fax: (49) 241-8082455
Liver Biology and Pathobiology
Cytokine-independent repression of rodent Ntcp in obstructive cholestasis†
Article first published online: 18 FEB 2005
Copyright © 2005 American Association for the Study of Liver Diseases
Volume 41, Issue 3, pages 470–477, March 2005
How to Cite
Geier, A., Zollner, G., Dietrich, C. G., Wagner, M., Fickert, P., Denk, H., van Rooijen, N., Matern, S., Gartung, C. and Trauner, M. (2005), Cytokine-independent repression of rodent Ntcp in obstructive cholestasis. Hepatology, 41: 470–477. doi: 10.1002/hep.20594
Conflict of interest: Nothing to report.
- Issue published online: 22 FEB 2005
- Article first published online: 18 FEB 2005
- Manuscript Accepted: 13 DEC 2004
- Manuscript Received: 29 JUL 2004
- Deutsche Forschungsgemeinschaft. Grant Numbers: SFB542 TP C1, DI 729/3-1
- Austrian Science Foundation. Grant Number: P15502
Cholestatic liver injury is associated not only with accumulation of bile acids but also with activation of proinflammatory cytokines. Common bile duct ligation (CBDL) induces sustained downregulation of the Na+/taurocholate cotransporter (Ntcp) in rodent liver. Although repression of Ntcp during endotoxemia is cytokine mediated, it is unclear whether inflammatory cytokines contribute to this downregulation in obstructive cholestasis. Cytokine inactivation in CBDL rats and mice was either performed directly with tumor necrosis factor alpha (etanercept) or interleukin 1 beta inactivation (anakinra/AMG 719) or indirectly Kupffer cell depletion via intraperitoneal administration of liposome-encapsulated dichloromethylene bisphosphonate. Protein and messenger RNA (mRNA) expression of Ntcp and short heterodimer partner (SHP) were analyzed via Western and Northern blotting. Key regulators of Ntcp (hepatocyte nuclear factor 1 alpha [HNF-1α], HNF-4α, retinoid X receptor alpha [RXRα]:retinoic acid receptor alpha [RARα]) were studied via electrophoretic mobility shift analysis and nuclear Western blot analysis. Both methods of cytokine inactivation failed to maintain Ntcp protein or mRNA expression within 3 days after CBDL in either rats or mice (20%-40% of sham controls), while SHP mRNA expression increased three- to five-fold. Decreased nuclear HNF-1α and HNF-4α protein levels (45% and 60% of sham controls, respectively) and HNF-1α binding activity (32% of sham controls) were not restored during cytokine inactivation after CBDL, indicating cytokine-independent mechanisms of Ntcp regulation. RXRα:RARα binding remained unchanged in all experimental conditions. In conclusion, during obstructive cholestasis accumulating bile acids per se, without major contribution of cytokines, leads to downregulation of Ntcp via repression of HNF-1α and HNF-4α. (HEPATOLOGY 2005;41:470–477.)
The Na+/taurocholate cotransporter (Ntcp; SLC10A1 in humans, Slc10a1 in rodents) represents the main hepatocellular sodium-dependent uptake system for conjugated bile acids from sinusoidal blood in human and rodent liver.1-3 Downregulation of Ntcp under cholestatic conditions has been identified as a common mechanism of protecting the liver from continued uptake of potentially toxic bile acids.4 Accordingly, Ntcp expression is decreased in various human cholestatic liver diseases5-7 and rodent models of cholestasis.4, 8-12
Downregulation of Ntcp in cholestasis primarily occurs at the transcriptional level.4, 10, 13 Hepatocyte nuclear factor 1 alpha (HNF-1α) and the retinoid X receptor alpha (RXRα):retinoic acid receptor alpha (RARα) heterodimer complex have been identified as the major transactivators in the proximal region of rat Ntcp.10, 14 Decreased binding activity at these two regulatory elements in vivo occurs either via induction of inflammatory cytokines or retention of bile acids and leads to downregulation of Ntcp gene expression.10, 15, 16 HNF-1α activity is decreased by tumor necrosis factor alpha (TNF-α) and interleukin 1 beta (IL-1β)16 or by accumulating bile acids which in turn inhibit HNF-4α–mediated HNF-1α transactivation.17 Accumulating bile acids bind to farnesoid X receptor, a bile acid receptor, thereby inducing the expression of short heterodimer partner (SHP),18, 19 which inhibits HNF-4α–mediated transactivation of the HNF-1α promoter.17 As another pathway, SHP interferes with RXRα:RARα binding to the Ntcp promoter.15 Alternatively, inflammatory mediators such as IL-1β mediate suppression of RXRα:RARα transactivation of the Ntcp promoter by a c-jun-N-terminal kinase–dependent inhibition of RXRα phosphorylation.20
Obstructive cholestasis is a complex pathophysiological situation associated not only with accumulation of bile acids but also with activation of proinflammatory cytokines,21 both of which may mediate Ntcp repression. In addition, hydrophobic bile acids per se have been shown to induce proinflammatory cytokines in macrophages.22 However, the potential role of cytokines in mediating Ntcp repression in obstructive cholestasis is currently unknown.
Therefore, the contribution of proinflammatory cytokines to Ntcp repression was analyzed after common bile duct ligation (CBDL). Using a pharmacological knockout approach, we characterized the effects of TNF-α and IL-1β in the downregulation of nuclear receptors (SHP, RXRα:RARα), hepatocyte-enriched transcription factors (HNF-1α, HNF-4α), and Ntcp expression in both rats and mice. In addition, Kupffer cell depletion experiments employing liposomal bisphosphonate targeting have been performed to determine a potential role of other inflammatory mediators as effectors of a bile acid–independent regulation.
Materials and Methods
Male Swiss albino mice (20-25 g, strain Him OF1 SPF; Institute of Laboratory Animal Research, Himberg, Austria) and male Sprague-Dawley rats (200-250 g; Janvier, Le Genest–St. Isle, France) were used. All animals were kept on a 12-hour light/dark cycle and had free access to water and standard mouse or rat diet. The experimental protocols were approved by the local Animal Care and Use Committee according to criteria outlined in the Guide for the Care and Use of Laboratory Animals prepared by the National Academy of Sciences, as published by the National Institutes of Health (NIH publication 86-23, revised 1985).
CBDL and Pharmacological Cytokine Inactivation.
All surgical procedures were performed under sterile conditions. Mice and rats (n = 3 to 4) were bile duct ligated as previously described.4, 23, 24 Briefly, the abdomen was opened through a midline incision. The bile duct was identified and isolated and double ligated in his proximal part. In addition, cholecystectomy was performed in mice after ligation of the cystic duct. Controls underwent a sham operation with exposure, but without ligation of the common bile duct and removal of the gallbladder. During the operation rats and mice received either 8 mg/kg body weight etanercept, a soluble p75 TNF receptor fusion protein,25 or 100 mg/kg body weight of a long acting formulation of anakinra (AMG 719, kindly provided by Amgen Inc., Thousand Oaks, CA), a specific IL-1 receptor antagonist.26 Etanercept was injected intraperitoneally once again 48 hours after surgery, and AMG719 was administered daily. The livers were harvested 1 (mice) or 3 (rats) days after CBDL under general anesthesia, immediately snap frozen, and stored in liquid nitrogen until RNA extraction or were freshly used for isolation of liver nuclei and isolation of nuclear proteins.
Kupffer Cell Depletion Via Administration of Liposome-Encapsulated Dichloromethylene Bisphosphonate.
Kupffer cells were depleted in Swiss albino mice (n = 7) via intraperitoneal injection of 4 mL/kg BW liposome-encapsulated dichloromethylene bisphosphonate (Cl2MBP) 48 hours before CBDL, and mice were sacrificed 24 hours after CBDL.27-29 CL2MBP was a gift from Roche Diagnostics (Mannheim, Germany); Cl2MBP liposomes were prepared as described previously.27 Controls received intraperitoneal treatment with vehicle (i.e., liposomes in phosphate-buffered saline). The duration of 24 hours for CBDL was chosen because depletion of Kupffer cells as well as systemic macrophages for a longer period would be expected to lead to pronounced bacterial translocation from the gut or to loss of regenerative mechanisms induced by cytokines.30 Kupffer cell elimination was confirmed by staining paraffin sections with an antibody recognizing the F4/80 antigen expressed by macrophages (Serotec, Oxford, UK; dilution 1:50). In brief, 3-μm paraffin sections were deparaffinated, rehydrated, and digested with 0.1% protease (type XXIV, Sigma) for 10 minutes. Endogenous peroxidase was blocked with 1% H2O2 in methanol for 10 minutes. Specific binding of the F4/80 antibody was detected using a biotinylated anti-rat immunoglobulin G (Dako, Glostrup, Denmark) and the ABC System (Dako) with β-amino-9-ethyl-carbazole as substrate.
RNA Preparation and mRNA Analysis.
RNA was isolated from mouse and rat livers by standard phenol chloroform extraction procedure. Total RNA (10-20 μg) was analyzed by Northern blotting with specific and constitutively expressed probes of mouse or rat Ntcp, mouse SHP and mouse or rat glyceraldehyde-3-phosphate dehydrogenase as previously described.23, 31 RNA was quantified spectrophotometrically at 260 nm and the quality of total RNA was controlled by denaturing formaldehyde agarose gel electrophoresis. Real-time polymerase chain reaction was performed on a Gene-Amp 5700 Sequence Detection System and with GeneAmp 5700 SDS software (Applied Biosystems, Vienna, Austria) to semiquantify mRNA levels of IL-1β and TNF-α after Kupffer cell depletion as described previously.32 The TaqMan oligonucleotides corresponding to published sequences were as follows:
IL-1β: forward primer, 536-560; reverse primer, 687-667; probe, 638-665, corresponding to the published sequence, GenBank accession number M15131;
TNF-α: forward primer, 408-432; reverse primer, 582-560; probe, 443-468, corresponding to the published sequence, GenBank accession number M13049.
mRNA data were normalized to those of 28S ribosomal RNA.32
Electrophoretic Mobility Shift Analysis.
To determine binding activity of liver-enriched transcription factors and nuclear hormone receptors, electrophoretic mobility shift assays were performed as previously described.31 Five to 10 micrograms of nuclear extracts (protein) were incubated on ice for 30 minutes with a 2 × 104 cpm [32P] end-labeled oligonucleotide probe representing the HNF-1α or DR-1 (direct repeat-1; binding site for RXRα:RARα) element of the proximal rat Ntcp promoter.10, 14 For competition assays, 100-fold molar excess of unlabeled oligonucleotides were coincubated with the labeled probe. For supershift experiments nuclear extracts were preincubated for 30 minutes on ice with 1 μg of a polyclonal antibody against HNF-4α (Santa Cruz Biotechnology, Santa Cruz, CA) before addition of the labeled oligonucleotide probe. Separation of protein–DNA complexes from unbound labeled probes was performed via electrophoresis using a nondenaturing 6% polyacrylamide gel and quantified via phosphorimaging (Biorad, Munich, Germany).
Western Blot Analysis.
Preparation of either liver microsomes or nuclear protein and analysis of rat Ntcp protein mass and protein mass of various transcription factors were performed as described previously.12, 32 Protein concentrations were determined according to Bradford33 and 75 μg of microsomal protein or 10 to 25 μg of nuclear protein extract was separated by SDS-PAGE as previously described.12, 32 After electrotransfer onto nitrocellulose membranes, the blots were incubated with anti-rat Ntcp fusion protein antiserum,34 monoclonal anti-Na+/K+-ATPase α-1 antibody (Upstate, Lake Placid, NY) or polyclonal antibodies against transcription factors HNF-1α, HNF-4α, RXRα, and RARα as well as histone (Santa Cruz Biotechnology). Immune complexes were detected according to the ECL kit (Amersham, Little Chalfont, England).
Three to four animals were studied in each group. Data are reported as the arithmetic mean ± SD. Differences between experimental groups were analyzed via ANOVA with Bonferroni posttesting using the SYSTAT statistic program (SYSTAT, Evanston, IL). A P value of less than .05 was considered significant.
Effects of Cytokine Inactivation on Ntcp Expression in CBDL Rats.
To determine whether reduced Ntcp expression, which has been observed after bile duct ligation in rats,4 was prevented by inactivation of TNF-α or IL-1β, we quantified Ntcp steady-state mRNA levels via Northern blot analysis (Fig. 1) and protein expression via Western blot analysis (Fig. 2). As previously described,4 rat Ntcp mRNA and protein expression declined to 37% ± 10% and 20% ± 11% of controls, respectively (P < .05; n = 4) (Figs. 1, 2). Inactivation of TNF-α or IL-1β by pretreatment with etanercept or AMG719 had no significant effect on steady state mRNA levels of Ntcp, which remained below 40% of controls (Fig. 1). Similarly, inactivation of either cytokine did not significantly restore Ntcp protein expression (Fig. 2). In contrast, Mrp2 mRNA expression was preserved after cytokine inactivation as a positive control (data not shown). These data indicate that Ntcp downregulation after CBDL occurs via cytokine-independent pathways.
Binding Activity and Nuclear Protein of Transcription Factors After Cytokine Inactivation in CBDL Rats.
Electrophoretic mobility shift assay and nuclear Western blotting were performed to further elucidate the consequences of biological inactivation of TNF-α or IL-1β as potential regulators of transcription factors involved in Ntcp downregulation. Oligonucleotides representing HNF-1α and RXRα:RARα (DR-1) responsive elements from the proximal rat Ntcp promoter sequence10, 14 were used for gel shift analysis (Fig. 3A-B). Analysis of nuclear extracts isolated 72 hours after CBDL showed decreased binding activity of HNF-1α to 32% ± 14% (P < .05; n = 4), whereas binding to the DR-1 element was unaltered. Supershift assays excluded significant HNF-4α binding to the Ntcp DR-1 element (data not shown). To characterize the nature of decreased HNF-1 binding to its respective element, HNF-1α and HNF-4α nuclear protein was quantified via Western blotting using nuclear extracts isolated at the same time point (Fig. 3C). CBDL leads to a significant decrease in nuclear protein expression of both hepatocyte-enriched transcription factors to 45% ± 28% and 58% ± 5%, respectively (P < .05 each; n = 4). In parallel to HNF-1 binding activity in electrophoretic mobility shift assay, nuclear protein abundance of both transcription factors remained suppressed despite either anticytokine treatment (Fig. 3C). Therefore, HNF-1α and HNF-4α appear to be decreased after CBDL, primarily via cytokine-independent but bile acid–dependent pathways.
CBDL-Mediated SHP Induction and Ntcp Repression in Mice Is Not Mediated by Cytokines.
Previous studies have shown concomitant induction of SHP associated with reduction of Ntcp levels in CBDL-treated mice,23 but the relative contribution of proinflammatory cytokines to this SHP-dependent pathway regulating Ntcp repression is unknown. To determine the effects of cytokine inactivation on Ntcp regulation in CBDL, mRNA levels of SHP and Ntcp were studied in CBDL mice 24 hours after surgery (Fig. 4). After CBDL, Ntcp steady state mRNA levels were reduced to 14% ± 10% of sham-operated controls (P < .05). Concomitant with reduction of Ntcp levels, SHP mRNA was induced 1 day (or 24 hours) after CBDL (272% ± 92%). As observed in CBDL-treated rats, inactivation of TNF-α or IL-1β did not significantly alter Ntcp mRNA expression, and similar results were obtained even after combined blockade of both cytokines (Fig. 4). In contrast, SHP mRNA expression was further increased by anticytokine treatment compared with both sham-operated controls and CBDL mice (506% ± 34% of controls after etanercept, 316% ± 101% after AMG 719, and 507% ± 74% after combined blockade vs. 282% ± 58% in CBDL mice, respectively; P < .05 each; n = 3). These data exclude a causative role of TNF-α and IL-1β in the SHP-dependent downregulation of Ntcp in CBDL mice.
Kupffer Cell Depletion Does Not Prevent Ntcp Repression in CBDL Mice.
To investigate the relative contribution of proinflammatory cytokines/mediators (other than TNF-α and IL-1β) in mediating Ntcp repression during CBDL, Kupffer cells, the main source of hepatic cytokines,35 were depleted via intraperitoneal injection of liposome-encapsulated CL2MBP before CBDL.27 Staining liver sections with red pulp F4/80 antibody revealed successful Kupffer cell depletion (Fig. 5). Three of 7 CBDL mice treated with CL2MBP developed portal vein thrombosis, and 1 mouse died, whereas no portal vein thrombosis or death occurred in vehicle (liposome)-injected animals. Kupffer cell depletion, however, neither restored Ntcp expression nor prevented induction of SHP 3 days after CBDL (Fig. 6). To investigate whether sustained Ntcp repression despite efficient Kupffer cell depletion was due to proinflammatory cytokines excreted by other cells (e.g., bile duct epithelial cells, endothelial cells),36 we compared hepatic TNF-α and IL-1β mRNA expression—both known to be involved in Ntcp regulation9, 16, 37—in Kupffer cell–depleted and naive CBDL animals. CL2MBP liposome administration before CBDL attenuated upregulation of TNF-α mRNA (229% ± 123% in Kupffer cell–depleted CBDL vs. 552% ± 64% in naive CBDL mice) (Fig. 6) and completely prevented IL-1β induction (47% ± 22% in Kupffer cell–depleted CBDL vs. 242% ± 29% in naive CBDL mice) (Fig. 7). The occurrence of portal vein thrombosis had no influence on Ntcp expression nor on TNF-α or IL1β mRNA levels. These results suggest that during CBDL effects of bile acid–induced SHP on Ntcp repression may dominate those of cytokines that were effectively suppressed by Kupffer cell depletion.
CBDL represents a model of obstructive cholestasis that is associated with both elevation of serum bile acid levels and induction of proinflammatory cytokines.23, 38 Rodent Ntcp gene expression is controlled by the hepatocyte-enriched transcription factors HNF-1α and HNF-4α as well as nuclear receptors, including RXRα:RARα and the transcriptional repressor SHP.10, 14, 15 Farnesoid X receptor–mediated SHP induction by bile acids has been shown to decrease RXRα:RARα binding to the Ntcp promoter and inhibits HNF-1α transactivation by HNF-4α.15, 17, 23, 39 In lipopolysaccharide-induced cholestasis, reduction of nuclear levels of HNF-1α and RXRα:RARα has previously been linked to Ntcp repression.10, 16, 20, 37 Increased levels of proinflammatory cytokines with concomitantly reduced binding activities of RXRα:RARα have also been suggested to downregulate Mrp2 (Abcc2) expression in livers and duodenums of CBDL rats.40, 41 However, the contribution of proinflammatory cytokines (including TNF-α and IL-1β) to Ntcp regulation during obstructive cholestasis has not been studied to date.
The current study yielded several important new findings: (1) down-regulation of Ntcp within 3 days of obstructive cholestasis does not involve TNF-α and IL-1β; (2) Kupffer cell depletion experiments without recovery of Ntcp expression after CBDL also exclude a role of other inflammatory mediators as effectors of bile acid–independent gene regulation; (3) decreased Ntcp mRNA expression upon CBDL is paralleled by a cytokine-independent reduction of both HNF-1α and HNF-4α; and (4) maintained RXRα:RARα binding despite reduced Ntcp expression supports a role for HNF-1α as the major gene regulator.
Selective inactivation of TNF-α and IL-1β without effects on Ntcp expression in the present study clearly demonstrates that downregulation of Ntcp in obstructive cholestasis is independent of these cytokines. Moreover, the lack of Ntcp restoration in Kupffer cell–depleted CBDL animals with attenuated (i.e., TNF-α) or even absent (i.e., IL-1β) cytokine induction further indicates that the effects of bile acids (e.g., via induced SHP) may dominate those of cytokines in mediating Ntcp repression in biliary obstruction. However, in contrast to CBDL, there may be a significant contribution of proinflammatory cytokines to Ntcp repression in rodent models of inflammation-induced cholestasis. In these models, Kupffer cell depletion and anticytokine strategies are able to prevent downregulation of Ntcp,16, 42 because inflammatory cytokines are higher in endotoxin-induced cholestasis while bile acid levels are much lower than in CBDL.
Elevated bile acid levels in CBDL lead to induction of SHP, which may subsequently reduce HNF-4α and HNF-1α transactivation.17, 43 However, this mechanism remains to be determined for β-muricholic acid, the major bile acid accumulating in rodents following CBDL.32, 44 HNF-1α and the RXRα:RARα heterodimer complex have been identified as the major transactivators of rat Ntcp.10, 14 In contrast, regulation of the mouse Ntcp promoter by HNF-1α and RXRα:RARα was recently questioned by Jung et al.,45 who did not identify respective response elements in a minimal mouse Ntcp promoter. However, the presence of specific binding elements upstream of the studied minimal promoter has to be postulated from other studies: A principal role of HNF-1α is suggested by findings in HNF-1α knockout mice that have demonstrated an almost absent baseline Ntcp mRNA expression in these animals.46 Furthermore, in extended studies on the full-length mouse Ntcp promoter gene, transactivation occurs by a more distal HNF-1α element which is also present in rat Ntcp (A. Geier et al., unpublished data, June 2004). The present study supports a role of HNF-1α as the predominant of these two Ntcp regulators, because gene expression in rats is remarkably reduced after CBDL even in the presence of maintained RXRα:RARα binding. However, RXRα:RARα may play a supplemental role in Ntcp regulation, which is in line with decreased binding activity and nuclear levels during endotoxin-induced cholestasis and prolonged (i.e., 7 days) bile duct obstruction.10, 40 During the first 3 days of CBDL, SHP mRNA expression correlates inversely with Ntcp expression, while after 7 days SHP mRNA levels approximate normal values while Ntcp remains repressed.23 This apparent discrepancy could be attributed to rising levels of proinflammatory cytokines such as TNF-α and IL-1β over time (G. Zollner et al., unpublished data, June 2004), because these cytokines are known to inhibit RXRα:RARα activity, which is decreased primarily during the later stages of biliary obstruction.20, 40 Moreover, bile acids have been shown to suppress HNF-4α transcription through SHP-independent mechanisms,17 which could explain the reduction of HNF-1α and subsequent Ntcp expression despite low SHP levels. We conclude that cytokine-independent effects are the dominant cause for Ntcp downregulation after CBDL. Such effects may involve bile acid–dependent suppression of the HNF-4α/HNF-1α pathway via SHP-dependent and independent mechanisms17 as a central pathway of Ntcp gene regulation.
In summary, taken together, our findings support a new model of distinct transcriptional pathways involved in the regulation of Ntcp expression during cholestasis depending on the type of cholestatic liver injury. In endotoxin-induced cholestasis, proinflammatory cytokines have been demonstrated to repress both HNF-4α/HNF-1α and RXRα:RARα pathways, resulting in downregulation of Ntcp.10, 16, 23, 47 Bile acids per se accumulating during the early stage of obstructive cholestasis without major contribution of cytokines may lead to Ntcp downregulation via cytokine-independent repression of HNF-4α/HNF-1α transactivation.
The excellent technical assistance of Sabine Beutelspacher, Andrea Fuchsbichler, Aline Müller, Petra Schmitz, Dagmar Silbert, and Sonja Strauch is gratefully acknowledged.
- 11Ethinyl estradiol-induced cholestasis involves alterations in expression of liver sinusoidal membrane transporters. Am J Physiol 1996; 34: G1043–G1052., , , , , .