The type 2 inositol 1,4,5-trisphosphate receptor, emerging functions for an intriguing Ca²⁺-release channel

Abstract

The inositol 1,4,5-trisphosphate (IP₃) receptor (IP₃R) type 2 (IP₃R2) is an intracellular Ca²⁺-release channel located on the endoplasmic reticulum (ER). It displays in many cell types a predominantly perinuclear or even nuclear localization. IP₃R2 is characterized by a high sensitivity to both IP₃ and ATP and is biphasically regulated by Ca²⁺. Interestingly, ATP stimulates IP₃R2 independently of the cytosolic [Ca²⁺]. Furthermore, IP₃R2 is modulated by phosphorylation events mediated by e.g. protein kinase A, Ca²⁺/calmodulin-dependent kinase II and protein kinase C. In addition to its regulation by protein kinase A, IP₃R2 forms a complex with adenylate cyclase 6 and is directly regulated by cAMP, thereby linking in a new way Ca²⁺-dependent and cAMP-dependent signalling. Finally, in the ER, IP₃R2 is less mobile than the other IP₃R isoforms, while its functional properties appear dominant in heterotetramers. These properties make the IP₃R2 a Ca²⁺ channel with exquisite properties for setting up intracellular Ca²⁺ signals with unique characteristics. IP₃R2 plays a crucial role in the function of secretory cell types (e.g. pancreatic acinar cells, hepatocytes, salivary gland, eccrine sweat gland). In cardiac myocytes, the role of IP₃R2 appears more complex, because, together with IP₃R1, it is needed for normal cardiogenesis, while its aberrant activity is implicated in cardiac hypertrophy and arrhythmias. Moreover, IP₃R2 expression is driven by IP₃-induced Ca²⁺ release leading to a self-perpetuating system of cardiac hypertrophy. Most importantly, its high sensitivity to IP₃ makes IP₃R2 a target for anti-apoptotic proteins (e.g. Bcl-2) in B-cell cancers. Disrupting IP₃R/Bcl-2 interaction therefore leads in those cells to increased Ca²⁺ release and apoptosis. Intriguingly, IP₃R2 is not only implicated in apoptosis but also in the induction of senescence, another tumour-suppressive mechanism. These results were the first to unravel the physiological and pathophysiological role of IP₃R2 and we anticipate that further progress will soon be made in understanding the function of IP₃R2 in various tissues and organs.

1. Introduction

The inositol 1,4,5-trisphosphate (IP₃) receptors (IP₃Rs) are ubiquitously expressed intracellular Ca²⁺-release channels. These channels are tetrameric in structure and predominantly localized in the endoplasmic reticulum (ER). IP₃, produced by phospholipase C after cell activation by hormones, growth factors or neurotransmitters, diffuses into the cytosol, binds to and activates the IP₃R, leading to Ca²⁺ release from the ER. This Ca²⁺ release is instrumental in the formation of the spatio-temporal Ca²⁺ signals, fundamental for the regulation of multiple cellular processes, including proliferation, differentiation, metabolism, secretion, cell fate and memory [1-3].

In all vertebrate organisms, three different genes encode IP₃Rs, leading to three main types of IP₃Rs, IP₃R1 (first fully cloned in 1989 [4]), IP₃R2 (first fully cloned in 1991 [5]) and IP₃R3 (first fully cloned in 1993 [6]). Each monomer is about 2700 amino acids in length and consequently has a predicted molecular mass of approximately 300 kDa. The IP₃R proteins are structurally and functionally divided into 5 distinct domains: the N-terminal coupling domain (a.k.a. suppressor domain), the IP₃-binding core, the internal coupling domain (a.k.a the modulatory and transducing domain), the channel domain and the C-terminal coupling domain (a.k.a. the gatekeeper domain) [7]. The three IP₃R isoforms share only 60-80 % overall similarity at the amino acid level, but the similarity is much higher in certain defined regions (e.g. the IP₃-binding site, the 5^th and 6^th transmembrane regions, putative ER-retention signals) while much lower in others, allowing for the existence of very distinct properties between the isoforms (see section 2). Finally, a further diversity can result from alternative splicing and the formation of heterotetramers [1, 8, 9].

The vast majority of cell types express more than one IP₃R isoform but their relative proportion can be highly variable [10-14]. Moreover, in various cases it was shown that the relative expression levels depended on the differentiation or developmental state of the cells or could be modulated by specific treatments (e.g. [14-21]).

The participation of the IP₃ Rs in establishing distinct patterns of Ca²⁺ signals resulting in different cellular outcomes depends therefore on the complement of the various IP₃R (splice) isoforms expressed, their intracellular location, the presence of regulatory factors, including associated proteins, and their phosphorylation status [1, 14, 22-25].

On the basis of their almost exclusive expression of a single IP₃R isoform, some cell types have been used as model system for the analysis of the role of the different IP₃R isoforms [8]. However, only a rather limited set of cell types expresses predominantly IP₃R2 (Table 1).

Table 1.

Cell types or tissues predominantly expressing IP₃R2.

Cell type/tissues (in alphabetical order)	Relevant references	Specific remarks
AR42J	[11, 12]	Pancreatoma cell line
Cardiac myocytes	[12, 13, 136, 189]
Glia	[117, 190]
Hepatocytes	[11-13, 29]	Polarized expression IP3R2
Intercalated cells of renal collecting duct	[191]
Pancreatic acinar cells	[12]	IP3R3 ≈ IP3R2
RBL-2H3	[11, 115, 192]	Mucosal mast cell line

Interestingly, in many cells IP₃R2 is expressed at a different subcellular location than the other IP₃R isoforms. For example, in bovine aortic endothelial cells, bovine adrenal glomerulosa cells and COS-7 cells [26], the HepG2 liver cell line [27] and the hippocampal cell line HT22 [28] IP₃R2 displayed a predominantly nuclear localization, while in hepatocytes IP₃R2 is confined to the apical pole of the cell, near the canalicular membrane [29].

The nuclear localization of IP₃R2 is particularly interesting with respect to the role of Ca²⁺ signalling in the nucleus, e.g. for gene transcription. Although the presence of IP₃Rs in the inner leaflet of the nuclear envelope remains the subject of debate [30-32], there is at least strong evidence that a subset of the IP₃R2 is facing the nucleoplasm in HepG2 cells [27], in SKHep1 cells [33] and in atrial myocytes [34] where they can control Ca²⁺ release directly into the nucleus.

Information concerning the regulation of IP₃R2 expression is likewise still quite limited. The 5′-flanking region of murine IP₃R2 has been sequenced and contained at least 7 transcription initiation sites with an upstream promoter containing no conventional TATA box but a GC box [35]. To the best of our knowledge, only two recent studies described pathways involved in the regulation of IP₃R2 expression. First, in the heart, direct binding of nuclear factor of activated T cells (NFAT) c1 to the IP₃R2 promoter drives IP₃R2 expression [36]. Second, in dendritic cells, IP₃R2 expression is controlled by the transcription factor ETS1, which itself depends on protein kinase B (AKT/PKB) 2 [37]. Finally, although not yet understood at the mechanistic level, in the HT22 cell line, oxidative stress leads to a specific upregulation of IP₃R2 [38] (see section 5).

Splicing of IP₃R2 is much less documented than that of IP₃R1, but two different splice variants have been described. One appears to be muscle-specific and is limited to the N-terminal 175 amino acids of IP₃R2 supplemented with 6 additional amino acids; although a regulatory role has been proposed, its function has not yet been elucidated. [39]. The second splice variant uses the same splice acceptor site and is lacking amino acids 176-208. The deletion is localized fully within the suppressor domain, which plays an important role in both the regulation of IP₃ binding and in the coupling of the IP₃R N-terminus to the channel region [40]. As a consequence, the resulting protein is defective in both IP₃ binding and Ca²⁺ release. However, its expression in cells prevents the agonist-dependent clustering of the endogenous IP₃Rs, probably via heterotetramerization, and can therefore impact intracellular Ca²⁺ signalling [41].

Finally, also at the protein level, IP₃R2 levels appear to be regulated in a different manner when compared to the other isoforms. While all IP₃R isoforms are downregulated under conditions of chronic stimulation [12, 42, 43], IP₃R2 appeared the least susceptible [12].

In spite of the unique molecular properties displayed by IP₃R2 (see sections 2.1-2.5), this protein is much less investigated than IP₃R1 or indeed, even IP₃R3. This apparent lack of progress was probably in retrospect multifaceted. The low general abundance of IP₃R2 and the lack of good model systems for its investigation partially explain the fewer studies directed specifically towards the IP₃R2. Additionally, but no less importantly, it appeared that it was historically difficult to make IP₃R2 expression constructs. Moreover, the quality of the antibodies raised against IP₃R2 was also generally poor. Finally, a recent analysis of classically used IP₃R inhibitors demonstrated that IP₃R2 expressed in DT40 triple knockout (DT40 TKO) chicken B lymphocytes was the least sensitive of all the IP₃R isoforms to heparin, caffeine and 2-aminoethoxydiphenyl borate [44]. This latter observation may explain why pharmacological approaches to discern IP₃R2 function have been largely unsuccessful.

Notwithstanding these issues, recent work has begun to unravel the significance of IP₃R2 in a number of physiological settings. The aim of this review is therefore to highlight these important functions of IP₃R2 and to so stimulate further research in the field.

2. Specific molecular and cellular properties of IP₃R2

IP₃R2 has a high sequence and structural homology with the other IP₃R isoforms and consequently shares a large number of properties with IP₃R1 and IP₃R3. In this chapter, we will therefore focus on specific properties in which the IP₃R2 clearly differs from the other isoforms (Figure 1).

Figure 1. Linear representation of the IP₃R2 based on the human sequence demonstrating the interaction sites for its major regulators.

The IP₃R2 is represented in blue; the 5 functional domains are indicated. In the channel domain, the 6 transmembrane helices as well as the connecting loops are depicted in green. The specific regulatory mechanisms discussed in the text are shown: identified phosphorylation sites are represented in dark blue, interaction sites for ATP and for Bcl-2 are in orange and the recently described Gly²⁴⁹⁸Ser mutation in the pore domain affecting IP₃R2 function [126] is depicted in red.

2.1 IP₃ affinity

Binding of IP₃ to the IP₃-binding core is the key step needed for the induction of IP₃-induced Ca²⁺ release. As stated above (see section 1), the three IP₃R isoforms all have a very similar structure [45]. For example, all IP₃R isoforms contain their IP₃-binding core towards their N-terminus, preceded by the so-called suppressor domain [46].

A striking property of the IP₃R2 is its much higher affinity for IP₃ when compared with the two other IP₃R isoforms. This was first observed in IP₃-binding experiments, which under various conditions demonstrated a rank-order of IP₃ affinities IP₃R2 > IP₃R1 > IP₃R3 [5, 10, 47]. Subsequent studies conclusively demonstrated that the IP₃-binding cores of each of the IP₃R isoforms display a similar affinity for IP₃ (about 2 nM), but as demonstrated by the analysis of normal and chimeric N-terminal domains of the different IP₃R isoforms the presence of the suppressor domain determines the specific IP₃ affinity of the isoform (50, 14 and 163 nM for IP₃R1, IP₃R2 and IP₃R3, respectively). These results underpin the importance of the suppressor domain for IP₃R function [46].

The difference in IP₃ affinity is also reflected in functional experiments. Analysis of IP₃-induced Ca²⁺ release in DT40 TKO cells expressing a single IP₃R isoform demonstrated that DT40 TKO cells heterologously expressing a single IP₃R isoform could sustain Ca²⁺ oscillations for an extended period after stimulation by an anti B-cell receptor antibody only if the expressed isoform was IP₃R2 [48]. Very similar results were obtained in vascular myocytes, whereby only the cells expressing IP₃R2 in addition to IP₃R1 displayed a Ca²⁺ oscillation pattern [49, 50]. Comparison of native IP₃R1 (from cerebellum) and IP₃R2 (from heart) [51] or the comparative analysis of each of the three IP₃R isoforms heterologously expressed in Sf9 insect cells [52] also confirmed the rank-order of the sensitivity of channel opening towards IP₃ as IP₃R2 > IP₃R1 > IP₃R3, when measured after incorporation in planar lipid bilayers.

It must however be pointed out that a number of studies found a rank-order of affinities that is different from that mentioned above [53, 54]. Although this discrepancy has never been fully clarified, it can be assumed that variability in the experimental conditions (e.g. pH, [Ca²⁺]) as well as in the state of the IP₃Rs (e.g. redox state, phosphorylation state, associated proteins, existence of heterotetramers) could explain this variability [1].

Taken together, most of the available evidence points to IP₃R2 as being the most sensitive IP₃R isoform. This observation raises the intriguing potential that IP₃R2 can be active in the presence of basal, resting IP₃ levels.

2.2. Regulation by cytosolic Ca²⁺

It has long been recognized that IP₃R activity can be biphasically regulated by cytosolic Ca²⁺, meaning that a relatively low [Ca²⁺] (usually less than 0.3 μM) potentiate IP₃-induced Ca²⁺ release, while higher [Ca²⁺] lead to an inhibition of the IP₃R. Plotting IP₃R activity against [Ca²⁺] therefore leads to a typical bell-shaped curve. The original observations were obtained in smooth muscle [55], neurons [56, 57] and oocytes [58], all tissues later shown to be particularly rich in IP₃R1. It was therefore a long standing question whether this property was uniquely related to this isoform or whether IP₃R2 and IP₃R3 shared this property.

Although the stimulatory effect of Ca²⁺ on IP₃R2 (and IP₃R3) activity has never been in doubt, the inhibitory action of a high [Ca²⁺] was not always clearly observed. The group of Mignery published single-channel data demonstrating that both IP₃R2 endogenously expressed in heart [51] or recombinantly expressed IP₃R2 [59] displayed a much broader bell-shaped dependence towards Ca²⁺, meaning that on the one hand the stimulatory phase starts at a much lower [Ca²⁺] than for IP₃R1, and on the other hand that IP₃R2 remains active at a [Ca²⁺] already fully inhibiting IP₃R1. In contrast to these reports, the group of Bezprozvanny found quite similar (and narrow) [Ca²⁺] response curves for each of the three IP₃R isoforms heterologously expressed in Sf9 insect cells and investigated in planar lipid bilayers [60].

As will be further explained below (see section 2.3), the sensitivity of the IP₃R2 towards Ca²⁺ is not dependent on the presence of ATP, but the latter will increase the likelihood of IP₃R2 being in an open state at all [Ca²⁺] (Figure 2).

Figure 2. Modulation of IP₃R2 single-channel activity by IP₃, Ca²⁺ and ATP.

Representative single channel recordings of IP₃R2 expressed in DT40 TKO cells using the “on-nucleus” configuration of the patch-clamp technique. In A, channel activity was stimulated with a maximal [IP₃] (10 μM) at the indicated [Ca²⁺] and [ATP] (10 μM, in blue; 5 mM, in black). The pooled data in B reveal that channel activity stimulated by maximal [IP₃] is modulated by [Ca²⁺] in a biphasic manner and that this relationship is unaffected by increasing the [ATP]. In C, channel activity was stimulated with a sub-maximal [IP₃] (1 μM) at the indicated [Ca²⁺] and [ATP] (10 μM, in blue; 5 mM, in black). The pooled data in D demonstrate that while channel activity is also biphasically regulated by [Ca²⁺] at sub-maximal [IP₃], the maximally achievable open probability, at each [Ca²⁺], is, in contrast to what happens at a maximal [IP₃], markedly potentiated in the presence of a high [ATP]. Modified from [79], with permission.

Taken together, these results indicate that IP₃R2, similarly to the other IP₃R isoforms, is regulated in a biphasic way by the cytosolic [Ca²⁺], though that depending on the exact state of the receptor, different sensitivities to Ca²⁺ can be observed.

2.3. Regulation by ATP

2.3.1. Regulation of IP₃R2 is distinct from other IP₃Rs

Adenine nucleotides were recognized by early studies [61-68] as important regulators of IP₃-induced Ca²⁺ release, raising the attractive possibility that channel activity could be fine-tuned to match the metabolic status of the cell. The diversity of cell types in which ATP modulates IP₃-induced Ca²⁺ release is consistent with a regulation affecting all IP₃R family members. However, two initial studies reported that in contrast to IP₃R1 and IP₃R3, IP₃R2 was not subject to modulation by adenine nucleotides [48, 52]. Specifically, studying individual mammalian isoforms reconstituted in planar lipid bilayers, Bezprozvanny and colleagues reported that under conditions optimal for channel activity IP₃R2 had no requirement for ATP [52]. This observation was independently confirmed studying Ca²⁺ release from DT40 cells expressing a single IP₃R isoform following genetic ablation of the other family members [48]. These reports are, in hindsight, important because they provide the first indication that ATP regulation of IP₃R2 was distinct from the other family members. A subsequent detailed analysis of individual mammalian isoforms expressed in the DT40 TKO IP₃R null background confirmed these earlier reports [69]. IP₃R2 was indeed, in contrast to other IP₃R family members, insensitive to ATP at maximal [IP₃]. Nevertheless, it was demonstrated that IP₃R2 activity, measured as Ca²⁺ release, or at the single channel level in “on-nucleus” patch clamp recordings, was markedly enhanced at sub-saturating [IP₃] [69]. Moreover, the sensitivity of ATP regulation of IP₃-induced Ca²⁺ release also differed between individual isoforms under identical conditions with IP₃R2 being strikingly more sensitive than IP₃R1 or IP₃R3 (EC₅₀ 40 μM, 100 μM and 500 μM for IP₃R2, IP₃R1 and IP₃R3 respectively) [69, 70]. An issue posed by these data is the concentration range of ATP that might be expected to dynamically regulate IP₃R activity. In turn, this raises the question whether modulation occurs at physiological levels of nucleotides or is only relevant under pathological conditions when ATP is depleted. Given the cellular levels of MgATP (~1 mM) and “free” ATP³⁻ and ATP⁴⁻ (10-100 μM) the answer is fundamentally dependent on the “species” of ATP that regulates IP₃R channel activity. Several studies have addressed this issue and have reached disparate conclusions [67, 71-73] and thus this important issue as well as the consequences of the high functional affinity of IP₃R2 remains to be resolved.

2.3.2. Putative peptide motifs in IP₃R2 responsible for ATP regulation

Modulation of IP₃R activity is widely believed to occur by ATP binding to glycine-rich domains (Gly-Xaa-Gly-Xaa-Xaa-Gly), reminiscent of Walker type A repeats, present in a number of proteins that utilize ATP in a catalytic manner [74-77]. Consistent with this idea, a number of studies using either photo-affinity or fluorescent ATP probes have demonstrated binding to regions of IP₃R or glutathione S-transferase-recombinant fragments harbouring these putative recognition sites [69, 75-77]. The primary sequence of IP₃R2 contains one such motif, Gly-Leu-Gly-Leu-Leu-Gly, spanning amino acids 1969-1974, which has been termed the “ATPB” site (Figure 1). Mutagenesis of three Gly residues to Ala in the motif eliminated binding of ATP and nucleotide regulation of Ca²⁺ release, confirming the functional importance of the ATPB site in IP₃R2 [69]. Moreover, in cells expressing IP₃R2 with an ATP binding-deficient ATPB motif, the frequency and amplitude of B cell receptor-activated Ca²⁺ oscillations were markedly reduced compared with wild-type IP₃R2, suggesting strongly that nucleotide regulation of Ca²⁺ release is at least, constitutively required to shape cytosolic Ca²⁺ signals at physiologically relevant ATP levels [69]. Unexpectedly, mutations of all known Walker A motifs in IP₃R1 and IP₃R3 failed to abrogate nucleotide modulation [70]. The somewhat surprising conclusion is therefore, that ATP regulation of IP₃R1 and IP₃R3 is independent of known ATP-binding motifs, and thus the identity of molecular sites of nucleotide regulation in these IP₃R remains to be elucidated. Consequently, the ATPB site in IP₃R2 is unique as the only molecular locus for regulation of IP₃R family members by adenine nucleotides that is defined unequivocally.

2.3.3. Mechanism of ATP regulation of IP₃R2

Several studies have investigated the biophysical basis for ATP regulation of IP₃R channel activity. In accordance with the singular features of IP₃R2 when compared with other family members, it also appears that ATP regulates IP₃R2 in a similarly distinctive manner. Using “on-nucleus” patch clamp single channel recordings of both endogenous Xenopus IP₃R1 or rat IP₃R1 expressed in DT40 TKO cells, elevating ATP levels increased the channel open probability (P_o) by modulating the sensitivity of the channel to both activating and inhibitory [Ca²⁺], essentially left-shifting the bell-shaped [Ca²⁺] versus P_o relationship at a given [IP₃] [78, 79].

In contrast, while IP₃R2 displays an identical biphasic Ca²⁺ sensitivity when exposed to saturating [IP₃] (conditions in which IP₃R2 is insensitive to ATP), at low [IP₃], the Ca²⁺ sensitivity of mouse IP₃R2 was not altered by increasing ATP [79]. Elevating ATP simply dramatically enhanced P_o, resulting in a marked increase in activity. A detailed kinetic analysis of the channel gating also indicated that IP₃R2 displayed “bursting” activity with properties distinct from IP₃R1 (Figure 2). Specifically, with elevated ATP, the number of bursting episodes of relatively constant duration was increased, while IP₃R1 bursts simply lengthened in time. By analogy to a gear change in a car, we have termed this the transition from ‘park’ into a ‘drive’ mode. A minimal scheme to describe the channel kinetics at sub-saturating [IP₃] suggests that both channels transition between single open and closed states during drive mode with relatively constant kinetics and then are “parked” in a longer-lived closed state in the interburst intervals [79]. In the case of the IP₃R1, increasing Ca²⁺ and ATP facilitates bursting by facilitating both the transition out of the parked state and also by decreasing the likelihood it will return to this state. In contrast, [Ca²⁺] does not influence the time the IP₃R2 spends in drive mode but simply destabilizes the parked state to initiate activity (Figure 3). Increasing ATP then appears to markedly increase overall channel P_o by prominently decreasing the amount of time in the parked state [79]. This unique property results in dissociation of the modulation of IP₃R2 activity by ATP from the [Ca²⁺] in its immediate environment and likely allows added flexibility for tuning Ca²⁺ signals to the needs of the cell.

Figure 3. “Park and Drive” model for IP₃R1 and IP₃R2 gating.

An increase in IP₃R1 (A) and IP₃R2 (B) channel activity in the presence of activating ligands is characterized by an increase in channel “bursting” without altering the intraburst kinetics. The bursts have subtype specific characteristics. A gating scheme for both channels can minimally be described by three states; one open state (in green) and two closed states (in red). Bursting activity is represented by rapid transitions between the open state (O) and a short-lived closed state (C₁) representing the “Drive Mode” of the channel. In the interburst intervals, the channel is effectively “Parked” in a long-lived closed state (C₂). For both IP₃R1 (A) and IP₃R2 (B) increasing the concentrations of activating ligands solely alters the transition from C₂ to C₁. However, ligands both increase the likelihood that IP₃R1 will leave the parked state to drive mode, as well as reciprocally decreasing the chances it will return to this state, thus extending the period of bursting (A). In the case of IP₃R2, ligands only destabilize the parked state resulting in an increase in bursting episodes of relatively constant duration (B).

2.3.4. Dominance of IP₃R2 ATP regulatory characteristics

An important question exists as to how the distinct features of individual IP₃R subtypes are reflected in the overall characteristics of Ca²⁺ release from heterotetrameric channels. Specifically, are the properties simply a blended integration of the individual subtypes or can a particular subtype dominate the overall characteristics? Studies investigating ATP regulation of Ca²⁺ release in cells expressing multiple IP₃R isoforms indicate that the latter possibility occurs, specifically when IP₃R2 is expressed. For example, the characteristics of ATP regulation of IP₃R2 (albeit the lack of regulation at saturating [IP₃]) were observed in DT40 cells engineered to express only IP₃R2, or in cells expressing both IP₃R2 and IP₃R1 or IP₃R3 [48]. Similarly, in salivary and pancreatic acinar cells [80, 81], which natively express IP₃R2 and IP₃R3 to approximately equal extents, the features of ATP regulation precisely match those documented for IP₃R2 stably expressed in isolation in either DT40 TKO [69] or in AR42J pancreatoma [80] cells (i.e. absence of regulation at saturating [IP₃] and EC₅₀ for ATP ~40 μM). Notably, similar experiments in pancreatic and parotid acinar cells prepared from IP₃R2 null animals revealed identical properties to IP₃R3 (i.e. regulation at all [IP₃] and EC₅₀ for ATP ~500 μM) [80, 81]. Conversely, “rescue” experiments ectopically expressing IP₃R2 in RINm5F insulinoma cells which predominately express IP₃R3, converted IP₃R3 characteristics to IP₃R2 [80]. While these data clearly indicate the dominant influence of IP₃R2 expression and are consistent with this occurring as a function of heterotetramer formation, these data could also formally be explained by an intermolecular interaction between clusters of homotetrameric IP₃R. This issue has recently been tackled by generating tetrameric IP₃R from concatenated IP₃R dimers connected by short flexible linkers [82]. Expression of dimers results in the assembly of tetramers where the subunit composition can be unequivocally defined. Expression of dimers of IP₃R1 or IP₃R2 exhibited the distinctive properties of ATP regulation typical of channels assembled from their respective monomeric parent subtype. Remarkably when heterodimers of IP₃R1 and IP₃R2 were expressed, resulting in assembly of channels consisting of equal numbers of IP₃R1 and IP₃R2 subunits, ATP regulation was indistinguishable from IP₃R2, thus recapitulating the dominant effects seen in cells expressing native receptors [82]. These data indicate that IP₃R2 in the context of a heterotetrameric channel exerts a dominant influence. Further work is needed to establish the number of monomers of IP₃R2 necessary to exert this influence and whether IP₃R2 similarly is the principle monomer that dictates the overall channel properties when subjected to other forms of regulation.

2.4. Regulation by phosphorylation

Like for many other ion channels, phosphorylation/dephosphorylation reactions provide a versatile, reversible form of acute regulation of IP₃R activity. IP₃Rs have been shown to be biochemical substrates for numerous families of serine/threonine and tyrosine directed kinases. In a more limited number of cases, a comprehensive documentation of the phosphorylation event, including location of the substrate motif and the subsequent functional consequences have been detailed. These studies have largely focused on IP₃R1 as a template. With some notable exceptions, for example the AKT/PKB site conserved in the C-termini of each IP₃R [83, 84], the amino acid motifs subject to phosphorylation events are not generally preserved between IP₃R subtypes (Figure 1). Therefore, this form of regulation has the capacity to provide modulation of activity in an IP₃R sub-type specific manner. Below we highlight reports that have specifically focused on regulation of IP₃R2 activity. The interested reader is directed to Vanderheyden et al. [23] and Betzenhauser and Yule [85] for detailed discussion of IP₃R phosphorylation and its functional consequences.

2.4.1. Regulation of IP₃R2 by protein kinase A (PKA)

Historically, perhaps the most exhaustive investigation of IP₃R modulation relates to PKA phosphorylation of IP₃R1. Indeed, IP₃R1 was identified as a major brain phosphorylated substrate even prior to the protein being appreciated as the receptor for IP₃ [86]. Subsequent studies demonstrated that IP₃R1 is phosphorylated at serine residues within two canonical consensus motifs (Bas-Bas-Xaa-Ser/Thr, where Bas = a basic residue) [87, 88] and phosphorylation is associated with markedly enhanced Ca²⁺ release [89, 90]. To complete the strong case for IP₃R1 being a functionally important PKA substrate, mutation of Ser¹⁵⁸⁹ and Ser¹⁷⁵⁵ to non-phosphorylatable alanine residues completely abrogates phosphate incorporation and the functional effects of PKA activation [90-92]. However, while PKA activation in cells that predominately express IP₃R2 such as hepatocytes, parotid acinar cells and AR42J similarly results in enhanced Ca²⁺ release [93-95], the PKA substrate motifs present in IP₃R1 are not conserved in IP₃R2 [5]. In addition, while PKA activation results in IP₃R2 phosphorylation, phosphate incorporation is non-stoichiometric and indeed much reduced in comparison to IP₃R1 [96]. Nevertheless, IP₃R2 contains approximately 30 serine or threonine residues, which constitute minimal PKA consensus motifs consisting of basic residues preceding the phosphorylated amino acid at the −2 and −3 positions (Bas-Bas-Xaa-Ser/Thr). Using an approach based on expressing consecutive domains of IP₃R2 with N-terminal epitope tags, it was shown that PKA could only specifically phosphorylate in vitro a peptide fragment consisting of amino acids 920-1583 [97]. Mutation of Ser⁹³⁷ to alanine (Ser⁹³⁷Ala) abrogated all phosphorylation of the fragment, pinpointing this residue as the PKA target site (Figure 1). Subsequently, it was shown that an antibody raised against phospho-Ser⁹³⁷ recognized IP₃R2 after forskolin treatment in cells expressing IP₃R2 but not cells expressing a mutant full-length receptor harbouring a Ser⁹³⁷Ala mutation [97]. Notably, Ser⁹³⁷ was independently identified as a phosphorylated residue in a proteomic screen of hepatocytes [98]. Importantly, PKA activation markedly potentiated IP₃-induced Ca²⁺ release in DT40 TKO cells expressing IP₃R2 but not Ser⁹³⁷ Ala IP₃R2 thus establishing this motif as likely solely responsible for the PKA-mediated phospho-regulation of IP₃R2 [97]. The Ca²⁺ signalling machinery is a rich source of substrates responsible for cross-talk between cAMP and Ca²⁺ signalling which ensure fine-tuning of the Ca²⁺ signal and appropriate activation of effectors [99]. PKA phosphorylation of IP₃R2 likely is an important site of this interaction in cells such as astrocytes, cardiac myocytes, hepatocytes and acinar cells that prominently express this family member.

2.4.2. Regulation of IP₃R2 by Ca²⁺/calmodulin-dependent protein kinase II (CaMKII)

CaMKII are a family of serine/threonine kinases assembled as either homo- or heteromultimers derived from the products of four closely related genes [100]. As a Ca²⁺/calmodulin-regulated enzyme, this kinase is an important primary effector of IP₃R-induced Ca²⁺ release and accordingly plays prominent roles in regulating various signal transduction pathways including the translocation of transcription factors and activity of ion channels [101, 102]. Notably, IP₃Rs are substrates for the kinase, which provides a regulatory loop following Ca²⁺ release. Early work suggested that IP₃R1 was a substrate for CaMKII in vitro and that the sites were distinct from those phosphorylated by PKA [103], however the functional consequences were poorly defined. Subsequently, studies based largely on pharmacology, concluded that Ca²⁺ release from Xenopus oocytes and HeLa cells was attenuated following CaMKII activation [104, 105]. More recently, a thorough characterization of the molecular sites and functional consequences of the CaMKII-mediated phopho-regulation of IP₃R2 has been reported. Using a similar approach to that used to identify PKA sites, the ability of CaMKII to phosphorylate IP₃R2 fragments in vitro was assessed. It was initially demonstrated that a candidate residue was present within a fragment encompassing the initial 1078 amino acids [106] and further refinement narrowed the potential phosphor-acceptor residue to within residues 134-338 [107]. Subsequent mutagenesis of potential serines/threonine residues in CaMKII consensus motifs (Ser/Thr-Xaa-Asp) within this region [107] identified Ser¹⁵⁰ as phosphorylated by CaMKII (Figure 1). This site is conserved in mammalian IP₃R family members and ryanodine receptor (RyR) 2, suggesting a common mode of regulation in these channels. When incorporated in bilayers, CaMKII phosphorylation reduced the P_o of IP₃R2 and this effect was reversed by the CaMKII inhibitor KN62. Importantly, the reduced channel activity was absent in Ser¹⁵⁰Ala IP₃R2, indicating that the site was functionally relevant [107]. Notably, CaMKII colocalizes and interacts with IP₃R2 in the nuclear envelope of cardiac myocytes [106, 108]. This interaction has been proposed to be functional important for cardiac remodelling during hypertrophy [109] (see section 4).

2.4.3. Regulation of IP₃R2 by protein kinase C (PKC)

Cerebellar IP₃R was also initially identified as a substrate for PKC with phosphorylation sites independent of those sites modified by PKA [103]. Interestingly, however, it was demonstrated that phosphorylation by PKC was enhanced by prior PKA phosphorylation indicating a potential additional layer of cross-talk between these prominent cellular signalling systems [110]. Similarly to CaMKII, because at least conventional PKC family members are regulated by an elevation in Ca²⁺, PKC phosphorylation provides a potential feedback loop to regulate IP₃R activity. To date however, the functional effects of PKC phosphorylation and the sites of phosphorylation are relatively poorly defined. Unfortunately despite the general appreciation of the amino acid motifs that constitutes a PKC consensus sequence and the presence of multiple such templates in IP₃R, none have been experimentally defined. Furthermore, while PKC activation results in enhanced Ca²⁺ release from liver nuclei, presumably reflecting IP₃R1 and IP₃R2 activity [111], Ca²⁺ release is inhibited in AR42J cells, which predominately express IP₃R2 [112]. These disparate findings may reflect subtype-specific regulation of IP₃R given that the PKC consensus motif numbers and location are different in each family member [9]. However an additional consideration is that numerous proteins in the signalling pathway from plasma membrane receptor occupation to the generation of Ca²⁺ signals are substrates for PKC and thus caution must be taken in interpreting data generated from indirect measurements of IP₃R activity. Our own experience is that IP₃R2 single channel activity recorded in either DT40 cell nuclei or in DT40 plasma membranes was unaffected by phorbol ester treatment or recombinant PKC exposure (Wagner, Chandrasekhar and Yule; unpublished observations). These data might indicate that IP₃R is not a direct substrate for PKC. Yet, we cannot formally exclude the possibility that a scaffolding or anchoring protein necessary for activity is absent from the DT40 system. Hence, further work is required to characterize the impact of PKC on IP₃R activity in general and on IP₃R2 in particular.

2.5. Other characteristics of IP₃R2

All the characteristics discussed above (see sections 2.1-2.4) have been the subject of extensive investigations. There are however, a few less studied properties, which nevertheless might be very interesting for understanding the cellular function of IP₃R2.

2.5.1. Regulation of IP₃R2 by cAMP

Similarly to the other IP₃R isoforms, many accessory proteins interact with, and modulate IP₃R2 function. These proteins include regulatory and structural proteins, many of which were also reported to interact with IP₃R1 and/or IP₃R3 [1, 2, 14]. An interesting exception is the interaction described between IP₃R2 and type 6 adenylate cyclase (AC6) [113]. In HEK 293 cells stably transfected with the type I parathyroid hormone receptor, a complex is formed between IP₃R2, AC6 and Gαs [113, 114]. This close association facilitates an exquisite regulation of IP₃R2 by cAMP, and in addition, Ca²⁺ released through IP₃R2 may control AC6 in a negative feedback loop. Importantly, the regulation by cAMP does not require the canonical ATP-binding site or the activity of PKA. Moreover, although all IP₃R isoforms are potentially sensitive to cAMP, only IP₃R2 has been unequivocally linked to a cAMP-producing enzyme (AC6) [114]. This mechanism can be of great general importance, since it provides a novel example of cross-talk between the cAMP- and the Ca²⁺-dependent pathways.

2.5.2. Clustering and mobility of IP₃R2

All IP₃R isoforms, including IP₃R2 [115] are known to cluster in an agonist-dependent way [116] but a punctate distribution of IP₃R2 has also been observed for native IP₃R2 [117] and heterologously expressed IP₃R2 [41] in resting cells. This property can be correlated to the higher affinity of IP₃R2 for IP₃ (see section 2.1), which may allow the clustering to occur at basal, resting [IP₃]. Interestingly, also other differences in behaviour were found between the different IP₃R isoforms. A recent study performed in COS-7 cells confirmed that heterologously expressed IP₃R2 showed a punctate distribution, in contrast to IP₃R1 and IP₃R3 that were uniformly distributed [118]. Moreover, IP₃R2 appeared much less mobile than either the other IP₃R isoforms or than other proteins involved in intracellular Ca²⁺ handling, such as RyR1 or sarco/endoplasmic reticulum Ca²⁺ ATPase (SERCA) 1. In addition, its mobility depended on its intracellular localization with the IP₃R2 located in the perinuclear region having the lowest mobility. As the IP₃R2 has the highest sensitivity to IP₃ (see section 2.1), its lesser mobility may determine the initiation sites for intracellular Ca²⁺ signals.

3. The function of IP₃R2 in secretory cells

IP₃R2 exhibits prominent expression in classical secretory cells (Table 1), including exocrine cells of the pancreas [119-121], salivary glands [81, 120-123], lacrimal gland [124], olfactory glands [125], liver [29], eccrine sweat glands [126] and the secretory epithelia of the biliary tree [127] and the intestine [18] and the goblet cells of the small intestine [128]. A common feature of these epithelial cells is that they are morphologically and functionally polarized to secrete fluid and protein across their apical pole into a lumen forming a duct. Notably, the Ca²⁺ signal is centrally important to the primary secretory function of these cells by virtue of directly activating ion channels and the exocytotic machinery necessary for vectoral fluid and protein secretion [129, 130]. In exocrine glands, IP₃R2 and IP₃R3 are expressed in approximately equal numbers [12] and both family members are co-localized to a region immediately below the apical plasma membrane [119, 120]. This region has been termed “the trigger zone” because Ca²⁺ signals are invariably initiated in this region prior to the signal spreading as a Ca²⁺ wave towards the basal aspects of the cell [131-133]. There appears to be substantial functional redundancy between IP₃R2 and IP₃R3 in exocrine cells as mice null for either IP₃R in isolation have no obvious phenotype. However, compound knockouts of IP₃R2 and IP₃R3 have severe exocrine deficiency manifested as dry mouth [121], dry eye [124], pancreatic insufficiency [121] and attenuated mucus secretion [125]. As such, double knockouts are born normally but demise soon after weaning [121]. Indeed a detailed analysis of these mice have shown that Ca²⁺ signals in pancreatic, salivary, lacrimal and mucus glands are essentially unaltered in IP₃R3 null mice [121, 124, 125] and only reduced to a modest degree at low [IP₃] in IP₃R2 null mice [80, 81, 121, 124, 125]. These data suggest that IP₃R2 is not generally essential for overall exocrine function.

A possible exception to this idea has been highlighted by a recent study, which investigated the cause of a severe congenital sweating defect in a Pakistani family [126]. Ca²⁺ signalling is known to be important for sweat secretion and both IP₃R2 and IP₃R3 are expressed in the secretory cells of the sweat gland. A screen based on identifying regions of autozygosity in the genome of afflicted individuals revealed a mutation (Gly²⁴⁹⁸Ser) targeting an amino acid predicted to be critical to the function of the selectivity filter in the pore region of IP₃R2. This mutation rendered the channel completely inactive when expressed in DT40 TKO cells, thus potentially explaining the defect observed in the patients. Consistent with this idea, subsequent studies showed that mice lacking IP₃R2 exhibited a decreased ability to sweat although the effect was more modest than observed in humans. The differences in severity between the phenotype observed between mouse and human may reflect the relative levels of these subtypes in mouse versus human. Alternatively, the relatively mild phenotype in the mouse might be related solely to the knockout of IP₃R2, reflecting some degree of compensation by the residual IP₃R3. In this scenario the more severe effect in human could be due to the combined effect of ablation of IP₃R2 pore function and a possible dominant negative effect of the mutant IP₃R2 when incorporated into heterotetramers containing IP₃R3.

The primary secretory function of hepatocytes is the secretion of bile and changes in intracellular Ca²⁺ play important regulatory roles in this process. Hepatocytes express predominantly IP₃R2 (Table 1) with smaller amounts of IP₃R1 and virtually no IP₃R3. However, in contrast to exocrine acinar cells, each isoform exhibits a distinct sub-cellular localization and therefore the isoforms appear not to have redundant functions. IP₃R2 is enriched at the canalicular membrane, whereas IP₃R1 has a more uniform distribution throughout structures in the cytosol [29]. Consistent with the sensitivity of IP₃R2, agonist-induced Ca²⁺ signalling is initiated through IP₃R2 localized to the canalicular membrane [29, 134, 135] and Ca²⁺ release through this isoform is necessary for trafficking of the bile salt export pump to the canalicular membrane [134].

4. The function of IP₃R2 in the heart

As indicated (Table 1), cardiomyocytes are one of the cell types in which IP₃R2 are highly expressed. Both atria and ventricles express IP₃R2 [136] and consequently IP₃R2 channels have been implicated in both physiological and pathophysiological signalling in the heart.

At the physiological level, Mikoshiba and co-workers showed that IP₃R2 channels, together with IP₃R1 channels, are critical for normal cardiogenesis [137]. Consistent with this, IP₃R2 and IP₃R1 are co-expressed in different parts of the embryonic heart, including atria, ventricle and atrioventricular canal, and in different cell types, including endothelial cells and cardiomyocytes, although timing differences in the appearance of IP₃R2 versus IP₃R1 exist. IP₃R1/IP₃R2 double knockout mice die in utero at embryonic stage E11.5 with major heart defects at the level of the ventricles (thin myocardial wall and poor trabeculation) and the atrioventricular canal (reduced number of cells). These defects were associated with a decrease in endocardial and myocardial cell proliferation. Furthermore, mesenchymal cells were lacking at the level of the developing atrioventricular canal. The authors hypothesized that this phenotype was due to the absence of IP₃R-mediated Ca²⁺ signalling, downstream of the activation of the Ca²⁺/calmodulin-dependent phosphatase calcineurin and of the translocation of NFATc to the nucleus. Indeed, in an ex vivo epithelial-mesenchymal transition (EMT) assay, the defect in EMT in atrioventricular explants derived from IP₃R1/IP₃R2 double knockout mice could be restored by transducing constitutively active calcineurin. Moreover, the phenotype of these mice resembled well the phenotypes of mice knockouts for calcineurin B [138] or for NFATc3/NFATc4 [139]. Interestingly, the defect in endocardial cells could also be observed in developing zebrafish exposed to calcineurin inhibitors, such as FK506 or cyclosporine A. Hence, from this study, it is clear that IP₃R1 or IP₃R2 channels are needed for activating calcineurin/NFATc signalling and endocardial cell proliferation in vertebrates. It is important to note that the presence of either IP₃R1 or IP₃R2 is sufficient to drive normal cardiac development, indicating redundant functions for these channels in this process. In addition to the cardiac crescent, or first heart field, giving rise to a linear beating tube, there is a second source of myocardial cells, which is termed the second heart field. However, with respect to the latter, it appears that there is a redundant role for IP₃R1 and IP₃R3 [140]. IP₃R1/IP₃R3 double knockout mice are characterized by hypoplasia of the outflow tract and the primitive right ventricle at E8.5-9.5, probably due to a defective Mef2c-Smyd1 transcriptional pathway.

At the functional level, IP₃Rs were first shown to impact contractility and arrhythmias. Atrial myocytes express much higher levels of IP₃R2 than ventricular myocytes [136]. They are activated in response to elevated extracellular agonist concentrations, e.g. after ischaemia or during disease. Endothelin-1-induced IP₃R activation promotes the inotropy and the occurrence of arrhythmic events in atrial myocytes [141, 142]. While in basal conditions, the cardiac function of IP₃R2 knockout mice was similar to one of wild-type mice, the positive inotropic effect and the arrhythmic events induced by endothelin-1 were absent in IP₃R2 knockout mice [143]. Hence, the presence of IP₃R2 appears to be not essential for the normal functioning of the rodent heart, which is in line with the normal phenotype of the IP₃R2 knockout mice generated by Chen and co-workers [143] and by Mikoshiba and co-workers [121].

In the ventricle, far fewer IP₃Rs are present but may still contribute to Ca²⁺ regulation under baseline conditions [144-147]. Moreover, a study by Roderick and co-workers revealed that increased IP₃R2-mediated Ca²⁺ signalling, in response to enhanced IP₃ signalling, is responsible for inducing hypertrophic pathways after prolonged endothelin-1 exposure of neonatal or adult rat ventricular cardiomyocytes [148]. Endothelin-1 triggered the expression of atrial natriuretic factor (ANF), a marker for hypertrophy. Endothelin-1-induced hypertrophy was independent of excitation-contraction coupling, but required IP₃ signalling and downstream IP₃R-mediated Ca²⁺ release. The latter occurred in the perinuclear region, but not in the cytosol, while Ca²⁺ transients linked to excitation-contraction coupling occurred throughout the cardiomyocyte. Specifically buffering nuclear Ca²⁺ by nuclear-targeted calbindin prevented endothelin-1-induced ANF expression. The nuclear Ca²⁺ signal was mediated by IP₃R2 channels, which were enriched in the perinuclear region and led to the activation of calcineurin and downstream NFATc1, which accumulated in the nucleus.

The IP₃R/calcineurin/NFATc1 hyperactivity also seems to be operative in response to prolonged β-adrenergic signalling, which occurs during workload-induced cardiac hypertrophy via enhanced excitation-contraction coupling. Interestingly, this model led to increased endothelin-1 signalling. This involved the release of endothelin-1 and autocrine/paracrine-mediated hyperactivation of its receptor (ET_AR), thereby triggering downstream IP₃ signalling and Ca²⁺-dependent calcineurin activation. In hypertrophic cardiomyocytes (e.g. derived from spontaneous hypertrophic rats or from aortic-banded mice), IP₃R2 not only played a role in the nucleus, where its hyperactivation via increased IP₃ signalling downstream of ET_AR led to calcineurin/NFATc1 activation and ANF expression, but also became upregulated in the junctional sarcoplasmic reticulum [146]. Here, localization of IP₃R2 coincides with RyR2 channels, thereby augmenting Ca²⁺ transients associated with excitation-contraction coupling or endothelin-1 exposure. As a consequence, the IP₃R2-mediated Ca²⁺ rise during diastole may activate or sensitize RyR2 channels, resulting in spontaneous extra-systolic Ca²⁺-release events and the occurrence of arrhythmias [147]. This increased “extra-nuclear” expression of IP₃R2 was also found in human heart samples derived from patients with heart failure after ischemic dilated cardiomyopathy [146].

The upregulation of IP₃R2 channels during cardiac hypertrophy was mediated via a dynamic and Ca²⁺-dependent regulation of miRNA-133a [149]. In normal physiological conditions, miRNA-133a expression is highly expressed in cardiomyocytes, thereby targeting the 3′ untranslated region of the IP₃R2 mRNA. As a consequence miRNA-133a reduces the basal expression of IP₃R2 and thereby avoids hypertrophy or arrhythmias resulting from excessive Ca²⁺ signalling. Interestingly, limiting the expression of miRNA-133a using an antagomir led to hypertrophic signalling (evident from the increased ANF expression), which was dependent on IP₃-induced Ca²⁺ release, since degrading IP₃ using IP₃ 5-phosphatase limited the increase of ANF by miR-133a antagomir.

The role of miR-133a in controlling IP₃R2 expression and the initiation of hypertrophic markers was found both ex vivo and in vivo. In isolated, hypertrophic cardiomyocytes from spontaneous hypertensive rats, IP₃R2 levels were elevated, while miRNA-133a was downregulated [149]. Overexpression of miRNA-133a in these hypertrophic cardiomyocytes reduced ANF expression to levels similar as control cardiomyocytes. In addition, in vivo application of miRNA-133a antagomir caused IP₃R2 upregulation and hypertrophic signalling. Interestingly, increased IP₃-induced Ca²⁺ release was also involved in the decreased miRNA-133a expression in hypertrophic models. Indeed, lowering IP₃ signalling by IP₃ 5-phosphatase transduction blunted the endothelin-1-induced decrease in miRNA-133a and the concomitant increase in IP₃R2 protein levels. Collectively, these findings indicate that during pathophysiological conditions associated with increased endothelin-1, increased IP₃ signalling can lead to downregulation of miRNA-133a. The latter will lead to upregulation of IP₃R2 protein levels, thereby further driving the downregulation of miRNA-133a by boosting IP₃-induced IP₃R2-mediated Ca²⁺ signalling. This perpetual feedback cycle will establish a new signalling network that favours the expression of hypertrophic genes like ANF (via hyperactivation of calcineurin/NFATc1) and the occurrence of arrhythmic events.

The mechanism by which IP₃-induced Ca²⁺ release controls miRNA-133a expression seems to involve transcription factors like the serum response factor (SRF), which is negatively regulated by the homeodomain-only protein. SRF induces miRNA-133a expression and subsequent IP₃R2 downregulation. However, during hypertrophy, IP₃-induced Ca²⁺ release may increase homeodomain-only protein expression, thereby recruiting class I histone deacetylase (HDAC) and limiting transcriptional activity of SRF.

The importance of IP₃ signalling and IP₃R2 has also been elegantly addressed by Molkentin and co-workers by the generation of transgenic mice overexpressing an IP₃ sponge, which represents a mutated, high-affinity form of the IP₃-binding core (to blunt endogenous IP₃-induced Ca²⁺ release by trapping IP₃), or overexpressing IP₃R2 (to boost IP₃-induced Ca²⁺ release) in cardiomyocytes [150]. Mice overexpressing the IP₃ sponge displayed reduced cardiac hypertrophy in response to chronic β-adrenergic stimulation and angiotensin II stimulation. In contrast, IP₃R2-overexpressing mice displayed only a mild cardiac hypertrophic phenotype under basal conditions. However, when cardiac hypertrophy was induced (e.g. using transverse aortic constriction, chronic β-adrenergic stimulation or overexpression of Gαq, an upstream phospholipase C activator) mice expressing high IP₃R2 levels demonstrated increased hypertrophic responses. Under these conditions, mice expressing low levels of IP₃R2 (except for transverse aortic constriction) also displayed enhanced cardiac hypertrophy. Moreover, IP₃R2 channels, which already display high sensitivity to IP₃, may be further sensitized by increased PKA signalling downstream of β-adrenergic receptor stimulation leading to hyperphosphorylation of IP₃R2 at Ser⁹³⁷ [97] (see section 2.4.1). The increased sensitivity of IP₃R2-expressing mice to cardiac hypertrophy-inducing conditions could be linked to increased calcineurin and NFAT signalling. Consistent with this, the augmented cardiac hypertrophic response in IP₃R2-overexpressing mice were completely blunted when these mice were crossed with calcineurin B-knockout mice, indicating an essential role of calcineurin/NFAT signalling in response to hyperactive IP₃R2-mediated Ca²⁺ signalling. While from the above studies, calcineurin emerged as the downstream target of increased IP₃R-mediated Ca²⁺ signalling, it is important to note that also nuclear CaMKIIδ has been implicated in altered transcription in response to cardiac hypertrophic endothelin-1 signalling [109]. Increased IP₃ signalling in response to endothelin-1 triggers a unique nuclear Ca²⁺ signalling that does not occur during excitation-contraction coupling but activates CaMKII, which together with protein kinase D results in the phosphorylation and nuclear export of class II HDAC5, a transcriptional repressor. In healthy conditions, nuclear HDAC5 forms a complex with the transcription factor MEF2, thereby preventing the transcription of hypertrophic genes. In hypertrophic conditions, HDAC5 is exported from the nucleus, leading to de-repression of MEF2 and the induction of hypertrophic genes. Interestingly, blocking IP₃Rs using chemicals like 2-aminoethoxydiphenyl borate or using IP₃R2-knockout mice, prevents the nuclear export of HDAC5 and subsequent activation of the hypertrophic transcription program.

All these studies are consistent with a critical role for Ca²⁺ signalling via IP₃R2 in cardiac hypertrophy, being in the nucleus and required for driving transcription of hypertrophic genes and in the junctional sarcoplasmic reticulum being responsible for driving extra-systolic Ca²⁺ rises and contractions. Moreover, these studies all support the concept of distinct Ca²⁺ signalling compartments in cardiomyocytes, either in the cytosol during physiological excitation-contraction coupling driven by RyR2 channels or in the nucleus during pathophysiological hypertrophic signalling driven by IP₃ and IP₃R2 channels [151].

5. The role of IP₃R2 in cell death and in senescence

Over the last 20 years, IP₃R channels have emerged as key regulators that control cell death and survival in a variety of cellular systems [14, 152-154]. T cells deficient in IP₃R1 are resistant to a variety of apoptotic triggers, including chemical stimuli, like corticoids, and biological stimuli, including excessive T-cell receptor stimulation and exposure to Fas ligand [155]. Interestingly, susceptibility to T-cell receptor stimulation could be restored by artificially rising the cytosolic [Ca²⁺] using the SERCA inhibitor, thapsigargin. Also, a role for IP₃R3 has emerged in pro-apoptotic Ca²⁺ signalling [156], because some studies proposed that this channel may be preferentially located in the mitochondrial ER-associated membranes. As such, IP₃R3 channels are thought to be part of the “quasi-synaptic” Ca²⁺-transport complex between the ER Ca²⁺ stores and the mitochondria that can involve IP₃Rs, GRP75 and VDAC1 [157, 158]. Nevertheless, it is becoming increasingly clear that all IP₃R isoforms participate in apoptotic Ca²⁺ signalling and/or influence the susceptibility of cells towards apoptotic stimuli. This can mean two things: i) not only IP₃R3, but also IP₃R1 and IP₃R2 channels can be part of the ER-mitochondrial junction complexes, and ii) not only direct Ca²⁺ transfer into the mitochondria, but also other downstream Ca²⁺-dependent signalling pathways participate in triggering mitochondrial outer membrane permeabilization, the point-of-no-return in apoptosis. Furthermore, it is important to emphasize that a complex interaction exists between IP₃Rs and proteins from the B-cell lymphoma (Bcl)-2 family involved in the control of apoptosis, whereby several interaction sites for such proteins have already been identified on the IP₃R [159-162].

IP₃R-mediated Ca²⁺ release can lead to calcineurin activation, which dephosphorylates the pro-apoptotic “sensitizer” BH3-only protein, Bad [163, 164]. Phosphorylated Bad is neutralized due to its scaffolding with 14-3-3 proteins and therefore it cannot form a complex with anti-apoptotic Bcl-Xl [165]. Dephosphorylation of Bad by calcineurin, e.g. in response to increases in cytosolic [Ca²⁺] mediated by IP₃Rs [164], results in Bad release from 14-3-3 proteins and its translocation from the cytosol to the mitochondrial membranes. Here, it can bind to and inhibit anti-apoptotic Bcl-Xl proteins [163], thereby displacing Bim/tBid, which then can activate Bax/Bak and induce apoptosis.

These data indicate that dampening the IP₃R-mediated Ca²⁺ rise, either by lowering IP₃R levels or altering the IP₃R-expression profile, by inhibiting the Ca²⁺-flux properties of IP₃Rs, or by lowering the ER Ca²⁺ content, which decreases the driving force for Ca²⁺ release into the cytosol upon IP₃R activation, will be cytoprotective [162]. Not surprisingly, different pro-survival signalling mechanisms, which are often oncogenic, appear to have exploited this concept to promote cell survival, including the survival of malignant or altered cells. In many cases, different mechanisms can be simultaneously operative. For instance, oncogenic KRAS mutations appear to switch the expression from IP₃R3 into IP₃R1 and to lower the ER Ca²⁺-store content, together suppressing agonist-induced Ca²⁺ release and mitochondrial Ca²⁺ accumulation and thus protecting cells against menadione exposure [166]. AKT/PKB phosphorylates all three IP₃R isoforms, thereby suppressing their pro-apoptotic Ca²⁺-release function [84, 156]. This mechanism is also exploited by tumour suppressors like the promyelocytic leukemia protein, which enhance IP₃R3 activity by counteracting PKB-mediated IP₃R3 phosphorylation [167]. Other survival/anti-apoptotic proteins, like Bcl-2, have been reported to lower ER Ca²⁺ store-content by sensitizing IP₃Rs to basal IP₃ levels and to directly suppress IP₃R-mediated Ca²⁺ release, thereby preventing toxic mitochondrial Ca²⁺ overload [168]. Evidently, these mechanisms will also result in reduced calcineurin activation, thereby limiting Bad dephosphorylation and its subsequent inhibitory effects on the anti-apoptotic Bcl-2 proteins.

While most studies have addressed the role of IP₃R1 and IP₃R3 channels in apoptosis, there is emerging evidence that IP₃R2 channels play a crucial role in mediating proapoptotic Ca²⁺ signalling. Definitely, IP₃R2 with its high sensitivity to IP₃ (see section 2.1) may actually be a very critical regulator of cell survival versus cell demise by rendering cells sensitive to basal IP₃ signalling. The role of IP₃R2 in cell death has been elucidated in different studies and using different approaches.

First of all, there is evidence that cell death triggered by cellular exposure to cytotoxic compounds or agents that induce oxidative stress has been associated with an increase in IP₃R2 levels and activity. lncreasing oxidative stress in a neuronal cell line exposed to sub-lethal concentrations of tert-butyl hydroperoxide-mediated oxidative stress led to prominent upregulation of IP₃R2 mRNA and protein levels, while IP₃R1 and IP₃R3-expression levels remained unaltered [38]. Consistent with elevated IP₃R expression levels, Ca²⁺ release from the nucleoplasm in response to a cell-permeable IP₃ ester was strongly potentiated in tert-butyl hydroperoxide-treated cells. Also, the nephrotoxic compound uranyl acetate has been shown to increase IP₃R2 mRNA and protein levels in human epithelial kidney cells, thereby increasing the basal cytosolic [Ca²⁺] and apoptosis levels [169]. Similar findings have been reported in HeLa cells exposed to fast H₂S donors, although in this case IP₃R1 expression levels were also increased [170]. Interestingly, IP₃Rs may also be directly affected by reactive oxygen species (ROS) [171]. In intact DT40 cells, superoxide anions caused Ca²⁺ release from the ER, likely via a mechanism that sensitizes IP₃Rs to basal levels of IP₃ signalling. In these DT40 cells, the presence of IP₃R2 and IP₃R1 isoforms, but not of IP₃R3, was required for superoxide anion-induced [Ca²⁺] rise in the cytosol.

The role of IP₃R2 channels in apoptotic Ca²⁺ signalling was also identified in B-cell cancer cells, in particular in a subset of “primed to death” diffuse large B-cell lymphoma cell lines [172]. Cells expressing high IP₃R2 levels seem “addicted” to the presence and recruitment of anti-apoptotic Bcl-2 proteins at the ER and especially in the IP₃R protein complex [173]. By interacting via its BH4 domain with the modulatory and transducing domain of the IP₃Rs, Bcl-2 inhibits IP₃-induced Ca²⁺ release [174-176]. The binding site for the BH4 domain of Bcl-2 (Figure 1) has been identified [175] and is conserved between the three IP₃R isoforms [177]. Importantly, a peptide tool designed to disrupt IP₃R/Bcl-2-complexes by targeting Bcl-2’s BH4 domain (see [175-178]) was very effective in inducing intracellular Ca²⁺ overload and provoking cell death in DL-BCL cells that express high levels of IP₃R2, like SU-DHL-4 cells [172]. In contrast, cells that expressed very low levels of IP₃R2 were virtually resistant to this peptide tool. The apoptotic resistance of these cells to this peptide was not due to a general defect in the initiation or execution of apoptosis, since staurosporine or BH3-mimetic drugs were very effective in these cells. We hypothesize that anti-apoptotic Bcl-2 is required at the ER to associate with the IP₃R2 to prevent its hyperactivity in response to the ongoing IP₃ signalling downstream of the B-cell receptor [173]. It remains to be elucidated whether these findings translate into primary B-cell cancer cells. In any case, disrupting IP₃R/Bcl-2 complexes results in excessive Ca²⁺-signalling patterns and apoptotic cell death in primary peripheral mononuclear blood cells (mainly B cells) isolated from chronic lymphocytic leukaemia patients [179]. Remarkably, a gene expression profile analysis using the GeneSapiens microarray database revealed an upregulation of IP₃R2 at the mRNA level in chronic lymphocytic leukaemia samples [173].

IP₃R2 channels are not only implicated in apoptosis but also play a role in cellular senescence. Stable cell cycle arrest is a key feature of cellular senescence, which is activated in response to cellular stress. Factors include oncogenic stress following loss of PTEN function, DNA damage or telomere attrition, oxidative stress and replicative stress [180]. The arrest in proliferation depends on the major tumour suppressor pathways involving p53/p21 and p16/Rb [180, 181]. At the physiological level, cellular senescence contributes to ageing at the level of the organism [182]. However, cellular senescence can also function as an important “health keeper” fighting pathophysiological conditions associated with oncogenic stress [183, 184]. As such, cellular senescence, in addition to apoptosis, is one of the pathways that counteract cancer cell initiation and tumour development [185, 186]. For instance, in pre-malignant hepatocytes, senescence led to the secretion of chemo- and cytokines, resulting in their clearance by CD4⁺ T cells [187]. Loss of immune surveillance caused the progression of the pre-malignant hepatocytes into hepatocellular carcinomas. Recently, Wiel et al. [188] performed an elegant shRNA-based screen to identify which “loss-of-function” genes can cause escape from oncogene-induced senescence in immortalized human mammary epithelial cells (HEC). Interestingly, the gene coding for IP₃R2 was identified as a prominent modulator of this form of senescence. These findings correlated with an analysis performed by the authors using the Oncomine database, which indicated that many malignant tumours displayed a decrease in IP₃R2 mRNA levels. IP₃R2 shRNAs alleviated the growth arrest in HEC exposed to oncogenic stress. Prolonged incubation of these cells with cell-permeable IP₃ repressed cell growth and induced pre-mature senescence. Oncogenic stress-induced senescence led to an increase in the Ca²⁺ accumulation in the mitochondria, a process that did not occur in the IP₃R2 shRNA-treated cells, and also boosted IP₃-induced mitochondrial Ca²⁺ uptake. This mitochondrial Ca²⁺ uptake was proposed to be responsible for the decrease in mitochondrial potential observed during oncogene-induced senescence, because shRNA against the IP₃R2 or against the mitochondrial Ca²⁺ uniporter prevented this decline in mitochondrial potential. Interestingly, chemical induction of mitochondrial depolarization blocked cell growth and induced pre-mature senescence. The role of IP₃R2 and of the subsequent mitochondrial Ca²⁺ accumulation was linked to an increase in ROS production, since anti-oxidants promoted oncogene-induced senescence escape. Finally, these concepts may not be limited to oncogene-induced senescence, but may also be applicable in models of replicative senescence. IP₃R2 knockdown counteracted the increase in mitochondrial Ca²⁺ and the decline in mitochondrial potential observed during replicative senescence, thereby delaying the occurrence of senescence in these models.

6. Conclusions

IP₃R2 is characterized by a number of important and specific properties, including, but not limited to, its high sensitivity to IP₃ and ATP. Other properties such as its regulation by protein kinases, its interaction with adenylate cyclase to couple to cAMP production, its ability to recruit associated proteins and its low mobility in the ER, remain underexplored. At the functional level, it is clear that IP₃R2 is not only important for regulating secretion, but also is implicated in health and disease, including prominent roles in cardiac function and tumour growth. The available evidence indicates that tumour cells either downregulate IP₃R2 expression or dampen its activity via Bcl-2, since IP₃R2 can promote senescence and/or apoptosis. It is now anticipated that further research will elucidate additional important functions of IP₃R2 in other tissues and organs and further that developing tools specifically targeting or impacting IP₃R2 will allow modulating its function in disease states.

Highlights.

• * The understanding of IP₃R2 has long lagged behind that of the other IP₃R isoforms

• * IP₃R2 is an intracellular Ca²⁺-release channel with important and unique properties

• * IP₃R2 performs crucial physiological functions in various cell types

• * IP₃R2 is implicated in health and disease, including cardiac hypertrophy and cancer

• * IP₃R2 forms an important target for further research

Abbreviations

AC6

Adenylate cyclase type 6

AKT/PKB

Protein kinase B

ANF

Atrial natriuretic factor

Bcl

B-cell lymphoma

BH

Bcl-2 homology domain

CaMKII

Ca²⁺/calmodulin dependent protein kinase II

DT40 TKO

DT40 triple knockout

EMT

Epithelial-mesenchymal transition

ER

Endoplasmic reticulum

ET_AR

Endothelin receptor type A

HDAC

Histone deacetylase

HEC

Immortalized human mammary epithelial cells

IP₃

Inositol 1,4,5-trisphosphate

IP₃R(1, 2, 3)

Inositol 1,4,5-trisphosphate receptor (type 1, type 2, type 3)

NFAT

Nuclear factor of activated T cells

PKA

Protein kinase A

PKC

Protein kinase C

P_o

Open probability

ROS

Reactive oxygen species

RyR

Ryanodine receptor

SERCA

Sarco/endoplasmic reticulum Ca²⁺ ATPase

SRF

Serum response factor