Calcium Signaling in the Liver

Abstract

Intracellular free Ca²⁺ ([Ca²⁺]i) is a highly versatile second messenger that regulates a wide range of functions in every type of cell and tissue. To achieve this versatility, the Ca²⁺ signaling system operates in a variety of ways to regulate cellular processes that function over a wide dynamic range. This is particularly well exemplified for Ca²⁺ signals in the liver, which modulate diverse and specialized functions such as bile secretion, glucose metabolism, cell proliferation, and apoptosis. These Ca²⁺ signals are organized to control distinct cellular processes through tight spatial and temporal coordination of [Ca²⁺]i signals, both within and between cells. This article will review the machinery responsible for the formation of Ca²⁺ signals in the liver, the types of subcellular, cellular, and intercellular signals that occur, the physiological role of Ca²⁺ signaling in the liver, and the role of Ca²⁺ signaling in liver disease.

Introduction

Many of the functions of the liver are regulated by increases in intracellular Ca²⁺ ([Ca²⁺]i; a list of abbreviations used in this article is provided in Table 1). This includes vesicular trafficking and canalicular contraction, which are regulated by cytosolic Ca²⁺ ([Ca²⁺]c) and participate in the control of bile secretion (252, 398, 400, 415); glucose and energy metabolism, which occur through the modulation of regulatory enzyme activity (35, 109, 200) and through changes in mitochondrial Ca²⁺ ([Ca²⁺]m) (154, 326); and control of the cell cycle, which includes modulation of transcription (157, 314) and anti- and proapoptotic proteins (17, 221, 258) to regulate cell proliferation and death, and is regulated by nucleoplasmic Ca²⁺ ([Ca²⁺]n) (332). To control, yet discriminate among so many functions simultaneously, [Ca²⁺]i signals are highly organized in time (e.g., through the amplitude of Ca²⁺ spikes and the frequency of Ca²⁺ oscillations) and space (e.g., through Ca²⁺ gradients and waves). These specialized features of Ca²⁺ signals are mediated by the PLC/InsP3/InsP3R cascade, which may be activated by Ca²⁺ mobilizing hormones or growth factors and leads to release of Ca²⁺ stored in the endoplasmic reticulum (ER) (24, 25, 27, 72, 105, 379, 381). The interplay between different organelles in liver cells also determines the proper spatio-temporal modulation of Ca²⁺ signals and the execution of specific functions (310, 326, 343). In addition, Ca²⁺ signaling patterns are not just organized within individual cells but are coordinated among cells as well, through a complex communication network dependent on signaling through gap junctions (105, 266, 344) and paracrine messengers (104, 115, 259, 351) that extends across the hepatic lobule. This article describes the basic components required for the formation of Ca²⁺ signals in the liver, the mechanisms of distinct Ca²⁺ signaling pattern formation, and their involvement in processes related to health and disease.

Table 1.

List of Abbreviations

[Ca2+]c	Cytosolic calcium
[Ca2+]i	Intracellular calcium
[Ca2+]m	Mitochondrial calcium
[Ca2+]n	Nucleoplasmic calcium
5-FU	5-fluorouracil
Ach	Acetylcholine
ADP	Adenosine diphosphate
AP-1	Activator protein 1
ATP	Adenosine triphosphate
ATF6α	Activating transcription factor 6α
AVP	Arginine vasopressin
BDL	Bile duct ligation
BSEP	Bile salt export pump
cADPR	Cyclic adenosine diphosphate ribose
CaMK	Calcium-calmodulin-dependent protein kinase
cAMP	Cyclic adenosine monophosphate
CD38	Adenosine diphosphate-ribosyl cyclase
CFTR	Cystic fibrosis transmembrane conductance regulator
CGamF	Cholylglycylamido-fluorescein
CICR	Calcium-induced calcium release
c-Met	Hepatocyte growth factor receptor
CMFDA	5-chloromethylfluorescein diacetate
cPKCα	Ca2+-dependent conventional protein kinase C alpha
CRAC	Calcium release activated channel
CREB	Cyclic adenosine monophosphate response element binding
CsA	Cyclosporin A
Cx26	Connexin 26
Cx32	Connexin 32
DAG	Diacylglycerol
DMSO	Dimethylsulfoxide
EGF	Epidermal growth factor
Elk1	E twenty-six-like transcription factor 1
ER	Endoplasmic reticulum
ERK	Extracellular signal-regulated kinase
FLZ	N-(2-(4-hydroxy-phenyl)-ethyl]-2-(2,5-dimethoxy-phenyl)-3-(3-methoxy-4-hydroxy-phenyl)-acrylamide
FXR	Farnesoid X receptor
GDP	Guanosine diphosphate
GFP	Green fluorescent protein
GPCR	G-protein-coupled receptor
HBV	Hepatitis B virus
HCC	Hepatocellular Carcinoma
HCV	Hepatitis C Virus
HGF	Hepatocyte growth factor
HSC	Hepatic stellate cells
IFNα	Interferon-alpha
IGF-1	Insulin-like growth factor 1
IMM	Inner mitochondrial membrane
INM	Inner nuclear membrane
InsP3	Inositol-1,4,5-trisphosphate
InsP3R	Inositol-1,4,5-trisphosphate receptor
IRE1α	Inositol requiring enzyme 1α
JNK	c-jun-N-terminal kinase
LCA	Lithocholic acid
LGMN	Legumain
LPS	Lipopolysaccharide
mCICR	Mitochondrial calcium induced calcium release
MCU	Mitochondrial calcium uniporter
MRP2	Multidrug resistance associated protein 2
NAADP	Nicotinic acid adenine dinucleotide phosphate
NAD(P)H	Nicotinamide adenine dinucleotide phosphate
NAFLD	Non-alcoholic fatty liver disease
NE	Nuclear envelope
NFAT	Nuclear factor of activated T-cells
NFKB	Nuclear factor kappa B
NPC	Nuclear pore complex
NR	Nucleoplasmic reticulum
OAP	Oct-1 associated protein
OMM	Outer mitochondrial membrane
ONM	Outer nuclear membrane
PDGF	Platelet-derived growth factor
PERK	Protein kinase ribonucleic acid-like endoplasmic reticulum kinase
PH	Partial hepatectomy
PIP2	Phosphotidylinositol-4-5-bisphosphate
PKA	Cyclic adenosine monophosphate-dependent protein kinase
PKC	Protein kinase C
PKG	Cyclic guanosine monophosphate-dependent protein kinase
PLC	Phospholipase C
PT	Permeability transition
PTB	Phosphotyrosine binding
PTP	Permeability transition pore
PV	Parvalbumin
RaM	Rapid uptake mode
RASAL	Ras guanosine triphosphatase activating protein
RaSH	Rapid Subtraction Hybridization
ROS	Reactive oxygen species
RTK	Receptor Tyrosine Kinase
RyR	Ryanodine receptor
SERCA	Sarco/endoplasmic reticulum calcium adenosine triphosphatase
SH	Src homology domain
SOC	Store-operated calcium channel
STIM1	Stromal interaction factor 1
TBH	Tertbutylhydroperoxide
tBHP	Tert-butyl hydroperoxide
TUDCA	Tauroursodeoxycholic acid
TLCA	Taurolithocholic acid
TPC	Two-pore channel
TRPV	Transient receptor potential vanilloid
UDCA	Ursodeoxycholic acid
UPR	Unfolded protein response
VDAC	Voltage-dependent anion channel

General Molecular and Cellular Mechanisms of Ca²⁺ Signaling

Molecular machinery for Ca²⁺ signal formation in the cytoplasm

The Ca²⁺ signaling system contains a wide molecular set of signaling components that act in synergy to drive Ca²⁺ signals with unique spatial and temporal properties (25, 28). [Ca²⁺]c signals can be derived from the extracellular medium or from internal stores. In the case of the latter, several intracellular messengers bind to and regulate the activity of specific intracellular receptors to promote Ca²⁺ release (25, 41). Inositol-1,4,5-trisphosphate (InsP3) is one such Ca²⁺-mobilizing messenger and acts through the InsP3 Receptor (InsP3R). Similarly, Ca²⁺ itself and cyclic ADP ribose (cADPR) promote [Ca²⁺]c signals through their interaction with ryanodine receptors (RyRs). Additionally, nicotinic acid adenine dinucleotide phosphate (NAADP) activates two-pore channels (TPCs) (25, 28, 41, 58, 178).

InsP3-mediated Ca²⁺ release is initiated by the binding of Ca²⁺ mobilizing hormones or growth factors to plasma membrane G-protein-coupled receptors (GPCRs) (23, 25, 357) or receptor tyrosine kinases (RTKs) (23, 25, 220), respectively. GPCRs are composed of seven hydrophobic, helical, membrane-spanning domains and a cytosolic G-protein binding site. G-proteins are heterotrimeric and consist of α, β, and γ subunits (357, 414). Each α-subunit has intrinsic GTPase activity at the guanine nucleotide binding site, and specific binding sites for effector proteins such as phospholipase C (PLC), adenylate cyclase, or other hydrolases. Once a ligand binds to its specific GPCR, guanosine diphosphate (GDP) is rapidly exchanged for GTP, which results in G-protein dissociation from the receptor and α-subunit dissociation from the β-γ complex. The α and β-γ complex each may then activate effector proteins in different biochemical pathways. When GTP is hydrolyzed by GTPase, the heterotrimeric G-protein complex reassociates and can be reactivated (188, 357). Several GPCRs are expressed and have been particularly well studied in the liver. These include the V1a vasopressin receptor, the α_1B adrenergic receptor, several subtypes of the P2Y class of purinergic receptors, and the angiotensin receptors (129, 337, 338, 381).

RTKs exhibit ligand-activated protein kinase activity and have only one membrane-spanning helix, connecting the extracellular ligand-binding domain to the cytoplasmic protein tyrosine kinase domain. Kinase activity is switched on by receptor dimerization and transphosphorylation of tyrosine residues within the kinase activation loop in the catalytic cytoplasmic domain. This in turn leads to the recruitment of a series of adaptor proteins containing an SH2 (Src homology domain 2), SH3 (Src homology domain 3), or PTB (phosphotyrosine binding) activating signaling pathways that generally regulate cell proliferation, migration, and survival. Protein phosphatases may terminate the signal by dephosphorylating tyrosine residues to disassemble the signaling complexes formed around the activated receptors (25, 220, 350). Several RTKs are expressed in the liver (40, 64, 170, 331, 363, 373) including the hepatocyte growth factor (HGF) receptor c-Met, the epidermal growth factor (EGF) receptor, and the insulin receptor, and their actions are implicated in glucose and lipid metabolism (254, 358), liver fibrosis (18), regeneration (389), and hepatocellular carcinogenesis (181, 383, 389).

Once Ca²⁺-mobilizing hormones or growth factors bind to their specific plasma membrane receptors, membrane-associated PLC is activated. There are multiple subtypes of PLC (323), and several isoforms of these subtypes have been identified; in particular, PLCβ1 and PLCβ2 are activated by GPCRs, while PLCγ is activated by RTKs (25, 323). Activation of PLC by GPCRs hydrolyzes phospholipid phosphotidylinositol-4-5-bisphosphate (PIP2) in the plasma membrane, resulting in the formation of diacylglycerol (DAG) and InsP3. DAG remains at the plasma membrane to activate protein kinase C (PKC), while InsP3 diffuses into the cytosol to bind to the InsP3R, causing receptor opening and subsequent Ca²⁺ release from intracellular stores (25). Activation of PLC by RTKs was initially thought to similarly act on the plasma membrane pool of PIP2 (25) but evidence suggests it may alternatively hydrolyze nuclear PIP2 to generate InsP3 and Ca²⁺ release within the nucleoplasm (137, 331) (Fig. 1). The InsP3R is generally found in the membrane of the ER (367) and the nuclear envelope (NE) (176, 249) although InsP3Rs have been localized to the plasma membrane in certain tissues (92), as well as along the nucleoplasmic reticulum (NR) (see below) (176, 249).

Figure 1.

Molecular machinery for Ca²⁺ signal formation in hepatocytes. Ca²⁺ signals may be generated by Ca²⁺-mobilizing hormones (through activation of G-protein-coupled receptors) or growth factors (through receptor tyrosine kinases). Binding of a ligand to its specific receptor leads to the activation of phospholipase C (PLC), which hydrolyzes Phosphotidylinositol-4-5-bisphosphate (PIP2) into diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (InsP3). DAG remains in the plasma membrane and InsP3 diffuses throughout the cytosol and binds to the InsP3 receptor (InsP3R) in the endoplasmic reticulum to allow Ca²⁺ release to the cytosol. Alternatively, receptor tyrosine kinases may translocate to the nucleus to locally activate PLC and induce Ca²⁺ release from the nucleoplasmic reticulum into the nucleoplasm. Nuclear PIP2 is depicted. However, its exact localization within the nucleus remains unknown [modified from references (211) and (137), with permission].

Three isoforms of the InsP3R have been identified and are denoted types I, II, and III (125, 240, 369). KO mice have been generated for each of the three isoforms (126, 248). InsP3R I deletion causes death of most of these mice in utero and the remaining have ataxia and tonic-clonic seizures. On the other hand, InsP3R II and InsP3R III double KO animals only show gross morphological or functional defects after postnatal day 20, when they manifest weight loss due to digestive system dysfunction (126). The InsP3R sequence contains approximately 2700 amino acids and its structure has six membrane-spanning domains with the large C-terminal end directed toward the cytoplasm and the narrower end facing the ER lumen (180). There are three regions within the InsP3R: an N-terminal InsP3-binding domain, a Ca²⁺ channel-forming C-terminal domain, and a regulatory domain flanked by the InsP3-binding domain and the channel region. Along their amino acid sequence, InsP3Rs contain several sites for phosphorylation by kinases such as cyclic AMP-dependent protein kinase (PKA), cyclic GMP-dependent kinase, PKC, Ca²⁺-calmodulin-dependent protein kinase II, and Akt kinase (191). Additionally, several protein partners interact with the regulatory domain to modulate the channel activity and with the C-terminal region to determine the subcellular localization of the receptor (32, 69). Upon InsP3 binding, the receptor undergoes a conformational change resulting in opening of the Ca²⁺ channel, to allow Ca²⁺ release from the ER into the cytosol. Each InsP3R isoform acts as an InsP3-gated Ca²⁺ channel (33, 151, 317). However, their sensitivity to InsP3 is not uniform, with type II being the most sensitive, followed by type I and lastly type III (277). Ca²⁺ release through the InsP3R requires InsP3 but the concentration of Ca²⁺ in the cytosol modulates the open probability of the Ca²⁺ channel (33, 150, 317). The open probability of Type I InsP3R exhibits a bell-shaped dependence on [Ca²⁺]i concentration. Therefore, incremental increases in [Ca²⁺]i concentration up to 1 to 2 μmol/L, sustain and amplify Ca²⁺ release, while higher concentrations become progressively inhibitory. This biphasic [Ca²⁺]i concentration dependence indicates that high concentrations of Ca²⁺ can feedback to turn the channel off, making the activity of the channel self-limiting (33). This property is responsible for Ca²⁺ oscillations (33, 88, 114). In contrast, the type III InsP3R exhibits a sigmoidal dependence on [Ca²⁺]i, and thus remains open in the presence of high Ca²⁺ concentrations. In this sense, activation of this isoform results in rapid and complete emptying of Ca²⁺ stores (150, 151). Single-channel studies of the type II InsP3R revealed that, like the type III lnsP3R, this isoform stays open in the presence of high Ca²⁺ concentrations (317). Unexpectedly, cells expressing only the type II isoform exhibit sustained [Ca²⁺]i oscillations, similar to what is observed in cells expressing only the type I InsP3R (261). This may reflect the fact that modifier proteins can alter the Ca²⁺ dependence of these isoforms (256).

The sequences of the InsP3R isoforms are considerably homologous but each subtype is expressed and regulated in a distinct fashion. There also are isoform-specific differences in tissue expression and subcellular distribution, suggesting that the various isoforms serve distinct roles in [Ca²⁺]i signaling (257). Particularly, hepatocytes express only type I and type II InsP3R, and type II is the major isoform (165, 409). The type III InsP3R is not detected in hepatocytes, but is the predominant isoform in the bile duct epithelial cells, cholangiocytes (163). Moreover, isoforms may be expressed in distinct sub-cellular locations. The InsP3R II is most concentrated in the apical region of hepatocytes, while InsP3R I is expressed throughout the cell (165, 267). In contrast, the InsP3R III is most concentrated in the apical region of cholangiocytes, while InsP3R I and II are dispersed throughout the cell (163). The apical localization of the type II and III InsP3R in hepatocytes and cholangiocytes, respectively, represents a specialized region for triggering polarized Ca²⁺ waves (see below) (163, 165).

Ca²⁺ can also be released from the ER through RyRs. These channels are activated by Ca²⁺ itself through a mechanism known as Ca²⁺-induced Ca²⁺ release (CICR), or by interaction with cADPR (25, 251, 253). Although RyRs seem to be absent from hepatocytes (135, 165), there is pharmacological evidence for RyR sensitivity, since RyR inhibition reduces the speed of hormone-induced [Ca²⁺]i waves (267). In addition, ryanodine binding sites have been identified in hepatic microsomes (356), and also a truncated RyR has been identified in hepatocytes, which increases the frequency of InsP3-induced oscillations, although it is not able to promote Ca²⁺ release (303). Moreover, there is evidence that CD38 (ADP-ribosyl cyclase), the enzyme that catalyzes the formation of cADPR from NAD+ is expressed in the liver (342). In rat hepatocytes, CD38 is mainly expressed in the NE, where it contributes to nuclear Ca²⁺ ([Ca²⁺]n) signaling (133, 192). Despite the different observations suggesting the presence of RyRs in the liver, their contribution to Ca²⁺ signaling in hepatocytes is an aspect that needs further investigation.

In addition to InsP3, Ca²⁺, and cADPR, NAADP can also promote Ca²⁺ release into the cytosol (25, 127). NAADP was originally found in sea urchin eggs (209) and is the most potent Ca²⁺ mobilizing messenger yet described (128). This intracellular messenger was initially shown to interact with lysosome-related acidic compartments (71, 418). Further evidence has revealed TPCs (7, 178) as the NAADP-activated Ca²⁺ release channels, with three differentially distributed isoforms (TPC 1, 2, and 3) (50, 58, 224, 340, 424). TPC2, in particular, has predominant lysosomal localization and has been identified in most human tissues, but with particularly abundant expression in the liver (58). NAADP interacts with high affinity with TPC2 (58), although recent evidence suggests that NAADP may bind to an accessory protein within a larger TPC complex (224). This interaction is tightly controlled by luminal pH and Ca²⁺ concentration (306). TPC2 mediates [Ca²⁺]i release after activation with low nanomolar concentrations of NAADP while micromolar concentrations desensitize the receptor. Moreover, NAADP-mediated Ca²⁺ release is subsequently amplified by CICR from InsP3Rs (58, 424). Therefore, this Ca²⁺ release from lysosomal stores may play an important role in shaping [Ca²⁺]c signals (60, 294). However, endogenous activators of NAADP and physiological functions of NAADP/TPC2-mediated Ca²⁺ release in the liver still remain elusive.

Mitochondrial Ca²⁺

In addition to the role that mitochondria play in the control of metabolism and respiration, these specialized organelles also contribute to the tight spatiotemporal control of Ca²⁺ signals by taking up Ca²⁺ from or releasing Ca²⁺ back to the cytosol (310, 343) (Fig. 2). In fact, mitochondria display an enormous capacity to accumulate Ca²⁺ in the very rapid time scale of hundreds of milliseconds (61) thus shaping and maintaining [Ca²⁺]c at resting levels.

Figure 2.

Physiological and pathophysiological actions of Ca²⁺ in hepatocytes. 1. Cytosolic and mitochondrial Ca²⁺ signals are closely interrelated. Ca²⁺ is transmitted from inositol 1,4,5-trisphosphate receptors (InsP3Rs) to mitochondria, leading to an increase in mitochondrial free Ca²⁺ and the formation of the permeability transition pore (PTP). Reversible opening of the PTP is observed in normal mitochondrial function and shapes Ca²⁺ signals, but persistent opening of the PTP leads to leakage of cytochrome c, augmenting the cellular sensitivity to apoptotic stimuli. 2. Type II InsP3R-mediated Ca²⁺ release regulates organic anion and bile salt secretion through the insertion of the transporters multidrug resistance-associated protein 2 (MRP2) and bile salt export pump (BSEP), respectively, into the canalicular membrane. Loss of this isoform results in impaired transporter activity and contributes to the pathophysiology of intrahepatic cholestasis. 3. Nuclear Ca²⁺ is essential for cell-cycle progression through the control of gene transcription and expression of proliferative proteins. Furthermore, nuclear Ca²⁺ may be associated with liver tumor growth. 4. The endoplasmic reticulum (ER) regulates protein folding, quality control, trafficking, and targeting, and these depend on adequate amounts of Ca²⁺ in the ER lumen. Unbalance between ER load and protein-folding capacity leads to ER stress. In obesity, liver expression and activity of SERCA are impaired, leading to ER stress and defective glucose metabolism.

Ca²⁺ uptake across the inner mitochondrial membrane (IMM) is mediated by the ruthenium-red-sensitive [Ca²⁺]m uniporter (61, 81, 237, 349) and/or the rapid uptake mode, a transport mechanism that might contribute to [Ca²⁺]m uptake from fast [Ca²⁺]c transients in vivo (147). In addition, the outer mitochondrial membrane voltage-dependent anion channel may modulate Ca²⁺ access to the mitochondrial intermembrane space (136). This type of mitochondrial transport is driven by the potential gradient across the mitochondrial membrane (146). Ca²⁺ efflux occurs through the electrogenic Na+-Ca²⁺ antiporter and through a Na+-insensitive but H+-sensitive Ca²⁺ efflux pathway (22, 167). These efflux mechanisms are active transport systems (146). The molecular identity of these transport systems has remained largely elusive. However, recent studies have revealed novel candidate proteins that may mediate Ca²⁺ transport across the IMM (152). Ca²⁺ can also be released from mitochondria through the permeability transition pore (PTP), a voltage-dependent, high conductance channel that requires a permissive load of matrix Ca²⁺ for opening (20, 21, 80). Formation of the PTP results from increases in mitochondrial free Ca²⁺, membrane depolarization, and alkaline matrix pH (147, 299), and induces the PT, a permeability increase of the IMM to ions and solutes with molecular masses up to 1500 Da (20, 21, 80). Activation of the PTP leads to rapid release of large amounts of [Ca²⁺]m into the cytosol (173, 183). The PTP can be inhibited by antioxidants, reducing agents, and the immunosuppressive peptide cyclosporin A (21, 146, 278). Endogenous inhibitors of PT include adenosine diphosphate (ADP), Mg²⁺, and increased mitochondrial membrane potential (146). Persistent opening of the PTP can induce irreversible cell injury and death (20, 21, 80, 146, 276). On the other hand, reversible activation of the PTP is observed in normal mitochondrial function and plays a role in [Ca²⁺]i signaling (146, 173, 174) (Fig. 2). This is possible through several signaling steps; the mitochondrial membrane potential gradient allows [Ca²⁺]m uptake through the uniporter; this influx of cations is compensated by H+ efflux resulting in increased mitochondrial pH, which in turn activates formation of the PTP. This alters the membrane potential gradient, allowing Ca²⁺ to be released from the mitochondria, triggering [Ca²⁺]m-induced Ca²⁺ release (mCICR) (173). mCICR, in regions where mitochondria are densely distributed, allows Ca²⁺ waves to propagate across the cytosol (153, 173). Indeed, mitochondria participate in hepatocyte [Ca²⁺]i signaling (85, 153). This ability to contribute to the spatiotemporal control of Ca²⁺ signaling is due to the close association between mitochondria and [Ca²⁺]i release sites. This has been confirmed by the observation that [Ca²⁺]m can closely parallel the [Ca²⁺]i increase induced by receptor activation. Furthermore, the increase in [Ca²⁺]m following release of Ca²⁺ from intracellular stores is faster and larger than the increase that follows influx of extracellular Ca²⁺ (324). There is not only close proximity but also dynamic physical interaction between mitochondria and the ER (84, 348). Additionally, there is functional evidence for microdomains where mitochondria and InsP3Rs are in close apposition, so that mitochondria take up a significant fraction of the Ca²⁺ released by the InsP3R (85, 153, 183, 324). Approximately 30% of the total cellular mitochondrial pool is located in these microdomains of high [Ca²⁺]i (367). Of note, sub-cellular regions that contain fewer mitochondria show greater sensitivity to InsP3 (153). This property of mitochondria can locally modulate InsP3R sensitivity and thus define the threshold for InsP3 to trigger [Ca²⁺]i signals in different subcellular regions (153).

[Ca²⁺]m uptake affects multiple factors in cell metabolism such as mitochondrial proton motive force, electron transport and the activity of dehydrogenases associated with the tricarboxylic acid cycle (154, 326). In isolated hepatocytes, slow or small [Ca²⁺]i elevations are not transmitted effectively into mitochondria, and therefore, are unable to activate mitochondrial metabolism (326). Conversely, [Ca²⁺]i oscillations trigger [Ca²⁺]m oscillations and sustained nicotinamide adenine dinucleotide NAD(P)H formation (311). Therefore, the frequency rather than the amplitude of [Ca²⁺]i oscillations regulates mitochondrial metabolism (154, 311). Indeed, Ca²⁺ signaling in the cytosol and mitochondria are closely related, allowing the tight regulation of spatiotemporal Ca²⁺ signaling and cell metabolism.

Nuclear Ca²⁺

Most Ca²⁺ signaling machinery found in the cytoplasm is duplicated in the nucleus, and there is evidence that [Ca²⁺]n signals can occur independent of [Ca²⁺]c signals (106, 215). The nucleus is bordered by the NE, a specialized region of the ER, which has a double function as it insulates the nucleoplasm from the cytoplasm and also has the ability to store and release Ca²⁺ (207, 249, 279). It is noteworthy that the NE is not simply a smooth-surfaced outer boundary but is interrupted by invaginations that reach deep within the nucleoplasm and could even traverse the nucleus completely. The existence of such a complex branched network of invaginations forming a NR represents an additional regulatory Ca²⁺ domain within the nucleus (106, 234). Phosphoinositides (as well as the enzymes responsible for their synthesis, namely, phosphatidylinositol and phosphatidylinositol 4-phosphate kinases) have long been known to be present in the nucleus (176), although their precise localization within the nucleus remains unknown (177). Furthermore, different PLC isoforms have been identified in the nucleus and may provide in situ production of IP3 and DAG (74, 394). Both the NE and NR express InsP3Rs and numerous lines of evidence support a role for nuclear InsP3Rs in regulating [Ca²⁺]n release (57, 160, 171, 215, 238, 362). Of note, InsP3Rs can be found in the outer (ONM) (62, 233, 362) and inner nuclear (241, 249) membranes. Additionally, sarco/ER Ca²⁺ ATPase (SERCA) uptake pumps are found in the ONM and have been shown to be identical to those in the ER (207). Therefore, the nucleus expresses the machinery necessary for local Ca²⁺ release and reuptake. Such machinery may be activated selectively through tyrosine kinase pathways (73). Indeed, IGF-1 and integrins cause PIP2 breakdown in the nucleus but not at the plasma membrane (73), while activation of GPCRs causes breakdown of PIP2 in the plasma membrane, but not in the nucleus (307). Similarly, activation of the HGF receptor c-Met or the insulin receptor causes their rapid translocation to the nucleus, to selectively hydrolyze nuclear PIP2 and generate InsP3-dependent [Ca²⁺]n signals (137, 331). It also has been hypothesized that relocation of MAP kinase to the nucleus activates nuclear PLC to generate InsP3 in the nucleus (347).

[Ca²⁺]n signaling may also occur by passive transmission of [Ca²⁺]i signals into the nucleus through the nuclear pore complexes (NPCs), which are involved in transport of macromolecules from the cytoplasm to nucleoplasm and vice versa (366). In fact, this mechanism was originally thought to be the principal one responsible for formation of [Ca²⁺]n signals. The ion conductance of NPC is very large and this should allow free solute and ion diffusion. Then, [Ca²⁺]c signals should generate identical [Ca²⁺]n signals (3). It has been reported, however, that the NPC conductance can be dramatically reduced by accumulation of Ca²⁺ inside nuclear cisterna (gating) or by transport of macromolecules through the NPC (plugging) (365), although this view is controversial (132). Opinions on the Ca²⁺ permeability of NPC are also far from uniform. Both reports of free (51, 355) and restricted (11, 134, 230) permeability of Ca²⁺ through NPC have been shown. Studies monitoring diffusion of photoactivatable GFP across the NE suggest that permeability of the nuclear pore may be regulated by [Ca²⁺]i rather than by Ca²⁺ stores within the NE (285). In either case, since EF-hand Ca²⁺-binding motifs are present in proteins of the nuclear pore, it is possible that these function as Ca²⁺ gating sensors for the pore (347).

Although [Ca²⁺]n signals can originate inside the nucleus, and the permeability of nuclear pores to Ca²⁺ is regulated, there is considerable evidence that [Ca²⁺]n can passively follow [Ca²⁺]i (160, 225). For example, stimulation of hepatocytes with vasopressin results in a Ca²⁺ wave that appears to spread from the cytosol into the nucleus (223). Moreover, a mathematical analysis of Ca²⁺ waves in hepatocytes stimulated with vasopressin suggests that [Ca²⁺]n signals can be described simply by diffusion of Ca²⁺ inward from the NE (118).

[Ca²⁺]n can also contribute to Ca²⁺ signals in the cytosol. For example, Ca²⁺ puffs, which are highly transient and localized [Ca²⁺]i signals that result from the coordinated opening of small clusters of InsP3Rs, can spread across the cell by diffusing across the nucleus (421). Puffs can be triggered by a subthreshold concentration of agonist and the resulting [Ca²⁺]i signal rapidly dissipates by diffusion in the cytosol and sequestration of Ca²⁺ into intracellular stores. However, since the range of diffusion of Ca²⁺ in the nucleus can be much greater than in the cytosol (118), Ca²⁺ puffs generated near the NE can spread into and across the nucleus to spread to other, more distant regions of the cytosol (225). Thus, the nucleus may function as a path that helps distribute Ca²⁺ to the cytosol.

Ca²⁺ signaling patterns in individual cells

Ca²⁺ signals are highly organized in both space and time, from the subcellular to the cellular to the whole tissue level, to ensure reliability and specificity (24, 72, 105, 381). In the liver, this organization is achieved in the form of single transient or sustained [Ca²⁺]i increases, Ca²⁺ oscillations and [Ca²⁺]i waves (105, 381). In hepatocytes, Ca²⁺ oscillations originate from the periodic opening of Ca²⁺ channels located in the membrane of the ER, following activation of the PLC/InsP3/InsP3R cascade, and their shape and duration depends on the concentration and the type of agonist in use (413). For example, populations of isolated hepatocytes respond to agonists such as vasopressin, phenylephrine, or angiotensin with regular biphasic increases in [Ca²⁺]i, composed by two main events. The first is a rapid peak and then fall in [Ca²⁺]i, which occurs over a period of seconds. This is the result of Ca²⁺ release from InsP3-sensitive stores and does not depend on extracellular Ca²⁺, since it can be observed in Ca²⁺-free medium (189, 337). The second event is the sustained plateau in [Ca²⁺]i, which follows the rapid peak. This occurs only in the presence of extracellular Ca²⁺ that is necessary to reload depleted intracellular stores (155). In the case of vasopressin, lower concentrations induce [Ca²⁺]i oscillations, while higher concentrations induce sustained increases in [Ca²⁺]i. In contrast, stimulation of hepatocytes with phenylephrine or the bile acid ursodeoxycholic acid (UDCA) typically evoke [Ca²⁺]i oscillations at all concentrations (30, 337). Additionally, ATP stimulation leads to complex oscillations corresponding to a large peak, followed by smaller amplitude oscillations superimposed on a plateau phase (98). In all cases, the information of the agonist-induced Ca²⁺ increase is encoded in the rate of Ca²⁺ events or the amplitude of the Ca²⁺ rise, a phenomenon known as frequency encoding (26, 104, 105, 129). The duration of the [Ca²⁺]i oscillations can also be determined by the agonist (413). For example, phenylephrine induces shorter [Ca²⁺]i spikes than vasopressin or angiotensin (412, 413).

Frequency-encoded Ca²⁺ signals are thought to play an important role in determining the extent and targeting of [Ca²⁺]i signals to downstream Ca²⁺-sensitive proteins, yielding higher fidelity than other types of [Ca²⁺]i signals (e.g., amplitude modulated) (379, 381). In this sense, the frequency of [Ca²⁺]i spiking has been shown to differentially induce gene expression through the transcription factors, NFAT, NFKB, and OAP (100, 222, 384), to modulate Ras and the MAP kinase pathway (384) and to regulate mitochondrial metabolism (154) with greater efficacy than a single Ca²⁺ pulse or a persistent increase in [Ca²⁺]i. Decoding the information contained in a [Ca²⁺]i spike is essential to ensure the precise processing of Ca²⁺ signals. The activation and inactivation rates of Ca²⁺/calmodulin-dependent protein kinase II (87) and the Ca²⁺-dependent translocation of calmodulin (79), PKC (16, 304, 393) or the RAS GTPase activating protein, RASAL (397), suggests that these Ca²⁺-sensitive proteins can sense the frequency of Ca²⁺ signals and discretely modulate enzymatic activity. Mitochondria sense the frequency of oscillatory [Ca²⁺]i signals as well (325) by increasing the dehydrogenase NAD(P)H production in a time-averaged manner, thus modulating mitochondrial oxidative metabolism (326, 327).

Changes in InsP3 concentration are not the sole determinant of Ca²⁺ oscillations (246). Thapsigargin (117), which releases Ca²⁺ from the ER through an InsP3-independent mechanism, and nonhydrolyzable InsP3 analogs (396) can trigger [Ca²⁺]i oscillations. The bell-shaped dependence of the InsP3R open probability on [Ca²⁺]i concentration is thought to be one of the regulatory mechanisms responsible for oscillatory Ca²⁺ signals (33, 88, 317). Additionally, the maintenance of hormone-induced Ca²⁺ oscillations requires the constant replenishment of the ER with Ca²⁺ entering the cell through Ca²⁺ entry channels in the plasma membrane. Although several types of Ca²⁺ entry channels may be involved in maintaining proper Ca²⁺ concentrations in the ER, store-operated Ca²⁺ channels (SOCs), which are activated by a decrease in Ca²⁺ in the ER lumen, play a major role in liver cells (14, 15, 182). The liver expresses the SOCs Orai1 (a member of the Ca²⁺ release-activated channel modulator family of proteins) and transient receptor potential vanilloid 1 (TRPV1), which are plasma membrane Ca²⁺ channels, and the ER protein stromal interaction factor 1 (STIM1). It has been shown that these SOCs have a high selectivity for Ca²⁺ and that the Orai1 and STIM1 activation mechanism involves Ca²⁺ release from a putative small subregion of the ER likely close to the plasma membrane (15). However, the nature of such an ER subregion, its relationship with the bulk of the ER, and the steps involved in STIM1 interaction with Orai1 and other proteins still need to be determined in liver cells.

Ca²⁺ signals can also encode information spatially, in the form of [Ca²⁺]i waves (105, 381). These are characterized by an initial local increase in Ca²⁺ concentration, which then propagates in the whole cell at an approximately constant rate (104). This spatial organization has been observed in isolated hepatocytes (338, 381) and in individual cells in the perfused liver (328, 382). In nonpolarized preparations of hepatocytes, vasopressin, and phenylephrine induce [Ca²⁺]i waves that begin at a single locus, and this site of origin remains constant regardless of the agonist (338). The subcellular distribution of the InsP3R has been implicated in this spatial organization of Ca²⁺ waves (161). In polarized hepatocytes, the type II InsP3R (the most abundant isoform) is localized in the pericanalicular region (267) and this localized expression establishes a trigger zone from which Ca²⁺ waves originate (165) (Fig. 3). Additionally, the function of this trigger zone as an initiation site for Ca²⁺ waves depends on the integrity of lipid rafts (263). Similarly, the apical localization of InsP3Rs determines the initiation of Ca²⁺ signals in other cell systems (272, 417) including cholangiocytes, which express mostly the type III InsP3R (163). In these systems, InsP3 is generated by PLC activation near plasma membrane hormone receptors, which are often on the basolateral membrane, so it would be counterintuitive that Ca²⁺ signals instead would begin at the apical membrane. Certain biophysical properties of the type II (317) and III (150) InsP3Rs may explain why these receptors initiate Ca²⁺ waves at the trigger zone, but an alternative explanation is that clustering of InsP3Rs in the apical region lowers the threshold for Ca²⁺ release (371, 421). Indeed, the rise time is prolonged and the Ca²⁺ wave speed is reduced in hepatocytes with decreased expression or in which there is redistribution of the type II InsP3R (161).

Figure 3.

Ca²⁺ waves begin in the apical region of hepatocytes. (A) Confocal images of an isolated rat hepatocyte couplet loaded with the Ca²⁺ sensitive dye Fluo-4/AM and stimulated with vasopressin. Serial images of the region of interest outlined in yellow show that a Ca²⁺ wave starts in the apical (a) region of the cell and spreads to the basolateral (b) region. Images were pseudocolored according to the scale shown at the bottom. (B) Graphical representation of the fluorescence increase in the apical and basolateral region shows that the apical Ca²⁺ signal precedes the basolateral Ca²⁺ signal [modified from reference (263), with permission].

The mechanism that governs apical-to-basolateral Ca²⁺ waves in hepatocytes remains elusive. In pancreatic acinar cells, the basolateral expression of the RyR Ca²⁺ release channel (213) may be the lead for Ca²⁺ wave propagation. RyR is activated by Ca²⁺ itself, which results in CICR. Therefore, InsP3 in the trigger zone initiates Ca²⁺ waves, which propagate across the cell by RyR-mediated CICR (172, 212). Cholangiocytes, on the other hand do not express the RyR but low levels of InsP3Rs are found in the basolateral region (163). Since Ca²⁺ can coactivate the InsP3R (33), Ca²⁺ waves in cholangiocytes likely propagate by a modified form of CICR in which Ca²⁺ released from the apical trigger zone sensitizes basolateral InsP3Rs allowing Ca²⁺ waves to form. In hepatocytes, pharmacological evidence using RyR inhibitors suggests that the spread of Ca²⁺ waves depends upon serial activation of apical InsP3Rs, and then basolateral RyRs (267), similar to what occurs in pancreatic acinar cells (212). However, RT-PCR studies suggest that hepatocytes lack RyRs (165). Conversely, a truncated but functional form of the RyR was detected in hepatocytes (303). The subcellular distribution of this novel RyR variant and its role in the formation of Ca²⁺ waves still needs to be determined. Of note, the speed or amplitude of [Ca²⁺]i waves in hepatocytes is not altered in Ca²⁺-free medium (267), which suggests that their propagation depends only on release of Ca²⁺ from intracellular stores. Furthermore, Ca²⁺ waves can be induced in hepatocytes by direct introduction of a nonmetabolizable analog of InsP3 (105) or by treating these cells with agents such as tert-butyl hydroperoxide or certain bile acids that raise Ca²⁺ in the cytosol independently of InsP3 (76, 336). These observations strongly suggest that Ca²⁺, and not InsP3, plays the major role in the propagation of the waves. However, a modeling system has suggested instead that InsP3, and not Ca²⁺, is necessary for InsP3-induced Ca²⁺ waves (179).

Intercellular communication: Spread of Ca²⁺ signals from cell to cell

[Ca²⁺]i waves are not confined to single cells, but rather a Ca²⁺ increase in one cell can be conveyed to neighboring cells giving rise to intercellular Ca²⁺ waves (129, 339). Ca²⁺ waves spread between hepatocytes (266), and this communication depends on gap junctions, which, in fact, are permeable to both Ca²⁺ and InsP3 (344). Either molecule could act, therefore, as the diffusible Ca²⁺-mobilizing second messenger (344). Additionally, hormone-induced [Ca²⁺]i signaling is highly coordinated in these cells, which depends upon gap junction conductance as well (266). Further evidence for a role of gap junctions has been provided by studies demonstrating blockage of intercellular Ca²⁺ waves by gap junction inhibitors (346, 385) or electroporation of antibodies to the gap junction components, connexins (38). Hepatocytes express two connexin isoforms, connexin 26 (Cx26), and connexin 32 (Cx32) (111), which is the most abundant in the liver (111, 204). Expression of both of these isoforms is dramatically reduced and coordination of [Ca²⁺]i signals is impaired after bile duct ligation (BDL) (111). Moreover, hepatocytes isolated from Cx32 KO mice have impaired intercellular propagation of Ca²⁺ signals (78, 282). Additionally, transmission of Ca²⁺ from cell to cell is possible by the expression of Cx32 or connexin 43 (Cx43) in a liver cell line where intercellular Ca²⁺ signals are not normally observed (214). Such findings show that gap junctions allow second-messenger communication to organize Ca²⁺ signals in adjacent cells. Furthermore, functional gap junction channels are essential for the spread of intralobular Ca²⁺ waves (129).

Other factors are required to support the spread of [Ca²⁺]i waves. Increases in InsP3 are necessary in each cell across which a Ca²⁺ wave spreads (39). Moreover, the presence of both InsP3 and Ca²⁺ is required to control the transmission of a [Ca²⁺]i wave across a hepatocyte (385). Additionally, Ca²⁺ waves only propagate if agonist is binding to its specific receptor at the surface of each cell. This likely means that the presence of hormone at each cell ensures that a sufficient level of excitability is achieved in the cytosol through the generation of the necessary levels of intracellular messengers, to support the propagation of a [Ca²⁺]i wave (328, 385). Intercellular Ca²⁺ waves are further organized by hepatocyte “pacemaker cells” (214). These cells have increased expression of hormone receptor, which makes them more sensitive to hormonal stimulation. Therefore, Ca²⁺ signals occur sooner in pacemaker cells than in other cells stimulated with the same concentration of hormone, due to higher InsP3 production that drives [Ca²⁺]i waves to neighboring cells (77). Of note, this increased sensitivity does not result from differences in downstream Ca²⁺ signaling components such as G-proteins or InsP3R (77, 387). This specialized organization might explain why vasopressin-induced waves begin in the region of the central venule, where the vasopressin V1a receptor is most heavily expressed, and then propagate to the portal region, where it is less heavily expressed (271). Additionally, pacemaker regions have been observed for other agonists like ATP and phenylephrine. Collectively, Ca²⁺ wave organization and propagation depends on several key aspects: InsP3 generation in each cell; a subset of cells with relatively increased expression of hormone receptor to initiate Ca²⁺ signals and thus serve as pacemakers for their neighbors; and gap junctions to coordinate the progression of Ca²⁺ signals among cells (56, 214).

Ca²⁺ waves may also be communicated between cells by paracrine signaling, involving the release of a messenger, its diffusion in the extracellular space, binding of the substance to a receptor, and activation of a signaling cascade that ultimately increases [Ca²⁺]c (104). One such extracellular messenger in the liver is ATP. This agonist is secreted by both hepatocytes (351) and bile duct cells (115, 259) and stimulates the G-protein-coupled P2Y receptors that are also expressed in both cell types (196, 250, 269). Therefore, secretion of ATP by hepatocytes stimulates InsP3-mediated Ca²⁺ signaling in neighboring hepatocytes, as well as in bile duct cells (351), creating a paracrine signaling system in which signal propagation to adjacent cells does not depend on intercellular communication through gap junctions.

Ca²⁺ signals also regulate specific functions in other liver cell types. For example, [Ca²⁺]i mobilization is implicated in the differentiation (activation) of hepatic stellate cells (HSCs) and portal (myo)fibroblasts during liver fibrosis (198, 292). HSC and portal fibroblasts express receptors for several Ca²⁺-mobilizing agonists, such as endothelin (198, 322), platelet-derived growth factor (PDGF) (86, 110), vasopressin (305) and ATP (102, 316), and upregulation of serotonin (5-HT2) receptors may be observed after HSC activation (292). Liver myofibroblast activation may exhibit different profiles of Ca²⁺ responses to endothelin and PDGF (198). Furthermore, CD38-mediated Ca²⁺ signals (through production of cADPR and NAADP) contribute to liver fibrosis via HSC activation (194). Moreover, activated fibroblasts and HSC are contractile due to their expression of motor proteins that can be regulated by Ca²⁺-dependent and Ca²⁺-sensitization mechanism pathways (175, 330, 420). In addition, InsP3R type 1 is shifted to the nucleus and to cell extensions upon HSC activation and Ca²⁺ release in these extensions promotes localized contractions (201).

Physiological Functions of Ca²⁺ Signals

Bile secretion

Bile secretion depends on the function of a number of membrane transport systems in hepatocytes and bile-duct epithelial cells (cholangiocytes) and on the structural and functional integrity of the bile-secretory apparatus (265). Ca²⁺ has been implicated in the regulation of such functions (Fig. 2). Agents that increase [Ca²⁺]c from either internal or external sources in isolated hepatocytes inhibit bile secretion in isolated perfused livers, independently of PKC or hemodynamic effects. Furthermore, inhibition of Ca²⁺ signals prevents this inhibition of bile secretion (273). These findings suggest that increased [Ca²⁺]c may play an inhibitory role in the regulation of hepatocyte bile production (273).

Possibly, Ca²⁺ inhibits bile flow by increasing paracellular permeability, which has been reported in isolated perfused livers and isolated hepatocytes (229), and would allow reflux of biliary constituents into the sinusoidal space, thereby dissipating the osmotic gradient that drives bile flow (274). In contrast to the inhibitory effect on net bile secretion, Ca²⁺ has been shown to mediate organic anion secretion through the insertion of the transporter Multidrug resistance-associated protein 2 (MRP2) into the canalicular membrane (83). Additionally, Ca²⁺ plays an important role in maintaining bile salt secretion through posttranslational regulation of the bile salt export pump (BSEP) (202) (Fig. 2). These effects are associated with pericanalicular Ca²⁺ release through the type II InsP3Rs that localize to this region in hepatocytes (165). These roles of Ca²⁺ in canalicular secretion can be supported by the observations that subplasmalemmal Ca²⁺ signals generally are obligatory for exocytosis (370), and Ca²⁺ acts as a regulator of exocytosis in the liver, in particular (53), suggesting that Ca²⁺ would be necessary for canalicular MRP2 and BSEP insertion. Moreover, the pericanalicular clustering of type II InsP3Rs creates a “trigger zone” that produces apical-to-basolateral Ca²⁺ waves in hepatocytes (165) which might promote bile secretion in a similar fashion to what has been observed in other polarized epithelia, including pancreatic acinar cells (187, 275) and cholangiocytes (116). These apparent discrepancies in the effects of Ca²⁺ on bile secretion in hepatocytes may be explained by differences in the nature of choleretic and cholestatic Ca²⁺ signals. Therefore, it may be important to consider that the magnitude, timing, and subcellular localization of Ca²⁺ signals all contribute to their physiological effects (72).

Canalicular contraction is a crucial step in conducting bile flow through the canalicular network and into the bile ducts (290, 399). Ca²⁺ mediates this process, since microinjection of Ca²⁺, or stimulation with vasopressin (266, 400) or ATP (398) promotes canalicular contraction in hepatocytes. Furthermore, [Ca²⁺]i increases result in the phosphorylation of myosin light-chain kinase which ultimately generates contractions in the pericanalicular actin network (415), which are organized within the hepatic lobule to form peristaltic waves (399). These waves are in turn essential to maintain bile flow (301).

The biliary tree plays an important role in conditioning the canalicular bile that is secreted by hepatocytes (265). Cholangiocytes perform this function through the regulation of biliary bicarbonate secretion (164). The prototypical and most extensively characterized pathway for bicarbonate secretion is initiated by stimulation of the secretin receptor which leads to formation of cAMP and activation of the cystic fibrosis transmembrane conductance regulator (CFTR). Chloride released into the ductular lumen then is thought to be exchanged for bicarbonate via a chloride-bicarbonate exchanger (186, 208). An alternative or secondary secretory pathway is stimulated by neurotransmitters such as acetylcholine (ACh) and autocrine agents such as ATP, and is mediated by InsP3 and [Ca²⁺]c (269). The downstream effect of Ca²⁺ is to activate Ca²⁺-dependent chloride channels on the apical plasma membrane, which in turn release chloride that is exchanged for bicarbonate via a chloride-bicarbonate exchanger (186, 208). Although these two pathways were initially thought to act separately, a cross-talk has been suggested by the observation that Ca²⁺ potentiates adenylyl cyclase activity and cAMP formation in cholangiocytes via a calcineurin-dependent pathway (6), which may represent an indirect mechanism for Ca²⁺-stimulated secretion. However, further evidence has suggested that cAMP- and CFTR-mediated secretion may have a more direct association with Ca²⁺ (259). Increases in cAMP result in apical, CFTR-dependent secretion of ATP, leading to stimulation of apical P2Y nucleotide receptors and subsequent activation of apical, type III InsP3Rs. This in turn activates Ca²⁺-dependent chloride channels and then bicarbonate secretion (101, 259). These observations suggest that cAMP acts through an autocrine loop that involves ATP release, activation of P2Y receptors, and then an increase in [Ca²⁺]c. Moreover, this might represent a convergence of secretory pathways in the cholangiocyte at the level of InsP3R-mediated Ca²⁺ release (259). Therefore, stimulation of basolateral M3 muscarinic receptors by ACh leads to type I and II InsP3R-mediated Ca²⁺ release. This Ca²⁺ in turn drives apical Ca²⁺-dependent chloride channels, resulting in biliary bicarbonate secretion. In contrast, stimulation of basolateral secretin receptors results in cAMP formation, subsequent activation of CFTR, and ATP-mediated bicarbonate secretion. Further studies revealed that this purinergic-signaling axis is present in both small and large cholangiocytes. ATP released from “upstream” small cholangiocytes lining the small intrahepatic bile ducts can be secreted into the bile, and can act as a paracrine signal to the large cholangiocytes lining the larger “downstream” bile ducts, stimulating Ca²⁺-dependent secretion. Moreover, small cholangiocytes, which do not express CFTR (120), respond to extracellular nucleotides through Ca²⁺-activated chloride efflux, representing an additional driving force for cholangiocyte secretion (411).

Primary cilia, which are sensory organelles expressed in the apical membrane of cholangiocytes, also relate to Ca²⁺ signaling and secretion. Primary cilia sense increases in osmolarity within the bile duct lumen, through the activation of the Ca²⁺-permeable ion channel TRPV4. This allows entry of extracellular Ca²⁺ into the cytoplasm, increases in bile flow and ATP release and bicarbonate secretion (141). Primary cilia can also sense and transduce mechanical stimuli from within the bile duct lumen, so that increases in bile flow activate both cAMP production and Ca²⁺ release (245).

Glucose metabolism

Glucose metabolism in the liver is controlled through the coordination of glycogen synthesis and breakdown. Ca²⁺ plays an important role in these events, modulating the phosphorylation and dephosphorylation, and hence the activity of the regulatory intracellular enzymes glycogen synthase and glycogen phosphorylase (35, 109, 200). Several Ca²⁺-mobilizing hormones have been implicated in these events. For example, vasopressin, angiotensin-II and α-adrenergic agonists stimulate PLC, InsP3 increase, and Ca²⁺ release from the ER, leading to the phosphorylation of glycogen phosphorylase, decrease in glycogen synthase activity, and glycogenolysis (107, 297). Nucleotides promote glycogenolysis in a similar fashion through the activation of P2Y nucleotide receptors in both rat and human hepatocytes (96, 97, 190). The bile acids UDCA, lithocholic acid, and taurolithocholic acid (TLCA) activate phosphorylase in a comparable way to vasopressin (48, 75), although through a Ca²⁺-dependent but InsP3-independent manner (48). Physiological levels of glucagon can evoke Ca²⁺ increases in the liver (67, 361) by activation of either InsP3-dependent or cAMP-dependent signaling pathways, resulting in glycogen phosphorylase activation (361, 395). The glycogenolytic capacity of the liver is regulated by several factors such as the distribution of gluconeogenic enzymes, which are localized in the periportal region (184). Perhaps for this reason, ATP and glucagon stimulate glucose release mostly in the periportal area. On the other hand, vasopressin and norepinephrine stimulate glucose release in pericentral hepatocytes, which can be explained by the fact that these cells are more sensitive to such agonists than periportal hepatocytes (129, 386). This differential sensitivity to glucose-mobilizing hormones is overcome by intercellular communication, which coordinates the regulation of glucose metabolism across the whole liver through gap junctions. For example, norepinephrine or glucagon-induced glucose release is impaired in isolated perfused livers from Cx32-deficient mice (368). In an analogous way, treatment of isolated perfused livers or isolated rat hepatocyte couplets (107) with a gap junction blocker decreases vasopressin or glucagon-induced glucose release. However, gap junction inhibition does not affect glycogen breakdown induced in a receptor-independent fashion. Moreover, hormone-induced glucose release is also impaired if hepatocytes are dispersed (107). Collectively, these observations demonstrate that gap junctions play a very important role in the control of hormone-induced glucose release in the liver in response to the heterogeneous distribution of hormone receptors. Stress conditions can challenge this coordination of metabolic function; Cx32-deficient mice become hypoglycemic in response to fasting, whereas WT animals do not, and endotoxin-induced hypoglycemia is aggravated in Cx32-deficient mice (78). Because the coordination of glucose storage and release assures balanced glucose metabolism in the liver, the actions of these two pathways are complemented by one another to maintain homeostasis. For example, insulin antagonizes the glycogenolytic activity of glucagon and α1-adrenergic agonists in the liver (34, 89, 407). It is thought that insulin interferes with glucagon by reducing elevated levels of cAMP (34, 159, 407), whereas insulin inhibits the effects of α1-agonists by interfering with their ability to elevate cytosolic free Ca²⁺ (380).

Cell proliferation

Ca²⁺ plays an important role in cell proliferation (143, 364) as well and in progression of the cell cycle (185, 308, 318, 334, 390). Temporal and spatial patterning of Ca²⁺ signals determines their specificity, and Ca²⁺ signaling patterns can vary in different parts of the cell (28, 72, 157, 215). For example, oscillation frequency, amplitude, and time delay of cytosolic calcium oscillations are essential kinetic parameters to determine entry into the cell cycle (403, 404). Furthermore, the proliferative ability of calcium is closely related to the subcellular compartments where it is released. The heterologous expression of the Ca²⁺-binding protein parvalbumin (PV) has been used to study the role of Ca²⁺ signaling in the regulation of the cell cycle. PV variants targeted to the nucleus or the cytoplasm were used to investigate the relative importance of Ca²⁺ signals in each of these cellular compartments for regulation of the cell cycle in the hepatoma cell lines SkHep1 and HepG2. It was found that nucleoplasmic rather than [Ca²⁺]c is essential for cellular proliferation, and is necessary, in particular, for progression through early prophase (332) (Fig. 2). On the other hand, both cytosolic and nuclear Ca²⁺ signals were important for proliferation of LX-2 immortalized human HSCs and primary rat HSCs through CaMK II-mediated regulation of Cdc25C phosphorylation (359). Additional studies led to the observation that receptors for HGF, EGF, and insulin, three potent growth factors in the liver, translocate to the nucleus to generate InsP3-dependent Ca²⁺ signals (9, 137, 331), suggesting that certain growth factors may stimulate proliferation of hepatocytes by selectively inducing Ca²⁺ signals in the nucleus.

What are the specific targets of Ca²⁺ signals in the nucleus that are responsible for cell proliferation? [Ca²⁺]n signals activate the transcription factors cAMP response element binding (CREB) (157) and Elk1 (314) and stimulate the intranuclear activity of PKC (106) and CaMK-IV (90), which would be expected to stimulate cell proliferation (Fig. 2). A recent study used a rapid subtraction hybridization (RaSH) approach to select genes whose expression was affected by a small alteration in [Ca²⁺]n concentration. Legumain (LGMN), an asparaginyl endopeptidase involved in tumor metastasis was identified in this screen as a novel specific target for [Ca²⁺]n signals in SkHep1 cells. [Ca²⁺]n buffering resulted in decreased LGMN expression. Furthermore, LGMN gene knockdown resulted in decreased cell proliferation, suggesting that expression of this protein is associated with increased [Ca²⁺]n-dependent proliferation. These findings highlight a new role for LGMN and provide evidence that [Ca²⁺]n signals regulate cell proliferation in part through the modulation of LGMN expression (8).

Ca²⁺ also plays an important role in the proliferation of bile duct cells. The proliferative activity of large cholangiocytes is regulated by the activation of cAMP-dependent mechanisms (4, 121, 219). On the other hand, small cholangiocyte proliferation is regulated by the activation of α1-adrenergic receptors and occurs through Ca²⁺/calcineurin-dependent activation of nuclear factor of activated T cells 2 and specificity protein 1. Furthermore, histamine stimulates small cholangiocyte proliferation through the activation of the H1HR receptor and InsP3/Ca²⁺ signaling (120, 122). These findings implicate α1-adrenergic and histamine receptors (in particular, H1HR) as key factors in the regulation of normal cholangiocyte proliferation and function. This is also important in pathological ductopenic conditions associated with damage of large ducts in which Ca²⁺-dependent small cholangiocyte proliferation may be a key compensatory mechanism for maintaining homeostasis and overall bile duct function (219, 239).

Apoptosis

Apoptosis can result from a variety of extracellular or intracellular stimuli, which converge with activation of caspases to lead directly to cell death (227). Apoptosis generally occurs through one of two pathways. The extrinsic pathway is initiated by death receptors of the tumor necrosis factor receptor superfamily (108). The intrinsic pathway involves the convergence of intracellular stress signals on mitochondria, resulting in the formation of the PTP (10, 63). In many cells, the death receptor pathway may act through mitochondria as well by inducing the cleavage of the proapoptotic molecule Bid and subsequent mitochondrial permeabilization (231). Therefore, mitochondrial permeabilization is key to both pathways. Apoptosis is modulated by a number of proteins in part through the InsP3R or InsP3R-induced Ca²⁺ release mechanisms (Fig. 2). These in turn mediate effects in [Ca²⁺]m, since a subset of mitochondria are in close proximity to InsP3Rs (324), and mitochondrial and [Ca²⁺]c signals are interrelated (154, 183). In particular, the mitochondrial apoptotic pathway is inhibited by members of the Bcl-2 family of proteins, including Bcl-2, Bcl-xL, and Mcl-1 (1, 144, 243, 321). Bcl-2 inhibits apoptosis in part by decreasing the size of ER Ca²⁺ stores (17), whereas Bcl-xL acts in part by inhibiting expression of the InsP3R (221, 405). Collectively, this reduces Ca²⁺ release from the ER into the cytosol and consequently decreases the amount of Ca²⁺ transmitted to surrounding mitochondria. Since mitochondria are less likely to become overloaded with Ca²⁺, the formation of the PTP is prevented and the overall result is a diminished cellular sensitivity to apoptotic stimuli (17, 221) (Fig. 2). In fact, overexpression of Bcl-xL provides protection against apoptotic cell death in rat hepatoma cells (124). On the other hand, neither InsP3R expression nor ER Ca²⁺ stores are affected by Mcl-1. Rather, Mcl-1 inhibits Ca²⁺ signaling directly within mitochondria (258). In contrast, the proapoptotic Bcl-2 family members Bax and Bak promote ER Ca²⁺ overload through the control of InsP3R phosphorylation (287), which leads to sensitization of mitochondria to calcium-mediated fluxes (284). These alterations serve as important upstream signals for cytochrome c release (283) (Fig. 2). Additionally, cytochrome c that has leaked from mitochondria via the PTP binds to the InsP3R, which facilitates release of toxic amounts of Ca²⁺ from the ER (36). This is thought to result in a positive feedback loop that causes further Ca²⁺ overload of mitochondria and then further leakage of cytochrome c (36). Inhibition of the interaction between cytochrome c and the InsP3R inhibits development of apoptosis by blocking this positive feedback loop (37). Therefore, Bcl-2 family members regulate their pro- or antiapoptotic functions at least in part through a range of complementary effects on Ca²⁺ signaling pathways.

Mitochondrial permeabilization and cell death can also be induced by dysregulation of redox transition, namely, the uncontrolled generation of reactive oxygen species (ROS) (216, 298). Oxidants increase [Ca²⁺]i load by stimulating IP3Rs in the ER, while inhibiting SERCA pumps and Ca²⁺ extrusion via plasma membrane channels (59). In turn, [Ca²⁺]m uptake stimulates oxidative metabolism, resulting in enhanced ATP production, and hence ROS generation as a byproduct of oxidative phosphorylation (320). Above a certain degree of [Ca²⁺]m and ROS rise, a progressive Ca²⁺ surge prompts a feed-forward loop that induces prolonged PTP opening and commits cells to death (298) by the combination of a set of events: IMM depolarization, which causes cessation of oxidative phosphorylation and ROS production; matrix swelling and cristae unfolding, and the ensuing breaches in the OMM with release of stored Ca²⁺ and of apoptogenic proteins (20, 80, 142, 319). In fact, Ca²⁺-induced production of ROS has been shown to promote cytochrome c release in rat liver mitochondria through mitochondrial permeability transition-dependent mechanisms (300). Furthermore, an analogue of the lipid peroxides formed during oxidative stress, tertbutyl-hydroperoxide, prompts NAD(P)H oxidation, increasing mitochondrial free Ca²⁺ and ROS, eventually leading to PTP opening and cell death (216).

It is important to note that when PTP induction is transient or limited to a fraction of mitochondria in a cell, a short burst of ROS production could occur. These ROS could serve as signaling components to propagate waves of short PTP openings in the surrounding mitochondria, producing additional ROS with possible functions in the regulation of chaperones, kinases, and gene expression. Hence, PTP could use Ca²⁺-dependent and independent ROS generation to tune many cellular routines unrelated to death induction (195).

Ca²⁺ Signals in Liver in Health and Disease

Liver growth: Regeneration versus hepatocellular carcinoma

The role of Ca²⁺ in cell proliferation assumes special importance in the process of liver regeneration. Liver regeneration after the loss of hepatic tissue is a fundamental parameter of the response to liver injury (255). It is induced by acute damage to the organ and is regulated in a well-orchestrated manner by cytokines, growth factors, hormones, and neurotransmitters that help promote massive proliferation of hepatocytes, which switch from a quiescent to a proliferative phenotype. This results in the restoration of liver mass and function within days after partial tissue loss (112, 377). A number of Ca²⁺ mobilizing hormones and growth factors are involved in this process of proliferative signaling. These include adenosine triphosphate (ATP) (138), arginine vasopressin (AVP) (281), noradrenaline (82), HGF (12, 260), EGF (260, 373), and insulin (331).

Alterations in components of the Ca²⁺ signaling machinery, namely, the different InsP3R isoforms and SERCA pumps, have been reported to occur during liver regeneration (95, 113, 232). Other reports have suggested that the sensitivity of liver cells to Ca²⁺ mobilizing agonists (169), as well as agonist-induced calcium signals are modified after partial hepatectomy (PH) in the rat (197). Another study suggested that these modifications reflect the remodeling of Ca²⁺ signals, which is thought to be essential for the initial regenerative response (280). Thus, Ca²⁺ signals are thought to adapt to the context of regulated proliferation induced by PH. Is there a fine regulation mechanism for proliferation during liver regeneration and does Ca²⁺ signal compartmentalization influence this process? Lagoudakis et al. investigated the role of [Ca²⁺]c during liver regeneration in the rat, targeting PV to this compartment (206). The study demonstrated in vivo, at the organ level that hepatocytes require unaltered cytosolic calcium signaling to progress through the cell cycle after PH, even though liver cell lines do not share this requirement, as discussed above (332). This may reflect the fact that primary hepatocytes are more sensitive than liver cell lines to Ca²⁺ buffering, since buffering of Ca²⁺ in the nucleus of primary hepatocytes in vivo leads to widespread apoptosis, liver failure, and death (unpublished observation). Alternatively, perhaps this is due to the fact that primary hepatocytes and liver cell lines do not express the same repertoire of intracellular calcium channels in the different cell compartments (215). Progression through the cell cycle from G₀ to G₁ and S phases is triggered presumably by Ca²⁺ mobilizing agonists that are known to be released in the early stages of liver regeneration after PH, such as HGF and EGF (260), ATP (138), AVP (281), and noradrenaline (82). Furthermore, [Ca²⁺]c signals may be important for other early events after PH, such as gene transcription and phosphorylation of ERK and CREB, which are crucial for hepatocyte progression after PH (341, 352, 372), and which are dependent on cytosolic and/or nuclear Ca²⁺ signaling (91, 99, 157, 205). Specific targets of [Ca²⁺]c signals during liver regeneration remain unknown.

A recent report has shown that [Ca²⁺]m signals also play an important role in the progression of liver regeneration after PH (145). Experiments in SkHep1 cells expressing PV in the mitochondria revealed that buffering of [Ca²⁺]m leads to decreased apoptosis through altered expression of proapoptotic and antiapoptotic proteins. In vivo studies showed that [Ca²⁺]m buffering accelerates liver regeneration after PH. Therefore, the regulation of apoptosis by [Ca²⁺]m signals is an additional mechanism by which hepatocyte proliferation and liver regeneration may be modulated. Future advances in this field should lead to a better understanding of how these various Ca²⁺ compartments act in an integrated manner to regulate liver regeneration.

Hepatocellular carcinoma (HCC) is the most common primary liver malignancy in adults and the third most common cause of cancer death worldwide (228). Its complex molecular pathogenesis is regulated by several signaling pathways (391, 392, 406). In particular, the role of growth factors, which act in part through [Ca²⁺]i release (28), has been highlighted in various studies and their signaling pathways are known targets for the discovery of new therapeutic approaches (391, 392, 406). In particular, HGF and its receptor c-met have been implicated in tumor invasion in HCC (247). The versatility of MET-mediated biological responses is modulated by subcellular compartmentalization of MET signaling pathways (389). In fact, c-met translocates to the nucleus upon HGF stimulation to generate InsP3-dependent Ca²⁺ signals there, and this process is thought to regulate cell proliferation (137). The invasive potential of HCC cells has also been linked to altered store-operated Ca²⁺ entry (148), through regulation by βig-h3, an extracellular matrix protein that is implicated in tumorigenesis and metastasis (374, 378). Selective buffering of nuclear but not [Ca²⁺]c also impairs growth of liver tumors in vivo (332) (Fig. 2). Additionally, [Ca²⁺]n signals mediate cell proliferation through the regulation of the asparaginyl endopeptidase LGMN (8). The observations that HGF stimulates increased expression of this protein, and expression also is increased in tumor cells compared to normal cells in liver specimens from patients with hepatocellular carcinona (HCC), suggests that LGMN’s positive effects in cell proliferation may promote carcinogenesis.

Two therapeutic agents, interferon-alpha (IFNα) and 5-fluorouracil (5-FU) are used experimentally in combination against HCC (264, 295, 345). A recent study revealed that [Ca²⁺]i might be a potent upstream proapoptotic messenger acting at the early stage of IFN-α/5-FU-induced apoptosis through the mitochondrial pathway (422). Furthermore, the novel squamosamide derivative and potent antioxidant N-(2-(4-hydroxy-phenyl)-ethyl]-2-(2,5-dimethoxy-phenyl)-3-(3-methoxy-4-hydroxy-phenyl)-acrylamide also known as FLZ, has been shown to have antitumor activity in HCC cells via modulation of the expression or activation of cell-cycle regulatory proteins, which are associated with decreased Ca²⁺/ROS levels. Therefore, FLZ represents a potential therapeutic agent for the treatment of HCC (315). Understanding the interplay between Ca²⁺ signaling and other signaling pathways, and how compartmentalized Ca²⁺ signals may control cell proliferation will reveal new clues of HCC pathogenesis, with the potential to lead to the discovery of more specific and effective therapeutic approaches.

Cholestasis

Bile secretion is one of the primary functions of the liver and defects in this process result in cholestasis. Chronic cholestatic conditions of the liver are among the most serious forms of liver disease. These conditions often result from diseases of the bile ducts (329, 388). Secretion in cholangiocytes can be mediated in part by Ca²⁺ (101, 164). More specifically, InsP3-mediated Ca²⁺ signals are important for normal bile secretion as discussed above, and loss of InsP3R is a common event in cholestasis (312). In particular, apical type III InsP3R expression is greatly decreased or absent in the bile ducts of patients with a range of cholestatic disorders including bile duct obstruction resulting from stone disease or malignancy, biliary atresia, primary biliary cirrhosis and sclerosing cholangitis (354). In accordance with these observations, animals subjected to common BDL or lipopolysaccharide (LPS) injection, which are accepted models of cholestasis (329, 354), had selective loss of InsP3R expression in biliary epithelia (354) (Fig. 4). As a result, Ca²⁺ signaling and Ca²⁺-mediated bicarbonate secretion is impaired. However, upstream components of the Ca²⁺ signaling pathway are preserved, which suggests that the selective decrease in InsPRs is the cause of defective Ca²⁺ signaling. Furthermore, this decrease in InsP3Rs is thought to interfere also with cAMP-mediated ductular secretion (218, 259). The loss of InsP3R from bile duct epithelia has been observed in other experimental animal models of cholestasis (5, 203, 217, 262, 360, 388). These findings may be explained by the fact that type III is the most abundant InsP3R isoform in cholangiocytes (163), and contribute to the formation of apical-to-basolateral Ca²⁺ waves (163, 166, 270, 409), which are critical for secretion in bile duct cells, due to their role in the transport of fluid and electrolytes across the apical membrane (116).

Figure 4.

Inositol 1,4,5-trisphosphate receptors (InsP3R) expression is lost in bile duct epithelia after bile duct ligation. Confocal immunofluorescence of liver sections from normal rats and rats subjected to bile duct ligation (BDL) labeled with isoform-specific InsP3R antibodies (green) and rhodamine phalloidin (red). Type 1 InsP3R labeling in normal bile duct cells (A) is found throughout each cell although is expressed at low levels, similar to that observed 2 weeks after BDL (B). Type 2 InsP3R labeling is seen throughout each cell in normal liver sections (C) and it is nearly absent 2 weeks after BDL (D). Type 3 InsP3R labeling is found predominantly in the apical region of bile duct cells in normal liver sections (E) and is also markedly reduced 2 weeks after BDL (F) [reprinted from reference (354), with permission].

Although cholestasis is often associated with defects in the bile ducts (329, 388), genetic or acquired disorders of hepatocyte transporters can also account for this condition (388). Recent studies have demonstrated that insertion of the transporter MRP2 in the plasma membrane is Ca²⁺ dependent and loss of type II InsP3R results in impaired canalicular organic anion secretion in hepatocytes (83) (Fig. 2). Moreover, loss of this isoform impairs the choleretic effect of tau-roursodeoxycholic acid (TUDCA) in cells treated with the cholestatic bile acid TLCA (83). Similarly, loss of type II InsP3R expression or loss of pericanalicular expression of the receptor leads to impaired BSEP-dependent secretion in hepatocytes (Fig. 2), and is observed in animal models of estrogen and endotoxin-induced cholestasis (202). Indeed, decreased type II InsP3R expression in the liver had been observed in an earlier study in animals subjected to BDL (103) and mistargeting of hepatocellular transporters is a common feature of cholestatic disorders (333). Being the most expressed isoform in this cell type, and localized preferentially to the canalicular region (267), the type II InsP3R regulates polarized Ca²⁺ waves (165), which in turn may regulate secretion in an analogous manner to what takes place is cholangiocytes (259, 354).

The mechanisms by which the decrease in InsP3R expression occurs remain unknown. Certain hepatic proteins can have decreased expression in BDL and other cholestatic conditions through the activation of nuclear hormone receptors like the farnesoid X receptor (FXR) (94). However, it is not known whether the promoter regions of any of the InsP3R isoforms contain response elements for FXR. Proinflammatory cytokines released during cholestasis are known to inhibit fluid secretion in cholangiocytes, so one could question their contribution to decreased expression of InsP3R. However, it has been shown that proinflammatory cytokines do not impair Ca²⁺-mediated secretion in these cells (360). Consistent with this observation, interleukin 1 increases rather than decreases transcription of the type I InsP3R gene (49). Perhaps increased degradation rather that decreased production could cause the observed loss of InsP3R expression. Indeed, InsP3R ubiquitination and proteasomal degradation are triggered upon secretagogue stimulation in pancreatic acinar cells (410).

These findings suggest that in canalicular cholestasis, decreased expression and localization of canalicular transporters results in part from loss of pericanalicular Ca²⁺ signaling. This has implications for the treatment of cholestatic disorders. UDCA, the only therapy of proven benefit for certain types of chronic cholestatic liver disease (296, 309), ameliorates liver function abnormalities in part by stimulating biliary exocytosis in a Ca²⁺-dependent fashion (30, 31). Moreover, these effects are partially mediated by the cooperative action of Ca²⁺-dependent conventional PKC alpha (cPKCα), which promotes transporter insertion into the canalicular membrane (29), and PKA (408), which phosphorylates the type II InsP3R potentiating InsP3-evoked Ca²⁺ release (52, 55). Furthermore, the cytoprotective effects of UDCA and its taurine conjugate TUDCA are associated with enhancement of Ca²⁺ signaling, inhibition of which blocks the activation of key survival pathways (244). In fact, UDCA stimulates secretion of the Ca²⁺-mobilizing agonist ATP from hepatocytes into bile, which may promote bile flow and act on downstream bile-duct epithelia by stimulating fluid and electrolyte secretion (268). Additionally, UDCA can locally stimulate cholangiocytes to release ATP and initiate P2Y receptor-mediated Ca²⁺-dependent fluid secretion (115). These findings are consistent with the hypothesis that stimulation of certain types of Ca²⁺ signaling in bile ducts and hepatocytes may provide a strategy for the treatment of cholestatic diseases. Taken together, loss of apical InsP3Rs in cholangiocytes and hepatocytes appear to be a common basis for impaired secretion and the development of cholestasis. These findings suggest a choleretic effect for Ca²⁺ through localized Ca²⁺ release in the pericanalicular region. In contrast, earlier studies had suggested a cholestatic effect, in which global Ca²⁺ signals induced by ionophores may promote cholestasis (273). Understanding the localized nature of these, Ca²⁺ signaling microdomains may provide insight into mechanisms of cholestasis and lead to the discovery of new alternatives to treat this condition.

The role of gap junctions in cholestasis has also been reported. Correa et al. investigated the involvement of Cx32, the major gap junction protein in the liver (111, 204), in the hepatic response to cholestasis (78). This study demonstrated that Cx32 deficiency impairs intercellular communication and the transmission of Ca²⁺ signals between hepatocytes after treatment with endotoxin. This finding is consistent with previous observations that Ca²⁺ waves and Ca²⁺ oscillations depend upon gap junctions in hepatocyte couplets (266), and that decrease in Cx32 protein levels after BDL is associated with impaired coordination of hormone-induced Ca²⁺ signals (111). Furthermore, Cx32 deficiency in mice leads to prolonged cholestasis when compared to wild type animals, suggesting that Cx32 expression is important for recovery from cholestasis after treatment with endotoxin (78). This report proposed an important role of gap junctions in the liver, to assure integrated and uniform tissue response in times of stress (78). Similarly, a subsequent study showed that cholestatic bile acids inhibit gap junction permeability and prevent the coordination of Ca²⁺ oscillations in hepatocytes and cholangiocytes (47). These findings suggest that this inhibitory effect might contribute to the cholestatic effect of these bile acids and may be an early event before the downregulation of Cx32 that has been observed previously in other models of cholestasis (111, 139). In contrast, opposing results were shown subsequently, in which TLCA-induced cholestasis was partially reversed by TUDCA in both wild type and Cx32-deficient mice, suggesting that gap junctional intercellular communication is neither essential for the cholestatic effect of TLCA nor needed for the choleretic effect of TUDCA (313). In aggregate, these observations are consistent with the idea that altered Ca²⁺ signaling is involved in the pathogenesis of cholestasis and can be modulated for therapeutic benefit, but careful examination of the response of patients to cholestasis may ultimately be necessary to establish a common basis for this disorder and the relevance of these observations for human disease.

Viral hepatitis

Viruses encode proteins and utilize host cellular proteins and signaling pathways to enhance viral replication, so regulating cellular Ca²⁺ signals is an ideal strategy for altering the intracellular environment to favor viral replication (66). Hepatitis B virus (HBV) is a para-retrovirus that affects more than 350 million people worldwide and chronic infection with HBV is the major cause for the development of HCC (158). The multifunctional HBV protein HBx is essential for replication (68, 425) and is thought to make significant contributions to the development of HBV-associated HCC (44). Elevation of [Ca²⁺]c signals as a result of HBx has been reported in several studies. HBx regulation of [Ca²⁺]c was shown to be essential for HBV replication in HepG2 cells and cultured primary rat hepatocytes (46, 130); calcium signals also promoted HBV capsid assembly in HepG2 cells (70). HBx elevation of [Ca²⁺]c signals is required for other activities of HBx, such as regulation of cell proliferation, activation of Pyk2/FAK-Src kinases, and activation of the transcription factor AP-1 (43, 45, 46, 131, 288). A very recent study added to these findings proposing that HBx stimulates HBV replication through increases in [Ca²⁺]m uptake, which promotes ER Ca²⁺ depletion and increased stored-operated Ca²⁺ entry, causing an increased plateau level of [Ca²⁺]c spikes (419). Indeed, continuing studies geared toward the discovery of the precise molecular basis of HBx regulation of Ca²⁺ signaling may help identify specific targets that regulate HBV replication and HCC development.

Hepatitis C Virus (HCV) also alters Ca²⁺ signals upon infection. HCV infection is a major cause of chronic hepatitis, liver cirrhosis, and HCC, affecting more than 170 million people worldwide (353). Several studies have suggested that HCV promotes alterations in mitochondrial oxidative metabolism (289, 376). Piccoli et al. provided insight into the mechanism by which this may occur, demonstrating that HCV protein expression results in Ca²⁺ overload in the mitochondrial compartment, which triggers alterations in the bioenergetic balance and nitro-oxidative stress (302). This study proposes that these effects are primed by an HCV protein pool localized on ER-associated mitochondrial membranes. In this model, HCV-induced ER stress, which has been documented previously (19, 375, 376), may provoke depletion of ER Ca²⁺, which in turn may be captured by the mitochondria to result in the described alterations (302). Cumulatively, these results suggest that restoring [Ca²⁺]m homeostasis might prevent or even reverse the effects of HCV.

ER stress

Effects of ER stress in the liver are important for the pathogenesis of a number of metabolic and inflammatory conditions. The ER represents a central regulator of protein folding, quality control, trafficking, and targeting. Therefore, the ability to adapt its capacity to manage synthetic, metabolic, and other unfavorable conditions is extremely important to maintain homeostasis in the cell. Disequilibrium between ER load and folding capacity establishes ER stress, and the ability to adapt to physiological levels of ER stress is important to all cells, but especially to professional secretory cells (156), like hepatocytes (Fig. 2). This adaptation is achieved by the activation of the unfolded protein response (UPR) which consists of the complex coordinated action of three distinct signal transduction pathways mediated by inositol requiring enzyme (IRE) 1α, protein kinase RNA-like ER kinase (PERK), and activating transcription factor 6α (ATF6α) (168, 236). Following activation, UPR signaling pathways act synergistically to encode functions that ameliorate the stressed state of the ER. If ER stress persists, the cell is functionally compromised and signal alarm pathways are activated to lead to apoptosis (168, 236, 335).

UPR activation has been reported in several physiological responses, including glucose and lipid metabolism, inflammation, autophagy, and apoptosis (168). The role of Ca²⁺ has been particularly noted in the process of ER stress-induced apoptosis. Disequilibrium in ER Ca²⁺ levels can activate the UPR, and constant or pronounced ER stress leads to apoptosis (168). ER stress-associated cell death results from the action of several apoptosis mediators. Some of these are activated by the UPR sensors, whereas others are related to Ca²⁺ and redox homeostasis (236). The proapoptotic proteins Bax and Bak as well as the antiapoptotic Bcl-2 that reside in the ER membrane and are known to regulate Ca²⁺ homeostasis, interact with members of the UPR cascades like IRE1α and c-jun-N-terminal kinase to mediate ER stress-induced apoptosis (162, 210, 401, 416). Furthermore, ER stress-inducing agents lead to sustained Ca²⁺ release from the ER, accumulation of Ca²⁺ in the mitochondria and subsequent mitochondrial permeabilization, resulting in release of apoptotic effectors from mitochondria into the cytosol (93). When the antiapoptotic protein Bcl-2 is over expressed, resting free ER Ca²⁺ and cell death are decreased (119). Moreover, deficiency of the proapoptotic proteins Bak and Bax results in lower free ER Ca²⁺ content, and reduced sensitivity to apoptotic agents (286). Furthermore, Bax Inhibitor-1 induces Ca²⁺ release under acidic conditions leading to increased cell death (193). In some cases during UPR activation, translation is resumed prematurely before ER stress is completely resolved. Consequently, newly synthesized proteins can undergo oxidative protein folding in the ER generating ROS (242). Additionally, protein folding in the ER can lead to activation of InsP3Rs and release of [Ca²⁺]i from the ER, which in turn is transmitted to the mitochondria to produce ROS (298). In both cases, the generation of ROS results in apoptosis. Thus, current evidence suggests that Ca²⁺ and Ca²⁺-sensitive proteins play a key role in the development of ER stress-induced apoptosis. However, the exact signaling pathways that mediate this process in hepatocytes and whether ER stress affects the Ca²⁺ signaling machinery and/or downstream functions of Ca²⁺ are subjects that need further investigation.

Sustained ER stress and activation of the UPR have also been observed in several liver diseases, and Ca²⁺ signaling-associated defects have been implicated, in particular, in the progression of obesity and the obesity-related disorder Nonalcoholic fatty liver disease (NAFLD). Indeed, incorporation of saturated fatty acids into the ER disrupt its structure and function (42, 226), affecting the folding capacity and chaperone function by altering ER Ca²⁺ homeostasis (54, 140, 149, 402). Several lines of evidence suggest that members of the Bcl-2 family of proteins, which directly alter ER Ca²⁺, may play a direct role in saturated fatty acid-induced hepatocyte cell death and the progression of NAFLD (2, 13, 65, 235, 291). Moreover, palmitic acid requires the flux of Ca²⁺ and cytochrome C release to mediate ER stress-induced apoptosis (423). Furthermore, a new link has been established between ER stress and obesity, in which protein and mRNA levels of SERCA2b are markedly reduced in the livers of obese mice and recovery of SERCA2b significantly ameliorates ER stress, increases glucose tolerance and achieves euglycemia in severely obese and diabetic mice (293) (Fig. 2). Subsequent studies demonstrated that alterations in the ER fatty acid and lipid composition result in the inhibition of SERCA activity and ER stress. Hence, abnormal lipid and Ca²⁺ metabolism are important contributors to hepatic ER stress in obesity (123).

Therapeutic agents that affect Ca²⁺ signaling are currently being used for conditions like spasticity, and other compounds have been tested for specific UPR proteins in the liver (199). However, further investigation is needed to better understand the molecular basis of ER stress-induced liver damage and the precise involvement of Ca²⁺ signaling in this process to develop targeted therapies for these conditions.

Conclusion

Ca²⁺ signaling in the liver results from the orchestrated action of multiple members of the Ca²⁺ signaling machinery in the different cells that form this highly specialized organ. Spatial and temporal Ca²⁺ signaling patterns should be considered not only at the subcellular and cellular level, but also in the context of the whole organ, in which cells maintain constant communication to achieve the proper performance of liver function. We furthermore are now beginning to understand how defects in Ca²⁺ signaling at these various levels play a role in the pathogenesis of a wide range of liver diseases. The understanding of how defects in the Ca²⁺ signaling machinery and in Ca²⁺ signaling patterns may lead to pathological conditions, may in turn result in the rational design of novel, more specific, and more effective therapeutic approaches.