Knockdown of hepatocyte aquaporin-8 by RNA interference induces defective bile canalicular water transport

Aquaporin-8 (AQP8) water channels, which are expressed in rat hepatocyte bile canalicular membranes, are involved in water transport during bile formation. Nevertheless, there is no conclusive evidence that AQP8 mediates water secretion into the bile canaliculus. In this study, we directly evaluated whether AQP8 gene silencing by RNA interference inhibits canalicular water secretion in the human hepatocyte-derived cell line, HepG2. By RT-PCR and immunoblotting we found that HepG2 cells express AQP8 and by confocal immunofluorescence microscopy that it is localized intracellularly and on the canalicular membrane, as described in rat hepatocytes. We also verified the expression of AQP8 in normal human liver. Forty-eight hours after transfection of HepG2 cells with RNA duplexes targeting two different regions of human AQP8 molecule, the levels of AQP8 protein specifically decreased by 60-70%. We found that AQP8 knockdown cells showed a significant decline in the canalicular volume of approximately 70% (P < 0.01), suggesting an impairment in the basal (nonstimulated) canalicular water movement. We also found that the decreased AQP8 expression inhibited the canalicular water transport in response either to an inward osmotic gradient (-65%, P < 0.05) or to the bile secretory agonist dibutyryl cAMP (-80%, P < 0.05). Our data suggest that AQP8 plays a major role in water transport across canalicular membrane of HepG2 cells and support the notion that defective expression of AQP8 causes bile secretory dysfunction in human hepatocytes.     

The water channel aquaporin-8 is mainly intracellular in rat hepatocytes, and its plasma membrane insertion is stimulated by cyclic AMP

We previously found that water transport across hepatocyte plasma membranes occurs mainly via a non-channel mediated pathway. Recently, it has been reported that mRNA for the water channel, aquaporin-8 (AQP8), is present in hepatocytes. To further explore this issue, we studied protein expression, subcellular localization, and regulation of AQP8 in rat hepatocytes. By subcellular fractionation and immunoblot analysis, we detected an N-glycosylated band of approximately 34 kDa corresponding to AQP8 in hepatocyte plasma and intracellular microsomal membranes. Confocal immunofluorescence microscopy for AQP8 in cultured hepatocytes showed a predominant intracellular vesicular localization. Dibutyryl cAMP (Bt(2)cAMP) stimulated the redistribution of AQP8 to plasma membranes. Bt(2)cAMP also significantly increased hepatocyte membrane water permeability, an effect that was prevented by the water channel blocker dimethyl sulfoxide. The microtubule blocker colchicine but not its inactive analog lumicolchicine inhibited the Bt(2)cAMP effect on both AQP8 redistribution to cell surface and hepatocyte membrane water permeability. Our data suggest that in rat hepatocytes AQP8 is localized largely in intracellular vesicles and can be redistributed to plasma membranes via a microtubule-depending, cAMP-stimulated mechanism. These studies also suggest that aquaporins contribute to water transport in cAMP-stimulated hepatocytes, a process that could be relevant to regulated hepatocyte bile secretion.     

Water transporting properties of hepatocyte basolateral and canalicular plasma membrane domains

Previous work from our laboratory supports an important role for aquaporins (AQPs), a family of water channel proteins, in bile secretion by hepatocytes. To further define the pathways and molecular mechanisms for water movement across hepatocytes, we directly assessed osmotic water permeability (Pf) and activation energy (Ea) in highly purified, rat hepatocytes basolateral membrane vesicles (BLMV) and canalicular membrane (CMV) vesicles by measuring scattered light intensity using stopped-flow spectrophotometry. The time course of scattered light for BLMV and CMV fit well to a single-exponential function. In BLMV, Pf was 108 +/- 4 mum.s-1 (25 degrees C) with an Ea of 7.7 kcal/mol; in CMV, Pf was 86 +/- 5 mum.s-1 (25 degrees C) with an Ea of 8.0 kcal/mol. The AQP blocker, dimethyl sulfoxide, significantly inhibited the Pf of both basolateral (81 +/- 4 mum.s-1; -25%) and canalicular (59 +/- 4 mum.s-1; -30%) membrane vesicles. When CMV were isolated from hepatocytes treated with dibutyryl cAMP, a double-exponential fit was needed, implying two functionally different vesicle populations; one population had Pf and Ea values similar to those of CMV from untreated hepatocytes, but the other population had a very high Pf (655 +/- 135 mum.s-1, 25 degrees C) and very low Ea (2.8 kcal/mol). Dimethyl sulfoxide completely inhibited the high Pf value in this second vesicle population. In contrast, Pf and Ea of BLMV were unaltered by cAMP treatment of hepatocytes. Our results are consistent with the presence of both lipid- and AQP-mediated pathways for basolateral and canalicular water movement across the hepatocyte plasma membrane barrier. Our data also suggest that the hepatocyte canalicular membrane domain is rate-limiting for transcellular water transport and that this domain becomes more permeable to water when hepatocytes are exposed to a choleretic agonist, presumably by insertion of AQP molecules. These data suggest a molecular mechanism for the efficient coupling of osmotically active solutes and water transport during canalicular bile formation.     

Establishment of plasma membrane domains in hepatocytes. I. Characterization and localization to the bile canaliculus of three antigens externally oriented in the plasma membrane

A membrane fraction denoted N2 upper was isolated from homogenates of rat liver by sucrose gradient centrifugation. This fraction, which was enriched 65-fold over the homogenate in 5'-nucleotidase activity, was used as an immunogen in goats. The antisera obtained contained antibodies to three predominant polypeptides in the N2 upper membrane fraction, as shown by crossed immunoelectrophoresis. These polypeptides had molecular weights of 105,000, 110,000, and 160,000 after recovery from the crossed immunoelectrophoretic gels and are denoted PM105, PM110, and PM160. Each was a distinct polypeptide, as shown by the distinct peptide patterns resulting from limited proteolysis in the presence of detergents. The three polypeptides were synthesized by primary cultures of hepatocytes and were externally oriented at the surface of these cells, as shown by their accessibility in situ to iodination catalyzed by lactoperoxidase. They were not detectable in the serum by crossed immunoelectrophoresis. The three antigens were present at very low (PM110) or nondetectable (PM105, PM160) concentrations in intracellular membrane fractions derived from the Golgi and smooth and rough endoplasmic reticulum of liver. The antigens also were reduced in concentration in a plasma membrane fraction most likely derived from the sinusoidal surface of the hepatocyte. The three membrane antigens bind to concanavalin A; hence, they are probably glycoprotein constituents of a discrete domain of the hepatocyte plasma membrane. Immune complexes were isolated after crossed immunoelectrophoresis and injected into rabbits. Each of the antisera obtained was reactive to one of the membrane polypeptides. Sections of fixed rat livers were reacted with each of the antibodies and then the primary antibody was localized by indirect immunocytochemical methods using horseradish peroxidase or colloidal gold as labels. Each of the three antigens was localized by this method to the bile canalicular domain of the hepatocyte plasma membrane.     

Identification of rat hepatocyte plasma membrane proteins using monoclonal antibodies

We have localized and identified five rat hepatocyte plasma membrane proteins using hybridoma technology in combination with morphological and biochemical methods. Three different membrane preparations were used as immunogens: isolated hepatocytes, a preparation of plasma membrane sheets that contained all three recognizable surface domains of the intact hepatocyte (sinusoidal, lateral, and bile canalicular), and a glycoprotein subfraction of that plasma membrane preparation. We selected monoclonal IgGs that were hepatocyte specific and localized them using both immunofluorescence on 0.5-micron sections of frozen liver and immunoperoxidase at the ultrastructural level. One antigen (HA 4) was localized predominantly to the bile canalicular surface, whereas three (CE 9, HA 21, and HA 116) were localized predominantly to the lateral and sinusoidal surfaces. One antigen (HA 16) was present in all three domains. Only one antigen (HA 116) could be detected in intracellular structures both in the periphery of the cell and in the Golgi region. The antigens were all integral membrane proteins as judged by their stability to alkaline extraction and solubility in detergents. The apparent molecular weights of the antigens were established by immunoprecipitation and/or immunoblotting. In a related study (Bartles, J.R., L.T. Braiterman, and A.L. Hubbard, 1985, J. Cell. Biol., 100:1126-1138), we present biochemical confirmation of the domain-specific localizations for two of the antigens, HA 4 and CE 9, and demonstrate their suitability as endogenous domain markers for monitoring the separation of bile canalicular and sinusoidal lateral membrane on sucrose density gradients.     

Endogenous and exogenous domain markers of the rat hepatocyte plasma membrane

We have used a combined biochemical and morphological approach to establish the suitability of certain endogenous and exogenous domain markers for monitoring the separation of rat hepatocyte plasma membrane domains in sucrose density gradients. As endogenous domain markers, we employed two of the integral plasma membrane protein antigens, HA 4 and CE 9, localized to the bile canalicular and sinusoidal/lateral domains, respectively, of the hepatocyte plasma membrane in rat liver tissue (Hubbard, A. L., J. R. Bartles, and L. T. Braiterman, 1985, J. Cell Biol., 100:1115-1125). We used immunoelectron microscopy with a colloidal gold probe to demonstrate that HA 4 and CE 9 retained their domain-specific localizations on isolated hepatocyte plasma membrane sheets. When the plasma membrane sheets were vesiculated by sonication and the resulting vesicles were centrifuged to equilibrium in sucrose density gradients, quantitative immunoblotting revealed that the vesicles containing HA 4 and those containing CE 9 exhibited distinct density profiles. The density profile for the bile canalicular vesicles (marked by HA 4) was characterized by a single peak at a density of 1.10 g/cm3. The density profile for the sinusoidal/lateral vesicles (marked by CE 9) was bimodal, with a peak in the body of the gradient at a density of 1.14 g/cm3 and a smaller amount in the pellet (density greater than or equal to 1.17 g/cm3). We used this sucrose gradient fractionation as a diagnostic procedure to assign domain localizations for several other hepatocyte plasma membrane antigens and enzyme activities. In addition, we used the technique to demonstrate that 125I-wheat germ agglutinin, introduced during isolated liver perfusion at 4 degrees C, can serve as an exogenous domain marker for the sinusoidal domain of the rat hepatocyte plasma membrane.     

CMP-N-acetylneuraminic acid hydrolase, an ectoenzyme distributed unevenly over the hepatocyte surface.

The regional localization of CMP-N-acetylneuramic acid hydrolase at the hepatocyte surface was studied by using plasma membranes and hepatocytes isolated from rat liver. 1. By homogenization of the rat liver plasma membrane preparations and subsequent discontinuous sucrose gradient centrifugation, one light and two heavy membrane fractions were obtained. The origin of these three subfractions is discussed based on the specific activities in the three fractions of 5'-nucleotidase, alakaline phosphatase and Mg2+-ATPase and on electron microscopic examination of the fractions. Evidence is given suggesting that the light fraction is derived from the bile canalicular surface of the plasma membrane, and that the heavy fractions are derived predominantly from the sinusoidal and lateral surfaces of the liver cell membrane. CMP-AcNeu hydrolase was present at highest specific activity in one of the heavy subfractions. Therefore it is concluded that CMP-AcNeu hdyrolase is located preferentially in the sinusoidal and/or lateral plasma membrane parts of the liver cell. 2. Experiments with intact and disintegrated hepatocytes isolated from rat liver indicated that CMP-AcNeu hydrolase is located at the surface of the cell membrane, with its functional group directed to the outside.     

Isolation and characterization of a liver plasma membrane fraction enriched in glucagon-sensitive adenylate cyclase

Partially purified liver plasma membranes were fractionated further on sucrose layers. Three membrane populations, numbered Peaks 1, 2 and 3, were isolated at densities of 1.23, 1.16, and 1.03, respectively. Peaks 1 and 2 were enriched to a similar degree in 5′-nucleotidase activity, a plasma membrane marker, relative to membranes in Peak 3. Electron micrographs indicated that Peak 1 possessed desmosomes and bile canaliculi, while Peak 2 contained large vesicles as well as smaller vesicular structures attached to membranes. The latter have been attributed to hepatocyte sinusoidal surfaces. All three membrane fractions contained adenylate cyclase activity with the highest specific activity found in Peak 2. The enzyme in all three peaks was F sensitive with higher sensitivity in Peaks 1 and 2. Glucagon sensitivity of adenylate cyclase in Peak 2 membranes was four times that of Peak 1. Only Peak 2 membranes were sensitive to epinephrine. The Peak 2 membranes were three times more sensitive to glucagon than the partially purified membranes from which they were derived. These findings indicate that, while both bile canalicular and sinusoidal faces of hepatocytes possess adenylate cyclase, the sinusoidal fraction is more sensitive to glucagon. Solubilized adenylate cyclase of the Peak 2 membranes, obtained as the 165,000g supernate of membranes treated with Lubrol-PX, was sensitive to stimulation by guanyl nucleotide analogs. Guanyl nucleotide sensitivity thus resides in the catalytic site and is not dependent on membrane integrity. All three membrane fractions possessed similar activities of nucleotide phosphohydrolase activity.     

Fluorescence anisotropy from diphenylhexatriene in rat liver plasma membranes

Rat livers were fractionated to obtain intracellular membrane preparations and a highly purified preparation of bile canaliculi. The fraction containing bile canaliculi was homogenized and subfractionated to give fractions representing fragments of contiguous membrane and of canalicular microvilli. The relative purity and extent of contamination of each preparation was determined. When the fluorescent probe 1,6-diphenyl-1,3,5-hexatriene was incorporated into aliquots of each fraction at the same probe: lipid ratio and the steady-state anisotropy of its fluorescence measured, it was found that the plasma membrane preparations were much more ordered than the intracellular membrane preparations. Of the plasma membrane preparations, that containing the canalicular microvilli was the most ordered, even allowing for any contribution of contaminants. Thus the microvillus membrane of the bile canaliculus appears to be the most ordered domain of the plasma membrane of the hepatocyte. The high order in this domain may be a factor in reducing the susceptibility to bile salt damage during bile secretion, since it is this region which is exposed to high concentrations of bile salts in vivo.     

Hepatocellular transport and secretion of biliary lipids.

Bile is the route for elimination of cholesterol from the body. Recent studies have begun to elucidate hepatocellular, molecular and physical-chemical mechanisms whereby bile salts stimulate biliary secretion of cholesterol together with phospholipids, which are enriched (up to 95%) in phosphatidylcholines. Active translocation of bile salts and phosphatidylcholines across the hepatocyte's canalicular plasma membrane provides the driving force for biliary lipid secretion. This facilitates physical-chemical interactions between detergent-like bile salt molecules and the ectoplasmic leaflet of the canalicular membrane, which result in biliary secretion of cholesterol and phosphatidylcholines as vesicles. Within the hepatocyte, separate molecular pathways function to resupply bile salts, phosphatidylcholines and cholesterol to the canalicular membrane for ongoing biliary lipid secretion.     

Isolation of rat hepatocyte plasma membranes. I. Presence of the three major domains

A rat liver plasma membrane preparation was isolated and characterized both biochemically and morphologically. The isolation procedure was rapid, simple and effective in producing a membrane fraction with the following biochemical characteristics: approximately 40-fold enrichment in three plasma membrane markers, 5'-nucleotidase, alkaline phosphodiesterase I (both putative bile canalicular membrane enzymes), and the asialo-glycoprotein (ASGP) receptor (a membrane glycoprotein present along the sinusoidal front of hepatocytes); a yield of each of these plasma membrane markers that averaged approximately 16%; and minimal contamination by lysosomes, nuclei, and mitochondria, but persistent contamination by elements of the endoplasmic reticulum. Morphological analysis of the preparation revealed that all three major domains of the hepatocyte plasma membrane (sinusoidal, lateral, and bile canalicular) were present in substantial amounts. The identification of sinusoidal membrane was further confirmed when ASGP binding sites were localized predominantly to this membrane in the isolated PM using electron microscope autoradiography. By morphometry, the sinusoidal front membrane accounted for 47% of the total membrane in the preparation, whereas the lateral surface and bile canalicular membrane accounted for 6.8% and 23% respectively. This is the first report of such a large fraction of sinusoidal membrane in a liver plasma membrane preparation.     

Phosphoinositide 3-kinase is involved in the glucagon-induced translocation of aquaporin-8 to hepatocyte plasma membrane

BACKGROUND INFORMATION: PI3K (phosphoinositide 3-kinase) mediates several signal transduction pathways in hepatocytes, including some involved in the regulation of vesicle trafficking. Hepatocytes express the water channel AQP8 (aquaporin-8) predominantly in an intracellular location, and it redistributes to the canalicular membrane, upon stimulation with the hormone glucagon, by a cAMP/protein kinase A-dependent mechanism. Since glucagon is capable of stimulating PI3K activity in hepatocytes and a cross talk between cAMP and PI3K has been suggested, in the present study, we examine whether PI3K activation is involved in the glucagon-induced translocation of AQP8. RESULTS: By quantitative immunoblotting of purified hepatocyte plasma membranes, we found that the preincubation of cells with two structurally different PI3K inhibitors, wortmannin or LY294002, prevented the glucagon-induced translocation of AQP8 to hepatocyte plasma membrane. Confocal immunofluorescence microscopy in cultured hepatocytes confirmed the dependence of the hormone-induced redistribution of AQP8 on PI3K activity. Functional studies showed that the PI3K inhibitors were also capable of preventing the glucagon-induced increase in hepatocyte osmotic membrane water permeability. CONCLUSIONS: Our results suggest that PI3K activation is involved in the glucagon-dependent signal transduction pathways leading to hepatocyte AQP8 translocation.     

Membrane microdomain components of Histoplasma capsulatum yeast forms, and their role in alveolar macrophage infectivity

Analysis of membrane lipids of Histoplasma capsulatum showed that ~40% of fungal ergosterol is present in membrane microdomain fractions resistant to treatment with non-ionic detergent at 4°C. Specific proteins were also enriched in these fractions, particularly Pma1p a yeast microdomain protein marker (a plasma membrane proton ATPase), a 30kDa laminin-binding protein, and a 50kDa protein recognized by anti-α5-integrin antibody. To better understand the role of ergosterol-dependent microdomains in fungal biology and pathogenicity, H. capsulatum yeast forms were treated with a sterol chelator, methyl-beta-cyclodextrin (mβCD). Removal of ergosterol by mβCD incubation led to disorganization of ergosterol-enriched microdomains containing Pma1p and the 30kDa protein, resulting in displacement of these proteins from detergent-insoluble to -soluble fractions in sucrose density gradient ultracentrifugation. mβCD treatment did not displace/remove the 50kDa α5-integrin-like protein nor had effect on the organization of glycosphingolipids present in the detergent-resistant fractions. Ergosterol-enriched membrane microdomains were also shown to be important for infectivity of alveolar macrophages; after treatment of yeasts with mβCD, macrophage infectivity was reduced by 45%. These findings suggest the existence of two populations of detergent-resistant membrane microdomains in H. capsulatum yeast forms: (i) ergosterol-independent microdomains rich in integrin-like proteins and glycosphingolipids, possibly involved in signal transduction; (ii) ergosterol-enriched microdomains containing Pma1p and the 30kDa laminin-binding protein; ergosterol and/or the 30kDa protein may be involved in macrophage infectivity.     

Preservation of hepatocyte plasma membrane domains during cell division in situ in regenerating rat liver

We have utilized antibodies against five domain-specific integral proteins of the rat hepatocyte plasma membrane to examine the fates of the plasma membrane domains during hepatocyte division in the regenerating rat liver. The proteins were quantified on immunoblots of liver homogenates prepared during the peak of hepatocyte mitotic activity, 28-30 hr after two-thirds hepatectomy. Two sinusoidal/lateral proteins, CE 9 and the asialoglycoprotein receptor, and one bile canalicular protein, dipeptidylpeptidase IV, were not changed significantly in amount; whereas one sinusoidal/lateral protein, the epidermal growth factor receptor, and one bile canalicular protein, HA 4, were reduced to less than or equal to 50% of control levels. Light microscopic examination of plastic sections of regenerating liver tissue revealed that the mitotic hepatocytes generally appeared to retain normal contacts with neighboring interphase hepatocytes. Immunofluorescence was used to localize the domain-specific proteins on mitotic hepatocytes identified in 0.5-micron frozen sections of 28- to 30-hr regenerating liver tissue. Independent of mitotic stage, the hepatocytes retained mutually exclusive bile canalicular and sinusoidal/lateral domains, as defined at the molecular level by the distributions of specific proteins, such as HA 4 and CE 9, respectively.     

Localization of plasma membrane CD38 is domain specific in rat hepatocyte

CD38 is a 42- to 45-kDa type II transmembrane glycoprotein with the ability to synthesize cADPR, a metabolite with potent calcium mobilizing properties independent of IP(3). We report here the primary characterization and localization of CD38 in the plasma membrane fraction of rat hepatocyte. Western blot analysis of a partially purified plasma membrane fraction with a panel of polyclonal antibodies against CD38 detected a 42- to 45-kDa protein band which is characteristic of CD38. ADP-ribosyl cyclase activity was found to be present in the plasma membrane fraction, indicating the presence of functionally active CD38. Subfractionation of the plasma membrane to the sinusoidal and bile canalicular membrane fractions showed the presence of ADP-ribosyl cyclase activity in both fractions with the sinusoidal membrane fraction having a 10-fold higher specific activity than the bile canalicular membrane fraction. Immunohistochemical staining with the same panel of polyclonal antibodies showed exclusive differential spatial localization to both the nuclei and sinusoidal domain of the plasma membrane. It is possible that the different spatial distribution of CD38 in the rat hepatocyte might be responsible for its myriad of previously known functional roles.     

Separation of hepatocyte plasma membrane domains by free flow electrophoresis

We have applied free flow electrophoresis to separate the canalicular and basolateral (sinusoidal and lateral) domains of rat hepatocyte plasma membranes. Hepatocyte plasma membranes were prepurified by rat zonal and discontinous sucrose gradient centrifugation. In electrophoretic separation, the canalicular membranes were more deflected toward the anode than the basolateral membranes. Na+-dependent taurocholate uptake could be measured in both membrane fractions, transport activity being highest in fractions containing the highest specific activity in the basolateral marker enzyme Na+-K+-ATPase. Thus, differences in electrophoretic mobility permit the separation of functional intact plasma membrane vesicles derived from basolateral and canalicular plasma membrane domains of rat hepatocyte.