Discovery of farnesoid X receptor and its role in bile acid metabolism

Abstract

In 1995, the nuclear hormone orphan receptor farnesoid X receptor (FXR, NR1H4) was identified as a farnesol receptor expressed mainly in liver, kidney, and adrenal gland of rats. In 1999, bile acids were identified as endogenous FXR ligands. Subsequently, FXR target genes involved in the regulation of hepatic bile acid synthesis, secretion, and intestinal re-absorption were identified. FXR signaling was proposed as a mechanism of feedback regulation of the rate-limiting enzyme for bile acid synthesis, cholesterol 7α-hydroxylase (CYP7A1). The primary bile acids synthesized in the liver are transformed to secondary bile acids by the gut microbiota. The gut-to-liver axis plays a critical role in the regulation of bile acid synthesis, composition and circulating bile acid pool size, which in turn regulates glucose, lipid, and energy metabolism. Dysregulation of bile acid metabolism and FXR signaling in the gut-to-liver axis contributes to metabolic diseases including obesity, diabetes, and non-alcoholic fatty liver disease. This review will cover the discovery of FXR as a bile acid sensor in the regulation of bile acid metabolism and as a metabolic regulator of lipid, glucose, and energy homeostasis. It will also provide an update of FXR functions in the gut-to-liver axis and the drug therapies targeting bile acids and FXR for the treatment of liver metabolic diseases.

1. Introduction

Farnesoid X receptor (FXR, NR1H4) is a member of a nuclear receptor superfamily highly expressed in the gastrointestinal tract, where bile acids are synthesized, metabolized, and reabsorbed for recirculation from the intestine to the liver. Bile acids are endogenous ligands of FXR, which acts as a bile acid sensor to control bile acid synthesis. FXR also serves as a metabolic regulator and integrator of glucose, lipid, and energy metabolism [1, 2]. FXR critically regulates the enterohepatic circulation of bile acids by controlling transcription of key regulatory genes in bile acid synthesis, biliary bile acid secretion, and trans-intestinal bile acid transport to the liver via portal blood circulation. This is a highly efficient physiological pathway that not only maintains bile acid homeostasis but also transports nutrients and drugs from the intestine to the liver for metabolism and distribution to other tissues [3]. The enterohepatic circulation of bile acids integrates bile acid signaling through FXR in the gastrointestinal tract and other organs to coordinately regulate metabolic functions in target organs, including cell growth and differentiation.

Bile acids are amphipathic detergent molecules that are required for intestinal absorption, digestion and solubilization of fats, lipid soluble vitamins, nutrients, and drugs, which are then transported to the liver for metabolism, detoxification and distribution to other tissues and organs [4]. Bile acids are the end products of cholesterol catabolism in the liver. The conversion of cholesterol to bile acids for biliary secretion is the major pathway to eliminate excessive cholesterol and maintain whole body cholesterol homeostasis, and the liver is the only organ that expresses all the enzymes needed for the synthesis of primary bile acids.

The gut microbiome plays a critical role in the biotransformation of primary bile acids to secondary bile acids, and it determines bile acid composition and circulating bile acid pool size. Bile acids, in turn, control the gut bacterial community (microbiome). Dysregulation of bile acid metabolism and FXR signaling alters the gut microbiome and may negatively impact host metabolism and homeostasis (dysbiosis) and contribute to liver metabolic diseases, obesity, Type 2 diabetes and non-alcoholic fatty liver disease (NAFLD), the prevalence of which has increased substantially to account for 25% of the population in the US and around the world [5–8]. Currently, there is no FDA approved drug to treat nonalcoholic steatohepatitis (NASH), a progressive form of NAFLD.

Cholestasis is a chronic liver condition resulting from obstructed hepatic bile flow, leading to accumulation of cytotoxic bile acids in the liver and increased bile acids in the systemic circulation [9]. Chronic cholestasis leads to fibrosis, cirrhosis, liver failure and higher risk of hepatocellular carcinoma (HCC) and cholangiocarcinoma [10–13]. FXR signaling plays a critical role in protection against cholestasis by regulating several hepatobiliary bile acid transporters; however, the underlying mechanism of hepatoprotection of FXR signaling is not entirely clear [14].

This review will describe the discovery of FXR as bile acid-activated receptor, the role of FXR in the regulation of bile acid synthesis and enterohepatic circulation of bile acids, the metabolic functions of FXR in the liver and intestine, bile acid-related liver diseases, and drug therapy targeting bile acids and FXR for cholestasis and NASH. Older original references on FXR, bile acid synthesis and FXR regulation of bile acid synthesis and transport and hepatic metabolism are cited, as are review articles mostly published within the last ten years on bile acid and FXR function in liver metabolism, diseases, and therapies.

2. Nuclear receptor gene superfamily

Nuclear hormone receptors are activated by endocrine hormones at physiological concentrations, which stimulates a cascade reaction to amplify hormone signals that regulate cell growth, development, and differentiation [15–17]. The nuclear receptor (NR) superfamily consists of 48 genes in the human genome [18] (Fig. 1).

Fig. 1.

Farnesoid X receptor (FXR) is a member of nuclear receptor superfamily. The domain structure of nuclear receptors is shown. Nuclear receptor superfamily consists of 48 nuclear receptors, which can be separated into three types based on ligand identity and 7 groups based on nucleotide sequence alignment. Nuclear receptor binds to the hormone response element, arranged as direct repeat (DR), everted repeat (ER) or inverted repeat (IR) as a monomer or homodimer, or heterodimer with RXR in the gene promoter. Without ligand binding, nuclear receptor binds co-repressors and is inactive. Upon ligand binding, co-repressors are released to allow recruitment of co-activators to stimulate RNA polymerase II transcriptional activity.

Details of each nuclear hormone receptor in the superfamily can be found in the Nuclear Receptor Signaling Atlas (NRSA) (signalingpathways.org) and the Concise Guide to Pharmacology.

(https://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=695).

Nuclear receptors have distinct domain structures termed A/B, C, D and E domains. The highly variable domain A/B is in the N-terminal domain. This domain contains the activation function-1 domain (AF-1), which interacts with cell-specific coregulators and is subject to posttranslational modifications (including phosphorylation, acetylation, SUMOylation, and others). The DNA-binding domain (DBD, domain C) contains the highly conserved zinc fingers that bind DNA. There are two adjacent zinc fingers, each containing four cysteine residues that coordinate a zinc ion in the center to bind to the major grove of DNA double helix. The D domain is the hinge region located between the DBD and ligand-binding domain (LBD, domains E/F) and contains a nuclear localization signal (NLS). The LBD is highly conserved and contains helices 3, 4 and 12, and contains the AF-2 domain (in the E domain).

NRs are classified according to their ligand identity: type 1 endocrine receptors (classic steroid hormone receptors), type 2 adopted orphan receptors (endogenous ligands identified) and type 3 orphan receptors (no identified ligand) (Fig. 1). Sequence alignment and phylogenetic analysis have classified 48 NRs to 7 groups: NR1, NR2, NR3, NR4, NR5, NR6 and NR0 (atypical receptors lacking a DBD).

FXR is a type 2 receptor that closely relates to liver X receptor α and β (LXRα, NR1H3; LXRβ, NR1H2), peroxisome proliferator-activated receptor α, γ and δ (PPARα, NR1C1; PPARγ, NR1C2; PPARδ, NR1C3), pregnane X receptor (PXR, NR1I2) and constitutive androgen receptor (CAR, NR1I3). They form a heterodimeric complex with a common partner, RXRα. The heterodimer binds to DNA sequences of AGGTCA-like motifs arranged in a direct-repeat, everted-repeat or inverted-repeat with various numbers of nucleotides in between two half sites (Fig. 1).

3. Discovery of FXR as a bile acid-activated receptor

In 1995, Forman et al. utilized a mammalian two-hybrid system to demonstrate that FXR and RXRα heterodimerize, and further identified farnesol metabolites as specific FXR ligands using luciferase reporter assays [19]. Farnesol metabolites are intermediates in the mevalonate pathway which synthesizes cholesterol, bile acids, steroids, and farnesylated proteins. They found that FXR mRNA expression was restricted to the liver, kidney, adrenal glands, and intestinal villi of rats. FXR and RXRα heterodimers preferentially bind to an inverted repeat with one nucleotide spacing (IR1, AGGTCA n TGACCT) in the target gene promoter.

In 1999, three laboratories reported that bile acids were endogenous FXR ligands that repress transcription of the gene encoding the rate-limiting enzyme for bile acid synthesis, cholesterol 7α-hydroxylase (CYP7A1) [20–22]. The efficacy of bile acids as activators of FXR has been determined by reporter assay, in the order of chenodeoxycholic acid (CDCA, EC₅₀=17 μM) > deoxycholic acid (DCA) > lithocholic acid (LCA) > cholic acid (CA, EC₅₀ = ~600 μM) [20]. Co-activator recruitment assays demonstrated that CDCA (EC₅₀ = 11.7 μM) is the most potent and efficacious bile acid to recruit steroid receptor co-activator-1 (SRC-1), followed by DCA (EC₅₀ = 19.0 μM), while ursodeoxycholic acid (UDCA) is very weak and CA is inactive. Efficacies are correlated to their ability to repress expression of CYP7A1 or other FXR target genes [23].

4. Bile acid synthesis in the liver

In the liver, cholesterol is catabolized to bile acids through two multi-step biosynthetic pathways [24]. The classic bile acid synthesis pathway (neutral pathway) is initiated by microsomal CYP7A1 and several enzymes in microsomes and the cytosol to modify the steroid ring, then followed by mitochondrial steroid 27-hydroxylase (CYP27A1) to oxidize the steroid sidechain. The alternative bile acid synthesis pathway (acidic pathway) is initiated by CYP27A1 to hydroxylate the steroid sidechain followed by modification the steroid ring as in the classic pathway. In the classic bile acid synthesis pathway, CYP7A1 is the first and rate-limiting enzyme which hydroxylates cholesterol specifically at the 7α-position to form 7α-hydroxycholesterol. Then, the bile acid-specific hydroxysteroid dehydrogenase 3B7 (HSD3B7) converts 7α-hydroxycholesterol to 7α-hydroxy-4-cholesten-3-one (C4). C4 is the common precursor for the two major primary bile acids, cholic acid (CA, 3α, 7α, 12α) and chenodeoxycholic acid (CDCA, 3α, 7α), and serum levels of C4 are a surrogate marker for the rate of bile acid synthesis [25]. Sterol 12α-hydroxylase (CYP8B1) catalyzes 12α-hydroxylation of C4 leading to synthesis of CA. Without this step, C4 is converted to CDCA. Aldo-keto-reductases (AKR1D1 and AKR1C4) convert the 3-keto group of C4 to 3β-HO and mitochondrial CYP27A1 catalyzes the sterol side-chain oxidation at C-27 (also named C-26S) to form tri-hydroxycholestanoic and di-hydroxycholestanoic acids, which are activated by bile acid-CoA synthase (BACS, SLC27A1) in the peroxisomal membrane to form acyl-CoAs. Acyl-CoAs enter the peroxisomes via bile acid-acyl transporter (ABCD3) and are modified in β-oxidation reactions involving α-methylacyl-CoA racemase (AMACR), acyl-CoA oxidase 2 (ACOX2), branch chain acyl-CoA oxidase, D-bifunctional protein (HSD17B4), and sterol carrier protein X (SCPX) to cleave propionyl-CoA, forming cholyl-CoA and chenodeoxycholyl-CoA, respectively [26]. Then, bile acid-CoA-amino acid transferase (BAAT) conjugates taurine or glycine to form conjugated bile acids, T/GCA and T/GCDCA (Fig. 2). Detailed illustrations of bile acid synthesis pathway can be found in several recent reviews [4, 27].

Fig. 2.

Bile acid synthesis and the role of FXR in the regulation of bile acid synthesis in hepatocytes and in the enterohepatic circulation of bile acids. Cholesterol is converted to cholic acid (CA) and chenodeoxycholic acid (CDCA) in human liver. The classic pathway is initiated by cholesterol 7α-hydroxylase (CYP7A1), while the alternative pathway is initiated by sterol 27-hydroxylase (CYP27A1) and followed by oxysterol 7α-hydroxylase (CYP7B1). Sterol 12α-hydroxylase (CYP8B1) catalyzes cholic acid (CA) synthesis. In mouse liver, CDCA is converted to α- and β-muricholic acids (αMCA and βMCA) by Cyp2c70 as primary bile acids. Details of bile acid synthesis pathway and enzymes are described in the text. Major regulatory enzymes are shown. CA and CDCA are conjugated to taurine (T) or glycine (G) and secreted into bile. Bile acids are reabsorbed in the ileum. In the colon, gut bacteria bile salt hydrolase (BSH) de-conjugates bile acids and 7α-dehydroxylase (7α-DH) converts CA and CDCA to deoxycholic acid (DCA) and lithocholic acid (LCA), respectively. Bile acids activate FXR, which plays a critical role in regulation of bile acid synthesis. Activation of FXR inhibits CYP7A1 and CYP8B1 through two pathways. In the liver, FXR induces SHP to inhibit CYP7A1 and CYP8B1 gene transactivation by HNF4α and LRH-1 (Pathway 1). In the intestine, FXR induces fibroblast growth factor 19 (FGF19), which activates hepatic FGF receptor 4 (FGFR4)/β-Klotho signaling mainly via ERK1/2 to inhibit CYP7A1 gene transcription (Pathway 2). FXR induces bile salt export pump (BSEP) to efflux bile acids into bile. ATP binding cassette G5 and G8 (ABCG5/G8) effluxes cholesterol and multidrug resistant protein 2/3 (MDR2/3) effluxes phospholipids into bile to form mixed micelles with bile acids. MDR related protein 2/3 (MRP2/3) effluxes bilirubin and glutathione-conjugated bile acids. In the enterocyte of ileum, bile acids are reabsorbed via apical sodium-dependent bile acid transporter (ASBT), which is inhibited by bile acids and FXR. FXR induces intestine bile acid binding protein (IBABP), which binds and transports bile acids across the enterocyte to the sinusoidal membrane to be secreted to portal blood via the organic solute transporter α/β (OSTα/OSTβ) dimer, which are induced by FXR. Bile acids circulated to the liver are taken up by hepatic sodium-dependent taurocholate co-transporting peptide (NTCP), which is inhibited by bile acids via short heterodimer partner (SHP). Organic anion transporting peptides (OATPs) and MRP4 uptake bile acids to hepatocytes independent of sodium. At the sinusoidal membrane, FXR induces OSTα/OSTβ or MRP4 (induced in cholestasis) to efflux bile acids into systemic blood circulation. (+) indicates stimulation, (−) indicates inhibition.

In the alternative bile acid synthesis pathway, CYP27A1 hydroxylates cholesterol to 27-hydroxycholesterol and 3β-hydroxy-5-cholestenoic acid, then oxysterol 7-hydroxylase (CYP7B1) converts the latter to 3β,7α-dihydroxy-5-cholestenoic acid, which can be converted to CDCA via the same enzymatic reactions as in the classic pathway [4]. The alternative pathway is thought to mainly produce CDCA; however, human patients with gene mutations in CYP7A1 have both CA and CDCA [28]. In mice deficient of Cyp7a1, bile acids are synthesized via the alternative pathway to produce more tauromuricholic acids (TMCAs) from TCDCA, yet still have a substantial amount of TCA in bile (~32%) [29]. Thus, the alternative pathway can also synthesize TCA. A recent in vitro study reported that human CYP8B1 could convert CDCA to CA [30]. CYP27A1 and CYP7B1 are widely expressed in most extrahepatic tissues that synthesize oxysterols that can be used for steroid hormone synthesis in the adrenal glands [31]. Cyp7b1 is non-specific and can hydroxylate 24-hydroxycholesterol produced in the brain and 25-hydroxycholesterol produced in the liver to produce bile acids in mice [27]. These investigators proposed that CYP7B1 might play a role in the regulation of oxysterol levels, which can be cytotoxic. Fig. 2 illustrates key enzymes in bile acid biosynthesis pathways in hepatocytes and the role of FXR in the regulation of hepatic bile acid synthesis and enterohepatic circulation of bile acids.

The classic bile acid synthesis pathway is highly regulated, while the alternative pathway is constitutively active. The classic pathway is the major bile acid synthetic pathway in humans, since endogenous production of 27-hydroxycholesterol accounts for ~8.7% of total bile acid synthesis in humans [32]. Cyp7a1 is expressed at very low levels in fetal rat liver and increases after birth, reaching adult levels by time of weaning [33, 34]. The alternative pathway may be the predominant pathway for bile acid synthesis in fetal liver. The classic and alternative pathways contribute about equally to produce bile acids in mice. In mice and rats, CDCA, a hydrophobic bile acid, can be 7β-isomerized to UDCA (3α,7β), a highly soluble bile acid. Recently Cyp2c70 has been identified as the 6β-hydroxylase that converts CDCA and UDCA to α-MCA (3α, 6β, 7α) and β-MCA (3α,6β,7β) in mouse liver [35]. In mice, but not humans, Cyp2a12 can re-hydroxylate DCA and LCA at the 7α-position to form CA and CDCA, respectively [36]. Thus, mice predominantly have TCA (~50%) and TMCAs (~50%) in a highly hydrophilic bile acid pool, whereas in humans, the bile acid pool is highly hydrophobic and consists of CA, CDCA and DCA.

Amino acid-conjugated bile acids form sodium salts at physiological conditions, thus reducing toxicity and increasing solubility for active transport into bile via bile salt export pump (BSEP or ABCB11) (Fig. 2). The ATP binding cassette B5 (ABCB5) and ABCB8 heterodimer effluxes cholesterol and multi-drug resistant protein 2/3 (MDR2/3 or ABCB4) effluxes phospholipids into bile to form mixed micelles with bile acids, which are stored in the gallbladder. MDR resistant associated protein 2/3 (MRP2/3 or ABCC2) effluxes bilirubin and glutathione conjugated-bile acids. After meal intake, bile acids are released into the gastrointestinal tract. In the ileum, bile acids are reabsorbed together with lipid-soluble fats, sterols, and drugs, and are transported via portal blood to the liver for metabolism and detoxification and these compounds are then distributed to other tissues. The enterohepatic circulation of bile acids from the intestine to the liver inhibits bile acid synthesis. This is a classic feedback mechanism, in that the end products of a metabolic pathway inhibit the first and rate-limiting enzyme, ultimately controlling the amount of bile acids in circulation and protecting against cytotoxicity (Fig. 2). However, the molecular mechanism of bile acid feedback regulation of bile acid synthesis is complex and not completely understood (sections 5 , 6 and 7 below).

5. The role of FXR in regulation of bile acid homeostasis

FXR plays a central role in the regulation of the enterohepatic circulation of bile acids to maintain bile acid homeostasis (Fig. 2). In the ileum, apical sodium-dependent bile acid transporter (ASBT) reabsorbs bile acids from the intestinal lumen into enterocytes. Bile acids inhibit ASBT expression likely via FXR/SHP or FXR/Fgf15. Bile acids activate intestinal FXR to induce ileum bile acid binding protein (IBABP), which binds and transports bile acids from the basolateral membrane to the sinusoidal membrane where FXR induces organic solute transporter α (Ostα/Ostβ) heterodimer to efflux bile acids into portal blood circulation back to the liver [37]. Sodium-taurocholate co-transport peptide (NTCP) takes up bile acids through the sinusoidal membrane of hepatocytes, and SHP inhibits RARα/RXRα transactivation of NTCP [38]. FXR induction of BSEP and inhibition of NTCP maintain very low bile acid levels in hepatocytes to protect against cholestatic liver injury. Organic anion transporting peptides (OATPs) take up bile acids into hepatocytes independent of sodium. FXR also induces BACS and BAAT to conjugate taurine or glycine to bile acids [39]. In the canalicular membrane, FXR induces BSEP to efflux conjugated bile acids into bile [40]. BSEP is an ATP-dependent bile acid efflux transporter that generates bile flow and is rate-limiting for the enterohepatic circulation of bile acids. FXR also induces MDR2/3, which effluxes phospholipids to facilitate bile acid secretion into bile. During cholestasis, FXR induces MRP4 as an adaptive mechanism to efflux bile acids into blood circulation, and FXR also induces the OSTα/OSTβ heterodimer to efflux bile acids from hepatocytes. The enterohepatic circulation of bile acids is highly efficient in recycling 95% of bile acids in each cycle. Bile acids lost in feces and urine are replenished by de novo synthesis.

6. Liver FXR/SHP-dependent mechanism

Chiang and Stroup first identified a putative bile acid response element in the Cyp7a1 gene promoter [41]. Cyp7a1 mRNA and protein levels in liver, and Cyp7a1 catalytic activity, are inhibited by bile acid feeding and induced by bile acid sequestrants and bile fistulae (biliary diversion), which interrupt the enterohepatic circulation of bile acids to reduce bile acid pool size. This suggests that Cyp7a1 is mainly regulated by gene transcription. DNase I foot printing assay of the rat Cyp7a1 gene promoter identified a cluster of overlapping hormone response element-like sequences (AGGTCA), which are known to bind the nuclear receptors HNF4α and LRH-1. These sequence motifs are responsive to cholic acid feeding and deoxycholic acid-induced inhibition of Cyp7a1 gene transcription. These investigators named these sequence motifs the bile acid responsive elements I and II (BAREI and BAREII) on the Cyp7a1 gene proximal promoter, which are highly conserved in rats, mice, and humans [41, 42]. They proposed that transcription factors bound to the BAREs were required for basal transcription of the Cyp7a1 gene. Further, they hypothesized that bile acids might bind to a bile acid receptor, which forms a complex with a bile acid responsive protein to block its binding to the BARE, thus inhibiting Cyp7a1 gene transcription [41]. These investigators demonstrated that FXR inhibited Cyp7a1 gene transcription without binding to the BAREs but mutagenizing these BARE sequences abolished the inhibitory effect of FXR on Cyp7a1 gene transcription [43].

In 2000, three laboratories reported a cascade mechanism in which FXR induction of the negative receptor SHP inhibited HNF4α and LRH-1 transactivation of the Cyp7a1 gene (Pathway 1, Fig 2) [44–46]. SHP negatively regulates the genes involved not only in bile acid synthesis, but also adipogenesis, conjugation and transport of bile acids, and lipogenesis and gluconeogenesis in the liver [47–52]. Rat, mouse and human CYP8B1 gene promoters have multiple HNF4α and LRH-1 binding sites that serve as BAREs and the FXR/SHP pathway also mediates bile acid inhibition of Cyp8b1 gene transcription [53–56]. However, differential regulation of the mouse Cyp7a1 and Cyp8b1 genes by the FXR/SHP pathway has been reported, and this pathway may be more important for inhibition of Cyp8b1 compared to Cyp7a1 [57].

To confirm the role of FXR in the regulation of bile acid synthesis, FXR deficient mice were generated [39]. FXR deficient mice had increased serum bile acid concentration and reduced bile acid pool and fecal bile acid excretion due to decreased canicular bile acid transport BSEP. These mice were not responsive to bile acid inhibition of Cyp7a1 and had increased serum and hepatic cholesterol and triglycerides, and a proatherogenic serum lipoprotein profile. This study demonstrated that FXR is a bile acid sensor that plays a critical role in bile acid and lipid homeostasis. As expected, SHP deficient mice had increased bile acid pool size and FXR agonism did not have a regulatory effect [58]. However, SHP deficient mice fed bile acids or GW4064, an FXR agonist, still had reduced Cyp7a1 expression [59], while the bile acid sequestrant cholestyramine reduced the bile acid pool but increased Cyp7a1 and Cyp8b1 gene expression. This indicated that SHP-independent mechanisms exist to inhibit Cyp7a1 gene expression and bile acid synthesis [60]. Several studies showed that PXR and CAR inhibited Cyp7a1 gene transcription by competing for HNF4α binding to BAREs on the Cyp7a1 gene promoter [61–64]. Further, bile acids have been shown to activate the c-Jun-N-terminal kinase (JNK) pathway to inhibit bile acid synthesis independent of SHP [57, 59, 65].

7. The intestinal FXR /liver FGFR4 mechanism

In 1995, Pandak et al. reported that intravenous infusion of taurocholate in biliary-diverted rats failed to inhibit Cyp7a1 specific activity, mRNA levels or transcriptional activity, whereas intraduodenal infusion of taurocholate repressed Cyp7a1 activity and gene expression [66]. This study suggested that an intestinal factor released or absorbed in the presence of bile acids in the intestinal lumen might play a role in the regulation of bile acid synthesis. Attempts to identify this intestinal factor had been unsuccessful until ten years later, when Inagaki et al. reported that intestinal FXR induced fibroblast growth factor 15 (FGF15), which functions as an enterohepatic signal to regulate bile acid synthesis in mouse liver [67].

In vivo imaging of FXR activity by FXR reporter assays demonstrate that the ileum is the primary bile acid signaling tissue [68]. Fgf15 activates membrane FGF receptor 4 (FGFR4)/β-Klotho complex to inhibit Cyp7a1 and Cyp8b1 gene expression [67] (Pathway 2, Fig. 2). This study suggested that Fgf15/FGFR5 signaling was mediated by cJun and is partially dependent on SHP to inhibit Cyp7a1 expression in mice. However, mouse Fgf15 cannot be detected in serum using commercial antibodies. A highly sensitive targeted mass spectrometry method has been developed to detect Fgf15 in mouse plasma using stable isotope standards and anti-peptide antibodies [69]. The differential role of liver FXR and intestinal FXR for the inhibition of Cyp7a1 and Cyp8b1 in mice has been reported [70]. These investigators showed that GW4064, an FXR agonist, inhibited Cyp7a1 in liver specific Fxr deficient mice but not in intestinal Fxr deficient mice, demonstrating that activation of intestinal FXR was required for inhibition of Cyp7a1 gene transcription. Furthermore, this study shows that the FXR/SHP pathway contributed more to inhibit Cyp8b1 gene transcription but had a minor role in the suppression of Cyp7a1 gene transcription. In vivo transfection of FGF19, the human orthologue of Fgf15, in mice strongly activated extracellular signal-regulated kinase (ERK1/2) but had much less effect on the JNK pathway independent of SHP, indicating FGF19 primarily activates ERK1/2 signaling to inhibit Cyp7a1 gene transcription [57].

Human FGF19 and mouse Fgf15 share about 50% amino acid sequence identity, but FGF19 is more stable than FGF15 and can be assayed using commercial antibodies [71]. FGF19 is highly expressed in the human gallbladder and in the common bile duct, and exhibits lower expression in the ileum [72]. FGF19 has been shown to strongly inhibit CYP7A1 mRNA expression in human primary hepatocytes [73]. This study showed that inhibition of ERK1/2, but not JNK or p38 kinase, blocked FGF19 inhibition of CYP7A1 mRNA expression. It was reported that bile acids activate ERK1/2 signaling to inhibit CYP7A1 gene transcription [74]. Furthermore, small interference RNA (siRNA) knockdown of SHP did not affect FGF19 inhibition of CYP7A1 in human hepatocytes or CDCA-stimulated tyrosine phosphorylation of FGFR4 in hepatocytes. Meanwhile, FGF19 antibody treatment and siRNA against FGFR4 abrogated FXR agonist-mediated inhibition of CYP7A1. A liver cytoplasmic non-receptor tyrosine kinase, Shp2, has been shown to mediate FGF19/FGFR4 signaling in the suppression of bile acid synthesis [75]. Furthermore, FXR/FGF19/FGFR4 signaling may activate a non-receptor tyrosine kinase, Src, which phosphorylates FXR for nuclear localization to inhibit bile acid synthesis during primary biliary cholangitis (PBC) [76]. All these studies suggest that bile acids activate hepatic FXR to induce FGF19 by autocrine/paracrine mechanisms to inhibit bile acid synthesis during cholestasis. This gut-to-liver signaling pathway may be the major mechanism responsible for bile acid feedback inhibition of bile acid synthesis. The liver FXR/SHP pathway may be activated as an adaptive response to reduce bile acid synthesis during cholestasis.

8. Cell signaling mechanisms

An overlooked mechanism for bile acid inhibition of bile acid synthesis is the cell signaling mechanism, which are independent of FXR, SHP and FGF15/19. The cell signaling mechanism was first reported by Roger Davis’s laboratory. They found that bile acids induced cytokine secretion from hepatic Kupffer cells and subsequently inhibited bile acid synthesis in hepatocytes [77]. They suggested that bile acids induced the inflammatory cytokines IL-1β and TNFα, which cross the sinusoidal membrane to activate Toll-like receptor and protein kinase C/JNK pathway to inhibit CYP7A1 expression and bile acid synthesis [78]. It is well documented that CDCA and proinflammatory cytokines strongly inhibit CYP7A1 and CYP7B1 gene transcription via activation of cJun/JNK to phosphorylate and inhibit HNF4α transactivation of CYP7A1 and CYP8B1 genes in human hepatocytes [65, 79]. Bile acids can activate PKCα and PKCα to inhibit CYP7A1 gene expression via activation of cJun/JNK in human hepatocytes [80, 81]. Bile acids activate phorbol ester-activated PKC to inhibit Cyp7a1 gene expression via the BAREI in the Cyp7a1 promoter [82]. The PKC activator phorbol 12-myristate 13-acetate can reduce Cyp7a1 mRNA expression through a phorbol ester response element located in the BAREII [83]. Hepatic growth factor (HGF) released from stellate cells activates c-Met to stimulate cJun/JNK and ERK1/2 signaling pathways and rapidly inhibits Cyp7a1 gene transcription [84]. These mitogen-activated kinase pathways (MAPK) including cJun N-terminal kinase (JNK) and extracellular regulated kinase (ERK1/2) may be rapidly activated during cholestasis to inhibit Cyp7a1 and reduce bile acid synthesis as an adaptive response to protect the liver from injury [74]. In human-like hydrophobic bile acid mouse models, Cyp2c70 knockout and Cyp2c70/Cyp2a12 double knockout mice, the bile acid pools were reduced but the FXR/SHP and Fgf15 pathways were not activated [36]. It was suggested that PXR and the cytokine/cJun/MAPK pathways might be the predominant mechanisms over the FXR/SHP and FXR/Fgf15 pathways to control bile acid synthesis within a hydrophobic bile acid pool. These cells signaling pathways can be rapidly activated without inducing transcription factors to inhibit bile acid synthesis in acute liver injury such as during cholestasis.

9. FXR and bile acid biotransformation by gut microbiota

The primary bile acids are mostly reabsorbed in the terminal ileum, while some primary bile acids are converted to secondary bile acids by the gut microbiome. In the colon, bacterial bile salt hydrolases (BSH) deconjugate taurine- and glycine-conjugated bile acids, then bacterial 7α-dehydroxylases remove the 7α-HO group from CA and CDCA to form deoxycholic acid (DCA) and lithocholic acid (LCA), respectively (Fig. 2). In humans, gut bacteria 7α- and 7β-hydroxysteroid dehydrogenase (7α/β-HSDH) can convert CDCA to UDCA via multi-step reactions. UDCA is a secondary bile acid synthesized in the intestine. In mice, β-MCA can be converted to its 6α-isomer ω-MCA (3α,6α,7β), the most soluble bile acid, for excretion into feces. Most LCA is sulfonated to increase its solubility for fecal or renal excretion. Some DCA in the colon is secreted into portal circulation. Thus, most bile acids in mice are TCA and TMCAs, with very little TLCA and TDCA. Bile acids are glycine- or taurine-conjugated in a ratio of ~ 3:1 in humans. The ratio of CA, CDCA and DCA is about 40:40:20 in the bile acid pool, making it highly hydrophobic. In mice, most bile acids (~95%) are taurine-conjugated, and the ratio of CA to MCA (α and β) is about 50:50 in bile acid pool, which is highly hydrophilic.

The gut microbiota plays a critical role in bile acid synthesis and in regulating bile acid composition and circulating pool size [4, 85]. Bile acids, especially DCA, act as bactericides to control gut bacteria overgrowth. The gut microbiota also plays important roles in intestine barrier function, pathogen defense and immunity. The gut microbiota converts polysaccharides to short-chain fatty acids (acetate, butyrate, and propionate) and utilizes amino acids to harvest energy. The gut microbiota also controls the release of gut hormones involved in satiety, such as GLP-1 and peptide YY, and may release lipopolysaccharides to cause low-grade inflammation. Bacterial BSHs are considered gatekeepers of bile acid metabolism and host-microbiome crosstalk in the gastrointestinal tract [86]. BSH plays a critical role in determining the ratio of conjugated to unconjugated bile acids in the circulating bile acid pool, which is altered in diabetes and metabolic liver disease and is linked to dysbiosis. BSH activity is limited to the anaerobic phyla Firmicutes (Gram positive, genera Clostridium, Enterococcus, Lactobacillus, Bifidobacterium, and Listeria) and Bacteroidetes (Gram negative, genus Bacteroidia). Primary and secondary bile acids can be converted to 3α, 7α or 12α-oxo-bile acids by 3α,7α, or 12α HSDH. Then 3β,7β or 12β-HSDH convert oxo-bile acids to β-hydroxylated bile acids, such as UDCA from CDCA [87]. BSH activity determines the bile acid composition in the bile acid pool, which in turn shapes the gut microbiome composition. CA administration to rats increases the Firmicutes to Bacteroidetes ratio and Clostridia BSH to convert CA to DCA and inducing gut microbiota changes, similar to those induced by high fat diets [88]. High saturated-fat diets induce TCA and expend pathobiont Bilophila wadsworthia to increase proinflammatory cytokines and promote inflammatory bowel disease (IBD) in IL10 deficient mice [89].

10. The role of FXR in gut microbiota

The secondary bile acids synthesized by gut bacteria control gut bacterial overgrowth and protect intestinal barrier function via FXR signaling [90]. A study of germ-free and conventionally raised mice showed that the gut microbiota altered bile acid composition and reduced bile acid pool size to promote weight gain and hepatic steatosis in germ-free mice and is dependent on FXR [85]. In germ-free mice, TMCAs are increased to antagonize intestinal FXR activity and increase hepatic bile acid synthesis and total bile acid pool size. FXR plays a critical role in shaping the gut microbiota and controlling hepatic lipid metabolism [91]. This study showed that oral administration of Gly-MCA, an intestinal FXR antagonist, altered the gut bacteria community by reducing the ratio of Firmicutes to Bacteroidetes and short-chain fatty acids, and improved metabolic functions in high fat-diet fed wild type mice but not in FXR deficient mice. In a phase 1 study, administration of obeticholic acid (OCA), a synthetic bile acid and potent FXR agonist, to healthy subjects increased Gram-positive bacteria in the small intestine, consistent with increased Firmicutes in mice treated with OCA [92].

11. FXR as a metabolic regulator in the gut-to-liver axis

The metabolic functions of FXR in glucose and lipid metabolism have been reviewed extensively [27, 93, 94]; the following is a brief review of the metabolic functions of FXR in the liver and intestine (Section 12 and 13). Beside its role in the regulation of bile acid metabolism, FXR has been shown to regulate lipoprotein, glucose, and triglyceride metabolism in the liver (Fig. 3).

Fig. 3.

The role of FXR signaling in metabolic regulation. In hepatocytes, activation of FXR inhibits bile acid synthesis and reduces bile acid pool size; reduces ApoA1 and HDL and reverse cholesterol transport by scavenger receptor b1 (Srb1) and apoptosis; reduces fatty acid uptake via CD36, lipogenesis and VLDL secretion and increases triglyceride clearance; reduces postprandial glucose by reducing glycolysis and stimulate fasting gluconeogenesis and glycogen synthesis. In Kupffer cells and macrophages, activation of FXR reduces liver inflammation and nonalcoholic steatohepatitis (NASH) fibrosis and primary biliary cholangitis (PBC). In adipose tissue, activation of FXR stimulates adipose tissue browning and energy metabolism. In the intestine, activation of FXR reduces NFκB and inflammatory cytokines, protects barrier function, induces FGF19 and ceramides productions, and glucagon-like peptide-1 (GLP-1) secretion. GLP-1 stimulates insulin secretion and increases insulin sensitivity in pancreatic β cells, where activation of FXR also stimulates insulin synthesis and improves insulin sensitivity.

4.1. Metabolic functions of FXR in the liver

Studies in tissue specific FXR deficient mice demonstrated the differential functions of liver and intestine FXR in liver metabolism, though many metabolic functions of FXR overlap in the liver and intestine. Bile acids also activate G protein-coupled receptor-1 (Gpbar-1), also known as Takeda G protein-coupled receptor 5 (TGR5) (Section 14). TGR5 and FXR are co-expressed in the enteroendocrine L cells, which secrete GLP-1, and FXR and TGR5 crosstalk to regulate metabolism and inflammation in the gut-to-liver axis (Section 15).

12. Metabolic functions of FXR in the liver

Hepatic FXR expression is regulated by glucose and insulin [95]. FXR deficient mice have increased glycolytic and lipogenic gene expression profiles and reduced gluconeogenesis upon carbohydrate feeding. These mice are glucose intolerant and insulin insensitive [96]. Activation of FXR by the agonist GW4064 or overexpression of FXR lowered serum glucose and improved insulin sensitivity in obese and diabetic db/db mice [97]. Activation of FXR reduces postprandial glucose levels by inhibiting glycolysis. FXR is expressed in pancreatic β cells and the expression of FXR is regulated by glucose-induced insulin synthesis which increases insulin sensitivity [98]. In the postprandial state, increased bile acids may activate FXR to regulate glucose homeostasis [99–101].

FXR has been shown to inhibit expression of apolipoprotein A1 (ApoA1), suggesting direct inhibition of high-density lipoprotein (HDL) synthesis [102]. Activation of FXR reduces serum HDL via FGFR4 signaling [103]. Importantly, activation of FXR promotes macrophage-to-feces reverse cholesterol transport by inducing the hepatic HDL receptor scavenger receptor-B1 (SR-B1) [104]. Activation of FXR induces SHP to inhibit hepatic fatty acid transporter CD36, lipogenesis enzymes (including fatty acid synthase, acetyl-CoA carboxylase, and stearyl-CoA desaturease-1), and microsomal triglyceride transport protein involved in very low-density lipoprotein (VLDL) production [47]. FXR stimulates lipoprotein lipase and ApoCII involved in triglyceride hydrolysis and clearance. In addition, bile acids and FXR also regulate apoptosis, which plays a role in lipid metabolism [105–108]. Activation of FXR has been shown to reduce cytokines and inflammation to preserve the intestinal barrier in inflammatory bowel disease [109].

Autophagy is a cellular response to nutrient deprivation and/or starvation to stimulate lipolysis and maintain energy homeostasis. Singh et al. were the first to report that autophagy regulates lipid metabolism and plays a role in metabolic diseases [110, 111]. Defective autophagy in obesity has been shown to promote ER stress and insulin resistance [112]. FXR has been implicated in autophagy; activation of FXR strongly represses the induction of autophagy in the fasting state and in FXR deficient mice this response is lost [113]. In the fed state, FXR and cAMP responsive element binding protein (CREB) coordinately regulate hepatic autophagy; pharmacological activation of FXR represses autophagy gene expression and autophagy even in the fasted state [114]. In FXR deficient mice, feeding-induced inhibition of autophagy was attenuated. Chromatin immunoprecipitation assays showed that FXR and CREB share several binding sites on the promoter of many autophagy-promoting genes. CREB upregulates these autophagy genes while FXR trans-represses CREB-regulated genes. CREB and FXR may compete for binding sites at the promoters of autophagy genes in response to nutrient availability to regulate lipid metabolism during fed vs. fasted stats. Bile acid signaling via Cyp7a1-AKT-mTOR axis induced hepatic autophagy in mice [115]. Increased intracellular cholesterol impaired lysosomal function and autolysosome accumulation in hepatocytes. Cholestyramine feeding induced Cyp7a1 and increased bile acid synthesis to reduce intracellular cholesterol levels and restore hepatic autophagy, ultimately improving metabolism in high fat diet-fed mice. Thus, bile acids and FXR signaling regulate autophagy via hepatic lipid metabolism.

13. FXR functions in the intestine

FXR is highly expressed in the enterocytes in the ileum, where bile acids are actively reabsorbed. Activation of intestinal FXR induces FGF19/15, which inhibits bile acid synthesis via FGF19/15/FGFR4 signaling as described in Section 7. Paradoxical roles of intestinal FXR in metabolism and disease have been reported. The Gonzalez laboratory published a series of studies to demonstrate that deficiency of intestinal FXR or antagonizing intestinal FXR protects against diet-induced obesity, insulin resistance and non-alcoholic fatty liver disease [116–118]. They found that FXR induces ceramide synthesis in the intestine. Ceramides attenuate citrate synthase activity through the induction of endoplasmic reticulum (ER) stress. This stimulates mitochondrial acetyl-CoA and pyruvate carboxylase activities to increase hepatic gluconeogenesis and caffeic acid phenethyl ester, a dietary supplement and an inhibitor of BSH activity, antagonizes intestinal FXR activity to increase TMCA and reduce intestinal ceramide synthesis, which is stimulated by FXR [118, 119] (Fig. 3). An antioxidant, tempol reduces obesity by decreasing Lactobacillus and BSH activity to increase TMCA and antagonize intestinal FXR [120]. In contrast, fexaramine, an intestine-restricted FXR agonist, has been shown to promote adipose tissue browning and reduce weight in high fat diet-fed mice [121]. Interestingly, fexaramine treatment increased serum LCA. Fexaramine alters the gut microbiome by increasing gut Acetatifactor and Bacteroides acidifaciens, both of which have 7α-and 7β-dehydroxylase activities that convert CDCA and UDCA to LCA to stimulate TGR5 signaling (Section 14). This ultimately stimulates GLP-1 secretion, promotes adipose tissue browning and increases insulin sensitivity in diet-induced and genetic obese mice [122].

Furthermore, deficiency of hepatic autophagy has been shown to cause liver injury via FXR/Fgf15-FGFR4 signaling to remodel the gut microbiome [123]. In this study, autophagy gene 5 (Atg5) deficiency in mice caused liver injury and altered the intestinal bile acid composition by increasing gut bile acid-metabolizing bacteria, unconjugated bile acids and Fgf15. FXR/Fgf15 signaling plays a critical role in autophagy as an adaptive response to protect against liver injury.

14. Bile acid-activated G protein-coupled receptors

In 2002, Maruyama et al. identified G protein-coupled bile acid receptor-1 (GPBAR-1) or Takeda G protein-coupled receptor 5 (TGR5) [124]. TGR5 is a Gαs GPCR, which is widely expressed in the gastrointestinal tract. TGR5 is also expressed in monocytes and macrophages, in the epithelial cells of the intestine and gallbladder, and in liver sinusoidal endothelial cells and Kupffer cells, but not in hepatocytes [125–128]. Secondary bile acids are more potent TGR5 activators than primary bile acids, in the order of LCA>DCA>CDCA>CA. A recent study reports UDCA and TUDCA are potent TGR5 activators with an EC₅₀ of 14 μM and 20 μM, respectively, which are more efficacious than TLCA [129]. TGR5 has several important functions in different organs and tissues (Fig. 4). Activation of TGR5 induces synthesis of cAMP by adenyl cyclase, which induces type 2 deiodinase to stimulate thyroxine synthesis and mitochondrial energy production in brown adipose tissue [130]. In the ileum and colon, activation of TGR5 reduces inflammation via inhibition of TNFα and NF-kB activity and protects against colitis, inflammatory bowel disease and colon cancer [131–133]. In intestinal L cells, TGR5 stimulates insulin sensitivity via GLP-1-induced insulin secretion, and intestinal TGR5 agonism improves hepatic steatosis and insulin sensitivity in diet-induced obese (DIO) mice [134]. In cholangiocytes, TGR5 is involved in bile acid-dependent proliferation, [135] while in Kupffer cells TGR5 mediates bile acid-induced cytokine production [136]. Reduced expression of TGR5 in biliary epithelial cells has been identified in PSC patients and in liver and extrahepatic bile ducts of Mdr2 knockout primary sclerosing cholangitis (PSC) mice [137]. In muscle tissue, TGR5 stimulates nitric oxide synthesis, smooth muscle relaxation, skeletal muscle atrophy, and decreases inflammation and atherosclerosis [138, 139]. Activation of TGR5 inhibits atherosclerosis by reducing macrophage inflammation and lipid loading [140]. In the brain, TGR5 signaling decreases satiety, inflammation, and oxidative stress and protects against obesity [141]. In the gallbladder, TGR5 stimulates refilling and exerts control over the bile acid composition. In monocytes and macrophages, activation of TGR5 reduces NLRP3 inflammasome inflammation and stimulates the M1 to M2 switch [142, 143]. A study of Atg7DeltaCD11c mice showed reduced body weight and fat mass and B. acidifaciens was significantly increased in feces compared to Atg7f/f mice [144]. B. acidifaciens has both 7α- and 7β-dehydroxylase activities, which may activate TGR5 signaling and GLP-1 secretion to stimulate energy expenditure and reduce weight.

Fig. 4.

The role of TGR5 signaling in different tissues and organs. In adipose tissue, activation of TGR5 induces cAMP to induce type 2 deiodinase (DIO2) to synthesis T3 and stimulates energy metabolism, and adipose tissue beiging and browning. In enteroendocrine L cells in ileum and colon, activation of TGR5 stimulates GLP-1 secretion to improve insulin sensitivity. In cholangiocyte, activation of TGR5 and S1PR2 stimulates proliferation and inflammation and primary sclerosing cholangitis (PSC), PBC and cholangiocarcinoma (CCA). In monocyte and macrophage, activation of TGR5 reduces NF-kB and Nrlp3 inflammasome, inflammatory cytokine production and M1 to M2 switch. In muscle, activation of TGR5 induces nitric oxide and smooth muscle relaxation and reduces inflammation and atherosclerosis. In the brain, TGR5 reduce satiety, inflammation, and oxidative stress. In the gallbladder, activation of TGR5 stimulates gallbladder refiling. In the intestine, activation of TGR5 reduces NF-kB inflammation and prevents colitis, IBD and colon cancer.

Another GPCR, sphingosine-1-phosphate receptor 2 (S1PR2) has been shown to be activated by conjugated bile acids, TCA and DCA. S1PR2 is the Gαi protein-coupled receptor in rodent hepatocytes [145]. S1PR2 is highly expressed in cholangiocytes and activation of S1PR2 can cause cholangiocyte proliferation and cholangiocarcinoma [146, 147]. S1PR2 activates sphingosine kinase 2 (Sphk2), a major sphingosine kinase expressed in the liver, and the AKT/insulin and ERK 1/2 signaling pathways to regulate hepatic lipid metabolism [145, 148]. It has been reported that ER stress via activation factor 4 activates Sphk2 and overexpression of Sphk2 ameliorates hepatic steatosis and insulin resistance in HFD-fed mice [149]. Bile acid activation of S1PR2 signaling has been shown to promote neuroinflammation in hepatic encephalopathy in mice [150].

15. FXR and TGR5 signaling crosstalk

Both FXR and TGR5 are expressed in several tissues in the gastrointestinal tract and are activated by bile acids. FXR is expressed in the cytosol and is transported to the nucleus by ligand binding while TGR5 is expressed at the cell surface membrane. FXR directly binds to DNA to regulate target gene transcription whereas TGR5 signals through cAMP, protein kinase C and CREB or phosphorylates and activates other transcription factors to indirectly regulate target gene transcription. FXR and TGR5 signaling have overlapping functions in gallbladder, intestine, and adipose tissues.

Both FXR and TGR5 signaling have been implicated in controlling gallbladder filling. After a meal, cholecystokinin released from the mucosal epithelial cells of the duodenum stimulates gallbladder contraction to release bile acids to the intestine for nutrient absorption. In the late postprandial state, FXR induces Fgf15, which activates the FGFR4/β-Klotho complex expressed in the mucosal epithelial cells of the gallbladder to stimulate gallbladder filling. Consistently, in FXR knockout mice Fgf15 is not induced and the gallbladder is empty [151]. TGR5 is highly expressed in human gallbladder epithelium [128]. Activation of TGR5 by LCA or the specific agonist INT-777 stimulates gallbladder filling by cAMP-induced stimulation of smooth muscle relaxation in wild type mice but not TGR5 deficient mice [152]. This study also showed that TGR5-stimulated gallbladder filling is independent of Fgf15. It is possible that in the gallbladder FXR induces TGR5, which produces cAMP to stimulate gallbladder filling.

Both TGR5 and FXR are involved in liver regeneration because deficiency of either FXR or TGR5 could reduce the rate of liver regeneration after partial hepatectomy [153, 154]. Double deletion of FXR and TGR5 in LDL receptor knockout mice has been shown to exacerbate atherosclerosis more than single deletion of FXR or TGR5 in LDL receptor knockout mice [155]. Double deficiency of both FXR and TGR5 (DKO) increased bile acid pool size with increased TCA, decreased TMCAs, and increased Cyp7a1 and Cyp8b1 expression as expected [156]. Cholestyramine feeding further increased Cyp7a1 mRNA, protein expression and bile acid pool size, while CA feeding unexpectedly also induced Cyp7a1 expression, indicating impaired bile acid metabolism and regulation in these mice. Transcriptome analysis revealed that fibrosis gene expression levels were increased in livers of chow fed DKO mice compared to wild type mice, and the top upregulated genes were involved in steroid synthesis, liver cirrhosis and connective tissue disease. Interestingly, Western diet feeding increased oxidative stress and markers of liver fibrosis but not steatosis in DKO mice, and liver fibrosis was exacerbated in DKO mice.

FXR and TGR5 are co-expressed in enteroendocrine L cells in the ileum and colon. In FXR deficient and TGR5 deficient mice glucose-induced GLP-1 secretion was abolished, indicating both FXR and TGR5 are involved in GLP-1 secretion and insulin sensitivity [157]. Interestingly, activation of FXR by INT-747 (obeticholic acid, OCA) induces TGR5 mRNA and protein expression in ileum and L cells [157]. In FXR deficient mice, TGR5 expression is reduced, and GLP-1 secretion is impaired. A functional FXR response element was identified in the human TGR5 gene promoter. The FXR and TGR5 dual agonist INT-767 was more effective than the FXR selective agonist OCA or TGR5 selective agonist INT-777 in increasing intracellular Ca²⁺ and cAMP levels, which stimulates GLP-1 secretion and improves hepatic glucose metabolism, lipid metabolism, and insulin sensitivity in diet-induced obese and diabetic mice [157]. Another study reported that the intestine-restricted FXR agonist fexaramine markedly increased TLCA in serum, liver, ileum, and colon, and increased Fgf15 and GLP-1 secretion, improved insulin tolerance, glucose tolerance, white adipose tissue browning, and energy metabolism in db/db mice [122]. Analysis of 16S ribosomal RNA of the gut microbiota identified fexaramine-induced and LCA-producing bacteria Acetatifactor and Bacteroides; these two bacteria have both 7α- and 7β-dehydroxylase activities that convert both CDCA and UDCA to LCA. Antibiotic treatment completely reversed the metabolic beneficial effects of fexaramine. Thus, activation of intestinal FXR reshapes the gut microbiota to activate TGR5 to improve metabolism in mice.

Both FXR and TGR5 are implicated in improving glycemic control after vertical sleeve gastrectomy (VSG) in mice [158–160]. Metabolic gastric surgery (including adjustable gastric band, VSG, and Roux-en-Y gastric bypass) rapidly improves insulin resistance and glycemic control for overly obese and diabetic patients. After surgery, serum bile acids (especially conjugated bile acids), FGF19, and GLP-1 are increased and correlated to diabetic remittance [161–164]. However, the underlying mechanism that improves insulin resistance after metabolic gastric surgery is not known. A recent study reported that VSG increased cholic acid 7-sulfate (CA-7S) in liver and cecal content in mice and human patients [165]. This study showed that CA-7S activated TGR5 in L cells to stimulate GLP-1 secretion and improve insulin resistance and glucose tolerance [165]. VSG increased CA-7S but decreased Clostridia and LCA production, and somehow increased portal blood LCA to activate liver VDR, which induced sulfotransferase 2A1 to convert CA to CA-7S [166]. This study also showed that CA-7S is a potent TGR5 agonist. Further study is needed to understand how VSG reduces LCA production in the gut but increases LCA in portal blood to the liver. VDR is not expressed in hepatocytes, but in stellate cells. VDR deficient mice should be used to verify the role of VDR in CA-7S synthesis in the liver. Additionally, the downstream molecular targets of metabolic surgery need to be further identified to develop non-invasive therapies for obesity, diabetes, and metabolic liver diseases [167].

16. Bile acids and FXR-based drug therapies

The discovery of FXR as a key metabolic regulator has received the attention of drug companies to develop bile acids and their derivatives as therapeutic drugs for metabolic diseases, including cholestatic liver diseases, gallstone disease, Type 2 diabetes, obesity and NAFLD. Maintaining bile acids homeostasis is essential for protecting against liver-related metabolic diseases. Bile acid synthesis pathways are remarkably similar between mice and humans, but bile acid composition, pool size, and the gut microbiome are very different. Studies in animal models have contributed enormously to the development of bile acids and derivatives targeting FXR signaling for metabolic disease therapy [168]. However, mouse and human FXR differ in their efficacies in bile acid binding and activation [169]. In addition, lipoprotein metabolism is very different between mice and humans; HDL is the predominant lipoprotein in mouse serum, whereas serum LDL-cholesterol is much higher than HDL-cholesterol in humans.

Increased cholesterol, reduced bile acids, and impaired gallbladder contractility contribute to cholesterol gallstone disease. In inborne errors of bile acid synthesis, deficiency of bile acid synthesis enzymes due to genetic mutations cause reduced bile acid synthesis, accumulation of toxic bile acid intermediates, and contributes to cholestatic liver diseases and injury. Bile acids have been used for decades as a replacement therapy for treating digestive diseases and for patients with inborne errors of bile acid synthesis [27, 170].

Intrahepatic cholestasis arises from genetic mutations of bile acid transporter genes and autoimmune destruction of small bile ductules, while extrahepatic cholestasis results from obstruction by stones or tumors of the bile duct [171]. Primary biliary cholangitis (PBC) is an acquired chronic cholestasis associated with autoimmune destruction of the small bile ducts, portal infiltration and fibrosis. Primary sclerosing cholangitis (PSC) is caused by injury and fibrosis of both intra- and extra-hepatic bile ducts and obstruction of bile blow [170]. Mutations in the FXR gene have been linked to progressive familial intrahepatic cholestasis type 2 (PFIC2) [172, 173]. These patients have undetectable hepatic expression of BSEP (a target of FXR), normal or near normal serum γ-glutamyl transferase activity, and rapid progression to end-stage liver disease.

UDCA (Actigall, Allergan Pharma) is an approve drug for cholesterol gallstone dissolution. UDCA increases the hydrophilicity of the bile acid pool to decrease toxicity and liver injury. UDCA also significantly improves liver function tests and delays the need for liver transplantation for intrahepatic cholestasis of pregnancy and PFIC3 (MDR3 mutation) patients [174]. Ursodiol (Dr. Falk Pharma) has been approved for the treatment of PBC. UDCA stimulates bile flow and promotes biliary HCO₃⁻ secretion [175]. UDCA also exhibits anti-inflammatory and anti-apoptosis effects [175, 176]. A side chain-shortened C₂₃ homologue of UDCA, nor-ursodeoxycholic acid (norUDCA), significantly reduced alkaline phosphatase in PSC patients [177]. NorUDCA undergoes cholehepatic shunting and promotes HCO₃⁻ secretion, exhibiting potent anti-cholestasis and anti-inflammatory effects in experimental models [178].

NAFLD is the most common chronic liver disease with prevalence in about 25% of the U.S. population, and is a risk factor for type 2 diabetes, obesity, and cardiovascular disease [179, 180]. Bile acid composition and pool size are altered in insulin-resistant patients, characterized by increased hydrophobic plasma bile acids and 12α-hydroxylated bile acids [181]. NASH is a progressive form of NAFLD that may lead to cirrhosis and HCC [5, 180, 182]. During the progression of cirrhosis, conjugated primary bile acids increased and secondary bile acids decreased. There is also an associated shift in the microbiome to increased Firmicutes (Clostridium cluster XIVa) and production of DCA, which increases LPS and inflammation to reduce bile acid synthesis [183]. A prospective, cross-sectional study showed increased CA and CDCA, increased bile acid synthesis, and increased ratio of primary bile acids to secondary bile acids in NASH patients compared to healthy control patients [184]. Another recent study also showed patients with NASH have increased circulating bile acids, especially conjugated primary bile acids, decreased secondary bile acids, and increased ratio of conjugated CA to CDCA [185]. A Finish population study reports Clostridium subclusters IV and XIVa are associated with fatty liver disease [186]. Interestingly, a recent transcriptome analysis of NAFLD liver samples revealed increased CYP7A1 expression, which was inversely correlated to expression of SHP and FGF19 in early-stage NAFLD but decreased with further progression of fibrosis [187]. High fat, high cholesterol and high fructose diets cause hepatic inflammation and NASH fibrosis in experimental animal studies [188–190]. The molecular mechanism of progression from simple steatosis to fibrosis in NASH is not completely understood. The role of FXR and the gut microbiome in NAFLD and NASH are described in several recent reviews [191–195].

Obeticholic acid (OCA, OCALIVA^®, Intercept Pharmaceuticals), 6α-ethyl CDCA, is a potent, selective FXR agonist (EC₅₀ = 0.099 μM), which decreased bile acid synthesis and alleviated inflammation in experimental cholestasis [196]. OCA therapy improved NASH scores in PBC patients who had inadequate responses to UDCA and in non-cirrhotic NASH patients [197–203]. However, OCA has several unwanted side effects, including pruritus and increased serum LDL-cholesterol and decreased HDL-cholesterol. The third phase of OCA clinical trials (25 mg/day) showed significant improvement of NASH fibrosis by 1 stage (regression from Stage 3 to Stage 2) without worsening fibrosis. However, the FDA recently rejected the new drug application for OCA for the treatment NASH-fibrosis because the benefits of OCA remain uncertain and do not outweigh the potential risks.

Several non-steroid FXR agonists are in clinical trials for NASH [204]. Cilofexor (Gilead) reduced steatosis and serum ALT levels but had adverse effects of increased cholesterol, reduced HDL levels, and pruritis in patients with PSC [205]. Maintaining bile acids homeostasis is essential for protecting against liver-related metabolic diseases. Combination therapies of Cilofexor with an acetyl-CoA carboxylase inhibitor (Firsocostat, Gilead) significantly reduced steatosis, inflammation, and serum AST and ALT in patients with bridging fibrosis and cirrhosis [206]. However, Cilofexor and Firsocostat combination therapy also failed in NASH trials. Tropifexor (Novartis) reduced liver fat and serum ALT but had adverse effects of pruritus, increased LDL and decreased HDL in NASH and PBC patients [204]. Tropifexor also failed in a clinical trial for NASH.

FXR antagonists have been shown to have metabolic benefits in NAFLD mouse models and in human patients [117, 207]. UDCA is a weak FXR antagonist proven to be of some therapeutic value in treating obese NAFLD patients [208]. Metformin, a first line drug for type 2 diabetes, decreased Bacteroides fragilis in the gut and increased GUDCA, which antagonized intestinal FXR to improve metabolism and hyperglycemia in diabetic patients and in high fat diet-induced obese and diabetic mice [207]. Antagonism of FXR reduces FGF19 production in the intestine and stimulates bile acid synthesis in the liver. Increasing bile acid synthesis reduces hepatic cholesterol. Many studies show increased serum FGF19 levels after metabolic gastric surgery, which are correlated to diabetes remission, suggesting that FGF19 may be used directly as a drug therapy for NAFLD. NGM282 (Aldafermin, NGM Pharma), an FGF19 variant without tumorigenic activity, has been shown to improve PBC in a phase 2 trial [209]. Aldafermin reduced bile acid synthesis and liver inflammation and significantly improved NASH fibrosis scores in 12 weeks [210]. However, Aldafermin failed to reach the primary endpoint of reducing fibrosis by greater than one stage without worsening of NASH in a phase 2B clinical trial.

17. Conclusion and future perspectives

Bile acids (UDCA, CA and CDCA) have been used for treating digestive diseases since the 1960s. The discovery of bile acids as specific FXR agonists in 1990 have accelerated the development of bile acid-based drugs for treating liver-related metabolic disease in last two decades. Disease mechanisms and pathogenesis of NASH are complex and not completely understood. Targeting to the gut microbiota by drugs, probiotics and diets is an emerging area of drug development for diabetes, obesity, NASH and NAFLD. FXR and TGR5 signaling pathways plays critical role in integrating liver metabolism and regulation of homeostasis. There are no FDA approved drugs for NASH. Many bile acid/FXR based drugs targeting to FXR are in clinical trials but many of them failed in the phase 2b/3 trials for NASH [211]. Most current drug therapies target FXR signaling to inhibit bile acid synthesis, but hepatic cholesterol and LDL-cholesterol may accumulate to cause dyslipidemia and lead to further adverse effects. Antagonizing intestine FXR stimulates bile acid synthesis to reduce hypercholesterolemia and protects against cardiovascular and chronic liver diseases. Cholesterol-lowering statins are effective for treating dyslipidemia and may be used to treat chronic liver diseases including NASH [212]. Statins reduce macrophage activation, reduce inflammatory cytokines, inflammatory cell migrations and anti-fibrogenesis. A recent study shows that OCA inhibited bile acid synthesis and Cyp7a1 expression but failed to reverse fatty liver, whereas atorvastatin reduced liver injury and vascular damage and stimulated energy metabolism via altering gut microbiome/TGR5 signaling in adipose tissue in a mouse model of NAFLD [213]. Thus, statins reduce hepatic cholesterol, reshape the gut microbiota to antagonize intestinal FXR and activate TGR5 signaling, which stimulates GLP-1 secretion and energy metabolism to improve NAFLD and protect against NASH. Further research into the roles of bile acids and cholesterol in FXR and TGR5 signaling crosstalk via the gut microbiota is needed for the development of effective drugs for NASH-fibrosis, NAFLD and HCC.

Highlights:

• FXR is a bile acid-activated receptor that acts as a metabolic sensor for integrating metabolic homeostasis.

• FXR in the gut-to-liver axis plays a central role in the regulation of bile acid synthesis and metabolic homeostasis.

• Dysregulation of bile acid and lipid homeostasis causes dysbiosis and contributes to the pathogenesis of liver-related metabolic diseases.

• Drugs targeting bile acids and FXR are in clinical trials for treating cholestatic liver diseases and non-alcoholic fatty liver diseases, diabetes, and obesity.

Abbreviations

CA

cholic acid

CDCA

chenodeoxycholic acid

CYP7A1

cholesterol 7α-hydroxylase

CYP7B1

sterol 12α-hydroxylase

CYP27A1

sterol 27-hydroxylase

DCA

deoxycholic acid

FGF15

fibroblast growth factor 15

FGF19

fibroblast growth factor 19

FGFR4

fibroblast growth factor receptor 4

FXR

farnesoid X receptor

HCC

hepatocellular carcinoma

LCA

lithocholic acid

MCA

muricholic acid

NAFLD

non-alcoholic fatty liver disease

NASH

non-alcoholic steatohepatitis

NR

nuclear receptor

OCA

obeticholic acid

PBC

primary biliary cholangitis

PFIC

progressive familial intrahepatic cholestasis

PSC

primary sclerosing cholangitis

TGR5

Takeda G protein-coupled receptor 5

UDCA

ursodeoxycholic acid