Cholic acid mediates negative feedback regulation of bile acid synthesis in mice

Cholesterol is converted into dozens of primary and secondary bile acids through pathways subject to negative feedback regulation mediated by the nuclear receptor farnesoid X receptor (FXR) and other effectors. Disruption of the sterol 12α-hydroxylase gene (Cyp8b1) in mice prevents the synthesis of cholate, a primary bile acid, and its metabolites. Feedback regulation of the rate-limiting biosynthetic enzyme cholesterol 7α-hydroxylase (CYP7A1) is lost in Cyp8b1^–/– mice, causing expansion of the bile acid pool and alterations in cholesterol metabolism. Expression of other FXR target genes is unaltered in these mice. Cholate restores CYP7A1 regulation in vivo and in vitro. The results implicate cholate as an important negative regulator of bile acid synthesis and provide preliminary evidence for ligand-specific gene activation by a nuclear receptor.

The synthesis and excretion of bile acids comprise the major pathway of cholesterol catabolism in mammals. Synthesis provides a direct means of converting cholesterol, a hydrophobic and insoluble molecule, into a water-soluble and readily excreted form, the bile acid. The biosynthetic steps that accomplish this transformation also confer detergent properties to the bile acid that are exploited by the body to facilitate the direct secretion of cholesterol from the liver and the solubilization of essential nutrients from the diet. These roles in the elimination of cholesterol are counterbalanced by the ability of bile acids to solubilize dietary cholesterol and to promote the delivery and accumulation of this sterol in the liver.

Bile acid synthesis is tightly regulated via a classical feedback mechanism to ensure that sufficient amounts of cholesterol are catabolized to maintain homeostasis and provide adequate emulsification in the intestine. The accumulation of bile acids causes a decrease in the transcription of the cholesterol 7α-hydroxylase gene (Cyp7a1), which encodes the first and rate-limiting enzyme in the major biosynthetic pathway (1). Conversely, the loss of bile acids results in enhanced transcription from this gene and a subsequent increase in biosynthetic output. Six members of the nuclear receptor family of transcription factors are known to regulate the expression of Cyp7a1 (2, 3). These include two proteins mediating repression of the gene, the farnesoid X receptor (also known as FXR, NR1H4) and the short heterodimer partner (SHP, NR0B2); three proteins that activate the gene, CYP7A1 promoter transcription factor (CPF, FTF, LRH-1, NR5A2), liver X receptor-α (LXR-α, NR1H3), and hepatic nuclear factor-4α (HNF-4α, NR2A1); and isoforms of the retinoid X receptor (RXR, NR2B1-3) that act as essential partners for FXR and LXR.

When an organism is replete with bile acids, further synthesis is prevented by FXR, which upon binding bile acids (4–6) activates transcription of the Shp gene. SHP in turn binds to and inhibits the positively acting CPF receptor, causing a decrease in Cyp7a1 expression (7, 8). This regulatory circuit is less active when bile acids are depleted, allowing CPF and HNF-4α to stimulate transcription of Cyp7a1 and to restore the bile acid pool (9). In some species, including rats and mice, the expression of Cyp7a1 is increased by the accumulation of cholesterol in the liver (1). LXR and its ligands are responsible for this enhancement (10–13). Studies in knockout mice lacking LXR (13), FXR (14), or SHP (15, 16) confirm these observations and suggest the existence of additional regulatory mechanisms affecting bile acid synthesis that may be receptor independent.

Bile acids are remarkably diverse in structure and exist in many different forms within vertebrates. This complexity is evident in the mouse, in which chemical analyses have identified two primary bile acids, cholate and β-muricholate, which are the immediate products of the biosynthetic pathways, and seventeen secondary bile acids, which arise from the actions of the gut flora on primary bile acids (17). A long-standing question in the field is why so many different types of bile acids are present in a single organism (18). A diverse mixture of bile acids with different detergent properties may be required to ensure solubilization of the many hydrophobic nutrients encountered by the small intestine upon consumption of a varied diet. This prediction is supported by studies in which the effects of dietary bile acids on the solubilization of cholesterol are monitored (19). In this experimental situation, the fed bile acid accounts for a majority of the pool, and depending on its structure, either increases or decreases the amount of cholesterol absorbed by the small intestine. Alternatively, individual bile acids may differ in their ability to activate FXR and hence to regulate the synthesis of bile acids.

In the current studies we characterize a line of mice in which the capacity to synthesize the primary bile acid cholate is eliminated by targeted disruption of the sterol 12α-hydroxylase gene (Cyp8b1). This alteration has some expected consequences for bile acid and cholesterol metabolism, including a change in the composition of the bile acid pool and a decrease in intestinal cholesterol absorption, as would be predicted from previous studies in mice with altered bile acid biosynthesis. Unexpectedly, the mutant mice have an expanded bile acid pool and an increased rate of bile acid synthesis. These phenotypic traits are traced to constitutive expression of cholesterol 7α-hydroxylase. Since other FXR target genes are expressed normally in these animals, and the capacity to respond to cholate by repression of 7α-hydroxylase remains in the Cyp8b1^–/– mice, the data suggest that cholate is responsible for feedback regulation of bile acid biosynthesis.

Deletion of the Cyp8b1 gene. Sterol 12α-hydroxylase is a microsomal cytochrome P450 that catalyzes the addition of a hydroxyl group to carbon 12 of a number of sterol intermediates destined to become cholate in the bile acid biosynthetic pathways (Figure 1a). The enzyme is encoded by a gene (Cyp8b1) that lacks introns and is located on mouse chromosome 9qF4 (20). To disrupt Cyp8b1, a plasmid for use in homologous recombination was constructed in which the coding region of the gene was replaced with the E. coli β-galactosidase gene (lacZ) (Figure 1b). This DNA was electroporated into embryonic stem cells, which were then grown in the presence of ganciclovir and G418. Screening of 1,300 colonies identified a single isolate with the desired recombination event. Injection of this colony into blastocysts produced 12 chimeric males (>90% agouti), of which several transmitted the mutation through the germline. Crosses between Cyp8b1^+/– mice produced the expected number of wild-type, heterozygous, and homozygous offspring, and the survival rates of mice of each genotype were indistinguishable. CYP8B1 expression is limited to the liver (29), and analysis of RNA from this tissue revealed that the introduced mutation eliminated the encoding mRNA (Figure 1c).

Bile acid composition. The consequences of loss of Cyp8b1 on the types of bile acids present in the mutant mice were determined by gas chromatography/mass spectrometry (Figure 2). The primary bile acids cholate and β-muricholate together made up about 85% of the bile acids present in a male wild-type mouse (Figure 2). ω-Muricholate accounted for approximately 6% of the pool, and the remainder consisted of small amounts of α-muricholate (4%), deoxycholate (1%), chenodeoxycholate (0.7%), and ursodeoxycholate (0.6%). In contrast to this makeup, no cholate or its secondary metabolite, deoxycholate, were detected in Cyp8b1^–/– mice (Figure 2), indicating that deletion of the gene eliminated CYP8B1 activity and that there were no compensating enzyme activities in the mutant mice. The fraction of the pool normally occupied by cholate was replaced mostly by increases in the muricholates. There were also substantial increases in the proportion of chenodeoxycholate (to 15%) and ursodeoxycholate (to 9%) in these animals (Figure 2). Similar alterations in the composition of bile acids were found in female knockout mice (data not shown).

Figure 2

Composition of gallbladder bile in male wild-type and Cyp8b1^–/– mice. Amounts of individual bile acids in the gallbladder bile of wild-type and Cyp8b1^–/– mice were determined by gas chromatography/mass spectrometry and are indicated as a percentage of the entire pool. The position and stereochemistry of hydroxyl groups on the ring structures of each bile acid are indicated in parentheses. Similar results were obtained when the total bile acid pool was analyzed by HPLC.

Bile acid pool size and synthetic rate. Changes in the types of bile acids present in Cyp8b1^–/– mice were accompanied by an increase in the pool size, that is, the total mass of bile acids present in the animal (Figure 3a). The pool size in Cyp8b1^–/– males was increased by 37% relative to that in wild-type controls, while that in Cyp8b1^–/– females was 20% larger than in the wild-type. Both of these increases were statistically significant (P < 0.05). Female wild-type mice typically have 35–40% larger bile acid pools than do males, as a consequence of increased output from the CYP7A1 pathway (30). This sexually dimorphic trait was maintained in the Cyp8b1^–/– mice (Figure 3a), indicating that male/female differences in bile acid synthesis cannot be ascribed to alterations in CYP8B1 activity.

Figure 3

Bile acid pool sizes and excretion rates in wild-type and Cyp8b1^–/– mice. (a) The mass of bile acid in the enterohepatic circulation of normal and mutant male and female mice (n = 6 per group) was determined by extraction, purification, and quantitation by HPLC. Values (mean ± SEM) determined in the Cyp8b1^–/– mice were significantly different (P = 0.0037 for males; P = 0.04 for females by two-tailed Student t test) from those of normal controls. (b) Fecal bile acid excretion rates, a measure of hepatic bile acid synthesis, were determined as described in Methods. Excretion in Cyp8b1^–/– mice was significantly greater (P = 0.003 for males; P = 0.0002 for females) than that in wild-type mice. (c) Radiolabeled cholesterol and a nonabsorbable recovery standard (sitostanol) were administered by gavage to male and female mice (n = 6 per group) of the indicated genotypes. The percentage of cholesterol absorbed was determined after extraction of sterols from the feces excreted from the animals over a 72-hour period. Absorption values measured in mutant mice were significantly smaller (P = 0.02 for males; P = 0.0007 for females) than those in wild-type controls. (d) Fecal neutral sterols (cholesterol, coprostanol, and cholestanone) were quantitated by gas chromatography. Reduced absorption (see c) led to a significantly increased sterol excretion rate in Cyp8b1^–/– mice (P < 0.0003 for males and females). Bw, body weight.

Rates of bile acid synthesis in the liver were indirectly assessed by measuring the amount of bile acid excreted into the stool per day. As shown by the data in Figure 3b, this parameter was significantly increased in both male and female Cyp8b1^–/– mice. Since a majority of bile acids are synthesized via the CYP7A1 pathway in the mouse (19), these results suggested that the output from this highly regulated pathway was elevated in the knockout animals.

Cholesterol absorption and excretion. Variations in bile acid pool size and composition affect the amount of cholesterol absorbed from the small intestine. In general, increases in the pool size and/or the content of hydrophobic bile acids in the pool cause increases in sterol absorption and vice versa. This relationship is illustrated in the sexes of wild-type mice; males absorb 44% of a bolus of cholesterol delivered intragastrically whereas females absorb 56% (Figure 3c). The difference is attributable to an increase in the content of cholic acid in the female pool (30) and to an increase in the pool size in this sex (Figure 3a).

Both male and female Cyp8b1^–/– mice absorbed less cholesterol than their wild-type counterparts (Figure 3c), despite having larger bile acid pool sizes. Furthermore, there was no statistical difference (P = 0.48) between the absorption values measured in knockout males (27%) and those in knockout females (30%), even though Cyp8b1^–/– females had a larger pool size than did males (Figure 3a). These data suggested that of the two parameters of bile acid metabolism affecting solubilization and intestinal cholesterol absorption, composition was more important than pool size.

The observed differences in cholesterol absorption were confirmed by measurement of neutral sterol levels in the feces. As illustrated in Figure 3d, male and female Cyp8b1^–/– mice excreted significantly more neutral sterols (cholesterol, cholestanone, and copro-stanol) than did wild-type mice. When the levels of total fecal lipid (fats plus neutral sterols) were determined, these too were found to be elevated in knockout males (6.1% ± 0.05% versus 5.0% ± 0.04%, P = 0.02) and females (6.3% ± 0.07% versus 5.5% ± 0.04%, P = 0.005). Approximately 2% of fecal lipid is composed of neutral sterols, while triacylglycerols and fatty acids account for about 98%. The increases in total fecal lipid in the Cyp8b1^–/– mice thus indicate that the animals exhibited malabsorption of both sterols and fats (steatorrhea).

The decrease in absorption of dietary cholesterol in the Cyp8b1^–/– mice was accompanied by a significant increase in the synthesis of cholesterol by the liver. The de novo rate of hepatic sterol synthesis in male knockout mice was 2.3 ± 0.4 μmol/g tissue/h versus 1.0 ± 0.2 μmol/g tissue/h in wild-type controls. Cholesterol biosynthetic rates were not significantly different in other tissues of the mutant mice, including the kidney and adrenal gland, and triglyceride synthesis rates were unchanged in all tissues examined. The compensatory increase in hepatic sterol synthesis in response to sterol malabsorption in the gut maintained whole-body cholesterol homeostasis, since no statistically significant differences were detected in serum cholesterol and triglyceride levels, and the serum lipoprotein profiles of normal and mutant mice were superimposable (data not shown).

Cyp7a1 expression. The increases in bile acid pool size and excretion in the Cyp8b1^–/– mice suggested that output from the CYP7A1 pathway was increased. The data for Figures 4a and 4c showed that CYP7A1 mRNA and protein levels were substantially increased in both male and female Cyp8b1^–/– mice, and that the increases approached those seen in sterol 27-hydroxylase–deficient (Cyp27a1^–/–) animals, which have a defect in bile acid synthesis (see below). Levels of CYP7A1 enzyme activity (Figure 4b) were also increased in the Cyp8b1^–/– mice, although not to the degree of the mRNA and protein (Figure 4, a and c). No CYP7A1 mRNA or protein was detected in mice bearing a null mutation in the gene (Cyp7a1^–/– mice), and no changes in the levels of a control mRNA (cyclophilin) or microsomal protein (immunoglobulin heavy-chain binding protein; BIP) were measured in the mutant mice (Figure 4, a and c).

Figure 4

Expression of CYP7A1 in wild-type and knockout mice. (a–c) CYP7A1 mRNA, enzyme activity (expressed as pmol product per minute per mg microsomal protein), and protein were assessed by blotting, mass spectrometry, and immunoblotting, respectively, in male and/or female mice (n = 6 per group) of the indicated genotypes. Cyp27a1^–/– mice have a reduced bile acid pool size and synthetic rate due to an induced mutation in the gene for sterol 27-hydroxylase, an enzyme involved in the synthesis of all primary bile acids. Cyp7a1^–/– mice served as negative controls in these experiments. BIP, immunoglobulin heavy-chain binding protein.

The expression levels of other enzymes involved in bile acid synthesis, including the CYP7B1 oxysterol 7α-hydroxylase, the CYP27A1 sterol 27-hydroxylase, and the CYP39A1 oxysterol 7α-hydroxylase, were unchanged in Cyp8b1^–/– mice based on RNA blotting and immunoblotting (data not shown, and see below). As CYP7B1 and CYP39A1 expression are regulated by the nuclear receptor HNF-4α (31) these findings suggested that the increase in CYP7A1 mRNA was specific and not related to a global change in hepatic transcription. Steady-state levels of CYP3A11 mRNA, which reflect the activity of the pregnane X receptor (32, 33), were the same in the livers of wild-type and Cyp8b1^–/– mice, and as expected were induced in Cyp27a1^–/– mice (34).

FXR target gene expression. Mice with deficiencies in bile acid biosynthetic enzymes, such as the Cyp7a1^–/– and Cyp27a1^–/– animals, have reduced bile acid pool sizes and biosynthetic rates. These reductions decrease ligand availability and impair the ability of FXR to activate SHP, which in turn leads to activation of bile acid synthesis of the Cyp27a1 gene (Figure 4, a and c) (7, 23, 35). The same mechanism could not explain the increased expression of the enzyme in the Cyp8b1^–/– mice, since these animals have an elevated bile acid pool size and biosynthetic rate (Figure 3). Never-theless, levels of hepatic SHP mRNA were reduced by about 50% in Cyp8b1^–/– mice as judged by RNA blotting (Figure 5a), and by 46% in real-time PCR experiments (Figure 5b). In contrast to these results, the hepatic expression of other FXR target genes (36, 37), including the bile salt export protein (BSEP, also known as ABCB11) and the Na⁺-taurocholate cotransporting polypeptide (Ntcp, or SLC10A1), were unchanged in the Cyp8b1^–/– mice (Figure 5a). The mRNA level of organic anion transporter 2 (Oatp2, or SLC21A3) was no different, and FXR mRNA was elevated about twofold in the mutant mice. The latter two genes, Slc21a3 and Fxr, are dependent on the transcription factor HNF-1α for expression (38), and thus these results indicated that the observed changes in SHP expression were not due to decreased HNF-1α activity. Similarly, the amounts of the scavenger receptor BI (SR-BI) and CYP7B1 mRNAs, which are indirect and direct targets of HNF-4α, respectively (31), were unchanged (Figure 5a).

Figure 5

Gene expression in wild-type and Cyp8b1^–/– mice. (a) mRNA levels for the indicated genes were determined by blotting in male mice (n = 6 per group) of the indicated genotypes. Autoradiographic exposure times varied from 7 to 18 hours. (b and c) SHP and IBABP mRNA levels were quantitated in the livers and/or five intestinal segments (proximal to distal, S1-1 to S1-5) of male mice of the indicated genotype by real-time PCR in the presence of the intercalating dye SYBR Green. Levels of cyclophilin mRNA in each sample were used to normalize the results. The numbers shown are the averages of triplicate determinations for each mRNA and tissue. Oatp2, organic anion transporter 2 (also known as SLC21A3); SR-BI, scavenger receptor BI.

Levels of the FXR-responsive IBABP mRNA in the small intestine were also reduced by about 50% in the distal 20% of the small intestine (Figure 5c). Data obtained from Cyp8b1^–/– and control Cyp27a1^–/– and Cyp7a1^–/– mice confirmed that the expression of SHP was reduced in the small intestine when the bile acid pool size was changed (Figure 5b), and that the expression of BSEP and Ntcp mRNAs was variably affected in mice of these genotypes (Figure 5a).

Real-time PCR measurements of mRNAs encoding cholesterol supply pathway genes indicated that hydroxymethylglutaryl (HMG) CoA reductase mRNA was increased 1.5-fold and HMG CoA synthase mRNA was increased 1.4-fold in the knockout mice. These results were in agreement with the observed increase in hepatic de novo cholesterol synthesis in the Cyp8b1^–/– mice. The levels of LDL receptor mRNA were elevated 1.1-fold in the mutant mice, suggesting that most of the lost dietary cholesterol was replaced via the de novo biosynthetic pathway and not via receptor-mediated endocytosis of lipoproteins.

In a final experiment, the hepatic expression of approximately 400 genes involved in lipid metabolism was compared between wild-type and Cyp8b1^–/– mice using cDNA microarray hybridization. The gene with the largest fold-increase (4.7-fold) was Cyp7a1, while the gene with the largest fold-decrease (to undetectable levels) was Cyp8b1. The chip hybridization results for other genes agreed with those obtained by RNA blotting (Figure 5a). The outcomes from these three types of gene expression experiments revealed that changes in the bile acid composition and increases in the pool size in the Cyp8b1^–/– mice had differential effects on cholesterogenic and FXR target gene expression, impairing the transcription of some FXR-responsive genes (e.g., Shp and Ibabp) but leaving that of others (e.g., Abcb11 and Slc10a1) neither increased nor decreased.

Cholate responsiveness in Cyp8b1^–/– mice. To determine whether regulatory responses to bile acids had been lost in the Cyp8b1^–/– mice, groups of wild-type and mutant animals were fed chow diets supplemented with different amounts of cholate or chenodeoxycholate for 4 days. As expected, consumption of these diets altered the levels of the fed bile acid in the pool (see Figure 6 legend) and decreased de novo bile acid synthesis. The addition of 0.1% cholate to the diet caused a 70% decrease in the level of CYP7A1 mRNA in wild-type mice, while the addition of 0.25% cholate caused a decline of greater than 98% (Figure 6). Chenodeoxycholate at 0.1% caused a 25% decrease in the mRNA, while diets with 0.25% caused a 90% decline. Similar results were obtained in the Cyp8b1^–/– mice (Figure 6). SHP mRNA levels correlated inversely with those of CYP7A1 in mice of both Cyp8b1 genotypes, and again, cholate appeared to be more efficacious than chenodeoxycholate in inducing SHP mRNA levels (Figure 6).

Figure 6

Responses to dietary bile acids in wild-type and Cyp8b1^–/– mice. Male animals (n = 4–5 per genotype) were fed the indicated amounts of cholate or chenodeoxycholate (cheno) for a period of 4 days. mRNA levels for CYP7A1 and SHP were determined by real-time PCR. The numbers shown are the averages of triplicate determinations for each mRNA sample. Error bars indicate SEM. In wild-type mice fed 0.1% cholate, the composition of gallbladder bile was 15% muricholates, 78% cholate, 5% deoxycholate, and 2% other; in those fed 0.25% cholate it was 5% muricholates, 73% cholate, 15% deoxycholate, and 7% other. In those fed 0.1% chenodeoxycholate it was 58% muricholates, 13% cholate, 17% chenodeoxycholate, and 12% other, and in those fed 0.25% chenodeoxycholate it was 36% muricholates, 2% cholate, 47% chenodeoxycholate, and 15% other. Bile acid composition was similar in Cyp8b1^–/– mice fed the same diets.

These data indicated that the capacity to respond to bile acids in the liver was intact in the Cyp8b1^–/– mice and that restoration of either cholate or chenodeoxycholate and their metabolites to the pool led to appropriate regulatory events. They also suggested that cholate was a more potent mediator of feedback regulation than chenodeoxycholate; however, this interpretation was tempered by the fact that the percentage of the fed bile acid in the pool differed between cholate and chenodeoxycholate, as did the amounts of other primary and secondary bile acids (Figure 6 legend).

Gene regulation in vitro. To examine the individual contributions of bile acids to regulation, we tested their abilities to activate different FXR target genes in vitro using hepatocytes prepared from wild-type and FXR^–/– mice. These experiments used the taurine-conjugated forms of the bile acids and thus were designed to mimic the in vivo situation. As indicated by the data shown in Figure 7a, taurocholic acid and the FXR agonist GW4064 (8) efficiently induced endogenous Shp expression in hepatocytes from wild-type mice. Chenodeoxycholate was less potent and less efficacious, whereas β-muricholate was inactive as an inducer. The levels of SHP mRNA were unchanged in response to bile acids or GW4064 in hepatocytes derived from FXR^–/– mice (Figure 7b).

Figure 7

Regulation of gene expression by different bile acids in primary hepatocytes. Hepatocytes were isolated from the livers of wild-type and FXR^–/– male mice and treated with the indicated concentrations of taurine-conjugated bile acids or the FXR-specific agonist GW4064 for 16 hours. Total RNA was isolated from the cells, and the levels of SHP and BSEP mRNA were quantitated by real-time RT-PCR. The numbers shown are the averages from three experiments and triplicate determinations for each mRNA. Error bars indicate SEM. TCA, taurocholic acid (circles); TCDCA, taurochenodeoxycholic acid (symbol X); TβMCA, tauro-β-muricholic acid (squares); GW4064, 3-(2,6-dichlorophenyl)-4-(3′-carboxy-2-chloro-stilben-4-yl)-oxymethyl-5-isopropyl-isoxazole (48) (triangles).

In contrast to the response of SHP, levels of BSEP mRNA were induced equally well by cholate, chenodeoxycholate, and GW4064 in hepatocytes from wild-type mice (Figure 7c). The addition of β-muricholate to these cells did not increase BSEP mRNA, and the responses of this mRNA to bile acids and GW4064 were reduced in the FXR^–/– mice (Figure 7d).

Expression of the CYP7A1 and CYP8B1 mRNAs was lost upon isolation and plating of hepatocytes from mice of both FXR genotypes, and therefore we were unable to determine whether expression of SHP in response to cholate led to repression of these genes.

The current results have implications for ligand-activated gene transcription mediated by nuclear receptors and for the regulation of bile acid metabolism. They indicate that cholate, a primary bile acid in the mouse, is an in vivo mediator of negative feedback regulation of bile acid biosynthesis in this species. Loss of cholate, one of many potential ligands for the transcription factor FXR, from the bile acid pool causes a decrease in the hepatic expression of a negative regulatory factor (SHP), which in turn causes a loss of repression of the Cyp7a1 gene and an increase in biosynthetic output. This dysregulation appears to be specific, since transcription from other FXR-dependent genes in the liver (e.g., BSEP) is unaffected by the removal of cholate, and all expected regulatory responses are restored upon feeding cholate. Based on current understanding, one interpretation of these results is that the expression of SHP requires the binding and activation of FXR by cholate, whereas expression of BSEP by this receptor is accomplished by the binding of other bile acids. Alternatively, cholate may modulate SHP-independent negative feedback pathways that regulate CYP7A1 expression.

With respect to bile acid synthesis, these data underscore the important role of CYP8B1 in regulating the pathway and indicate why the expression of this enzyme is subject to negative feedback regulation. The inhibition of CYP8B1 expression by bile acids accomplishes three important goals. First, it acutely decreases the synthesis of a ligand required by FXR for feedback regulation. Second, it ensures that the system will gradually return to normal as the level of cholic acid is progressively reduced with each cycle of the enterohepatic circulation; the rate at which normalcy is reestablished becomes a function of the half-life of cholic acid in the system, which is among the shortest of all bile acid half-lives (39, 40). Third, as the half-life of a bile acid is largely dependent on factors in the gut rather than the liver (41), it allows one component tissue of the enterohepatic system to modulate a regulatory event occurring in another.

How might different bile acid ligands change the target specificity of FXR? Unique transcriptional responses could be mediated at the DNA level by the particular sequence or context of the response element (42). The FXR response element consists of two inverted repeats separated by one base pair (43), and variants of this sequence have been identified in the mouse SHP gene (AGGTGAATAACCT) (7, 8), the IBABP gene (AGGTGAATAACCT) (4), and the BSEP gene (GGGACATTGATCC) (36). The response elements in the SHP and IBABP genes are identical in sequence, and we note that the expression of both of these genes is decreased in the absence of cholate (Figure 5). Their response elements thus may be recognized by FXR in a ligand-dependent manner.

Bile acid–dependent responses could be accomplished at the protein level by the recruitment of different coregulators to a target gene following activation by a ligand. The ability of different bile acids to bind and activate FXR has been examined in transfection and binding experiments (4–6, 44). These studies show that chenodeoxycholate is a far better ligand than cholate, and in fact, in some (4, 5, 44) but not all (6) transfection assays with monkey CV-1 and human HepG2 cells, cholate is inactive. Cholate was similarly unable to stimulate the binding of FXR to peptides derived from the p160 family of coactivator proteins. The inactivity of cholate in these assays contrasts with its very broad distribution in the bile of mammals (18), the elevated levels of CYP7A1 mRNA observed in Cyp8b1^–/– mice (Figure 4), its efficient repression of bile acid synthesis in the whole animal (Figure 6), and the ability to preferentially stimulate SHP expression in primary hepatocytes (Figure 7). The discrepancies between the earlier findings and those presented here suggest there may be a hepatic coactivator protein that is missing from the immortalized cell lines and that is required for FXR to specifically stimulate SHP transcription.

Deletion of the Cyp8b1 gene prevents the synthesis of cholate in the mutant animal and at the same time eliminates all of the secondary bile acids derived from cholate. Chief among these is deoxycholate, which is the most abundant bile acid excreted in the stool of the mouse (17). This bile acid was toxic at concentrations over 10 μM in the in vitro hepatocyte experiments reported in Figure 7, and for this reason, we cannot be certain that cholate is the only ligand responsible for the gene-specific regulation reported here. Nevertheless, it is clear that a 12α-hydroxyl group must be present on the active metabolite for proper regulation of bile acid synthesis. If a secondary bile acid like deoxycholate is responsible for FXR-mediated activation of SHP, then the flora of the gut play an even more important role in bile acid and sterol metabolism than previously suspected (45).

The existence of multiple ligands for a single receptor extends the dynamic range over which transcription from a target gene can be regulated. The response to a particular ligand is unlikely to be all or none, as indicated by the results obtained with chenodeoxycholate feeding in Cyp8b1^–/– mice (Figure 6) and the results obtained with this ligand in hepatocytes (Figure 7). Chenodeoxycholate clearly is able to mediate feedback repression of CYP7A1 when elevated in the pool by feeding, and to induce SHP expression at higher concentrations in vitro, but in both of these systems it is less potent than cholate. This differential response is most clearly observed in Cyp8b1^–/– mice maintained on normal chow. In these mice, chenodeoxycholate accounts for about 15% of the bile acid pool (Figure 2), yet this level is unable to suppress output from the Cyp7a1 gene. In contrast, BSEP mRNA expression is normal in the mutant mice (Figure 5a). These concentration-dependent effects suggest that the binding of more potent ligands like cholate to FXR amplifies the activation properties of the receptor.

An analogous situation exists in the case of the androgen receptor, which upon binding testosterone or dihydrotestosterone mediates different developmental programs involved in formation of the male phenotype. Mutations in the gene that synthesizes dihydrotestosterone block one of these pathways and typically lead to male pseudohermaphroditism (46); however, if the concentration of testosterone is elevated in dihydrotestosterone target tissues, the effects of the mutation are ameliorated and male development proceeds (47).

Studies in SHP knockout mice underscore the important role of this nuclear receptor in regulating bile acid synthesis and indicate that there are several FXR/SHP-independent pathways that negatively regulate the system (15, 16). The latter pathways may also respond to cholate, since the levels of CYP7A1 are consistently higher in Cyp8b1^–/– mice (five- to sevenfold higher than in the wild-type) than they are in SHP^–/– mice (two- to threefold higher than in the wild-type). In vitro, the differential response to bile acids appears to be largely through FXR, since these ligands were inactive in hepatocytes lacking the receptor (Figure 7, b and d). How cholate affects the FXR-independent pathways that regulate bile acid synthesis in vivo is the subject of current research.

We thank Kevin Anderson, Norma Anderson, Lorena Avila, and Scott Clark for excellent technical assistance; Jay Horton and Young-Ah Moon for help with in vivo sterol synthesis and hepatocyte preparation; Margrit Schwarz for GW4064; Frank Gonzalez for FXR^–/– mice; Alan Hofmann for discussions on the half-life of cholate; and Joe Goldstein and David Mangelsdorf for critical reading of the manuscript. This research was supported by grants from the NIH (HL-20948), the Keck Foundation, the Perot Family Foundation, and the Swedish Medical Research Council.

See the related Commentary beginning on page 1067.

Jia Li-Hawkins and Mats Gåfvels contributed equally to this work.

Erik G. Lund’s present address is: Merck and Co., Rahway, New Jersey, USA.Conflict of interest: No conflict of interest has been declared.Nonstandard abbreviations used: cholesterol 7α-hydroxylase (CYP7A1); farnesoid X receptor (FXR); short heterodimer partner (SHP); CYP7A1 promoter transcription factor (CPF); liver X receptor (LXR); hepatic nuclear factor (HNF); sterol 12α-hydroxylase (CYP8B1); ileal bile acid binding protein (IBABP); sterol 27-hydroxylase (CYP27A1); oxysterol 7α-hydroxylase (CYP7B1); bile salt export protein (BSEP); Na⁺-taurocholate cotransporting polypeptide (Ntcp).

1. Russell, DW, Setchell, KDR. Bile acid biosynthesis. Biochemistry 1992. 31:4737-4749.
   View this article via: PubMed CrossRef Google Scholar
2. Lu, TT, Repa, JJ, Mangelsdorf, DJ. Orphan nuclear receptors as eLiXiRs and FiXeRs of sterol metabolism. J Biol Chem 2001. 276:37735-37738.
   View this article via: PubMed Google Scholar
3. Chawla, A, Repa, JJ, Evans, RM, Mangelsdorf, DJ. Nuclear receptors and lipid physiology: opening the X-files. Science 2001. 294:1866-1870.
   View this article via: PubMed CrossRef Google Scholar
4. Makishima, M, et al. Identification of a nuclear receptor for bile acids. Science 1999. 284:1362-1365.
   View this article via: PubMed CrossRef Google Scholar
5. Parks, DJ, et al. Bile acids: natural ligands for an orphan nuclear receptor. Science 1999. 284:1365-1368.
   View this article via: PubMed CrossRef Google Scholar
6. Wang, H, Chen, J, Hollister, K, Sowers, LC, Forman, BM. Endogenous bile acids are ligands for the nuclear receptor FXR/BAR. Mol Cell 1999. 3:543-553.
   View this article via: PubMed Google Scholar
7. Lu, TT, et al. Molecular basis for feedback regulation of bile acid synthesis by nuclear receptors. Mol Cell 2000. 6:507-515.
   View this article via: PubMed CrossRef Google Scholar
8. Goodwin, B, et al. A regulatory cascade of the nuclear receptors FXR, SHP-1, and LRH-1 represses bile acid synthesis. Mol Cell 2000. 6:517-526.
   View this article via: PubMed CrossRef Google Scholar
9. Nitta, M, Ku, S, Brown, C, Okamoto, AY, Shan, B. CPF: an orphan nuclear receptor that regulates liver-specific expression of the human cholesterol 7α-hydroxylase gene. Proc Natl Acad Sci USA 1999. 96:6660-6665.
   View this article via: PubMed CrossRef Google Scholar
10. Janowski, BA, Willy, PJ, Devi, TR, Falck, JR, Mangelsdorf, DJ. An oxysterol signaling pathway mediated by the nuclear receptor LXRα. Nature 1996. 383:728-731.
    View this article via: PubMed CrossRef Google Scholar
11. Lehmann, JM, et al. Activation of the nuclear receptor LXR by oxysterols defines a new hormone response pathway. J Biol Chem 1997. 272:3137-3140.
    View this article via: PubMed CrossRef Google Scholar
12. Forman, BM, Ruan, B, Chen, J, Schroepfer (Jr), GJ, Evans, RM. The orphan nuclear receptor LXRα is positively and negatively regulated by distinct products of mevalonate metabolism. Proc Natl Acad Sci USA 1997. 94:10588-10593.
    View this article via: PubMed CrossRef Google Scholar
13. Peet, DJ, et al. Cholesterol and bile acid metabolism are impaired in mice lacking the nuclear oxysterol receptor LXRa. Cell 1998. 93:693-704.
    View this article via: PubMed CrossRef Google Scholar
14. Sinal, CJ, et al. Targeted disruption of the nuclear receptor FXR/BAR impairs bile acid and lipid homeostasis. Cell 2000. 102:731-744.
    View this article via: PubMed CrossRef Google Scholar
15. Kerr, TA, et al. Loss of nuclear receptor SHP impairs but does not eliminate negative feedback regulation of bile acid synthesis. Dev Cell 2002. 2:713-720.
    View this article via: PubMed CrossRef Google Scholar
16. Wang, L, et al. Redundant pathways for negative feedback regulation of bile acid production. Dev Cell 2002. 2:721-731.
    View this article via: PubMed CrossRef Google Scholar
17. Schwarz, M, et al. Disruption of cholesterol 7α-hydroxylase gene in mice. II. Bile acid deficiency is overcome by induction of oxysterol 7α-hydroxylase. J Biol Chem 1996. 271:18024-18031.
    View this article via: PubMed CrossRef Google Scholar
18. Hofmann, A.F. 1999. Bile acid science (cholanology) at the dawn of a new millennium: past progress and the challenges for the future. In Bile acids in hepatobiliary disease. T.C. Northfield, H.A. Ahmed, R.P. Jazrawi, and P.L. Zentler-Munro, editors. Kluwer Academic Publishers. New York, New York, USA. 303–332.
19. Schwarz, M, Russell, DW, Dietschy, JM, Turley, SD. Marked reduction in bile acid synthesis in cholesterol 7α-hydroxylase-deficient mice does not lead to diminished tissue cholesterol turnover or to hypercholesterolemia. J Lipid Res 1998. 39:1833-1843.
    View this article via: PubMed Google Scholar
20. Gafvels, M, et al. Structure and chromosomal assignment of the sterol 12α-hydroxylase gene (CYP8B1) in human and mouse: eukaryotic cytochrome P-450 gene devoid of introns. Genomics 1999. 56:184-196.
    View this article via: PubMed CrossRef Google Scholar
21. Mansour, SL, Thomas, KR, Capecchi, MR. Disruption of the proto-oncogene int-2 in mouse embryo-derived stem cells: a general strategy for targeting mutations to non-selectable genes. Nature 1988. 336:348-352.
    View this article via: PubMed CrossRef Google Scholar
22. Sambrook, J., and Russell, D.W. 2000. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press. Plainview, New York, USA. 2400 pp.
23. Rosen, H, et al. Markedly reduced bile acid synthesis but maintained levels of cholesterol and vitamin D metabolites in mice with disrupted sterol 27-hydroxylase gene. J Biol Chem 1998. 273:14805-14812.
    View this article via: PubMed CrossRef Google Scholar
24. Turley, SD, Dietschy, JM. Re-evaluation of the 3α-hydroxysteroid dehydrogenase assay for total bile acids in bile. J Lipid Res 1978. 19:924-928.
    View this article via: PubMed Google Scholar
25. Schwarz, M, Davis, DL, Vick, BR, Russell, DW. Genetic analysis of cholesterol absorption in inbred mice. J Lipid Res 2001. 42:1801-1811.
    View this article via: PubMed Google Scholar
26. Jeske, DJ, Dietschy, JD. Regulation of rates of cholesterol synthesis in vivo in the liver and carcass of the rat measured using [3H]water. J Lipid Res 1980. 21:364-376.
    View this article via: PubMed Google Scholar
27. Li-Hawkins, J, Lund, EG, Turley, SD, Russell, DW. Disruption of the oxysterol 7α-hydroxylase gene in mice. J Biol Chem 2000. 275:16536-16542.
    View this article via: PubMed CrossRef Google Scholar
28. Andersson, U, et al. Thyroid hormone suppresses hepatic sterol 12α-hydroxylase (CYP8B1) activity and messenger RNA in rat liver: failure to define known thyroid response elements in the gene. Biochim Biophys Acta 1999. 1438:167-174.
    View this article via: PubMed Google Scholar
29. Eggertsen, G, et al. Molecular cloning and expression of rabbit sterol 12α-hydroxylase. J Biol Chem 1996. 271:32269-32275.
    View this article via: PubMed CrossRef Google Scholar
30. Schwarz, M, Russell, DW, Dietschy, JM, Turley, SD. Alternate pathways of bile acid synthesis in the cholesterol 7α-hydroxylase knockout mouse are not upregulated by either cholesterol or cholestyramine feeding. J Lipid Res 2001. 42:1594-1603.
    View this article via: PubMed Google Scholar
31. Hayhurst, GP, Lee, Y-H, Lambert, G, Ward, JM, Gonzalez, FJ. Hepatocyte nuclear factor 4α (nuclear receptor 2A1) is essential for maintenance of hepatic gene expression and lipid homeostasis. Mol Cell Biol 2001. 21:1393-1403.
    View this article via: PubMed CrossRef Google Scholar
32. Kliewer, SA, et al. An orphan nuclear receptor pathway activated by pregnanes defines a steroid signaling pathway. Cell 1998. 92:73-82.
    View this article via: PubMed CrossRef Google Scholar
33. Xie, W, et al. Humanized xenobiotic responses in mice expressing nuclear receptor SXR. Nature 2000. 406:435-439.
    View this article via: PubMed CrossRef Google Scholar
34. Honda, A, et al. Side chain hydroxylations in bile acid biosynthesis catalyzed by CYP3A are markedly up-regulated in Cyp27-/- mice but not in cerebrotendinous xanthomatosis. J Biol Chem 2001. 276:34579-34585.
    View this article via: PubMed CrossRef Google Scholar
35. Repa, JJ, et al. Disruption of the sterol 27-hydroxylase gene in mice results in hepatomegaly and hypertriglyceridemia: reversal by cholic acid feeding. J Biol Chem 2000. 275:39685-39692.
    View this article via: PubMed CrossRef Google Scholar
36. Ananthanarayanan, M, Balasubramanian, N, Makishima, M, Mangelsdorf, DJ, Suchy, FJ. Human bile salt export pump promoter is transactivated by the farnesoid X receptor/bile acid receptor. J Biol Chem 2001. 276:28857-28865.
    View this article via: PubMed CrossRef Google Scholar
37. Denson, LA, et al. The orphan nuclear receptor, shp, mediates bile acid-induced inhibition of the rat bile acid transporter, ntcp. Gastroenterology 2001. 121:140-147.
    View this article via: PubMed CrossRef Google Scholar
38. Shih, DQ, et al. Hepatocyte nuclear factor-1α is an essential regulator of bile acid and plasma cholesterol metabolism. Nat Genet 2001. 27:375-382.
    View this article via: PubMed CrossRef Google Scholar
39. Danielsson, H, Eneroth, P, Hellstrom, K, Lindstedt, S, Sjovall, J. On the turnover and excretory products of cholic acid and chenodeoxycholic acid in man. J Biol Chem 1963. 238:2299-2304.
    View this article via: PubMed Google Scholar
40. Hulzebos, CV, et al. Measurement of parameters of cholic acid kinetics in plasma using a microscale stable isotope dilution technique: application to rodents and humans. J Lipid Res 2001. 42:1923-1929.
    View this article via: PubMed Google Scholar
41. Carey, M.C. 1982. The enterohepatic circulation. In The liver: biology and pathobiology. Raven Press. New York, New York, USA. 429–465.
42. Hall, JM, McDonnell, DP, Korach, KS. Allosteric regulation of estrogen receptor structure, function, and coactivator recruitment by different estrogen responsive elements. Mol Endocrinol 2002. 16:469-486.
    View this article via: PubMed CrossRef Google Scholar
43. Forman, BF, et al. Identification of a nuclear receptor that is activated by farnesol metabolites. Cell 1995. 81:687-693.
    View this article via: PubMed CrossRef Google Scholar
44. Bramlett, KS, Yao, S, Burris, TP. Correlation of farnesoid X receptor coactivator requirements and cholesterol 7α-hydroxylase gene repression by bile acids. Mol Genet Metab 2000. 71:609-615.
    View this article via: PubMed CrossRef Google Scholar
45. Hylemon, PB, Harder, J. Biotransformation of monoterpenes, bile acids, and other isoprenoids in anaerobic ecosystems. FEMS Microbiol Rev 1998. 22:475-488.
    View this article via: PubMed CrossRef Google Scholar
46. Wilson, JD, Griffin, JE, Russell, DW. Steroid 5α-reductase 2 deficiency. Endocr Rev 1993. 14:577-593.
    View this article via: PubMed CrossRef Google Scholar
47. Mahendroo, MS, Cala, KC, Hess, DL, Russell, DW. Unexpected virilization in male mice lacking steroid 5α-reductase enzymes. Endocrinology 2001. 142:4652-4662.
    View this article via: PubMed CrossRef Google Scholar
48. Maloney, PR, et al. Identification of a chemical tool for the orphan nuclear receptor FXR. J Med Chem 2000. 43:2971-2974.
    View this article via: PubMed CrossRef Google Scholar