Final Report Summary - MECHFXR (Towards FXR-mediated therapeutic intervention: Understanding how FXR integrates metabolic, endocrine and inflammatory signaling.)
Towards FXR-mediated therapeutic intervention: Understanding how FXR integrates metabolic, endocrine and inflammatory signaling.
mechFXR
Dr A. Milona
The work on this project has focused on FXR, a nuclear receptor (NR) activated by bile acids, which regulates transcription of genes involved in bile salt, cholesterol, lipid and glucose homeostasis by classical transactivation mechanisms. FXR also regulates inflammation, possibly most probably via a mechanism involving FXR-mediated transrepression of NF-κB. These findings identify FXR as a potential therapeutic target for cholestasis, diabetes as well as inflammatory diseases of the gastrointestinal tract. Importantly, the anti-cholestatic and metabolic effects of FXR have been attributed to ligand-dependent gene transactivation, whereas the inflammatory actions are mediated by ligand dependent transrepression of Nf-κB. As with other NRs and preliminary data from clinical trials of an FXR full agonist (INT747), we know that global activation of FXR will cause side effects. This project has aimed to functionally dissect the different molecular mechanisms of FXR action during ligand-dependent transactivation and transrepression in order to gain essential insight into the basic biology of NR function and thereby improve NR drug design and efficacy.
Different post-translational modifications and interacting proteins are thought to discriminate between transactivation and transrepression. Post-translational modifications have been described in FXR, but how these modifications contribute to transcriptional changes leading to different biological responses, is unknown. In order to identify novel post-translational modifications in FXR, cells were transfected with FLAG-FXRa2 and were treated with or without GW4064 (synthetic FXR ligand) and TNFα (pro-inflammatory cytokine) for 24 hours. We performed FLAG-immunoprecipitations and identified post-translational modifications by Orbitrap mass-spectrometry.
Three novel acetylation sites were identified, Lysine-132, Lysine-251 and Lysine-353, all in phylogenetically conserved motives. The potential role of these acetylation sites is under investigation. A novel phosporylation site was also detected at Serine-224. The S224A mutant (defective in phosphorylation) resulted in abrogated transcriptional activity of FXR on SHP, IBABP and BSEP promoters in reporter assays. In contrast, transactivation capacity of FXR-S224D (phospho-mimicking) was similar or higher than wt FXR. Electro-mobility shift assays show absence of binding of FXR-S224A to BSEP, SHP and IBABP oligos, in contrast to FXR wt and FXR-S224D. However, the ability of FXR-S224A to inhibit NF-κB signalling by transrepression is unaffected in reporter assays. Phosphorylation of FXR-S224 is important for transactivation of FXR target genes, but not NF-κB transrepression in vitro. Selective FXR ligands that do not result in S224 phosphorylation can be useful to treat intestinal inflammation without interfering with bile salt, glucose and fat metabolism.
FXR interaction partners are also likely to be differential determinants of transactivation versus transrepression. We have set up a SILAC (Stable Isotope Labeling of Amino acids in cell Culture) assay to explore the profile of proteins interacting with FXR in transactivation and/or transrepression. Confirming the validity of the assayFrom initial results, we founddetected the well-established the Retinoid X Receptor as a binding partner in the transactivation situation, to which FXR is known to heterodimerize. In addition, we have detected 5 other proteins, that are currently validated.as well as other DNA binding proteins. However, these proteins seem to be in complex with FXR independent of ligand activation. We are currently optimizing the conditions to measure the dynamic ligand-mediated FXR interactions using RIME (rapid immunoimmune-precipitation mass spectrometry) that utilizes chemical cross-linking of the protein interactions before immunoprecipitation and identification of interacting proteins. The advantage of this method is that it includes the combination of formaldehyde crosslinking and on-bead digestion, permitting rapid and sensitive purification of interacting proteins endogenously in the livers of mice subjected in different treatments (FXR activation or in an inflammation induced model). .
In order to further exploit the potential of selective FXR ligands we also have set-up a high-throughput screening assay to identify FXR ligands, which allow transrepression of NF-κB signalling (anti-inflammatory), without affecting transactivation (metabolic) capacities of FXR. We have set up an automated luciferase assay to screen large chemical libraries to identify compounds which have the potential to modulate FXR function. HEK293T cells were transfected with an NFB promoter, TK renilla, FXR and RXR expression. Stimulation with chemical compounds was performed 24 hours post-transfection. NFB transcriptional activity was determined using luciferase assay the next day. Currently, we are analyzing the identified compounds for their transrepression capacity without influencing FXR transactivation actions.
In addition, we are interested to further explore the function and regulation of FXR in vivo, which could identify FXR to be implicated in novel pathways or diseases. We have previously reported that pregnancy has an important effect on liver physiology. We showed that pregnant mice have raised hepatic bile acid levels and pro-cholestatic gene expression, which could contribute in the development of a disease called Intrahepatic Cholestasis of Pregnancy (ICP). FXR regulates the transcription of the majority of these genes suggesting that the function of FXR might be perturbed during pregnancy. Further experiments indicated that estradiol might contribute to this effect. Also, estrogen receptor alpha (ERα) was shown to interact with FXR in vitro and suppress its function. However studies of their precise genome-wide interaction are lacking. Thus, we compared the genome-wide binding of FXR and ERα in livers of non-pregnant and pregnant mice and characterized their cooperative activity on binding to and regulating target gene transcription by ChIP-sequencing analysis. This identified 1864 FXR-binding sites in the non-pregnant livers and 25% were located within 2kb of a transcription start site, which is higher than predicted by random occurrence. Gene Ontology analysis showed FXR-binding sites close to genes involved in lipid and steroid metabolism. In addition, 75% of ERa binding sites overlaps with FXR in the non-pregnant livers. Remarkably, in pregnancy FXR but not ERa binding sites were significantly reduced. Furthermore, the signal of the ERα binding sites increased during pregnancy, including sites close to genes involved in carbohydrate and lipid metabolism as well as regulation of transcription. This study provides evidence on a genome-wide scale of both cooperative and independent interactions between FXR and ERα in regulating gene transcription in the liver.
This significant implication of ERa in several liver metabolic pathways (many common with FXR) led us to further investigate the effect and involvement of steroids in liver metabolism. Both steroid hormones and bile acids are made by cholesterol. In the liver, cholesterol escapes the body via bile acid formation and secretion into bile. Although, the liver is not considered a steroidogenic tissue, it expresses several steroidogenic genes under normal conditions. We have shown that pregnancy (high steroid concentrations) can expose cholestatic disease in genetically predisposed individuals and that estrogens likely interfere with the FXR function during pregnancy. Now, we extend this paradigm and present a novel, estrogen-modulated, hepatic target-gene of FXR/SHP, which may be involved in the pathology of cholestasis during pregnancy. The steroidogenic enzyme, Cyp17a1, is expressed in female, but not male mouse liver. Interestingly, hepatic Cyp17a1 is upregulated during pregnancy, Fxr deficiency and upon increased estrogen concentrations. In addition, bile acids robustly repress hepatic Cyp17a1 in vivo in an FXR-dependent manner in the liver, but not in the adrenal gland and the ovary. Cyp17a1 levels are mediated by a nuclear receptor cascade involving LRH-1, FXR and SHP. Adenoviral overexpression and knockdown of Cyp17a1 specifically in the liver resulted in altered lipid profile of triglycerides and lipoproteins in the serum and differential expression of genes involved in the regulation of lipid metabolism in the liver. Taken together, these data suggest that Cyp17a1 might have an important novel role in lipid metabolism in the liver, additional to that in steroidogenesis.
mechFXR
Dr A. Milona
The work on this project has focused on FXR, a nuclear receptor (NR) activated by bile acids, which regulates transcription of genes involved in bile salt, cholesterol, lipid and glucose homeostasis by classical transactivation mechanisms. FXR also regulates inflammation, possibly most probably via a mechanism involving FXR-mediated transrepression of NF-κB. These findings identify FXR as a potential therapeutic target for cholestasis, diabetes as well as inflammatory diseases of the gastrointestinal tract. Importantly, the anti-cholestatic and metabolic effects of FXR have been attributed to ligand-dependent gene transactivation, whereas the inflammatory actions are mediated by ligand dependent transrepression of Nf-κB. As with other NRs and preliminary data from clinical trials of an FXR full agonist (INT747), we know that global activation of FXR will cause side effects. This project has aimed to functionally dissect the different molecular mechanisms of FXR action during ligand-dependent transactivation and transrepression in order to gain essential insight into the basic biology of NR function and thereby improve NR drug design and efficacy.
Different post-translational modifications and interacting proteins are thought to discriminate between transactivation and transrepression. Post-translational modifications have been described in FXR, but how these modifications contribute to transcriptional changes leading to different biological responses, is unknown. In order to identify novel post-translational modifications in FXR, cells were transfected with FLAG-FXRa2 and were treated with or without GW4064 (synthetic FXR ligand) and TNFα (pro-inflammatory cytokine) for 24 hours. We performed FLAG-immunoprecipitations and identified post-translational modifications by Orbitrap mass-spectrometry.
Three novel acetylation sites were identified, Lysine-132, Lysine-251 and Lysine-353, all in phylogenetically conserved motives. The potential role of these acetylation sites is under investigation. A novel phosporylation site was also detected at Serine-224. The S224A mutant (defective in phosphorylation) resulted in abrogated transcriptional activity of FXR on SHP, IBABP and BSEP promoters in reporter assays. In contrast, transactivation capacity of FXR-S224D (phospho-mimicking) was similar or higher than wt FXR. Electro-mobility shift assays show absence of binding of FXR-S224A to BSEP, SHP and IBABP oligos, in contrast to FXR wt and FXR-S224D. However, the ability of FXR-S224A to inhibit NF-κB signalling by transrepression is unaffected in reporter assays. Phosphorylation of FXR-S224 is important for transactivation of FXR target genes, but not NF-κB transrepression in vitro. Selective FXR ligands that do not result in S224 phosphorylation can be useful to treat intestinal inflammation without interfering with bile salt, glucose and fat metabolism.
FXR interaction partners are also likely to be differential determinants of transactivation versus transrepression. We have set up a SILAC (Stable Isotope Labeling of Amino acids in cell Culture) assay to explore the profile of proteins interacting with FXR in transactivation and/or transrepression. Confirming the validity of the assayFrom initial results, we founddetected the well-established the Retinoid X Receptor as a binding partner in the transactivation situation, to which FXR is known to heterodimerize. In addition, we have detected 5 other proteins, that are currently validated.as well as other DNA binding proteins. However, these proteins seem to be in complex with FXR independent of ligand activation. We are currently optimizing the conditions to measure the dynamic ligand-mediated FXR interactions using RIME (rapid immunoimmune-precipitation mass spectrometry) that utilizes chemical cross-linking of the protein interactions before immunoprecipitation and identification of interacting proteins. The advantage of this method is that it includes the combination of formaldehyde crosslinking and on-bead digestion, permitting rapid and sensitive purification of interacting proteins endogenously in the livers of mice subjected in different treatments (FXR activation or in an inflammation induced model). .
In order to further exploit the potential of selective FXR ligands we also have set-up a high-throughput screening assay to identify FXR ligands, which allow transrepression of NF-κB signalling (anti-inflammatory), without affecting transactivation (metabolic) capacities of FXR. We have set up an automated luciferase assay to screen large chemical libraries to identify compounds which have the potential to modulate FXR function. HEK293T cells were transfected with an NFB promoter, TK renilla, FXR and RXR expression. Stimulation with chemical compounds was performed 24 hours post-transfection. NFB transcriptional activity was determined using luciferase assay the next day. Currently, we are analyzing the identified compounds for their transrepression capacity without influencing FXR transactivation actions.
In addition, we are interested to further explore the function and regulation of FXR in vivo, which could identify FXR to be implicated in novel pathways or diseases. We have previously reported that pregnancy has an important effect on liver physiology. We showed that pregnant mice have raised hepatic bile acid levels and pro-cholestatic gene expression, which could contribute in the development of a disease called Intrahepatic Cholestasis of Pregnancy (ICP). FXR regulates the transcription of the majority of these genes suggesting that the function of FXR might be perturbed during pregnancy. Further experiments indicated that estradiol might contribute to this effect. Also, estrogen receptor alpha (ERα) was shown to interact with FXR in vitro and suppress its function. However studies of their precise genome-wide interaction are lacking. Thus, we compared the genome-wide binding of FXR and ERα in livers of non-pregnant and pregnant mice and characterized their cooperative activity on binding to and regulating target gene transcription by ChIP-sequencing analysis. This identified 1864 FXR-binding sites in the non-pregnant livers and 25% were located within 2kb of a transcription start site, which is higher than predicted by random occurrence. Gene Ontology analysis showed FXR-binding sites close to genes involved in lipid and steroid metabolism. In addition, 75% of ERa binding sites overlaps with FXR in the non-pregnant livers. Remarkably, in pregnancy FXR but not ERa binding sites were significantly reduced. Furthermore, the signal of the ERα binding sites increased during pregnancy, including sites close to genes involved in carbohydrate and lipid metabolism as well as regulation of transcription. This study provides evidence on a genome-wide scale of both cooperative and independent interactions between FXR and ERα in regulating gene transcription in the liver.
This significant implication of ERa in several liver metabolic pathways (many common with FXR) led us to further investigate the effect and involvement of steroids in liver metabolism. Both steroid hormones and bile acids are made by cholesterol. In the liver, cholesterol escapes the body via bile acid formation and secretion into bile. Although, the liver is not considered a steroidogenic tissue, it expresses several steroidogenic genes under normal conditions. We have shown that pregnancy (high steroid concentrations) can expose cholestatic disease in genetically predisposed individuals and that estrogens likely interfere with the FXR function during pregnancy. Now, we extend this paradigm and present a novel, estrogen-modulated, hepatic target-gene of FXR/SHP, which may be involved in the pathology of cholestasis during pregnancy. The steroidogenic enzyme, Cyp17a1, is expressed in female, but not male mouse liver. Interestingly, hepatic Cyp17a1 is upregulated during pregnancy, Fxr deficiency and upon increased estrogen concentrations. In addition, bile acids robustly repress hepatic Cyp17a1 in vivo in an FXR-dependent manner in the liver, but not in the adrenal gland and the ovary. Cyp17a1 levels are mediated by a nuclear receptor cascade involving LRH-1, FXR and SHP. Adenoviral overexpression and knockdown of Cyp17a1 specifically in the liver resulted in altered lipid profile of triglycerides and lipoproteins in the serum and differential expression of genes involved in the regulation of lipid metabolism in the liver. Taken together, these data suggest that Cyp17a1 might have an important novel role in lipid metabolism in the liver, additional to that in steroidogenesis.