Altered hepatobiliary gene expressions in PFIC1: ATP8B1 gene defect is associated with CFTR downregulation

Authors

  • Christine Demeilliers,

    1. Université de Cergy-Pontoise, GRP2H, Département de Biologie, Errmece, Cergy-Pontoise, France
    2. Inserm, U680, Paris, France
    Search for more papers by this author
  • Emmanuel Jacquemin,

    1. AP-HP, Centre Hospitalo-Universitaire Bicêtre, Service d'Hépatologie Pédiatrique, Paris, France
    Search for more papers by this author
  • Véronique Barbu,

    1. Inserm, U680, Paris, France
    2. Université Pierre et Marie Curie (UPMC-Paris 6), Faculté de Médecine, Paris, France
    Search for more papers by this author
  • Martine Mergey,

    1. Inserm, U680, Paris, France
    2. Université Pierre et Marie Curie (UPMC-Paris 6), Faculté de Médecine, Paris, France
    Search for more papers by this author
  • François Paye,

    1. Inserm, U680, Paris, France
    2. Université Pierre et Marie Curie (UPMC-Paris 6), Faculté de Médecine, Paris, France
    3. AP-HP, Hôpital Saint-Antoine, Service de Chirurgie, Paris, France
    Search for more papers by this author
  • Laura Fouassier,

    1. Inserm, U680, Paris, France
    2. Université Pierre et Marie Curie (UPMC-Paris 6), Faculté de Médecine, Paris, France
    Search for more papers by this author
  • Nicolas Chignard,

    1. Inserm, U680, Paris, France
    2. Université Pierre et Marie Curie (UPMC-Paris 6), Faculté de Médecine, Paris, France
    Search for more papers by this author
  • Chantal Housset,

    1. Inserm, U680, Paris, France
    2. Université Pierre et Marie Curie (UPMC-Paris 6), Faculté de Médecine, Paris, France
    3. AP-HP, Hôpital Tenon, Service de Biochimie-Hormonologie, Paris, France
    Search for more papers by this author
  • Nour-Eddine Lomri

    Corresponding author
    1. Université de Cergy-Pontoise, GRP2H, Département de Biologie, Errmece, Cergy-Pontoise, France
    2. Inserm, U680, Paris, France
    • Université de Cergy-Pontoise, UFR Sciences et Techniques, Site Saint-Martin, 2 avenue Adolphe Chauvin BP-222, 95302 Cergy-Pontoise, France
    Search for more papers by this author
    • fax: 33(1) 34 25 66 93.


  • Potential conflict of interest: Nothing to report.

Abstract

Recent reports in patients with PFIC1 have indicated that a gene defect in ATP8B1 could cause deregulations in bile salt transporters through decreased expression and/or activity of FXR. This study aimed to: (1) define ATP8B1 expression in human hepatobiliary cell types, and (2) determine whether ATP8B1 defect affects gene expressions related to bile secretion in these cells. ATP8B1 expression was detected by RT-PCR in hepatocytes and cholangiocytes isolated from normal human liver and gallbladder. ATP8B1 mRNA levels were 20- and 200-fold higher in bile duct and gallbladder epithelial cells, respectively, than in hepatocytes. RT-PCR analyses of the liver from two patients with PFIC1, one with PFIC2, one with biliary atresia, showed that, compared to normal liver, hepatic expressions of FXR, SHP, CYP7A1, ASBT were decreased at least by 90% in all cholestatic disorders. In contrast, NTCP transcripts were less decreased (by ≤30% vs. 97%) in PFIC1 as compared with other cholestatic disorders, while BSEP transcripts, in agreement with BSEP immunohistochemical signals, were normal or less decreased (by 50% vs. 97%). CFTR hepatic expression was decreased (by 80%), exclusively in PFIC1, while bile duct mass was not reduced, as ascertained by cytokeratin-19 immunolabeling. In Mz-ChA-2 human biliary epithelial cells, a significant decrease in CFTR expression was associated with ATP8B1 invalidation by siRNA. In conclusion, cholangiocytes are a major site of ATP8B1 hepatobiliary expression. A defect of ATP8B1 along with CFTR downregulation can impair the contribution of these cells to bile secretion, and potentially explain the extrahepatic cystic fibrosis–like manifestations that occur in PFIC1. (HEPATOLOGY 2006;43:1125–1134.)

Progressive familial intrahepatic cholestasis (PFIC) represents a heterogeneous group of inherited disorders, in which cholestasis starts in infancy and generally leads to liver failure in childhood. Three types of PFIC, caused by mutations in three separate genes, are currently recognized. The first two types, PFIC1 and 2, share common phenotypic features, including normal or nearly normal serum γ-glutamyl transpeptidase activity and little bile duct proliferation, that are unusual in other types of cholestatic liver diseases.1 In PFIC2, mutations affect the ABCB11 gene, which encodes a bile salt transporter, the canalicular bile salt export pump (BSEP).2 Thus, a defect in the extrusion of bile salts across the canalicular membrane of hepatocytes is the primary cause of cholestasis in PFIC2. In PFIC1, mutations affect the ATP8B1 gene, which encodes a protein also called FIC1.3 ATP8B1 protein belongs to a subfamily of P-type adenosine triphosphatases. Two members of this subfamily including ATP8B1 have been reported to mediate aminophospholipid translocation in plasma membranes.4, 5 The physiological function of ATP8B1 protein and the mechanisms by which ATP8B1 deficiency leads to PFIC1 disease remain poorly understood. It was previously shown that ATP8B1 is expressed in different tissues such as the liver, the intestine, and the pancreas.3, 6 This tissue distribution is consistent with the fact that in PFIC1 cholestasis may be associated with extrahepatic manifestations such as pancreatic dysfunction or chronic diarrhea.7, 8 In rat and mouse species, ATP8B1 protein has been detected by immunochemical analyses in the two epithelial cell types of the liver (hepatocytes and cholangiocytes), and has been localized at the canalicular/apical domain of these cells,5, 9 while in the human liver, only an immunolocalization of ATP8B1 at the canalicular membrane of hepatocytes has been reported.5, 9

Evidence has been provided to indicate that ATP8B1 could interfere with transcriptional activity of the farnesoid X receptor (FXR). FXR is a nuclear receptor for bile salts and a key transcriptional regulator of genes involved in the synthesis, conjugation and transport of bile salts, in the liver and in the intestine.10 FXR regulates the expression of BSEP11 and of the apical sodium-dependent bile salt transporter (ASBT), a protein involved in the intestinal absorption of bile salts.12 In response to elevated bile salt levels, FXR is activated, which subsequently stimulates BSEP expression in the liver and represses ASBT expression in the intestine.11, 12 This latter effect is mediated by the small heterodimer partner (SHP), a gene repressor induced by FXR.12 It has been previously shown that loss of ATP8B1 in the ileum of patients with PFIC1 impaired FXR expression and activation, leading to an overexpression of ASBT, which would presumably cause intestinal bile salt hyperabsorption in these subjects.13 Subsequently, it has been shown by another group that, in the liver from one patient with PFIC1, the expression of FXR and of FXR target genes including BSEP was reduced.14 These findings have suggested that ATP8B1 deficiency could cause FXR deregulation at hepatic and intestinal levels.

Different cell types at the hepatic level, i.e., hepatocytes and cholangiocytes (both in bile ducts and in the gallbladder), contribute to bile secretion. Bile is initially formed by an osmotic process in hepatocytes. In bile ducts and in gallbladder, bile undergoes modifications including concentration, dilution, and alkalinization upon activation of the chloride channel cystic fibrosis transmembrane conductance regulator (CFTR), while bile salts may be reabsorbed through ASBT.15, 16 We have also recently shown that FXR and its target gene SHP, represent transcriptional gene regulators in hepatocytes as well as cholangiocytes (both in bile ducts and in the gallbladder).17

To better define at which level ATP8B1 contributes to the regulation of bile secretion, we examined its expression in human hepatocytes, bile duct and gallbladder epithelial cells. To determine whether ATP8B1 deficiency may affect biliary functions in hepatocytes or in cholangiocytes, we examined the expression of genes involved in different aspects of bile secretion: (1) in the liver of patients with PFIC1 as compared to a normal liver and to the liver of patients with other cholestatic liver diseases (PFIC2 and biliary atresia), and (2) in a biliary epithelial cell line in which ATP8B1 endogenous expression was invalidated by small interfering RNA (siRNA).

Abbreviations

PFIC, progressive familial intrahepatic cholestasis; BSEP, bile salt export pump; FXR, farnesoid X receptor; ASBT, apical sodium-dependent bile salt transporter; SHP, small heterodimer partner; CFTR, cystic fibrosis transmembrane conductance regulator; siRNA, small interfering RNA; RT-PCR, reverse transcription polymerase chain reaction; NTCP, Na+-taurocholate cotransporting polypeptide; CYP7A1, cytochrome P450 7A1.

Patients and Methods

Patients' Data.

Liver samples were collected at the time of liver transplantation, from two patients with PFIC1, and two patients with other cholestatic liver diseases (PFIC2 and biliary atresia). Informed consent was obtained from the patients' families, in accordance with the guidelines of local ethical committee. Clinical and biochemical data of the patients at the time of liver transplantation are shown in Table 1. Both patients with PFIC1 have been presented in a previous report.8 Patient 1 was homozygous for a splice donor site mutation in the ATP8B1 gene. This mutation, within the 3′ exon/intron boundary of exon 18, is predicted to cause an in-frame deletion of amino acids 645-699 in the cytoplasmic domain of ATP8B1 protein.3, 8 Patient 2 was compound heterozygous for a missense mutation in exon 13 and a nonsense mutation in exon 20. These mutations cause an amino acid substitution (G446R) and truncation (Y757X), respectively, in the cytoplasmic domain of the protein.8 Both patients presented with failure to thrive. They developed cholestatic jaundice at ages 4 and 2 months and received liver transplantation at ages 5.5 and 4 years, in patients 1 and 2, respectively. In both cases, liver transplantation was indicated for severe cholestasis. Histology of the explanted liver showed extensive fibrosis without ductopenia or ductular proliferation in patient 1, and cirrhosis with little ductular proliferation in patient 2. Extrahepatic manifestations, present in both patients, consisted in chronic diarrhea and sensorineural deafness in patient 2, as reported.8 In addition, since the first report, patient 1 has also developed sensorineural deafness and both patients have developed pancreatic dysfunction, as attested by serum amylase concentrations of 405 and 135 IU/L (N100), and elastase concentrations in stools of 138 and 180 μg/g (N200), in patients 1 and 2, respectively. Because pancreatic dysfunction may be caused by genetic defects in the CFTR gene, a screening for the 30 most common CFTR mutations (INNO-LiPA-CFTR set, Innogenetics, Gent, Belgium) was performed and was negative in both patients. In patient 2, a sweat chloride test was also performed and was normal (50 mEq/L, normal <60 mEq/L).

Table 1. Clinical and Biochemical Data of the Patients at the Time of Liver Transplantation
No.DiagnosisGenderAgeSerumBile
γGT (U/L)ALP (U/L)Bilirubin (μmol/L)BS (mmol/L)
  1. NOTE. Normal values: Serum γ-glutamyl transpeptidase (γGT) < 45 U/L, alkaline phosphatase (ALP) < 360, bilirubin < 17 μmol/L; bile salt (BS) concentration in bile > 10 mmol/L.

  2. Abbreviations: M, male; F, female; ND, not determined.

1PFIC1M5.5 years4977678ND
2PFIC1M4 years225061183
3PFIC2M4.4 years571831570.02
4Biliary AtresiaF5 months36916510ND

The patient with PFIC2 (patient 3) whose case has been previously reported18 was homozygous for a nonsense mutation (R1090X) in the ABCB11 gene. He presented with persistent jaundice at age 1-2 months and received liver transplantation for altered hepatic function at age 4.4 years. Histology of the explanted liver showed cirrhosis without ductular proliferation. In patient 4, neonatal cholestasis and polysplenia syndrome were evocative of biliary atresia, a diagnosis which was confirmed by cholangiography during laparotomy at age 3 months. This patient underwent liver transplantation for liver failure at age 5 months. Histology of the liver showed cirrhosis with intense ductular proliferation.

Liver samples defined as normal were obtained from patients who underwent liver surgery for focal lesions and whose liver function tests were normal. They were taken at a distance (>3 cm) from focal lesions and displayed no significant histological abnormality.

Cell Isolation.

Hepatocytes and intrahepatic bile ducts were isolated from samples of normal human liver as defined above, using procedures that we previously described.17 More than 90% of isolated bile duct cells were cholangiocytes as ascertained by cytokeratin 19 immunolabeling and γ-GT cytochemical staining17, 19 Gallbladder epithelial cells were isolated as described.17, 20, 21

Reverse Transcription PCR.

Total RNA was extracted from liver tissue or from cell homogenates using RNA Plus reagent (Qbiogene, Illkirch, France). Complementary DNA was prepared from 1 to 2 μg total RNA, using pd(N)6 primers (Amersham Biosciences, Piscataway, NJ) and the Moloney murine leukemia virus reverse transcriptase (Invitrogen, Cergy-Pontoise, France).

The Na+-taurocholate cotransporting polypeptide (NTCP), BSEP, CFTR, ATP8B1 and β-actin transcripts were detected by conventional PCR, using the following primers : 5′-TCCTGGTTCTCATTCCTTGC-3′ (forward) and 5′-TATGCCAATGAGGAGAAGCC-3′ (reverse) to amplify a 419-bp fragment of NTCP cDNA; 5′-ACAGACCAGGATGTTGACAGG-3′ (forward) and 5′-GCTCTTCCAAGAGCTGTTGC-3′ (reverse) to amplify a 351-bp fragment of BSEP cDNA; 5′-AACTGCTGAACGAGAGGAGC-3′ (forward) and 5′-TTGACTATTGCCAGGAAGCC-3′ (reverse) to amplify a 367-pb fragment of CFTR cDNA; 5′-TTCAATGGCTACTCTGCGC-3′ (forward) and 5′-AGCACACAGCAACAGTCAGG-3′ (reverse) to amplify a 559-pb fragment of ATP8B1 cDNA; 5′-CCTCATGAAGATCCTCACCG-3′ (forward) and 5′-CAGTGATCTCCTTCTGCATCC-3′ (reverse) to amplify a 660-pb fragment of β-actin cDNA. PCR was performed using puRe Taq Ready-To-Go PCR beads (Amersham Biosciences). PCR products obtained after completion of 30 cycles were separated by electrophoresis through a 2% agarose gel stained with ethidium bromide.

Quantitative real-time PCR was performed using the Sybr Green PCR Core Reagents Kit (PE Applied Biosystems, Foster City, CA) on a ABI SDS 7000 Sequence Detector (Applied Biosystems), to measure ASBT, BSEP, CFTR, cytochrome P450 7A1 (CYP7A1), ATP8B1, FXR, NTCP and SHP transcripts. Calibrated human total RNA (Applied Biosystems, Warrington, UK) was used to standardize 18S transcripts. Primers were designed using published human cDNA sequences in Genbank database with Primer Express software v1.5 (Applied Biosystems) (Table 2). All reactions were run in duplicate, using 200 nmol/L of each target forward and reverse primers and 50 nmol/L of each 18S forward and reverse primers. The amplification conditions were the following: 2 minutes at 50°C, 10 minutes at 95°C, 40 cycles at 95°C for 15 s and 60°C for 1 minute. Data were collected, analyzed with Sequence Detector v1.7 software (Applied Biosystems) and expressed as a copy number reported to the amount of 18S RNA or as a relative amount (2−ΔΔCT), as reported.22

Table 2. Primers Used in Real-Time PCR
GeneForward primer (5′-3′)Reverse primer (5′-3′)Genbank Database Accession No.
18SGAGCGAAAGCATTTGCCAAGGGCATCGTTTATGGTCGGAAM10098
ASBT (SLC10A2)GGGACCAGCGCTTCTGTGGATGCACCAGAGCAAACTGTTGNM_000452
BSEP (ABCB11)GAAAGGAGAGGCGGTTCATTGTAAACATGATGCACTGGGCAANM_003742
CFTR (ABCC7)CCATCAGCCCCTCCGACAAAGCCTTGTATCTTGCACCTCTNM_000492
CYP7A1TGTCCTGGAAGATTGTTCGCTGGACATTTAGCTTGGCCCTCTNM_000780
ATP8B1GCGCAGACTGCATACGAGGTCACTCACATCCTGGTCGAGNM_005603
FXR (NR1H4)GCGACTGAGAAAAAATGTGAAGCTGCATGACTTTGTTGTCGAGGNM_005123
NTCP (SLC10A1)CAGTGTGGCCGTCACAGTTCTAAAAGGCATCAGGGAGGAGGL21893
SHPAAGGTAACACAGCCAGGTAGGGGGTCAGAGCAAATTTCTCAGGCAB058644

Immunohistochemistry.

BSEP immunolabeling was performed on liver tissue cryosections using an anti-BSEP polyclonal antibody, and a three-step immunoperoxidase method, as reported.18 Cytokeratin 19 immunolabeling was performed to visualize bile ducts on paraffin-embedded liver tissue sections using an anti-cytokeratin 19 monoclonal antibody (Amersham Life Science, Little Chalfont, UK), and an avidin-biotin-peroxidase method, as described.19

Transient Transfection of Cell Lines With siRNA.

The human biliary epithelial cell line Mz-ChA-223 was maintained in DMEM with 1 g/L glucose, 10 mmol/L Hepes and 10% fetal bovine serum, under 95% air and 5% CO2 at 37°C. The cells were transfected with a synthetic ATP8B1 siRNA or with a silencer negative control RNA, using PolyFect Transfection Reagent (Qiagen, Valencia, CA). ATP8B1 siRNA (Dharmacon Research Inc., Lafayette, CO) consisted of a 21-nucleotide sequence (5′-AAGCAAACGAUCGCAAGUAdTdT-3′), which corresponded to position 201-218. The silencer negative control consisted of a 19-nucleotide scrambled sequence (Ambion, Cambridgeshire, UK).

Statistical Analysis.

Comparisons were performed using one way anova. P values less than .05 were considered as significant.

Results

Expression Pattern of ATP8B1 Gene in Human Hepatobiliary Cells.

To better define ATP8B1 contribution to bile formation, we first examined its expression in the different cell types which cooperate to ensure bile production in the human liver and biliary tract. The expression of ATP8B1 along with that of NTCP/BSEP and CFTR transporters, which are specifically expressed in hepatocytes and cholangiocytes, respectively, was analyzed by RT-PCR in hepatocyte, intrahepatic bile duct and gallbladder epithelial cell preparations, issued from the same donors. The purity of the cell preparations was ascertained by the fact that NTCP and BSEP transcripts were detected only in hepatocyte preparations. CFTR transcripts were detected in bile duct and gallbladder epithelial cells and not in hepatocytes (Fig. 1A). ATP8B1 expression was detected in all three cell types (Fig. 1A). The amounts of ATP8B1 transcripts, measured by quantitative real-time PCR, were 20- and 200-fold higher in bile duct and gallbladder epithelial cells, respectively, than in hepatocytes (Fig. 1B). These results indicate that gallbladder and bile duct epithelial cells (or cholangiocytes) are over hepatocytes, the predominant sites of ATP8B1 expression in the human biliary tree. They suggest that, besides hepatocytes, cholangiocytes may be a target of ATP8B1 dysfunction in patients with PFIC1.

Figure 1.

ATP8B1 expression in hepatobiliary cell types. (A) Total RNA from human hepatocytes (lane 1), intrahepatic bile duct (lane 2) and gallbladder epithelial cells (lane 3) prepared from the same subjects were submitted to conventional RT-PCR with primers designed to amplify Na+-taurocholate cotransporting polypeptide (NTCP), bile salt export pump (BSEP), cystic fibrosis transmembrane conductance regulator (CFTR), ATP8B1 or β-actin. After 30 cycles, samples of amplified cDNA were separated on ethidium bromide-stained agarose gels. Representative gels of cell preparations obtained from three different subjects are shown; (B) Quantification of ATP8B1 mRNA by real-time RT-PCR in human hepatocytes (1), intrahepatic bile duct (2) and gallbladder epithelial cells (3), prepared from the same three subjects. Data are expressed as ATP8B1 mRNA copy number per μg of 18S RNA, and are shown on a logarithmic scale. They represent the means of three preparations analyzed in triplicate. *P < .05 vs. hepatocytes; ** P < .001 vs. hepatocytes and intrahepatic bile duct.

Expression of Genes Involved in Hepatocyte Biliary Functions in PFIC1 and Other Cholestatic Diseases.

Next, we examined the possibility that ATP8B1 defect may affect FXR regulation of bile formation in hepatocytes, and measured the expression of FXR and of FXR target genes in hepatocytes (SHP, CYP7A1, NTCP and BSEP), within the liver of patients with PFIC1 and other cholestatic diseases.

FXR, SHP and CYP7A1 gene expressions were downregulated by approximately 90% in patients with PFIC1 as compared to controls, and to a similar extent in patients with PFIC2 or biliary atresia (Fig. 2A-C). A downregulation of NTCP and BSEP expressions was also detected, but was less pronounced in PFIC1 than in PFIC2 or biliary atresia (Fig. 3A-B). NTCP mRNA levels were decreased at most by 30% in patients with PFIC1 and by 97% in patients with PFIC2 or biliary atresia (Fig. 3A). Also, whereas BSEP mRNA levels were decreased by 97% in patients with PFIC2 or biliary atresia, they were decreased by approximately 50% in one of the patients with PFIC1 (patient 1), while in patient 2, they were not different from control levels (Fig. 3B, closed and open circles, respectively).

Figure 2.

Variations in hepatic expressions of FXR and FXR target genes in PFIC1 as compared to other cholestatic diseases. The mRNA levels of (A) farnesoid X receptor (FXR), (B) small heterodimer partner (SHP), and (C) cytochrome P450 7A1 (CYP7A1), were measured by real-time RT-PCR in normal human livers from three controls, in the liver from two patients with PFIC1, a patient with PFIC2, and a patient with biliary atresia. Results are expressed as mRNA levels normalized to 18S, and relative to the mean control value. Each circle represents the result of analyses performed in triplicate in one subject. The three controls are represented by three different symbols; patients with PFIC1 1 and 2 are represented by closed and open circles, respectively.

Figure 3.

Variations in hepatic expressions of hepatocyte bile salt transporters, NTCP and BSEP, in PFIC1 as compared to other cholestatic diseases. The mRNA levels of (A) Na+-taurocholate cotransporting polypeptide (NTCP), and (B) bile salt export pump (BSEP), were measured by real-time RT-PCR in normal human livers from three controls, in the liver from two patients with PFIC1, a patient with PFIC2 and a patient with biliary atresia. Results are expressed as mRNA levels normalized to 18S, and relative to the mean control value. Each circle represents the result of analyses performed in triplicate in one subject. The three controls and the two patients with PFIC1 are represented by the same symbols as in Fig. 2.

Because BSEP mediates the rate-limiting step of bile salt transport in hepatocytes, the expression of BSEP was also examined at the protein level, by immunohistochemistry, in the liver of patients with PFIC1, as compared to PFIC2 and controls. As shown in Fig. 4, the results of protein detection were fully concordant with the mRNA data. In normal liver tissue, BSEP protein was clearly detected as a network outlining bile canaliculi (Fig. 4A). In the patient with PFIC1 whose BSEP mRNA levels were not different from normal (patient 2), BSEP was immunodetected with a similar pattern and the same intensity as in control liver (Fig. 4B). In the patient with PFIC1 whose BSEP mRNA levels were lower than in controls but higher than in PFIC2 (patient 1), BSEP immunoreactivity was detected with lower intensity than in controls (Fig. 4C), while in the patient with PFIC2, BSEP immunolabeling was very sparse (Fig. 4D).

Figure 4.

Immunohistochemical detection of BSEP in PFIC1 and PFIC2. BSEP immunolabeling was performed on cryosections of normal liver tissue (control) and of liver tissue from PFIC patients, using an anti-BSEP polyclonal antibody and an indirect immunoperoxidase method, followed by hematoxylin conterstaining. BSEP immunoreactivity outlining the network of bile canaliculi is detected with high intensity in the control (A) and in PFIC1 patient 2 (B), with lower intensity in PFIC1 patient 1 (C) and is very sparse and faint in the PFIC2 patient (D). BSEP, bile salt export pump.

Altogether, these results suggest that FXR, SHP and CYP7A1 gene expressions are downregulated to a similar extent in PFIC1 and various other cholestatic diseases, while BSEP and NTCP are less downregulated in PFIC1.

Expression of Genes Involved in Cholangiocyte Transport Functions in PFIC1 and Other Cholestatic Diseases.

The evidence of a high endogenous expression of ATP8B1 in cholangiocytes inferred that these cells may be important targets of ATP8B1 dysfunction in the liver. To address this possibility, we examined the expression of ASBT and of CFTR, two major transporters in cholangiocytes, within the liver of patients with PFIC1 and other cholestatic diseases.

ASBT mRNA levels were markedly decreased (by 96%-97%) both in patients with PFIC1 and in the patients with other cholestatic diseases (Fig. 5A). In the liver of patients with PFIC1, CFTR mRNA levels were also markedly decreased (by 80%) as compared to controls (Fig. 5B). However, as opposed to ASBT, CFTR downregulation was detected selectively in PFIC1. In the PFIC2 patient, CFTR mRNA level was within normal range and in the patient with biliary atresia, the level of CFTR mRNA level was, by contrast, increased by 39% (Fig. 5B). The decrease in CFTR expression within the liver of patients with PFIC1 was not explained by a reduction in bile duct mass, as ascertained by cholangiocyte immunolabeling with an anti-cytokeratin 19 antibody on liver tissue sections. Cytokeratin labeling revealed no ductopenia in either patient, and the presence of a mild ductular reaction in patient 2 (Fig. 6). Thus, the fact that CFTR expression was diminished selectively in patients with PFIC1, while the bile duct mass was maintained or even increased in these patients, suggested that CFTR gene expression was affected by ATP8B1 defect in cholangiocytes either directly or indirectly as a consequence of the PFIC1 phenotype.

Figure 5.

Variations in hepatic expressions of cholangiocyte transporters, ASBT and CFTR, in PFIC1 as compared to other cholestatic diseases. The mRNA levels of (A) apical sodium-dependent bile salt transporter (ASBT), and (B) cystic fibrosis transmembrane conductance regulator (CFTR), were measured by real-time RT-PCR in normal human livers from three controls, in the liver from two patients with PFIC1, a patient with PFIC2, and a patient with biliary atresia. Results are expressed as mRNA levels normalized to 18S, and relative to the mean control value. Each circle represents the result of analyses performed in triplicate in one subject. Controls and patients are represented by the same symbols as in Fig. 2.

Figure 6.

Immunhistochemical detection of bile duct sections in PFIC1. Cytokeratin 19 immunolabeling was performed to visualize bile ducts on paraffin-embedded liver tissue sections, using an anti-cytokeratin 19 monoclonal antibody and an indirect immunoperoxidase method. The presence of moderate ductular reaction in patient 2 is shown.

Effect of ATP8B1 Gene Invalidation on CFTR Expression in Human Biliary Epithelial Cells.

To further address the possibility of a link between ATP8B1 defect and a deregulation of CFTR expression, we performed an invalidation of ATP8B1 endogenous expression by siRNA in the Mz-ChA-2 human biliary epithelial cell line. In negative control experiments, Mz-ChA-2 cells transfected with a scrambled siRNA showed no change in ATP8B1 expression, and a statistically nonsignificant increase in CFTR expression (Fig. 7A-B). By contrast, after transfection of Mz-ChA-2 cells with ATP8B1 siRNA, the level of ATP8B1 mRNA was reduced by 40% (P < .05, Fig. 7A), and following ATP8B1 invalidation in these cells, the levels of CFTR transcripts were significantly reduced (50% as compared to controls, P < .05, Fig. 7B), providing further evidence that specific inhibition of ATP8B1 gene expression is associated with a downregulation of CFTR expression in biliary epithelial cells.

Figure 7.

Effect of ATP8B1 invalidation on CFTR expression in human biliary epithelial cells. Mz-ChA-2 cells were transfected with a scrambled siRNA (si Control) or with ATP8B1 siRNA (si ATP8B1) for 24 hours. The mRNA levels of (A) ATP8B1 and (B) CFTR were measured by real-time RT-PCR in these cells. Results are expressed as mRNA levels normalized to 18S, and relative to the mean control value. Bar graphs represent means ± SEM of 3 experiments performed in duplicate. *P < .05 vs. other conditions.

Discussion

The molecular mechanisms underlying disease phenotypes linked to ATP8B1 gene mutations are poorly understood. More specifically, it remains to better define: (1) which cell types in the organism express ATP8B1 and may thus represent direct targets of ATP8B1 deficiency in human disease, and (2) what are the functions of ATP8B1 gene in these cells and what are the consequences of ATP8B1 defect on their physiology. Previous studies have outlined putative functions of ATP8B1 in hepatocytes and in enterocytes.5, 13, 14 In the present study, we provide evidence that cholangiocytes are a site of high endogenous expression of ATP8B1 in the liver and biliary tract. Furthermore, the ATP8B1 defect was associated with a downregulation in the expression of CFTR, a key player in the physiology of these cells.

Based on quantitative analyses, we were able to show that ATP8B1 transcripts were by more than one and two orders of magnitude more abundant in bile duct and gallbladder epithelial cells, respectively, than in hepatocytes. This indicates that within the liver, cholangiocytes are the predominant site of ATP8B1 endogenous expression, an observation in line with personal immunohistochemical data showing that on rat liver tissue sections, ATP8B1 immunolabeling is detected predominantly in bile ducts. This finding also implies that cholangiocytes should be given at least as much attention as hepatocytes, when considering the consequences of the ATP8B1 defect on the liver.

Strong evidence indicates that in PFIC1, bile salt transport is altered in the liver, as attested by low concentrations of bile salts in bile,24–26 and in the intestine, as suggested by the frequent exacerbation of diarrhea following the restoration of normal bile salt secretion by liver transplantation.8, 27 Recently, two groups reported on decreased FXR mRNA levels in the liver and intestine of patients with PFIC1, with subsequent downregulation of BSEP mRNA in the liver, and upregulation of ASBT mRNA in the intestine.13, 14 Both groups have suggested that impaired ATP8B1 function resulted in reduced FXR expression and/or activity with subsequent impairment of canalicular and ileal bile salt transport, which would be causative for the cholestatic phenotype. Yet, in the present study, patients with PFIC1 as well as patients with other types of cholestatic diseases (PFIC2 and biliary atresia) showed reduced hepatic expression of FXR, suggesting that the drop in FXR levels is secondary to cholestasis. In support of this view, studies in the Atp8b1G308V/G308V mutant mouse, the mouse model for PFIC1, did not reveal a significant change in FXR expression levels unless serum bile salts were increased by bile salt feeding.28

To further examine the possibility that ATP8B1 impairment could reduce the transcriptional activity of FXR in the liver, we compared the hepatic mRNA levels of FXR target genes in PFIC1 to those in other conditions. The mRNA levels of SHP, a direct target gene of FXR transcriptional activity both in hepatocytes and in cholangiocytes,17, 29, 30 were equally decreased in PFIC1, PFIC2 and biliary atresia, and thus mirrored the changes in FXR expression in all three cholestatic diseases. CYP7A1, the key enzyme of bile salt synthesis, NTCP, the basolateral bile salt transporter in hepatocytes, and ASBT, the apical bile salt transporter expressed in cholangiocytes, are targets of FXR-induced repression through SHP-mediated mechanism.12, 29, 30 Despite decreased SHP expression, CYP7A1 and NTCP genes were downregulated in all the patients investigated here, which may be attributable to residual SHP activity or, as previously demonstrated in Shp knockout mice, to additional pathways involved in the negative feedback of these genes.31 However, while CYP7A1 and ASBT were downregulated to a similar extent in all three types of cholestatic diseases, NTCP was downregulated to a lesser extent in PFIC1 than in PFIC2 or biliary atresia. Conversely, the mRNA levels of BSEP, a direct FXR target gene in hepatocytes,11 were maintained at higher levels in PFIC1 than in the other cholestatic diseases. While we cannot exclude altered transcriptional activity of the FXR/SHP pathway in PFIC1, the circulating bile salt pool composition, which plays a critical role in determining the nuclear activation state of FXR,32 undergoes changes which may be less dramatic in PFIC1 than in PFIC2 or biliary atresia, and may account for these differences. Thus, it was previously demonstrated that chenodeoxycholate and lithocholate acting through FXR cause upregulation and downregulation of BSEP expression, respectively.33 The concentrations of bile salts in bile, available in the present study from one patient with PFIC1 (patient 2) and from the patient with PFIC2 (3 and 0.02 mmol/L, respectively), are consistent with a less severe defect of bile salt secretion in PFIC1 than in PFIC2. Of additional interest, the defect of bile salt secretion in patient 2, contrasted with normal BSEP mRNA levels and immunolabeling in this patient, supports the hypothesis that ATP8B1 loss may impair BSEP activity itself. According to this hypothesis, the absence of ATP8B1 putative translocase activity would impede the bilayer lipid asymmetry required for the activity of membrane-associated proteins, such as BSEP at the canalicular membrane of hepatocytes, or ASBT at the apical membrane of enterocytes.34

An important new observation in this study is that CFTR, a gene expressed exclusively in cholangiocytes within the liver,35, 36 was downregulated selectively in patients with PFIC1. In the patient with biliary atresia, intense ductular reaction was associated with increased hepatic mRNA levels of CFTR, as previously reported in experimental bile duct obstruction.37 In the patient with PFIC2 who had no ductular reaction, CFTR mRNA levels were unchanged. By contrast, CFTR mRNA levels were profoundly decreased in both patients with PFIC1, irrespective of whether they had or not bile duct proliferation. Because low CFTR expression in these patients was not explained by bile duct loss or by CFTR mutations, as verified by genotyping, we raised the possibility that CFTR downregulation might be linked to the ATP8B1 defect in cholangiocytes. We found that in cultures of human biliary epithelial cells, partial inhibition of the ATP8B1 gene was associated with decreased CFTR expression. Because ATP8B1 may function as a translocase, transferring phosphatidylserine from the outer to the inner hemileaflet of cell membranes, one may hypothesize that ATP8B1 defect affects the activation of regulatory proteins, like PKC isoforms, which translocate to cell membranes and bind to phosphatidylserine upon activation.38 It is of particular interest that the same (classical) isoforms of PKC have the potential to regulate CFTR expression.39 Alternatively, ATP8B1 defect could affect the transcriptional activity not only of FXR,13 but also of other transcription factors, like HNF1, which regulate the tissue-specific expression of CFTR.40 Because CFTR is a major determinant of hydroelectrolytic secretion in the biliary tract,20, 41–43 CFTR downregulation may contribute to impaired bile secretion through bile inspissation in PFIC1. This would explain the typical ultrastructural abnormality of bile in PFIC1 termed “Byler bile”, characterized by a coarsely granular appearance.7, 26 ATP8B1-related CFTR downregulation in extrahepatic epithelia could also explain that diarrhea, pancreatitis, and in some cases, elevated sweat chloride concentrations, that are common manifestations in cystic fibrosis, may also occur in PFIC1.7, 8, 27, 44, 45

We conclude from this study that a downregulation of CFTR expression in patients with PFIC1 might be associated with a defect in the ATP8B1 gene. This may impair the contribution of cholangiocytes to bile secretion and also potentially explain some of the extrahepatic manifestations in PFIC1 and related disorders. The gallbladder, a site of high endogenous expression of ATP8B1 and of CFTR,20, 21, 42 fully takes part to the regulation of the bile acid pool,46 and thus may be of particular importance in the pathogenenesis of PFIC1.

Acknowledgements

We thank Dr. Peter L. M. Jansen (University Hospital Groningen, Groningen, The Netherlands) and Dr. Monique Fabre (Hôpital Bicêtre, service anatomie pathologique, Paris, France) for providing BSEP and CK19 immunohistochemical data, respectively; Drs. Dale Leitman (University of California, San Francisco, CA), Christian Hulen and Claire Dunois-Lardé for their helpful input; Noura Mebirouk, Mélanie Levasseur and Olivier Giraudier for their technical assistance.

Ancillary