|Home | About | Journals | Submit | Contact Us | Français|
The membrane lipid phosphatidylinositol(3,4,5)trisphosphate (PtdInsP3) regulates membrane receptor signalling in many cells, including immunoreceptor signalling. Here, we review recent data that have indicated essential functions for the soluble PtdInsP3 analogue inositol(1,3,4,5)tetrakisphosphate (InsP4) in T cell, B cell and neutrophil development and function. Decreased InsP4 production in immunocytes causes immunodeficiency in mice and might contribute to inflammatory vasculitis in Kawasaki disease in humans. InsP4-producing kinases could therefore provide attractive drug targets for inflammatory and infectious diseases.
Because it can be combinatorially phosphorylated on six hydroxyl groups, the cyclic poly-alcohol D-myo-inositol is an ideal cellular information carrier. Its many derivatives form an “inositol code”, whose most prominent members are soluble inositol phosphates (InsPs) [G] and their membrane-lipid counterparts, the phosphoinositides [G]1, 2 (Fig. 1, Box 1). In 1983, the identification of inositol(1,4,5)trisphosphate [Ins(1,4,5)P3, InsP3 here] as a second-messenger [G] that mediates the receptor-induced Ca2+ mobilization in most mammalian cells first demonstrated physiological importance for soluble InsPs3. More recently, intriguing functions for several higher-order InsPs have been identified (Box 1). Moreover, the phosphoinositide-lipids phosphatidylinositol(4,5)bisphosphate (PtdIns(4,5)P2) and PtdIns(3,4,5)P3 (PtdInsP3) regulate signalling by many receptors in many cell types4–9. PtdInsP3 is generated by phosphoinositide 3-kinase (PI3K) and metabolized by lipid-phosphatases, including phosphatase-and-tensin-homologue (PTEN), SH2-domain-containing inositol polyphosphate 5-phosphatase 1 (SHIP1) and SHIP210. The key to phosphoinositide function is the phosphorylation status of their cytosol-exposed InsP headgroups. These bind to specific domains in signalling or cytoskeletal proteins to control their membrane recruitment. Defects in phosphoinositide metabolism or function contribute to many diseases, including cancer, leukaemias, immunodeficiencies, autoimmune, neurodegenerative, allergic and inflammatory disorders4–9.
The figure below summarizes important aspects of the mammalian higher-order InsP metabolism with known enzymes. For additional details, see REF.44, 74, 76, 114 Orange, enzymes with genetically documented functions in immunocytes. Purple, InsPs found in immunocytes. Bold purple, antigen-receptor modulated InsPs17, 18, 23, 25, 45, 58–66, 115.
Higher-order InsPs have recently been implicated in regulating major cellular processes in eukaryotes:
InsPs may modulate protein function by acting as
Most of these findings were obtained in yeast or non-immunocytes, reviewed in REFs.1, 14, 33, 44–47, 74, 76, 117, 137. InsP-pyrophosphorylation further expands the inositol code and enablespyrophosphate-transfer onto protein substrates. It will be interesting to determine whether InsP3 3-kinase deficiency affects specific higher-order InsP-isomer levels in immunocytes, and to analyze whether altered isomers mediate aspects of InsP4 function. In one study, ConA-stimulatedItpkB−/− thymocytes showed altered InsP4, Ins(1,3,4)P3 and InsP2 levels. InsP3 and bulk InsP5 appeared unaffected23.
Here, we review recent findings that have unveiled important roles for soluble inositol(1,3,4,5)tetrakisphosphate (Ins(1,3,4,5)P4, InsP4) in the immune system. Decreased InsP4 production in immunocytes causes immunodeficiency in mice and might contribute to inflammatory vasculitis in Kawasaki disease [G]. We first briefly introduce InsP4 and the kinases that produce it; REF.11–15 provide more details. Next, we review recently discovered roles for InsP4 in controlling T cell, B cell and neutrophil development and function16–24. We propose a model where InsP4 at least in part governs immunocyte development by controlling PI3K, PTEN, SHIP1 and SHIP2 function as a soluble PtdInsP3 analogue, and by limiting store-operated Ca2+-entry (SOCE) [G] in B cells and neutrophils. We then discuss how seemingly opposite InsP4 functions in thymocytes and developing B cells establish appropriate selection windows that are crucial for the generation of competent, self-tolerant T and B cell repertoires. Finally, we review the therapeutic opportunities of targeting InsP4 and summarize important open questions about this intriguing soluble messenger molecule.
The cellular concentrations of multiple InsPs are altered after immunoreceptor stimulation (Box 1). For example, T or B cell receptor (TCR or BCR) stimulation activates phospholipases Cγ (PLCγ1, PLCγ2), which then hydrolyse membrane PtdIns(4,5)P2 into two second messenger molecules: the membrane-lipid diacylglycerol (DAG) and soluble InsP3 (Fig. 1)25–28. DAG recruits and activates key signalling mediators such as the RAS-activator RASGRP1 and protein kinases C (PKCs). InsP3 binding to InsP3-receptors mobilizes Ca2+. PLCγ, RASGRP1, PKCs and Ca2+ mobilization are all important for lymphocyte development and function29, 30.
TCR/BCR-engagement also triggers membrane recruitment and activation of PI3Ks, which phosphorylate PtdIns(4,5)P2 at the 3-position of its inositol-ring into PtdInsP3. PtdInsP3 then recruits several important signalling proteins to cellular membranes by binding to their pleckstrin-homology (PH) [G] or other domains31. Among these, the TEC-family protein-tyrosine-kinases (TFKs) ITK [G], TEC and BTK are essential for antigen-receptor signalling. In particular, TFKs phosphorylate and activate PLCγ1 and PLCγ227, 28. Furthermore, PtdInsP3-binding to the AKT PH domain controls recruitment and activation of this key regulator of lymphocyte survival, proliferation, differentiation and activation5, 6, 32.
The crucial roles of PI3K, PTEN, SHIP1 and SHIP2 in immunocyte development, survival, proliferation and function, and in preventing leukaemia, indicate the importance of carefully controlling PtdInsP3 turnover in immunocytes (Fig. 1, Box 1)5, 6, 32.
Besides mediating Ca2+ release, InsP3 can also be converted to InsP4 by one of four mammalian InsP3 3-kinases: the IP3-3K/IP3K/ITPK gene-family members A, B and C, and the distantly related, predominantly nuclear InsP multikinase (IPMK/IPK2) (Fig. 1,,2,2, Box 1, Table 1)11–15, 25, 33–39. ITPKA is abundant in the nervous system. It was also found in myeloid precursors with little expression in lymphoid tissues18, 20. ItpkA−/− mice showed altered long-term potentiation [G] in the hippocampus and memory deficits without reported immunological phenotypes40, 41. ITPKB is abundant in brain and haematopoietic cells18, 33. ItpkB−/− mice showed defects in lymphocyte development and myelopoiesis16–23. Published data do not indicate an important role for the ubiquitously expressed ITPKC in mouse development, viability, fertility or immune function23. ItpkC−/− thymocytes showed unaltered InsP3 3-kinase activity23. In humans, however, genetic data may indicate an intriguing ITPKC role in inhibiting peripheral T cell function42. No roles for IPMK in the immune system have been reported (Table 1), although IPMK might regulate JAK–STAT signalling in flies43.
Biochemical studies and genetic studies in yeast suggest many potential molecular functions for InsP4, reviewed in REFs.1, 11–14, 33–35, 44–47. InsP4 binding to their PH domains may regulate membrane recruitment, activation or protein-interactions of Ras GTPase-activating proteins (RASA2/GAP1m, RASA3/GAP1IP4BP)48, 49, BTK, AKT, or certain regulators of vesicular trafficking (synaptotagmins, centaurin-α1/p42IP4). A long-standing controversy concerns InsP3 3-kinase roles in regulating Ca2+ mobilization in mammalian cells14, 44, 45. Early studies suggested that they limit InsP3 accumulation by converting it into InsP4. Others suggested that InsP4 or a metabolite inhibit Ca2+ mobilization, possibly by anatagonizing with InsP3-binding to InsP3-receptors50, 51. In rat hepatocytes, InsP4can induce nuclear Ca2+-uptake52. In other studies, InsP4 augmented InsP3-mediated Ca2+ mobilization through an enigmatic mechanism or might even mimick InsP344, 45. Genetic data may argue against RASA3-involvement53. In RBL-2H3 or Jurkat cells, exogenous InsP4 facilitated InsP3-induced SOCE by inhibiting an InsP3-metabolizing 5-phosphatase (Fig. 3)51. Interestingly, InsP4 alone could not induce SOCE in these cells51. Conversely, InsP4 directly activated Ca2+-channels in endothelial or neuronal cells54–57. Thus, InsP3 3-kinases and InsP4 can have diverse effects on Ca2+-mobilization depending on experimental conditions, cell type, upstream-receptor, their subcellular localization, composition of the InsP3-receptors or Ca2+-channels involved, or other factors51. This demonstrates the need for genetic and biochemical in vivo studies with purified, primary cells.
Similar to PtdInsP3, InsP4 can be metabolized through dephosphorylation. The best characterized route is its dephosphorylation by a 5-phosphatase into Ins(1,3,4)P3, a precursor for many other InsPs (Box 1)17, 18, 23, 25, 45, 58–66. Several phosphatases can dephosphorylate the 5-positions of InsP3 and InsP467. Which phosphatases do this in immunocytes is unknown. In vitro, the immunocyte-expressed SHIP1, SHIP2 or PTEN can dephosphorylate InsP4 to Ins(1,3,4)P3 or Ins(1,4,5)P3/InsP3, respectively (Fig. 1)68–71. However, the PTEN/SHIP-deficiency of Jurkat cells72 which contain an InsP4-inhibitable 5-phosphatase51 indicates relevance for other enzymes.
Besides acting as an InsP3 3-kinase, or a non-conventional PI3K of uncertain physiological relevance73, IPMK can also phosphorylate InsP3, InsP4, Ins(1,4,5,6)P4 or Ins(1,3,4,6)P4 at the 3, 5 or 6-positions to generate Ins(1,3,4,5,6)P511–14, 25, 33–36, 39, 74. This likely is the major eukaryotic pathway for producing higher-order InsPs (Box 1)14, 75–77. Because of its early position in this pathway, some InsP4 functions might really be mediated by other InsPs.
Studies of ItpkB−/− mice recently unveiled important in vivo functions of ITPKB and InsP4 and provided novel insight into how InsP4 acts molecularly.
ItpkB−/− mice independently generated by two groups show severe peripheral T cell lymphopaenia and immunodeficiency. The cause is a profound block of T cell development at the CD4+CD8+ double-positive (DP) thymocyte-stage due to impaired positive selection17, 18, 23. Retrovirally expressed17 wildtype but not catalytically impaired ITPKB restored the development of ItpkB−/− thymocytes. Hence, the catalytic activity of ITPKB is required in thymocytes. During positive selection, low avidity/affinity TCR engagement induces DP thymocyte maturation. High avidity/affinity TCR stimulation induces apoptosis during negative selection. This prevents the maturation of detrimental self-reactive T cells. Most TCR signalling-mediators are involved in both processes. One study reported defective negative selection in ItpkB−/− mice expressing the transgenic HY TCR23. However, the higher deletion in the controls compared to ItpkB−/− thymocytes likely reflects the presence of mature thymocytes in the control but not ItpkB−/− sample and not increased negative selection, because mature T cells can induce DP cell death through a cytokine storm30. Moreover, the HY-TCR system may have limited physiological relevance, because expression of this transgenic mature TCR starts before the DP stage, alters the phenotype of DP cell precursors and causes deletion at the DP stage78. In contrast, endogenous mature TCRs are expressed mainly on DP and later stages, and significant negative selection occurs after the DP stage. Therefore, detailed studies in more appropriate models are required to conclusively determine whether ITPKB has a role in negative selection. An essential role for ITPKB in positive but not negative selection would be intriguing. Of all the TCR signalling-mediators studied, only RASGRP1, calcineurin B and ERK1/2 are exclusively required for positive but not negative selection30, 79, 80.
ItpkB−/− thymocytes show strongly decreased TCR-induced InsP4 production23. Total InsP3 3-kinase activity was reduced by ~50%, indicating that ITPKB might be redundant with other InsP3 3-kinases in thymocytes. InsP3 production and Ca2+-mobilization were normal18, 23. This was unexpected, because InsP3 mobilizes Ca2+ and ITPKB converts InsP3 into InsP4, which can itself control Ca2+ mobilization (discussed above)14, 44, 45. Although more detailed analyses are needed, the InsP4-role in a positive feedback-loop of PLCγ1 activation discussed below (Fig. 1,,3)3) may explain this conundrum. Impaired InsP3-turnover through ItpkB could allow the residual PLCγ1-activity in ItpkB−/− thymocytes to produce normal InsP3 levels17. Because of the unchanged InsP3 levels, the failure of catalytically inactive ITPKB to restore ItpkB−/− thymocyte development indicated a specific requirement for InsP4 for positive selection. Indeed, addition of a cell-permeable InsP4-ester can allow ItpkB−/− thymocyte maturation in NTOC [G] (Y. H. Huang and K. Sauer, unpublished observations). Mechanistic analyses showed that in ItpkB−/− DP cells, TCR-induced ERK activation was defective due to impaired ITK membrane recruitment and activation, causing decreased PLCγ1 activation and DAG production17, 18. TCR stimulation primarily activates ERK through RASGRP1-mediated RAS-activation26, 81, 82. RASGRP1 is recruited to cellular membranes through DAG. The ability of exogenously provided phorbol-12-myristate-13-acetate (PMA), a DAG-analogue, to restore ERK activation in ItpkB−/− thymocytes and allow their maturation showed, for the first time, that DAG is essential for thymocyte positive selection. It moreover confirmed InsP4 as an essential mediator of ITK- and PLCγ1-dependent DAG production, RAS and ERK activation. But how does InsP4 act?
A first clue came from the finding that InsP4 binds to the ITK PH domain in vitro17. At high, probably super-physiological, concentrations, InsP4 competed with PtdInsP3 for ITK PH domain binding. Surprisingly, physiological, low μM concentrations of soluble InsP4 promoted ITK PH domain binding to PtdInsP3-coated beads17. DP thymocytes contain an ITK pool whose ability to bind PtdInsP3 is transiently increased by TCR stimulation. In ItpkB−/− thymocytes, which produce decreased InsP4 after TCR stimulation, basal ITK binding to PtdInsP3 is reduced and TCR stimulation does not increase it. Moreover, TCR-induced ITK membrane recruitment and activation were perturbed. This suggests that TCR-induced InsP4 promotes Itk–PtdInsP3 interactions, and therefore ITK membrane recruitment and activation in DP thymocytes. The precise underlying mechanism remains to be determined. However, the ability of full-length ITK or its PH domain alone to oligomerize opens the possibility that InsP4 binding to one ITK subunit induces allosteric changes in the other ITK subunits that cooperatively increase the affinity of their PH domains for PtdInsP317, 27, 28, 83. These findings identified the first physiological “InsP4 receptor” and the first in vivo function for InsP4. Moreover, they indicate that in some cases, oligomerization might be important for the functions of PH domains, which mediate protein membrane recruitment downstream of many different receptors in many cell types17, 31, 84.
Based on these data, InsP4 establishes a positive feedback-loop [G] of PLCγ1 activation by ITK downstream of the TCR that is required for production of sufficient DAG to activate RAS and ERK and trigger positive selection (Fig. 3)17. Rescued ItpkB−/− thymocyte maturation upon retroviral expression of wildtype but not CaM-binding-deficient ITPKB indicates that Ca2+–CaM-binding is required for ITPKB activation in vivo17, 85. In vitro data support this view11–15, 85. TCR-induced InsP4 production requires Ca2+ and ITPKB18, 23, 25, 62, 63 and can be inhibited by CaM inhibitors (J. Huang and K. Sauer, unpublished observations). These findings add a feedforward [G] loop of ITPKB activation by Ca2+–CaM downstream of ITK and PLCγ1 to our model. In DP thymocytes, the combination of feedback- and feedforward-activation loops establishes an InsP4- and Ca2+-dependent signal amplifier that is required for mild TCR stimuli to trigger positive selection.
Depending on ligand affinity/avidity, an enigmatic signal-splitter downstream of Ca2+ has been postulated to direct the outcome of TCR engagement towards positive or negative selection80. The Ca2+ dependence of ITPKB and its requirement for positive selection make the InsP4-mediated signal amplifier an attractive candidate signal-splitter component. A prediction would be that InsP4 is only required for ERK activation in response to mild TCR stimuli, but dispensable for negative selection in response to stronger TCR stimuli. Interestingly, “strong” CD3/CD4-costimulation induced DAG production and ERK activation in ItpkB−/− thymocytes17, 18 (and Y.H. Huang and K. Sauer, unpublished observations). Thus, it will be interesting to analyse in more physiological models whether negative selection is intact in ItpkB−/− mice.
ItpkB−/− thymocytes show 30–40% residual InsP3 3-kinase activity and InsP4 production18, 23. Thus, complete InsP4-deficiency could have more profound effects on TCR signalling and might impact processes that are intact in ItpkB−/− mice. Reminiscent of the effects of PtdInsP3 modulation on multiple aspects of thymocyte development5–9, 32, Itk−/− mice have defects in preTCR signalling, positive and negative selection. In particular, they accumulate innate-like CD8+ T cells [G] and untypical T helper 2 (TH2)-phenotype γδ T cells [G]27, 28, 86, 87. It will be interesting to investigate whether co-disruption of multiple InsP3 3-kinases completely inhibits TCR-induced InsP4 accumulation and ITK activation in thymocytes and elicits similar phenotypes. However, the positive selection defect in Itk−/− and even Itk−/−Rlk−/− mice is milder than in ItpkB−/− mice. Thus, other InsP4 targets likely contribute to InsP4-function in positive selection. Indeed, InsP4 modulates binding of PtdInsP3 to other thymocyte-expressed InsP4-binding proteins, including TEC and RASA348, with similar dose-responses as ITK17. RASA3 inactivates RAS. Its PH domain can bind InsP4, PtdIns(4,5)P2 or PtdInsP344, 45, 49, 88, 89. Impaired RASA3-sequestration from RAS by InsP4 has been proposed to explain the decreased RAS/ERK activation in ItpkB−/− DP cells18. However, no in vivo data are available and further studies are needed. RAS is primarily activated by DAG-recruited RASGRP1 in T cells (Fig. 3). Impaired DAG production in ItpkB−/− DP cells17 blocks RASGRP1 recruitment and RAS activation upstream of RASA3 and should thus be dominant over any potential effects of InsP4-deficiency on RASA3. Hence, other InsP4 effectors are more likely to mediate ITK-unrelated InsP4 functions in positive selection.
Altogether, ITPKB and InsP4 are essential for thymocyte development. A profound block of positive selection in ItpkB−/− mice results primarily from defective TCR-induced DAG production. This in turn reflects the disruption of an essential signal-amplifier composed of a feedforward loop of ITPKB activation by Ca2+–CaM, and of a positive feedback loop of ITK and PLCγ1 activation by InsP4. The amplifier is probably needed to allow mild TCR stimuli to elicit positive selection. Phenotypic differences between Itk−/− and ItpkB−/− mice indicate involvement of other InsP4-effectors. Whether InsP4 has any role in thymocyte negative selection is still unclear.
TCR stimulation induces InsP3 3-kinase activity and InsP4 production in human Jurkat T cells25, 62. Peripheral T cells express all three ITPKs. Owing to the developmental defects and severe T cell lymphopenia in ItpkB−/− mice, analyses of specific ITPKB roles in peripheral T cells will require conditional ItpkB−/− mice, RNAi [G] or selective small-molecule inhibitor studies.
A recent study found an intriguing association of a human ITPKC allele that decreased ITPKC mRNA splicing efficiency and abundance with susceptibility to Kawasaki syndrome, the leading cause of childhood-acquired-heart-disease in developed countries42. Acute-phase patients have T cell infiltration into the coronary artery wall and increased IL-2 production, suggesting that T cell hyperactivation could contribute to the disease. Suggesting a potential ITPKC-function in limiting peripheral T cell activation that could be perturbed in Kawasaki-disease patients, ITPKC mRNA levels increased after PMA/Ionomycin stimulation33, 42. ITPKC overexpression decreased, and ITPKC knockdown increased, phytohaemagglutinin- and PMA-induced NFAT activation and IL-2 mRNA expression in Jurkat cells42. The mechanism is unknown, but could involve some of the aforementioned, controversial InsP3 3-kinase/InsP4-roles in Ca2+-mobilization, or other functions of ITPKC, InsP4 or its metabolites (Fig. 3, Box 1)14, 44, 45. Mechanistic studies of ITPKC in primary T cells are needed to eludicate it.
A negative InsP4 role in peripheral T cells would contrast with its positive role in thymocytes. Several potential explanations could be envisioned: First, the functions of InsP4 could differ depending on TCR signal intensity or the developmental stage of T cells. Positive selection is triggered by mild and/or transient TCR signals in DP thymocytes. Negative selection is mediated by strong and/or sustained TCR signals and might be less impaired in ItpkB−/− mice17, 23, 30. Peripheral T cells also generate strong TCR signals that might be impacted differently by InsP4-deficiency. Alternatively, the effects of InsP4 might depend on its cellular concentration, subcellular localization or on the specific InsP4-effectors or metabolites present in a cell. It is unknown how ITPKC-modulation affects the levels of InsP4 and its metabolites in peripheral T cells, and whether their InsP4-effector complement differs from that in thymocytes. Finally, different ITPKs might have distinct roles that could include noncatalytic functions and might depend on their subcellular localization41, 90, 91. Testing all of these possibilities will require simultaneous ITPK-inhibition through small molecules, or their conditional co-disruption.
ItpkB−/− mice contain substantial B cell numbers. B cell development in the bone marrow of chemically-induced ItpkB−/− mice is unchanged. However, the numbers of all splenic B cell subsets are markedly decreased, including transitional B cells undergoing selection, mature follicular and marginal zone B cells. Follicular B cells in ItpkB−/− mice have downregulated IgM levels. Half of the splenic IgD+ cells express the plasma cell marker CD13816, 19. They represent a pre-plasma cell checkpoint-stage enriched in models of B cell tolerance92. Indeed, mature B cells from ItpkB−/− mice share many features with B cells from soluble hen-egg lysozyme (sHEL)/HEL-specific immunoglobulin (IgHEL) and Ars/IgArs transgenic models of B cell anergy93. These include IgM-downregulation, impaired BCR-driven proliferation, reduced BCR-induced upregulation of surface-CD69, CD86 and MHCII, and decreased antibody-responses to T cell-independent antigens16, 19, 21. These defects are specific to BCR stimulation. ItpkB−/− B cell responses to LPS or CD40-specific antibodies were normal. These findings indicate that ITPKB might control B cell selection and tolerance induction.
In the sHEL/IgHEL and Ars/IgArs anergy models, constitutive self-antigen expression causes BCR desensitization, impaired BCR-activation of proximal tyrosine-kinases (LYN, SYK) and of downstream effectors including PLCγ2, reduced InsP3 production and impaired Ca2+ influx94, 95. In contrast, anergic B cells from ItpkB−/− mice showed normal BCR-activation of LYN, BTK, PLCγ2, ERK1, ERK2, IKKα and IKKβ with unaltered InsP3-production, but increased BCR-induced SOCE (Fig. 4)16. Normal Ca2+ responses to thapsigargin, which opens SOCE-channels by depleting ER Ca2+-stores, indicated that SOCE-channels were unaltered. Addition of cell-permeable InsP4-esters decreased thapsigargin-induced SOCE and normalized BCR-induced SOCE in ItpkB−/− B cells16. How InsP4 inhibits SOCE in B cells is unknown. The mechanism might include perturbations of some of the aforementioned, controversial InsP3 3-kinase/InsP4-roles in Ca2+-mobilization or other functions of InsP4 or its derivatives (Fig. 4, Box 1)1, 11–14, 33–35, 44–47. Altogether, B cells from ItpkB−/− mice are anergic but have a dramatically different biochemical signature than other models of B cell anergy. How enhanced BCR-induced Ca2+ signaling causes ItpkB−/− B cell anergy remains an important unresolved question.
An independent study described similar alterations in B cell development and T cell-independent antibody-responses in ItpkB−/− mice21. The authors also reported decreased survival of ItpkB−/− B cells in vitro and attributed this to accumulation of the pro-apoptotic protein BIM. Indeed, transgenic expression of the anti-apoptotic protein BCL-2 or BIM haplo-insufficiency increased B cell numbers in ItpkB−/− mice. The authors further reported impaired BCR- or Ionomycin-induced Ca2+ influx in ItpkB−/− B cells, and decreased BCR-induced ERK activation in BCL-2-expressing ItpkB−/− B cells. They also showed that ATP-stimulation, or cell-permeable InsP4-ester-treatment of non-lymphoid COS cell-transfectants expressing ITPKB sequestered RASA348 from the plasma membrane and proposed that InsP4 antagonizes RASA3 PH domain-binding to membrane-PtdIns(4,5)P2 (Fig. 1)21, 49. They proposed that InsP4 increases B cell survival by sequestering RASA3 in the cytosol, resulting in sustained RAS/ERK activation, BIM-phosphorylation and -degradation. However, RASA3-localization and -function were not studied in B cells. The relevance of the COS cell results for lymphocytes is unclear. Detailed analyses of RASA3-localization and -function in wildtype, ItpkB−/− or InsP4-ester-treated B cells are required to conclusively determine potential RASA3-roles in InsP4 signalling in B cells.
Because ITPKB appears to control B cell selection, it was unclear if the signalling alterations in B cells from ItpkB−/− animals reflect the induction and selective survival of anergic B cells, or an additional ITPKB function in BCR-signalling in mature B cells. To address this issue, ItpkB−/− mice were crossed with IgHEL-transgenic mice. B cells from ItpkB−/− IgHEL mice are also anergic19. They showed surface IgM downmodulation, IgD+CD138+ B cell accumulation and decreased proliferative antigen-responses. Early BCR signalling was unaltered. BCR-induced Ca2+ responses were greatly increased. This could be reversed by cell-permeable InsP4-ester treatment. Transgenic soluble sHEL-antigen expression in IgHEL mice leads to anergy, characterized by surface IgM downregulation and impaired HEL-induced proliferation. In sHEL/IgHEL double-transgenic mice, ItpkB deficiency resulted in a tenfold decrease in the number of splenic IgHEL-specific B cells, and two-fold fewer immature IgHEL-specific B cells in the bone marrow19. Thus, ITPKB-deficiency converted anergy to deletion in the sHEL/IgHEL model, reminiscent of the impact of losing other negative regulators of BCR signalling, including CD22, SHP1 or LYN.
These data indicate that the main ITPKB-function in B cells is to inhibit BCR-induced Ca2+ signalling. They help to reconcile the discrepancies between the previous reports by suggesting that the increased BIM-expression in polyclonal ItpkB−/− B cells in REF.21 may result from augmented negative selection of a B cell subset. Consistent with this model, ItpkB−/−IgHEL mice show neither increased negative selection nor BIM accumulation. Similar considerations may explain the decreased ERK activation and Ca2+ influx in polyclonal ItpkB−/− B cells in REF.21. In particular, the altered ionomycin-response in REF.21 but not REF.16 indicates a more extreme phenotype, perhaps owing to the selection of aberrant B cells over time or to differences in housing, health status, genetic background or age of the mice used in both studies.
Based on these data, we propose a model in which ITPKB, via production of InsP4, functions as a feedback-inhibitor of BCR-induced SOCE (Fig. 4). Increased SOCE in ItpkB−/− mice alters B cell selection and tolerance induction. The outcome depends on the BCR-affinity for self-antigens. ItpkB-deficiency enhances Ca2+ signaling and promotes positive selection for low-affinity self-specific BCRs such as IgHEL. It induces anergy of B cells expressing low-to-moderately self-reactive BCRs in non-transgenic or IgHEL backgrounds, but deletion of normally anergic B cells with more strongly self-reactive BCRs, as seen in the sHEL/IgHEL system19. By feedback-inhibiting SOCE, ITPKB therefore extends the repertoire of immature B cells that survive negative selection. Interestingly, all B cells that mature in non-transgenic ItpkB−/− mice become anergic regardless of their affinity for self. This might reflect distinct ITPKB or InsP4-roles in mature B cells whose elucidation will require inducible ITPKB-deletion or the use of specific small-molecule ITPKB inhibitors.
B cells from ItpkB−/− mice have features of anergic B cells that arise in models of self-tolerance, suggesting that ITPKB-deficiency can allow the positive selection of autoimmune-prone B cells. If some of these could escape anergy, they might cause autoimmune disease. However, in contrast to mice lacking other negative regulators of BCR signalling, including CD22, LYN or PTP1, ItpkB−/− mice show no overt signs of autoimmunity. This could simply reflect a requirement for other immunocytes such as mature T cells, which are absent in ItpkB−/− mice. However, mixed radiation-chimeras of ItpkB−/− bone marrow with wild-type T, B and myeloid cells also lacked autoimmune pathology (A. Miller and M. Cooke, unpublished observations). This might indicate a lack of required additional autoimmune insults. Alternatively, increased BCR-induced SOCE in ItpkB−/− B cells might prevent autoimmunity by enabling more efficient deletion of autoimmune-prone B cells as seen in the sHEL/IgHEL system. Crossing ItpkB−/− mice with appropriate autoimmune-prone strains or BCR-transgenic mice will address conclusively whether ItpkB-deficiency promotes or prevents autoimmune disease.
Recent studies have shown important functions of ITPKB in the innate immune system. Contrasting with its positive roles in lymphocyte development, ITPKB may limit myelopoiesis. It moreover controls neutrophil function.
ItpkB mRNA is expressed by all murine haematopoietic stem/progenitor cell populations (Fig. 5)20. ItpkC mRNA is expressed by all myeloid lineages, and minimally by haematopoietic stem cells and common lymphoid progenitors. Common myeloid progenitors and megakaryocyte–erythrocyte progenitors express low ItpkA mRNA levels. ItpkB−/− mice show increased proliferation and expansion of granulocyte–monocyte progenitors, increased neutrophil production and elevated peripheral blood neutrophil numbers (Fig. 5)20. Thus, ITPKB limits myeloid differentiation in vivo. The precise mechanism is unclear, but might involve inhibition of the PI3K target AKT, whose PH domain can bind PtdInsP3, PtdIns(3,4)P2 or InsP4, which can antagonize PtdInsP3-binding22, 31, 96–98. AKT is essential for myelopoiesis and can promote neutrophil and monocyte development6. Haematopoietic progenitor cell-enriched bone marrow cells from ItpkB−/− mice show increased phosphorylation (activation) of AKT, and of the AKT target-site Y145 on the cell-cyle inhibitor p21Cip1. Y145-phosphorylation decreases cell cycle-inhibition by p21Cip1. It will be exciting to analyze whether AKT-inhibition normalizes myelopoiesis in ItpkB−/− mice, and to explore potential ItpkA/C-roles in myelopoiesis.
Stimulation with chemoattractants including N-formyl-methionyl-leucyl-phenylalanine (fMLP) induces InsP3 3-kinase activity and InsP4 accumulation in neutrophils, which express mainly ITPKB22, 47, 99. Bone-marrow-derived neutrophils (BMNs) from ItpkB−/− mice show increased chemotaxis and superoxide production after treatment with fMLP or the complement-factor C5a. This correlates with increased AKT phosphorylation and actin polymerization22. Treatment with cell-permeable InsP4-ester had opposite effects and inhibited fMLP-induced membrane-recruitment of the AKT PH domain in neutrophilic HL60 promyelocytic leukemia cells. AKT-PH domain–GFP fusion proteins co-precipitated InsP4, InsP5 and InsP6. This suggests that ITPKB limits chemoattractant-induced neutrophil activation, possibly by inhibiting AKT membrane recruitment and activation through InsP4-antagonism with AKT PH domain binding to PtdInsP3 or PtdIns(3,4)P2. Whether InsP4 limits Akt-recruitment in other cell types in vivo is unknown. Overall, fMLP-induced Ca2+ responses were normal in ItpkB−/− neutrophils22. However, the authors later mentioned decreased Ca2+ store-release but enhanced SOCE as unpublished data20, 47. Defective Ca2+ homeostasis, Ca2+ mobilization or other ITPKB/InsP4 functions could possibly explain the paradoxical finding of decreased viability of ItpkB−/− BMNs in vitro20 although PI3K–AKT signalling (which is increased in ItpkB−/− BMNs) inhibits neutrophil spontaneous death47.
Interestingly, ItpkB deficiency promoted neutrophil recruitment into inflamed peritoneal cavities in an acute peritonitis model. However, clearance of the injected bacteria was not substantially improved22. Subsequent studies showed slightly decreased bacterial clearance in vivo, but increased killing of serum-opsonized bacteria in vitro by ItpkB−/− neutrophils. Serum from ItpkB−/− mice, which contains less opsonizing IgG than wildtype serum due to the aforementioned B cell defects, facilitated killing of bacteria less efficiently than wildtype serum47. This indicates that ITPKB limits neutrophil recruitment and in vitro function. However, in vivo, ITPKB probably acts in several different cell types to control peritonitis responses. Although not investigated in detail, the decreased IgG titres in ItpkB−/− mice16, 19, 21 might mask or counteract the neutrophil hyperactivity by preventing efficient opsonization of bacteria. Finally, altered Ca2+ homeostasis and mobilization in ItpkB−/− neutrophils47 and impaired additional functions of ITPKB/InsP4 probably contribute to the complex effects of ItpkB deficiency on neutrophil function in vivo.
Its similarity to the cytoplasm-exposed PtdInsP3 tetraphospho-inositol-headgroup allows InsP4 to bind to the PH domains of the lymphoid-expressed PI3K-effectors ITK, TEC, BTK and AKT (Figs. 1, ,33)17, 31, 96–98. In thymocytes, InsP4 binding to the ITK PH domain promotes PtdInsP3 binding to ITK, probably through a cooperative-allosteric mechanism involving PH domain oligomerization17. In neutrophils, InsP4 inhibits AKT PH domain –membrane interactions, probably by competing with PtdInsP3- and PtdIns(3,4)P2-binding. Overall, the phenotypes of ItpkB−/− mice are consistent with one role for InsP4 as a soluble regulator of PI3K, PTEN, SHIP1 and SHIP2 function. InsP4 does this by controling PH domain interactions of their lipid-substrates or products (Fig. 1). The phenotypic differences between mice lacking individual InsP3 3-kinases versus PI3K subunits probably reflect incomplete abrogation of InsP4 or PtdInsP3 production, respectively, the fact that InsP4 can either promote or inhibit PtdInsP3 function and might only control a subset of PtdInsP3 targets, PtdInsP3-independent functions of InsP4 or its metabolites (Box 1), or non-catalytic ITPKB functions3, 4791.
How InsP4 inhibits SOCE in B cells and possibly neutrophils in unknown. The mechanism might include perturbations of some of the aforementioned, controversial InsP3 3-kinase/InsP4-roles in Ca2+-mobilization, or other InsP4 functions (Fig. 4)1, 11–14, 33–35, 44–47.
InsP4 acts as a precursor for multiple other InsPs, many of which have been found in immunocytes (Box 1)17, 18, 23, 25, 58–66. Some InsP4 functions could thus be mediated by InsP4-derived InsPs rather than by InsP4 itself. Based on the recently identified, intriguing higher-order InsP functions in non-lymphoid cells (Box 1), higher-order InsP studies in immunocytes should be very exciting.
Quantitative kinetic analyses of InsP4, PtdInsP3 and higher-order InsP-levels in ItpkB−/− mice, the identification of all relevant InsP-targets through bioinformatics or affinity-chromatography, ITPKB structure-function analyses and conditional InsP3 3-kinase co-disruption in relevant cell types will be required to fully understand how InsP3 3-kinases govern immunocyte development and function, and to identify all underlying molecular mechanisms.
Although ITPKB is required for efficient selection of both T and B cells, its positive role in thymocyte positive selection strikingly contrasts with its negative role in B cell selection17, 18, 23. In T cells, ITPKB likely extends the repertoire by amplifying weak TCR signals that would otherwise fail to elicit positive selection. In B cells, ITPKB expands the repertoire by inhibiting BCR signalling and preventing the deletion of B cells that would otherwise fail to develop. That a single gene achieves repertoire extension in opposite manners in T and B cells is fascinating. It probably reflects important differences in the rate-limiting steps for selection of T cells (positive selection) versus B cells (negative selection). Interestingly, these differences mirror the differential impact of ITPKB on SOCE in both cell types. Feedback-inhibition of SOCE by InsP4 does not control thymocyte positive selection but is crucial for extending the B cell repertoire. It will be interesting to determine whether this difference results from differences in InsP4 or InsP4-metabolite concentrations, InsP4-effector or Ca2+-channel repertoires33, 100 between both cell types.
Another difference between ItpkB−/− thymocytes and splenic B cells is the defective TCR-induced ITK membrane recruitment and activation in ItpkB−/− thymocytes17, which contrasts with the normal BCR-induced phosphorylation of BTK and PLCγ in ItpkB−/− splenic B cells16, 19. The highly related ITK, TEC and BTK PH domains can all bind PtdInsP3 or InsP45, 17, 33, 46, 47. The differences between ItpkB−/− thymocytes and B cells could possibly be explained by differing affinities of the various PH domains for InsP4 versus PtdInsP3101–103, or by PH domain-independent BTK activation in B cells5. Moreover, the possibility that sorted splenic B cell subpopulations from ItpkB−/− mice might show stronger differences in BTK and PLCγ phosphorylation than the subtle effects found in total splenic B cells16 remains to be investigated.
T cells, B cells and neutrophils are key mediators of inflammatory diseases, including rheumatoid arthritis, diabetes or asthma. Defects in their development or function can cause immunodeficiencies such as AIDS or X-linked agammaglobulinemia, or promote cancer. T cells moreover cause organ-transplant rejection. All three cell types require ITPKB for their development or function. We therefore next discuss the therapeutic opportunities and liabilities of targeting ITPKs pharmacologically.
Because of its key roles in immunocyte development and function, cancer cell survival, proliferation and migration, and in other pathologic processes such as obesity, PI3K is a popular therapeutic target. The first PI3K inhibitors recently entered clinical trials for inflammatory disease and cancer indications. However, the ubiquitous functions of PI3K provide significant potential for toxicity, which might limit therapeutic applications5, 7, 8, 104. Autoimmune-phenotypes in mice with altered PI3K function raise additional concerns for long-term therapies5. Moreover, the ability of IPMK to function as a Wortmannin-insensitive PI3K73 might limit the efficacy of classic PI3K inhibitors that do not inhibit IPMK. Because their product, InsP4, controls PI3K function at the level of PtdInsP3, ITPKs are potential alternative targets. However, InsP4 can affect PtdInsP3 function positively, negatively or not at all depending on the specific PtdInsP3 target involved17, and ITPKs have PtdInsP3-independent functions. Hence, the effects of ITPK inhibition are difficult to predict.
Based on the phenotypes of ItpkB−/− mice (Table 1)16–18, 21, 23, selective ITPKB-inhibition should primarily affect B cell, T cell and neutrophil development and function without significant effects on other tissues. However, broad ITPK-family inhibitors might have effects on tissues expressing other ITPKs, including the brain. Although it might be possible to prevent ITPK inhibitors from passing the blood–brain barrier, analyses of ItpkA/B/C double- or triple-deficient mice and studies with proof-of-concept ITPK inhibitors are required to address potential liabilities. Although InsP4 can act as a precursor for higher-order InsPs with incompletely understood but important functions in mammalian cells (Box 1)74, mouse embryonic fibroblasts lacking detectable ItpkA, ItpkB and ItpkC-expression are viable and still produce InsP5 and InsP623, 75. This suggests that preservation of higher-order InsP production through the broadly expressed, structurally diverse IPMK77 might reduce the risk of potential side-effects of selective ITPK-inhibition.
Their biochemical properties and available high-throughput assays make ITPKs highly “druggable” [G]11–13, 25, 33, 105. Compared with protein and lipid kinases, distinct structural features of the ITPKs could allow the rational design of highly selective, ATP- or substrate-competitive or allosteric ITPK inhibitors11, 33, 85, 106–108. Indeed, several groups have identified small-molecule ITPK inhibitors. Some show high ITPK-selectivity over IPMK33, 109–112.
The strong inhibitory effects of ItpkB-deficiency on thymocytes and B cells make broad immunosuppression for the treatment of transplant-rejection or severe autoimmune diseases the most likely indication for ITPKB inhibitors. A true evaluation of their therapeutic potential awaits validation with human cells. ITPKB is highly expressed by human peripheral T cells33. The recent implication of genetic ITPKC polymorphisms in Kawasaki disease42 indicates the pathophysiological importance of ITPKs. Although the precise roles of ITPKs in disease aetiology are unknown, a negative role for ITPKC in human T cell activation42 could pose a potential liability for broad Itpk-inhibitors. Alternatively, this negative function of ITPKC might allow the development of ITPK inhibitors as adjuvants to promote T cell function in immunodeficiencies and infectious diseases.
Neutrophils have important roles in infections, allergies, inflammatory diseases and autoimmunity113. Hence, the pathophysiological relevance of the increased neutrophil chemotaxis and superoxide production, but decreased neutrophil viability in ItpkB−/− mice22 requires exploration. Increased neutrophil recruitment in peritonitis models in ItpkB−/− mice22 could possibly indicate opportunities for ITPKB-inhibition for anti-infective therapies, but might pose liabilities for immunosuppression. The feasibility of neutrophil modulation depends on the involvement of T and B cells in the targeted diseases, as despite the increased neutrophil number, clearance of intraperitoneal-injected bacteria was unaltered or even decreased in ItpkB−/− mice through mechanisms that might involve decreased serum IgG titres22, 47. Altogether, the validation of ITPKB as a therapeutic target will require more detailed analyses of its role in innate immunocytes and in peripheral T cells, analyses of ItpkB−/− mice in relevant disease models and proof-of-concept studies with potent and selective small-molecule ITPKB inhibitors.
Two decades after its discovery, ItpkB−/− mice have demonstrated important in vivo signalling functions for InsP4 and unveiled the first physiological InsP4 effectors84. Like many other important discoveries, these findings pose more questions than they answer (Box 2). In non-immune cells and yeast, higher-order InsPs have recently been implicated in several important cellular processes (Box 1)1, 25, 74, 84. Thus, it will be particularly exciting to investigate the functions of the higher-order InsPs in immunocytes (Box 1) and their potential contributions to InsP4-signaling25. Without any doubt, deciphering the lymphocyte “inositol code” will remain a rich area of research for years to come.
We thank N. Gascoigne, G. Mayr and Y. Hsing Huang for critical reading of the manuscript and valuable comments. We apologize to our colleagues for citing reviews instead of many important original references due to limitations to the number of references allowed. K.S. is supported by NIH grants AI070845 and GM088647.
Karsten Sauer received his Ph.D. in Biochemistry from the University of Tuebingen, Germany. After postdoctoral training at the Merck Research Laboratories, Rahway, New Jersey, USA, he joined the Genomics Institute of the Novartis Research Foundation, San Diego, California, as a group leader. He now is an Associate Professor at The Scripps Research Institute, La Jolla, California, USA, which he joined in 2006. His research interests are the signal transduction mechanisms controlling lymphocyte development and function, and the mechanisms through which signalling defects cause immune disorders.
Michael P. Cooke
Michael P. Cooke received his Ph.D in Biochemistry from the University of Washington, Seattle. After postdoctoral work at Stanford University, he was Director of Functional Genomics at SyStemix Inc., Palo Alto, California. Since 1999, he is Director of Immunology at the Genomics Institute of the Novartis Research Foundation, San Diego, California. His research includes the application of functional genomics to study the biology of haematopoietic stem cells and the adaptive immune system and the translation of these findings into novel therapeutics.
Entrez Gene: http://www.ncbi.nlm.nih.gov/gene
ItpkB | Itpkb | ITPKB | InsP3KB
NCBI Structure: http://www.ncbi.nlm.nih.gov/Structure
ItpkB | Itpkb | ITPKB
Kawasaki syndrome | Kawasaki disease | Mucocutaneous Lymph Node Syndrome | Infantile Polyarteritis
Web site links
Karsten Sauer’s homepage: http://www.scripps.edu/ims/sauer
Michael P. Cooke’s homepage: http://www.gnf.org/technology/immunology