Novel bile acid therapeutics for the treatment of chronic liver diseases

Abstract

Recent developments in understanding the role of bile acids (BAs) as signalling molecules in human metabolism and inflammation have opened new avenues in the field of hepatology research. BAs are no longer considered as simple molecules helping in fat digestion but as agents with real therapeutic value in treating complex autoimmune and metabolic liver diseases. BAs and their receptors such as farnesoid X receptor, transmembrane G protein-coupled receptor 5 and peroxisome proliferator-activated receptor have been identified as novel targets for drug development. Some of these novel pharmaceuticals are already in clinical evaluation with the most advanced drugs having reached phase III trials. Chronic liver diseases such as primary biliary cholangitis, primary sclerosing cholangitis and nonalcoholic fatty liver disease, for which there is no or limited pharmacotherapy, are most likely to gain from these developments. In this review we discuss recent and the most relevant basic and clinical research findings related to BAs and their implications for novel therapy for chronic liver diseases.

Introduction

The incidence and prevalence of liver diseases in general is rising and significantly contributing to the increasing burden on health care [Williams and Horton, 2013]. To deal with the challenge of treating patients with liver diseases, clinicians need effective therapies that target the underlying disease pathology as well as symptoms and complications associated with the disease. Whilst great success has been achieved in the last few years in the pharmacological treatment of viral hepatitis [Lam et al. 2015], the need for effective drug therapies in metabolic and cholestatic liver diseases has been only partially met. As obesity is becoming increasingly prevalent, there is a clear and pressing need for drug therapies in patients with nonalcoholic fatty liver diseases (NAFLDs) and nonalcoholic steatohepatitis (NASH). At the same time, patients with primary sclerosing cholangitis (PSC) and high-risk primary biliary cholangitis (PBC) are also in need of better second-line disease-modifying drugs as well as effective therapies to manage associated symptoms (e.g. pruritus).

Bile acids (BAs) have always been of interest to gastroenterologists and hepatologists with their traditional role in fat digestion known for more than 50 years [Borgstrom et al. 1957]. In recent years, there has been a growing interest in BAs as signalling molecules and they are emerging as the key players in the quest for novel drug therapies in liver diseases. Major developments achieved in the basic and clinical research related to BAs have augmented our interest in exploiting their physiological role for therapeutic benefit in liver diseases. This has effectively set the stage to identify novel targets for treating patients with PBC, PSC and NAFLD for which there is currently limited or no effective drug therapy.

An in-depth review of BAs with respect to their chemistry, synthesis, transport and regulation is beyond the scope of this article and has been extensively reviewed elsewhere [Dawson et al. 2009; Monte et al. 2009; Schaap et al. 2014]. Therefore, in the first part of this review we provide an overview of BA synthesis, transport, signalling and regulation. We then describe recent advances in the understanding of BA receptors in relation to cholestasis, glucose and lipid metabolism, immune function and antifibrotic actions. These relevant updates will help the reader to better understand the therapeutic benefits of BAs in chronic liver diseases covered in the later part of this paper. In this review of potential novel BA-based therapeutic agents we mainly focus on the treatment of patients with PBC, PSC and NAFLD.

BA synthesis and transport

BAs along with phospholipids and cholesterol are major constituents of bile. BAs are amphipathic molecules (i.e. with both hydrophilic and hydrophobic regions) with detergent-like actions and are synthesized from enzymatic catabolism of cholesterol by the hepatocytes [Monte et al. 2009]. BA synthesis is a complex process involving at least 17 different enzymes but can be summarized into three main steps: modification of the steroid ring, cleavage of the side chain and conjugation with glycine or taurine [Russell, 2003]. Two pathways exist for BA synthesis. The classical (‘neutral’) pathway is responsible for the production of cholic acid (CA) and chenodeoxycholic acid (CDCA) which accounts for 90% of primary BA synthesis in humans [Anderson et al. 1972]. The other 10% is produced by the alternative (‘acidic’) pathway which can only produce CDCA. Cytochrome P450 7A1 (CYP7A1) is the rate-limiting enzyme in BA synthesis. After their synthesis, unconjugated CA and CDCA are targeted to the peroxisomes where they are conjugated (amidation) with glycine and taurine that renders them more hydrophilic and more readily secretable in the bile.

In humans, predominant conjugated BAs are glycoconjugates and under physiological pH conditions these conjugated BAs exist as anionic salts and are therefore called ‘bile salts’ (BS). These BS are stored in the gallbladder and upon ingestion of a meal they are released into the intestinal lumen where they facilitate absorption of fat and fat-soluble vitamins. Conjugated primary BAs present in the intestinal lumen are modified by the intestinal bacteria by deconjugation, oxidation and dehydroxylation to produce secondary BAs: lithocholic acid (LCA) and deoxycholic acid (DCA) [Ridlon et al. 2006]. Human bile predominantly contains CDCA and DCA and a very small amount of ursodeoxycholic acid (UDCA). Hydrophobicity of the BAs is a determinant of their cytotoxicity which increases in the order of LCA > DCA > CDCA > CA > UDCA [Carey, 1983].

After their normal physiological function is completed in the intestine, BAs reach the ileum where most are reabsorbed efficiently via a sodium-dependent process. The apical sodium-dependent BA transporter (ASBT, gene symbol SLC10A2) expressed in the distal ileum is the predominant transporter mediating the ileal uptake of conjugated BAs [Craddock et al. 1998]. ASBT mediates active transfer of BAs across the luminal plasma membrane to an intracellular protein called ileal BA binding protein (IBABP) that facilitates intracellular diffusion of BAs to the basolateral membrane (Figure 1). Then, BAs exit the enterocyte across the basolateral plasma membrane mediated by the organic solute transporter (OSTα/β) and enter the portal bloodstream [Rao et al. 2008]. BAs circulating in the portal circulation are transported across the basolateral membranes of the hepatocytes via sodium taurocholate cotransporting polypeptide transporter (NTCP, gene symbol SLC10A1) [Hagenbuch and Meier, 1994]. Finally, BAs are transported across the canalicular plasma membrane of the hepatocytes via the bile salt export pump (BSEP) and secreted into bile. This efficient cycle between small intestine and liver is repeated several times a day to ensure 95% of BAs re-enter the liver, leaving only approximately 5% (or approximately 0.5 g/day) in the intestinal lumen [Hofmann, 1984].

Figure 1.

Enterohepatic circulation of bile acids (BAs) via enterocyte in terminal ileum. (1) Primary BAs synthesized in liver and excreted into duodenum as constituent of bile; (2) BAs avidly and actively reabsorbed in the terminal ileum via ASBT (apical sodium bile acid transporter); (3) BAs transported intracellularly by IBABP (intracytosolic bile acid binding protein); (4) BAs free to bind with nuclear receptor FXR (farnesoid X receptor); (5) BAs released into portal venous circulation via organic solute transporter (OST) α/β and circulated back to liver. SHP, small heterodimer partner.

BA regulation

BA homeostasis is under tight regulation mediated by the negative feedback effect of BAs on the activity and expression of CYP7A1 as well as signalling via farnesoid X receptor (FXR) and fibroblast growth factor (FGF19).

FXR

FXR was first identified in 1995 as an orphan receptor. It is a member of the nuclear receptor superfamily and acts as key regulator in a diverse range of cell functions including development, differentiation and metabolism [Mangelsdorf et al. 1995]. FXR was so named because a supraphysiological concentration of farnesol, an intermediate in the mevalonate pathway, was found to demonstrate weak agonism [Modica et al. 2010]. The discovery that primary BAs were the natural, endogenous ligands for FXR was reported at the turn of the century [Makishima et al. 1999; Parks et al. 1999]. Soon FXR was implicated in BA homeostasis following the discovery that FXR knockout (FXR^–/–) mice showed diminished ability to downregulate CYP7A1 mRNA in response to BAs [Sinal et al. 2000]. Although FXR is expressed in the ileum, liver, adrenal glands and the kidneys, the intestine (mainly ileum) seems to have the most intense FXR expression [Inagaki et al. 2006]. Indeed, among all the nuclear hormone receptors, FXR is the most dedicated to BA signalling [Schaap et al. 2014]. The activation of FXR by primary BAs has several downstream effects on both hepatocytes and enterocytes. In the hepatocytes, FXR activation induces small heterodimer partner (SHP) that downregulates the synthesis of BAs by inhibiting CYP7A1 [Kir et al. 2012]. Also, FXR activation downregulates NTCP (reduced BA uptake) and upregulates BSEP (increased export of BAs) [Martinez-Augustin and Sanchez De Medina, 2008]. In the enterocytes, FXR activation reduces ASBT expression (inhibits BA absorption) and increases expression of IBABP and OSTα/β (prevents intracellular BA accumulation) [Martinez-Augustin and Sanchez De Medina, 2008]. FXR activation also prevents BA toxicity by transcriptional induction of detoxification enzymes and canalicular secretion of BAs via upregulation of BSEP.

FGF19

A second FXR-dependent mechanism to reduce primary BA synthesis is through the production of enterokine FGF15 (in rodents) or FGF19 (in humans) [Holt et al. 2003; Inagaki et al. 2005]. When there is a high BA load in the ileum, activated FXR induces transcription of the FGF19 in the ileum. FGF19 is able to travel in the bloodstream and bind to its receptor hepatocyte FGF receptor 4 (FGFR4) and initiate a SHP-independent downregulation of CYP7A1, resulting in inhibition of BA synthesis [Jones, 2012; Kir et al. 2012].

In summary, FXR and FGF19 activation reduces endogenous BA synthesis, protects hepatocytes from BA toxicity and promotes secretion of BAs. Therefore the ‘anticholestatic’ effects of the BA-FXR-FGF19 signalling cascade have potential therapeutic implications in cholestatic diseases.

Transmembrane G protein coupled receptor 5

Transmembrane G protein coupled receptor 5 (TGR5, also called Gpbar-1) is a G protein coupled BA receptor that was identified as the first cell surface receptor for BAs [Maruyama et al. 2002; Kawamata et al. 2003]. TGR5 is not expressed in the hepatocytes but found in the cholangiocytes as well as a variety of other cell types such as brown adipose tissue, brain, gall bladder epithelium, intestines, spleen, endothelial cells, Kupffer cells and CD14+ cells [Maruyama et al. 2002; Kawamata et al. 2003; Keitel et al. 2007]. Although many BAs are capable of activating TGR5, the most potent natural ligands are taurine-conjugated secondary BAs, such as taurolithocholate [Keitel et al. 2008]. It has been shown that BA-induced TGR5 activity plays a major role in glucose homeostasis, increased energy expenditure, oxygen consumption and gallbladder filling [Katsuma et al. 2005; Watanabe et al. 2006; Thomas et al. 2009; Li et al. 2011]. In addition, TGR5 activation improves hepatic steatosis and insulin sensitivity and protects biliary epithelium against the detergent effect of BAs. More recently, BA-mediated TGR5 activation has also been shown to have an anti-inflammatory role by reducing nuclear factor κB (NFκB) translocation [Li et al. 2011]. Due to its diverse and favourable effects, TGR5 is an emerging target for drug discovery with a potentially beneficial role of TGR5 agonists in the treatment of type 2 diabetes mellitus (T2DM) and inflammation-driven metabolic diseases such as NASH.

Scientific rationale for BA-based therapy

In addition to their key role in BA homeostasis, FXR activity and FGF19 signalling are involved in diverse biological pathways. Essentially, FXR exerts its functions by eliciting transcriptional alterations and controls a number of important metabolic pathways [Schaap et al. 2014]. In this section, we briefly review the key biological processes modulated by FXR and FGF19 with respect to their attractive therapeutic implications.

Glucose and lipid metabolism

Activation of FXR inhibits the expression of hepatic sterol regulatory element-binding protein 1c (SREBP-1c) [Watanabe et al. 2004]. SREBPs are transcription factors that act as master regulators of lipid metabolism. They act to control the biogenesis of cholesterol and also control the expression of genes involved in lipogenesis. In an animal model, activation of FXR by CA inhibited hepatic SREBP-1c expression in a SHP-dependent manner, leading to reduction in serum triglyceride levels [Watanabe et al. 2004]. In addition, FXR can induce the expression of apolipoprotein C-II [Houten et al. 2006] which is a coactivator for lipoprotein lipase that acts to clear serum triglyceride from the circulation. FGF19 is also shown to regulate key enzymes in hepatic lipid synthesis [Miyata et al. 2011].

FXR signalling is also essential to maintain glucose homeostasis. In an FXR^–/– mouse model, elevated serum glucose and impaired glucose and insulin tolerance were demonstrated by Ma and colleagues [Ma et al. 2006]. This study also demonstrated that administration of CA repressed the expression of gluconeogenic genes and decreased serum glucose in wild-type mice. In addition, FGF19 is involved in glucose metabolism with its actions resembling that of insulin; that is, inhibition of gluconeogenesis and stimulation of glycogen synthesis [Potthoff et al. 2011].

Immune functions

It has long been recognized that BAs are bacteriostatic (but weakly bactericidal) and that decrease in BAs within the small bowel leads to bacterial overgrowth [Floch et al. 1971; Berg, 1995]. This has been confirmed recently by the experimental studies of obstructive cholestasis and cirrhosis that showed oral administration of BAs reduced bacterial overgrowth as well as maintained the intestinal barrier function and prevented endotoxaemia [Lorenzo-Zuniga et al. 2003; Ogata et al. 2003]. BA-FXR signalling has been proposed as the key mechanism by which BAs control bacterial overgrowth and maintain the epithelial barrier. A landmark study showed that BAs regulate an anti-inflammatory response via FXR in the terminal ileum [Inagaki et al. 2006]. In this study of a rodent model of cholestasis FXR agonist treatment protected the epithelial barrier by increasing the expression of several genes associated with intestinal mucosal defence pathways and decreased the number of bacteria isolated from mesenteric lymph nodes. Similarly, treatment with a FXR agonist has been shown to maintain the epithelial barrier in an animal model of colitis in wild type, but not FXR−/− mice [Gadaleta et al. 2011]. In addition, FXR activation rendered several different immune cell types refractory to stimulation with lipo-polysaccharide (LPS), a bacterial cell wall component. More recent evidence shows FXR agonism reduces the LPS-induced production of proinflammatory cytokines by macrophages, whilst maintaining the production of anti-inflammatory interleukin (IL)-10 [Haselow et al. 2013] and attenuates the chemoattractant IL-8 response to stimulation with tumour necrosis factor α (TNFα) [Speight et al. 2015]. Finally, FXR and FGF19 exert anti-inflammatory activity via suppression of NFκB, which is a key nuclear receptor in both acute and chronic inflammatory processes. This is supported by observations that FXR and NFκB mutually antagonize each other [Wang et al. 2008] and activation of FXR and FGF19 inhibit the expression of NFκB controlled inflammatory genes [Drafahl et al. 2010; Gadaleta et al. 2010; Zhou et al. 2014].

Taken together, these results suggest FXR agonists have anti-inflammatory actions and may have potential therapeutic utility in preventing bacterial translocation and reducing spontaneous bacterial peritonitis in patients with cholestasis and cirrhosis.

Liver fibrosis and carcinogenesis

Evidence shows that FGF19 increases proliferation of hepatocytes through activation of FGFR4, and BA-FXR-FGF19 signalling is essential for normal liver regenerative process [Huang et al. 2006; Wu et al. 2010; Zhang et al. 2012; Uriarte et al. 2013; Kong et al. 2014]. Therefore, inhibition of FGF19 signalling could be a potential therapy for hepatocellular carcinoma (HCC). Also the observation that FXR^–/– mice have high incidence of HCC [Kim et al. 2007] suggests the potential regulatory role of FXR in tumour suppression. This has recently been corroborated by two animal studies in which FXR agonist treatment prevented development of liver cancer and reduced liver tumour size and metastasis [Deuschle et al. 2012; Jiang et al. 2013].

FXR activation is also implicated in the inhibition of fibrotic mechanisms within the liver via hepatic stellate cells (HSCs) [Fiorucci et al. 2005; Renga et al. 2011]. In a mouse model, administration of a FXR agonist for 12 weeks promoted resolution of liver fibrosis [Fiorucci et al. 2004]. However, a more recent study has contradicted this by showing a low level of FXR expression in HSCs in liver fibrosis and suggesting HSCs may not represent direct therapeutic targets for FXR ligands [Fickert et al. 2009]. Therefore, the current evidence on the antifibrotic effect of FXR is equivocal and merits further investigation.

Novel BA-based therapies

PBC, PSC and NASH represent complex, multifactorial diseases in which effective drug management remains an unmet clinical need. There is a clear need beyond UDCA in patients with PBC and PSC and beyond diet and lifestyle modifications in patients with NASH. In this section we review innovative BA-based therapeutic approaches being investigated for these diseases (summary in Table 1).

Table 1.

Novel bile acid based therapeutic approaches in chronic liver diseases.

Class of molecule	Example molecules	Therapeutic rationale	Target disease	Phase of development	ClinicalTrials.gov identifier
FXR agonists	INT-747, INT-767, GW4064, GSK2324, PX-102, Way362450, fexaramine, LJN452	↓BA synthesis and promotion of BA excretion	PBC	Phase II completed	NCT00570765
				Phase II completed	NCT00550862
				Phase III (POISE)	NCT01473524
				Phase III	NCT02308111
				Phase II	NCT02516605
				Phase II	NCT02516605
			PSC	Phase II (AESOP)	NCT02177136
		↓lipogenesis, gluconeogenesis and liver inflammation	T2DM with NAFLD	Phase II completed	NCT00501592
			NASH	Phase II	NCT01265498
		↑tumour suppression and ↓liver fibrosis	HCC	–	–
TGR5 agonist	INT-767, INT-777	↓BA synthesis, ↑bile flow, anti-inflammatory	PSC
		↑energy expenditure, anti-inflammatory	NASH	–	–
FGF-19 analogue	NGM-282	↓BA synthesis, anti-inflammatory, antifibrotic	PBC	Phase II	NCT02026401
ASBT inhibitor	LUM-001, GSK2330672	↓enterohepatic circulation of BAs and ↑faecal excretion	Pruritus in PBC	Phase II (CLARITY)	NCT01904058
				Phase II	NCT01899703
PPAR agonists	Fenofibrate, bezafibrate	↓BA synthesis and promotion of BA excretion, anti-inflammatory, may ↑FXR activity	PBC, PSC	Phase III (BEZURSO)	NCT01654731
	GFT505	↑insulin sensitivity and glucose homeostasis, hepatoprotective, anti-inflammatory	NASH	Phase II (GOLDEN)	NCT01694849
UDCA related	norUDCA	cholangioprotective, stabilizes ‘biliary bicarbonate umbrella’	PSC	Phase II (NUC-3)	NCT01755507
	Tauroursodeoxycholate (TUDCA)	more hydrophilic than UDCA	PBC	Phase III	NCT01857284
Fatty acid–bile acid conjugate	Aramchol	Stearoyl coenzyme A desaturase 1 (SCD1) inhibitor, ↓liver fat	NASH	Phase II (ARREST)	NCT02279524

ASBT, apical sodium dependent BA transporter; BA, bile acid; FGF, fibroblast growth factor; FXR, farnesoid X receptor; HCC, hepatocellular carcinoma; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; PBC, primary biliary cholangitis; POISE, PBC OCA International Study of Efficacy; PPAR, peroxisome proliferator-activated receptor; PSC, primary sclerosing cholangitis; T2DM, type 2 diabetes mellitus; TGR, transmembrane G protein-coupled receptor; UDCA, ursodeoxycholic acid.

UDCA: ‘the current BA therapy’

UDCA is the only US Food and Drug Administration (FDA) approved drug for PBC and it is the current standard of care for patients with this condition. The mechanism of action of UDCA has been well established and comprehensively explained in an excellent recent review [Beuers et al. 2015]. In brief, UDCA has potent anticholestatic, antiapoptotic and anti-inflammatory properties. Notably, UDCA is a weak FXR and TGR5 ligand [Parks et al. 1999; Halilbasic et al. 2013]. The optimum dose of UDCA in treating patients with PBC is 13–15 mg/kg/day and guidelines recommend initiating treatment at a low dose and increasing it gradually to the optimum dose [Beuers et al. 2009; Lindor et al. 2009].

Multiple lines of evidence confirm that UDCA improves biochemical markers of cholestasis and may delay the progression of PBC [Beuers et al. 2009; Lindor et al. 2009]. However, unfortunately the response to UDCA is not universal and up to 40% of patients with PBC do not improve their liver biochemistry on UDCA. Indeed a substantial proportion of patients have disease progression despite UDCA therapy. Patients with incomplete biochemical response to UDCA are referred to as ‘UDCA nonresponders’ based on different treatment response biochemical criteria (e.g. Paris and Barcelona criteria) [Corpechot, 2012]. UDCA nonresponders are at higher risk of disease progression, symptom burden and poor prognosis compared with ‘UDCA responders’. In addition, the Global PBC Study Group recently demonstrated that serum levels of alkaline phosphatase (ALP) and bilirubin are surrogate endpoints of outcomes in PBC and patients with an ALP greater than 1.67 × upper limit of normal (ULN) or bilirubin greater than ULN have increased risk of transplantation or death [hazard ratio (95% confidence interval, CI): 2.83 (2.4–3.4); p < 1 × 10⁻³⁴] [Lammers et al. 2014].

In PSC patients the conventional dose of UDCA (10–15 mg/kg/day) is safe but high dose (28–30 mg/kg/day) has been shown to be harmful [Lindor et al. 2009]. Moreover, the long term efficacy of UDCA therapy in PSC is unclear as the evidence suggests that UDCA improves liver biochemistry but has no significant effect in slowing disease progression [Poropat et al. 2011]. Therefore, the current guidelines recommend against the use of UDCA as medical therapy in PSC [Beuers et al. 2009; Chapman et al. 2010].

In PBC and PSC, UDCA is not effective in improving cholestasis-associated symptoms such as pruritus and fatigue. Due to these limitations of UDCA, there is an urgent need for novel therapy in PSC and second line therapies in patients with PBC with UDCA nonresponse status. The potential therapeutic role of UDCA in other but rare liver diseases is also being explored. Recent experimental models of polycystic liver disease (PLD) have shown that UDCA inhibits hepatic cystogenesis by inhibiting the proliferation of polycystic cholangiocytes [Munoz-Garrido et al. 2015]. Therefore, UDCA may be an effective therapeutic option in reducing liver volume in PLD and help to improve symptoms caused by the mass effect of polycystic liver. An international, multicentre, randomised controlled trial is currently recruiting patients to assess the efficacy of UDCA (15–20 mg/kg/day UDCA for 24 weeks) in reducing total liver volume in PLD patients (CURSOR study; NCT02021110).

FXR agonists

As noted above, among all nuclear receptors FXR has emerged as a prime therapeutic target due to its diverse functions in the regulation of BAs, metabolism of glucose and lipids, and anti-inflammatory activity. Several FXR agonists have been developed with two basic structures: small molecule, steroidal semisynthetic ligands and nonsteroidal, fully synthetic ligands. The most clinically advanced therapeutic FXR agonist is INT-747 [obeticholic acid (OCA), 6α-ethyl-chenodeoxycholic acid, Intercept Pharmaceuticals, New York, USA]. OCA is a steroidal semisynthetic BA in which CDCA has been modified by the addition of an alkyl group to form a more potent FXR agonist [Pellicciari et al. 2002].

GW4064 (GlaxoSmithKline, NC, USA) is a nonsteroidal fully synthetic FXR agonist first developed in 2000 [Crawley, 2010]. Animal studies have shown that GW4064 prevents diet-induced hepatic steatosis and insulin resistance, and attenuates endotoxin-induced hepatic inflammation by repressing macrophage activation [Ma et al. 2013; Yao et al. 2014]. As an FXR agonist, GW4064 has greater potency than CDCA but is currently not being evaluated for clinical use. It can be commercially obtained from Sigma Aldrich (St. Louis, USA) for experimental use.

PX-102 is a therapeutic nonsteroidal compound manufactured as a modification of GW4064 (Phenex Pharmaceuticals AG, Ludwigshafen, Germany). It has demonstrated some efficacy in the mouse models of NAFLD by decreasing levels of serum and liver cholesterol and triglyceride [Hambruch et al. 2013].

WAY-362450 (Exelixis Inc., California) is a fully synthetic agonist manufactured as an azepino derivative [Flatt et al. 2009]. Studies on animal models of NAFLD and NASH demonstrated that treatment with WAY-362450 reduced liver inflammation and fibrosis with an associated decrease in serum liver enzymes [Zhang et al. 2009].

Finally, fexaramine is another nonsteroidal FXR ligand shown to have distinct genomic targets and favourable metabolic effects in a mice model but currently is not being studied in humans [Downes et al. 2003].

FXR therapeutics in PBC

By virtue of its potent FXR agonist action and resulting effects on BAs, OCA is emerging as a promising second-line agent for treating patients with PBC who are ‘UDCA nonresponders’. Two phase II randomized placebo-controlled trials have evaluated OCA in PBC with improvement in liver enzymes as the primary endpoint and as a surrogate marker for patient outcome. The first study [ClinicalTrials.gov identifier: NCT00570765] showed that in comparison to placebo, monotherapy with OCA 10 or 50 mg daily for 12 weeks (n = 59) reduced ALP from a mean of 3.9 × ULN to 1.9 × ULN [Kowdley et al. 2011]. Significant improvements were also seen in the levels of serum γ-glutamyl transpeptidase (γ-GT), bilirubin, C-reactive protein, immunoglobulin M (IgM) and TNFα. More recently, in a phase II multicentre double-blind efficacy trial, 165 patients with PBC and inadequate response to UDCA (serum alkaline phosphatase levels > 1.5 × ULN) were randomly treated with 10 mg, 25 mg or 50 mg doses of OCA or placebo once daily for 3 months [ClinicalTrials.gov identifier: NCT00550862] [Hirschfield et al. 2015]. Compared with placebo, OCA significantly decreased ALP levels with average decline of 20–25% from baseline. Significant improvements were also seen in the levels of serum bilirubin, γ-GT and ALT. The biochemical benefit of OCA was maintained in 61 patients who had completed a 12-month open-label extension of the study. Currently two phase III studies of OCA in PBC are ongoing [ClinicalTrials.gov identifiers: NCT01473524, NCT02308111] with early results from the first phase III PBC OCA International Study of Efficacy (POISE) study showing clinically meaningful biochemical improvement [Nevens et al. 2014].

LJN452 (Novartis, Basel, Switzerland) is a non-BA FXR agonist currently entering a phase II trial to assess safety, tolerability and efficacy in patients with PBC [ClinicalTrials.gov identifier: NCT02516605].

FXR therapeutics in NAFLD

NAFLD is the most common cause of chronic liver disease in the developed world, affecting up to a third of the population. It is characterized by the accumulation of hepatic fat and presents as a spectrum from simple steatosis to NASH, fibrosis and eventually cirrhosis [Sattar et al. 2014]. NAFLD is strongly associated with T2DM and patients with NASH have an increased risk for the development of progressive fibrosis, cirrhosis and HCC. So far, OCA is the only FXR agonist studied in patients with NAFLD, with encouraging results from the FLINT (Farnesoid X nuclear receptor ligand obeticholic acid for non-cirrhotic, non-alcoholic steatohepatitis) study reported recently [Neuschwander-Tetri et al. 2015]. In this double-blind, randomized, placebo-controlled trial, patients with NASH without cirrhosis were randomized to receive either 25 mg of OCA (n = 141) or placebo (n = 142) for 72 weeks. The primary outcome measure was reduction in the histological score for fibrosis (NAFLD activity score) from baseline to the end of treatment. The trial was stopped early as 45% of patients in the OCA group had reached the primary endpoint compared with 21% in the placebo group (relative risk 1.9; 95% CI 1·3–2·8). The OCA intervention also significantly improved the serum ALT levels. These results suggest that the FXR agonist OCA may be beneficial in NAFLD and prevent progression of fibrosis in patients with NASH.

Safety of FXR therapy

Pruritus is the most common adverse event reported to be associated with OCA in PBC and NAFLD trials. In the phase II PBC trial, frequency of pruritus was 47% with 10 mg, 87% with 25 mg and 80% with 50 mg of OCA [Hirschfield et al. 2015]. Similarly in the FLINT study, 23% of patients in the OCA arm developed pruritus compared with 6% in the placebo group [Neuschwander-Tetri et al. 2015]. Data from both trials suggested that pruritus was directly related to the OCA dose. This issue is being investigated in the phase III (POISE) study, with early results suggesting the frequency of pruritus is lowest when OCA is started at 5 mg and titrated upwards [Nevens et al. 2014].

A potentially more concerning adverse event in the FLINT study was dyslipidaemia seen in patients treated with OCA. Increase in low-density lipoprotein and decrease in high-density lipoprotein (HDL) levels were observed in the OCA group. Although dyslipidaemia is a well-known cardiovascular risk factor, whether this is true for patients treated with an FXR agonist is unclear. Interestingly animal models of atherosclerosis have provided some evidence to suggest that FXR agonists are protective [Hartman et al. 2009]. Clearly more research is needed to assess the cardiovascular risk of FXR agonist therapy in NASH. Interestingly in the PBC study OCA treatment was associated with decrease in serum levels of total and HDL cholesterol [Hirschfield et al. 2015].

An unproven safety concern is the potential for FXR agonists to predispose people to the development of HCC induced by FXR overexpression and elevation of circulating FGF19 levels. FGF19 transgenic mice develop HCC [Nicholes et al. 2002] and overexpression of FGF19 has been associated with development of HCC in both animal models and potentially in humans [Lin and Desnoyers, 2012; Schaap et al. 2015]. Therefore, it is feasible that overpromotion of FXR with FXR agonist therapy could lead to high levels of FGF19, which in turn is carcinogenic. This concern merits further clinical evaluation. Similar procarcinogenic concern applies to FGF19 analogues. One proposal to overcome this undesirable effect is to engineer FGF19 variants that have lost the mitotic activity but still are effective in their metabolic activity [Wu et al. 2010].

INT-767

INT-767, a semisynthetic BA analogue, is a 23-sulphate derivative of OCA with dual FXR and TGR5 agonist actions but with a higher affinity to FXR [Rizzo et al. 2010]. Prominent features of INT-767 include inhibition of BA synthesis, simulation of bicarbonate-rich choleresis by enhancing biliary bicarbonate secretion and immune modulation via inhibition of NFκB. In the mouse model of sclerosing cholangitis, INT-767 reduced serum levels of ALT, ALP as well as liver inflammation and fibrosis [Baghdasaryan et al. 2011]. Interestingly these effects were shown to be mediated exclusively by FXR and not by TGR5. Therefore, INT-767 may be a potential therapeutic agent in treating patients with cholestasis and it is currently entering phase I clinical trials.

INT-777/TGR5 agonists

INT-777 [6α-ethyl-23(S)-methylcholic acid] is a potent, semisynthetic and selective TGR5 agonist [Pellicciari et al. 2009]. Animal studies have shown that INT-777 increases bile flow, produces significant reduction in weight gain and adiposity as well as improvement of liver function with concomitant reductions in steatosis and fibrosis [Pellicciari et al. 2009; Thomas et al. 2009]. These results suggest promising therapeutic potential for INT-777 in the treatment of obesity and related disorders such as NASH. A theoretical concern with the use of TGR5 agonists is the potential for aggravating pruritus, a common symptom in patients with cholestatic liver disease. Although animal studies have shown that activation of TGR5 can induce itch [Alemi et al. 2013; Lieu et al. 2014] it remains unknown whether therapeutic use of TGR5 agonists will have a similar effect on pruritus. Currently there are no ongoing clinical trials with INT-777.

NGM-282

NGM-282 (NGM Biopharmaceuticals, Inc., San Francisco, CA, USA), a biological drug, is a recombinant variant of FGF19 with potential anticholestatic properties. A phase II clinical trial evaluating the safety and tolerability of 28 days of treatment of NGM-282 with UDCA in patients with PBC has recently been completed [ClinicalTrials.gov identifier: NCT02026401] and results are awaited.

Norursodeoxycholic acid

24-norUrsodeoxycholic acid (norUDCA) is a synthetic, side-chain-shortened UDCA homologue. With its hepatocyte and cholangiocyte protective properties it has recently emerged as an attractive therapeutic candidate for cholestatic liver diseases, especially for PSC. NorUDCA differs from UDCA in metabolism and therapeutic mechanisms, with important clinical consequences. Like UDCA, norUDCA is not a direct FXR or TGR5 ligand but it is significantly more hydrophilic and less toxic than UDCA [Fickert et al. 2013]. NorUDCA is superior to UDCA in the treatment of sclerosing cholangitis, attributed largely to its ability to increase the hydrophilicity of biliary BAs, stimulate bile flow with flushing of injured bile ducts, and induce detoxification and elimination routes for BAs [Hofmann et al. 2005; Fickert et al. 2006]. Because of its shortened side chain norUDCA is not conjugated with taurine or glycine in the liver and is secreted into the bile in its unconjugated form. Human studies have shown norUDCA induces a sustained bicarbonate-rich hypercholeresis, the increased bile flow being attributed mainly to ‘cholehepatic shunting’ [Hofmann et al. 2005]. The cholangioprotective effect is mainly due to profound alkalinization of bile which stabilizes the ‘biliary bicarbonate umbrella’ [Hohenester et al. 2012] and in turn reduces ductular reaction, inflammation and fibrosis [Halilbasic et al. 2009; Fickert et al. 2013]. In addition to the anticholestatic and cholangioprotective mechanisms, rodent studies have suggested that norUDCA has potential antiproliferative, anti-inflammatory and antifibrotic properties [Fickert et al. 2006; Halilbasic et al. 2009] which could be beneficial in both cholestatic and noncholestatic conditions. Recent data also suggest norUDCA produces significant suppression of lipogenesis and normalization of BA metabolism through mechanisms involving crosstalk between CYP7A1 and SHP [Beraza et al. 2011].

Due to its multiple beneficial properties suggested by experimental data, norUDCA is a promising drug therapy to attenuate the progression of complex disorders such as PSC and NASH. Also the potential beneficial effects of combined therapy with norUDCA (bicarbonate-rich choleresis) and FXR agonist (suppression of BA synthesis) in cholestatic liver disease merits further exploration. Currently the optimal dose of norUDCA for therapeutic benefit is not known and a large, multicentre, double-blind, placebo-controlled, randomized dose-finding phase II trial (Dr Falk Pharma GmbH, Freiburg im Breisgau, Germany) is evaluating the efficacy of three different doses of norUDCA for the treatment of PSC [ClinicalTrials.gov identifier: NCT01755507].

Aramchol

Aramchol [(3β-arachidyl-amido, 7α-12α-dihydroxy, 5β-cholan-24-oic acid), Trima Israel Pharmaceutical Products, Maabarot, Israel] is a fatty acid–BA conjugate currently being investigated for NAFLD and NASH. Aramchol is a novel synthetic lipid molecule obtained by conjugating two natural components, CA and arachidic acid (saturated fatty acid). The main mechanism of action of Aramchol is via inhibition of the stearoyl coenzyme A desaturase 1 (SCD1) activity which is a key enzyme modulating fatty acid metabolism in the liver [Dobrzyn and Ntambi, 2005; Leikin-Frenkel et al. 2008]. SCD1 inhibition decreases the synthesis and increases β oxidation of fatty acids, resulting in decreased hepatic storage of triglycerides and fatty acid esters.

Aramchol has received a fast-track status from the FDA. In a phase IIa placebo-controlled trial of 58 patients [ClinicalTrials.gov identifier: NCT01094158], 3 months of treatment with single daily dose of aramchol (100 or 300 mg) was safe and well tolerated and produced significant and dose-dependent reduction in liver fat [Safadi et al. 2014]. A large multicentre, double-blind, placebo-controlled phase IIb study (ARREST trial) is currently evaluating the safety and efficacy of two aramchol doses in patients with NASH without cirrhosis [ClinicalTrials.gov identifier NCT02279524] with reduction in hepatic steatosis as the primary endpoint.

ASBT inhibitor

As noted above, ASBT is a key protein involved in the enterohepatic circulation of BAs and maintaining the BA pool. Physiological effects of ASBT inhibition include lack of ileal BA uptake, increased faecal BAs, reduced FXR stimulation and reduced FGF19 levels. In animal studies, SC-435 (an ASBT inhibitor) produced BA malabsorption (resulting in diarrhoea) and lowered plasma cholesterol [West et al. 2003]. As ileal BA uptake is upregulated in PBC [Lanzini et al. 2003], pharmacological inhibition of ASBT on the circulating levels of BAs in PBC has generated considerable interest. Recently in a bile duct ligated mice model, treatment with ASBT inhibitor A4250 was shown to attenuate BA-mediated cholestatic liver injury by reducing biliary BA output [Baghdasaryan et al. 2014].

Since BAs have been proposed as potential direct or indirect pruritogens in cholestasis, ASBT inhibitors may also have a role in treating pruritus. Recently two large multicentre, randomized phase II clinical trials evaluating the safety and efficacy of ASBT inhibitor drugs (GSK2330672 and LUM001) in patients with PBC and pruritus have completed recruitment [ClinicalTrials.gov identifiers: NCT01899703, NCT01904058]. The results of these studies are likely to inform the safety and therapeutic potential of ASBT inhibitors in the treatment of cholestatic pruritus.

Fibrates

Fenofibrate and bezafibrate are fibric acid derivatives that have been in use for over two decades primarily to treat hyperlipidaemia in patients with cardiovascular and metabolic diseases. Following the first study in 1993 that suggested fibrates improve liver biochemistry [Day et al. 1993], they have been actively pursued as potential adjuvants to UDCA therapy to improve cholestasis. The primary mechanism of anticholestatic effect of fibrates is through inhibition of BA synthesis mediated via nuclear receptor peroxisome proliferator-activated receptor (PPAR). Fenofibrate is a PPARα selective agonist and bezafibrate is a ‘pan-PPAR’ agonist as it activates all three isoforms (α, γ and δ). PPARα plays a key role in maintaining BA homeostasis by regulating genes responsible for BA synthesis and transport [Ghonem et al. 2015]. Therefore, by activating PPARα, fenofibrate and bezafibrate reduce BA synthesis (downregulate CYP7A1), decrease BA secretion into bile, facilitate elimination of toxic BA and increase biliary phospholipid output. Fibrates also have immune modulator function via inhibition of NFκB, increased apolipoprotein AII (inhibits lymphocyte activation), and suppression of lymphocyte proliferation [Vu-Dac et al. 1995; Schoonjans et al. 1996]. In addition, the crosstalk between PPARα and FXR may enhance PPARα transcription in HSCs, leading to decreased liver fibrosis [Pineda Torra et al. 2003].

To date, a number of studies (mostly uncontrolled and small case series) with bezafibrate and fenofibrate have consistently reported significant improvement in ALP, ALT and IgM in patients with PBC. Recent systematic reviews and meta-analyses also support the use of fibrates as adjuvant treatment in patients with PBC who are UDCA nonresponders despite treatment with the optimum dose of UDCA for 12 months [Grigorian et al. 2015; Zhang et al. 2015]. However, fibrates currently carry contraindications for use in patients with hepatic or severe renal dysfunction, including in patients with PBC. Therefore, safety and efficacy of fibrates need further evaluation in prospective studies before they can be used in routine clinical practice for PBC. Currently, a large multicentre, prospective, double-blind, placebo-controlled phase III study of bezafibrate in combination with UDCA (BEZURSO study) in PBC is recruiting [ClinicalTrials.gov identifier: NCT01654731] and the results are eagerly awaited to clarify the true effect of fibrates in cholestasis.

GFT505

GFT505 (GENFIT, Loos, France) a novel PPARα/δ agonist, is currently being developed as a novel therapy for NASH and has received FDA fast-track status. Preclinical studies showed hepatoprotective effects of GFT505 as it decreased hepatic lipid accumulation, improved liver dysfunction markers, and inhibited proinflammatory gene expression [Staels et al. 2013]. GFT505 treatment also decreased plasma concentrations of ALT, ALP and γ-GT. In addition, GFT505 has an insulin-sensitizing effect which in combination with hepatoprotective effects makes it a potential therapy for NAFLD [Cariou and Staels, 2014]. Currently an international placebo-controlled phase IIb study (GOLDEN trial) is investigating the safety and efficacy of GFT505 (80 and 120 mg once daily for 52 weeks) in patients with NASH without cirrhosis, with resolution of NASH without worsening of fibrosis as the primary endpoint [ClinicalTrials.gov identifier: NCT01694849].

BA therapy in malignancy

A few years ago BAs were proposed as potential shuttles to deliver chemotherapy agents to treat tumours such as HCC, cholangiocarcinoma (CCA) and colorectal carcinoma [Kramer et al. 1992]. The rationale for this proposal was that the tumours in the enterohepatic circulation maintain good expression of BA transporter proteins (especially ASBT) and the ‘BA-drug couple’ would be efficiently taken up by carrier proteins expressed in the tumour cells. Therefore specific organotropic cytostatic BA derivatives (called ‘Bamets’) were developed to enhance anti-tumour drug delivery and increase tumour sensitivity to chemotherapy [Criado et al. 1997]. Conjugates of cisplatin with glycocholate (Bamet-R2) and with UDCA (Bamet-UD2) have been proved useful in the chemotherapy of experimental models of HCC [Macias et al. 1998; Larena et al. 2002]. More recently, Bamet-UD2 has been shown to inhibit tumour growth in CCA cells expressing ASBT [Lozano et al. 2015]. Results of these experimental models are encouraging but the possibility of using BA derivatives to treat malignancies needs further clinical evaluation.

Conclusion

The landscape of treating chronic liver diseases is rapidly changing largely due to recent advances in unravelling the role of BAs as signalling molecules in metabolic and cholestatic diseases. The understanding of diverse role of BAs in biological pathways is not complete and is continuing to evolve. Novel BA-based pharmaceuticals have raised hope for availability of better therapies in the near future with several compounds already in clinical trials. The current and future development of drugs based on the therapeutic concept of BAs is most likely to benefit patients with PBC, PSC and NASH.