Expression and localization of hepatobiliary transport proteins in progressive familial intrahepatic cholestasis


  • Verena Keitel,

    1. Klinik für Gastroenterologie, Hepatologie und Infektiologie, Heinrich Heine Universität, Düsseldorf, Germany
    Search for more papers by this author
  • Martin Burdelski,

    1. Pädiatrische Gastroenterologie und Hepatologie, Universitätsklinikum Hamburg-Eppendorf, Germany
    Search for more papers by this author
  • Ulrich Warskulat,

    1. Klinik für Gastroenterologie, Hepatologie und Infektiologie, Heinrich Heine Universität, Düsseldorf, Germany
    Search for more papers by this author
  • Thomas Kühlkamp,

    1. Klinik für Gastroenterologie, Hepatologie und Infektiologie, Heinrich Heine Universität, Düsseldorf, Germany
    Search for more papers by this author
  • Dietrich Keppler,

    1. Abteilung Tumorbiochemie, Deutsches Krebsforschungszentrum Heidelberg, Germany
    Search for more papers by this author
  • Dieter Häussinger,

    1. Klinik für Gastroenterologie, Hepatologie und Infektiologie, Heinrich Heine Universität, Düsseldorf, Germany
    Search for more papers by this author
  • Ralf Kubitz

    Corresponding author
    1. Klinik für Gastroenterologie, Hepatologie und Infektiologie, Heinrich Heine Universität, Düsseldorf, Germany
    2. Pädiatrische Gastroenterologie und Hepatologie, Universitätsklinikum Hamburg-Eppendorf, Germany
    3. Abteilung Tumorbiochemie, Deutsches Krebsforschungszentrum Heidelberg, Germany
    • Klinik für Gastroenterologie, Hepatologie und Infektiologie, Universität Düsseldorf, Moorenstrasse 5, D-40225 Düsseldorf, Germany
    Search for more papers by this author
    • fax: (49)211-8118838

  • Potential conflict of interest: Nothing to report.


Mutations of the bile salt export pump (BSEP) or the multidrug resistance P-glycoprotein 3 (MDR3) are linked to impaired bile salt homeostasis and lead to progressive familial intrahepatic cholestasis (PFIC)-2 and -3, respectively. The regulation of bile salt transporters in PFIC is not known. Expression of hepatobiliary transporters in livers of ten patients with a PFIC phenotype was studied by quantitative reverse transcription polymerase chain reaction, Western blotting, and immunofluorescence microscopy. PFIC was diagnosed by clinical and laboratory findings. All patients could be assigned to PFIC-2 or PFIC-3 by the use of BSEP- and MDR3-specific antibodies and by MDR3 gene-sequencing. Whereas in all PFIC-2 patients, BSEP immunoreactivity was absent from the canalicular membrane, in three PFIC-3 livers, canalicular MDR3 immunoreactivity was detectable. Serum bile salts were elevated to 276 ± 233 and to 221 ± 109 μmol/L in PFIC-2 and PFIC-3, respectively. Organic anion transporting polypeptide OATP1B1, OATP1B3, and MRP2 mRNA and protein levels were reduced, whereas sodium taurocholate cotransporting polypeptide (NTCP) was only reduced at the protein level, suggesting a posttranscriptional NTCP regulation. Whereas MRP3 mRNA and protein were not significantly altered, MRP4 messenger RNA and protein were significantly increased in PFIC. In conclusion, PFIC-2 may be reliably diagnosed by immunofluorescence, whereas the diagnosis of PFIC-3 requires gene-sequencing. Several mechanisms may contribute to elevated plasma bile salts in PFIC: reduced bile salt uptake via NTCP, OATP1B1, and OATP1B3, decreased BSEP-dependent secretion into bile, and increased transport back into plasma by MRP4. Upregulation of MRP4, but not of MRP3, might represent an important escape mechanism for bile salt extrusion in PFIC. Supplementary material for this article can be found on the HEPATOLOGY website ( (HEPATOLOGY 2005;41:1160–1172.)

Bile salt homeostasis is achieved by the coordinate action of bile salt transporters. At the sinusoidal membrane the Na+-taurocholate cotransporting polypeptide (NTCP, SLC10A1) mediates sodium-dependent bile salt uptake,1 whereas sodium-independent bile salt uptake is achieved by organic anion transporting polypeptide OATP1B1 (OATP-C/OATP2/SLC21A6)2, 3 and OATP1B3 (OATP8/SLC21A8).4 At the canalicular membrane, bile salts are actively secreted from hepatocytes by different members of the adenosine triphosphate–binding-cassette (ABC) transporter superfamily. Under normal conditions, the bile salt export pump BSEP (ABCB11) mediates the canalicular secretion of taurine- and glycine-conjugated bile salts.5 Besides the secretion of bile salts by BSEP, the multidrug resistance associated protein 2 (MRP2, cMOAT, ABCC2)6, 7 transports bile salts after they have been sulfated or glucuronidated.8

Several transporters from the ABCC subfamily are localized at the sinusoidal membrane of hepatocytes, where they mediate efflux of substrates back into blood.9, 10 These ABC-transporters include MRP3 (ABCC3)11.12 and MRP4 (ABCC4).10 MRP3 transports bile salts such as taurocholate, glycocholate, taurochenodeoxycholate-3-sulfate, and taurolithocholate-3-sulfate.11, 13 In Dubin-Johnson patients with defective MRP2, the MRP3 protein is upregulated.9 Similarly, in rats with Mrp2-deficiency14 or with obstructive cholestasis,15 an increased expression of Mrp3 was observed. Mrp3 upregulation was suggested to represent a protective mechanism16 against toxic effects of certain bile salts. Human MRP4 cotransports monoanionic bile salts together with glutathione,10 and sulfated bile salts were shown to competitively inhibit transport of estradiol 17-beta-D-glucuronide by MRP4,17 indicating that more than one possible escape route for bile salts across the sinusoidal membrane may exist.

Genetic abnormalities of BSEP are the cause of progressive familial intrahepatic cholestasis type 2 (PFIC-2).18 A similar form of cholestasis is observed in patients with mutations of the familial intrahepatic cholestasis (FIC) 1-gene product (ATP8B1), leading to PFIC-1 or the less severe form termed benign recurrent intrahepatic cholestasis 1 (BRIC-1). Both forms of inherited cholestasis, PFIC-1 and -2, are characterized by low gamma-glutamyltransferase (γGT) serum activities.19 In contrast, mutations of the multidrug resistance P-glycoprotein 3 (MDR3), which acts as a phospholipid flippase,20 lead to a third form of severe cholestasis during childhood (PFIC-3), characterized by high γGT levels.21 Increased γGT levels are attributed to a low phosphatidylcholine concentration in bile, which normally protects the biliary epithelium from bile salt toxicity.22

Human studies should help to better understand the pathophysiology of human cholestatic conditions as a prerequisite for a more directed treatment. Therefore, this study investigated naturally occurring “knock-outs” in humans. We were interested in how the impairment of a single transporter affects other transport systems, and we aimed to identify changes in transporter expression that may account for increased bile salt levels observed in PFIC.


OATP, organic anion transporting polypeptide; ABC, adenosine triphosphate–binding cassette; NTCP, sodium taurocholate cotransporting polypeptide; BSEP, bile salt export pump; MRP, multidrug resistance associated protein; PFIC, progressive familial intrahepatic cholestasis; BRIC, benign recurrent intrahepatic cholestasis; γGT, gamma-glutamyltransferase; MDR, multidrug resistance; HPRT1, hypoxanthine phosphoribosyltransferase 1; mRNA, messenger RNA.

Materials and Methods


A rabbit polyclonal antibody (K168) was raised against a 15–amino acid oligopeptide containing amino acids 688 to 702 from the linker region of human BSEP, which comprise only 3 and 2 identical amino acids of the corresponding sequences of human MDR1 and MDR3, respectively, the two most related proteins.

The rabbit anti-human BSEP K165 antibody was raised against the N-terminus and purified as described recently.23 The anti-human antibodies against MRP2 (EAG5), MRP4 (SNG), OATP1B1 (ESL), and OATP1B3 (SKT) were raised in rabbits.4, 6, 10 The rabbit anti-human NTCP antibody (K9) was donated by Drs. B. Stieger and P. Meier (Kantonsspital, Zürich, Switzerland).24 The mouse monoclonal antibodies M2III6 and M2I4, P3II26, and M3II9 against human MRP2, MDR3, and MRP3, respectively, were from Alexis (Grünberg, Germany).

Sampling of Liver Tissues.

Small liver pieces from 10 children with inherited cholestatic liver diseases were collected between 1994 and 2003 when children underwent liver transplantation. Laboratory parameters were determined shortly before transplantation. All children were treated with ursodeoxycholic acid at the time of transplantation. Liver samples were kept at −70°C. They were partly cut by cryosectioning and partly homogenized for protein or RNA preparation. One PFIC-2 sample (C4) was not suitable for RNA extraction, and another PFIC-2 sample (C1) was too small for Western blotting. Serological and clinical parameters of the patients are summarized in Table 1.

Table 1. Clinical and Laboratory Parameters of PFIC Patients
PatientsClinical ParametersLaboratory ParametersDiagnosis
Onset of SymptomsAge at LTxBS (<8 μM)γGT (<25 U/l)Bilirubin (<1 mg/dL)MDR3-MutationMDR3-IFBSEP-IFPFIC Type
  • NOTE. Data of 10 patients with a PFIC phenotype are presented.

  • Abbreviations: LTx, liver transplantation; BS, plasma bile salt concentrations; γGT, γ-glutamyltransferase (normal values in brackets; values were obtained at the time of liver transplantation); IF, immunofluorescence; y, years; m, months.

  • *

    Patient described in Jacquemin et al.39

  • Patient described in De Vree et al.21

C1F.A.4 m3 y 5 m4771.3None detected+2
C2A.A.3 m2 y 5 m1441922.6None detected+2
C3S.H.3 m3 y 3 m6167722.7None detected+2
C4M.D.6 m10 y 7 m16075.4None detected+2
C5T.A.6 m4 y 2 m414115.4555 A → G/T175A Heterozygous+2
C6M.D.4 m7 y 9 m1632611.21069 G → T/S346I++3
C7S.F.birth8 y 7 m2992501.91069 G → T/8346I*++3
C8K.Y.6 m3 y 10 m852770.52890 C → A/A953D++3
C9K.B.4 m3 y 8 m3601651.1426–432 del+3
C10B.D.7 y11 y 4 m1992420.7None detected+3

A first control group (control I) was established from organ donors when such livers were not completely used for transplantation. Liver pieces were snap-frozen in liquid nitrogen and stored at −70°C. A second group of controls (control II) was collected from six patients (with normal bilirubin levels) who had a liver needle biopsy for diagnostic reasons. The liver diseases of these patients are summarized in Table 2. All patients had given their written consent, and studies on excess material were approved by the local ethics committee. Samples were snap-frozen in liquid nitrogen and were subjected to RNA preparation, when liver histology showed no, minimal, or mild fibrosis.

Table 2. Clinical and Laboratory Data and Histology of Control Group II
SampleDiagnosisAgeSexGPT (<23 U/L)γGT (<18 U/L)Bilirubin (<1.1 mg/dL)Histology (° Fibrosis)
  1. NOTE. Histology revealed no (0), minimal (I°), or mild (II°) fibrosis.

  2. Abbreviations: AIH, autoimmune hepatitis; C2, alcohol; HBV, hepatitis B virus infection; TA, transaminases; NASH, nonalcoholic steatohepatitis; ALT, alanine amino transferase; γGT, γ-glutamyltransferase (normal values in brackets).

2C2 abuse56F3150.21
4C2 abuse61F261890.41
5Elevated TA33M43350.48

Sequence Analysis of MDR3.

Sequencing of MDR3 (Genbank accession NM_000443) was performed at the genomic level by the use of DNA from lymphocytes and specific primers, which enclosed all exons and exon/intron boundaries of MDR3. For primer sequences, please refer to the supplemental data file.

Quantitative Reverse Transcription Polymerase Chain Reaction.

Total RNA was isolated using the RNA extraction kit (Qiagen, Hilden, Germany), and complementary DNA was obtained with the first strand complementary DNA synthesis kit (Roche, Mannheim, Germany). The levels of gene expression were measured by real-time SYBR Green PCR with the Gene Amp 5700 Sequence Detection System (Applied Biosystems, Foster City, CA) according to the manufacturer's instructions. Primers for transporter and housekeeping genes are summarized in Table 3. For each gene, the number of cycles was determined until fluorescence reached a threshold value within the exponential increase in fluorescence. Data were produced in triplicates for each gene. Mean values of cycle numbers of each transporter gene were subtracted from the mean of cycle numbers of the housekeeping gene (HPRT1) for the respective patient's sample. This value taken to the power of 2 is the mRNA expression of a gene in relation to HRPT1 expression, termed “relative mRNA unit” (RRU) of the respective gene.

Table 3. RT-PCR Primers Used in This Study
GeneProteinForward PrimerReverse Primer
  1. NOTE. Nucleotide sequences of the forward and reverse primers, which were used for quantitative RT-PCR measurements.

ABCB1MDR1, P-gpggcaaagaastaaagcgactgaagcccaggtgtgctcgg
ABCB11BSEP, SPGPactagatgaagccacttctgccttatgcaccgtcttttcactttctgt
ABCC2MRP2, cMRP, cMOATggacagtgacaaggtaatggtcctgccgcactctataatcttccc
SLCO1B1, SLC21A6OATP1B1, OATP2tgaaatcacttgcactgggtttctcctagtgctcgtataaccattgagt
SLCO1B3, SLC21A8OATP1B3, OATP8ttccctctaatctgcgaaagcaggtcaaggttaggccgg
OATP1A2, SLC21A3OATP1A2, OATP-Acagttcttgggctgtgcgactggagcatcaaggaacagtc
SLCO2B1, SLC21A9OATP2B1, OATP-Battggacggctttaaccctgttgtattccacacgagtgctgg

Immunofluorescence and Confocal Laser Scanning Microscopy.

Immunofluorescence of tissue sections was performed as recently described.23 The primary antibodies were diluted as follows: EAG5 at 1:150; SNG, M2III6, M2I4, P3II26, K168, and K165 at 1:25; K9 at 1:200; M3II9 at 1:10; ESL and SKT at 1:100. Fluorescein or Cyanin 3 conjugated secondary antibodies (Jackson Immuno Research Laboratories, West Grove, PA) were diluted 1:100 and 1:500, respectively. Immunostained liver-samples were analyzed on a Zeiss LSM 510 META confocal system mounted on an Axioscope 100 M inverted microscope (Zeiss, Oberkochen, Germany). To compare the immunofluorescence intensity of different livers, instrument settings were constant along the complete series of livers.

Western Blot Analysis and Densitometry.

For Western blot analysis, human liver pieces were homogenized in hypotonic buffer (0.1 mmol/L EDTA, 0.5 mmol/ sodium phosphate, pH 7.0) supplemented with protease inhibitors using a tight-fitting potter (glas/teflon, 10 strokes, 1,000 rpm). After centrifugation (100.000g, 1 hour, 4°C), pellets were resuspended in hypotonic buffer. These crude membrane fractions were frozen and stored at −20°C. Protein concentrations were measured in triplicates by the Advanced Protein Assay (Cytoskeleton, Denver, CO). Equal amounts of proteins were separated by SDS-PAGE and blotted on nitrocellulose membranes. Proteins were detected by specific antibodies and by the use of the enhanced chemiluminescence detection kit (Amersham-Pharmacia, Freiburg, Germany). Densitometry was performed with a Kodak Image Station 440 CF and the Kodak 1D 3.5 Software.


Values from reverse transcription polymerase chain reaction and densitometric analysis are given as means ± standard deviations. The Mann-Whitney U test (Wilcoxon test) was used for statistical analysis, with a P value less than .05 considered statistically significant.


Diagnosis of Low- and High-γGT PFIC.

Diagnosis of progressive familial intrahepatic cholestasis was based on clinical and laboratory parameters. Onset of symptoms of cholestasis was observed within the first 6 months of life in nine of ten patients (Table 1). Progression toward liver cirrhosis necessitating liver transplantation was observed at 5.9 ± 3.3 years of age (Table 1). According to serum γGT activities, children were grouped into low- and high-γGT cholestasis. Six children (C1, C2, C4, C5, C6, and C10) had normal or near normal γGT activities, and 4 children (C3, C7, C8, and C9) had elevated γGT activities (Table 1).

BSEP expression and localization in PFIC livers were investigated using the K165 and K168 antibodies, which were directed against different epitopes of human BSEP. K168 but not its pre-immune serum resulted in a typical canalicular staining pattern. The canalicular immunoreactivity could be blocked by the peptide used for antibody generation, confirming the specificity of K168 (see images from the supplemental data file at the HEPATOLOGY website: Canalicular BSEP expression was considered negative, when both BSEP antibodies did not show any canalicular immunoreactivity despite a canalicular expression pattern of MRP2.

In four of six patients with low-γGT cholestasis (C1, C2, C4, and C5), no canalicular BSEP immunoreactivity was detected, suggesting impaired BSEP protein synthesis or targeting in line with a PFIC-2 phenotype (Fig. 1). In a fifth patient with high-γGT cholestasis (C3), BSEP immunoreactivity was absent despite clear MRP2 staining (Fig. 1). Therefore, this patient was also diagnosed as PFIC-2. In line with this diagnosis, no MDR3 mutation was detected by sequencing, and canalicular immunoreactivity to the MDR3 antibody was present in this patient as in all other PFIC-2 patients (Fig. 2).

Figure 1.

BSEP expression in PFIC-2 and PFIC-3 livers. PFIC liver sections were stained with the BSEP antibody K165 (red) and the MRP2 antibody M2III6 (green). Livers C1, C2, C3, C4, and C5 displayed no BSEP immunoreactivity despite a canalicular MRP2 staining pattern and were therefore diagnosed as PFIC-2. Livers C6, C7, C8, C9, and C10 were diagnosed as PFIC-3 due to MDR3 gene-mutations or absence of MDR3 expression. They showed colocalization of BSEP and MRP2 similar to control livers (representative control liver in upper panel, K168 BSEP antibody). Bars = 10 μm. The scattered red dots, which were observed in some PFIC-2 livers, were almost indistinguishable from autofluorescence observed in several livers. Therefore, it cannot be excluded that BSEP formed intracellular deposits in PFIC-2. BSEP, bile salt export pump; PFIC, progressive familial intrahepatic cholestasis; MRP, multidrug resistance associated protein; MDR, multidrug resistance.

Figure 2.

MDR3 expression in PFIC-2 and PFIC-3 livers. (A) PFIC liver slices were stained with the MDR3 antibody P3II26 (red) and the MRP2 antibody M2III6 (green). PFIC-2 livers C1, C2, C3, C4, and C5 showed colocalization of MDR3 and MRP2. Livers C6, C7, and C8 showed a clear canalicular staining pattern of MDR3 despite a MDR3 mutation (Table 1). Livers C9 and C10 displayed no canalicular MDR3 but MRP2 staining. Bars = 10 μm. (B) Western blot analysis of MDR3 in PFIC and controls. MDR, multidrug resistance; PFIC, progressive familial intrahepatic cholestasis; MRP, multidrug resistance associated protein.

BSEP mRNA expression was 5.8 ± 5.3 (n = 4) relative mRNA units (in the following termed “RRU”; see Materials and Methods) in PFIC-2 livers and 6.2 ± 3.0 RRU (n = 5) in PFIC-3 livers (concerning PFIC-3 diagnosis: see later discussion). The mRNA values were insignificantly lower or similar to the values of control group I (9.4 ± 0.7 RRU, n = 3) or control group II (4.0 ± 1.8 RRU, n = 6), respectively (Table 4).

Table 4. mRNA Expression of Hepatobiliary Transporters in PFIC Livers
 PFIC-2 (n = 4)PFIC-3 (n = 5)Control I (n = 3)Control II (n = 6)
Mean ± SDMean ± SDMean ± SDMean ± SD
  • NOTE. Relative mRNA levels of transporters were measured in PFIC-2 and PFIC-3 livers and compared to controls. mRNA levels of individual genes are expressed as x-fold copies compared to the patient's individual HPRT-mRNA-level.

  • Abbreviation: n.d., not determined.

  • *

    Significantly different from control I (P < .05)

  • **

    significantly different from control II (P < .05)

  • #

    MRP3 mRNA of C10 was an outlier (70.3 RRU) which was not included, otherwise mean MRP3 expression in PFIC-3 was 20.6 ± 27.9 RRU.

BSEP5.8 ± 5.36.2 ± 3.09.4 ± 0.74.0 ± 1.8
MDR338.8 ± 22.134.9 ± 18.9**15.0 ± 9.717.3 ± 17.5
MRP244.2 ± 52.840.9 ± 30.370.3 ± 55.632.3 ± 7.3
MDR14.0 ± 1.97.5 ± 3.3**3.3 ± 3.82.6 ± 1.7
NTCP36.7 ± 29.240.1 ± 20.133.5 ± 17.356.4 ± 17.5
OATP1B147.2 ± 34.0**74.7 ± 29.9**75.4 ± 31.4123.2 ± 28.2
OATP1B31.2 ± 1.0**1.9 ± 1.83.5 ± 1.54.4 ± 1.9
OATP2B11.2 ± 0.51.3 ± 0.61.2 ± 0.4n.d.
OATP1A20.06 ± 0.040.04 ± 0.030.08 ± 0.08n.d.
MRP11.3 ± 1.42.2 ± 1.5*/**0.7 ± 0.40.4 ± 0.3
MRP34.4 ± 4.48.2 ± 3.3#13.2 ± 15.55.3 ± 1.9
MRP42.8 ± 2.4**6.0 ± 2.6*/**1.0 ± 0.80.7 ± 0.4

In three of four patients with high-γGT cholestasis, MDR3 gene-sequencing showed 3 different homozygous mutations listed in Table 1. A fourth child (C6) carried a disease-causing MDR3 mutation25 despite normal γGT. Interestingly, this child had the same mutation as child C7 with 10-fold increased γGT activities (Table 1). Patient C10, who had normal serum γGT activities, displayed clear BSEP immunoreactivity (Fig. 1) but no canalicular MDR3 immunoreactivity (Fig. 2). Although no MDR3 mutation was detectable by sequencing in this patient, he was diagnosed as PFIC-3 on the basis of the immunofluorescence pattern.

Despite homozygous mutations in the MDR3 gene, canalicular localization of MDR3 could be detected in 3 of these 5 livers (C6, C7, C8; Fig. 2). Two different mutations were found in the three patients with detectable canalicular MDR3 immunoreactivity (S346I and A953D). In all PFIC-3 livers, BSEP immunoreactivity showed a canalicular distribution and colocalized with other canalicular proteins such as MRP2 (Fig. 1).

MDR3 mRNA was detected in all PFIC-2 and PFIC-3 livers and was increased compared with controls (Table 4). MDR3 protein amounts were quite variable in PFIC-2 and PFIC-3 livers and showed no clear trend. MDR3 protein was detected in PFIC-3 livers C6, C7, and C8 (Fig. 2B), in line with the immunofluorescence data.

Expression of Other Transporter Proteins in PFIC-2 and PFIC-3.

NTCP mRNA levels in PFIC livers were not significantly different from those of control groups I and II (Table 4). Interestingly, Western blot analysis showed a significant reduction of NTCP protein by approximately 50% in both PFIC-2 and PFIC-3 livers (Fig. 3B, Table 5). Accordingly, reduced NTCP immunoreactivity was observed in PFIC-2 and -3 (Fig. 3A).

Figure 3.

NTCP expression in PFIC-2 and PFIC-3 livers. (A) NTCP was localized in the basolateral membrane of both control and PFIC livers. Reduced immunoreactivity to the anti-NTCP antibody was observed for most PFIC livers compared with controls. Bars = 10 μm. (B) NTPC expression is reduced in PFIC livers as shown by Western blot analysis. Images from the remaining patients are included in the supplemental data file. NTCP, sodium taurocholate cotransporting polypeptide; PFIC, progressive familial intrahepatic cholestasis.

Table 5. Protein Expression of Hepatobiliary Transporters in PFIC Livers
 PFIC-2 (n = 4)PFIC-3 (n = 5)Control I (n = 3)
Mean ± SDMean ± SDMean ± SD
  • NOTE. Relative protein expression was quantified by Western blot analysis and densitometry. Values were calculated to be x-fold as compared with the protein expression of control liver Con1 (from control group I), which was set to 1.0.

  • *

    Significantly different from control I with P < .05.

  • A shift of OATP1B1 to a lower molecular weight was observed in control patient Con1 (Fig. 5B); therefore, densitometry was normalized to Con2.

MDR30.98 ± 0.520.50 ± 0.471.09 ± 0.08
MRP20.57 ± 0.390.62 ± 0.36*1.14 ± 0.16
NTCP0.46 ± 0.17*0.47 ± 0.14*1.01 ± 0.37
OATP1B10.28 ± 0.42*0.42 ± 0.38*1.21 ± 0.18
OATP1B30.17 ± 0.05*0.32 ± 0.32*1.07 ± 0.25
MRP30.52 ± 0.250.76 ± 0.20.81 ± 0.17
MRP44.99 ± 5.989.27 ± 3.14*0.78 ± 0.55
Figure 5.

OATP1B1 and OATP1B3 expression in PFIC-2 and PFIC-3 livers. (A) OATP1B1 immunoreactivity in liver sections of PFIC patients was noticeably reduced compared with controls (identical settings were used for all pictured. Representative pictures are shown; PFIC-3 patients with [C7] or without [C9] canalicular MDR3 expression). OATP1B3 immunoreactivity was reduced in PFIC livers compared with controls. In some PFIC livers such as C4, intracellular OATP1B3 was increased at the expense of membrane-bound OATP1B3. (B) OATP1B3 protein expression is reduced in most PFIC patients. Similarly, OATP1B1 protein was lower in PFIC patients compared with controls. Of note, the molecular mass of OATP1B1 was diminished in Con1 (designated as OATP1B1*), which was not observed in any other control or PFIC patient. The reason for this shift in molecular mass is not known. A representative loading control (GAPDH) is shown. Images from the remaining patients are included in the supplemental data file. OATP, organic anion transporting polypeptide; PFIC, progressive familial intrahepatic cholestasis; MDR, multidrug resistance.

MRP3 transports conjugated monoanionic bile salts and was suggested to play a role in bile salt homeostasis under cholestatic conditions.11, 13 Relative MRP3 mRNA levels were 4.4 ± 4.4 RRU in PFIC-2 and 8.2 ± 6.6 RRU in PFIC-3 livers, respectively. They were not significantly different from controls (control I: 13.2 ± 15.5 RRU; control II: 5.3 ± 1.9 RRU). Protein expression of MRP3 was unaltered in PFIC-2 and -3 livers and was 0.81 ± 0.17 RDU in control I and 0.52 ± 0.25 in PFIC-2 and 0.76 ± 0.20 in PFIC-3 livers, respectively (Fig. 4B). In all livers examined, MRP3 was localized at the basolateral membrane of hepatocytes (Fig. 4A).

Figure 4.

Expression of MRP isoforms in PFIC-2 and PFIC-3 livers. (A) MRP3 and MRP4 were costained in PFIC or control livers. Immunoreactivity to MRP3 was similar in control and in PFIC livers. Basolateral MRP3 was found over the entire sinusoid without evident zonal distribution. In contrast to MRP3, MRP4 displayed a stronger immunofluorescence signal in PFIC livers as compared with controls. Bars = 10 μm. (B) Western blot analysis of MRP2, MRP3, and MRP4 in PFIC livers. Whereas MRP3 protein showed no evident differences in control and diseased livers, MRP4 was upregulated in most PFIC livers. MRP2 expression was diminished in approximately half of the livers. Images from the remaining patients are included in the supplemental data file. MRP, multidrug resistance associated protein; PFIC, progressive familial intrahepatic cholestasis.

MRP4 is also localized at the sinusoidal membrane of hepatocytes. Recently human MRP4 has been shown to transport bile salts together with glutathione.10 In control livers, relative MRP4 mRNA expression was 1.0 ± 0.8 RRU (control I) and 0.7 ± 0.4 RRU (control II). Relative MRP4 mRNA levels were significantly elevated in PFIC-3 (6.0 ± 2.6 RRU) and in PFIC-2 (2.8 ± 2.4 RRU) livers compared with control group II, but the increase in MRP4 mRNA levels did not reach significance in PFIC-2 when compared with control I (Table 3). MRP4 protein expression was increased 5- and 10-fold in PFIC-2 and PFIC-3 livers, respectively, compared with control livers (Table 5). MRP4 expression was more variable in the PFIC-2 livers examined here (Fig. 4B). Immunofluorescence microscopy confirmed the basolateral localization of MRP4 in both control and diseased livers (Fig. 4A).

MRP2 transports mono- and bisglucuronosyl bilirubin and certain conjugated bile salts into bile.26 MRP2 mRNA levels appeared lower in PFIC livers (Table 3), but the differences did not reach significance because of considerable variation in MRP2 mRNA and protein levels (Fig. 4B). MRP2 protein was detected in all control and diseased livers at the canalicular membrane (Figs. 1 and 2).

The mRNA expression of OATP1B1, the counterpart of MRP2 in terms of bilirubin glucuronide transport,3 was significantly reduced in PFIC-2 and -3 compared with control II (Table 4) but was almost unchanged in PFIC-3 livers when compared with control I. Immunofluorescence staining of OATP1B1 at the sinusoidal membrane was notably reduced in diseased livers (Fig. 5A). Protein levels were significantly reduced in PFIC livers (Table 5).

Expression of OATP1B1 mRNA was 40 to 60 times higher than OATP1B3 mRNA and 20 to 40 times higher than OATP2B1 mRNA expression. OATP1B3 mRNA expression in diseased livers was lowered, whereas OATP2B1 was almost unaltered in PFIC livers compared with controls. Western blot analysis of OATP1B3 clearly showed a reduced expression of OATP1B3 in PFIC-2 and -3 livers (Table 5, Fig. 5B). In accordance with the Western blot data, OATP1B3 was reduced in most and hardly detectable in some PFIC livers as detected by immunofluorescence (Fig. 5A).


This study investigates the expression of hepatobiliary transporters in livers of children with PFIC.19, 27 PFIC livers are of interest because they represent human gene “knock-outs” and because results from animal studies are not entirely valid for humans. For example, Bsep knock-out mice develop only mild liver disease in contrast to humans.28 Other examples of species differences regarding bile salt homeostasis are the different suppression of Cyp7A1, Cyp27A1, and Cyp8B1 by bile salts in human hepatocytes compared with rodent hepatocytes,29 a differential regulation of Cyp7A1 by L×R,30 and marked species differences in the binding of bile salts to farnesoid X receptor.31

Our results suggest that PFIC-2 can be diagnosed on the basis of absent canalicular BSEP immunoreactivity (Table 1). This is in line with previous reports32, 33 suggesting that immunofluorescence staining might be a valuable diagnostic tool for PFIC-2.34 Although our children with a low-γGT cholestasis having a PFIC-1 genotype was not ruled out, this seems unlikely, because BSEP expression and its canalicular localization were reported to be unaltered in PFIC-1 disease.35 Milder forms of BSEP mutations were recently identified to cause a second form of benign recurrent intrahepatic cholestasis (BRIC-2).36 However, BSEP localization in BRIC-2 patients was not investigated in that study.

Impaired processing or targeting of mutated BSEP might be the cause of absent BSEP from the canaliculi in these patients. Defective BSEP targeting as an important mechanism underlying PFIC-2 was reported recently: 5 of 7 mutations prevented rat Bsep from trafficking to the apical membrane.37 Similarly, increased intracellular degradation and reduced canalicular targeting of mouse Bsep carrying the common D482G mutation was demonstrated recently.38

MDR3 immunofluorescence is of limited value in the diagnosis of PFIC-3, as shown by the association of disease-causing MDR3 mutations and normal canalicular MDR3 localization. The mutation S346I was detected in 2 children (C6 and C7). Normal canalicular MDR3 localization associated with this mutation was already reported for child C7.39 A second MDR3 mutation (A953D, C8) identified in this study has not been described so far. Similar to S346I, a canalicular staining pattern of MDR3 was observed in this patient. Additional PFIC-3 causing MDR3 mutations, which do not prevent MDR3 targeting to the canaliculi, include T424A and V425M.22, 39 Preserved expression and targeting suggest that these MDR3 mutations cause decreased transporter activity. One mutation found in patient C9 led to a 7–base pair deletion, resulting in a premature stop-codon, and was recently described.21 In line with an early truncation of the protein, no canalicular immunoreactivity was detected in the liver of this child. It is of note, however, that the MDR3 mRNA level was not significantly lowered in this patient compared with controls. The mutation A555G (PFIC-2 patient C5) results in an amino acid change in codon 175 from threonine to alanine. It has not been described so far in PFIC-3 patients,21 patients with MDR3-associated cholelithiasis,40 or patients with intrahepatic cholestasis of pregnancy,25 all of which are linked to MDR3 mutations.22 The pathophysiological relevance of this mutation remains to be determined; however, based on the absence of canalicular BSEP immunoreactivity, BSEP deficiency is likely the more relevant alteration in this patient.

Taken together, PFIC-3 cannot be diagnosed on the basis of MDR3 immunofluorescence as suggested for PFIC-234 but requires gene-sequencing or single nucleotide polymorphism analysis.

Our results further suggest that γGT levels are of limited specificity in the diagnosis of the PFIC subtype. For example patient C3 could only be diagnosed as PFIC-2 on the basis of absent BSEP immunoreactivity despite elevated γGT. In contrast, patient C6 had a homozygous MDR3 mutation and almost normal γGT, whereas patient C7 with the same mutation had 10-fold increased γGT levels. Conversely, in patient C10, MDR3 was not detectable in the canaliculi or on Western blots, but he had normal γGT levels.

In this study, 2 control groups were established. One control group consisted of liver samples of organ donors. The second control group was composed of liver biopsy specimens from adult patients with different liver diseases. This approach—taking mildly diseased livers as controls—was already used in other studies of human cholestatic conditions.41–43 Despite the limitations and small sample sizes, some clear differences were detectable between PFIC livers and controls.

Serum concentrations of primary bile salts are elevated in PFIC-2 and PFIC-3 compared with healthy individuals. Bile salt concentrations in bile of PFIC-3 patients are normal,32–39 which might be explained by preserved canalicular BSEP expression as demonstrated in this study. Elevated bile salt concentrations in the blood of PFIC-3 patients consequently must be attributable to other mechanisms than BSEP downregulation and might include impaired sinusoidal uptake of bile salts or increased efflux back into blood. Both possibilities are supported by our data: Downregulation of the NTCP protein occurred on a posttranscriptional level in PFIC-2 and -3 livers and might explain the decreased bile acid uptake. In individual patients, an apparently low correlation between Western blot and immunofluorescence data was observed. This might be attributable to the general difficulties in protein quantification from microscopic images. For example, sinusoidal transporters distribute over a large area, and small or moderate changes in their amount cannot be recognized by eye. Furthermore, even within a single tissue section, staining quality can vary considerably, and transporter density in livers with cirrhosis may be heterogeneous. However, immunofluorescence data complement Western blot analysis, in that they give information about the subcellular transporter localization.

The main transporters from the OATP family involved in bile salt uptake were also downregulated in PFIC-livers. Reduction of OATP1B3 and OATP1B1 mRNA was more pronounced in PFIC-2 compared with PFIC-3 livers, whereas mRNA levels of OATP2B1, which is not involved in bile salt transport, was hardly changed. It might be speculated that absence of a functionally active BSEP protein from the canalicular membrane of hepatocytes in PFIC-2 causes a stronger accumulation of bile salts compared with PFIC-3 hepatocytes, making a reduction of bile salt uptake even more essential. Downregulation of NTCP, OATP1B1, and OATP1B3 was similarly observed in advanced stages of primary biliary cirrhosis.43, 44 The subcellular distribution of hepatobiliary transporters is of major significance for overall bile formation. The intracellular occurrence of transporter proteins as observed in PFIC livers (e.g., Fig. 3 for NTCP or Fig. 5A, patient C4, for OATP1B3) or in control livers (e.g., Fig. 2, patient Con1 for MDR3 and MRP2 or Fig. 3 for NTCP) indicates that transporter localization needs to be considered to better understand the pathophysiology of cholestatic syndromes. However, “dynamic” models (such as cell lines, primary hepatocytes, or perfused animal livers) are required to investigate this complex matter.

An increased reflux of bile salts from hepatocytes into plasma might further contribute to increased plasma bile salt concentrations in PFIC patients: In animal models, Mrp3 was found to be upregulated in certain cholestatic and icteric conditions to maintain bile salt or bilirubin glucuronide homeostasis.11, 15, 45 In humans, MRP3 expression is increased in patients with Dubin-Johnson-syndrome9 to compensate for reduced MRP2 activity at the canalicular membrane. Despite the role of MRP3 under these conditions, no changes of MRP3 mRNA or protein were observed in this study. In line with the Western blot data, immunoreactivity to MRP3 was indistinguishable in liver sections of PFIC and control patients (Fig. 4A). The data suggest that upregulation of MRP3 does not play a major role as an escape route for bile salts in PFIC patients, which is in line with the lower affinity for conjugated bile salts of human MRP3 compared with rat Mrp3.13 Similar to our findings, MRP3 was not upregulated in patients with advanced primary biliary cirrhosis.43

In contrast to MRP3, MRP4 is strongly upregulated at the mRNA and at the protein level in PFIC. This upregulation was most pronounced in PFIC-3 livers. Assuming that MRP4 may compete for bile salts with BSEP,46 the high MRP4 expression may partly explain the elevated plasma bile salt concentrations in PFIC-3 patients despite normal bile salt concentrations in bile as mentioned previously.

The mechanism underlying MRP4 upregulation in PFIC patients remains elusive. The increase in MRP4 protein might represent an adaptive response to increased bile salts, as reported recently in obstructive cholestasis in rat47 or in farnesoid X receptor knock-out mice with cholestasis due to severely reduced BSEP expression.48 All children of this study received tauroursodeoxycholate medication, and it was shown recently that tauroursodeoxycholate administration upregulates Mrp4/MRP4 in mice49 and in human livers.50 Therefore, tauroursodeoxycholate treatment may in part explain the increased MRP4 expression in our study. In all instances, increased bile salts (due to cholestasis or due to oral administration) seem to be associated with increased MRP4. In line with this conclusion, Mrp4 expression further increased, when farnesoid X receptor knockout mice received bile salts orally.48 Interestingly, Mrp3 was hardly affected in these cholestatic animals,48 which is in good agreement with our results.

In conclusion, upregulation of MRP4 in human liver disease likely represents a major escape route for bile salts. Together with the concerted downregulation of sinusoidal bile acid uptake transporter, MRP4 is likely to protect hepatocytes from toxic intracellular bile salt concentrations in cholestasis.


Technical assistance by A. Knoke is acknowledged.