How sweet it is to activate FXR


  • Peter L.M. Jansen,

    Corresponding author
    1. Division of Gastroenterology and Hepatology, Academic Medical Center, Amsterdam, The Netherlands
    2. Department of Surgery, NUTRIM School of Nutrition, Toxicology and Metabolism, Maastricht University, Maastricht, The Netherlands
    • Address reprint requests to: Peter L.M. Jansen, M.D., Ph.D., Division of Gastroenterology and Hepatology, Academic Medical Center, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands. E-mail:; fax: +31 20 5668278.

    Search for more papers by this author
  • Frank G. Schaap

    1. Division of Gastroenterology and Hepatology, Academic Medical Center, Amsterdam, The Netherlands
    Search for more papers by this author

  • See Article on Pages 1761 and 2023

  • Potential conflict of interest: Nothing to report.




farnesoid X receptor


nonalcoholic fatty liver disease.

The farnesoid X receptor (FXR) has emerged as a bile acid sensor and master regulator of diverse metabolic processes.[1, 2] FXR is a ligand-activated transcription factor, and is highly expressed in tissues with a high bile salt flux, i.e., the small intestine and liver. Both conjugated and unconjugated bile acids bind to FXR.[1] Since FXR is an intracellular receptor, membrane-impermeable conjugated bile acids have to be taken up by way of an appropriate carrier protein.[3] A main function of hepatic FXR is to prevent the harmful consequences of pathological bile acid overload in the hepatocyte.[1, 2] This is achieved through repression of bile acid uptake and synthesis, and stimulation of bile acid detoxification and secretion at the canalicular and basolateral poles. An important function of intestinal FXR is the suppression of bile acid synthesis in the postprandial phase. The FXR-inducible enterokine Fgf15/FGF19 plays a key role in this negative feedback loop by activating signaling pathways in the liver that target CYP7A1 expression, the main determinant of bile acid synthesis.[4] Apart from their effects on bile acid homeostasis, intestinal (by way of Fgf15/FGF19) and hepatic FXR control the expression of numerous metabolic genes in the liver, thus affecting postprandial metabolism.[1, 2, 4] Affected pathways include gluconeogenesis and lipogenesis, which are both repressed by activated FXR.

Bile acids are the major endogenous ligands of FXR and determine its activation. During the day, portal and systemic bile acid levels are subject to major diurnal and meal-induced fluctuations. Since the influence of FXR extends far beyond control of bile acid homeostasis, crosstalk with other metabolic pathways that modulate its activity can be expected. A number of recent studies have shed light on this issue.[5] FXR controls target gene expression in conjunction with a number of transcriptional cofactors (e.g., PGC1α, SIRT1) that affect interaction with the basal transcriptional machinery and modulate accessibility of the chromatin. FXR and its interacting cofactors are subject to (reversible) posttranslational modifications including methylation, acetylation and phosphorylation (Fig. 1). This couples the cellular metabolic state (i.e., availability of donor substrates and enzyme cofactors) to regulation of (metabolic) gene expression. The activity of FXR is greatly affected by these modifications either through effects on protein stability, interaction with cofactors, binding to genomic FXR-response elements, or transactivation potential.[5]

Figure 1.

Posttranslational modifications of FXR and hepatic processes controlled by FXR. Unliganded FXR (depicted as a heterodimer with the retinoid X receptor [RXR]) is bound to target promoter elements typically consisting of an inverted repeat with a single nt spacer (IR-1 motif).[1] Binding of ligand (here CDCA) induces a conformational change that results in release of co-bound corepressors and recruitment of transcriptional cofactors (e.g., PGC-1α, p300, SRC-1, SIRT1). FXR activity is modulated by a number of posttranslational modifications including phosphorylation, acetylation, and O-GlcNAcylation, providing crosstalk with metabolic signaling pathways and the cellular metabolic state. O-GlcNAcylation increases the transcriptional activity of FXR.[6] Acetylation (Ac) interferes with DNA binding of FXR and reduces its transcriptional activity.[5] Phosphorylation (P) increases transactivation activity of FXR.[5] Phosphorylation and acetylation sites are from Kemper,[5] the O-GlcNac site from Berrabah et al.[6] Positional information refers to the FXR isoform of 476 amino acids. AF1, activation function domain-1; DBD, DNA binding domain; LBD, ligand binding domain.

Berrabah et al.[6] now report on the posttranslational modification of FXR by O-GlcNAcylation at presumably multiple residues including a highly conserved serine residue located in its N-terminal activating function domain (AF1, Fig. 1). O-GlcNAcylation of specific serine/threonine residues is a common and dynamic modification that serves to modulate signaling and transcriptional networks in response to nutrients and cellular stress.[7] The synthesis of the donor substrate for O-GlcNAcylation, i.e., UDP-GlcNAc, is increased at high glucose levels when more glucose is shunted into the hexosamine biosynthetic pathway. Moreover, the enzyme responsible for O-GlcNAcylation (i.e., OGT) is activated by insulin signaling.[7] Targeting proteins that direct OGT to its protein substrates include the transcriptional cofactor PGC1α, a common component of nuclear receptor transcriptional complexes.[5, 7] Berrabah et al. show that the degree of O-GlcNAcylation of FXR is increased by high glucose levels in vitro or by a fasting-refeeding regimen in vivo. O-GlcNAcylated FXR is more responsive to FXR agonists such as the synthetic compound GW4064.6 Thus, high glucose conditions confer increased sensitivity of FXR-regulated genes to agonist challenge. Transcript profiling indicated that augmented agonist-response under these conditions affects the vast majority of FXR-driven genes in HepG2 cells and immortalized human hepatocytes.

The findings of Berrabah et al. interconnect glucose metabolism and FXR-regulated processes like bile acid homeostasis, and call for studies addressing how FXR activity is affected by chronic elevation of glucose levels and insulin resistance. It has already been established that the genomic binding sites of hepatic FXR are altered in obese mice, with likely—yet unknown—consequences for FXR target gene expression and FXR-regulated pathways. Moreover, FXR acetylation (a modification that impairs its transactivation potential) is elevated in diet-induced or genetically obese mice, and is accompanied by altered gene expression and an unfavorable metabolic outcome.[5] It remains to be determined if perturbed O-GlcNAcylation and/or enhanced removal of O-GlcNAc residues by O-GlcNAcase provides an additional level of FXR dysfunction in the insulin-resistant state.

Given the plethora of metabolic and other actions (e.g., antiinflammatory, stimulation of liver repair), FXR has been an attractive candidate for drug development. Treatment with the potent FXR agonist obeticholic acid causes an increase of insulin sensitivity and a reduction of inflammation and fibrosis markers in patients with nonalcoholic fatty liver disease (NAFLD) and type 2 diabetes.[8] In obese Zucker rats, treatment with obeticholic acid causes a reduction of hepatic fat content and improvement of hepatocyte ballooning.[9] The exact mechanism of how FXR agonism decreases hepatic triglycerides is debated. The transcriptional repressor SHP, a direct FXR target, has been implicated in this process. Watanabe et al.[10] attributed the FXR-mediated reduction of liver steatosis to SHP-mediated repression of SREBP1c, the master regulator of hepatic lipogenesis. However, transgenic expression of SHP failed to elicit downregulation of SREBP1c and lowering of hepatic triglycerides.[11] Additional studies (discussed in Xu et al.[12]) have cast doubt on the importance of the SHP/SREBP1c pathway in FXR-mediated lowering of hepatic lipid content.

In this issue, Xu et al.[12] report on a novel mechanism governing hepatic lipid homeostasis that involves carboxylesterase-1 (CES1). CES1 is highly expressed in the liver where it hydrolyzes a variety of esters of endogenous and xenobiotic origin including cholesteryl esters and certain prescription drugs. Xu et al.[12] now demonstrate that Ces1 is a triglyceride hydrolase, and a direct target of FXR in mice. Through gain- and loss-of-function experiments these authors elegantly show that Ces1 is required for FXR-mediated improvement of hepatic steatosis in lean and genetically obese mice. Adenoviral overexpression of Ces1 increases hepatic triglyceride hydrolysis, stimulates fatty acid oxidation, and, interestingly, led to a reduction of blood glucose levels. This apparent improvement of hepatic insulin signaling likely relates to reduced level of free fatty acids in the liver. Conversely, silencing of hepatic Ces1 had profound effects on de novo synthesis of palmitate, cholesterol and triglycerides and resulted in hepatic steatosis. This was attributed to loss of cholesterol esterase activity, with the perceived (unesterified) sterol-depleted state triggering activation of SREBP-mediated lipogenesis and cholesterologenesis. Hepatic Ces1 deficiency had no impact on blood glucose levels but raised non-high-density lipoprotein (HDL) cholesterol. Quiroga et al.[13] noted similar effects of global Ces1 deficiency on serum cholesterol. Intriguingly, these mice had reduced blood glucose levels, which was attributed to defective lipid utilization and switch to carbohydrate utilization. Treatment with obeticholic acid resulted in a moderate reduction of serum total cholesterol in wild-type but not in mice with hepatic Ces1 deficiency.[12] Such an effect of FXR agonism is, however, not apparent in patients who display a slight elevation of serum (LDL) cholesterol.[8]

The finding that hepatic Ces1 is a direct FXR target raises the intriguing possibility that decreased Ces1 levels observed in mouse models of obesity, insulin-resistance, or NAFLD[13] is related to FXR dysfunction. It remains to be determined if human CES1 is regulated by FXR as well, and if so, whether drug-drug interactions resulting from CES1-catalyzed deacylation of certain (pro)drugs can be expected in patients treated with FXR agonists. CES1, which is expressed in HepG2 cells, was not identified as a GW4064-regulated gene in the microarray experiments of Berrabah et al., and CES1 expression in human HuH7 hepatoma cells was unaffected by the endogenous FXR ligand CDCA.[14] Other aspects to be addressed in follow-up studies include the involvement of intestinal FXR in CES1-mediated control of chylomicron production, and the conservation of triglyceride hydrolyzing activity in human CES1.

In conclusion, these two studies add important insights into the regulation of FXR activity and diversity of its targets. An emerging concept is that FXR is dysfunctional in disorders associated with obesity. Stimulation of FXR activity by potent agonists may be beneficial in these disorders.

  • Peter L.M. Jansen, M.D., Ph.D.1,2 Frank G. Schaap, Ph.D.2

  • 1Division of Gastroenterology and Hepatology Academic Medical Center, Amsterdam The Netherlands

  • 2Department of Surgery, NUTRIM School of Nutrition, Toxicology and Metabolism, Maastricht University, Maastricht, The Netherlands