Potential conflict of interest: Nothing to report.
Supported by National Institutes of Health Grants GM551888, CA100070, AG039885, AG028865, CA159942 (to N. A. T.), and AG028865 (to G. J. D.).
One of the early events in the development of liver cancer is a neutralization of tumor suppressor proteins Rb, p53, hepatocyte nuclear factor 4α (HNF4α), and CCAAT/enhancer binding protein (C/EBP) α. The elimination of these proteins is mediated by a small subunit of proteasome, gankyrin, which is activated by cancer. The aim of this study was to determine the mechanisms that repress gankyrin in quiescent livers and mechanisms of activation of gankyrin in liver cancer. We found that farnesoid X receptor (FXR) inhibits expression of gankyrin in quiescent livers by silencing the gankyrin promoter through HDAC1-C/EBPβ complexes. C/EBPβ is a key transcription factor that delivers HDAC1 to gankyrin promoter and causes epigenetic silencing of the promoter. We show that down-regulation of C/EBPβ in mouse hepatoma cells and in mouse livers reduces C/EBPβ-HDAC1 complexes and activates the gankyrin promoter. Deletion of FXR signaling in mice leads to de-repression of the gankyrin promoter and to spontaneous development of liver cancer at 12 months of age. Diethylnitrosoamine (DEN)-mediated liver cancer in wild-type mice also involves the reduction of FXR and activation of gankyrin. Examination of liver cancer in old mice and liver cancer in human patients revealed that FXR is reduced, while gankyrin is elevated during spontaneous development of liver cancer. Searching for animal models with altered levels of FXR, we found that long-lived Little mice have high levels of FXR and do not develop liver cancer with age and after DEN injections due to failure to activate gankyrin and eliminate Rb, p53, HNF4α and C/EBPα proteins. Conclusion: FXR prevents liver cancer by inhibiting the gankyrin promoter via C/EBPβ-HDAC1 complexes, leading to subsequent protection of tumor suppressor proteins from degradation. (HEPATOLOGY 2013)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The development of hepatocellular carcinoma is a multistep process that includes the progressive alterations of gene expression leading to liver proliferation and liver cancer.1 The studies of liver regeneration after partial hepatectomy identified several critical steps of the initiation of liver proliferation.2 However, molecular mechanisms that trigger liver proliferation during development of liver cancer are not known. The quiescent stage of the liver is supported by a member of the CCAAT/enhancer binding protein (C/EBP) family, C/EBPα.3 Because three other tumor suppressor proteins—p53, Rb, and p16—protect the liver from the development of cancer,1 one would assume that the liver is well protected. Moreover, the growth inhibitory activities of some of these proteins are increased with age.3, 4 Despite these activations, the frequency of liver cancer increases with age,5, 6 suggesting that the tumor suppressor proteins are eliminated by a specific mechanism. We recently found that the age-associated development of liver cancer is mediated by activation of gankyrin,5 which is a component of 26S proteasome.7 Gankyrin also eliminates the growth inhibitory activities of Rb, p53, and p16. Elimination of C/EBPα and Rb is mediated by a direct interaction of gankyrin with these proteins and their subsequent degradation.5, 8 Gankyrin-mediated elimination of p53 involves activation of MDM2 ligase, which triggers degradation of p53 through a ubiquitin proteasome system.9 Gankyrin also neutralizes p16 by the replacement of p16 from cdk4.10
Gankyrin has been first discovered as a small non–adenosine triphosphate (ATP) subunit of 26S proteasome and as a protein that is increased in human hepatocellular carcinoma.7, 11 It has been shown that the development of liver cancer in animal models of carcinogenesis involves activation of gankyrin.1, 11 Moreover, the short hairpin RNA (shRNA)-mediated inhibition of gankyrin reduces the development of liver cancer in the nude mice.12 Recent studies have shown that gankyrin expression is increased in colorectal carcinoma samples, in pancreatic cancer, and in human lung cancers.13-15 In the present study, we found that farnesoid X receptor (FXR) represses gankyrin in quiescent livers and that liver cancer activates gankyrin via a release of this repression.
All animal studies were approved by the Institutional Animal Care and Use Committee at Baylor College of Medicine (protocol AN-1439). The generation and characterization of FXR/SHP knockout (KO) mice have been described,15 and Little mice have been characterized.16, 17 Protein extracts from human liver tumors and from healthy patients were obtained from OriGene (Rockville, MD).
Liver tumors were induced in wild-type (WT) and Little mice via diethylnitrosoamine (DEN) tumor liver induction as described.5 For FXR agonist treatment experiment, 8-week-old mice were injected with FXR agonist GW4064 intraperitoneally (30 mg/kg body weight). Control mice were injected with vehicle (corn oil).
Isolation of Nuclear and Cytoplasmic Extracts and Western Blotting.
Nuclear and cytoplasmic extract isolation and western blot analysis were performed as described our previous publications.18, 19 A typical picture of the quality of the separation of cytoplasmic and nuclear proteins is shown in Supporting Fig. 2.
Reverse-Transcription Polymerase Chain Reaction.
Total RNA from liver tissues or Hep3B2 cells was extracted with an RNeasy Mini Kit (Qiagen, Germantown, MD) according to the manufacturer's instructions. Complementary DNA was synthesized using SuperScript III First-strand (Invitrogen) and random primer hexamers. The primer sequences used in the studies are presented in the Supporting Information.
Chromatin Immunoprecipitation Assay.
Chromatin immunoprecipitation assay (ChIP) was performed as described5, 18 using the ChIP-IT kit (Active Motif, Carlsbad, CA).
Electrophoretic Mobility Shift Assay.
Electrophoretic mobility shift assay was performed as described.19
Antibodies and Reagents.
Antibodies to FXR (C20 and H130), gankyrin, C/EBPβ (C19), C/EBPα (A144), cdk4, cdc2, cyclin D3, Rb, p53, and HDAC1 (H-51) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies to acetyl-histone H3 (Lys9) and histone H3 trimethyl Lys9 were obtained from Abcam (Cambridge, MA). Monoclonal anti–β-actin antibody was from Sigma (St. Louis, MO). The bromodeoxyuridine (BrdU) uptake assay kit was obtained from Invitrogen (Carlsbad, CA). Co-immunoprecipitation studies were performed using TrueBlot reagents as described.5, 19
Knockdown of C/EBPβ in Hepa 1-6 Cells and in Liver.
Hepa 1-6 cells were transduced with the shRNA-expressing lentivirus (Sigma-Aldrich, St. Louis, MO), and stable cell lines were generated by selection with puromycin for 2 weeks. For in vivo silencing experiments, 3-month-old mice were injected via the tail vein with C/EBPβ siRNA or nontarget siRNA (50 μg of siRNA from Dharmcon complexed with in vivo-jet PEI, N/P ratio of 6 per mouse).
The Development of Liver Cancer in FXR/SHP KO Mice Involves Activation of Gankyrin and Elimination of C/EBPα.
FXR/SHP KO mice have hepatobiliary dysfunctions, including increased liver proliferation16 and development of liver cancer at age of 12 months (Anakk et al., submitted for publication). Because gankyrin-mediated elimination of C/EBPα is one of the key events in the development of liver cancer,5 we examined whether this pathway is activated in the livers of FXR/SHP KO mice. Figure 1A shows a typical liver of a 17-month-old FXR/SHP KO mouse with advanced cancer. BrdU uptake confirmed that liver proliferation was increased in these animals (Fig. 1B). Because C/EBPα needs to be phosphorylated at S193 by cdc2 and cdk4 to be degraded by gankyrin,5 we examined the expression of C/EBPα, gankyrin, cdc2, and cdk4 in livers of FXR/SHP KO mice. Figure 1C shows that gankyrin was elevated in the livers of FXR/SHP KO mice. The elevation of gankyrin correlated with elimination of C/EBPα protein, but not C/EBPα messenger RNA (mRNA) (Fig. 1D). Protein levels of cdc2, cdk4, and cyclin D3 were increased in livers of FXR/SHP KO mice (Fig. 1C). We next asked whether gankyrin is activated in FXR/SHP KO mice during the early stages of liver cancer. Examination of 6-month-old mice revealed that gankyrin increased significantly in livers of FXR/SHP KO mice; however, C/EBPα levels were reduced only slightly (Fig. 1E). Because the ph-S193 isoform of C/EBPα is a target of gankyrin, we suggested that the remaining 40%-50% of C/EBPα might not be phosphorylated at S193. We have shown that the phosphatase PP2A eliminates the phosphate from S193.20 Our studies of FXR/SHP mice revealed that PP2A was increased and the ph-S193 isoform of C/EBPα was not detectable in the nuclear extracts of livers from 6-month-old FXR/SHP KO mice (Fig. 1E). We also found that the enzymes, which phosphorylate C/EBPα at S193, were weakly activated at this age in FXR/SHP KO mice (Supporting Fig. 1A,B).
FXR Is Reduced and Gankyrin Is Elevated in Spontaneously Developed Mouse and Human Liver Tumors.
We next examined whether spontaneous liver tumors might have reduced FXR. Western blotting with proteins from liver tumors of 24-month-old mice revealed a reduction of FXR and elevation of gankyrin (Fig. 2A,B). Consistent with data in FXR/SHP KO mice, protein levels of C/EBPα were reduced in these tumor samples, whereas the levels of C/EBPα mRNA were unchanged (data not shown). We further examined expression of FXR, gankyrin, and C/EBPα in the livers of four patients with advanced liver cancer and in four normal patients. Figure 2C shows that FXR was reduced to 15%-20% in all examined tumor samples and that gankyrin was elevated in these samples. Western blot analysis revealed that C/EBPα was dramatically reduced in all human tumor samples. Thus, these studies revealed that spontaneous development of liver cancer in mice and in humans involves reduction of FXR, elevation of gankyrin, and reduction of C/EBPα.
FXR Inhibits Expression of Gankyrin.
The search for the FXR binding sites revealed no consensuses within the 1.4-kb region of the mouse gankyrin promoter, suggesting indirect mechanisms of the FXR-mediated repression of the promoter. Previous studies revealed that FXR directly binds to the C/EBPβ promoter21 and that C/EBPβ-HDAC1 complexes are abundant in the liver and repress C/EBP-dependent promoters.19 Therefore, we hypothesized that FXR might repress the gankyrin promoter through C/EBPβ-HDAC1 complexes. We found that the gankyrin promoter contained two consensuses for C/EBPβ and that C/EBPα and C/EBPβ bound to the gankyrin promoter in vitro (Fig. 3A,B). ChIP assay revealed that C/EBPα, C/EBPβ, and HDAC1 occupied the gankyrin promoter in the livers of WT animals. However, C/EBPβ and HDAC1 were not observed on the gankyrin promoter in livers of FXR/SHP KO mice (Fig. 3C). In agreement with these data, the activation of FXR in cultured mouse Hepa 1-6 cells by the ligands chenodeoxycholic acid (CDCA) and GW4064 reduced levels of gankyrin protein (Fig. 3D) and gankyrin mRNA (Fig. 3E). We observed that activation of FXR also increased levels of C/EBPβ and, surprisingly, HDAC1 (Fig. 3D).
FXR-Mediated Inhibition of Gankyrin Requires C/EBPβ.
We next examined whether the inhibition of gankyrin involves C/EBPβ-HDAC1 complexes. We found that activation of FXR in Hepa 1-6 cells increased amounts of the C/EBPβ-HDAC1 complexes (Fig. 4A) and that C/EBPβ-HDAC1 complexes occupied the gankyrin promoter (Fig. 4B). To examine whether the FXR-dependent inhibition of gankyrin requires C/EBPβ, we generated two cell lines (C3a and C4a) expressing shRNA to C/EBPβ, which dramatically inhibits C/EBPβ (Fig. 4C). The activation of FXR by CDCA in the control clone inhibited expression of gankyrin; however, FXR failed to inhibit gankyrin in clones C3a and C4a (Fig. 4D).
To determine whether C/EBPβ is required for the repression of gankyrin in quiescent livers, we inhibited C/EBPβ by siRNA as shown in Fig. 4E. The down-regulation of C/EBPβ led to a significant reduction of C/EBPβ-HDAC1 complexes. The reduction of C/EBPβ-HDAC1 complexes correlated with the elevation of gankyrin mRNA and protein (Fig. 4E,F). These studies show that FXR represses the gankyrin promoter and that this repression requires C/EBPβ.
Gankyrin Promoter Is Activated in Liver Cancer via Release of FXR-C/EBPβ-HDAC1–Mediated Repression.
We next examined the mechanisms that activate gankyrin during development of liver cancer after DEN injection. Because gankyrin is elevated during the early stages of DEN-mediated cancer,5 we examined the FXR-C/EBPβ-gankyrin pathway at days 2, 4, and 7 after DEN injection. FXR and C/EBPβ were reduced, whereas expression of gankyrin was elevated at days 2 and 4 (Fig. 5A, upper). The decline of FXR and C/EBPβ led to a reduction of the C/EBPβ-HDAC1 complexes (Fig. 5A, bottom). Examination of C/EBPβ and HDAC1 in FXR/SHP KO mice revealed that, at the age of 12 months, C/EBPβ expression was elevated in the livers of these mice, and amounts of C/EBPβ-HDAC1 complexes increased as well (Fig. 5B). However, these complexes were not bound to the gankyrin promoter (Fig 5C). We next examined the status of the gankyrin promoter and found that C/EBPα/β-HDAC1 complexes occupied and repressed the gankyrin promoter in quiescent liver, since histone H3 was trimethylated at K9 on the promoter (Fig. 5C). However, C/EBPβ and HDAC1 were removed from the gankyrin promoter in livers of DEN-injected mice, which led to acetylation of histone H3 at K9. Consistent with these data, the gankyrin promoter is also activated in FXR/SHP KO mice.
Activation of FXR in DEN-Treated Mice Inhibits Elevation of Gankyrin.
To determine whether the reduction of FXR is responsible for the elevation of gankyrin after DEN injection, we activated FXR by GW4064 and then treated mice with DEN. In control animals treated with corn oil, the expression of FXR, C/EBPβ, HDAC1, and gankyrin was similar to that observed in mice without GW4064 treatment (Fig. 5D). However, the activation of FXR by GW4064 supported high levels of C/EBPβ and C/EBPβ-HDAC1 complexes that correlated with the lack of activation of gankyrin (Fig. 5E). ChIP assay revealed that the C/EBPβ-HDAC1 complexes occupied and repressed the gankyrin promoter in GW4064-treated mice (Fig. 5F).
Little Mice Express High Levels of FXR and Do Not Develop Liver Cancer After DEN Injection.
Previous studies have shown that long-lived Little mice have increased levels of genes involved in the xenobiotic detoxification and that crossing these mice with FXR KO mice corrected their expression.17 We performed western blot analysis and found a four- to five-fold elevation of FXR in 24- to 36-month-old Little mice (Fig. 6A,B). It has been shown that the frequency of liver tumors increases with age and reaches around 30% at the age of 24 months.5 However, Little mice do not develop liver cancer with age. Therefore, we tested the hypothesis that high levels of FXR in old Little mice protect the liver from development of cancer. WT and Little mice were treated with DEN, and liver tumors were examined 35-36 weeks after DEN injection. We examined five WT mice and five Little mice and found that all WT animals developed advanced liver cancer, whereas only two Little mice had few tumor nodules of a very small size (Fig. 6C). Three other Little mice did not have liver cancer. Examination of liver sections via hematoxylin and eosin staining revealed that the livers of WT mice contained multiple diverse nodules of proliferating hepatocytes, including enlarged cells with moderate anisonucleosis on the left and a cluster of small, uniform, deeply basophilic cells on the right (Fig. 6D). In contrast, livers of Little mice treated with DEN showed unremarkable architecture and cytology, with uniform hepatocytes containing minimal cytoplasmic lipid and glycogen. We found that the number of replicating hepatocytes increased significantly in WT mice (up to 25%-30%), while around 5% of hepatocytes were BrdU-positive in the livers of Little mice (Fig. 6E,F). These data show that Little mice are resistant to the development of liver cancer after DEN treatment.
Little Mice Do Not Activate Gankyrin.
We next determined the molecular mechanisms by which Little mice are protected from liver cancer. A recent report showed that gankyrin causes degradation of the liver-specific transcription factor hepatocyte nuclear factor 4α (HNF4α).22 Therefore, we included this protein in our studies. We found that gankyrin was elevated and that it caused reduction of C/EBPα, Rb, HNF4α, and p53 in control WT mice (Fig. 7A,B). FXR was slightly reduced in WT mice; however, in Little mice, FXR levels remained at high levels, leading to the lack of activation of the gankyrin and to no reduction of C/EBPα, Rb, HNF4α, or p53. The reduction of the tumor repressor proteins in WT mice took place on the levels of protein degradation, since levels of mRNA were not changed significantly (Fig. 7C). To determine whether gankyrin is responsible for the degradation of tumor suppressor proteins, we examined interactions of these proteins with gankyrin. In these experiments, we used up to 1 mg of nuclear extracts for the co-immunoprecipitation studies. The remaining C/EBPα, Rb, p53, and HNF4α proteins were bound to gankyrin in WT mice but were not detected in gankyrin IPs from Little mice (Fig 7A, bottom). Examination of C/EBPβ and HDAC1 revealed that C/EBPβ and HDAC1 were increased in the livers of Little mice (Fig. 7D). We found that the amounts of C/EBPβ-HDAC1 complexes are higher in Little mice and that these complexes occupy and repress the gankyrin promoter in Little mice treated with DEN (Fig. 7E).
Gankyrin is a protein that is activated in liver cancer and causes degradation or elimination of activities of five tumor suppressor proteins; Rb, p53, C/EBPα, HNF4α, and p16.1, 5-7, 22 This places gankyrin in a unique position to be a target for therapeutic approaches in the prevention of liver cancer. In this study, we elucidated the mechanisms of activation of gankyrin during the development of liver cancer. Four lines of evidence show that development of liver cancer involves the reduction of FXR and subsequent activation of gankyrin. First, DEN-mediated carcinogenesis in WT mice reduces FXR, leading to the reduction of HDAC1-C/EBPβ complexes and activation of the gankyrin promoter. Second, the deletion of FXR signaling in FXR/SHP KO mice activates gankyrin in the liver, leading to development of liver cancer. Third, high levels of FXR in Little mice prevent development of age-associated liver cancer and development of cancer under DEN protocol. Fourth, levels of FXR are reduced in spontaneously developed mouse and human liver tumors, whereas gankyrin is elevated. Fig. 7F summarizes our studies and presents our hypothesis, according to which the elevation of gankyrin triggers degradation of four tumor suppressor proteins and leads to liver cancer. Based on the literature and our observations, we suggest that the gankyrin-mediated elimination of C/EBPα is associated with phosphorylation at S193, while other proteins might be degraded by additional mechanisms such as activation of MDM2 (for p53) and direct interactions of gankyrin with Rb. These findings provide a basis for the generation of gankyrin-based therapeutic approaches in the prevention of liver cancer.