Biology and pathobiology of gap junctional channels in hepatocytes

Authors

  • Mathieu Vinken,

    Corresponding author
    1. Department of Toxicology, Vrije Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium
    • Department of Toxicology, Vrije Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium
    Search for more papers by this author
    • fax: +32-2-4774582

    • Mathieu Vinken and Tamara Vanhaecke are postdoctoral research fellows of the Fund for Scientific Research Flanders (Fonds Voor Wetenschappelijk Onderzoek Vlaanderen), Belgium.

  • Tom Henkens,

    1. Department of Toxicology, Vrije Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium
    Search for more papers by this author
  • Evelien De Rop,

    1. Department of Toxicology, Vrije Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium
    Search for more papers by this author
  • Joanna Fraczek,

    1. Department of Toxicology, Vrije Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium
    Search for more papers by this author
  • Tamara Vanhaecke,

    1. Department of Toxicology, Vrije Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium
    Search for more papers by this author
    • Mathieu Vinken and Tamara Vanhaecke are postdoctoral research fellows of the Fund for Scientific Research Flanders (Fonds Voor Wetenschappelijk Onderzoek Vlaanderen), Belgium.

  • Vera Rogiers

    1. Department of Toxicology, Vrije Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium
    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

Abstract

The present review provides the state of the art of the current knowledge concerning gap junctional channels and their roles in liver functioning. In the first part, we summarize some relevant biochemical properties of hepatic gap junctional channels, including their structure and regulation. In the second part, we discuss the involvement of gap junctional channels in the occurrence of liver cell growth, liver cell differentiation, and liver cell death. We further exemplify their relevance in hepatic pathophysiology. Finally, a number of directions for future liver gap junctional channel research are proposed, and the up-regulation of gap junctional channel activity as a novel strategy in (liver) cancer therapy is illustrated. (HEPATOLOGY 2007.)

A major hallmark of multicellularity is the ability to communicate. In fact, the maintenance of homeostasis in multicellular organisms is governed by three major communicative networks, namely, extracellular, intracellular, and intercellular mechanisms. Extracellular signals (for example, hormones) trigger intracellular messengers (for example, signal transduction mediators), which, in turn, affect intercellular communication. Direct intercellular communication is mediated by gap junctional channels, which represent a particular type of cell junction that specializes in the direct exchange of essential cellular metabolites.1, 2 The liver was among the first organs in which gap junctional channels were studied. In 1967, Loewenstein and Kanno3 demonstrated that cancerous rat hepatocytes, as opposed to their healthy counterparts, are incapable of exchanging ions. Pioneering work of Revel and Karnovsky in the same year,4 using a negative staining protocol, showed the presence of hexagonal arrays of subunits in intercellular junctions in isolated liver plasma membranes. These structures were designated gap junctional channels and are also referred to as nexus or maculae communicantes. In 1974, Goodenough5 isolated two gap junctional channel proteins from mouse liver and named them connexins. The cloning of these first two connexins from rat liver by the mid 1980s6–9 was the actual start of 2 decades of extensive gap junctional channel research. Through the years, the liver has remained a popular reference organ in this research field, and as a result, a wealth of knowledge is available concerning gap junctional channels in liver. In this article, the biochemistry and (patho)physiology of hepatic gap junctional channels are reviewed. Special attention is paid to the involvement of gap junctional channels in the control of liver homeostasis and in the occurrence of liver disease and toxicity.

Biochemical Properties of Liver Gap Junctional Channels

Distribution and Formation.

Gap junctional channels are composed of two connexons of adjacent cells, which in turn are built up by six connexin proteins. At present, more than 20 connexin species have been cloned from humans and rodents. They all share a similar structure of four membrane-spanning domains, two extracellular loops, one cytoplasmic loop, one cytosolic N-terminal tail, and one C-terminal region (Fig. 1). Connexins are named after their molecular weight (expressed in kilodaltons), as predicted from complementary DNA cloning, and are expressed in a cell-specific way.2, 10 Liver vascular cells mainly express connexin 37 (Cx37) and Cx40,11 whereas most other nonparenchymal liver cells produce Cx439 (Table 1 and Fig. 2). With the exception of stellate cells,12 however, the presence of gap junctional channels between hepatic nonparenchymal cells still remains to be undeniably demonstrated. Hepatocytes, on the other hand, are loaded with gap junctional channels. Gap junctional channels occupy as much as 3% of the hepatocyte membrane surface13 and are composed of Cx32 and Cx26 (Table 1 and Fig. 2), which represent about 90% and 5%, respectively, of the total connexin amount in rat and human livers.14, 15 The Cx26:Cx32 protein ratio in rat liver is likely to be regulated by posttranscriptional mechanisms [for example, messenger RNA (mRNA) stability] because the Cx26 mRNA content is lower in comparison with Cx32 mRNA despite its faster transcription rate.16 Both connexins display different expression patterns, as Cx32 is uniformly distributed throughout the liver, whereas Cx26 is preferentially expressed in the periportal acinar area.13 In rat hepatocytes, Cx32 is also colocalized with other cell junction proteins, such as the tight junction building blocks occludin, claudin 1, zona occludens 1, and zona occludens 2.17

Figure 1.

Molecular architecture of gap junctional channels. Gap junctional channels are grouped in plaques at the membrane surface and are composed of 12 connexin proteins organized as two hexameric connexons. The connexin structure consists of four membrane-spanning domains (TM1-TM4), two extracellular loops (EL1 and EL2), one cytoplasmic loop (CL), one cytoplasmic amino tail (NT), and one cytoplasmic carboxy tail (CT).

Table 1. Connexins Expressed in Rodent and Human Livers
ConnexinLocalizationReference
  • See also Fig. 2.

  • Abbreviations: AEC, hepatic artery endothelial cell; BEC, biliary epithelial cell; GC, Glisson's capsule; HP, hepatocyte; KC, Kupffer cell; NS, not specified; PEC, portal vein endothelial cell; SC, stellate cell; and SEC, sinusoidal endothelial cell.

  • *

    Staining for this connexin is very weak in this cell type.

  • Cholangiocytes.

  • Cx43 is predominantly found in its nonphosphorylated variant in liver homogenates.

  • §

    Mesothelial cells.

  • Cx31.9 and Cx30.2 are expressed in human and murine livers, respectively.

Cx32HP, BEC,* SEC*6–8, 12, 77
Cx26HP, KC,* SC, SEC8, 9, 12
Cx43AEC, BEC, GC,§ KC, PEC, SC, SEC*11, 72, 77, 122–124
Cx40AEC, PEC11, 122
Cx37AEC, PEC11, 122, 125, 126
Cx39NS127
Cx31.9/Cx30.2NS128, 129
Figure 2.

Schematic overview of the most relevant connexin species expressed in rodent and human livers (see also Table 1).

Figure 3.

Involvement of gap junctional channels in the control of the hepatic life cycle. In the liver, hepatocytes directly communicate with one another through gap junctional channels. GJIC involves the direct exchange of essential messengers (for example, cAMP, ATP, IP3, and Ca2+), which in turn regulate key aspects of the hepatic life cycle, including cell growth and proliferation, cellular differentiation and functioning, and (apoptotic) cell death. Gap junctional channels are therefore considered goalkeepers of liver homeostasis.

Most connexins follow a classical secretory pathway during the biogenesis of gap junctional channels. Thus, connexins are synthesized by membrane-bound ribosomes and cotranslationally integrated into the endoplasmic reticulum, where they oligomerize to form a hexameric connexon. The latter can be homomeric, formed by only a single connexin species, or heteromeric, containing different types of connexins. The exact localization of oligomerization depends on the connexin species in question.18, 19 It is thought that both Cx26 and Cx32 assemble in the endoplasmic reticulum, whereas Cx43 rather oligomerizes in the trans-Golgi apparatus.20 Connexons are further transported in vesicles to the plasma membrane. Upon arrival at the cell membrane, connexins can either reside in nonjunctional regions or dock with an opposing connexon, thereby creating a narrower extracellular space (a 2- to 3-nm “gap”). The resulting gap junctional channel can be homotypic or heterotypic; that is, it can consist of homomeric or heteromeric connexons, respectively. Docking of connexons implies their peripheral insertion into so-called plaques (Fig. 1), which comprise punctuated regions of various sizes at the junctional membrane regions.18, 19 In liver, gap junctional channel plaques range from 0.2 to 1 μm in diameter21 and contain about 10 to more than 10,000 channels.22 The formation of functional gap junctional channels requires the presence of appropriate cell adhesion mediated by adherens junctions. Indeed, N-cadherin antibodies inhibit the formation of Cx43-containing gap junctional channels in Novikoff hepatoma cells.23

Connexin proteins usually display rapid turnover rates in comparison with other plasma membrane proteins. Both in vitro (primary cultured hepatocytes)24 and in vivo (regenerating liver),25 the half-lives of Cx26 and Cx32 have been found to be 2 and 3 hours, respectively, whereas the turnover times of other integral membrane proteins in primary hepatocyte cultures generally range from 17 to 100 hours.26 Upon degradation, gap junctional channels are internalized by one of the two opposing cells, and this results in the formation of so-called annular (gap) junctional channels. These structures are further degraded by both lysosomes and proteasomes, although the exact degradation pathway depends on the cell type and on the connexin species.18, 19 Degradation of Cx32-containing gap junctional channels in rat liver mainly occurs via the lysosomal pathway.27

Function and Regulation.

Gap junctional channels provide a pathway for communication between adjacent cells, about 180 Å in length and 15 Å in diameter.28 The flux of molecules through these channels is called gap junctional channel intercellular communication (GJIC). GJIC includes the passive diffusion of small (<1 kDa) and hydrophilic molecules, such as glucose, glutamate, glutathione, cyclic adenosine monophosphate (cAMP), adenosine trisphosphate (ATP), inositol trisphosphate (IP3), and ions (for example, Ca2+), whereas large polypeptides, nucleic acids, and cellular organelles are excluded.29 The biophysical properties of a gap junctional channel depend on the nature of the connexin species that form the channel. Homotypic Cx26 gap junctional channels favor cation transfer, whereas homotypic Cx32 gap junctional channels promote anion passage.30 Likewise, ATP passes more than 300 times better through gap junctional channels formed by Cx43 compared with Cx32-based gap junctional channels.31

GJIC is regulated at several levels ranging from connexin gene transcription to gap junctional channel degradation. Long-term control of GJIC mainly concerns regulation at the transcriptional level of connexin expression.32 The genomic structure of most connexin genes is rather simple, that is, a first exon (E1) containing the 5′-untranslated region (UTR), which is separated by an intron of varying length from a second exon (E2) bearing the complete coding sequence, and the 3′-UTR. At least two exceptions to this general connexin gene structure have been reported, namely, the interruption of the coding sequence by introns and the differential splicing of the 5′-UTR.32, 33 The latter has been most exemplified for the Cx32 gene. In humans and rats, the Cx32 gene consists of three exons (E1, E1B, and E2), whereas its mouse and bovine counterparts contain four exons (E1, E1A, E1B, and E2).34–36 Transcription is initiated from at least two tissue-specific promoters. It can be generalized that the P1 promoter, located upstream of E1, is active in liver (hepatocytes) and pancreas (secretory acinar cells), yielding the E1-E2 transcript. The P2 promoter, located upstream of E1B, is active in peripheral nerves, yielding the E1B-E2 transcript.35, 37 In mice and cows, a third transcript (E1A-E2) is found in embryonic cells, oocytes, and adult liver.34, 35 Differential promoter usage and alternative splicing have also been reported for the mouse and rat Cx43 genes.38 Furthermore, connexin gene promoters contain several binding sites for ubiquitous transcription factors, such as specificity protein 1 (Cx26, Cx32, and Cx43), yin yang 1 (Cx26 and Cx32), and activator protein 1 (Cx43).32 In addition, a number of cell type–specific transcription factors control connexin gene expression. Liver-specific expression of Cx32, for instance, has been found to depend on the binding of hepatocyte nuclear factor 1α at the P1 promoter.39, 40 Epigenetic mechanisms also seem to be critical determinants of (hepatic) connexin gene expression.32 In this respect, our group showed that trichostatin A, a prototypic histone deacetylase inhibitor, enhances GJIC in primary hepatocyte cultures, which is associated with differential effects on the expression of Cx26, Cx32, and Cx43.41 In addition, Piechocki and coworkers42 found that the absence of Cx43 and Cx32 expression in rat MH1C1 hepatoma cells and rat WB-F344 liver epithelial cells, respectively, is associated with DNA methylation in the corresponding gene promoters.

Regulation of GJIC at the posttranslational level, so-called gating, is controlled by a number of factors, including transmembrane voltage (voltage gating) and Ca2+ and H+ ions (chemical gating).43 Among all gating mechanisms, connexin phosphorylation has been most extensively studied. With some exceptions (for example, Cx26), most connexins are phosphoproteins. Phosphorylation has been implicated in the regulation of a broad variety of connexin processes, such as the trafficking and degradation as well as gating of gap junctional channels.18, 44 The outcome of the phosphorylation event depends on both the connexin species and the kinase involved. Cx32 can be phosphorylated in vitro by protein kinase A, protein kinase C, Ca2+-calmodulin–dependent protein kinase II, and the epidermal growth factor receptor.45 In primary cultured hepatocytes, protein kinase A–mediated phosphorylation of Cx32, induced by cAMP analogues, results in enhanced GJIC.46 Phosphorylation of rodent Cx32 by protein kinase C, on the other hand, prevents proteolysis by calpains.47 Cx32 also displays intrinsic kinase activity in rat liver, which suggests that this connexin can control its own phosphorylation status.48 Within the research field of connexin phosphorylation, however, most attention has been paid to Cx43. This particular connexin is a substrate for several kinases, including protein kinases A and C, members of the mitogen-activated protein kinase (MAPK) family, casein kinase 1, the cyclin-dependent kinase (Cdk) 1/cyclin B complex, and v-src. A number of phosphorylated Cx43 variants have been characterized, some of which give rise to typical shifts in electrophoretic mobility as seen during sodium dodecyl sulfate–polyacrylamide gel electrophoresis analysis. In contrast to the wealth of knowledge that has been gained concerning Cx43 phosphorylation, little is known about its dephosphorylation. Nevertheless, inhibitors of protein phosphatases are known to affect GJIC, and this suggests that these enzymes are involved in the control of the connexin phosphorylation/dephosphorylation equilibrium.44, 45

Abbreviations

AEC, hepatic artery endothelial cell; ATP, adenosine trisphosphate; BEC, biliary epithelial cell; cAMP, cyclic adenosine monophosphate; Cdk, cyclin-dependent kinase; Cx, connexin; CYP, cytochrome P450; GC, Glisson's capsule; GJIC, gap junctional channel intercellular communication; HCC, hepatocellular carcinoma; HP, hepatocyte; IP3, inositol trisphosphate; KC, Kupffer cell; MAPK, mitogen-activated protein kinase; mRNA, messenger RNA; PEC, portal vein endothelial cell; SC, stellate cell; SEC, sinusoidal endothelial cell; UTR, untranslated region.

Gap Junctional Channels and the Hepatic Life Cycle

Liver Cell Growth.

In normal conditions, the adult liver displays very low proliferative activity. Upon partial hepatectomy, however, the remaining intact hepatic lobes start to grow, and the original size becomes restored within 7 to 10 days.49 Although some conflicting results have been published, most investigators reported transiently increased GJIC activity in the G1 phase, followed by a dramatic decrease upon initiation of the S phase of the hepatocyte cell cycle25, 50–59 (Table 2). Similar alterations were observed at the level of Cx32 expression and, in many cases, in Cx26 levels, whereas Cx43 remained unchanged25, 54, 58 (Table 2). It has further been shown that the reduced expression of both Cx26 and Cx32 results from decreased mRNA stabilities of the corresponding transcripts.54 Similar findings are produced when an in vitro model of hepatocyte proliferation, namely, mitogen-stimulated primary hepatocytes, is used.51, 60, 61 In this system, MAPK directly phosphorylates Cx32, and this results in its decreased expression.61 In fact, connexin phosphorylation could represent a major mechanism responsible for GJIC alterations during liver cell cycling. In serum-stimulated rat liver clone 9 cells, progression from the G0 state to the S phase is also related to protein kinase C–dependent phosphorylation of Cx43 and disruption of GJIC.62

Table 2. Alterations in Connexin Expression During Liver Cell Proliferation In Vivo and In Vitro
ModelLevelObservation*Reference
  • *

    Modifications observed during hepatocyte DNA synthesis. For a detailed overview of changes occurring during separate stages of the cell cycle, the reader is referred to the original publications.

  • Most hepatocytes replicate at least once or twice following partial hepatectomy. A first peak of hepatocyte DNA synthesis is observed around 24 hours post-hepatectomy, and a second (smaller) one is seen between 36 hours and 48 hours post-hepatectomy, corresponding to DNA synthesis of nonparenchymal cells. Hepatocyte mitosis occurs 6 to 8 hours after DNA synthesis.49

  • The connexin gene transcription rate remains unchanged.54

  • §

    The connexin mRNA half-life strongly decreases.54

  • p38 MAPK mediates the decrease in Cx32 expression.53

  • Upon isolation, hepatocytes re-enter the cell cycle in the G1 phase and progress toward a restriction point (corresponding to 40–44 hours post-plating), which they can pass (that is, enter the S phase) only after growth factor stimulation.10

Regenerating rat liverGJICDecreased50, 55
 Gap junctional channel numberDecreased51, 55, 57, 59
 Connexin expression  
  Cx43Unchanged54
  Cx32§Decreased25, 50, 53, 54, 56–58
  Cx26§Unchanged25
  Decreased54, 58
Mitogen-stimulated primary rat hepatocytesGJICDecreased51
 Gap junctional channel numberDecreased60, 61
 Connexin expression  
  Cx43Increased52
  Cx32Decreased51, 52, 60, 61
  Cx26Decreased51, 60, 61

The relevance of altered GJIC during cell cycling remains elusive. In the regenerating liver of rats treated with a p38 MAPK inhibitor, the disappearance of Cx32 is inhibited without hepatocyte proliferative activity being affected.53 This indicates that down-regulation of gap junctional channels can occur independently of cellular proliferation and, consequently, may be considered a minor part of the growth response. On the other hand, in the regenerating liver of Cx32 knockout mice, the G0/S transition of the cell cycle and thus the proliferative activity of the hepatocytes are not promoted, but the extent of synchronous initiation and termination of DNA synthesis is decreased.58, 63 In this view, reduction of GJIC does not provide a direct signal for cells to divide but rather permits cell cycle progression upon mitogenic stimulation. Here, gap junctional channel activity seems to be coordinated with cell growth and serves a purpose other than triggering proliferation. Such a purpose may include the functional segregation of the metabolic pools in dividing cells from their quiescent neighbors in order to avoid homeostatic imbalance.50, 51, 64 Others strongly believe that gap junctional channels fulfill a determinate function in cell proliferation control, rather than merely an assisting role in growth progression. Gap junctional channels indeed provide a pathway for the direct exchange of essential growth mediators, such cAMP (Fig. 3).65 Curiously, interfering with connexin gene expression often reveals additional mechanisms involved in gap junctional channel–mediated control of cell proliferation (Table 3). Indeed, transfection of liver-derived cell lines with connexin genes can directly alter gene expression patterns. For instance, forced expression of Cx32 and Cx26 in rat WB-aB1 liver epithelial cells and human HepG2 hepatoma cells triggers the production of p27 and E-cadherin, respectively, which, in turn, negatively affect cell proliferation.66, 67

Table 3. Effects of Interfering with Connexin Gene Expression on Liver Cell Growth and Live Cell Death
ConnexinModelManipulation*ObservationMechanismReference
  • *

    Genetic manipulation of connexin expression.

Cx43Rat WB-F344 liver epithelial cellsExpression ↑Growth ↓ 130
Cx32Human PLC/PRF/5 hepatoma cellsExpression ↑Growth ↓ 86
 Mouse CHST8 hepatocytesExpression ↑Growth ↓ 131
 Rat WB-F344 liver epithelial cellsExpression ↑Growth ↓ 132
 Rat WB-aB1 liver epithelial cellsExpression ↑Growth ↓Cell cycle arrest at G1/S transition66
    No effect on G2/M transition 
    Reduced cyclin D1 expression 
    Induced p27 expression 
 Mouse primary hepatocytesExpression ↓Growth ↑Enhanced G1/S transition133
 Mouse liver tissueExpression ↓Growth ↑ 104
Cx26Human PLC/PRF/5 hepatoma cellsExpression ↑Growth ↓/death ↑ 86
 Human HepG2 hepatoma cellsExpression ↑Growth ↓Induced E-cadherin expression67, 134
    No effect on β-catenin expression 
    Induced E-cadherin/β-catenin interaction 
    Reduced matrix metalloproteinase-9 activity 

Liver Cell Differentiation.

Using a 2-acetylaminofluorene/partial hepatectomy model, several groups have shown that connexin expression is modulated during differentiation of early rat hepatic progenitor cells into adult liver parenchymal cells. Oval cells were indeed repeatedly found to switch from Cx43 to Cx26 expression and, in particular, to Cx32 expression upon differentiation into hepatocytes, both in vivo (2-acetylaminofluorene/partial hepatectomy model)15, 68, 69 and in vitro (liver epithelial cell line models).69, 70 Alterations in connexin expression are also seen during liver ontogenesis. In rat liver, Cx26 and Cx32 become detectable in the late stage of gestation and reach maximal levels about 1 week after birth. At this time, the adult patterns of connexin distribution are established; that is, Cx32 is uniformly distributed along the hepatic acinus, whereas Cx26 becomes preferentially located in the periportal area.71 This process coincides with the establishment of the glucagon receptor zonation pattern. The latter is mainly detected in the perivenous region, whereas the inverse holds for its ligand.72 On the other hand, glucagon was found to enhance gene transcription of Cx26 and, to a lesser extent, that of Cx32.73 On the basis of these findings, Cx26 zonation is believed to be generated at the transcriptional level, and glucagon is likely to play a major role in this process.71, 73

In adult liver, gap junctional channels fulfill a pivotal function in the maintenance of the differentiated phenotype of hepatocytes. The establishment of GJIC between hepatocytes is indispensable for the performance of their functionality (Fig. 3). Several liver-specific processes are known to rely on GJIC, including albumin secretion,74 ammonia detoxification,74 glycogenolysis,75, 76 bile secretion,77–79 and xenobiotic phase I biotransformation.80–83 With respect to the latter, it was found that both the constitutive and drug-induced expression of cytochrome P450 (CYP) isoenzymes (that is, CYP3A4 and CYP2B6) in primary human hepatocyte cultures requires the presence of Cx32-containing gap junctional channels.80 Administration of β-naphthoflavone, phenobarbital, or clofibrate to rats resulted in a decrease in perivenous Cx32 expression. This was associated with the induction of CYP1A1/2 (β-naphthoflavone) or CYP2B1/2 (phenobarbital and clofibrate), which are both preferentially localized in the perivenous area.81–83 The relevance of this interrelationship, however, is not clear. Hypothetically, colocalized alterations in Cx32 and CYP expression are important for the effective biotransformation of xenobiotics by restricting the cytoplasmic diffusion of reactive intermediates.81

A handful of reports have described the mechanistic basis of the involvement of GJIC in glycogenolysis. This process, involving the enzymatic degradation of glycogen to glucose, is activated by both hormonal and nervous stimuli and mainly occurs in the periportal acinar region. Perivenous hepatocytes also show glycogenolytic activity, albeit to a lesser extent in comparison with their periportal counterparts.76, 84 Gap junctional channels play a key role in the propagation of the glycogenolytic response from the periportal to perivenous regions. More specifically, gap junctional channels control the intercellular passage of IP3, which triggers Ca2+ release from the endoplasmic reticulum, in turn causing Ca2+ waves along the acinar tract.84, 85 It has indeed been shown that Cx32 knockout mice exhibit decreased levels of glucose release into blood upon glycogenolytic stimulation.75, 76 Bile secretion also relies on GJIC-dependent Ca2+ signaling. Bile flow consists of both canalicular secretion from hepatocytes and ductular secretion from cholangiocytes. The latter mainly express Cx43, and as mentioned for hepatocytes, the propagation of Ca2+ waves between these cells controls their secretory activity.77, 78

Liver Cell Death.

In recent years, it has become clear that gap junctional channels are involved in programmed cell death or apoptosis. The research field is still in its infancy, and as a result, specific data in relation to liver are yet very scarce (Table 4). In human PLC/PRF/5 hepatoma cells, apoptotic cell death was accelerated upon forced expression of Cx26.86 On the other hand, during exogenously induced apoptosis in human Hep3B hepatoma cells and in rat WB-F344 liver epithelial cells, a decline in GJIC activity was observed. This was associated with cytoplasmic redistribution of Cx43 without alterations in its expression.87 A more detailed study was carried out by Wilson and coworkers.88 They demonstrated that GJIC is induced in the early phases of apoptosis in serum-deprived rat WB-F344 liver epithelial cells and coincides with increased Cx43 expression and phosphorylation. The latter might be mediated by the Cdk 1/cyclin B complex, which also controls the G2/M transition of the cell cycle. Upon further progression of cell death, GJIC activity declines, as evidenced by the absence of communication between apoptotic bodies. These modifications in GJIC during apoptotic cell death in serum-deprived rat WB-F344 liver epithelial cells are not fully understood. The transient induction of GJIC in the early phases of apoptosis could point to a role for gap junctional channels in the initial spread of a death wave from cell to cell. In this context, Ca2+ ions are thought to be the killing messengers. The onset of apoptosis is generally associated with drastic alterations in the concentration of Ca2+, an ion that is intercellularly exchanged via gap junctional channels. The subsequent reduction in GJIC activity may possibly serve to reduce the flux of toxic metabolites (for example, nitric oxide and superoxide ions) and thus to protect a healthy cell from its dying neighbor.89, 90

Table 4. Alterations in Connexin Expression During Liver Cell Death In Vitro
ModelCell Death InductionObservationReference
Rat WB-F344 liver epithelial cellsSerum deprivationInduced GJIC and Cx43 expression in early cell death phase88
  Reduced GJIC during cell death progression 
 Choline depletionReduced GJIC87
  No effect on Cx43 expression 
  Cytoplasmic Cx43 redistribution 
 N-acetylsphingosineNo effect on GJIC and Cx43 phosphorylation135
 LA-12Reduced GJIC136
  Induction of Cx43 phosphorylation 
  Cytoplasmic Cx43 redistribution 
Human Hep3B hepatoma cellsCholine depletionReduced GJIC87
  No effect on Cx43 expression 
  Cytoplasmic Cx43 redistribution 

Gap Junctional Channels and Liver Dysfunction

Liver Toxicity and Chronic Liver Disease.

Because of the key roles of gap junctional channels in the control of hepatic homeostasis, it is not surprising that these structures are frequently affected upon impairment of this critical balance, as occurs during liver toxicity.64 The liver (and the hepatocyte in particular) is a primary target of systemic toxicity, and this is mainly due to its unique localization and function in the organism.10 Hepatic injury, induced by xenobiotics such as carbon tetrachloride, is usually associated with down-regulation of connexin expression.64, 91 The exact role of connexins in hepatotoxicity is, however, not clear. Cx32 dominant-negative mutant transgenic rats were found to be more resistant to chemically induced hepatic injury. This suggests that gap junctional channels could spread damage signals from cell to cell, rather than playing a cytoprotective role.92

Upon progression of liver disease (hepatitis, fibrosis, and cirrhosis), Cx32 protein levels gradually decrease, and it becomes located in the cytoplasm of the hepatocytes.93 This down-regulation is also observed at the (post)transcriptional level and is caused by increased degradation of Cx32 mRNA.94 Although a similar decrease in protein content is seen for Cx26, its mRNA levels are not affected and even increase during liver inflammation.95 Furthermore, alterations in connexin expression upon disturbance of liver homeostasis are not restricted to hepatocytes. In the injured liver and during culture, hepatic stellate cells transform into myofibroblast-like cells that exhibit increased proliferative activity and high production of extracellular matrix components. During this activation of hepatic stellate cells, Cx43 and Cx26 expression is induced, with the latter being preferentially found near the nucleus.12

Liver Cancer.

Primary liver cancer is the fifth most common cancer worldwide and the third most common cause of cancer mortality. Hepatocellular carcinoma (HCC) accounts for as many as 90% of primary liver cancers, and in most cases, HCC occurs within an established background of chronic liver disease. A number of relevant intracellular signaling transduction cascades, including phosphoinositide 3-kinase/protein kinase B signaling and the Wnt/β-catenin pathway, were found to be dysregulated in HCC.96 With respect to intercellular communication, it has been well documented that tumor cells such as HCC generally display reduced GJIC activity. Numerous mechanisms underlie the loss of GJIC in carcinogenesis. The occurrence of mutations in connexin genes, however, is a rare event. Rather, epigenetic modifications, such as DNA hypermethylation in the gene promoter region, can trigger silencing of connexin production.97, 98 Indeed, reduction of Cx26 expression in HCC was recently found to correlate with the methylation status of its gene promoter.99 Inappropriate phosphorylation is also frequently observed in tumor cells.97, 98 In this light, several well-known tumor-promoting agents, such as phorbol esters, interfere with the connexin phosphorylation status, generally resulting in down-regulation of GJIC.1, 65 Aberrant localization of connexin proteins is another regularly seen feature of tumor cells.97, 98 Thus, most Cx32 is found within the cytoplasm of cancerous hepatocytes.100 Curiously, down-regulated connexins in cancer are often replaced by other connexin species, whether or not physiologically present in the tissue in question. In HCC, for instance, Cx32 expression is strongly reduced, whereas Cx43 becomes detectable. The latter is preferentially present in the cytoplasm of hepatocytes in its nonphosphorylated form.101, 102 The induced expression of Cx43, reflecting a recapitulation of liver development, is not fully understood. In a recent study, it was found that Cx43-silenced human HuH7 hepatoma cells display less proliferative activity and a higher degree of differentiation, whereas the inverse observations were seen in their Cx43-overexpressing counterparts. Moreover, both Cx32 expression levels and GJIC negatively correlated with Cx43 production. Cx43 might therefore directly be responsible for the malignancy of hepatoma cells.103 By contrast, Cx32 can be considered a liver tumor suppressor gene, as Cx32 knockout rodents display increased susceptibility to chemically induced hepatocarcinogenesis.63, 104–106

Conclusions and Perspectives

The liver is a unique organ endowed with a plethora of specialized functions. The establishment of communicative networks between the different hepatic cells is a key feature in this process. Nonparenchymal liver cells are preferentially in paracrine or juxtactrine contact with themselves and other hepatic cells.107 Hepatocytes, the true workhorses of the liver, directly communicate with one another through gap junctional channels. These structures are composed of connexin proteins, of which at least seven family members are expressed in hepatic tissue (Table 1). More recently, a new type of gap junctional channel–forming protein has been discovered, namely, the pannexins.108 Although pannexins 1 and 2 have been identified in rat liver,109 it still remains to be investigated to what extent these proteins contribute to communication between hepatic cells and thus to liver functioning. To date, a number of prominent liver-specific functions have been found to be governed by connexin-based gap junctional channels. More specifically, hepatic expression of Cx32 positively correlates with CYP-mediated xenobiotic biotransformation,80–83 albumin secretion,74 ammonia detoxification,74 and glycogenolysis.75, 76 The contribution of Cx26 and other connexin species to these functions has been less documented thus far. Several novel technologies, such as RNA interference, now are available that will clarify these issues. These tools could also provide more insight into the involvement of gap junctional channels in other facets of the hepatocellular life cycle. In this respect, the exploration of the link between gap junctional channels and cell growth and, in particular, cell death at the molecular level has been initiated only in the last few years. Nevertheless, some striking findings have already been reported from these new research fields. Such studies frequently demonstrate very specific functions of individual connexin species as well as GJIC-independent actions of connexin proteins. Thus, connexins as such can directly trigger gene expression of proteins that are related to cell growth and cell death regulation. Although controversial at present, connexins could also provide a pathway for the extracellular release of essential homeostasis regulators, such as ATP.2, 110–112 Extensive research is required to shed more light onto the functional relevance of these GJIC-independent activities of connexins in liver. The knowledge that will be gained from such experiments is not only of key importance to (liver) molecular biologists but is also of great interest to clinical scientists. Liver pathologies, including hepatitis, liver fibrosis, cirrhosis, and cancer, are usually accompanied by dysregulation of connexins and their channels,93 and this renders these structures potential targets for clinical therapy. This has been most exemplified for hepatocarcinogenesis, which is associated with the abrogation of GJIC.98 In fact, restoration of GJIC by targeting connexin expression is a very attractive anticancer strategy. Several natural and synthetic compounds, such as glucocorticoids, retinoids, carotenoids, and flavonoids, are known to up-regulate connexin production in liver cancer cells.1, 113 A particular group of connexin inducers are the epigenetic modulators of gene expression. In a number of tumor cells, 5-aza-2′-deoxycytidine (decitabine), an inhibitor of DNA methyltransferase enzymes, was found to induce connexin re-expression.114–118 Likewise, inhibitors of histone deacetylase enzymes modulate connexin expression in a variety of cell lines. For instance, suberoylanilide hydroxamic acid strongly enhances connexin expression and GJIC in ras-transformed rat WB-F344 liver epithelial cells.119 An alternative and more fundamental means of regaining GJIC includes connexin gene therapy. This method not only results in the direct restoration of GJIC but also takes advantage of the tumor-suppressing actions of connexin genes as such.97, 120 Overexpression of connexins in liver cancer cells indeed strongly counteracts the tumorigenic phenotype by inhibiting cell growth and promoting cell death (Table 3). In addition, exogenous enhancement of GJIC can potentiate the efficiency of other anticancer therapies. In this light, combining connexin gene transfection or chemical induction of connexin expression with the herpes simplex virus thymidine kinase/ganciclovir strategy seems extremely promising.113, 121 Further exploration of these innovative anticancer strategies in the upcoming years will undoubtedly open new perspectives for (liver) site-directed clinical therapy.

Ancillary