numbers of single-positive thymocytes in Itpkb–/– LckPr-Bcl2 mice. However, Bcl2 expression failed to rescue this T cell subset and the phenotype of the mutant mice, confirming the idea that an increased sensitivity to apoptotic stimuli was not responsible for the thymic phenotype in Itpkb–/– mice. Calcium mobilization in Itpkb–/– thymocytes The physiological function of Ins(1,3,4,5)P4 remains unclear, but most experiments so far indicate direct or indirect involvement in the control of calcium mobilization after cell stimulation and Ins(1,4,5)P3 production1,2,10,20–22. In our mutant mice, we found no effect of Itpkb deficiency on membrane TCR, CD4 or CD8 expression by double-positive thymocytes, or on TCR internalization after incubation of anti–H-Y TCR-transgenic thymocytes with EL-4 mouse lymphoma cells pulsed with the H-Y peptide as antigen-presenting cells (Fig. 3b and data not shown). Our results indicate that the defect most likely lies downstream of Ins(1,4,5)P3 in the TCR signaling cascade, as Ins(1,4,5)P3 production was normal in mutant thymocytes stimulated by Con A or CD3 mAb (7D6; Fig. 4). Thus, we monitored intracellular free calcium concentrations in thymocytes stimulated with the CD3 mAb 7D6 (ref. 23). Unexpectedly, we found no obvious difference between Itpkb+/+ and

1140

Itpkb–/– calcium responses in double-positive thymocytes. We obtained similar results over a wide range of CD3 mAb 7D6 concentrations (Fig. 7a), as well as after the more physiological stimulation of anti– H-Y (Fig. 7b) or OT1 (data not shown) TCR-transgenic thymocytes through contact with EL-4 cells pulsed with the KCSRNRQYL (H-Y) or the SIINFEKL (OT1) peptide. We found no difference in calcium mobilization between Itpkb–/– and Itpkb+/+ thymocytes after incubating the cells with ionomycin (Fig. 7c), or after stimulating them with the CD3 mAb 7D6 in the presence of EGTA (Fig. 7d) or with thapsigargin (Fig. 7e). Collectively, these observations indicate that in the experimental conditions tested, Ins(1,3,4,5)P4 might not have a principal function in calcium mobilization in thymocytes. NFATs are transcription factors linked to cytokine and early response gene expression in activated lymphocytes24. NFAT4 (or NFATc3), in contrast to NFATp and NFATc, is preferentially expressed in double-positive thymocytes, and is essential for the generation of single-positive thymocytes, as mice lacking this factor have impaired development of these cells25. The calcium-dependent nuclear transport of NFAT4 was still present in Itpkb–/– thymocytes in response to stimulation with CD3 and CD4 mAbs, with Con A or with calcium ionophore and phorbol 12-myristate 13-acetate (Supplementary Fig. 2 online). These results indicate that the phenotypic alterations in Itpkb–/– thymocytes might be the consequence of a signal transduction defect distal to the production of Ins(1,4,5)P3, which does not seem to affect calcium mobilization or NFAT4 transcription factor translocation in the experimental conditions tested. Experiments with mice deficient for the third isoform of the Ins(1,4,5)P3 3-kinase, the ubiquitous Itpkc6, excluded the possibility of involvement of this isoform during T cell development (Fig. 8). Thymocyte development in Itpkc–/– and Itpkc–/– anti–H-Y TCR mice was normal (data not shown). T cell development in

VOLUME 4 NUMBER 10 OCTOBER 2003 NATURE IMMUNOLOGY

ARTICLES InsP6 are produced in large amounts in thymocytes, but no distinct physiological functions have been attributed to these potential messengers3,4. Our results in Itpkb+/– and Itpkb–/– mice emphasize the importance of this enzyme and its reaction product during final thymocyte differentiation; it is likely that TCR engagement on double-positive thymocytes activates Itpkb and the Ins(1,3,4,5)P4 production required for final maturation to single-positive cells. The reduced number of UEA-1+ medullary epithelial cells in the thymi of Itpkb–/– mice is most likely secondary to the nearly complete absence of single-positive thymocytes. It may reflect the interdependence of thymocytes and thymic epithelial cell development, as thymocyte–thymic epithelial cell interactions not only promote thymocyte differenc b d tiation but also determine thymic epithelial cell differentiation26. Although we have now defined an important physiological function for Ins(1,3,4,5)P4 in thymocytes, the exact mechanism that mediates its action in these cells is still unknown. Ins(1,3,4,5)P4 is classically preFigure 8 Targeted disruption of Itpkc. (a) Wild-type allele, targeting vector, predicted structure of a sented as a modulator of calcium mobilizamutated allele, probe used in Southern blot analysis, DNA fragments generated after digestion with tion in various cells1,2,10,22. For example, in EcoRI and BamHI, and primers used in RT-PCR. Filled boxes, exons. E, EcoRI; H, HindIII; B, BamHI; S, HeLa cells, electroporation of Ins(1,3,4,5)P4 SalI; P, PstI. (b) Southern blots of a recombinant embryonic stem clone. Left and right margins, length of causes a transient increase in the frequency of the genomic DNA fragments generated after digestion with restriction enzymes (above blot). (c) RT-PCR analysis of testis mRNA with primers located in the first and in the third exons. (d) Ins(1,4,5)P3 3-kinase calcium oscillations in response to histaactivity, shown as the amount of Ins(1,3,4,5)P4 produced by intestine or thymic cells from Itpkc+/+, mine20. In RBL-2H3 cells, Ins(1,3,4,5)P4 Itpkc+/– and Itpkc–/– mice. Data represent the means of three measurements. *, P < 0.0001. facilitates store-operated calcium influx after carbachol stimulation, by inhibition of Itpkb–/–Itpkc–/– mice did not demonstrate additional alterations beyond Ins(1,4,5)P3 5-phosphatase, thereby protecting Ins(1,4,5)P3 against those of Itpkb–/– mice, and we found no calcium-mobilization defect in hydrolysis21. In platelets, application of Ins(1,3,4,5)P4 on purified the thymocytes of these mice after CD3 mAb 7D6 or ionomycin stimu- plasma membranes induces a calcium flux, distinguishable from that lation (Fig.7a,c and data not shown). RNA hybridization analyses have induced by Ins(1,4,5)P3 (ref. 27). Although in other experimental sysshown that this isoform is not strongly expressed in the thymus and, tems a definite function for Ins(1,3,4,5)P4 has been demonstrated in accordingly, Ins(1,4,5)P3 3-kinase activity was not diminished in the modulation of calcium mobilization, here stimuli like Con A, CD3 thymic extracts from Itpkc–/– mice compared with that of Itpkc+/+ mice mAb or pulsed antigen-presenting cells induced a similar calcium (Fig. 8d and data not shown). However, in the intestine, where Itpkc is mobilization in Itpkb+/+ and Itpkb–/– thymocytes. These results indiwell expressed, Ins(1,4,5)P3 3-kinase activity was notably reduced in cate that modulation of calcium mobilization is not the mechanism that mediates the Ins(1,3,4,5)P4 action in these cells. However, we canthe mutant mice, confirming inactivation of Itpkc (Fig. 8d). not definitively exclude the possibility that our experimental design DISCUSSION masks subtle calcium mobilization defects. Indeed, the stimuli used in We report here that mice deficient in Itpkb, the B isoform of the our studies might be too strong compared with those occurring in vivo Ins(1,4,5)P3 3-kinase that generates Ins(1,3,4,5)P4 by phosphorylation during double-positive thymocyte selection. They still could generate of the second messenger Ins(1,4,5)P3, have a complete and specific T sufficient Ins(1,3,4,5)P4 to enable a normal calcium mobilization; cell deficiency because of a developmental block at the double-positive decreased Ins(1,3,4,5)P4 production might induce calcium-mobilizathymocyte stage. The block is associated with a notably decreased con- tion defects only when it falls beneath a certain threshold. In addition centration of Ins(1,3,4,5)P4 in stimulated thymocytes, and with a to the type of stimuli used in our study, the method itself has some reduced responsiveness of Itpkb–/– thymocytes to peptide antigens, limitations; monitoring calcium concentrations in a population of leading to substantial alterations in positive and negative selection. cells precludes precise analysis in single cells of calcium oscillation patThus, our results define Ins(1,3,4,5)P4 as an essential messenger dur- terns, whose frequencies are important for transcription factor activation28. ing T cell selection and differentiation in the thymus. The Ins(1,4,5)P3 calcium-signaling pathway in immature and Ins(1,3,4,5)P4 production is not always associated with modification mature T cells is a chief pathway that, after activation by TCR stimu- in calcium concentration, and control of calcium mobilization is not the lation, leads to specific gene expression and cytokine production as sole function proposed for Ins(1,3,4,5)P4. Adriamycin, an anthracycline well as cell differentiation and proliferation. Ins(1,3,4,5)P4 and its antibiotic used in anti-tumor treatment, is a potential and nonspecific downstream phosphorylation products such as Ins(1,3,4,5,6)P5 and Ins(1,4,5)P3 3-kinase inhibitor29. Although Ins(1,3,4,5)P4 concentration

© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

a

NATURE IMMUNOLOGY VOLUME 4 NUMBER 10 OCTOBER 2003

1141

© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

ARTICLES in Jurkat T lymphocytes is considerably decreased after incubation with adriamycin, it is unable to modify calcium concentration or calcium influx after T cell receptor stimulation. Candidate receptors have been described for Ins(1,3,4,5)P4 and may mediate its cellular effects. Among these, GAP1IP4BP is a ubiquitously expressed GTPase-activating protein of the GAP1 family that is highly expressed in peripheral blood leukocytes22,30,31. This GAP protein is active against Ras and Rap, both of which seem to be involved in thymocyte selection32–34. Various in vitro approaches have been used to manipulate the concentration of endogenous GAP1IP4BP in HeLa cells. However, no detectable effect of this protein has been noted on intracellular calcium concentration after histamine stimulation35. Thus, the GAP1IP4BP receptor might represent a good candidate for being responsible for the effects of Ins(1,3,4,5)P4 in thymocytes, as it is expressed in leukocytes, is active on Ras and Rap molecules, which are important during thymocyte differentiation, and does not interfere with calcium mobilization at least in HeLa cells stimulated by histamine. RNA hybridization studies have shown that Itpkb and Itpkc are both expressed in the thymus (ref. 9 and V.P., unpublished data). Our results in Itpkb–/–, Itpkc–/– and Itpkb–/– Itpkc–/– mice emphasize the exquisite specificity of the B isoform of Ins(1,4,5)P3 3-kinase in the observed thymic defects. They indicate that the proper intracellular localization10 and hence Ins(1,3,4,5)P4 production, or specific protein interactions mediated by the unique N-terminal end of these isoforms6,7, are of great importance for the in vivo functions of these enzymes. In conclusion, using genetically modified mice, we have defined the physiological function of Itpkb and its reaction product Ins(1,3,4,5)P4 in the thymus. Itpkb and Ins(1,3,4,5)P4 were essential during positive and negative selection of double-positive thymocytes, and in the control of thymocyte reactivity to antigens. As Itpkb is also expressed in mature T cells and Ins(1,3,4,5)P4 is produced in these cells after TCR stimulation (C.E., unpublished data), it is important now to define the putative signaling molecules downstream of Ins(1,3,4,5)P4 and its potential receptors in thymocytes and mature T cells, as well as to understand the signals mediated by Ins(1,3,4,5)P4 that control their responsiveness to antigens. METHODS Mice. For the production of Itpkb–/– mice, the targeting vector replaced a 0.3-kb genomic fragment from the first coding exon with a neomycin resistance (neor) cassette in reverse orientation relative to Itpkb transcription. A similar strategy was used for the production of Itpkc–/– mice, except that the neor cassette replaced a 1.1-kb genomic fragment comprising the end of the first coding exon. Transfection of R1 embryonic stem cells and production of chimeric mice were done as described36. Transgenic mice expressing anti–H-Y TCR (specific for the male H-Y peptide KCSRNRQYL (amino acids 738–746 of Smcyencoded protein) presented by H-2Db MHC class I antigens)16, OT1 TCR (specific for the peptide SIINFEKL (amino acids 738–746 of ovalbumin) presented by H-2Kb MHC class I antigens)17, and LckPr-Bcl2 transgenic mice19 have been described. C.B.-17 SCID mice were purchased from Harlan. Itpkb+/+, Itpkb+/– and Itpkb–/– mice were analyzed on a 129/CD1 background, except for crossings with transgenic mice, for which mice backcrossed on a C57BL/6 background were used. Mice 4 weeks of age were used, unless otherwise specified. Mice were maintained in a conventional nonspecific-pathogen-free animal room. All animal studies were authorized by the Animal Care Use and Review Committee of the Free University of Brussels. Statistics. Student’s t-test was used for comparison of means. Ins(1,4,5)P3 3-kinase activity. The assay of Ins(1,4,5)P3 3-kinase activity was done as described37. Staining of cells with fluorescent antibodies. Cells were stained with fluorescein isothiocyanate– or phycoerythrin-conjugated antibodies to CD4 (RM4-5), CD8 (53-6.7), TCRβ (H57-597), CD69 (H1.2F3), B220 (RA3-6B2) and CD5

1142

(53-7.3; all from BD Pharmingen), and CD3 mAb (2C11; ref. 38), detected by flow cytometry in a FACScan and analyzed with WinMDI software (http://facs.scripps.edu/software.html). Immunohistochemistry of spleen, lymph nodes and thymus. Samples were processed as described39. UEA-1 was purchased from Vector Laboratories. Determination of inositol phosphates in intact cells. Mouse thymocytes (3 × 107 cells) were labeled for 20 h with [3H]inositol (50 µCi/ml; NEN). Cells were stimulated for 5 min with Con A (10 µg/ml, Sigma) or with CD3 mAb 7D6 in a total volume of 0.6 ml. Reactions were quenched with 133 µl perchloric acid (9%) and were analyzed by HPLC on Partisphere SAX column (25 cm × 4.6 mm). Samples were eluted as follows with gradients generated by mixing buffer A (1 mM Na2 EDTA) with buffer B (buffer A plus 2 M (NH4)H2PO4, pH 3.35, with H3PO4): 0–3 min, 0% buffer B; 3–123 min, 0–65% buffer B; 123–140 min, 65–100% buffer B; 140–145 min 100–0% buffer B. Radioactivity was detected with an online detector from Raytest. Data were normalized for the incorporation of tritiated phosphoinositides as determined for each condition. The identity of each peak was confirmed with labeled standards prepared as reported6. Reconstitution of SCID mice. Livers from embryonic day 12.5 embryos were disrupted using a syringe with 100 µl PBS. The homogenate was injected retroorbitally into 6- to 8-week-old SCID mice irradiated at 200 rads. The lymphoid compositions of thymi, spleens and lymph nodes were determined 4 weeks later by flow cytometry. Intracellular calcium analysis. Intracellular calcium concentrations were estimated by the determination of fluorescence emitted by the calciumbinding dye Fluo-3, measured with a FACScan in either total thymocytes or on flow cytometry–sorted or gated double-positive thymocytes. The dyeloading procedure was adapted from a published protocol40. Cells were washed twice in calcium- and magnesium-free HBSS (Life Technologies) and were incubated at 37 °C for 30 min at a density of 1 × 107 cells/ml with pluronic acid (100 µg/ml; Sigma) and Fluo-3 (7.5 µM; Molecular Probes). Cells were then washed twice in complete medium and resuspended at a density of 5 × 105 cells/ml. Non-TCR-transgenic thymocytes were stimulated with CD3 mAb (7D6; ref. 23) cross-linked with rabbit anti-mouse IgG in the presence or absence of EGTA (1 mM), with ionomycin (300 ng/ml; A23187; Sigma) or with thapsigargin (1 µM, Sigma). For stimulation by antigenpresenting cells, TCR-transgenic thymocytes were mixed on ice with an equal number of EL-4 mouse lymphoma cells loaded for 2 h at 37 °C with the appropriate peptide. The cells were pelleted together by centrifugation at 4 °C for 1 min at 2,000g. Cells were then resuspended, warmed to 37 °C and immediately analyzed by FACScan. Confocal microscopy. After overnight culture, thymocytes were left unstimulated or were stimulated for 1 h at 37 °C with a combination of phorbol 12-myristate 13-acetate (5 ng/ml; Sigma) and calcium ionophore (130 ng/ml; A23187; Sigma), with CD3 mAb (4 µg/ml; 7D6) and CD4 mAb (GK1.5) cross-linked with donkey anti-mouse IgG (1:100 dilution; Jackson ImmunoResearch Laboratories), or with Con A (5 µg/ml; Sigma). Cells were then centrifuged on coated slides (Shandon) in a cytospin at 100g for 5 min, permeabilized and fixed with methanol at –20 °C for 6 min. The preparations were washed three times with PBS and blocked for 30 min in PBS containing 0.5% gelatin and 0.25% BSA. Specimens were incubated for 1 h at room temperature with the primary antibody rabbit polyclonal antiNFATc3 (1:500 dilution; M-75X; Santa Cruz), washed three times with 0.2% gelatin in PBS, and incubated for another 1 h with the secondary antibody goat anti-rabbit IgG Alexa 488 (1:100 dilution; Molecular Probes). Samples were washed three times and then mounted in a solution of 1 mg/ml of p-phenylenediamine in 90% glycerol. Preparations were visualized by confocal microscopy (Zeiss LSM510). Note: Supplementary information is available on the Nature Immunology website. ACKNOWLEDGMENTS We thank E. Marion, O. Giot, C. Jacques and C. Moreau for technical assistance; the staff of the Laboratoire de Physiologie Animale for immunohistochemistry,

VOLUME 4 NUMBER 10 OCTOBER 2003 NATURE IMMUNOLOGY

ARTICLES technical discussions and help; A. Nagy for R1 embryonic stem cells; K. Rajewsky and R. Merino for anti–H-Y TCR and LckPr–Bcl2 transgenic mice, respectively; and J. Penninger for reviewing the manuscript. Supported by the Fondation David et Alice Van Buuren and the Fondation Hoguet (V.P.), and the Fonds de la Recherche Scientifique Médicale de Belgique, Interreg II (cofinanced by the Région Wallonne and the European Commission, FEDER), Action de Recherche Concertée of the Communauté Française de Belgique, and The Free University of Brussels.

© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

COMPETING INTERESTS STATEMENT The authors declare that they have no competing financial interests. Received 14 May; accepted 26 August 2003 Published online at http://www.nature.com/natureimmunology/

1

Irvine, R.F. & Schell, M.J. Back in the water: the return of the inositol phosphates. Nat. Rev. Mol. Cell Biol. 2, 327–338 (2001). 2. Shears, S.B. The versatility of inositol phosphates as cellular signals. Biochim. Biophys. Acta 1436, 49–67 (1998). 3. Zilberman, Y., Howe, L.R., Moore, J.P., Hesketh, T.R. & Metcalfe, J.C. Calcium regulates inositol 1,3,4,5-tetrakisphosphate production in lysed thymocytes and in intact cells stimulated with concanavalin A. EMBO J. 6, 957–962 (1987). 4. Guse, A.H., Greiner, E., Emmrich, F. & Brand, K. Mass changes of inositol 1,3,4,5,6pentakisphosphate and inositol hexakisphosphate during cell cycle progression in rat thymocytes. J. Biol. Chem. 268, 7129–7133 (1993). 5. Communi, D., Vanweyenberg, V. & Erneux, C. Molecular study and regulation of D-myo-inositol 1,4,5-trisphosphate 3-kinase. Cell Signal. 7, 643–650 (1995). 6. Dewaste, V. et al. Cloning and expression of a cDNA encoding human inositol 1,4,5trisphosphate 3-kinase C. Biochem. J. 352, 343–351 (2000). 7. Dewaste, V., Roymans, D., Moreau, C. & Erneux, C. Cloning and expression of a fulllength cDNA encoding human inositol 1,4,5-trisphosphate 3-kinase B. Biochem. Biophys. Res. Commun. 291, 400–405 (2002). 8. Takazawa, K., Perret, J., Dumont, J.E. & Erneux, C. Molecular cloning and expression of a human brain inositol 1,4,5-trisphosphate 3-kinase. Biochem. Biophys. Res. Commun. 174, 529–535 (1991). 9. Vanweyenberg, V., Communi, D., D’Santos, C.S. & Erneux, C. Tissue- and cell-specific expression of Ins(1,4,5)P3 3-kinase isoenzymes. Biochem. J. 306, 429–435 (1995). 10. Dewaste, V. et al. The three isoenzymes of human inositol 1,4,5-trisphosphate 3kinase show specific intracellular localization but comparable Ca2+ responses upon transfection in COS-7 cells. Biochem. J. 374, 41–49 (2003). 11. Mailleux, P., Takazawa, K., Erneux, C. & Vanderhaeghen, J.J. Inositol 1,4,5-trisphosphate 3-kinase distribution in the rat brain. High levels in the hippocampal CA1 pyramidal and cerebellar Purkinje cells suggest its involvement in some memory processes. Brain Res. 539, 203–210 (1991). 12. Jun, K. et al. Enhanced hippocampal CA1 LTP but normal spatial learning in inositol 1,4,5-trisphosphate 3-kinase(A)-deficient mice. Learn. Mem. 5, 317–330 (1998). 13. Singer, S.M. & Nash, T.E. T-cell-dependent control of acute Giardia lamblia infections in mice. Infect. Immun. 68, 170–175 (2000). 14. Farr, A.G. & Anderson, S.K. Epithelial heterogeneity in the murine thymus: fucose-specific lectins bind medullary epithelial cells. J. Immunol. 134, 2971–2977 (1985). 15. Starr, T.K., Jameson, S.C. & Hogquist, K.A. Positive and negative selection of T cells. Annu. Rev. Immunol. 21, 139–176 (2003). 16. Kisielow, P., Bluthmann, H., Staerz, U.D., Steinmetz, M. & von Boehmer, H. Tolerance in T-cell-receptor transgenic mice involves deletion of nonmature CD4+8+ thymocytes. Nature 333, 742–746 (1988). 17. Hogquist, K.A. et al. T cell receptor antagonist peptides induce positive selection. Cell 76, 17–27 (1994).

NATURE IMMUNOLOGY VOLUME 4 NUMBER 10 OCTOBER 2003

18. Swat, W., Dessing, M., von Boehmer, H. & Kisielow, P. CD69 expression during selection and maturation of CD4+8+ thymocytes. Eur. J. Immunol. 23, 739–746 (1993). 19. Sentman, C.L., Shutter, J.R., Hockenbery, D., Kanagawa, O. & Korsmeyer, S.J. bcl-2 inhibits multiple forms of apoptosis but not negative selection in thymocytes. Cell 67, 879–888 (1991). 20. Zhu, D.M., Tekle, E., Chock, P.B. & Huang, C.Y. Reversible phosphorylation as a controlling factor for sustaining calcium oscillations in HeLa cells: Involvement of calmodulin-dependent kinase II and a calyculin A-inhibitable phosphatase. Biochemistry 35, 7214–7223 (1996). 21. Hermosura, M.C. et al. InsP4 facilitates store-operated calcium influx by inhibition of InsP3 5-phosphatase. Nature 408, 735–740 (2000). 22. Loomis-Husselbee, J.W. et al. Modulation of Ins(2,4,5)P3-stimulated Ca2+ mobilization by ins(1,3,4,5)P4: enhancement by activated G-proteins, and evidence for the involvement of a GAP1 protein, a putative Ins(1,3,4,5)P4 receptor. Biochem. J. 331, 947–952 (1998). 23. Coulie, P.G. et al. Identification of a murine monoclonal antibody specific for an allotypic determinant on mouse CD3. Eur. J. Immunol. 21, 1703–1709 (1991). 24. Serfling, E. et al. The role of NF-AT transcription factors in T cell activation and differentiation. Biochim. Biophys. Acta 1498, 1–18 (2000). 25. Oukka, M. et al. The transcription factor NFAT4 is involved in the generation and survival of T cells. Immunity. 9, 295–304 (1998). 26. van Ewijk, W., Shores, E.W. & Singer, A. Crosstalk in the mouse thymus. Immunol. Today 15, 214–217 (1994). 27. El Daher, S.S. et al. Distinct localization and function of (1,4,5)IP3 receptor subtypes and the (1,3,4,5)IP4 receptor GAP1(IP4BP) in highly purified human platelet membranes. Blood 95, 3412–3422 (2000). 28. Dolmetsch, R.E., Xu, K. & Lewis, R.S. Calcium oscillations increase the efficiency and specificity of gene expression. Nature 392, 933–936 (1998). 29. da Silva, C.P., Emmrich, F. & Guse, A.H. Adriamycin inhibits inositol 1,4,5-trisphosphate 3-kinase activity in vitro and blocks formation of inositol 1,3,4,5-tetrakisphosphate in stimulated Jurkat T-lymphocytes. Does inositol 1,3,4,5-tetrakisphosphate play a role in Ca2+-entry? J. Biol. Chem. 269, 12521–12526 (1994). 30. Cullen, P.J. et al. Identification of a specific Ins(1,3,4,5)P4-binding protein as a member of the GAP1 family. Nature 376, 527–530 (1995). 31. Cullen, P.J. Bridging the GAP in inositol 1,3,4,5-tetrakisphosphate signalling. Biochim. Biophys. Acta 1436, 35–47 (1998). 32. Amsen, D., Kruisbeek, A., Bos, J.L. & Reedquist, K. Activation of the Ras-related GTPase Rap1 by thymocyte TCR engagement and during selection. Eur. J. Immunol. 30, 2832–2841 (2000). 33. Dong, C., Davis, R.J. & Flavell, R.A. MAP kinases in the immune response. Annu. Rev. Immunol. 20, 55–72 (2002). 34. Werlen, G., Hausmann, B., Naeher, D. & Palmer, E. Signaling life and death in the thymus: timing is everything. Science 299, 1859–1863 (2003). 35. Walker, S.A. et al. Analyzing the role of the putative inositol 1,3,4,5-tetrakisphosphate receptor GAP1IP4BP in intracellular Ca2+ homeostasis. J. Biol. Chem. 277, 48779–48785 (2002). 36. Clement, S. et al. The lipid phosphatase SHIP2 controls insulin sensitivity. Nature 409, 92–97 (2001). 37. Takazawa, K. et al. Rat brain inositol 1,4,5-trisphosphate 3-kinase. Ca2+-sensitivity, purification and antibody production. Biochem. J. 268, 213–217 (1990). 38. Leo, O., Foo, M., Sachs, D.H., Samelson, L.E. & Bluestone, J.A. Identification of a monoclonal antibody specific for a murine T3 polypeptide. Proc. Natl. Acad. Sci. USA 84, 1374–1378 (1987). 39. Pajak, B. et al. Immunohistowax processing, a new fixation and embedding method for light microscopy, which preserves antigen immunoreactivity and morphological structures: visualisation of dendritic cells in peripheral organs. J. Clin. Pathol. 53, 518–524 (2000). 40. Baus, E., Urbain, J., Leo, O. & Andris, F. Flow cytometric measurement of calcium influx in murine T cell hybrids using Fluo-3 and an organic-anion transport inhibitor. J. Immunol. Methods 173, 41–47 (1994).

1143