Defective bile salt biosynthesis and hydroxylation in mice with reduced cytochrome P450 activity


  • Potential conflict of interest: Nothing to report.

  • The authors thank Professor Bert Groen for providing us the negative controls for the Abcg5/8 blot. The authors are grateful to CNPq (Brazil) for providing a postdoctoral fellowship to Dr. Alexandra Acco (process no.: 200266/2008-0). The authors thank Dr. Colin Henderson (Dundee) for kind supply of Hrn mice. This study was partially funded by the Maag Lever Darm Stichting (MLDS; Dutch Digestive Diseases Foundation) (WO 08-63).


The difference in bile salt (BS) composition between rodents and humans is mainly caused by formation of muricholate in rodents as well as by efficient rehydroxylation of deoxycholic acid. The aim of this study was to characterize bile formation in a mouse model (Hrn mice) with hepatic disruption of the cytochrome p450 (CYP) oxidoreductase gene, encoding the single electron donor for all CYPs. Bile formation was studied after acute BS infusion or after feeding a BS-supplemented diet for 3 weeks. Fecal BS excretion in Hrn mice was severely reduced to 7.6% ± 1.8% of wild-type (WT), confirming strong reduction of (CYP-mediated) BS synthesis. Hrn bile contained 48% ± 18% dihydroxy BS, whereas WT bile contained only 5% ± 1% dihydroxy BS. Upon tauroursodeoxycholate infusion, biliary BS output was equal in WT versus Hrn, indicating that canalicular secretion capacity was normal. In contrast, taurodeoxycholic acid (TDC) infusion led to markedly impaired bile flow and BS output, suggesting onset of cholestasis. Feeding a cholate-supplemented diet (0.1%) resulted in a completely restored bile salt pool in Hrn mice, with 50% ± 9% TDC and 42% ± 10% taurocholic acid in bile, as opposed to 2% ± 1% and 80% ± 3% in WT mice, respectively. Under these conditions, biliary cholesterol secretion was strongly increased in Hrn mice, whereas serum alanine aminotransferase levels were decreased. Conclusion: Hrn mice have strongly impaired bile salt synthesis and (re)hydroxylation capacity and are more susceptible to acute TDC-induced cholestasis. In this mouse model, a more-human BS pool can be instilled by BS feeding, without hepatic damage, which makes Hrn mice an attractive model to study the effects of human BS. (HEPATOLOGY 2013)

Bile is a potentially highly toxic fluid. Bile formation is tightly regulated and its principal solutes are bile salts, phospholipids, and cholesterol. Impaired formation of bile can lead to severe hepatic toxicity, caused by the apoptotic action of accumulated bile salts. It is known that bile cytotoxicity, determined by its bile salt composition, differs between various species. Rodents, in contrast to humans,1, 2 are capable of rehydroxylating bile salts (BSs) that reenter the liver after dehydroxylation by intestinal bacteria, resulting in a more-hydrophilic and less-cytotoxic bile salt pool. Hence, human bile contains substantial amounts of the secondary cytotoxic BS, deoxycholic acid (DCA), whereas this is very low in rodent bile. Like humans, mice synthesize chenodeoxycholic acid, which is another cytotoxic dihydroxy BS, but they further hydroxylate it to less cytotoxic BSs.

The hepatic cytochrome P450 (CYP)-dependent monooxygenase system plays a central role in the metabolism of drugs, toxins, and carcinogens, but it also acts in key steps of cholesterol, steroid hormone, and BS synthesis. The exact role of many individual enzymes is known, although there are still many orphan enzymes in the CYP-family of which the exact function is unknown.3, 4

The family of CYPs contains 103 enzymes in mice5 and 57 in humans,6 with the drug-metabolizing CYPs having strongly overlapping functions. To eliminate both synthesis and rehydroxylation of bile salts, one would have to knock out multiple CYPs. However, all CYPs receive their electron from one single electron donor: the CYP oxidoreductase (CPR).7-9 Therefore, deletion of this gene will inactivate all CYPs. By using the Cre/loxP system, a hepatocyte-specific knockout (KO), hepatic reductase null (Hrn), has been generated.3

The aim of this study was to investigate the role of hepatic CYPs in the biosynthesis and rehydroxylation of BSs. BS synthesis is a hepatocyte-specific and partly CYP-mediated process and is therefore strongly impaired in Hrn mice.3 However, a restored BS pool can be artificially instilled by BS feeding. In that case, a change in composition of the BS pool is expected because there is extensive bacterial dehydroxylation and recirculation of BSs without CYP-mediated rehydroxylation. In the present article, we studied the effect of hepatic CYP inactivation on BS synthesis, excretion, and enterohepatic circulation. As expected, the absence of hepatic CYP activity severely affects bile formation and BS composition. Strikingly, these differences in bile formation and BS composition induced increased biliary cholesterol secretion, but did not lead to increased hepatic damage.


ALT, alanine aminotransferase; AMC, Academic Medical Center; Asbt, apical sodium-dependent BS transporter; AST, aspartate aminotransferase; BS, bile salt; BSEP, BS export pump; CA, cholic acid; CYP, cytochrome P450; GC, gas chromatography; GCDCA, glycochenodeoxycholic acid; HPLC, high-performance liquid chromatography; Hrn, hepatic reductase null mice; KO, knockout; mRNA, messenger RNA; SR-B1, scavenger receptor class B1; SEM, standard error of the mean; TC, taurocholic acid; TDC, taurodeoxycholic acid; TMC, tauromuricholic acid; TUDC, tauroursodeoxycholic acid; WT, wild type.

Materials and Methods


All experiments were performed with age-matched 2-6-month-old male C57bl/6 Cprlox/lox (wild-type; WT) or C57bl/6 Cprlox/lox + CreALB (Hrn) mice.3 Animals were kept in a pathogen-free environment on a controlled 12-hour light/dark regimen in the animal facility of the Academic Medical Center (AMC; Amsterdam, The Netherlands). All mice were fed a standard rodent chow and water ad libitum before the BS infusion experiments. In all other experiments, mice were first fed a purified semisynthetic diet for 2 weeks (K4068.02; Arie Blok Diervoeders, Woerden, The Netherlands), followed by a 3-week feeding period with a purified semisynthetic diet (control), either or not supplemented with 0.1% cholic acid (CA) (Calbiochem, Bad Soden, Germany) or 0.1% glycochenodeoxycholic acid (GCDCA) (Sigma-Aldrich, St. Louis, MO). This feeding regime corresponds to a BS dosing of 200 mg/kg*day. At 48 hours before the end of the experiment, mice were put in a metabolic cage and feces was collected. All animal experiments were approved by the institutional animal care and use committee of the AMC.

Bile Collection and In Vivo BS Infusions.

To obtain bile, mice were anesthetized and placed on a thermostatted heating pad to maintain body temperature. The common bile duct was closed with a ligature, and the gallbladder was canulated. Bile was collected in fractions of 10 or 15 minutes. In the BS infusion experiments, endogenous BSs were depleted for 90 minutes before infusion. At 90 minutes, either tauroursodeoxycholic acid (TUDC; Calbiochem) or taurodeoxycholic acid (TDC; Sigma-Aldrich) was infused through the jugular vein. Infusion rates were increased every 30 minutes in the case of TUDC infusion with 600 nmol/min*100 g (600, 1,200, 1,800, and 2,400 nmol/min*100 g) and every 60 minutes in the case of TDC infusion with 83 nmol/min*100 g (83 and 166 nmol/min*100 g). At the end of the experiment, blood, liver, and other organs were harvested.

Serum Biochemistry.

All serum biochemistry was performed by the routine laboratory of clinical chemistry of the AMC.

Enzymatic Assays.

BS, choline containing phospholipids, and cholesterol, were determined enzymatically. BSs were extracted from feces using 50% tert-butanol (Merck, Darmstadt, Germany). All measurements were done on a Novostar analyzer (BMG Labtech GmbH, Offenburg, Germany). Biliary BS composition was determined by reverse-phase high-performance liquid chromatography (HPLC).10 In brief, 100 μL of diluted bile or liver extract was applied to a Hypersil C18 (3 μm, 15 cm) HPLC column (Thermo Scientific, Breda, The Netherlands). The starting eluent consisted of 6.8 mM of ammoniumformate (pH 3.9), followed by a linearly increasing gradient of acetonitrile (Biosolve, Valkenswaard, The Netherlands). Detection was performed using a nano-quantity analyte detector (based on condensation nucleation) (QT-500; Quant Technologies, Blaine, MN). Response factors for different BS species were determined by calibration curves for each species upon HPLC. Fecal cholesterol was extracted using a liquid/gas chromatography (GC) method described by Gerhardt et al.11

Radiolabeled Determination of the BS Pool.

Mice were fed a control diet for 3 weeks, followed by oral gavage of 1 μCi of 3H-labeled taurocholic acid (TC) in saline 18 hours before collection of blood and bile. Total biliary BS content was measured enzymatically, and 3H levels in bile were determined by scintillation counting. The obtained specific radioactivity was used to calculate the total BS pool from the injected amount of radioactivity.

Protein Analysis.

Protein levels of Abcg5 were determined by western blotting in total liver homogenates using rabbit anti-Abcg55 in a 1:1,000 dilution. Livers were homogenized in ice-cold buffer (5 mmol/L of Tris [pH 7.5] and 250 mmol/L of sucrose with protease inhibitors). Intensities of blotting signal were normalized by those for the plasma membrane protein, Atp1a1, used as a loading control. Also, liver homogenate from Abcg5 KO mice (Abcg5−/−), kindly donated by Prof. Dr. Bert Groen (Groningen University, Groningen, The Netherlands), was used as a negative control in Abcg5 blottings.

Quantitative Polymerase Chain Reaction Analysis of Gene Expression.

Liver tissue was snap-frozen and stored at −80°C. RNA was isolated from tissue using Trizol reagent (Invitrogen, Leiden, Netherlands). Complementary DNA synthesis was transcribed using locked oligo-dT primers and Superscript III reverse transcriptase (Invitrogen). Quantitative polymerase chain reaction was performed using a Lightcycler 2.0 with the Fast Start DNA MasterPlus SYBR Green I kit (Roche, Woerden, Netherlands). Expression levels were normalized for 36B4.

Statistical Analysis.

All data are given as means ± standard error of the mean (SEM). Significance was tested by use of one-way analysis of variance with Bonferroni's correction for multiple testing considering statistic significance when P < 0.05.


Characteristics of Hrn Versus WT Animals.

Mice lacking hepatic CYP reductase (C57bl/6 Cprlox/lox + CreALB, further designated as Hrn) showed normal growth and normal appearance.3 Serum aspartate aminotransferase (AST) (aspartate aminotransferase) and alanine aminotransferase (ALT) were significantly increased in Hrn versus WT mice (Table 1). Also, livers of Hrn mice were enlarged (Table 1) and appeared to be fatty, as described by Henderson et al.3 Serum bilirubin was not increased (data not shown), indicating that there was no cholestasis.

Table 1. Phenotypic Characteristics of WT and Hrn Mice Under Basal Conditions
ParameterUnitWTHrnP Value
  1. Hepatic bile was collected during the first 10 minutes after canulation of the gallbladder. Hereafter plasma and livers were collected. Values are given as mean ± SEM.

 ASTU/L251 ± 43.8 (n = 9)608 ± 148 (n = 9)0.0024
 ALTU/L82 ± 18 (n = 10)330 ± 91 (n = 10)<0.0001
 FlowμL/min*100 g7.4 ± 0.5 (n = 17)4.3 ± 0.5 (n = 15)<0.0001
 BSnmol/min*100 g266 ± 57 (n = 17)61 ± 20 (n = 15)0.0022
 Cholesterolnmol/min*100 g4.0 ± 0.9 (n = 16)1.8 ± 0.4 (n = 14)0.0328
 Phospholipidsnmol/min*100 g28.4 ± 4.6 (n = 17)11.8 ± 1.7 (n = 15)0.0031
 Weight% of body weight4.5 ± 0.1 (n = 17)7.8 ± 0.3 (n = 19)<0.0001
 BSnmol/g26.7 ± 3.1 (n = 5)53.4 ± 6.4 (n = 5)<0.0100

To investigate bile secretion in Hrn mice, bile was collected for 10 minutes after cannulation of the gallbladder and distal ligation of the common bile duct. As expected, Hrn had a decreased biliary BS output (Table 1), because CYP7A1, the rate-limiting enzyme in BS synthesis, should not be functioning.

Hrn Mice Fed a Control Diet Have Reduced Biliary Cholesterol and Phospholipid Excretion.

Besides BSs, biliary cholesterol and phospholipid excretion were also low in Hrn versus WT mice (Table 1). However, when normalized for BS excretion, neither cholesterol nor phosholipid excretion were significantly changed (data not shown).

Hrn Mice Have a Decreased BS Pool.

To further investigate impaired BS synthesis, we measured the total BS pool by measuring tritium and total biliary BS after injection of a single dose of 3H-TC. Under control conditions (3 weeks feeding with the semisynthetic diet), the total BS pool in Hrn was significantly reduced, compared to WT (43 ± 5 and 107 ± 16 μmol/100g, respectively; Fig. 1A). Fecal output of BSs was reduced to 7.6% ± 1.8% in Hrn mice, compared to WT (Figure 1B). Because fecal excretion is, by far, the biggest route by which BSs can leave the body, fecal BS excretion equals hepatic BS synthesis. Hence, overall BS synthesis was decreased by more than 90% in Hrn mice. Relative messenger RNA (mRNA) levels of Asbt (apical sodium-dependent BS transporter) were equal in WT and Hrn, indicating that ileal uptake of BSs was not affected (Fig. 1C). Most likely, the strong reduction in the BS pool of Hrn mice would lead to decreased cholesterol absorption in the small intestine, because this process depends on proper cholesterol solubilization in the gut.12 To test this, we measured fecal sterol excretion and confirmed that this is the case, whereas fecal plant sterol excretion was unaffected (Fig. 1D). The increased fecal cholesterol excretion was not accompanied by an increase in hepatic 3-hydroxy-3-methyl-glutaryl coenzyme A reductase mRNA (data not shown).

Figure 1.

Total BS pool and fecal excretion in WT and Hrn mice. A single oral dose of 1 μCi of [3H]-TC was given 18 hours before sacrifice of animals, which allows complete mixture of the radioactive BS with the endogenous BS pool. Biliary BS concentration was measured enzymatically, and radioactivity was measured in all organs of the enterohepatic circulation (blood, intestine plus content, gallbladder plus content, and liver). The obtained specific radioactivity of biliary BS was used to calculate BS pool size (WT, n = 3; Hrn, n = 5) (A). During 48 hours, feces was collected in mice fed a semisynthetic diet. WT and Hrn mice were separated to avoid coprophagy contamination. BSs were extracted from feces with tert-butanol and measured quantitatively in a NOVOstar analyzer (BMG LABTECH GmbH, Ortenberg, Germany). Feces was collected from four cages per group of WT or Hrn (B). mRNA expression was determined for ileum of control-fed mice (WT, n = 5; Hrn, n = 4) (C). Fecal neutral sterols were measured with GC (D).

Excretion and Rehydroxylation of BSs.

To exclude an affected retention time of BSs in livers of Hrn mice, we infused a hydrophilic BS, TUDC, at rates up to 2,400 nmol/min*100 g and measured BS excretion in time (Fig. 2A). Compared to WT mice, BS excretion in Hrn was not changed, indicating that function of the Na+-dependent taurocholate cotransporting protein and the BS export pump (BSEP), involved in hepatic uptake and biliary secretion of BSs, respectively, were not or minimally decreased. The relative biliary output of cholesterol (Fig. 2B) and phospolipids (data not shown) was not different in TUDC-infused WT and Hrn mice. After infusion of a more-hydrophobic BS (TDC) at limited rates (166 nmol/min*100 g), bile flow was strongly diminished in Hrn, but not in control, mice. Higher infusion rates (200 nmol/min*100 g) led to acute death of the Hrn, but not WT, mice during the experiment (data not shown). Analysis of BS, cholesterol, and phospholipid output in bile showed a decreased output of these three components in Hrn mice in the highest infusion rate (Fig. 2C,E). Analysis of BS composition of secreted bile revealed a large difference between WT and Hrn mice. Whereas TDC was virtually completely converted to TC in WT, the majority of BS was secreted unaltered as TDC in Hrn bile. Hence, the absence of hepatic CYP activity in Hrn led to a much higher TDC/TC ratio in bile than in WT (Fig. 2F).

Figure 2.

Time course of biliary BS output. Bile was collected during 90 minutes to deplete endogenous BSs; subsequently, TUDC or TDC was infused through the jugular vein. TUDC infusion increased stepwise every 30 minutes with 600 nmol/min*100 g (600-1,200-1,800-2,400 nmol/min*100 g [A and B]; n = 4), and TDC infusion increased stepwise every 60 minutes (83-166 nmol/min*100 g [C-F]). (A) Biliary BS output during depletion phase and TUDC infusion. (B) Biliary cholesterol output versus BS output during depletion phase and TUDC infusion. (C) Biliary BS output during depletion phase and TDC infusion. (D) Biliary cholesterol output during depletion phase and TDC infusion. (E) Biliary phospholipid output during depletion phase and TDC infusion. (F) Ratio TDC/TC, as measured by HPLC. Open squares: WT mice (n = 13); open circles: Hrn mice (n = 11). As a result of severe cholestasis in Hrn mice after 150 minutes in the TDC-infusion experiment, the amount of collected bile was too little to determine BS species. WT mice did not suffer from cholestasis.

Biliary BS Composition in BS-Fed Mice.

Control-fed Hrn animals showed a decreased biliary BS output (Figure 2A,C; Table 2). Analysis of the BS composition of control-fed Hrn animals showed a decrease in the primary trihydroxy BSs, tauromuricholic acid beta (TMC-β) and TC (Table 2). The amount of dihydroxy BSs was 48% ± 18% in Hrn mice, compared to 5% ± 1% in WT mice. To instill a more-humanized BS pool, we fed Hrn and WT mice a BS-supplemented diet containing either 0.1% CA or 0.1% GCDCA. Supplementation with CA resulted in a relative amount of TC of 42% ± 10% in Hrn, compared to 80% ± 3% in WT, mice. The concentration of the secondary BS (TDC), formed from TC by bacterial dehydroxylation, was 50% ± 9% in Hrn, compared to 2% ± 1% in WT, mice (Table 2). TDC is rehydroxylated in WT livers, but not in Hrn mice, because of a lack of CYP activity. Feeding a GCDCA-supplemented diet also led to high dihydroxy BS levels in bile of Hrn mice. After 3 weeks of a 0.1% GCDCA-supplemented diet, the relative amount of GCDCA in WT was 16% ± 1%, compared to 59% ± 2% in Hrn mice, and relative amounts for TCDCA were 2% ± 1%, compared to 25% ± 3% in WT versus Hrn (Table 2).

Table 2. Biliary BS Composition in WT and Hrn Mice Fed a Control Diet or a Diet Supplemented With 0.1% CA or 0.1% GCDCA
BS species (output nmol/min*100 g)WT (n = 5)Hrn (n = 4)P Value
  1. After 3 weeks of feeding a semisynthetic control diet (WT, n = 5; Hrn, n = 4), a purified semisynthetic diet supplemented with 0.1% CA (WT, n = 5; Hrn, n = 4) or a semisynthetic diet supplemented with 0.1% GCDCA (WT, n = 4; Hrn, n = 4), bile was collected and BS composition was determined using an HPLC analyzer. BS species are listed from most hydrophilic to most hydrophobic. Values for output are given as mean ± SEM; between parentheses are the percentages of total. P value for significance is given for the differences in percentages in Hrn mice, compared to WT.

  2. Abbreviations: ND, not detectable; NS, not significant.

TMC-α, %5.9 ± 1.6 (6.5 ± 2.2)6.2 ± 1.2 (19.4 ± 4.0)<0.001
TMC-β, %46.9 ± 5.9 (54.6 ± 5.9)7.5 ± 3.3 (21.8 ± 14.9)0.003
TUDC, %3.6 ± 0.6 (4.1 ± 0.4)6.3 ± 2.2 (19.0 ± 9.4)0.008
TC, %29.9 ± 6.3 (33.6 ± 4.4)3.9 ± 1.3 (11.2 ± 6.4)<0.001
TCDC, %1.1 ± 0.3 (1.2 ± 0.6)8.2 ± 1.0 (28.6 ± 14.2)0.003
Diet0.1% CA0.1% CA0.1% CA
BS species (output nmol/min*100 g)WT (n = 4)Hrn (n = 4)P Value
TMC-α, %1.8 ± 0.6 (1.7 ± 0.7)0.8 ± 0.3 (1.1 ± 0.5)NS
TMC-β, %16.3 ± 3.7 (15.6 ± 3.6)0.3 ± 0.1 (0.4 ± 0.2)<0.001
TUDC, %0.6 ± 0.1 (0.6 ± 0.19)0.6 ± 0.2 (0.9 ± 0.5)NS
TC, %81.4 ± 9.0 (80.3 ± 3.8)35.1 ± 14.4 (42.4 ± 10.3)<0.001
TCDC, %0.3 ± 0.1 (0.3 ± 0.1)3.7 ± 1.2 (5.1 ± 1.1)<0.001
TDC, %1.6 ± 0.4 (1.6 ± 0.8)37.1 ± 11.3 (50.1 ± 9.2)<0.001
Diet0.1% GCDCA0.1% GCDCA0.1% GCDCA
BS species (output nmol/min*100 g)WT (n = 4)Hrn (n = 4)P Value
TMC-α, %13.6 ± 2.4 (6.8 ± 2.5)20.5 ± 8.7 (5.9 ± 3.8)NS
TMC-β, %78.1 ± 6.1 (39.6 ± 11.3)7.5 ± 3.1 (3.1 ± 3.3)<0.001
TUDC, %36.8 ± 4.1 (12.2 ± 8.1)23.1 ± 6.3 (5.1 ± 3.4)NS
TC, %63.0 ± 4.1 (21.1 ± 4.1)4.7 ± 4.4 (1.7 ± 3.2)0.040
TCDC, %4.9 ± 0.7 (2.5 ± 0.7)80.0 ± 21.9 (25.0 ± 2.5)<0.001
TDC, %5.4 ± 0.5 (2.1 ± 1.5)0.9 ± 0.1 (0.2 ± 0.2)0.040
GCDCA, %30.3 ± 5.5 (15.8 ± 8.5)186.2 ± 49.8 (59.1 ± 1.6)<0.001

In both CA- and GCDCA-fed Hrn mice, total BS output was increased, compared to control-fed Hrn mice, and not different from that of CA- or GCDCA-fed WT animals (Fig. 3A). Markedly, serum ALT levels decreased upon BS feeding in Hrn mice, compared to control-fed Hrn mice; in fact, on a 0.1% cholate diet, ALT levels normalized (Fig. 3B).

Figure 3.

Biliary BS species. After 3 weeks of feeding a semisynthetic control diet (WT, n = 5; Hrn, n = 5), a purified semisynthetic diet supplemented with 0.1% CA (WT, n = 5; Hrn, n = 5), or a semisynthetic diet supplemented with 0.1% GCDCA (WT, n = 4; Hrn, n = 4), total biliary BS output was measured (A). (B) ALT levels on different BS-supplemented and control diets for Hrn mice. Open bars: WT mice; closed bars: Hrn mice.

In Hrn mice fed a control diet, biliary cholesterol output was decreased, as expected (Table 1). However, feeding a CA diet for 3 weeks resulted in a very strong increase in biliary cholesterol secretion, even when related to the increased BS secretion. The ratio of cholesterol over BS concentration was 3-fold higher than in WT (Fig. 4A). mRNA levels for the main apical cholesterol (half)transporter in hepatocytes (Abcg5) were not significantly different between WT and Hrn mice (Fig. 4C). The same was found for Abcg8 (Fig. 4E). Protein content of Abcg5 (Fig. 4B) was not significantly different between WT and Hrn mice (Fig. 4B,D). Because cholesterol secretion is up-regulated in CA-fed Hrn mice, we also measured hepatic cholesterol. Hepatic cholesterol was not significantly different in Hrn mice versus WT (data not shown). Furthermore, we measured Npc2 mRNA levels, because Npc2 was postulated to act as a cholesterol acceptor in bile.12 Also, Npc2 tended to be higher (P = 0.08) in Hrn on control diet, but was not significantly different from WT on BS-supplemented diet (Fig. 4F). Scavenger receptor class B1 (SR-B1) expression levels did not show any differences between WT and Hrn, neither on a control nor on a BS-supplemented diet (data not shown).

Figure 4.

Biliary cholesterol excretion after BS feeding. Biliary BS versus biliary cholesterol output (A) in WT (open squares) and Hrn (open circles), relative Abcg5, Abcg8, and Npc2 mRNA expression (C, E, and F), and Abcg5 protein expression (B and D). All animals were fed control of CA 0.1% diet for 3 weeks (WT, n = 5; Hrn, n = 5). (C-F) Mean ± SEM. (B) KO indicates samples from Abcg5−/− mouse. Open bars: WT mice; closed bars: Hrn mice.


In this study, we characterized bile formation in the Hrn mouse model with defective CYP activity by hepatic KO of the single electron donor, CYP oxidoreductase.3 This defect leads to a strongly reduced BS synthesis and BS rehydroxylation capacity. The total BS pool in Hrn mice is reduced to 40% of that in WT mice. The defect in BS synthetic capacity can be better evaluated by fecal BS secretion. Fecal excretion is the only way by which BSs can leave the body and there is little breakdown of BSs in the intestine. Hence, fecal BS output is a direct measure of hepatic BS synthesis, and we observed that fecal output in Hrn mice was reduced to 8% of WT. This coincides with the reduction of cytochrome reductase activity to 5% of normal, as reported on by Henderson et al.3 The discrepancy between the very low BS synthetic capacity and the mildly reduced BS pool is most likely caused by highly efficient reabsorption of BSs in the terminal ileum by the apical BS transporter, Asbt. Expression of this transporter was found to be normal in Hrn mice, compared to WT mice (Fig. 1C).

The advantage of the residual BS synthetic capacity is that Hrn mice do not have the severe phenotype that mice with a complete lack of BSs display. Mice with a gene disruption of Cyp7a1 suffer from neonatal death, and even when mothers were supplemented with fat-soluble vitamins, they still displayed oily coats, hyperkeratosis, and apparent vision defects and steatorrhoea.13, 14 As a result of the small residual BS pool, Hrn mice will not suffer from fat-soluble vitamin deficiency. However, these mice do have increased fecal cholesterol excretion, as might be expected in a BS-deficient state. Insufficient amounts of BSs in the small intestine will lead to reduced cholesterol absorption and, as a consequence, to increased fecal cholesterol excretion.12

Apart from reduced synthetic capacity and in contrast to other mouse models with BS synthetic deficiencies, this mouse model importantly also lacks rehydroxylation capacity for secondary BSs that are reabsorbed in the gut. In healthy mice, content of primary dihydroxy BSs is very low and the secondary dihydroxy BS, deoxycholate, is rapidly rehydroxylated in the liver and is therefore also low. As a consequence, WT mice have approximately 5% dihydroxy BSs, whereas Hrn mice have up to 43% of dihydroxy BSs. The lack of hydroxylation capacity was most strongly illustrated when we intravenously infused taurodeoxycholate and analyzed the secreted biliary BS species. In WT mice, there was a quantitative conversion into taurocholate, but in Hrn mice, this conversion was virtually absent (Fig. 2F). By feeding BSs to these mice, the BS pool can be completely restored in terms of mass (Fig. 3A). When feeding a diet supplemented with cholate or glycochenodeoxycholate, content of dihydroxy BSs can be considerably increased up to levels that also prevail in the human BS pool.

The lack of (re)hydroxylation allowed us to test the classical concept of BS-induced cholestasis in mice more thoroughly than in the past.15 Whereas infusion of the hydrophilic BS, TUDC, evoked identical bile flow and BS secretion in WT and Hrn mice (Fig. 2A), infusion of relatively low amounts of the more-hydrophobic BS, TDC, resulted in a rapid decrease in bile flow in Hrn mice and not in WT mice. A likely explanation for this TDC-induced cholestasis is that as a result of hydrophobic BS excretion, a concequence of the lack of rehydroxylation, excessive amounts of cholesterol are extracted from the membrane. Indeed, cholesterol was initially excreted at normal rates during TDC infusion in Hrn mice, whereas BS and phospholipid excretion were already significantly lowered (Fig. 2C,E). We think that onset of cholestasis is most likely caused by reduced BSEP function during high TDC excretion in Hrn mice. This hypothesis is in line with previous findings of Paulusma et al.,16 who found a very strict correlation between membrane cholesterol content and BSEP activity.

Interestingly, such TDC-induced cholestasis was not observed when Hrn mice were chronically fed with CA, whereas this led to considerable amounts of TDC in bile (50% of total, which is comparable with human BS composition). Apparently, in the presence of healthy defensive machinery17 that mitigates the detergent action of BSs, the hepatocanalicular membrane of the mouse is able to adapt to the excretion of relatively high amounts of TDC. Thus, TDC-induced cholestasis apparently only occurs in a setting of acute infusion of relatively high doses of TDC.

We observed that in Hrn mice fed a cholate-supplemented diet, cholesterol excretion is increased more than in WT animals (Fig. 4A). Biliary cholesterol transport, for a large part, depends on the cholesterol excretion pump, ABCG5/8, and expression of this exporter is increased upon cholate feeding, albeit to similar extent in WT and Hrn mice. Kosters et al. presented evidence for another, ABCG5/8-independent route to secrete cholesterol into bile.18 Wiersma et al showed that this may be partly mediated by SR-B1.19 However, expression of SR-B1 was not significantly different between WT and Hrn mice either (data not shown). Finally, it is possible that increased cholesterol secretion is caused by an increased amount of Npc2 protein in bile. NPC2 is thought to act as an intracellular cholesterol acceptor and as an extracellular cholesterol acceptor in bile and could therefore increase cholesterol output. According to Yamanashi et al., biliary Npc2 acts as an accelerator of cholesterol efflux, and suggests that an increased biliary Npc2 activity can be positively regulated by BSs and phospholipids.20 However, we also found no difference in Npc2 expression between WT and Hrn mice. Given the fact that all these transporters only display marginal differences in expression between WT and Hrn mice, we attribute the increased cholesterol excretion into bile mainly to the higher hydrophobicity of the Hrn BS pool. It has been established that more-hydrophobic BSs induce higher rates of biliary cholesterol excretion, which is likely caused by their higher cholesterol-solubilizing capacity.15, 21-25 This explanation is supported by the observation that, upon infusion of the hydrophilic BS, TUDC, no increase in cholesterol was observed (Fig. 2B).

In conclusion, we show that Hrn mice have an impaired BS synthesis and rehydroxylation capacity. It was possible to feed these animals hydrophobic BSs so as to instill a more-human BS pool. Although acute secretion of large amounts of TDC induced cholestasis, there was no sign of cholestasis during chronic excretion of TDC or GCDCA. Back-crossing these mice with any model for intrahepatic cholestasis would lead to a more-humanized model, gaining better insights in the mechanisms behind cholestasis and hopefully finding a way to improve these devastating diseases.