Claudin-1 involved in neonatal ichthyosis sclerosing cholangitis syndrome regulates hepatic paracellular permeability

Authors

  • Brigitte Grosse,

    1. INSERM UMR-S757, Orsay, University Paris-Sud 11, Paris, France
    Search for more papers by this author
  • Doris Cassio,

    1. INSERM UMR-S757, Orsay, University Paris-Sud 11, Paris, France
    Search for more papers by this author
  • Nadya Yousef,

    1. INSERM UMR-S757, Orsay, University Paris-Sud 11, Paris, France
    2. Pediatric Intensive Care Unit, National Reference Center for Biliary Atresia, CHU Bicêtre, Assistance Publique–Hôpitaux de Paris, University Paris-Sud 11, Paris, France
    Search for more papers by this author
  • Céline Bernardo,

    1. Pathology Unit, and National Reference Center for Biliary Atresia, CHU Bicêtre, Assistance Publique–Hôpitaux de Paris, University Paris-Sud 11, Paris, France
    Search for more papers by this author
  • Emmanuel Jacquemin,

    1. INSERM UMR-S757, Orsay, University Paris-Sud 11, Paris, France
    2. Pediatric Hepatology Unit and National Reference Center for Biliary Atresia, CHU Bicêtre, Assistance Publique–Hôpitaux de Paris, University Paris-Sud 11, Paris, France
    Search for more papers by this author
  • Emmanuel Gonzales

    Corresponding author
    1. INSERM UMR-S757, Orsay, University Paris-Sud 11, Paris, France
    2. Pediatric Hepatology Unit and National Reference Center for Biliary Atresia, CHU Bicêtre, Assistance Publique–Hôpitaux de Paris, University Paris-Sud 11, Paris, France
    • Service d'Hépatologie Pédiatrique, Hôpital Bicêtre, 78 rue du Général Leclerc, Le Kremlin-Bicêtre, 94275 Paris, France===

    Search for more papers by this author
    • fax: +33 1 45 21 28 16


  • Potential conflict of interest: Nothing to report.

  • N.Y. was supported by the Groupe Francophone d'Hépatologie, Gastroentérologie et Nutrition Pédiatriques (GFHGNP), France.

Abstract

Neonatal ichthyosis and sclerosing cholangitis (NISCH) syndrome is a liver disease caused by mutations of CLDN1 encoding Claudin-1, a tight-junction (TJ) protein. In this syndrome, it is speculated that cholestasis is caused by Claudin-1 absence, leading to increased paracellular permeability and liver injuries secondary to paracellular bile regurgitation. We studied the role of claudin-1 in hepatic paracellular permeability. A NISCH liver and polarized rat cell lines forming TJs, the hepatocellular Can 10 and the cholangiocellular normal rat choloangiocyte (NRC), were used. In contrast to NRC, Can 10 does not express claudin-1. Can 10 cells were transfected with a plasmid encoding Claudin-1, and stable Claudin-1-expressing clones were isolated. Claudin-1 expression was silenced by transfection with short interfering RNA in Can 10 clones and with short hairpin RNA in NRC. Claudin-1 expression was evaluated by quantitative reverse-transcriptase polymerase chain reaction, immunoblotting, and immunolocalization. Paracellular permeability was assessed by fluorescein isothiocyanate-dextran passage in both lines and by transepithelial resistance measurements in NRC. In the NISCH liver, Claudin-1 was not detected in hepatocytes or cholangiocytes. In Claudin-1 expressing Can 10 clones, Claudin-1 was localized at the TJ and paracellular permeability was decreased, compared to parental Can 10 cells, this decrease correlating with claudin-1 levels. Silencing of Claudin-1 in Can 10 clones increased paracellular permeability to a level similar to that of parental cells. Similarly, we observed an increase of paracellular permeability in NRC cells silenced for claudin-1 expression. Conclusion: Defect in claudin-1 expression increases paracellular permeability in polarized hepatic cell lines, supporting the hypothesis that paracellular bile leakage through deficient TJs is involved in liver pathology observed in NISCH syndrome. (Hepatology 2012)

Neonatal ichthyosis and sclerosing cholangitis (NISCH) syndrome (MIM 607626) is a rare autosomal recessive disease caused by mutations in CLDN1 encoding Claudin-1, an integral membrane protein of tight junctions (TJs).1, 2 To date, 12 patients have been reported as presenting with neonatal-onset sclerosing cholangitis, ichthyosis, hypotrichosis, and dental anomalies.2-5 Two deletions resulting in premature stop codon have been identified.2, 3 In skin and fibroblasts of NISCH patients, Claudin-1 was not detected by immunoblotting or immunostaining.2, 3, 6 In the 2 patients reported on, so far, in whom liver claudin-1 expression was studied, expression in cholangiocytes was not studied. Claudin-1 was not detected by immunoblotting using a whole liver extract in 1 patient or by immunofluorescence in hepatocytes in the other patient.2

TJs are junctional complexes that contribute to the formation of polarized epithelial barriers by controlling the extent and selectivity of permeability along the paracellular pathway (i.e., the gate function) and by forming an apical/basolateral intramembrane diffusion barrier to membrane proteins in the outer leaflet of the plasma membrane (i.e., the fence function).7, 8 The structure of TJs is complex, involving integral membrane proteins (e.g., occludin, claudins, and junction adhesion molecules; JAMs) linked to a dense cytoplasmic network of scaffolding and adaptor proteins, signaling components, and actin-binding cytoskeletal linkers. In various diseases, including familial hypercholanemia, causative mutations in genes encoding TJ proteins have been identified.9 Moreover, claudin defects have been reported in skin, kidney, inner ear, and eye genetic diseases, revealing the crucial role played by TJs and claudins in epithelia.10 In the liver, TJs separate bile from plasma. They play a critical role in establishing the biliary canalicular pole, maintaining canalicular localization of specific proteins, sealing bile canaliculi (BC), and protecting liver cells from toxic bile components, in particular bile acids.11 Secondary changes in the expression pattern of TJ proteins and their transcripts have been described in different animal models of cholestasis.12-14 In various human cholestatic diseases, secondary alterations of TJs have been reported, both in hepatocytes and in cholangiocytes, suggesting that paracellular permeability plays a role in the mechanism of cholestasis.15-19

In patients with NISCH syndrome, it is speculated that increased paracellular permeability caused by primary Claudin-1 defect leads to bile leakage, resulting in bile duct and hepatocellular injury. Unfortunately, the basis for the sclerosing cholangitis in NISCH patients could not be studied in Cldn1-knockout mice because of severe impairment of epidermal barrier function and subsequent dehydration, which results in early lethality.10, 20 Claudin-1-related modifications of paracellular permeability have been reported in Madin-Darby canine kidney cells,21, 22 but, so far, comparable studies in liver cell lines have not been carried out. Herein, we studied the effect of claudin-1 expression on paracellular permeability in rat hepatocellular (Can 10) and biliary (i.e., normal rat cholangiocyte; NRC) polarized cell lines. Our results show that the level of Claudin-1 expression correlates with paracellular permeability in both lines, supporting the pathophysiological mechanism of bile leakage caused by a primary Claudin-1 defect in the liver disease observed in NISCH syndrome.

Abbreviations

BC, bile canaliculi; BSEP, bile salt export pump; DMEM, Dulbecco's modified Eagle's medium; FITC, fluorescein isothiocyanate; HCV, hepatitis C virus; JAMs, junction adhesion molecules; MDR3, multidrug resistance 3 P-glycoprotein; MEM, minimal essential medium; mRNA, messenger RNA; NISCH, neonatal ichthyosis and sclerosing cholangitis; NRC, normal rat cholangiocyte; qRT-PCR, quantitative reverse-transcriptase polymerase chain reaction; shRNA, short hairpin RNA; siRNA, short interfering RNA; TER, transepithelial resistance; TJ, tight junction; ZO, zona occludens.

Materials and Methods

Chemicals.

Fetal calf serum, Dulbecco's modified Eagle's medium (DMEM) F12 medium, some supplements for NRC culture (minimal essential medium [MEM] non-amino-acid, lipid concentrate, vitamin solution, and L-glutamine), Opti-MEM, and Geneticin were from Gibco (Grand Island, NY). Puromycin was from Ozyme (Paris, France). Type I collagen was from BD Biosciences (Franklin Lakes, NJ). Coon's modified F12 medium, some supplements for NRC culture (e.g., spite medium, trypsin inhibitor soybean, bovine pituitary extract, dexamethasone, and triodothyronine), TRI Reagent, 4-kDa-fluorescein isothiocyanate (FITC)-dextran, and antibody directed against actin were from Sigma-Aldrich (St. Louis, MO). Antibodies directed against human or mouse claudin-1, -2, -3, -4, and -5, occludin, and JAM-A were from Zymed (South San Francisco, CA). Antibodies against E-cadherin and zona occludens (ZO)-2 were from Transduction Laboratories (Lexington, KY), and anti-ZO-1 was from Dr. Bruce Stevenson (Salk Institute, La Jolla, CA).23 FITC-dextran (3-kDa) and secondary antibodies conjugated to Alexa Fluor and horseradish peroxidase were from Molecular Probes (Eugene, OR).

Cell Lines and Cell Culture.

Can 10 cells were cultured as previously described.24 The NRC cell line, kindly provided by Dr. Nicholas LaRusso (Mayo Clinic, Rochester, MN), was cultured on type I collagen (except for inserts), as previously described.25 Cells were cultured on plastic dishes (RNA and protein extraction), on glass coverslips (immunostaining and paracellular permeability measurement for Can 10), and on inserts (transepithelial resistance [TER] and fluorescent dextran passage measurements of NRC).

Generation of Claudin-1 Expressing Can 10 Clones (Can 10-CLDN-1).

pCMV6-A, encoding human Claudin-1 (OriGene Technologies, Inc., Rockville, MD), was transfected into Can 10 cells, using the Microporator (MP-100; Labtech International, East Sussex, UK), according to the manufacturer's instructions. Clonal selection of stably transfected cells was obtained with 400 μg/mL of Geneticin, added 2 days after transfection. Stable Geneticin-resistant Can 10-CLDN-1 clones were obtained at a frequency of 8 × 10−5; their chromosome content, determined as previously described,24 was similar to that of parental Can 10 cells (mean number, 50; range, 46-52). Claudin-1 expression was examined by quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR), immunoblotting, and immunolocalization in cells plated at 2-4 × 103/cm2 and cultured for 7-8 days.

Silencing of Claudin-1.

Short interfering RNA.

Can 10-CLDN-1 clones were transfected with short interfering RNA (siRNA) (5′-GAUCCAGUGCAAAGUCUU-3′) duplex targeting claudin-1 and with scrambled siRNA (Eurogentec, San Diego, CA). Cells were plated at 2.5 × 104/cm2 and immediately transfected using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA), according to the manufacturer's instructions. Final concentrations were 40 nM for siRNA and 2 μL/mL for Lipofectamine RNAiMAX. Cells were studied 2-3 days after siRNA transfection.

Short hairpin RNA.

NRC cells were transfected 24 hours after plating (2 × 104/cm2) with a pool of three mouse claudin-1 short hairpin RNA (shRNA)-lentiviral vector plamids (Santa Cruz Biotechnology, Santa Cruz, CA); among them, two were identical in sequence to rat claudin-1 and one was not. NRC cells were also transfected with scrambled shRNA (Santa Cruz). Transfection was performed using FuGENE HD reagent (Roche, Mannheim, Germany), according to the manufacturer's instructions (transfection reagent:DNA = 5:2). Puromycin (10 μg/mL) was added 48 hours after transfection; two independent stably transfected populations (claudin-1 shRNA pop-A and pop-B) and one scrambled shRNA population were isolated after 1-2 months in puromycin containing medium.

The silencing of claudin-1 expression was studied in cells plated at high density (3 × 104/cm2 for Can 10-CLDN-1 clones and 4-5 × 105/cm2 for NRC) after 2-3 days (Can 10-CLDN1 clones) and 3-4 days (NRC) of culture.

Real-Time qRT-PCR.

Total RNA was isolated using TRI Reagent and was converted to complementary DNA (cDNA) using SuperScript II Reverse Transcriptase (Invitrogen). Real-time PCR was performed using iQ SYBR Green SuperMix (Bio-Rad, Hercules, CA), according to the manufacturer's instructions. Relative expression of claudin-1 was quantified using glyceraldehyde 3-phosphate dehydrogenase and βactin as housekeeping genes. Primers are detailed in Supporting Table 1.

Immunoblotting.

Protein lysates were processed as previously described.26 Primary antibody dilutions were as follows: ZO-1, undiluted; JAM-A, 1/100; claudin-1, -2, -3, and -5, 1/125; ZO-2, 1/200; claudin-4, 1/250; occludin, 1/500; and actin, 1/1000. ImageJ 1.37v software was used for densitometric analysis.

Immunolocalization.

Cells.

Cells were fixed in ethanol (4°C, 30 minutes), then in acetone (room temperature, 1 minute). Thereafter, indirect immunofluorescence staining was performed as previously described.26 Antibodies against claudin-1, E-cadherin, and ZO-1 were used at dilutions of 1/100, 1/200, and undiluted, respectively. Appropriate Alexa Fluor–conjugated secondary antibodies were used at a dilution of 1/500. Cells were observed using a Zeiss Axioskop fluorescence microscope (Carl Zeiss, Inc., Göttingen, Germany). Confocal analysis was performed using a Zeiss LSM 510 confocal microscope equipped with a ×63 objective, and xy sections were taken in 0.3-μm steps.

Liver sections.

Immunostaining was performed as previously described.27 Liver samples of a NISCH patient (patient II-1, family 2, reference 2) with an homozygous CLDN1 truncating mutation were studied.2 A healthy human liver served as a control. Anti-claudin-1 antibody was used, at a dilution of 1/100. Anti-MDR3 (multidrug resistance P-glycoprotein) (Sigma-Aldrich) and anti-BSEP (bile salt export pump) antibodies were used as previously described.27

Epithelial Permeability Assay in Can 10 Cells.

Paracellular permeability of Can 10 cells was assessed by measuring paracellular diffusion into the BC of 3- and 4-kDa FITC-dextrans, as previously described.28 Cells were plated on glass coverslips at 4 × 103/cm2, and paracellular permeability was examined after 7-8 days. For cells transfected with siRNA, a higher plating density (3 × 104cells/cm2) was used, and cells were examined after 2-3 days. Cells were incubated with 2 mg/mL of FITC-dextran for 10 minutes at 37°C, washed twice with cold, serum-free medium, and observed within 5-10 minutes after mounting. Fluorescent BC were counted and expressed as a percentage of total number of BC observed by phase contrast. Three coverslips were examined from each culture (four fields per coverslip and at least 200 BC were observed).

TER and Paracellular Permeability of NRC Cells.

Paracellular permeability of wild-type NRC, NRC transfected with scrambled shRNA, and the two populations transfected with claudin-1 shRNA was assessed by measuring the TER and transepithelial flux of 4-kDa FITC-dextran. Experiments were performed on confluent monolayers obtained by culturing on inserts (0.4 μM pore size, BD Falcon; BD Biosciences), for 3-4 days, with cells plated at 3-4 × 105/cm2. TER was measured once a day using a Millicell-ERS volt-ohm meter (Millipore, Billerica, MA). TER was expressed as ohm.cm2 after background subtraction. On the last day, 4-kDa FITC-dextran (2 mg/mL of serum-free culture medium) was added to the upper chamber. Samples were collected from the lower compartment at 30-minute intervals for 2 hours. Fluorescence intensity was measured using a Wallac fluorometer (VICTOR3; PerkinElmer, Walthamn, MA). FITC-dextran flux was determined by linear regression after correction for background and expressed as ng/h.cm2, using a standard curve.

Statistical Analysis.

Data are expressed as means ± standard deviation. Statistical analyses were performed using the Student's t test, with a P value <0.05 being considered statistically significant.

Results

We first studied Claudin-1 expression in healthy human liver and in a liver of a NISCH patient. In healthy human liver, Claudin-1 was detected in hepatocytes (Fig. 1A) and cholangiocytes (Fig. 1B). In NISCH patient liver, we confirmed the absence of Claudin-1 in hepatocytes (Fig. 1C), we showed the absence of claudin-1 in cholangiocytes (Fig. 1D), and we observed a normal canalicular localization of BSEP (Fig. 1E) and MDR3 (Fig. 1F). To test the role of claudin-1 in paracellular permeability in the liver, we used two polarized rat cell lines: the hepatocellular Can 10 and the biliary NRC. Both form TJs. However, in contrast to NRC,29 Can 10 cells do not express claudin-1, although they express and properly target a large panel of TJ proteins.24, 30 Therefore, we experimentally manipulated claudin-1 levels in these cells and subsequently determined their paracellular permeability.

Figure 1.

Claudin-1 expression in healthy human liver and in a NISCH patient liver. Liver sections from healthy human liver (A and B) or a NISCH patient (C-F) were immunostained with anti-claudin-1 (A-D), anti-BSEP (E), and anti-MDR3 (F) antibodies. In healthy human liver, Claudin-1 was concentrated at the TJ, sealing canaliculi, in hepatocyte (A, arrows), and at the upper part of the lateral plasma membrane in cholangiocyte (B, arrowhead). In a NISCH patient liver, Claudin-1 was not detected in hepatocytes (C) or cholangiocytes (D), whereas BSEP (E) and MDR3 (F) were appropriately expressed at the canalicular membrane. Bar, 30 μm.

Generation and Characterization of Stable Can 10 Clones Expressing Claudin-1.

Can 10 is derived from the rat hepatoma Fao line after a transient culture in spheroids.24 It exhibits the typical hepatocyte polarity with functional long BC mimicking in vivo BC. It has been shown to be an appropriate model to study canalicular vectorial transport.24, 30-32

The pCMV6-CLDN-1 plasmid was successfully transfected into Can 10 cells. We isolated a total of 23 stable Geneticin-resistant Can 10-CLDN-1 clones. Among those, nine clones expressed Claudin-1 and four were selected for further analysis based on their different expression levels of claudin-1. High levels of claudin-1 messenger RNA (mRNA) and protein were found in clone 10 and lower levels in clones 1, 21, and 23 (Fig. 2A,B). One negative clone, clone 6, was used as a control. As expected, Claudin-1 mRNA and protein were not detected in negative clone 6 and parental Can 10 cells (Fig. 2A,B). All clones expressed other TJ proteins, claudin-2, -3, -4, and -5, JAM-A, occludin, ZO-1, and ZO-2 (Fig. 2B). Minimal changes were observed in the expression of these TJ proteins, except for claudin-4, which was decreased in claudin-1-expressing Can 10-CLDN-1 clones. To determine the localization of Claudin-1 and homogeneity of cell populations, Can 10-CLDN-1 clones were analyzed by immunofluorescence, using ZO-1 as a TJ control marker. In all four expressing clones, Claudin-1 was present in most cells and colocalized with ZO-1 at the TJ, as illustrated for clones 10 and 21 (Fig. 2C).

Figure 2.

Can 10-CLDN1 clones express various levels of Claudin-1 at the TJ. (A) Expression levels of Claudin-1 mRNA were determined in five Can 10-CLDN1 clones (including negative clone 6) and in Can 10 parental cells, using qRT-PCR. Claudin-1 mRNA level is expressed in arbitrary units, with a level of 1 being attributed to clone 1, the lowest producer of claudin-1. (B) Immunoblotting. Each lane was loaded with 25 μg of proteins. Claudin-1, -2, -3, -4, and -5, JAM-A, occludin, ZO-1, and ZO-2 were detected. Actin was used as a loading control. (C) Immunolocalization of Claudin-1 and ZO-1 in three Can 10-CLDN1 clones (including negative clone 6) and in Can 10 parental cells. The corresponding phase-contrast image is shown on the left. Some BC, well visible in phase contrast, are indicated by arrows. Bar, 10 μm.

Paracellular Permeability Is Decreased in Can 10 Clones Expressing Claudin-1.

Paracellular permeability was assessed by measuring the passage of FITC-dextrans into the BC. All Can 10-CLDN-1 clones expressing Claudin-1 showed reduced paracellular passage of 3- (Fig. 3A) and 4-kDa (Fig. 3B), FITC-dextrans, when compared to negative clone 6 and parental Can 10 cells. Moreover, the decrease in permeability correlates with Claudin-1 expression level (Fig. 3C) (for Claudin-1 expressing clones, r2 = 0.992 and 0.953 for 3- and 4-kDa dextrans, respectively).

Figure 3.

Claudin-1 expression decreases paracellular permeability of Can 10-CLDN-1 clones. (A and B) Paracellular permeability was assessed by measuring paracellular diffusion into the BC of 3- and 4-kDa FITC-dextrans. Values are means from three to five independent cell cultures. **P < 0.01; ***P < 0.001, as compared with negative clone 6 and parental Can 10 cells. (C) Relation between the paracellular permeability and the level of Claudin-1, calculated by densitometric analysis of immunoblots. Claudin-1 level is expressed in arbitrary units, with a level of 1 being attributed to clone 1, the lowest producer of Claudin-1.

Claudin-1 Silencing by siRNA Increases Paracellular Permeability to the Level of Parental Can 10 cells.

To further demonstrate that Claudin-1 expression is directly implicated in the decrease of paracellular permeability, we used siRNA to silence Claudin-1 in three Can 10-CLDN-1 clones expressing various levels of Claudin-1. As assessed by qRT-PCR (data not shown) and immunoblotting (Fig. 4A), treatment with siRNA targeting claudin-1 resulted in a 70%-90% decrease of Claudin-1 expression, when compared to scrambled siRNA. This decrease in Claudin-1 expression was also observed by immunofluorescence (Fig. 4B). In the three clones tested, 4-kDa FITC-dextran passage increased in Claudin-1-silenced clones, reaching the level observed in parental Can 10 cells (Fig. 4C).

Figure 4.

Silencing of Claudin-1 in Can 10-CLDN1 clones increases paracellular permeability to the level of parental Can 10 cells. Cells transfected by lipofectamine with no siRNA, scrambled siRNA, and Claudin-1 siRNA were examined 2-3 days after transfection. (A) Immunoblotting. Each lane was loaded with 25 μg of proteins. Claudin-1 was codetected with actin. (B) Immunolocalization of Claudin-1 and ZO-1 with the corresponding phase-contrast image shown on the left. Some BC, well visible in phase contrast, are indicated by arrows. Bar, 10 μm. (C) Paracellular permeability (4-kDa FITC-dextrans). Values are means from three independent cell cultures of each clone. ***P < 0.001, as compared with no siRNA and scrambled siRNA-transfected cells.

Claudin-1 Silencing by shRNA Increases Paracellular Permeability of NRC Cells.

NRC is a cell line derived from a primary culture of normal rat cholangiocytes. It exhibits the typical simple polarity of cholangiocytes and has been shown to be an appropriate model to study the TJ barrier function of the bile duct epithelium.25, 29 Expression of TJ proteins in NRC cells was only partially characterized.29 Using immunoblotting, we determined that in addition to claudin-1 and -4, ZO-1, and occludin,29 NRC cells also express claudin-2, -3, and -5, ZO-2, and JAM-A (Fig. 5A). Using confocal microscopy, we studied the localization of claudin-1 in NRC cells (Fig. 5B). As expected, claudin-1 was properly targeted and colocalized with ZO-1 at the upper part of the lateral plasma membrane (i.e., the TJ level), whereas E-cadherin was present throughout the lateral plasma membrane.

Figure 5.

Expression of TJ proteins and localization of claudin-1 in NRC cells. (A) Immunoblotting. Each lane was loaded with a cell lysate aliquot containing 25 μg of proteins. (B) Immunolocalization examined by confocal microscopy of claudin-1 with the TJ-associated protein, ZO-1, and with the lateral protein, E-cadherin. Z series (20 xy sections) and xz sections taken, as indicated (white arrows) on the corresponding xy compilations, are shown. Bar 10 μm.

To evaluate the role of claudin-1 in NRC paracellular permeability, cells were transfected with claudin-1 shRNA. Two independent stably transfected populations, claudin-1 shRNA pop-A and pop-B, were isolated and characterized (Fig. 6). In both populations, we obtained a 50% decrease of claudin-1 mRNA (data not shown) and claudin-1 protein (Fig. 6A), when compared to cells transfected with scrambled shRNA and wild-type NRC cells. Using immunolocalization, we found that in contrast to wild-type cells and cells transfected with scrambled shRNA, claudin-1 is not homogeneously expressed in claudin-1 shRNA pop-A and pop-B. In both populations, negative fields for claudin-1 were observed (Fig. 6B). Paracellular permeability was examined by measuring TER (Fig. 6C,D) and FITC-dextran passage (Fig. 6E,F). In wild-type NRC and in NRC transfected with scrambled shRNA, TER increased with cell growth from days 2 to 4, whereas this increase was not observed in NRC transfected with shRNA targeting claudin-1 (Fig. 6C). At confluency, the TER was reduced in both cell populations transfected with claudin-1 shRNA (200-220 ohm.cm2), compared to the TER in cells transfected with scrambled shRNA or wild-type cells (550-570 ohm.cm2) (Fig. 6D).

Figure 6.

Silencing of claudin-1 increased paracellular permeability in NRC cells. NRC cells were transfected with scrambled or claudin-1 shRNA. Two populations, claudin-1 shRNA pop-A and pop-B, were isolated. (A) Immunoblotting. The upper panel shows a representative immunoblot. The lower panel illustrates densitometric analysis of scanned immunoblots obtained from at least three different cultures of each type of cells. *P < 0.05, as compared to scrambled shRNA-transfected cells and wild-type NRC. (B) Immunolocalization of claudin-1 and ZO-1. In each case, the corresponding phase-contrast image is shown on the left. Bar, 10 μm. (C and D) TER. (C) A representative experiment is presented (cells plated at 3-4 × 105/insert; confluent cell monolayers obtained 3-4 days after seeding). (D) Compilation of TER values obtained at confluency from three different cultures of each type of cells. ***P < 0.001, as compared to scrambled shRNA-transfected cells and wild-type NRC. (E and F) Dextran flux measurements. The passage of 4-kDa FITC-dextran from the apical to the basal side of confluent cell monolayers was followed during 2 hours. (E) A representative experiment is shown. (F) Mean values of dextran flux obtained from three different confluent cultures of each type of cells are presented. *P < 0.05; **P < 0.01, as compared to scrambled shRNA-transfected cells and wild-type NRC. The same inserts were used for TER and dextran flux measurements.

Consistent with the decrease of TER, 4-kDa FITC-dextran passage was altered in claudin-1 shRNA pop-A and pop-B (Fig. 6E), resulting in a 2- to 3-fold increase in paracellular flux of 4-kDa FITC-dextran in claudin-1 shRNA pop-A and pop-B (Fig. 6F).

Discussion

NISCH syndrome is caused by mutations in CLDN1 encoding claudin-1, an integral TJ protein. In humans and rats, claudin-1 is expressed by hepatocytes throughout the whole liver lobule and by cholangiocytes.11, 33 NISCH patients show radiological and histological features of sclerosing cholangitis, whereas absence of Claudin-1 has not been proven in cholangiocytes.2, 5 It is believed that absence of Claudin-1 in the liver alters TJ integrity, which is crucial in preventing paracellular passage of bile components, and that the ensuing leakage and paracellular regurgitation of toxic bile components (i.e., bile acids) leads to bile duct injury and cholestasis in NISCH patients. Here, we showed in a NISCH patient, that claudin-1 was not detected in hepatocytes or cholangiocytes. We showed, in hepatocellular and biliary cell lines, the two main liver cell types affected in NISCH syndrome, that a decrease in the expression level of claudin-1 leads to an increase in paracellular permeability, whereas an increase in claudin-1 expression leads to paracellular permeability decrease. These data reinforce the primary pathogenic role of mutated Claudin-1 in NISCH syndrome.

First, we used the Can 10 line as a hepatocellular polarized model to study hepatocellular paracellular permeability.24, 30-32 Can 10 cells express claudin-2, -3, -4, and -5, occludin, JAM-A, E-cadherin, and ZO-1 and -2 proteins.30 With the exception of the absence of claudin-1 and expression of claudin-4, the Can 10 TJ protein repertoire closely resembles that of human hepatocytes.11 In stable Can 10 clones expressing human Claudin-1, paracellular permeability was decreased. This decrease was proportional to the level of Claudin-1. Silencing Claudin-1 with siRNA restored paracellular permeability to the level observed in Claudin-1 nonexpressing cells. Collectively, these findings show that Claudin-1 expression regulates paracellular permeability in Can 10 cells and point out the role of Claudin-1 in maintaining TJ gate function in a hepatocellular cell type. Our expression studies of other TJ proteins showed minimal changes, except for claudin-4, which was decreased in Claudin-1-expressing Can 10-CLDN-1 clones. However, silencing claudin-1 using siRNA and shRNA in Can 10-CLDN-1 clones (Supporting Fig. 1A) and NRC (Supporting Fig. 1B), respectively, did not alter claudin-4 expression, whereas paracellular permeability was modified. Thus, variation of claudin-4 expression in Can 10-CLDN-1 clones is unlikely to account for the observed changes in paracellular permeability. Because TJ is a dynamic multiprotein complex, whether alteration of Claudin-1 expression alters TJ structure or TJ assembly remains to be determined.

Moreover, BSEP and MDR3, two canalicular proteins that play key roles in bile secretion, were correctly targeted to the canalicular membrane in a NISCH patient (Fig. 1E,F), suggesting that the fence function of TJ and the localization of bile secretion machinery are unaltered and do not account for cholestasis in NISCH patients. Biliary bile acid concentration (35 mmol/L), as well as biliary lipid composition (bile acids, 77%; phospholipids, 20%; cholesterol, 3%), were normal in the NISCH patient reported here, whereas serum bile acid concentration was elevated (106 μmol/L).27

Second, we used NRC, a rat polarized biliary line, to study biliary paracellular permeability. We showed that, in addition to claudin-1 and -4, ZO-1, and occludin,29 NRCs also express claudin-2, -3, and -5, ZO-2, and JAM-A. This claudin expression pattern is close to the one described in healthy human biliary cells, which express Claudin-1, -3, -4, -8, and -10 and ± -2.33 We confirmed that claudin-1 localized at the TJ in NRCs. We showed that a 50% silencing of claudin-1 expression induced a 2- to 3-fold increase of paracellular permeability, as assessed by TER and dextran-passage measurements. These data point out the role of claudin-1 in maintaining TJ gate function in biliary cells. Because dextrans used in the study had molecular weights higher than all types of conjugated bile acids found in humans, this suggests that in the case of Claudin-1 deficiency, such as in NISCH syndrome, biliary bile acids could regurgitate through leaky TJ.

In NISCH syndrome, cholestasis presents as a cholestasis of biliary type (i.e., obstructive).2, 5 We propose that at the level of cholangiocytes, the defect of gate function leads to biliary bile acid regurgitation in the mesenchymal periductular environment, resulting in pericholangiocellular inflammation, fibrosis, and, eventually, in sclerosing cholangitis.34 At the level of hepatocytes, the defect of gate function, possibly enhanced by obstructive cholestasis,15-19 could have (at least initially) less damaging consequences, because bile regurgitation could directly drain in the sinusoidal vascular space surrounding hepatocytes. This difference might explain the presenting cholestasis phenotype of NISCH syndrome. The reason why the sclerosing process seems to affect more intrahepatic bile ducts than extrahepatic bile ducts is unknown.2, 4 It might be a bias resulting from the fact that the few reported cholangiogramms have been performed very early in the disease course.2, 4, 5 Moreover, abnormally thin and/or irregular extrahepatic bile ducts have been observed during adolescence and young adulthood in 2 of the 4 originally reported patients with NISCH syndrome2 (unpublished results, E.J.). This observation is in accord with the expression of claudin-1 in extrahepatic bile ducts.33

It was found that discrete residues within the first extracellular loop of claudin-1 are critical for hepatitis C virus (HCV) entry.35 Monoclonal antibodies targeting claudin-1 have been shown to prevent HCV entry in human primary hepatocytes and the HepG2 line.36, 37 TJ was not altered by such treatment in the poorly polarized HepG2 line.37 However, our results obtained in highly polarized hepatocellular and stable claudin-1-expressing Can 10 cells indicate that inhibition of claudin-1 expression alters paracellular permeability. Consequently, the effect of anti-claudin-1 antibody therapy on TJ functional integrity should be studied in highly polarized hepatocellular cells expressing claudin-1, before translating such findings to human studies.

Herein, we showed that claudin-1 expression level correlates with paracellular permeability, both in hepatocellular and biliary cell lines. Our results support a pathophysiological model for the liver disease observed in NISCH syndrome, in which a bile leakage through Claudin-1-deficient TJs of hepatocellular and biliary cells might result in direct hepatocellular and/or biliary injuries and in cholestasis. Such a pathophysiological model fits well with the sclerosing cholangitis features observed in NISCH syndrome. More generally, our results support a primary role for paracellular permeability modifications induced by abnormal TJ protein expression in the pathophysiology of cholestatic liver diseases, especially sclerosing cholangitis. Nevertheless, because TJ proteins may be involved in cellular mechanisms other than regulation of paracellular permeability, such as cell proliferation and adhesion, we cannot exclude that other mechanisms might contribute to the pathogenesis of NISCH syndrome.8 We suggest that the cell systems we have generated represent useful tools for studying hepatic paracellular permeability.

Acknowledgements

The authors thank Dr. A. Hubbard (Johns Hopkins University, Baltimore, MD) for fruitful discussion and editing advice, Dr V. Nicolas (IFR 141, Plateforme d'Imagerie Cellulaire, Chatenay-Malabry, France) for help with confocal analysis, Dr B. Stieger (Division of Clinical Pharmacology, University Hospital, Zurich, Switzerland) for providing anti-BSEP antibody, S. Prigent (INSERM UMR-S757, University Paris-Sud 11, Orsay, France) for technical support, and M.J. Redon and Dr. C. Guettier (Pathology Unit, CHU Bicêtre, Assistance Publique–Hôpitaux de Paris, University Paris-Sud 11, Paris, France) for technical help with liver immunostaining.

Ancillary