Identification of bile acid precursors as endogenous ligands for the nuclear xenobiotic pregnane X receptor

Abstract

Sterol 27-hydroxylase (CYP27A1) is required for bile acid synthesis by both the classical and alternate pathways. Cyp27a1(-/-) mice exhibit a dramatic increase in the activity of cytochrome P450 3A (CYP3A), which catalyzes side-chain hydroxylations of bile acid intermediates, thereby facilitating their excretion in the bile and urine. We examine the role of the nuclear xenobiotic receptor PXR (pregnane X receptor) in this process. We demonstrate that expression of Cyp3a11 and other established PXR target genes is increased in the Cyp27a1(-/-) mice. WhenCyp27a1(-/-) mice are fed a diet containing either cholic acid or chenodeoxycholic acid, expression of CYP7A1, which catalyzes the rate-limiting step in bile acid biosynthesis, is strongly suppressed. In parallel, the induction of Cyp3a11 observed in these mice is reversed, suggesting that bile acid intermediates serve as PXR activators. In support of this hypothesis, three potentially toxic sterols (7alpha-hydroxy-4-cholesten-3-one, 5beta-cholestan-3alpha,7alpha,12alpha-triol, and 4-cholesten-3-one), including two that are known to accumulate in Cyp27a1(-/-) mice, are efficacious activators of mouse PXR. All three compounds are more potent activators of mouse PXR than of human PXR, which may explain in part why humans who lack functional CYP27A1 do not display a corresponding increase in CYP3A activity and are stricken with the disease cerebrotendinous xanthomatosis. Taken together, these results reveal the existence of a feedforward regulatory loop by which potentially toxic bile acid intermediates activate PXR and induce their own metabolism. In addition, this study demonstrates that animal models with alterations in gene expression can be used to identify endogenous ligands for orphan nuclear receptors.

Figure 1

Hepatic mRNA expression levels of several enzymes involved in bile acid biosynthesis are altered in the Cyp27a1^−/− mouse. Relative hepatic RNA levels were determined by Northern analysis from male, mixed-strain, wild-type and Cyp27a1^−/− mice fed a standard rodent diet. The values shown in boxes represent the mean ± SEM of the fold change in mRNA levels (n = 6 mice per genotype). Bile acid intermediates that would be predicted to accumulate upstream of CYP3A11 because of these changes in enzyme levels include the following: 1, 7α-hydroxycholesterol (5-cholesten-3β,7α-diol); 2, 7α-hydroxy-4-cholesten-3-one (4-cholesten-7α-ol-3-one); 3, 5β-cholestan-3α,7α,12α-triol; 4, 4-cholesten-3-one; and 5, cholestanol (5α-cholestan-3β-ol).

Figure 2

PXR target genes show enhanced hepatic expression in the Cyp27a1^−/− mouse. mRNA was isolated from the livers of chow-fed, male, wild-type and Cyp27a1^−/− mice of a mixed-strain background (C57BL/6:129Sv). Northern analysis was performed, quantified by PhosphorImager, standardized against β-actin, and mathematically adjusted to yield a value of 1 for the wild-type group (white bars). The relative expression values (black bars) are depicted as the mean and SEM for each genotype (n = 6), and asterisks denote statistically significant results as determined by Student's t test (*, P < 0.05; ***, P < 0.0001).

Figure 3

Bile acid feeding relieves the elevated hepatic CYP3A11 mRNA levels seen in the Cyp27a1^−/− mouse. Male, mixed-strain wild-type and Cyp27a1^−/− mice were fed normal chow (white bars) or chow diets containing CA (black bars) or CDCA (hatched bars) at 0.1% or 0.2% (wt/wt) for 16 days. Total RNA was isolated from individual livers, and equivalent amounts from each animal (n = 5–7) were pooled in groups for mRNA purification. Northern analyses were performed on the pooled mRNA (shown as blots in the upper part of the panel for each mRNA), quantified by PhosphorImager, and standardized against β-actin (bottom blots), and the results are shown as histograms.

Figure 4

Identification of bile acid intermediates that activate PXR. (A) CV-1 cells were cotransfected with mouse or human PXR and the XREM-luciferase reporter (17), mouse FXR and the FXRE_IBABPx3-tk-luciferase reporter (34), mouse CAR and DR3_{rat CYP3A1}-tk-luciferase, or mouse VDR and mSPPx3-tk-luc (19). Cells were exposed to various ligands, all provided at 10 μM, except 4-cholesten-3-one (lower bar of the pair; 33 μM), CDCA (100 μM), TCPOBOP (0.5 μM), and 1,25-dihydroxyvitamin D₃ (100 nM) for ≈40 h before assaying for luciferase activity. Transfection efficiency was corrected by analysis of β-galactosidase activity because of the cotransfection of a constitutive lacZ expression reporter. See Fig. 1 for the structures of bile acid intermediates. (B) Dose–response analysis of 5β-cholestan-3α,7α,12α-triol and 7α-OH-4-cholesten-3-one activation of mouse PXR and human PXR. Transfection experiments were performed as in A, except secreted placental alkaline phosphatase was used to correct for transfection activity and results are expressed as relative luciferase units (RLU).

Figure 5

5β-Cholestan-3α,7α,12α-triol and 7α-hydroxy-4-cholesten-3-one directly bind mouse PXR as determined by fluorescence polarization assay. A fluorescein-tagged SRC1 peptide (ILRKLLQE) was incubated with bacterially expressed, purified GST-mPXR (black bars) or GST (white bars) in the presence of vehicle (DMSO), PCN (10 μM), or bile acid intermediates at 10 μM. Ligand-induced recruitment of the fluorescein-tagged SRC1 peptide was monitored by an increase in millipolarization fluorescence units (mP). Data shown represent the mean ± SEM of six samples.