Potential conflict of interest: Nothing to report.
The orphan nuclear hormone receptor small heterodimer partner (SHP) regulates the expression of several genes involved in bile acid homeostasis in the liver. Because bile acid toxicity is a major source of liver injury in cholestatic disease, we explored the role of SHP in liver damage induced by common bile duct ligation (BDL). Shp−/− mice show increased sensitivity in this model of acute obstructive cholestasis, with greater numbers of bile infarcts and higher mortality than wild-type C57BL/6 mice. This increased sensitivity could not be accounted for by differences in expression of bile acid homeostatic genes 2 or 5 days after BDL. Instead, higher basal expression of such genes, including the key biosynthetic enzyme cholesterol 7α hydroxylase (Cyp7A1) and the bile salt export pump, is associated with both an increase in bile flow prior to BDL and an increase in acute liver damage at only 1.5 hours after BDL in Shp−/− mice, as shown by bile infarcts. At 3 hours, Cyp7A1 expression still remained elevated in Shp−/− with respect to wild-type mice, and the hepatic and serum bile acid levels and total hepatobiliary bile acid pool were significantly increased. The increased sensitivity of mice lacking SHP contrasts with the decreased sensitivity of mice lacking the farnesoid X receptor (FXR; nuclear receptor subfamily 1, group H, member 4) to BDL, which has been associated with decreased intraductal pressure and fewer bile infarcts. Conclusion: We propose that differences in acute responses to BDL, particularly the early formation of bile infarcts, are a primary determinant of the differences in longer term sensitivity of the Fxr−/− and Shp−/− mice to acute obstructive cholestasis. (HEPATOLOGY 2008.)
Bile acids are produced from cholesterol in the liver by series of enzymes, including a variety of cytochrome P450s such as cholesterol 7α-hydroxylase (Cyp7A1).1, 2 Once secreted into the intestine from the gall bladder, bile acids act as physiological detergents to promote the absorption of dietary lipids and fat-soluble vitamins. Because 95% to 97% of bile acids are reabsorbed, primarily in the terminal ileum, and recirculated via portal blood back to the liver, only 3% to 5% of daily bile acids are replaced by synthesis from cholesterol. As detergents, bile acids are potentially toxic, and their overall hepatic levels are tightly regulated. Thus, hepatic synthesis and the enterohepatic circulatory system work coordinately to maintain physiological homeostasis.1, 2 Hereditary and acquired defects in several important hepatic bile transporters such as bile salt export pump (Bsep), canalicular phospholipid flippase (Mdr2/MDR3), and multidrug resistance associated protein 2 (Mrp2) result in the development of cholestatic disorders known as progressive familial intrahepatic cholestasis 2, progressive familial intrahepatic cholestasis 3, and Dubin-Johnson syndrome, respectively.3
A variety of nuclear hormone receptors are involved in bile acid homeostasis.1, 2 Farnesoid X receptor (FXR; nuclear receptor subfamily 1, group H, member 4) is the primary hepatic bile acid sensor4–6 and has been studied most extensively for its regulation of numerous genes involved in bile acid homeostasis. Of particular importance is its induction of the orphan receptor and transcriptional repressor small heterodimer partner (SHP) in response to high hepatic bile acid levels.7, 8 In this autoregulatory paradigm, SHP up-regulation represses expression of Cyp7A1, the rate-limiting enzyme in the neutral pathway of bile acid biosynthesis, as well as several other targets, including the sodium-dependent bile salt importer Na+-taurocholate cotransporting polypeptide [Ntcp; solute carrier family 10 (sodium/bile acid cotransporter family), member 1].9 This model was strongly supported by results with Shp−/− mice, but these studies also revealed additional SHP-independent mechanisms for bile acid repression.10, 11 More recently, it has been suggested that the FXR-dependent induction of expression of fibroblast growth factor 15/19 (FGF15/19) in the intestine also contributes to this negative pathway via activation of the FGFR4 receptor in hepatocytes.12
To further investigate the role of SHP in bile acid homeostasis, we examined the response of C57BL/6 and congenic C57BL/6 Shp−/− mice to acute cholestasis induced by ligation of the common bile duct. In contrast to recent results demonstrating relative resistance of Fxr−/− mice to bile duct ligation (BDL),13, 14 the Shp−/− mice showed increased sensitivity, which was associated with increased bile acid levels and acute liver damage at very early times after BDL.
Shp null mice on a C57BL/6 and 129/Sv mixed background10 were backcrossed with C57BL/6 mice 10 times to generate congenic C57BL/6 Shp−/− mice. Age-matched male Shp−/− congenic and C57BL/6 wild-type (WT) mice, 8 to 16 weeks old, were used throughout this experiment. WT control mice (C57BL/6) were purchased from Harlan Sprague Dawley, Inc. (Indianapolis, IN). Mice were housed in a temperature-controlled room (23°C) in virus-free facilities on a standard 12-hour light/dark cycle (07:00 on, 19:00 off) and were fed standard chow (no. 5001, test diet) and water ad libitum. Prior to BDL, mice were fasted overnight. All animals received humane care according to the criteria outlined in the “Guide for the Care and Use of Laboratory Animals” prepared by the National Academy of Sciences and published by the National Institutes of Health (National Institutes of Health publication 86-23, revised 1985).
Mice were anesthetized by intraperitoneal injection of avertin and aseptically subjected to double ligation of the common bile duct below the bifurcation. Sham-treated animals underwent the same procedure with bile duct exposure, but without ligation. Mice were fasted after BDL or sham operations in the 3-hour and 2-day experiments to match conditions observed with bile duct–ligated Shp−/− mice, which did not consume any food for 2 days after BDL. For 5-day BDL, mice were fed ad libitum.
Following BDL, mice were weighed and anesthetized, and blood was collected from the orbital plexus and centrifuged at 6,000 g for 5 minutes to separate serum. Serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), or bilirubin levels were measured in the Comparative Pathology Laboratory at Baylor College of Medicine. Serum bile acids were measured with a bile acid L3K assay kit (Diagnostic Chemicals, Oxford, CT). After blood collection, mice were euthanized, and tissues were excised and frozen in liquid nitrogen for RNA isolation and bile acid analysis. Liver fragments were fixed in a 4% paraformaldehyde or 10% formalin solution and embedded in paraffin for hematoxylin and eosin staining as described.10
Tissue Bile Acid Analysis.
Liver fragments (50-100 mg) were weighed and homogenized in 1 mL of 75% ethanol. The homogenate was incubated at 50°C for 2 hours to extract bile acids and centrifuged at 6,000 g for 10 minutes. The bile acid content of the supernatant was determined with the bile acid L3K assay kit. For the hepatobiliary bile acid content, the whole liver plus gallbladder was processed via the same protocol. To determine the biliary bile acid level, the gallbladder content was recovered with a 1-mL syringe with a 30-gauge needle. The bile acid concentration was measured with the L3K assay kit as described previously.
Measurement of Bile Flow.
Bile flows of overnight fasted WT and Shp−/− mice were measured as previously described15 with slight modification. Briefly, fasted mice were anesthetized by intraperitoneal injection of avertin. The chest area was opened with midline incision, and the cystic duct and common bile duct were ligated. The common bile duct was cannulated just above the ligated area with a preweighed 30-gauge needle. Bile was collected for 10 to 15 minutes, and bile flow was determined as microliters per minute per 100 g of body weight, assuming a density of 1 g/mL.
Northern Blot Analysis.
Total RNA was isolated from liver and intestinal ileum with Trizol reagent (Invitrogen, Carlsbad, CA). Individual RNA quality was tested in denatured gel electrophoresis, and then equal amounts of total RNA from each tissue were pooled for use as total RNA. Total RNA (15 or 20 μg) from 4 to 6 animals was pooled and subjected to northern blot analysis. Relative expression was determined by densitometry with β-actin as the control. Probes were used as previously described.10, 16
Ten micrograms of liver or intestinal ileum RNA was treated with DNase and reverse-transcribed with Superscript II (Invitrogen), and complementary DNAs were subjected to real-time PCR as described.17 The PCR primer and probe sets were as follows: for Mrp4, forward primer 5′-tttacaagatggttcagcaactgg-3′, reverse primer 5′-ctgtccattggaggtgttcataac-3′, and probe 5′-cgaagccgctgccctcaccgaaac-3′; for Mdr1, forward primer 5′-gacttgatcgtggtgattgagaac-3′, reverse primer 5′-tggaccattgagaagtagatgcc-3′, and probe 5′-caaggagcacggcacccaccagc-3′; and for FGF15, forward primer 5′-gattgccatcaaggacgtcag-3′, reverse primer 5′-tcagcccgtatatcttgccg-3′, and probe 5′-tgcggtacctctgcatgagcgc-3′.
Western Blot Analysis.
Liver protein (30 μg) from naïve WT and Shp−/− mice was subjected to western blot analysis as described.18 Bsep (Kamiya Biomedical Co., Seattle, WA) and β-actin (Cell Signaling Technology, Inc., Danvers, MA) antibodies were used after 1:250 and 1:1000 dilution, respectively.
Analysis of Hepatic and Serum Bile Acid Composition.
Serum or livers from 5 mice of each genotype were pooled and processed for bile acid determination by high-performance liquid chromatography (Shimadzu, Japan) as described.19
Shp−/− Mice Are More Sensitive to BDL.
Initial results revealed that Shp−/− mice showed ill health and increased mortality with respect to congenic C57BL/6 WT mice following BDL. Consistent with this, the Shp−/− mice showed greater increases in serum ALT, AST, and bilirubin levels 5 days after BDL (Table 1). This increased liver damage is consistent with previous observations with mixed-background Shp−/− mice fed a cholic acid–containing diet for relatively short times,11 suggesting that elevated bile acids could contribute to the BDL-induced liver damage. In contrast to previous results in untreated mice,10 serum bile acid levels were not elevated in sham-operated (Table 1) or naïve Shp−/− mice (data not shown) with respect to their Shp+/+ counterparts, and biliary levels were also comparable. In agreement with previous results,10 however, basal hepatic bile acids were modestly elevated in sham-operated or naïve Shp−/− mice. As expected, serum bile acid levels were dramatically elevated in WT mice 5 days after BDL, whereas hepatic levels were more modestly increased and biliary levels were decreased. All these parameters were significantly elevated in the Shp−/− mice.
Table 1. Serum Biochemistry and Bile Acid Levels from Different Tissues Were Compared Between WT and Shp−/− Mice 5 Days After BDL
5 Days After BDL
Values are expressed as means ± standard deviation (n = 3-8 mice per group).
A simple explanation for these results would be increased expression of SHP target genes, particularly the rate-limiting biosynthetic enzyme Cyp7A1, in the Shp−/− mice. Instead, normalized mRNA expression of Cyp7A1 was lower in Shp−/− animals than in WT animals 5 days after BDL (Fig. 1A). Expression of the bile salt exporter Bsep was somewhat elevated, but most other genes associated with bile acid homeostasis, including biosynthetic enzymes and transporters, showed comparable expression in the two strains 5 days after BDL (Fig. 1). The observation that the loss of SHP function did not prevent Cyp7A1 repression in the 5-day BDL mice is consistent with previously described repression of such targets in Shp−/− mice fed bile acid–containing diets for comparable periods of time.10, 11 However, both the decreased Cyp7A1 expression and increased Bsep expression are more consistent with responses to elevated hepatic bile acid levels than to loss of SHP, and these results cannot account for the increased BDL sensitivity of the Shp−/− mice.
Because the increased mortality in the Shp−/− mice was observed predominantly 1 to 2 days after BDL, we focused on earlier time points. At 2 days after surgery, mortality of the WT mice (16%) was comparable to that of previous reports (for example, Stedman et al.14), whereas that of the Shp−/− mice was approximately 2-fold higher (33%) (Table 2). Histologic examination of the livers of Shp−/− mice showed extensive areas of bile infarcts, indicating severe liver injury, which were not observed in WT livers (Fig. 2A). As expected from the 5-day results, serum ALT, AST, and bilirubin were elevated in both genotypes 2 days after BDL, but levels were higher in Shp−/− mice (Fig. 2B). Serum and hepatic bile acid levels were also higher in the Shp−/− mice (Fig. 2C).
Table 2. Comparison of the Mortality of WT and Shp−/− Mice 2 Days After BDL
2 Days After BDL
Mice were subjected to common BDL or a sham operation. The number of dead mice from each genotype 2 days after BDL or the sham operation was counted and depicted.
At the level of gene expression, Cyp7A1 was more dramatically repressed in both strains at 2 days than 5 days (Fig. 3A). There were a number of other differences with respect to the later time point, including comparably elevated expression of Mdr2 in both strains at 2 days rather than the comparably decreased expression at 5 days (Fig. 3B). In addition, several important nuclear receptors known to regulate bile acid metabolism failed to show significant changes in expression in response to BDL or genotype, with the notable exception of a strong down-regulation of Fxr expression after BDL in both Shp−/− and WT mice (Fig. 3C). Overall, these results again fail to identify SHP-dependent alterations in genes associated with bile acid homeostasis that could account for the increased BDL sensitivity of the Shp−/− mice.
Acute Response to BDL in Shp−/− Mice.
The lack of apparent gene expression differences 2 days after BDL suggested that the increased sensitivity was dependent on much earlier differences, and histologic examination of livers obtained 1.5 hours after BDL showed that both bile infarcts (Fig. 4A) and neutrophil infiltration (data not shown) were already evident at this early time point in the Shp−/− mice but not in the WT mice. Consistent with this increased acute liver damage, serum levels of hepatic enzymes (AST and ALT) and bilirubin were elevated in the Shp−/− mice with respect to WT mice 3 hours after BDL (Fig. 4B). Total hepatobiliary bile acid levels were also increased in both sham-operated and BDL Shp−/− mice. Bile acid levels measured in the liver alone or in serum were not elevated in the sham-operated Shp−/− mice but were markedly increased in the Shp−/− animals 3 hours after BDL (Fig. 4C).
As expected from both the increased hepatobiliary bile acid pool and previous results, expression of the key bile acid biosynthetic enzyme Cyp7A1 was substantially elevated in naïve or sham-operated Shp−/− mice with respect to WT mice 3 hours after surgery, and this elevated expression was maintained in the Shp−/− mice 3 hours after BDL (Fig. 5A). A lesser increase in Cyp8B1 expression was also maintained in the Shp−/− mice. Among the transporters, Bsep, Ntcp, and possibly Mrp3 expression was more modestly elevated in the naïve or sham-operated Shp−/− mice with respect to the WT mice. Three hours after BDL, this differential expression appeared to be blunted in the Shp−/− mice. Expression of Mrp2 was comparable in naïve or sham-operated WT and Shp−/− mice but showed a more evident decrease following BDL in both genotypes. In the ileum, expression of the transporters organic solute transporter alpha (Ostα) and ileal bile acid transporter (iBAT) did not differ in either genotype 3 hours or 1 day after BDL. Expression of the FXR target FGF15 decreased in both genotypes at both time points but was also higher in the Shp−/− mice than in the WT animals (Fig. 5B,C).
These results suggest that the increased sensitivity of the Shp−/− mice to BDL is based on preexisting alterations in bile acid homeostasis, particularly increased bile flow due to elevated Bsep mRNA expression. This was confirmed by the elevation in not only Bsep protein expression in the Shp−/− livers (Fig. 6A) but also direct measurement of bile flow (Fig. 6B). As expected, this was associated with increased gallbladder volume (Fig. 6C).
We also measured the basal bile acid composition in the serum and liver of these two genotypes to determine whether alterations in this parameter could also contribute to the observed liver damage in Shp−/− mice. As shown in Fig. 7, the contributions of hydrophobic bile acids (cholic acid, chenodeoxycholic acid, and deoxycholic acid) to the total bile acid pools are somewhat higher in both the liver and serum of Shp−/− mice than in WT mice. This corroborates the previous data obtained from mixed-background Shp−/− mice11 and suggests that the increased hydrophobicity of bile acids may contribute to the increased liver damage observed in the BDL model of Shp−/− mice.
SHP is an atypical orphan nuclear receptor that lacks the conserved DNA binding domain.20 It is a transcriptional repressor and exerts its regulatory functions through protein-protein interactions with other nuclear hormone receptors and possibly other transcription factors, which can inhibit or even reverse their transactivation.20–25 Regulation of bile acid metabolism is the most prominent of a number of physiological roles proposed for SHP. The primary basis for this effect is the induction of SHP by FXR in the presence of elevated bile acid levels, inhibiting transactivation by the nuclear receptors liver receptor homolog-1 (LRH-1) and possibly hepatocyte nuclear factor 4 alpha (HNF4α) and repressing expression of Cyp7A1, the rate-limiting enzyme of bile acid biosynthesis, as well as a number of other bile acid biosynthetic enzymes.7, 8, 23–25 This function of SHP has been strongly supported by results with Shp−/− mice, which reveal defective repression in response to synthetic FXR agonists10, 11 or very short term dietary bile acid treatment.26 However, such studies have also demonstrated the importance of SHP-independent pathways that can efficiently repress Cyp7A1 and other negative targets in response to longer term treatments with dietary bile acids.10, 11
The increased expression of Cyp7A1 and other targets in the Shp−/− mice results in an increase in the hepatobiliary bile acid pool prior to BDL, as demonstrated here and in previous studies.10, 11 Both this increase and the potential defect in negative feedback regulation of bile acid production would be expected to increase the sensitivity of Shp−/− mice to the toxic consequences of a further increase in bile acid levels, and previous results have confirmed such increased sensitivity of Shp−/− mice to short-term effects of dietary bile acids,11 although they are relatively resistant to chronic (3-month) treatments.16 The current studies demonstrate that Shp−/− mice also show markedly increased sensitivity to acute obstructive cholestasis.
However, the increased sensitivity of the Shp−/− mice is clearly not due to defective repression of Cyp7A1 or other targets subsequent to BDL, at least at the time points examined. Thus, 2 and 5 days after BDL, the repression of SHP-negative target genes is, if anything, greater in the Shp−/− mice, perhaps as a consequence of increased hepatic bile acid levels (Table 1 and Fig. 2C). At the much earlier time point of 1.5 hours after BDL, when the WT mice show no histologic evidence of liver damage, the Shp−/− mice show numerous bile infarcts, zones of severe necrosis produced by rupture of bile ducts due to elevated hydrostatic pressure (Fig. 4A). This is well before the expected impact of differential transcriptional responses of the two genotypes to BDL. More specifically, it also precedes the decrease in mRNA levels of SHP targets such as Cyp7A1, which is present at apparently unaltered levels in both genotypes 3 hours after BDL. The preexisting increase in the bile acid pool was associated with an increase in the bile flow rate (Fig. 6B), which is presumably due to the increased expression of Bsep, considered the primary driver of bile flow.27 It is likely that this increased flow compensates for the increased Cyp7A1 levels and accounts for the normal hepatic bile acid levels in the Shp−/− mice. Previous results with ursodeoxycholic acid–treated Mdr2−/− mice elegantly demonstrated that an increase in bile acid infarcts upon BDL was a consequence of increased bile flow and subsequently increased intraductal pressures.28 Results of genetic and pharmacologic manipulation of apoptotic pathways confirm the dominant role of necrosis in BDL-induced liver damage in mice.29 Thus, we suggest that increased biliary pressure in response to BDL accounts for the increased formation of bile infarcts and liver damage in the Shp−/− mice.
The increased sensitivity of the Shp−/− mice to BDL provides a dramatic contrast with the decreased sensitivity of Fxr−/− mice,13, 14 which seems paradoxical on the basis of the coordinate function of the two receptors in negative regulation of bile acid production. At later times after BDL, the decreased sensitivity of the Fxr−/− mice has been linked to increased detoxification of bile acids within hepatocytes and also increased transport into the blood and clearance by the kidneys.13 More importantly, in the context of the current results, Fxr−/− mice also show substantially decreased basal expression of Bsep both prior to and following BDL, which was directly linked to decreased intrabiliary hydrostatic pressure.30 This decreased pressure can neatly account for the reduction in bile infarcts and liver damage in the Fxr−/− mice.14, 30 More generally, we note that both the predicted and observed effects of loss of the transactivator FXR or the repressor SHP are similar for negative targets but opposite for positive targets. Thus, Cyp7A1 expression increases in basal circumstances in both cases, as expected. However, the just noted decrease in Bsep expression in Fxr−/− mice contrasts with the increase in the Shp−/− mice, which could be due to loss of direct repressive effects, as suggested by results with SHP transgenic mice,31 increased FXR transactivation due to elevated bile acid levels, or both. We suggest that differences in such positive targets, particularly opposite effects on Bsep and intraductal pressures, are primary determinants in the opposite responses of the Shp−/− and Fxr−/− mice to BDL and that manipulation of SHP gene expression is a potential therapeutic target to treat chronic cholestasis. We have recently generated doubly mutant Fxr−/−Shp−/− mice, which show low basal Bsep expression, as expected from the strong regulation of this target by FXR. Preliminary results confirm the clear prediction that their acute response to BDL follows that of their Fxr−/− progenitors rather than their Shp−/− progenitors.
We thank Dr. Marie-Louise Ricketts and Dr. Vijay Yechoor for technical advice on real-time PCR and bile duct cannulation, respectively, and Akira Sasaki for technical support on high-performance liquid chromatography.