Novel and Emerging Therapies for Cholestatic Liver Diseases

Jordan Goldstein; Cynthia Levy


Liver International. 2018;38(9):1520-1535. 

In This Article

Novel Therapies

Bile Acids: TUDCA and norUDCA

Tauroursodeoxycholate (TUDCA) is the taurine conjugated form of UDCA, which has increased hydrophilicity. TUDCA was evaluated in a non-inferiority phase III trial including 199 Chinese patients with PBC treated over a 24-week period with TUDCA or UDCA. A similar proportion of patients in both groups achieved a 25% or 40% reduction in ALP compared to baseline values. Likewise, rates of study-related adverse events were similar except for itching, which appeared to occur more frequently in UDCA-treated subjects.[37] The study suffered from several limitations including a heterogeneous population, short treatment duration, lack of bioequivalence between doses of TUDCA and UDCA and lack of appropriate instruments to evaluate pruritus and other quality of life measures. Overall, the drug offers no advantage over the well-established use of UDCA.

Twenty four nor-ursodeoxycholic acid (norUDCA) is a side-chain shortened UDCA, which prevents it from undergoing amidation with glycine or taurine.[16] This modification is thought to reinforce the bicarbonate umbrella via cholehepatic shunting. Since this molecule escapes conjugation with glycine or taurine, it is passively reabsorbed within the biliary canaliculi, augmenting the secretion of bile acids and bicarbonate.[10] norUDCA is also more hydrophilic and therefore less toxic to the epithelial cells. A phase II study of 159 patients with PSC treated with placebo vs 500, 1000 or 1500 mg of norUDCA showed that, at 12 weeks, norUDCA reduced ALP levels in a dose-dependent manner.[38] The dose reductions were 12.3%, 17.3%, and 26.0% in the 500, 1000, and 1500 mg groups, respectively, compared to a 1.2% ALP increase in the placebo group. Notably, this effect was independent of disease stage, duration of disease or prior use of UDCA. Adverse events and pruritus occurred at similar rates in all groups. NUC-5, a larger phase III randomized controlled trial with norUDCA, is in progress in Europe (EudraCT Number: 2016-003367-19).

FXR Agonists: LJN-452, GS-9674, EDP-305

In addition to OCA, other FXR agonists are emerging as potential therapies for cholestatic liver diseases. In mouse models, 2 non-steroidal FXR agonists have shown promising results with reduction in liver biochemistries, improvement of inflammatory infiltrates and improvement in fibrosis (Figure 2).[39] These FXR agonists, GS-9674 by Gilead Sciences and Tropifexor (otherwise known as LJN-452) by Novartis, are currently undergoing phase II trials for PBC. Both are non-bile acid formulations, thus expected to cause less pruritus and hyperlipidaemia compared to OCA.[40] In addition, a newer non-bile acid FXR agonist, EDP-305, recently demonstrated a promising safety and tolerability profile in a phase I study including healthy individuals and patients with presumed non-alcoholic fatty liver disease.[41] The drug is now under evaluation for patients with PBC (NCT03394924).

Figure 2.

Bile acid binding to nuclear receptors such as FXR, PPAR, and PXR as well as the surface receptor TGR5 leads to various downstream effects on bile acids, metabolism, and inflammation. BA, bile acid; FXR, farnesoid X receptor; GLP-1, glucagon-like peptide-1; PPAR-α, peroxisome proliferator-activated receptor-α; PXR, pregnane X receptor; TGR-5, transmembrane G protein-coupled receptor 5

A proof-of-concept phase II trial assessing GS-9674 in patients with PSC in an open-label fashion is also ongoing (NCT02943460). All ongoing trials for PBC and PSC are summarized in Table 3 and Table 4.

FGF-19 Analogues: NGM282

As previously mentioned, FGF-19 expression results directly from FXR activation and leads to downregulation of CYP7A1.[42] In a phase II multicenter randomized controlled trial of NGM282, a synthetic analogue of FGF-19 with a modified N terminus, 45 patients with PBC and an incomplete response to UDCA received UDCA plus NGM282 in doses of 0.3 mg or 3.0 mg vs placebo for 28 days. A dose-dependent reduction in ALP and other liver biochemistries was observed in treated patients at day 28: 15.2% in the 0.3 mg group, 19.2% in the 3.0 mg group and 1.2% in the placebo group. The ALP reduction was more pronounced in the subgroups where the baseline ALP was greater than 3× the ULN. Main side effects included diarrhoea, nausea and headaches.[43] NGM282 was also evaluated in a randomized controlled trial including 62 patients with PSC (NCT02704364). A press release note from NGM states that the study did not meet the primary endpoint of a statistically significant change in ALP (

PPAR Agonists: Fibrates, Seladelpar

The nuclear receptor peroxisome proliferator-activated receptor (PPAR) occurs in 3 different isoforms: α, γ and δ. PPAR agonists have different affinities for the various isoforms. For instance, fenofibrate binds PPAR-α with the greatest specificity, bezafibrate binds all 3 subtypes comparably, and seladelpar binds the δ isoform selectively. Furthermore, each isoform has variable expression in different organs, with PPAR-α being predominantly expressed in the liver and other organs participating in fatty acid catabolism, PPAR-γ is found in adipose tissues and immune cells and the δ isoform being ubiquitously expressed.[44,45] PPAR agonists will have different effects depending on which isoforms are activated.

Peroxisome proliferator-activated receptor agonists have a multitude of effects thought to be beneficial in cholestatic diseases. First, they upregulate MDR3, which promotes excretion of phosphatidylcholine in bile causing reduced bile salt cytotoxicity. Furthermore, they modulate expression of MRP2, MDR1, NTCP and ASBT and upregulate sulphotransferase family 2A member 1 (SULT2A1) and UDP glucuronosyltransferase family 1 member A1 (UGT1A1), among other genes which help detoxify bile acids. Fibrates also suppress NF-κB transcription (Figure 2), thus interfering with inflammatory pathways, and cross-react with other receptors, including pregnane X receptor (PXR) and FXR, further enhancing its anticholestatic and anti-inflammatory properties.[16,46]

Fibrates. Several studies evaluated the use of bezafibrate alone or in combination with UDCA for PBC and consistently showed improvement in serum liver biochemistries.[47–56] Typically, these studies included a small sample size (10–30 patients) and were of relatively short duration. The recently completed BEZURSO trial, included a much larger sample size of 100 PBC patients (50 in each group), treated for 2 years with either UDCA/bezafibrate 400 mg/d or UDCA/placebo. Normalization of ALP and normalization of all liver chemistries were seen in 67% and 30% of bezafibrate-treated patients, respectively, compared to none on placebo. Additionally, liver stiffness by transient elastography and the itch scores also improved in treated patients.[57]

Similarly, Reig and colleagues reported their experience in treating 48 patients with PBC and an incomplete response to UDCA with the addition of bezafibrate 400 mg/d for a median of 38 months. Notably, 54% of patients normalized serum ALP. In this group of so-called responders, investigators observed significant improvement in pruritus, jaundice and liver stiffness measured by elastography. Predictors of non-response to bezafibrate included liver stiffness >7.3 kPa and younger age (<53 years). The drug was safe; serious adverse events were not reported; and the kidney function was stable. Side effects were infrequent and included mild gastrointestinal discomfort and myalgias.[58]

The long-term efficacy and safety of bezafibrate should be evaluated further. In that regard, a trial by Hosonuma and colleagues randomized 27 patients with PBC and incomplete response to UDCA to receive UDCA + bezafibrate (400 mg/d) vs UDCA alone for up to 8 years.[59] Despite improvement in serum ALP and Mayo risk score, survival was not better in bezafibrate-treated patients. In addition, patients receiving bezafibrate were more likely to discontinue the study due to renal dysfunction or myalgias. However, one should recognize that this study enrolled only patients with PBC who also had dyslipidaemia; thus, the patient population may not be representative of all PBC patients.

While bezafibrate is not available in the United States, fenofibrate is available. Treatment of PBC patients refractory to UDCA monotherapy with the addition of fenofibrate 160 mg/d led to significant reductions in serum ALP and other liver biochemistries as well as improvement in serum IgM and inflammatory markers. Side effects included mostly heartburn, myalgias and transient transaminase elevation.[60]

In contrast to the growing experience with PPAR agonists in PBC, summarized in Table 5, our understanding of their effect on patients with PSC is just beginning. A Japanese open-label study prospectively evaluated use of bezafibrate 200 mg twice daily for 12 weeks in 11 patients with PSC.[61] Serum ALP and ALT improved in 7 patients (64%). Use of fenofibrate was also evaluated for PSC in 2 small open-label studies from France and the US published only in abstract format, again showing improvement of serum liver biochemistries.[62,63] Additional studies including larger sample size and longer follow-up are warranted.

Other PPAR agonists. Seladelpar is a selective PPAR-δ agonist. It was recently evaluated in a multicenter trial for PBC patients with an incomplete response to UDCA, where addition of seladelpar 50 mg/d or 200 mg/d was compared to placebo. The study showed a significant reduction in ALP of 53% and 63% in the 50 and 200 mg groups, respectively, compared to 2% in the placebo group. However, the trial was stopped early after 1 patient in the 50 mg group and 2 in the 200 mg group developed grade 3 increases in aminotransferases.[64] An ongoing phase II/III trial is testing seladelpar at lower doses of 2, 5 and 10 mg (NCT03301506 & NCT02955602). Interim analyses suggest maintained efficacy at these lower doses, without the previously observed elevation in transaminases.[65]

Elafibranor (GFT505), a dual PPAR-α and PPAR-γ agonist, is thought to have antifibrotic activity and is currently being evaluated in non-alcoholic fatty liver disease and in PBC (NCT03124108).

PXR Agonists: Budesonide

The main role of PXR is in detoxification and metabolism of bile acids via CYP3A4 and SULT2A1, decreasing bile acid synthesis through CYP7A1, and upregulation of MRP3, MRP2 and MDR1 (Figure 2).[16,66] Glucocorticoids are thought to possibly improve liver chemistries and histology in PBC but are not recommended given their extensive side effect profile.[67] However, budesonide, a steroid with greater than 90% first-pass hepatic metabolism, may be a more tolerable alternative. In vitro, use of budesonide in combination with UDCA caused stimulation of the apical chloride/bicarbonate exchanger (AE2), producing a bicarbonate-rich choleresis and strengthening the bicarbonate umbrella.[68] Three small studies have evaluated the use of budesonide in PBC and suggest that budesonide + UDCA may be slightly superior to UDCA alone in non-cirrhotic patients with respect to improvement in inflammation and fibrosis.[67,69,70] Despite the high first-pass metabolism, side effects are still reported after long-term use of budesonide, including bone mass loss, hyperglycaemia and cosmetic changes.[69] Furthermore, increased rates of portal vein thrombosis have been reported in cirrhotic patients receiving budesonide.[71] Results of an European phase III multicentre randomized controlled trial comparing UDCA + budesonide to UDCA + placebo are awaited (Eudra CT number 2007-004040-70).

Budesonide has been previously evaluated in patients with PSC and found not to be beneficial at doses of 3 mg/d or 9 mg/d.[72,73]

TGR5 Agonists: INT-767 & INT-777

Transmembrane G-protein-coupled receptor 5 (TGR5) is a cell surface receptor which is widely distributed in the gastrointestinal tract, especially in the submucosal nervous plexus in the small and large intestine.[74] In the liver, we find TGR5 in the sinusoidal endothelial cells, cholangiocytes and Kupffer cells, but not in the hepatocytes. Bile acid binding to TGR5 results in activation of intracellular cAMP signalling and other cell-specific signalling cascades. Among other effects, there are inhibition of NF-kB pathway, secretion of glucagon-like peptide 1 (GLP-1) and activation of iodothyronine deiodinase, which result in anti-inflammatory properties, improvement of diabetes control and enhancement of basal metabolism.[75] In the biliary epithelial cell, TGR5 also mediates increased bicarbonate secretion through the cystic fibrosis transmembrane conductance regulator (CFTR), thus protecting against bile cytotoxicity. Furthermore, activation of TGR5 has been linked to the known immunosuppressive effect that bile acids exert on immune cells.[74]

In mice, TGR5 was found in dorsal root neurons that mediate itch and pain, and mice overexpressing TGR5 have spontaneous itching. Bile acids can activate TGR5 and induce secretion of neuropeptides that transmit itch and analgesia.[76] Therefore, modulating TGR5 could represent a novel approach to treat itching in cholestatic diseases. Mouse models with overexpression of TGR5 have shown protective effects in 3,5-diethoxycarbonyl-1,4-dihydrocollidine induced sclerosing cholangitis, and likewise, mice with lower levels of TGR5 expression showed worsening fibrosis.[16]

The wide ranging and favourable activity of TGR5 makes it a potential treatment target, with 2 drugs currently being investigated. INT-767, a 23-sulphate derivative of OCA, is an agonist of both FXR and TGR5, and is 3 times as potent an FXR agonist as OCA. INT-767 has been shown to decrease bile acid synthesis, increase bicarbonate excretion and inhibit NF-kB.[77] INT-777, α-ethyl-23(S)-methylcholic acid, is specific for TGR5. Both molecules have shown promising results in animal studies.[78] However, while TGR5 agonists are potentially beneficial to their anti-apoptotic and proliferative effects in cholangiocytes, certain cancerous cells including cholangiocarcinoma tissue have increased expression of TGR5 receptors. Consequently, activation by TGR5 agonists may promote cholangiocarcinoma progression, thus making these drugs less attractive for patients with PSC.[79]

All Trans Retinoic Acid (ATRA)

As FXR, PPAR and PXR all work via forming heterodimers with retinoid X receptor (RXR), with subsequent binding to promoter regions on DNA,[80] retinoic acid has been evaluated as potential therapy for PSC. A pilot study treated 15 patients with PSC who failed UDCA monotherapy with combination UDCA + ATRA (45 mg/m2/d divided into 2 doses) for 12 weeks. Investigators observed a significant decrease in the bile acid precursor C4 as well as in ALT levels, but no significant change was noted in ALP levels. Additionally, 63% and 26% of patients experienced headaches and tinnitus respectively.[81] A phase II study is ongoing to evaluate lower dose ATRA (10 mg twice daily) over 24 weeks in 20 PSC patients (NCT03359174).

Immunomodulatory Therapies

Ustekinumab. Genome-wide association studies found significant associations with the interleukin 12 (IL 12)–signal transducer and activator of transcription (STAT) 4 signalling pathway in PBC.[82] Furthermore, mouse models with deletion of IL-12p40 have marked reduction of pro-inflammatory Th1 cytokines in the liver and less pronounced destructive cholangitis.[83] With that in mind, a phase II open-label trial evaluated the effect of ustekinumab, an IL12/23 monoclonal antibody, in 20 patients with PBC and failed to demonstrate a sizable improvement in ALP after 28 weeks of treatment.[84] Ustekinumab is not being evaluated further for PBC.

Abatacept. Abatacept is a fusion protein of the Fc region of IgG1 and cytotoxic T-lymphocyte antigen 4 (CTLA-4), where CTLA-4 is required for costimulatory activation of T cells via CD80/CD86. In a mouse model of PBC, pretreatment with a CTLA-4 antibody abrogated development of cholangitis, and treatment after cholangitis had already developed led to reduction of the intrahepatic T-cell infiltrates and bile duct damage.[85] A phase II study of abatacept 125 mg weekly for 24 weeks in 20 patients with PBC failed to meet the primary endpoint of either normalization or a decrease >40% in ALP from baseline.[86]

Rituximab. Rituximab is an anti-CD20 monoclonal antibody which has been evaluated in 2 small studies for PBC, unfortunately with only modest effect on serum ALP.[87,88] Nevertheless, the drug was overall safe and well tolerated.[87] While Rituximab may not be beneficial in preventing disease progression, a study was recently completed in the UK exploring the use of rituximab for the treatment of fatigue in PBC (NCT02376335) and results are awaited.

Cenicriviroc. Cenicriviroc (CVC) is an antagonist of chemokine receptor 2 and chemokine receptor 5, which acts by blocking the recruitment of monocytes and macrophages. CVC has shown anti-inflammatory and antifibrotic effects in animal models of fibrosis and NASH.[89] A phase II trial (PERSEUS) evaluating the effect of a 24-week treatment with CVC 150 mg/d in patients with PSC (NCT02653625) was completed December 2017, but results are yet to be published.

Vedoluzimab. The aberrant homing theory asserts that extraintestinal manifestations of IBD, including PSC, are secondary to atypically expressed adhesion molecules. The mucosal addressin cell adhesion molecule-1 (MADCAM-1), usually confined to the gut, can be found in the hepatic endothelium in patients with IBD-associated liver disease. MADCAM-1 is bound by α4β7 integrin (found on T cells) causing migration of gut-primed, mucosally activated T cells to the liver.[90,91] These are long-lived memory cells, which can explain why liver disease can occur years apart from the bowel disease and run a separate disease course.

Vedoluzimab is an anti-α4β7 integrin monoclonal antibody. Conceptually, this drug could prevent abnormal lymphocyte trafficking from the gut to the liver and mitigate inflammation, cell death and the eventual development of fibrosis in PSC. A retrospective study of 34 PSC patients receiving vedoluzimab for concomitant IBD [16 with Crohn's disease (CD) and 8 with ulcerative colitis (UC)] did not demonstrate a decline in liver chemistries or Mayo PSC risk score at 30 weeks. However, 55% and 29% of CD and UC patients, respectively, reached clinical remission of their IBD symptoms.[92]

Manipulation of gut Microbiome

The strong association between PSC and IBD as well as the growing evidence suggesting a role for the microbiome on the pathogenesis of PSC has led many investigators to pursue the use of antibiotics in PSC.[93] Several antibiotics have been tested including vancomycin, metronidazole, minocycline and rifaximin. All but the latter have demonstrated at least some beneficial effect on liver biochemistries.[94–97] Among the antibiotics in use, vancomycin is the most promising.[97] It has been shown to improve liver biochemistries and markers of inflammation, lower the prognostic index mayo risk score (MRS), reduce symptoms and exert immunomodulatory effects on regulatory T cells, especially in the paediatric population without cirrhosis. Vancomycin is also better tolerated, with a significant proportion of patients on metronidazole or minocycline having dropped out of studies due to adverse events. However, long-term use of vancomycin induces drastic changes in the microbiome and prompts concerns about infections with resistant bacteria.[98] Furthermore, nephrotoxicity, neutropaenia and ototoxicity are notorious side effects of vancomycin.

Small randomized controlled trials of both low and high doses of vancomycin for PSC over 12 weeks have shown a decrease in baseline ALP of 18%–43%, improvement of the MRS and improvement in symptoms including pruritus, fatigue, diarrhoea and anorexia.[99,100] These results show promise and larger trials over longer periods are needed to assess the efficacy of vancomycin. Currently, 2 clinical trials looking at vancomycin in patients with PSC are underway (NCT02605213 & NCT01802073).

Finally, in an ongoing open-label trial examining the safety and efficacy of faecal microbiota transplant in 6 PSC patients, 5 having concomitant ulcerative colitis, an interim analysis showed that 50% reached the primary endpoint of a decline in ALP ≥50% from baseline, without reported adverse events (NCT02424175).[101]

Antifibrotic Therapies: Simtuzumab

Lysyl oxidase-like 2 is a copper-dependent amine oxidase which plays an important role in fibrosis by catalyzing the cross links in both collagen and elastin. Simtuzumab is an IgG4 monoclonal antibody directed against lysyl oxidase like-2 (LOXL2). In a phase 2b study which included 234 PSC patients (51% of patients at baseline had bridging fibrosis or cirrhosis), simtuzumab was administered at dosages of 75 mg or 125 mg vs placebo over 96 weeks. After 96 weeks, there was no significant difference in hepatic collagen content, Ishak fibrosis stage, decrease in ALP or adverse event rate compared to placebo. Despite simtuzumab being well tolerated, there was no overall benefit in patients with PSC and there are no other clinical trials evaluating this drug for PBC or PSC.[102]