Promotion of liver regeneration/repair by farnesoid X receptor in both liver and intestine in mice

Authors

  • Lisheng Zhang,

    1. Division of Gene Regulation and Drug Discovery, Department of Diabetes and Metabolic Diseases, Beckman Research Institute, City of Hope National Medical Center, Duarte, CA
    Search for more papers by this author
  • Yan-Dong Wang,

    1. Division of Gene Regulation and Drug Discovery, Department of Diabetes and Metabolic Diseases, Beckman Research Institute, City of Hope National Medical Center, Duarte, CA
    Search for more papers by this author
  • Wei-Dong Chen,

    1. Division of Gene Regulation and Drug Discovery, Department of Diabetes and Metabolic Diseases, Beckman Research Institute, City of Hope National Medical Center, Duarte, CA
    Search for more papers by this author
  • Xichun Wang,

    1. Division of Gene Regulation and Drug Discovery, Department of Diabetes and Metabolic Diseases, Beckman Research Institute, City of Hope National Medical Center, Duarte, CA
    Search for more papers by this author
  • Guiyu Lou,

    1. Division of Gene Regulation and Drug Discovery, Department of Diabetes and Metabolic Diseases, Beckman Research Institute, City of Hope National Medical Center, Duarte, CA
    Search for more papers by this author
  • Nian Liu,

    1. Division of Gene Regulation and Drug Discovery, Department of Diabetes and Metabolic Diseases, Beckman Research Institute, City of Hope National Medical Center, Duarte, CA
    Current affiliation:
    1. Department of Hematology and Oncology, First Affiliated Hospital of Jilin University, 71 Xinmin Street, Changchun City, Jilin 130041, China
    Search for more papers by this author
  • Min Lin,

    1. Division of Gene Regulation and Drug Discovery, Department of Diabetes and Metabolic Diseases, Beckman Research Institute, City of Hope National Medical Center, Duarte, CA
    Search for more papers by this author
  • Barry M. Forman,

    1. Division of Gene Regulation and Drug Discovery, Department of Diabetes and Metabolic Diseases, Beckman Research Institute, City of Hope National Medical Center, Duarte, CA
    Search for more papers by this author
  • Wendong Huang

    Corresponding author
    1. Division of Gene Regulation and Drug Discovery, Department of Diabetes and Metabolic Diseases, Beckman Research Institute, City of Hope National Medical Center, Duarte, CA
    • Division of Gene Regulation and Drug Discovery, Department of Diabetes and Metabolic Diseases, Beckman Research Institute, City of Hope National Medical Center, 1500 E. Duarte Road, Duarte, CA 91010. Fax: 626–256–8704;

    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

  • L.Z. and W.H. supported by Ibrahim training grant and NCI R01-139158.

Abstract

Farnesoid X receptor (FXR) is a member of the nuclear receptor superfamily and is the primary bile acid receptor. We previously showed that FXR was required for the promotion of liver regeneration/repair after physical resection or liver injury. However, the mechanism by which FXR promotes liver regeneration/repair is still unclear. Here we show that both hepatic-FXR and intestine-FXR contributed to promote liver regeneration/repair after either 70% partial hepatectomy or carbon tetrachloride-induced liver injury. Hepatic FXR, but not intestine FXR, is required for the induction of Foxm1b gene expression in liver during liver regeneration/repair. In contrast, intestine FXR is activated to induce FGF15 expression in intestine after liver damage. Ectopic expression of FGF15 was able to rescue the defective liver regeneration/repair in intestine-specific FXR null mice. Conclusion: These results demonstrate that, in addition to the cell-autonomous effect of hepatic FXR, the endocrine FGF15 pathway activated by FXR in intestine also participates in the promotion of liver regeneration/repair. (HEPATOLOGY 2012;56:2336–2343)

Farnesoid X receptor (FXR) is a member of the nuclear receptor superfamily. It is now well known that FXR plays a critical role to protect cells against bile acid-induced toxicity. FXR is not only a master regulator of bile acid homeostasis and detoxification,1–4 but also mediates a novel role of bile acid signaling in promoting liver regeneration/repair.5, 6

Liver regeneration or repair is a compensatory regrowth of liver after liver damage, including physical resection or toxic injury. Many genes and signaling pathways, such as cytokines and growth factors, have been identified to initiate or promote this process of liver regrowth. We recently showed that bile acid signaling was activated after 70% partial hepatectomy (PH) and FXR was required to promote liver regeneration.5–7 FXR has at least two roles during liver regeneration. One role is to suppress cholesterol 7α-hydroxylase (CYP7a1) expression and reduce bile acid stress. The other role is to promote hepatocyte proliferation by directly activating Foxm1b, which is a key regulator of hepatic cell cycle progression.7 Moreover, FXR is critical to promote liver repair after injury induced by a liver toxin, carbon tetrachloride (CCl4).8, 9 More interestingly, FXR null mice spontaneously develop liver tumors when they age.10, 11 Because bile acids are known to cause DNA damage and induce cell transformation if their levels are not controlled, FXR's roles in suppressing bile acid synthesis as well as promoting liver repair could be an intrinsic mechanism to protect liver from tumorigenesis.12, 13

Although bile acids are synthesized in the liver, FXR in both liver and intestine are required to control levels of bile acids. FXR represses CYP7a1 gene expression through the coordinated induction of fibroblast growth factor 15 (FGF15) in intestine and short heterodimer partner (SHP) in liver. FGF15 and SHP then act cooperatively to repress CYP7a1 transcription through a mechanism that is not yet understood.14 Mice with deletion of either FGF15 or SHP have markedly elevated basal CYP7a1 expression. Mice with intestine-specific deletion of FXR lost the suppression of CYP7a1 expression after treatment with an FXR ligand, GW4064, suggesting that FXR in the gut is key to regulate bile acid synthesis in the liver.15 Moreover, FGF15 has been shown to promote hepatocyte proliferation through its receptor (FGFR4) in liver.16 FGFR4-deficient mice exhibited increased liver injury and delayed liver repair after injury.17 All these results highlight an endocrine role of FGF15 from intestine to the liver. However, whether FGF15 has a role in liver regeneration/repair is unclear.

In this study, we took advantage of liver- and intestine-specific FXR null mice and showed that both hepatic FXR and intestinal FXR contributed to promoting liver regeneration/repair. We further demonstrated that FGF15 induced by intestine FXR was an endocrine pathway to promote liver regrowth.

Abbreviations

BrdU, 2-bromodeoxy-uridine; CCl4, carbon tetrachloride; CYP7a1, cholesterol 7α-hydroxylase; FGF15, fibroblast growth factor 15; FXR, farnesoid X receptor; PH, partial hepatectomy; SHP, short heterodimer partner.

Materials and Methods

Animal Maintenance and Treatments.

FXR whole-body knockout mice (KO) were described.5 Liver-specific FXR null mice (ΔL-FXR) and intestine-specific FXR null mice (ΔIN-FXR) were generated at the University of Southern California. All procedures followed National Institutes of Health (NIH) guidelines for the care and use of laboratory animals. Mice were housed in a pathogen-free animal facility under a standard 12-hour light/dark cycle and fed standard rodent chow and water ad libitum. Male mice between 8 and 10 weeks old were used in each group of experiments; 3-7 mice were used in each group.

Characterization and Genotyping Protocol.

Total proteins from livers or ileal mucosa of ΔL-FXR and ΔIN-FXR FXR-null mice and FXR flox/flox (FXR Fl/Fl) controls were extracted and subjected to western blot analysis. Tail biopsies from animals were analyzed by polymerase chain reaction (PCR). The presence of the cre allele was detected by primers ML136 and ML137, resulting in a 500-bp PCR product. Primer sequence information is provided in Supporting Table 1.

Hepatectomy and Liver Regeneration.

Partial hepatectomy (PH) was done according to the method of Higgins and Anderson.18 Left lateral, caudate, and median lobes were completely excised and the gallbladder was left intact, as described.5

CCl4 and Liver Regeneration.

For acute CCl4-induced liver damage and liver regeneration study, a single dose of 1.5 mL/kg of body weight was administered by intraperitoneal injection as described.13

Liver Histology.

As described by Huang et al.5 and Zhang et al.,7 briefly, after mice were euthanized their livers were removed and small pieces from different lobes of the livers were fixed in 4% formaldehyde-phosphate-buffered saline (PBS) solution, embedded in paraffin, sectioned at 5 μm, and stained with hematoxylin and eosin (H&E). For 2-bromodeoxy-uridine (BrdU) staining, mice were injected intraperitoneally with BrdU solution (10 mg/kg body weight) 2 hours before euthanasia. Liver sections were prepared and stained using a BrdU staining kit (Roche, Indianapolis, IN). The number of positively stained cells was counted in at least three randomly selected fields for each tissue section. The percentage of liver necrosis areas was assigned a score on a semiquantitative scale where 0 is defined as no necrosis area at 0 hours after CCl4 treatment: 1 is mild (30%-40%), 2 is moderate (40%-50%), 3 is severe (50%-60%), and 4 is the most severe (60%-80%).

Adenovirus FGF15 Purification and Mice Treatment.

Viruses were propagated in 911 cells as reported14 and purified by using the adenovirus purification kit (ClonTech). Mice were infected with adenovirus by injection into the tail vein as described.14 Each mouse received 1.0 × 109 particles/10g body weight in 0.1 mL of saline. Three days later, mice were either euthanized as control group (0 hours), treated with CCl4 (40 hours), or subjected to 70% PH (40 hours). Total RNA and liver sections were prepared at 0 hours and 40 hours after liver regeneration.

RNA Analysis.

Total liver RNA was extracted using TRIzol Reagent from Invitrogen (Carlsbad, CA) according to the manufacturer's instructions. Quantitative real-time PCR was performed using SYBR Green PCR Master Mix and an ABI prism 7300 Sequence Detection System (Applied Biosystems, Foster City, CA). Murine 36B4 was used as internal control. PCR primers specific for each gene are listed in Supporting Table 1.

Western Blotting.

Livers or ileums were homogenized in protein lysis buffer (50 mM Tris-HCl, pH 7.4, 1% NP-40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EDTA, 1 mM PMSF, 1 mM Na3VO4, 1 mM NaF, and proteinase inhibitor cocktail). Proteins were resolved by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred to nitrocellulose membrane, and detected by chemiluminescence (Supersignal, Pierce). Western blotting was performed using antibodies (anti-FXR and β-actin) from Santa Cruz Biotechnology (Santa Cruz, CA).

Serum Bile Acids.

Serum after 70% PH or CCl4 treatment were collected and the bile acids and measured using a kit from Diagnostic Chemicals (Charlottetown, PE, Canada).

Statistic Analysis.

Data are expressed as means ± standard deviation (SD). Two-tailed Student's t test was used to determine significant differences between data groups. All analyses were performed using one-way analysis of variance (ANOVA). P < 0.05 (*) was considered statistically significant.

Results

Generation of the ΔL-FXR Mice and ΔIn-FXR Mice. A conditional FXR gene allele (FXR neo-flox) was generated in embryonic stem cells.

Two loxP sequences flank exons 4, 5, and 6 of the murine FXR allele (FXR Fl/Fl). FXR Fl/Fl mice were crossed with the albumin-cre or villin-cre mice to delete FXR gene specifically in liver or intestine, respectively. After correct genotyping, western blotting to measure FXR protein was performed using total protein extracted from liver and ileum. The results indicated no FXR expression in the liver of ΔL-FXR mice. However, liver FXR protein levels were comparable between the FXR Fl/Fl and ΔIN-FXR mice (Fig. 1A). Similarly, no ileum FXR expression was detected in the ΔIN-FXR mice (Fig. 1B).

Figure 1.

Generation of ΔL-FXR and ΔIn-FXR mice. Western blot analysis of FXR protein levels for the indicated mice. Liver protein extracts (A) and ileum protein extracts (B) were prepared and immunoblotted with FXR and β-actin antibodies.

Defective Liver Regeneration in ΔL-FXR Mice.

We previously showed that FXR in liver was required for promoting liver regeneration. To confirm the previous observation that hepatic FXR is required to promote liver regeneration, we compared the liver regeneration after 70% PH in FXR Fl/Fl, ΔL-FXR, and FXR KO mice. As expected, a significant delay in hepatocyte proliferation was observed in ΔL-FXR animals compared to FXR Fl/Fl mice at 24 hours, 36 hours, and 72 hours after surgery. Fewer BrdU-positive hepatocytes were present in ΔL-FXR mice than in FXR Fl/Fl mice (Fig. 2A). In FXR Fl/Fl mice, the hepatocyte proliferation peaked at 36 hours after 70% PH, but this peak was strongly reduced in ΔL-FXR mice compared to the FXR Fl/Fl mice (Fig. 2A). These results suggest that hepatic FXR is required to promote liver regeneration. However, to our surprise, compared to ΔL-FXR mice, FXR KO mice showed significantly decreased BrdU incorporation in the liver at 36 hours and 72 hours (Fig. 2A), suggesting that FXR in other tissues may also contribute to a maximum effect on promoting liver regeneration.

Figure 2.

Defective liver regeneration in ΔL-FXR and FXR KO mice after 70% PH. BrdU incorporation was counted on FXR Fl/Fl, ΔL-FXR, and FXR KO mice at indicated times after 70% PH (24, 36, 48, and 72 hours) (A). Total bile acid levels were measured in serum (B). CYP7a1 (C) and Foxm1b (D) gene expression were measured at indicated times after PH. ΔL-FXR: ΔL; FXR KO: KO.

We also compared the serum bile acid levels in FXR Fl/Fl, ΔL-FXR, and FXR KO mice. As expected, serum bile acid levels were significantly higher in the FXR KO and ΔL-FXR mice compared to the FXR Fl/Fl mice at 24 hours and 36 hours after 70% PH. On day 3, serum bile acid levels in ΔL-FXR mice returned to a comparable level compared to the control mice. However, bile acid levels were still significantly higher in FXR KO mice at day 3 (Fig. 2B). This suggests that, although hepatic FXR plays a role in suppressing bile acid levels after 70% PH, FXR in other tissues such as intestine may be required to suppress bile acid levels at later stages after 70% PH. Consistently, the gene encoding the rate-limiting enzyme of bile acids synthesis, CYP7a1, was suppressed in all three groups of mice after 70% PH, but CYP7a1 messenger RNA (mRNA) levels were much higher in ΔL-FXR and FXR KO mice compared to the FXR Fl/Fl mice (Fig. 2C). FXR was shown previously to directly activate the Foxm1b gene after 70% PH.6 We measured the mRNA levels of Foxm1b after 70% PH and found that the induction of Foxm1b was blocked in both ΔL-FXR and FXR KO mice compared to that in FXR Fl/Fl mice at 36 hours after PH (Fig. 2D). Therefore, hepatic FXR is responsible for the hepatic induction of Foxm1b gene expression during liver regeneration.

Impaired Liver Repair in ΔL-FXR Mice After CCl4-Induced Injury.

We previously showed that FXR is also required to promote liver repair following CCl4-induced liver injury.8, 19 We therefore asked whether hepatic FXR plays the same role in this liver repair model. Following a single dose of CCl4 injection, both ΔL-FXR and FXR KO mice displayed defective liver regeneration compared to the FXR Fl/Fl mice during the first 3 days (Supporting Fig. 1A). But there were significantly fewer BrdU-positive hepatocytes in FXR KO mice compared to ΔL-FXR mice (Fig. 3A). We next analyzed the serum bile acid levels after CCl4 treatment and found that FXR KO mice had much higher bile acid levels in serum than the other two groups of mice during the first day after CCl4 injection (Fig. 3B). But in all three groups of mice, CYP7a1 mRNA levels were dramatically suppressed (Fig. 3C). Consistently, the induction of Foxm1b gene expression also mainly depended on hepatic FXR activation because its mRNA levels were significantly lower in both ΔL-FXR and FXR KO mice compared to FXR Fl/Fl mice (Fig. 3D). Similarly, the cyclin D1 expression levels were much lower in ΔL-FXR and FXR KO mice than in FXR Fl/Fl mice (Fig. 3E).

Figure 3.

Impaired liver regeneration of ΔL-FXR mice and FXR KO mice after CCl4-induced liver injury. (A) BrdU-positive hepatocytes were counted at 24, 40, 48, and 72 hours after CCl4 injection. For each group of mice, 3-7 mice were used for each timepoint. Total serum bile acids were measured (B). Gene expression of CYP7a1 (C), FOXm1b (D), and CyclinD1 (E) were measured in FXR Fl/Fl, ΔL-FXR, and FXR KO mice after CCl4 treatment at indicated timepoints. (F) Semiquantification of liver necrosis areas of H&E staining in FXR Fl/Fl, ΔL-FXR, and FXR KO mice at 24, 40, 48, and 72 hours after CCl4 injection.

However, H&E staining showed that FXR KO mice displayed much more extensive liver injury compared to that in ΔL-FXR mice (Supporting Fig. 1B). Scores of the liver necrosis areas showed significant differences between ΔL-FXR and FXR KO mice (Fig. 3F). These results suggest that, in addition to liver, FXR in other tissues may be important to protect liver injury and promote liver repair.

Intestine FXR Contributes to Liver Regeneration/Repair After Either 70% PH or CCl4-Induced Injury.

Because FXR in intestine is crucial for the feedback regulation of bile acid synthesis in liver and FXR KO mice display more severe defects of liver regeneration compared to ΔL-FXR mice, we hypothesized that intestine FXR may also play roles in liver regeneration/repair. Therefore, we compared the liver regeneration between ΔIN-FXR and FXR Fl/Fl mice after 70% PH. The hepatic BrdU incorporation was significantly lower in ΔIN-FXR mice compared to FXR Fl/Fl mice at 48 hours after 70% PH (Fig. 4A). Consistent with a key role of intestine FXR in regulating bile acid levels, ΔIN-FXR mice had higher serum bile acid levels comparing to that in FXR Fl/Fl mice (Fig. 4B). Although the CYP7a1 gene was suppressed in both ΔIN-FXR and FXR Fl/Fl mice, the expression levels were much higher in ΔIN-FXR mice at 48 hours and 72 hours after 70% PH (Fig. 4C). Interestingly, we observed a strong induction of FGF15 and SHP gene expression in the intestine of FXR Fl/Fl mice on the first 2 days of liver regeneration (Fig. 4D,E). However, this induction was absent in ΔIN-FXR mice (Fig. 4D,E).

Figure 4.

Defective liver regeneration in ΔIN-FXR mice after 70% PH. PH was performed in FXR Fl/Fl and ΔIN-FXR mice and BrdU staining was measured at 24, 48, and 72 hours after PH (A), and total serum bile acids was measured at the indicated times (B). Liver CYP7a1 mRNA levels were measured (C). Ileum FGF15 (D) and SHP (E) were measured by quantitative real-time PCR.

Similarly, the number of BrdU-positive hepatocytes was lower on the first 2 days after CCl4-induced liver injury. At 40 hours, there were significantly fewer proliferative hepatocytes in ΔIN-FXR mice compared to that in FXR Fl/Fl mice (Fig. 5A). We also compared the necrosis areas in liver induced by CCl4. As shown in Fig. 5B and Supporting Fig. 2, CCl4 caused more severe liver injury in ΔIN-FXR mice than in FXR Fl/Fl mice. Although the CYP7a1 expression levels were decreased in both ΔIN-FXR and FXR Fl/Fl mice after CCl4 injection, the expression levels of CYP7a1 in the ΔIN-FXR mice were significantly higher compared to that in FXR Fl/Fl mice (Fig. 5C). This confirms that intestine FXR plays an important role in the regulation of CYP7a1 expression. We next measured the FGF15 expression levels in intestine and found that the induction of the FGF15 in the FXR Fl/Fl mice was blocked in ΔIN-FXR mice (Fig. 5D).

Figure 5.

Impaired liver regeneration in ΔIN-FXR mice after CCl4-induced liver injury. (A) Quantification of BrdU staining of hepatocytes after CCl4 injection in FXR Fl/Fl and ΔIN-FXR mice. (B) Semiquantification measurements of necrosis areas of H&E staining liver sections from FXR Fl/Fl and ΔIN-FXR mice at indicated timepoints after CCl4 treatment. Total liver RNA was isolated for quantitative real-time PCR analysis of liver CYP7a1 (C) and ileum FGF15 (D) mRNA levels at 24, 40, 48, and 72 hours after CCl4 injection in FXR Fl/Fl and ΔIN-FXR mice.

Ectopic Expression of FGF15 Rescues the Defective Liver Regeneration/Repair in ΔIN-FXR Mice.

FGF15 is a hormone that can mediate the effect of intestine FXR to regulate bile acid levels in liver. Because we observed that intestine-specific deletion of FXR resulted in greater defective liver regeneration/repair induced by 70% PH and CCl4, we therefore used both of the models to ask whether FGF15 plays a role in promoting liver regeneration/repair. ΔIN-FXR and FXR KO mice were injected with either a recombinant adenovirus that expresses FGF15 or a control adenovirus, and then 70% PH was performed or a single dose of CCl4 was administered. We first confirmed that the FGF15 adenovirus infection increased FGF15 expression in ΔIN-FXR and FXR KO mice (Fig. 6A,B). We then observed that hepatic BrdU incorporation was significantly increased in ΔIN-FXR and FXR KO mice after FGF15 adenovirus injection compared with the control mice receiving the adenovirus alone after 70% PH at 40 h (Fig. 6C). Similar results were also observed in a toxic CCl4-induced liver injury model (Fig. 6D; Supporting Fig. 3). BrdU incorporation was significantly increased in adenovirus FGF15 expression group comparing with the control group in ΔIN-FXR and FXR KO mice. CYP7a1 expression levels were down-regulated in the FGF15-infected mice compared to the controls in either the 70% PH model (Fig. 6E) or CCl4 model (Fig. 6F). These results indicate that FGF15 activated by intestine FXR indeed participates in promoting liver regeneration/repair.

Figure 6.

Ectopic expression of FGF15 rescued defective liver repair. Measurement of FGF15 levels after FGF15 adenovirus infection in ΔIN-FXR and FXRKO mice before and after 70% PH (40 hours) (A) and measurement of FGF15 levels after FGF15 adenovirus infection in ΔIN-FXR and FXR KO mice before and after CCl4 (40 hours) injection (B). Quantification of the BrdU-positive hepatocytes at 0 hours and 40 hours after 70% PH in ΔIN-FXR and FXR KO mice (C) and quantification of the BrdU-positive hepatocytes at 40 hours after CCl4 injection with and without FGF15 adenovirus injection (D). CYP7a1 gene expression levels were analyzed in ΔIN-FXR and FXR KO mice after virus infection in 70% PH model (E) and in CCl4 liver toxic model (F).

Discussion

We previously showed that FXR was required for normal liver regeneration and liver repair after injury. However, the mechanism by which FXR regulates this process is still unclear. In this report we show that hepatic and intestine FXR use distinct mechanisms to promote liver regeneration/repair.

Liver regeneration is regulated by many signals from the hepatic environment. Different signal pathways will lead to the activation of transcription factors that either stimulate hepatocyte proliferation or promote cell survival to promote liver regrowth.5, 20 We previously showed that FXR bound to an FXRE in Foxm1b intron 3 and induced Foxm1b gene transcription during liver regeneration.6 In FXR KO mice, this Foxm1b induction was blocked and liver regeneration was delayed. Consistent with these results, we observed that the induction of Foxm1b expression is dramatically reduced in ΔL-FXR mice compared with the control mice after either 70% PH or CCl4-induced liver injury. In contrast, the induction of Foxm1b was not affected in ΔIn-FXR mice after liver damage, indicating the requirement of a cell autonomous mechanism for hepatic FXR to activate Foxm1b and potentially other factors that are involved in regulating cell cycle in liver.

Bile acids are potentially toxic and substantial increases in hepatic bile acid levels will induce hepatocyte death.21 We previously demonstrated that FXR was activated by elevated bile acid influx during liver regeneration.5 The importance for a stringent control of bile acid levels is highlighted by a delicate regulation of CYP7a1 expression. The identified regulators of CYP7a1 expression include cytokines, growth factors,22–26 and nuclear receptors.27, 28 During liver regeneration, hepatic bile acid levels need to be suppressed rapidly to prevent the toxic effect of increased bile acids in liver, as shown by a dramatic down-regulation of CYP7a1 mRNA levels.5, 7 We previously showed that, in addition to the FXR-SHP axis, hepatocyte growth factor and JNK pathways were involved in suppressing CYP7a1 expression during the acute phases of liver regeneration.7 In the current study, we now further demonstrate that, during liver regeneration/repair, FXR also activates the expression of FGF15 in the intestine to suppress CYP7a1 transcription. Consistently, several reports also suggest that FGF15 secreted from ileum has profound effects on liver metabolism.14, 29, 30, 31 Because we previously showed that the suppression of CYP7a1 expression and decreased bile acid synthesis was beneficial for liver regeneration, we therefore conclude that FGF15 induction after liver damage may also contribute to normal liver regeneration.

The most novel observation in this report is the delayed liver regeneration/repair and increased liver injury in ΔIN-FXR mice compared to FXR Fl/Fl control mice after either 70% PH or CCl4 injection. There results identify an unexpected role of intestine FXR in regulating liver regeneration/repair. It is clear that intestine FXR is key to control bile acid levels. Thus, higher levels of bile acids in ΔIN-FXR mice after liver injury may hamper normal liver regeneration/repair. Besides its effect on bile acid levels, the metabolic and mitogenic activities of FGF15 cannot be excluded. Moreover, the hydrophobic bile acid, deoxycholic acid (DCA) is significantly increased in fecal extracts from intestine FXR null mice but not from FXR KO or liver FXR null mice,15 and DCA may cause hepatocyte apoptosis and colon inflammation and necrosis.32, 33, 34 This may also be a protective function of intestine FXR during liver regeneration/repair. We further showed that intestine FXR induced FGF15 expression after liver injury, which in turn suppressed the CYP7a1 transcription and lowered serum bile acid levels. Exogenous delivery of FGF15 rescued the defect of liver repair in ΔIN-FXR and FXR KO mice. However, we would like to mention that ectopic expression of FGF15 by way of adenovirus, which is an effective model to overexpress FGF15, may generate some side effects and liver toxicity due to virus infection. A better delivery approach of FGF15 will be needed in the future.

Our results strongly suggest a promotive effect of FGF15 in liver regeneration/repair. FGF15 has also been shown to down-regulate Foxo1 gene expression and Foxo1 is associated with cell cycle arrest and growth inhibition.35 This may also contribute to the overall effect of intestine FXR and FGF15 in promoting liver regeneration/repair. Furthermore, a recent report indicates that selective activation of intestine FXR or treating mice with FGF19 could reduce liver necrosis and inflammatory cell infiltration in cholestasis mouse models.36 Taken together, we conclude that intestine FXR and its induction of FGF15 may have more important roles in liver protection than we previously thought.

In summary, our results confirm a critical role of hepatic FXR in inducing Foxm1b expression and promoting liver regeneration/repair. Moreover, our studies demonstrate that intestine FXR activates FGF15 expression in the intestine to promote liver regeneration/repair. Therefore, in addition to the cell-autonomous effect of hepatic FXR, the endocrine FGF15 pathway induced by FXR in intestine also participates in the promotion of liver regeneration/repair.

Acknowledgements

We thank Dr. Steve Kliewer for providing adeno-FGF15. We thank the people in W.H.'s lab for technical assistance and scientific discussion.

Ancillary