Mechanism of Tissue-specific Farnesoid X Receptor in Suppressing the Expression of Genes in Bile-acid Synthesis in Mice

Abstract

Activation of farnesoid X receptor (Fxr, Nr1h4) is a major mechanism in suppressing bile-acid synthesis by reducing the expression levels of genes encoding key bile-acid synthetic enzymes, CYP7A1/Cyp7a1 and CYP8B1/Cyp8b1. FXR-mediated induction of hepatic small heterodimer partner (SHP/Shp, Nr0b2) and intestinal fibroblast growth factor 15 (Fgf15, FGF19 in human) has been shown to be responsible for this suppression. However, the exact contribution of Shp/Fgf15 to this suppression and the associated cell signaling pathway is unclear. By using novel genetically modified mice, the current study showed that the intestinal Fxr/Fgf15 pathway was critical for suppressing both Cyp7a1 and Cyp8b1 gene expression, but the liver Fxr/Shp pathway was important for suppressing Cyp8b1 gene expression and had a minor role in suppressing Cyp7a1 gene expression. Furthermore, in vivo administration of Fgf15 protein to mice led to a strong activation of ERK, and to a smaller degree, JNK, in the liver. In addition, deficiency of either the ERK or JNK pathway in mouse livers reduced the basal, but not the Fgf15-mediated suppression, of Cyp7a1 and Cyp8b1 gene expression. However, deficiency of both ERK and JNK pathways prevented Fgf15-mediated suppression of Cyp7a1 and Cyp8b1 gene expression.

Conclusion

the current study clearly elucidates the underlying molecular mechanism of hepatic vs. intestinal Fxr in regulating the expression of genes critical for bile-acid synthesis and hydrophobicity in the liver.

Introduction

Bile-acid synthesis is the major mechanism to remove extra cholesterol from the body. Bile acids are required for the absorption of lipids and lipid-soluble vitamins from the intestine. Bile acids activate members of the nuclear receptor superfamily including farnesoid X receptor (FXR/Fxr, encoded by the NR1H4/Nr1h4 gene), pregnane X receptor, vitamin D receptor,^1–5 and a G-protein coupled receptor, TGR5, ⁶ which is a critical mechanism for maintaining endobiotic and xenobiotic homeostasis.

Two enzymatic pathways are responsible for bile-acid synthesis in the liver. The classical pathway generates cholic acid (CA) and the alternative pathway produces chenodeoxycholic acid (CDCA). The cholesterol 7α-hydroxylase, encoded by the cytochrome P450 7A1/7a1 (CYP7A1/Cyp7a1) gene, is the rate-limiting enzyme in the classical pathway. The sterol 12α-hydroxylase, encoded by the CYP8b1/Cyp8b1 gene, mediates the production of CA, and cholesterol will be converted only to CDCA when CYP8b1/Cyp8b1 is deficient. Because CAis less hydrophobic than CDCA, CYP8B1/Cyp8b1 is critical in regulating hydrophobicity of the bile-acid pool by regulating the CA/CDCA ratio.

Bile-acid synthesis is tightly regulated because disruption of bile-acid homeostasis leads to hepatobiliary and intestinal disorders, including cholestasis, gallstone disease, and inflammatory bowel disease, as well as systemic diseases such as atherosclerosis.⁷ Feedback suppression of CYP7A1/Cyp7a1 and CYP8B1/Cyp8b1 gene transcription by nuclear receptors and inflammatory cytokines is the most important mechanism in maintaining bile-acid homeostasis in humans and mice.⁸ Under physiological conditions, activation of FXR is the major mechanism to suppress bile-acid synthesis by directly inducing target genes in both the liver and intestine, including small heterodimer partner (SHP/Shp, encoded by the NR0B2/Nr0b2 gene) and fibroblast growth factor (Fgf) 15 (FGF19 in humans), which in turn inhibits, or activates signaling pathways to inhibit, CYP7A1/Cyp7a1 and CYP8B1/Cyp8b1 gene transcription.^9–13 Specifically, FXR/Fxr induces the expression of SHP/Shp in the liver, which will in turn inhibit liver homologue-1 (LRH-1/Lrh-1)-induced gene transcription of CYP7A1/Cyp7a1 and CYP8b1/Cyp8b1.^{12, 13} By contrast, Fxr induces Fgf15 in the intestine and released Fgf15 activates its receptor, fibroblast growth factor receptor 4 (Fgfr4) in the liver, to activate mitogen-activated protein kinase (MAPK) signaling pathways. It has been shown that activation of extracellular signal-regulated kinase (ERK) in humans and Jun N-terminal kinase (JNK) in rats mediates the suppression of CYP7A1/Cyp7a1 gene expression.^{11, 14} Furthermore, Shp is reported to be required by Fgf15 to suppress Cyp7a1 gene expression.⁹ However, the exact contribution of Shp and Fgf15 to the suppression of bile-acid synthesis following tissue-specific activation of FXR is not known. Furthermore, the exact signaling pathways in the liver with Fgf15 activation are not clear in mice. Elucidating these underlying molecular mechanisms will help to establish the mechanism of regulation of bile-acid synthesis by FXR/Fxr activation, which will aid in determining future strategies in the treatment of cholesterol and bile-acid disorders by tissue-specific activation of FXR to increase efficacy, and meanwhile, reduce toxic effect due to non-specific activation of FXR.

In the current study, by using novel genetically modified mice with various combinations of Fxr, Shp, Fgfr4, early growth response 1 (Egr1) or cJun deletion, and purified Fgf15 protein,¹⁵ the contribution of the intestinal Fxr/Fgf15 and the hepatic Fxr/Shp pathways to the suppression of Cyp7a1 and Cyp8b1 in mice was determined. In addition, whether Shp is required for Fgf15-mediated suppression of Cyp7a1 and Cyp8b1 gene expression was established in mice. Finally, the effect of MAPK pathway activation on basal and Fgf15-mediated suppression of Cyp7a1 and Cyp8b1 gene expression was determined.

Materials and Methods

Reagents

GW4064 was provided by the University of Kansas. Other reagents, unless mentioned, were obtained from Sigma (St. Louis, MO).

Animals

The whole-body Fxr knockout (KO) mice (Fxr WB KO) were reported and are on a pure C57BL/6J genetic background.^{16, 17} The generation of tissue-specific Fxr KO mice on a mixed genetic background has been described previously using the loxP/Cre technology with specific disruption of the Nr1h4 gene in hepatocytes (Fxr Liv KO) or in enterocytes (Fxr Int KO).¹⁸ Specifically, the Fxr Liv KO and Fxr Int KO mice were generated by crossbreeding Fxr floxed/floxed mice with albumin cre (+) or villin cre (+) mice. But these mice were on a mixed genetic background with variable basal expression of bile acid synthetic genes. So in the current study, the congenic Fxr Liv KO and Fxr Int KO mice in the C57BL/6J genetic background were produced. Shp KO mice and hepatocyte-specific Shp transgenic mice (Albumin promoter derived, Shp Tg) have been reported.^{19, 20} Fxr WB KO mice with hepatocyte-specific Shp over-expression (Fxr WB KO/Shp Tg) were generated via crossing Fxr WB KO mice with Shp Tg mice, with all three strains on the pure C57BL/6J genetic background. Fgfr4 KO mice on a mixed C57/129SvJ background were provided by Dr. Curtis Klaassen (University of Kansas Medical Center). Fgfr4/Shp double KO (Fgfr4/Shp DKO) mice were generated by cross breeding Fgfr4 KO and Shp KO mice. Egr1 KO mice on a C57BL/6 genetic background were obtained from Taconic (Hudson, NY). C57BL/6J mice bred in the same animal facility were used as wild-type (WT) controls for KO mice on C57BL/6J background. If KO mice were on a mixed genetic background (Fgfr4 KO and Fgfr4/Shp DKO), littermates were used as controls.

Animal treatment

Mice were bred and maintained in the Laboratory Animal Research facility at the University of Kansas Medical Center in rooms under a 12-hr light/dark cycle. All protocols were approved by the IACUC. All experiments used 10–16 week old male mice and all mice were sacrificed within a 30-min period in the morning. In addition, all treatments were repeated twice.

Activation of Fxr in vivo was achieved by treatment with an Fxr synthetic agonist, GW4064, at 150 mg/kg. GW4064 or vehicle was administered by oral gavage at 6 PM, followed by a second administration at 8 AM next morning. Two hrs later, liver and ileum were harvested. The generation of purified Fgf15 has been reported previously.¹⁵ Fgf15 protein was injected to mice through tail vein at a dosage of 10 μg/kg. Two hrs or indicated time points (for time course study) after the injection, livers were collected.

The total bile-acid pool size was determined by measuring bile acids of the small intestine, gallbladder, liver, and their contents. Ten-sixteen week-old mice were fed a chow diet with 2% cholestyramine for 10 days. Total bile acids were determined using a kit from Bio-quant Kits Inc. (San Diego, CA), and the pool size was expressed as micromoles of bile acids/100 g of body weight.

Lentivirus-mediated shRNA knockdown of cJun in vivo

Lentiviral vectors (6×10⁹ IU/ml) expressing shRNA against cJun (pLKO.1-puro vector, containing the sequence CCGGCAGTAACCCTAAGATCCTAAACTCGAGTTTAGGATCTTAGGGTTACT GTTTTTG) was produced by Capitol Science Inc. Lentiviral particles containing the cJun-shRNA were injected into WT and Egr1 KO mice at 6x10⁹ IU/kg of body weight. Ten days later, WT and Egr1 KO mice were separated into two groups and treated with saline or purified recombinant Fgf15 protein, with liver collected 2 hrs later.

RNA isolation and quantitative PCR (qPCR) analysis

Total RNAisolation and mRNA levels were determined by standard qPCR method. 18S RNA levels were used as normalization control with primers listed in Supplemental Table 1.

Western blotting

Protein from livers was extracted and phosphorylated JNK1/2, ERK1/2, and p38 as well as total cJun were determined by Western Blot using anti-phospho-MAPKs-specific Abs and anti-cJun antibody (Cell Signaling Technology; Beverly, MA).

Statistical analysis

All experimental data are expressed as the mean ± SD. Difference among multiple groups was tested using two-way ANOVA. Difference between two groups was tested by Student’s t-test. A p value of < 0.05 was considered statistically significant.

Results

Fxr in enterocytes suppressed both Cyp7a1 and Cyp8b1 gene expression, but Fxr in hepatocytes mainly suppressed Cyp8b1 gene expression

Aslight increase in bile-acid pool size was observed in Fxr Liv KO and Fxr Int KO mice (Fig 1B). Cholestyramine decreased pool size by 50% in all mice (Fig 1A and 1B). In intestine, Fxr activity was decreased with the decreased pool size (Fig 1C), indicated by decreased Ibabp and Fgf15 levels, while Fxr mRNA levels didn’t change. The hepatic Fxr mRNA levels were increased 2-fold in WT and Fxr Int KO mice by cholestyramine (Fig 1D). Moreover, hepatic Shp mRNA levels were increased about 3-fold by cholestyramine through an Fxr-independent mechanism, as similar increase was observed in Fxr Liv KO mice. Furthermore, decrease of pool size promotes bile-acid synthesis, revealed by 8–10 fold induction of Cyp7a1 in WT mice and FXR Liv KO mice and 3-fold induction of Cyp8b1 in all mice (Fig 1D), indicating increase in hepatic bile acids and/or Shp induction is not responsible for suppressing bile-acid synthesis. In addition, basal mRNA levels of Cyp7a1 increased in Fxr Int KO mice, but not in Fxr Liv mice, which is reciprocal to the reduced Fgf15 levels in the intestine. Cholestyramine didn’t further increase Cyp7a1 mRNA levels in FXR Int KO mice (Fig 1D). In contrast, Cyp8b1 mRNA levels were increased upon cholestyramine in both Fxr Liv KO and Int KO mice (Fig 1D), suggesting that Fxr in enterocytes suppresses both Cyp7a1 and Cyp8b1, but in hepatocytes, Fxr mainly suppresses Cyp8b1. The effect of tissue-specific activation of Fxr on Cyp7a1 and Cyp8b1 mRNA levels was also determined by activating Fxr using GW4064. Cyp7a1 mRNA levels were markedly reduced in WT and Fxr Liv KO mice, but not in Fxr Int KO mice (Supplemental Fig 1).

Figure 1. Effect of 2% cholestyramine feeding for 10 days on bile acid pool size in tissue-specific Fxr KO mice.

(A) Cholestyramine is a bile acid sequestrant that binds bile acids in the intestine to prevent their reabsorption. As a result, this decreases the feed-back suppression of hepatic bile acid synthesis from cholesterol. (B) Effect of cholestyramine feeding on bile acid pool size in Fxr Liv KO and Fxr Int KO mice. (C) Effect of cholestyramine feeding on mRNA levels of Fxr, Fgf15 and Ibabp in the ileum in Fxr Liv KO and Fxr Int KO mice. (D) Effect of cholestyramine feeding on mRNA levels of Fxr, Cyp7a1, Cyp8b1, and Shp in the liver in Fxr Liv KO and Fxr Int KO mice. n=5 male mice/group. An asterisk (*) indicates p < 0.05 between vehicle- and cholestyramine-fed mice within the same genotype.

Fxr-mediated Shp induction contributed to the suppression of Cyp8b1, but not Cyp7a1, gene expression

Induction of Shp and Fgf15 by Fxr activation is known to suppress Cyp7a1 and Cyp8b1 gene expression in the liver. However, it is unclear to what degree Shp and Fgf15 contributes to this suppression. Therefore, this study used Shp KO and Shp Tg mice to determine the contribution of Shp in suppressing Cyp7a1 and Cyp8b1 gene expression. In Shp KO mice, activation of Fxr still markedly suppressed Cyp7a1 to about 10% of the levels in vehicle-treated mice (Fig 2A), though the degree was smaller than that in WT mice (99% suppression).

Figure 2. Effect of Shp deficiency or hepatic overexpression on relative mRNA levels of Cyp7a1, Cyp8b1 and Shp in the liver, as well as on mRNA levels of Fgf15 in the intestine.

(A) The mRNA levels of hepatic Fxr, Cyp7a1, Cyp8b1, Shp, and intestinal Fgf15 in WT, Fxr WB KO and Shp KO mice with or without treatment with GW4064. (B) The mRNA levels of hepatic Fxr, Cyp7a1, Cyp8b1, Shp, and intestinal Fgf15 in WT, Fxr WB KO, Shp Tg and Fxr KO/Shp Tg mice with or without GW4064 treatment. n=5 male mice/group. A pound sign (#) indicates p < 0.05 between genetically modified and WT mice, and an asterisk (*) indicates p < 0.05 between GW4064-treated and vehicle-treated mice within the same genotype.

In Shp Tg mice with Shp over-expressed in hepatocytes, basal Cyp7a1 mRNA levels did not change, nor did the degree of Cyp7a1 suppression after Fxr activation (Fig 2B). Furthermore, in the FXR KO/Shp Tg mice that are deficient in Fxr but over-expressed Shp, the levels of Cyp7a1 mRNA were not affected by Shp over-expression either (Fig 2B). These results indicate that with Fxr activation Shp may play only a minor role in suppressing Cyp7a1 gene expression. Interestingly, in Shp KO mice, the basal Cyp8b1 mRNA levels were 2.5-fold higher than those in WT mice. Fxr activation suppressed Cyp8b1 mRNA levels in Shp KO mice, but the degree of suppression (about 30% decrease) in Shp KO mice was smaller than that in WT mice (50% decrease) (Fig 2A). Furthermore, in Shp Tg mice, Cyp8b1 mRNA levels were only slightly reduced, and were further decreased upon Fxr activation, but the degree of suppression was similar to that in WT mice (Fig 2B).

Exogenous Fgf15 protein administration reduced both Cyp7a1 and Cyp8b1 mRNA levels

Fgf15 has been shown to be important in suppressing Cyp7a1 and Cyp8b1 gene expression. Furthermore, it was reported that the Fgf15-mediated suppression requires Shp.^9–11 Therefore, in the current study, we determined to what degree Fgf15 mediates the suppression of Cyp7a1 and Cyp8b1 gene expression after Fxr activation. In addition, by using Shp KO mice, we tested whether the Fgf15-mediated suppression of Cyp7a1 and Cyp8b1 requires Shp. The effect of purified recombinant Fgf15 protein on Cyp7a1 mRNA levels was determined in WT, Fxr WB KO, Shp KO, Shp Tg and Fxr KO/Shp Tg mice. Compared with vehicle treatment, the Fgf15 protein resulted in strong reduction of Cyp7a1 mRNA levels in all strains (Fig 3), with 1%, 10%, 10%, 5%, and 10% of Cyp7a1 mRNA left respectively, indicating that Fgf15 is down-stream of Fxr and Shp. Fgf15 protein also suppressed Cyp8b1 gene expression in Fxr WB KO, Shp KO, Fxr WB KO/Shp Tg mice, in addition, Fgf15 tended to suppress Cyp8b1 gene expression in WT mice, but not in Shp Tg mice (Fig 3). Interestingly, Shp mRNA levels were also reduced markedly by Fgf15 treatment, indicating that Fgf15 regulates Shp expression in the liver.

Figure 3. Effects of exogenous Fgf15 protein on relative mRNA levels of Cyp7a1, Cyp8b1 and Shp in the liver.

Relative mRNA levels of Cyp7a1, Cyp8b1 and Shp in livers of WT, Fxr WB KO, Shp KO, Shp Tg and Fxr KO/Shp Tg mice at 2 hrs after tail-vein injection of purified recombinant Fgf15 protein (10 μg/kg). n=5 male mice/group. An asterisk (*) indicates p < 0.05 between Fgf15-treated and saline-treated group.

Next, by using Fgfr4 KO mice, we determined to what degree induction of Fgf15 by Fxr activation contributes to the suppression of Cyp7a1 and Cyp8b1 gene expression. Fgfr4 is a trans-membrane protein highly expressed in the liver and is the major receptor for Fgf15/FGF19, while β-Klotho acts as co-receptor to form the Fgfr4-β-Klotho complex and response to Fgf15/FGF19. The Fgf15 KO mice on the C57BL/6 background are embryonic lethal, so we used Fgfr4 KO mice as surrogates of Fgf15 KO mice in the current study. Consistent with previous findings, Fgfr4 KO mice expressed higher basal Cyp7a1 mRNAlevels (Fig 4A). Treatment with GW4064 in Fgfr4 KO mice only moderately suppressed 40% Cyp7a1 gene expression, compared with the 95% suppression in WT mice (Fig 4A). Interestingly, Fgfr4 deficiency led to decreased basal Shp gene expression in the liver (Fig 4B), and decreased GW4064-induced Shp expression as well, indicating that Shp expression may be regulated by MAPK signaling pathways. Though bile acid synthesis and Cyp7a1 gene expression was increased in β-Klotho knockout mice ²¹, the current study did not show changes the hepatocyte β-Klotho mRNAlevels after treatment with GW4064 (Fig 4B), indicating that β-Klotho may not be regulated by Fxr.

Figure 4. Effects of double deficiency of Fgfr4 and Shp on Fxr activation-mediated suppression of Cyp7a1 and Cyp8b1 gene expression.

The relative mRNA levels of hepatic Cyp7a1, Cyp8b1, Fgfr4, β-Klotho and Shp as well as intestinal Fgf15 in WT, Fgfr4 KO, Shp KO and Fgfr4/Shp DKO mice after activation of FXR by the GW4064 treatment. n=5 male mice/group.A pound sign (#) indicates p < 0.05 between genetically modified and WT mice, and an asterisk (*) indicates p < 0.05 between GW4064-treated and vehicle-treated mice within the same genotype.

Induction of Fgf15 and Shp may be the only two mechanisms responsible for Fxr-mediated suppression of the Cyp7a1 and Cyp8b1 gene expression

Besides Fgf15 and Shp, additional factors may be involved in suppressing Cyp7a1 and Cyp8b1 gene expression after Fxr activation. To test this possibility, we generated mice deficient in both Fgfr4 and Shp (Fgfr4/Shp DKO mice). A marked increase about 4-fold in Cyp7a1 but not Cyp8b1 mRNA levels was observed in these DKO mice (Fig 4). Surprisingly, activation of Fxr in these mice did not reduce the mRNA levels of Cyp7a1 or Cyp8b1 (Fig 4), indicating that Fgf15 and Shp may be the only two factors involved in mediating suppression of Cyp7a1 and Cyp8b1gene expression following Fxr activation.

Time course of Cyp7a1 and Cyp8b1 suppression as well as ERK and JNK activation by exogenous Fgf15 treatment

In vitro, activation of either JNK1/2 or ERK 1/2 following Fgfr4 activation has been shown to suppress Cyp7a1/CYP7A1 gene expression.^{10, 11} To clarify which of these two pathways is activated in vivo after Fgfr4 activation in mice, we determined Cyp7a1 and Cyp8b1 mRNA levels at 30 min, 1, 2, 3, 4, 6, and 8 hrs after exogenous Fgf15 protein treatment. As mentioned earlier, the expression of the Cyp7a1 gene is subject to circadian regulation, After Fgf15 injection, Cyp7a1 mRNA levels started to rise during the experimental duration even with vehicle treatment. With Fgf15 administration, the mRNA levels of Cyp7a1 decreased at 1 hr, reached their lowest point at 2 hrs, stayed low for 3 and 4 hrs, and returned to normal at 6 and 8 hrs (Fig 5A). The Cyp8b1 mRNA levels also showed circadian change with the degree much smaller than those of Cyp7a1. With the Fgf15 treatment, the Cyp8b1 mRNA levels started to decrease at 2 hrs, and remained reduced during the entire time course examined (Fig 5A).

Figure 5. Time course of the effects of recombinant Fgf15 protein administration on Cyp7a1 and Cyp8b1 mRNA levels and on activation of MAPK in the livers of WT mice.

(A) Relative mRNA levels of Cyp7a1 and Cyp8b1 in livers of mice at 30 min, 1, 2, 3, 4, 6, and 8 hrs after tail-vein injection of purified recombinant Fgf15 protein (10μg/kg). Please note the expression of Cyp7a1 is strongly, and the expression of Cyp8b1 is weakly, subject to circadian rhythm regulation. n=5 male mice/group. An asterisk (*) indicates p < 0.05 between Fgf15-treated and control vehicle-treated group. (B) From the same time-course liver samples at 30 min, 1, 2, and 3 hrs, hepatic protein levels of phosphorylated JNK1/2 (p-JNK1/2) and total JNK1/2 (T-JNK1/2), phosphorylated ERK1/2 (p-ERK1/2) and total ERK1/2 (T-ERK1/2), and phosphorylated p38 (p-p38) and total p38 (T-p38) were determined by Western blot analysis.

Once the time course of suppression of Cyp7a1 and Cyp8b1 gene expression by exogenous Fgf15 protein was established, the protein levels of total and phosphorylated (active form) mitogen-activated protein kinase (MAPK) family, including JNK, ERK and p38, at 30 min, 1 , 2, and 3 hrs after Fgf15 injection, were determined. The protein levels of total JNK1/2 and ERK1/2, as well as total and phosphorylated p38, remained unchanged during the time points measured. However, Fgf15 treatment strongly increased the phosphorylated ERK1/2 at both 30 min and 1 hr, and slightly increased the phosphorylated JNK1/2 at 1 hr (Fig 5B).

Both JNK and ERK supported the basal, but suppressed the Fgf15-mediated, expression of Cyp7a1 and Cyp8b1 gene

The downstream target of JNK is cJun that is a component of activating protein 1 (AP1), the downstream target of ERK is Egr1, and both AP1 and Egr1 are transcription factors. As shown earlier, following Fgf15 treatment, ERK, and to a much smaller degree, JNK, were activated in WT mice; therefore, the degree to which JNK and ERK activation contributed to the suppression of Cyp7a1 and Cyp8b1 gene expression was determined in mice with cJun knock-down or Egr1 deletion. The results showed that 2 hrs after Fgf15 treatment, the mRNA levels of cJun increased in WT mice, but not in Egr1 KO mice, indicating that Egr1 mediates the induction of cJun after MAPK activation (Fig 6B). Knock-down of cJun by shRNA markedly reduced cJun mRNA and protein levels in both WT and Egr1 KO mice (Fig 6B and 6D). Surprisingly, Cyp7a1 mRNA levels were about 5-fold reduced with cJun knock-down and Fgf15 treatment led to only a small, additional suppression in the cJun knock-down mice, which was not statistically significant (Fig 6A). Similarly, the expression of Cyp7a1 in Egr1 KO mice was about 10-fold reduced, and treatment with Fgf15 further reduced Cyp7a1 expression without statistical significance. Furthermore, cJun knock-down in Egr1 KO mice led to a similarly lower basal mRNA levels of Cyp7a1, but treatment with the Fgf15 protein in these mice did not further reduce Cyp7a1 mRNA levels (Fig 6A). The effects of cJun and Egr1 deficiency on Fgf15-mediated suppression of Cyp8b1 gene expression were similar to those of the Cyp7a1 gene, except for an even smaller degree of suppression after Fgf15 treatment (Fig 6A).

Figure 6. The effects of cJun knock-down and Egr1 knockout in vivo on basal, as well as Fgf15-mediated suppression of, mRNA levels of Cyp7a1 and Cyp8b1 in mouse livers.

(A) Relative mRNA levels of Cyp7a1, Cy8b1; (B) Relative mRNA levels of cJun and (C) Relative mRNA levels of Egr1, Shp in mouse livers with in vivo knock-down of cJun or knockout of Egr1. n=5 male mice/group. A pound sign (#) indicates P < 0.05 between knock-down/out mice and WT mice receiving the same treatment and an asterisk (*) means P < 0.05 between Fgf15-treated and vehicle-treated group. (B) From the same mouse livers, Western blot analysis of hepatic protein levels of activated JNK1/2 (p-JNK1/2) and total JNK1/2 (T-JNK1/2), activated ERK1/2 (p-ERK1/2) and total ERK1/2 (T-ERK1/2), and cJun.

Because deficiency of cJun or Egr1 not only led to reduced suppression of Cyp7a1 and Cyp8b1 gene expression after Fgf15 treatment, but also resulted in marked reduction of basal Cyp7a1 and Cyp8b1 gene expression, it is possible that JNK and ERK compensate each other’s function. To examine this possibility, we tested protein levels of cJun, total and activated JNK and ERK in cJun and Egr1 deficient mice. With cJun knock-down in WT mice,there was a trend towards increase in ERK phosphorylation (Fig 6D). Likewise, Egr1 deficiency led to marked increase in cJun, and total and activated ERK protein levels, as well as a slight increase in total JNK protein levels (Fig 6D). When cJun was further knocked down in Egr1 KO mice, the only protein that was changed was activated ERK, which was decreased (Fig 6D).

Discussion

This study presents tissue-specific roles for Fxr, Shp, and Fgf15 in suppressing Cyp7a1 and Cyp8b1 gene expression in mice (Fig 7). Intestinal Fxr activation predominately suppresses Cyp7a1 gene expression via the induction of Fgf15. Hepatic Fxr activation, via the induction of Shp, is less important in suppressing Cyp7a1 expression. In contrast, both intestinal Fxr/Fgf15 and hepatic Fxr/Shp pathways are important in suppressing Cyp8b1 gene expression. In addition, activation of both JNK and ERK is associated with the suppression of Cyp7a1 and Cyp8b1 gene expression. Finally, cJun and Egr1, the downstream targets of JNK and ERK, respectively, compensate each other in suppressing Cyp7a1 and Cyp8b1 gene expression, as well as in maintaining basal expression of Cyp7a1 and Cyp8b1 gene.

Figure 7. Model of negative feedback regulation of hepatic bile-acid synthesis via nuclear receptor Fxr in a tissue-specific manner.

Bile acids in hepatocytes activate Fxr to induce Shp expression, which inhibit Lrh-1 and/or Hnf4α induced gene transcription of Cyp7a1 and Cyp8b1, to mainly result in decreasing the expression of the Cyp8b1 gene, and to a minor extent, the Cyp7a1 gene. Bile acids in enterocytes activate Fxr to induce Fgf15 expression. Fgf15 is released into the circulation and is transported to the liver followed by binding to its cognate tyrosine kinase receptor, Fgfr4, and Fgfr4 co-receptor, β-Klotho on hepatocyte cell membrane. Activation of Fgfr4 activates mainly the ERK1/2 and to a less extent, the JNK1/2 to down-regulate both Cyp7a1 and Cyp8b1 gene expression. Furthermore, Shp appears to be regulated by Fgf15/Fgfr4.

It is commonly considered that increased bile acids in the liver initiate the suppression. Recently, this concept has been challenged by studies showing that an increase in intestinal bile acids is critical in suppressing bile-acid synthesis in the liver.^22–24 In addition, activation of bile acid-sensing nuclear receptor, FXR/Fxr, is the most important mechanism in suppressing Cyp7a1 and Cyp8b1.¹⁷ Furthermore, at least in mice, intestinal Fxr is important in suppressing Cyp7a1 gene expression and both hepatic and intestinal Fxr activation is important in suppressing Cyp8b1 gene expression.¹⁸ It is apparent that activation of Fxr in the intestine is predominant in regulating the amount of bile acids synthesized under physiological conditions, but activation of Fxr in both the liver and intestine is critical in regulating bile-acid hydrophobicity. This concept is further supported by a recent study showing that activation of FXR in intestine protects cholestasis by increasing Fgf15 to suppress Cyp7a1 and Cyp8b1 expression.²⁵

Furthermore, a significant contribution of this study to understanding bile acid synthesis is that Fgf15, but not Shp, predominantly suppresses Cyp7a1 gene expression. However, both Fgf15 and Shp are important for suppressing Cyp8b1 gene expression. Fgf15/FGF19 is one of the most strongly induced Fxr/FXR target genes and its role in suppressing bile-acid synthesis is novel.^{9, 11, 18, 26, 27} The current study clearly demonstrates that Fgf15 is largely responsible for suppressing Cyp7a1 gene expression in mice following Fxr activation in the intestine. Shp has been shown to inhibit Cyp7a1/CYP7A1 and Cyp8b1/CYP8B1 gene expression in rats and primary human hepatocytes by inhibiting LRH-1-mediated transcriptional activation of Cyp7a1/CYP7A1 and Cyp8b1/ CYP8B1 genes. ^{12, 13} In addition, Shp has been shown to be required for Fgf15-mediated suppression of Cyp7a1 gene expression in mice.⁹ In the current study, Shp seems to contribute a minor role (~15%) to the suppression of Cyp7a1 gene expression, but Shp plays an equally important role as Fgf15 in suppressing Cyp8b1 gene expression. In support of this conclusion, studies have shown that hepatic Lrh-1 gene deletion and Shp reduction did not affect Cyp7a1 gene expression, but instead, markedly reduced Cyp8b1 gene expression.^{28, 29} Furthermore, Shp has been shown to be required to suppress Cyp7a1 gene expression by the Fxr-Fgf15/Fgfr4 pathway.⁹ However, in the current study, exogenous Fgf15 protein treatment strongly inhibited Cyp7a1 gene expression in Shp KO mice as well, indicating that Fgf15 suppresses Cyp7a1 gene expression independent of Shp. This conclusion is also supported by a study that showed in primary human hepatocytes knock-down of SHP did not affect FGF19-mediated suppression of CYP7A1 gene expression.¹¹

MAPK activation is associated with the suppression of the Cyp7a1/CYP7A1 gene expression by Fgf15/FGF19. p38 activation indirectly increases Cyp7a1 expression via increasing HNF4α-mediated induciton of Cyp7a1 gene,³⁰ but our results showed that Fgf15 did not activate p38. It was also reported that over-expression of FGFR4 in mice reduces Cyp7a1 expression, which is associated with activation of JNK in the liver,¹⁴ and in primary human hepatocytes, treatment with FGF19 selectively activates ERK1/2 to suppress CYP7A1 expression.¹¹ The downstream target of JNK and ERK is cJun and Egr1, respectively. This study showed that in mice, treatment with Fgf15 rapidly mainly increased ERK activation, and to a smaller degree, JNK activation. However, knock-down/knockout of cJun or Egr1 markedly reduced basal expression of Cyp7a1 and Cyp8b1, but did not prevent suppression of Cyp7a1 and Cyp8b1 by Fgf15 protein. These data also indicate that the ERK and JNK pathways tend to compensate each other in suppressing Cyp7a1 and Cyp8b1 gene expression. Egr1 deficiency led to a strong induction of cJun, and it was reported that ERK activation leads to induction of c-fos, a partner of cJun to form AP1.³¹ Therefore, increased ERK activation with Egr1 deletion may increase AP1, which may result in suppression of Cyp7a1 expression. These data may suggest species difference in MAPK pathway-mediated suppression of Cyp7a1/CYP7A1 and Cyp8b1/CYP8B1 gene expression between mice and humans. In addition, despite cJun and Egr1 support basal expression of Cyp7a1 and Cyp8b1, Fgf15-activated JNK and ERK activation results in suppression of Cyp7a1 and Cyp8b1 expression, indicating that Fgf15 may switch the effects of MAPK on regulating Cyp7a1 and Cyp8b1 expression.

Another novel finding from this study is that in addition to regulating Cyp7a1 expression, the Fgf15/Fgfr4 pathway affects liver Shp expression. Basal and Fxr-induced Shp mRNA levels were decreased in the Fgfr4 KO mice. It will be interesting to determine the mechanisms of Fgf15/Fgfr4 regulation of Shp expression. This mechanism may involve post-translational modification of Fxr after MAPK activation. In fact, post-translational modification of Fxr by acetylation or phosphorylation has been reported to affect Fxr function. ^{32, 33}

In conclusion, this study provides several significant findings: (1) at least in mice, intestinal, but not hepatic, Fxr is important in suppressing Cyp7a1 gene expression. (2) Fgf15 is the major and Shp is a minor mediator in suppressing Cyp7a1, but both are equally important in suppressing Cyp8b1, gene expression in mice. (3) Likely, Fgf15 and Shp are the only two factors in mice to suppress Cyp7a1 and Cyp8b1 gene expression after Fxr activation. (4) ERK, and to a small degree, JNK, are involved in Fgf15-mediated suppression of bile-acid synthesis. However, deficiency of both ERK and JNK markedly reduced the basal expression of the Cyp7a1 and Cyp8b1 genes, adding another layer of complexity in regulating bile-acid synthesis following MAPK activation. Our study suggests that activating Fxr in the intestine may result in stronger suppression of bile-acid synthesis, which may be used as a strategy to inhibit bile-acid synthesis to treat diseases with overt bile-acid production. In contrast, inhibiting Fxr in the intestine may lead to enhanced cholesterol conversion to bile acids, which may be used as a useful strategy to reduce cholesterol levels.

Supplementary Material

Supp FIg S1. Supplemental Fig 1.

Effect of tissue-specific Fxr deficiency on relative mRNA levels of Fxr, Cyp7a1, Cyp8b1 and Shp in liver, as well as on mRNA levels of Fxr and Fgf15 in intestine. All four strains of mice (WT, Fxr WB KO, Fxr Liv KO, and Fxr Int KO) were bred to the C57BL/6J genetic background. The livers and ileums of male mice were collected within a 2-hr period in the morning. The mRNA levels of Fxr, Cyp7a1, Cyp8b1, and Shp in the liver, as well as mRNA levels of Fxr and Fgf15 in the ileum, were determined by q-PCR. Furthermore, the effects of tissue-specific activation of Fxr on gene expression were determined in each strain of mice treated twice (6 PM and next morning 8 AM) with GW4064 or vehicle, with livers and ileums collected 2 hrs after the second treatment. n=5 male mice/group. A pound sign (#) indicates P < 0.05 between genetically modified and WT mice, and an asterisk (*) indicates p < 0.05 between GW4064-treated and vehicle-treated mice within the same genotype

Abbreviations

AP1

activating protein 1

Cyp7a1

cholesterol 7α-hydroxylase

Cyp8b1

sterol 12α-hydroxylase

Egr1

early growth response 1

ERK

extracellular signal-regulated kinase

FXR

farnesoid X receptor

FGF15/19

fibroblast growth factor 15/19

FGFR4

Fibroblast growth factor receptor 4

JNK

Jun N-terminal kinase

LRH-1

liver homologue-1

MAPK

mitogen-activated protein kinase

pERK

phosphorylated ERK

pJNK

phosphorylated JNK

SHP

small heterodimer partner

shRNA

short hairpin RNA