Obeticholic acid, a farnesoid X receptor agonist, improves portal hypertension by two distinct pathways in cirrhotic rats


  • See Editorial on Page 2072

  • Potential conflict of interest: Nothing to report.

  • L.V. is an aspirant researcher and was granted a Ph.D. fellowship by the Fund for Scientific Research–Flanders (FWO Vlaanderen). R.F. was granted a post-Ph.D. fellowship by the Fund for Scientific Research–Flanders (FWO Vlaanderen). J.T. was supported by grants from the Deutsche Forschungsgemeinschaft (SFB TRR57 project 18). T.V. is an aspirant researcher and was granted a Ph.D. fellowship by the Fund for Scientific Research–Flanders (FWO Vlaanderen). W.L. and F.N. are senior clinical investigators for the Fund for Scientific Research–Flanders (FWO Vlaanderen).


The farnesoid X receptor (FXR) is a nuclear bile acid receptor involved in bile acid homeostasis, hepatic and intestinal inflammation, liver fibrosis, and cardiovascular disease. We studied the effect of short-term treatment with obeticholic acid (INT-747), a potent selective FXR agonist, on intrahepatic hemodynamic dysfunction and signaling pathways in different rat models of cirrhotic portal hypertension (PHT). For this, thioacetamide (TAA)-intoxicated and bile-duct–ligated (BDL) rats were used as models. After gavage of two doses of 30 mg/kg of INT-747 or vehicle within 24 hours, in vivo hemodynamics were assessed. Additionally, we evaluated the direct effect of INT-747 on total intrahepatic vascular resistance (IHVR) and intrahepatic vascular tone (endothelial dysfunction and hyperresponsiveness to methoxamine) by means of an in situ liver perfusion system and on hepatic stellate cell contraction in vitro. FXR expression and involved intrahepatic vasoactive pathways (e.g., endothelial nitric oxide synthase [eNOS], Rho-kinase, and dimethylarginine dimethylaminohydrolase [DDAH]) were analyzed by immunohistochemistry, reverse-transcriptase polymerase chain reaction, or western blotting. In both cirrhotic models, FXR expression was decreased. Treatment with INT-747 in TAA and BDL reactivated the FXR downstream signaling pathway and decreased portal pressure by lowering total IHVR without deleterious systemic hypotension. In the perfused TAA and BDL cirrhotic liver, INT-747 improved endothelial vasorelaxation capacity, but not hyperresponsiveness. In both groups, this was associated with an increased eNOS activity, which, in TAA, related to down-regulation of Rho-kinase and in BDL to up-regulation of DDAH-2. Conclusion: FXR agonist INT-747 improves PHT in two different rat models of cirrhosis by decreasing IHVR. This hemodynamic effect relates to increased intrahepatic eNOS activity by pathways that differ depending on the etiology of cirrhosis. (Hepatology 2014;59:2286–2298)






asymmetric dimethylarginine


bile duct ligated


cholesterol 7α-hydroxylase


dimethylarginine dimethylaminohydrolase


enzyme-linked immunosorbent assay


fetal calf serum


farnesoid X receptor


hypoxanthine-guanine phosphoribosyltransferase


hepatic venous pressure gradient


hepatic stellate cell




intrahepatic vascular resistance


obeticholic acid


isosorbide mononitrate


National Center for Biotechnology Information


mean arterial pressure


myosin light-chain phosphatase


nitric oxide


(phosphorylated) endothelial nitric oxide synthase


(phosphorylated) moesin


(phosphorylated) vasodilator-associated protein


portal hypertension


protein kinase G


portal pressure


reverse-transcriptase polymerase chain reaction


standard error of the mean


small heterodimer partner


alpha-smooth muscle actin




vasodilator-stimulated protein


vascular endothelial cells

Portal hypertension (PHT) is the driving force behind many of the lethal complications of cirrhosis, such as gastroesophageal variceal bleeding, hepatic encephalopathy, ascites, hepatorenal syndrome, and bacterial infections.[1-4] Despite its effect on morbidity and mortality, current pharmacotherapy remains almost exclusively restricted to the use of nonselective beta-adrenergic blockers. These drugs indisputably lower PHT by decreasing portal inflow,[5-9] but are hampered by insurmountable side effects and an insufficient hemodynamic response in 10%-15% and 60% of treated patients, respectively.[1, 10] In search of expansion of the therapeutic armamentarium, the research focus has shifted beyond splanchnic vasoconstrictors toward potential modification of increased intrahepatic vascular resistance (IHVR) because this not only addresses PHT as such, but might also improve liver perfusion and thereby liver function.[1, 11, 12] Most scientific interest has been dedicated to the functional component of IHVR, which appears more amenable to pharmacological intervention. This increased intrahepatic vascular tone is driven by activated hypercontractile hepatic stellate cells (HSCs)[13, 14] subjected to an impaired intrahepatic vasodilator capacity resulting from a reduced endothelial nitric oxide (NO) synthase (eNOS) activity[15-18] combined with decreased NO bioavailability[19] and an exaggerated response to vasoconstrictors[20, 21] mediated through up-regulated intracellular Rho-kinase–dependent pathways.[22]

In recent years, the farnesoid X receptor (FXR; NR1H4), a bile-acid–responsive transcription factor and member of the nuclear hormone receptor superfamily, has been found to be crucial in numerous hepatobiliary[23-25] and gastrointestinal diseases.[26] FXR is highly expressed in the normal liver, kidney, and intestine.[27] In normal liver, FXR has been shown to be a master transcriptional regulator of genes involved in bile acid homeostasis, but also a moderator of lipid and carbohydrate metabolism.[23-25, 27] More interesting from the perspective of targeting PHT, FXR has also been implicated in vasoregulation, because FXR agonism could prevent HSC activation[25] and endothelin-induced contraction[28] in vitro. However, this former finding was opposed by failure to detect physiologically important levels of FXR expression in both murine and human HSC by others.[29] Additionally, FXR was also shown to be involved in regulation of eNOS because it not only appeared to induce eNOS expression at the transcriptional level in aortic vascular endothelial cells (VECs),[30] but also was capable of decreasing systemic levels of asymmetric dimethylarginine (ADMA), a known endogenous NOS inhibitor, in Zucker diabetic fatty rats.[31] We recently described ADMA to be involved in intrahepatic endothelial dysfunction in bile-duct–ligated (BDL) cirrhotic rats.[18] Finally, preliminary data in an uncontrolled series of alcoholic patients with cirrhosis with established PHT suggest an improvement in hepatic venous pressure gradient (HVPG) after 7 days of treatment with the FXR agonist, obeticholic acid (INT-747).[32]

Given this background, we aimed to study the effects of INT-747, a highly potent and selective FXR receptor agonist on PHT and to investigate its exact effect on intrahepatic hemodynamic dysfunction and involved signaling pathways in two different rat models of cirrhosis.

Materials and Methods

Animal Models

Male Wistar rats (Janvier, France), weighing 200-250 g, were divided in three experimental groups. The first group served as untreated healthy controls (n = 24). The second group was administered weekly adjusted doses of thioacetamide (TAA) in their drinking water for 18 weeks, as previously described.[14, 18, 21, 33] At 18 weeks, when the experiments were performed, all animals (n = 64) had macroscopic and histology-proven toxic cirrhosis. The third group underwent double ligation of the common bile duct (BDL). After 4 weeks, animals (n = 53) were sacrificed. All had developed biliary cirrhosis, as confirmed on histology. For the in vitro experiments, normal controls were used for isolation of HSCs. All animal experiments were performed according to guidelines of the local ethics committee. In all experiments, rats from both the TAA and BDL groups were randomly allocated to either a control group receiving vehicle or a treatment group receiving the FXR agonist, INT-747 (kindly provided by Intercept Pharmaceuticals, Inc., New York, NY). The study protocol consisted of two doses of 30 mg/kg of INT-747 or vehicle by gavage 24 and 4 hours before hemodynamic measurements or collection of samples. Sustainability of response was assessed over a 10-day treatment period in 7 TAA rats receiving 30 mg of INT-747 daily and 6 BDL rats receiving 5 mg/kg every 2 days, based on our previous dose-safety studies. For gavage, the INT-747 compound was dissolved in 0.75-1.0 mL of freshly prepared methylcellulose (1%), with equal volumes of vehicle administered to controls.

In Vivo Experiments

Measurement of portal pressure (PP) and mean arterial pressure (MAP; Dataq Instruments Akron, OH) was carried out by cannulation of the carotid artery and portal vein (Insyte-W 24 Gauge; Becton Dickinson Benelux nv, Erembodegem, Belgium). Blood samples were collected in heparinized tubes (Becton Dickinson Vacutainer; Becton Dickinson Benelux nv) by puncture of aortic bifurcation. Tissue samples were snap-frozen for molecular analyses (cfr infra).

Reverse-Transcriptase Polymerase Chain Reaction of FXR, Its Downstream Signaling Molecule, Small Heterodimer Partner, and the HSC Hypercontractability Marker, Alpha-Smooth Muscle Actin

Relative expression of FXR, small heterodimer partner (SHP), and alpha-smooth muscle actin (α-SMA) was assessed by quantative reverse-transcriptase polymerase chain reaction (RT-PCR). For this, total RNA from culture-activated rat primary HSCs and liver samples from both treated and untreated animals from the different groups was extracted in TRIzol reagent (N.V. Invitrogen SA, Merelbeke, Belgium) and reverse transcribed. Subsequently, complementary DNA was amplified by PCR using the following sense and anti-sense primers: α-SMA: 5′-TCCCAGCACCATGAAGATCAA-3′ and 3′-AACATTCACAGTTGTGTGCTAGA-5′; FXR: 5′-CATTAACAACGCGCTCACCTG-3′ and 3′-TTCCTTAGCCGGCAATCCTG-5′; SHP: 5′-CTTGAGCTGGGTCCCAAGGA-3′ and 3′-CTAGCTGGGTACCAGGGCTC-5′. Primers were rat specific and designed using sequence data and Nucleotide BLAST software from the National Center for Biotechnology Information (NCBI; Bethesda, MD) database. Primers were subsequently manufactured by TIB MOLBIOL GmbH (Berlin, Germany). All samples, including negative control from rat brain, were run in duplicates with LightCycler 480 equipment (Roche Applied Science, Penzberg, Germany). Samples were normalized to hypoxanthine-guanine phosphoribosyltransferase (HPRT) housekeeping gene and fold-change was calculated with respect to the control group using the 2-ΔΔCt method, as previously described.[34]

Immunohistochemistry for FXR

Fresh liver samples from untreated controls, TAA, and BDL rats were snap-frozen in liquid-nitrogen–cooled isopentane and stored at −80°C. Four-micrometer-thick frozen sections were cut, dried overnight at room temperature, fixed in acetone, and then rinsed in wash buffer (EnVision FLEX [20×]; DakoCytomation A/S, Glostrup, Denmark) before use. Frozen rat liver sections were incubated for 30 minutes at room temperature with a 1:20 dilution of FXR primary antibody (Ab [H130]; Santa Cruz Biotechnology, Santa Cruz, CA). The second and third steps consisted of peroxidase-labeled rabbit anti-mouse and peroxidase-labeled swine anti-rabbit immunoglobulins (both from DakoCytomation A/S) for rat liver. Secondary and tertiary Abs were diluted (1:50 and 1:100, respectively) in Antibody diluent (Dako), containing 10% normal human serum. Negative controls consisted of omission of the primary Ab. FXR staining on culture-activated HSCs was performed after acetone fixation on glass Lab-Tek chamber slides (Thermo Fisher Scientific, Inc., Rockford, IL).

In Situ Liver Perfusion Study

To evaluate the intrahepatic vascular effect of INT-747 (30 mg/kg by gavage 24 and 4 hours before measurements) on total IHVR, vascular hyperresponsiveness, and endothelial dysfunction, a flow-controlled liver perfusion system was applied in both the TAA and the BDL models, as previously described.[18, 21, 33] To evaluate the effect of treatment on total IHVR, flow rates were increased by 5 mL/min each 5 minutes starting at a 30-mL/min flow rate comparing subsequent elevations in perfusion pressure. To evaluate vascular responsiveness to vasoconstrictors, methoxamine (10−4 M) was added to the perfusate after the stabilization period at a 35-mL/min flow rate. Increase in perfusion pressure was then evaluated after 5 minutes of exposure. Endothelial function was assessed by performing dose-response curves to increasing concentrations of acetylcholine (Ach) of 10−7, 10−6, and 10−5 M every 1.5 minutes after a 5-minute incubation period with methoxamine (10−4 M) as previously described.[18, 21, 33]

HSC Isolation and In Vitro Contraction Assay

Primary HSCs were isolated by in situ proteinase/collagenase perfusion, followed by density centrifugation, as previously described.[14] Viability was routinely above 95%. Cells were cultured in a William's E medium enriched with 10% inactivated fetal calf serum (FCS) and 2 mM of glutamin, containing 0.6 IU/mL of insulin and 1% antibiotic/antimycotic solution (N.V. Invitrogen SA). Characterization of rat liver-derived myofibroblast-like cultures, established by culturing enriched HSC fractions on plastic, was performed by staining with anti-α-SMA, as previously described.[14] Collagen lattice contraction experiments were performed according to a previously described protocol.[14] After 24 hours, a dose-effect experiment was performed by exchanging the medium for medium containing increasing concentrations of INT-747, ranging from 10−3 to 102 μM and with FCS (10%) as a procontractile agonist. Full medium without INT-747 and cell-free gels served as positive and negative controls, respectively. Gels were then dislodged from the well surface to allow contraction during the next 24 hours. Concentrations of INT-747 and dimethyl sulfoxide were kept below 103 μM and 0.5%, respectively, to avoid toxic interference, as assessed by XTT assay (Roche Diagnostics, Vilvoorde, Belgium). Change in gel surface, with control cell-free gels as estimates for the precontraction surface, was used to calculate contraction (expressed as percent change in surface). HSC hypercontractability was additionally assessed by means of α-SMA expression after 24 hours of incubation with either vehicle or 1 or 100 μM of INT-747 in culture-activated HSCs. All in vitro experiments were performed in triplicates using culture-activated HSCs from three different rat HSC isolations.

Western Blotting of Considered Involved Intrahepatic Vasoactive Pathways and Determination of ADMA Levels

Liver samples were snap-frozen and homogenized in buffer containing 25 mM of Tris/HCl, 5 mM of ethylenediaminetetraacetic acid, 10 μM of phenylmethanesulfonyl fluoride, 1 mM of benzamidine, and 10 μg/mL of leupeptin. Protein concentrations were assessed with a Dc-Assay kit (Bio-Rad Laboratories GmbH, Munich, Germany). Samples (20 μg of protein/lane) were subjected to sodium dodecyl sulfate polyacrylamide electrophoresis (15% gels for RhoA; 8% gels for Rho-kinase, eNOS, and phosporylated-eNOS [P-eNOS]; and 10% gels for moesin, phosporylated moesin (p-moesin), DDAH-1 and −2), vasodilator-stimulated protein [VASP] and phosporylated-VASP [P-VASP]. Proteins were blotted on nitrocellulose membranes. The membranes were blocked and incubated with the primary Abs: Rock-2 H-85; NOS3; moesin; p-moesin; VASP (Santa Cruz Biotechnology); p-eNOS (Ser 1177; Cell Signaling Technology, Danvers, MA); p-VASP (16C2; Calbiochem, San Diego, CA); and DDAH-1 (H-70) and DDAH-2 (H-85; Tebu-Bio, Boechout, Belgium). Thereafter, the membranes were incubated with the corresponding secondary peroxidase-coupled Abs (Calbiochem). After enhanced chemiluminescence (Amersham, Bucks, UK), digital detection was evaluated using Chemi-Smart (PeqLab Biotechnologie GmbH, Erlangen, Germany).[22, 35] Blood-plasma ADMA levels were assessed by competitive enzyme-linked immunosorbent assay (ELISA; Enzo Life Sciences BVBA, Zandhoven, Belgium).

Statistical Analysis

Statistical data were subjected to equal variance and normality testing. Data were all normally distributed. As such, data were expressed as means ± standard error of means (SEM) and subjected to parametrical statistics. Comparing two unpaired groups, the Student t test was applied; for comparison of multiple groups, a one-way analysis of variance was applied. If positive, a subsequent pair-wise comparison was performed by means of Bonferroni t testing. P values below 0.05 were considered statistically significant. Statistical analysis was performed by means of SigmaStat (edition 3.5 [2006]; Systat Software Inc., Chicago, IL) and GraphPad Prism software (version 6.02 [2013]; GraphPad Software Inc., San Diego, CA).


The FXR Pathway Is Down-Regulated in Both Toxic and Biliary Experimental Cirrhosis, but Can Be Reactivated by Means of INT-747

Basal expression of FXR was 2.8 times lower in vehicle-treated TAA and 8.6 times lower in vehicle-treated BDL rats, compared to untreated healthy controls (Fig. 1A; n = 8 per group; P < 0.001 in both groups). After 24 hours of treatment with the FXR agonist, INT-747, hepatic FXR expression was not significantly altered in liver-diseased groups, compared to their untreated cirrhotic counterparts (n = 8 in all groups; P = not significant). Inversely, INT-747 administration induced a 3.9-fold increase in its downstream effector, SHP, in TAA rats and and an 8.3-fold SHP increase in BDL rats (Fig. 1A; TAA vs. TAA+INT-747 and BDL vs. BDL+INT-747; n = 8 per group; P < 0.001) implying important reactivation of the downstream FXR-signaling pathway upon INT-747 treatment. Immunohistochemical (IHC) staining comfirmed an important decrease of the typical hepatocytic nuclear FXR receptor staining pattern in all untreated TAA and BDL rats, compared to healthy controls (Fig. 1B; n = 5 in all groups), corroborating the formentioned RT-PCR data in untreated animals on the protein level.

Figure 1.

(A) RT-PCR data showing significantly reduced hepatic FXR expression in both TAA and BDL rats (***P < 0.001 in both groups vs. CTRL; n = 8 in all groups). After treatment with the FXR agonist, obeticholic acid (INT-747), there was no significant difference in FXR expression. The FXR downstream target receptor, SHP, is significantly more expressed in both TAA and BDL rats upon obeticholid acid treatment, compared to untreated animals (***P < 0.001 in TAA vs. TAA+INT-747 and BDL vs BDL+INT-747; n = 8 in all groups). (B) Comparative IHC FXR staining showing a decrease below detection level of the typical nuclear staininig pattern in both TAA and BDL rats (n = 5 in all groups). Data are indicated as mean ± SEM.

Treatment With the FXR Agonist, INT-747, Reduces PP in the TAA and BDL Models Without Systemic Hemodynamic Impact

As expected, PP was significantly increased in the vehicle-treated cirrhotic rat models, whereas MAPs were lower, compared to healthy control animals, indicating the presence of the hyperdynamic circulatory state (Table 1). In contrast, treatment with two doses of INT-747 caused a decrease of PP in both the TAA and BDL models (13.9 ± 0.6 vs. 11.6 ± 0.8 mmHg; n = 13 and n = 7, P < 0.05; and 11.9 ± 0.6 vs. 10.1 ± 0.6 mmHg; n = 9 in both groups, P = 0.038, respectively). MAP was not affected by FXR agonist treatment, suggesting a primary beneficial intrahepatic effect. In 7 TAA and 6 BDL rats receiving 10 days of INT-747 treatment, PP was sustainably decreased, compared to their vehicle-treated cirrhotic counterparts (13.9 ± 0.6 vs. 11.9 ± 0.4 mmHg; P < 0.05; and 11.9 ± 0.6 vs. 9.4 ± 0.5 mmHg; P < 0.01, respectively), again without any adverse systemic hemodynamic effect.

Table 1. Hemodynamic and Morphological Features Among Different Experimental Rat Groups
 Ctrl (n = 5)TAA (n = 13)TAA+INT-747 (n = 7)P Value TAA vs. TAA+INT-747BDL (n = 9)BDL+ INT-747 (n = 9)P Value BDL vs. BDL+INT-747
Body weight (g)412 ± 7336 ± 15328 ± 200.84334 ± 16363 ± 130.17
MAP (mm Hg)130 ± 1268.0 ± 3.666.7 ± 7.80.8754.0 ± 4.157.6 ± 5.40.61
PP (mm Hg)5.6 ± 0.413.9 ± 0.611.6 ± 0.80.04111.9 ± 0.610.1 ± 0.60.038
IHVR (mmHg.min.100g/mL)0.476 ± 0.0050.811 ± 0.0110.678 ± 0.018<0.0012.009 ± 0.0741.548 ± 0.035<0.001

Treatment With the FXR Agonist, INT-747, Reduces the Total IHVR in Experimental Cirrhosis

Both the TAA and BDL cirrhotic rats showed a significantly higher perfusion pressure both in general and at each flow rate, compared to control animals, confirming the increased total IHVR typical of cirrhotic PHT (P < 0.001). After treatment with INT-747, perfusion pressure decreased at each flow rate in both the TAA and BDL model (P < 0.001 in TAA vs. TAA+INT-747; P < 0.001 in BDL vs. BDL+INT-747; n = 5 in all groups), indicating a decrease in total IHVR by INT-747 treatment (Fig. 2A,B).

Figure 2.

The FXR receptor agonist, obeticholic acid (INT-747), induces a significant decrease in perfusion pressure at each flow rate in both the TAA (A) and BDL (B) cirrhotic models (*P < 0.001 in TAA vs. TAA+INT-747 and BDL vs. BDL+INT-747; n = 5 in all groups). Data are indicated as mean ± SEM.

Treatment With the FXR Agonist, INT-747, Restores Endothelial Function in Experimental Cirrhosis

In the normal control liver, a dose-dependent intrahepatic vasorelaxation could be achieved with increasing concentrations of ACh compatible with normal vasoactive endothelial function. In both the TAA and BDL models, there was an impaired vasorelexation at any given dose of ACh compatible with endothelial dysfunction.[18, 21] In both TAA (Fig. 3A) and BDL (Fig. 3B) rats, after two doses of INT-747, a significant decrease in perfusion pressure could be observed at each given concentration of ACh.

Figure 3.

(A and B) In the healthy control group, a significant decrease in perfusion pressure was observed at all ACh concentrations ranging from 10−7 to 10−5 M after preincubation with methoxamine 10−5 M (P < 0.001). In the TAA+vehicle and BDL+vehicle groups, there was no significant change in perfusion pressure in response to increasing concentrations of ACh in the perfusate compatible with endothelial dysfunction. After two-dose gavage with 30 mg/kg of obeticholic acid (INT-747), endothelial dysfunction was restored fully in the TAA+INT-747 group and partially in the BDL+INT-747 group (P < 0.003 and P < 0.05 in the TAA+vehicle vs. TAA+INT-747 and BDL+vehicle vs. BDL+INT-747, respectively, at all ACh concentrations; n = 5 in all groups except BDL+INT-747: n = 4). Data are indicated as mean ± SEM.

Treatment With the FXR Agonist, INT-747, Does Not Affect Vascular Hyperresponsiveness

An exaggerated intrahepatic vascular response to vasoconstrictors such as methoxamine is characteristic in PHT, as previously described.[20-22] This vascular hyperresponsiveness could be reproduced in both the BDL and TAA models, but was not significantly affected by gavage with INT-747 (data not shown).

Culture-Activated HSCs Only Marginally Express FXR and SHP and Show Limited Functional Response Upon FXR Agonist Exposure

On the messenger RNA level, both FXR and its downstream effector, SHP, appeared only marginally expressed in isolated primary HSCs, in comparison to full healthy liver tissue (Fig. 4A). On the protein level, qualitative assessment of FXR expression by means of IHC was below detection level in culture-activated HSCs (Fig. 4B). Functionally, a small decrease of only 10% in the in vitro HSC contraction capacity could be observed, after the addition of INT-747 to the culture medium in the dose ranges of 100, 10, and 1 μM, compared to vehicle alone (−50.1 ± 3.5% [n = 5], −52.3 ± 3.1% [n = 10], −52.8 ± 2.5% [n = 13] vs. −60.7 ± 2.1% [n = 13], respectively; all P < 0.03). At lower concentrations (0.1, 0.01, and 0.001 μM), HSC contraction was unaffected (Fig. 4C). SHP and α-SMA expression in culture-activated HSCs remained unaffected after 24 hours of in vitro stimulation with 1 or 100 μM of INT-747 (Fig. 4D).

Figure 4.

(A) RT-PCR showing marginal expression of both FXR and SHP in isolated HSCs, compared to healthy control liver tissue. (B) Absence of clear FXR staining positivity on culture-activated primary rat HSCs. (C) FXR receptor agonist obeticholic acid (INT-747) reduces FSC-induced HSC gel contraction at concentrations of 1, 10, and 100 μM (*P < 0.03 at all concentrations). (D) RT-PCR data showing nonsignificant differences in SHP and α-SMA expression, compared to HPRT after 24-hour incubation of cultured HSCs with 1 and 100 μM of INT-747.

The FXR-Agonist–Induced Intrahepatic Hemodynamic Effects Relate to Increased eNOS-Activation, but Differ in Mechanism Depending on the Etiology of Cirrhosis

In both cirrhotic models (TAA and BDL), there was a significant down-regulation of P-eNOS, indicative of the considered fundamental reduced intrahepatic eNOS activity[15-18] typical of cirrhosis (P < 0.05 vs. healthy control; Fig. 5A). After treatment with INT-747, P-eNOS levels were restored back to the level of healthy controls for TAA (P < 0.05; Fig. 5A) and to above these latter for BDL (P < 0.005; Fig. 5A). As a logical consequence, levels of P-VASP, a key downstream target of endothelium-derived NO, were elevated in both TAA and BDL rats after treatment, compared to their untreated cirrhotic counterparts (P < 0.05 and P < 0.001, respectively, data not shown). Given the restoration of eNOS activity by INT-747, we next evaluated its effect on two considered pivotal inhibitory pathways on eNOS function: the RhoA/Rho-kinase pathway[36, 37] and DDAH-2. DDAH-2 is a key catabolic enzyme of ADMA, a major endogenous NOS inhibitor.[18] In both the TAA and BDL models, there was a significant up-regulation and activation of the Rho-kinase pathway, as shown by the increased phosphorylation of its substrate, moesin (P < 0.001 in both TAA and BDL; Fig. 5A). However, in the TAA model, there was a significant decrease of P-moesin in the TAA+INT-747 vs. TAA group, indicating a reduced Rho kinase activity after FXR agonist treatment (P < 0.05; Fig. 5A). In contrast, in the BDL model, P-moesin levels remained up-regulated irrespective of treatment, compared to healthy controls (P < 0.001; Fig. 5A). With regard to DDAH, both isoforms were only slightly decreased in TAA (DDAH-1 and −2) and showed no change after INT-747-treatment. In contrast, in BDL, DDAH-1 and −2 were strongly decreased, of which DDAH-2 levels were restored to normal values after FXR agonism with INT-747 (P < 0.02; Fig. 5A). Consistent with these findings, blood-plasma ADMA levels were increased in both vehicle-treated TAA and BDL rats, compared to healthy controls (0.32 ± 0.03 and 0.63 ± 0.04 vs. 0.24 ± 0.02 μmol/L, respectively; n = 8 in all groups, P < 0.001). However, only in BDL rats was a significant ADMA level decrease observed after INT-747 treatment (0.63 ± 0.04 vs. 0.54 ± 0.01 μmol/L; n = 8 in both groups; P < 0.05).

Figure 5.

(A) Western blotting analysis showing significant increase in P-eNOS protein levels in hepatic tissue of both TAA and BDL rats after two-dose gavage of 30 mg/kg of the FXR agonist, obeticholic acid (INT-747). In both TAA and BDL rats, there was an increase in P-moesin expression suggestive of increased Rho kinase activity. This was partially reversible after INT-747 gavage in the TAA model, but not in the BDL model. In turn, in the BDL rats, there was an increased DDAH-2 expression after FXR stimulation, which was not present in the TAA model. (B) Both in the BDL and TAA models after FXR agonist gavage, a common NO pathway through P-eNOS and subsequent PKG was activated, resulting in increased MLCP activity and subsequent HSC relaxation. In the TAA model, this P-eNOS increase was related to deactivation of the Rho-kinase pathway after INT-747 gavage. In the BDL model, increased P-eNOS was associated with up-regulated DDAH-2, a known ADMA inhibitor. Data are indicated as mean ± SEM.


The central homeostatic regulating role of FXR, a ligand-activated nuclear receptor belonging to the nuclear receptor superfamily, is receiving growing interest.[23-31, 38, 39] Besides its considered traditional function as a chief regulator of bile acid homeostasis, the FXR system has also been implicated in other physiological events, including regulation of lipid and carbohydrate metabolism, liver repair, and hepatoprotection.[23, 24, 27, 40, 41] As a result, FXR dysfunction has been suggested to play a crucial role not only in cholestatic disease, but also nonalcoholic fatty liver disease, impaired liver regeneration, liver fibrosis, and hepatocellular carcinoma.[40-44] The apparent cardinal role of FXR dysfunction and the unmet medical need in these different fields has led to the development of a first-in-class FXR agonist, 6-alpha-ethyl-chenodeoxycholic acid or INT-747, a semisynthetic bile acid derivative with a 50% effective concentration of 99 nM, which is 100 times more potent than its endogenous ligand, chenodeoxycholic acid.[45, 46] Endowed with anticholeretic, metabolic regulatory, anti-inflammatory, and, potentially, antifibrotic effects, FXR agonists are increasingly also suggested to be involved in vascular homeostasis, in particular, in regulation of eNOS function.[28, 30, 31, 47]

Decreased intrahepatic eNOS activity is considered an irrefutable dogma in the pathogenesis of increased IHVR in cirrhotic PHT.[48] Current innovative treatment strategies attempt to reduce PP in cirrhosis by increasing hepatic NO availability either by substitution liver-specific NO donors,[21] blocking inhibitory pathways (e.g., Rho) through statins,[37] improving NO bioavailability through decreased scavenging by superoxide,[49] or adenoviral hepatic gene transfer of eNOS[17] or eNOS-regulating genes.[22] However, all these treatment strategies either have been shown to insufficiently decrease PP, further aggravate systemic hypotension, or are not applicable in a clinical setting. In current clinical practice, the nonselective NO donor, isosorbide mononitrate (ISMN), is the only drug in this context that somewhat lives up to the expected standard, but only in combination with beta-adrenergic blockers and solely in the prevention of variceal rebleeding for patients noneligible for standard endoscopic or pharmacological intervention.[50, 51] However, concerns remain about its effectiveness and the occurrence of adverse events.[51] Our present study substantiates the hypothesis that FXR deficiency is involved in cirrhotic PHT and, more specifically, the increased IHVR, which is considered the primary factor in the etiopathogenesis of cirrhotic PHT. Not only did we demonstrate defectiveness of the FXR system in a cholestatic and noncholestatic rat model of cirrhosis (BDL and TAA, respectively), we also showed that functional restoration of FXR activity in these models by means of short-term oral treatment with INT-747 improved PHT by decreasing IHVR and without deleterious systemic hemodynamic effects. Although we demonstrated the sustainability of PP reduction in both models during 10 days of treatment, the short-term treatment approach allowed us to exclusively study the potential hemodynamic effects of FXR agonism in cirrhosis apart from possible long-term antifibrotic, -cholestatic, and -inflammatory effects on the liver. Functional reactivation of the FXR pathway was demonstrated by an important increase in the FXR direct target gene, SHP, expression after FXR agonist stimulation in both models of cirrhosis. SHP is an FXR downstream orphan nuclear receptor that inhibits numerous other nuclear receptors and cholesterol 7α-hydroxylase (CYP7A1), among others.[52] The obtained beneficial effects relate hemodynamically to correction of intrahepatic endothelial dysfunction and molecularly to restoration of impaired intrahepatic eNOS activity, which—according to the dogma mentioned earlier[48]—were highly dysfunctional in both the vehicle-treated TAA and BDL group models, as previously described.[18, 33, 37] Consistent with correction of impaired eNOS activity, we observed a significant increase in phosphorylated eNOS in both TAA and BDL rats. Activation of the NO-dependent pathway was further substantiated downstream by an increased phosphorylation of VASP (vasodilator-stimulated phosphoprotein), indicating increased PKG (protein kinase G) activity, directly influencing myosin light-chain phosphatase (MLCP) activity, which finally results in HSC relaxation.

Even more interesting, the pathways leading to improved eNOS activity differ depending on the etiology of cirrhosis. In the BDL model, but not in TAA, the increase in P-eNOS after INT-747 treatment was related to a decrease of ADMA blood levels and an increase of DDAH-2, an enzyme responsible for the degradation of ADMA, as we have demonstrated earlier in BDL.[18, 53] ADMA is a competitive inhibitor of the eNOS substrate, L-arginine, and has been shown recently to also reduce eNOS phosphorylation in VECs in vitro and in vivo.[54] Similar beneficial effects of FXR agonism on systemic ADMA metabolism have been described in Zucker diabetic fatty rats and insulin-resistant high-salt–fed Dahl Rats.[31, 55] Conversely, in TAA, but not in BDL, the Rho-kinase pathway was significantly down-regulated after INT-747 treatment, which might explain reconstitution of eNOS activity by FXR agonism in this particular model.[37] Although we could not assess the reason for this differential regulation in both models, the downstream effects prove the same and the overall hemodynamic effects remain comparable. Subsequently, we assessed whether INT-747 could also exert an endothelium-independent direct effect on culture-activated HSCs, the cellular effectors of increased intrahepatic vascular tone. For this purpose, we performed an in vitro collagen lattice contraction assay, in which we could confirm only a marginal loss of contractility at the higher concentrations of INT-747, as described by Li et al.[28] However, on the molecular level, this did not coincide with increased SHP expression nor with decreased expression of the HSC activation marker, α-SMA. Furthermore, we and others[29] found a very low expression level of FXR and SHP in culture-activated HSCs, which makes a significant FXR-mediated effect on HSCs unlikely. Contrarily, given the major beneficial effects of INT-747 on endothelial function and hepatic P-eNOS levels, the observed decrease in IHVR is likely to be confined to FXR-related endothelial effects, as suggested by other researchers in different experimental conditions.[30, 31]

In summary, we have shown that FXR is defective in two experimental rat models of cirrhosis and that restoration of its activity by INT-747, an FXR agonist, improved PHT in both models by decreasing IHVR and without deleterious systemic hemodynamic effects. This hemodynamic effect relates to correction of eNOS dysfunction by two different mechanisms. Given their previously reported antifibrotic and -cholestatic effects in addition to now substantiated portal hypotensive potential, FXR agonists might play an important role in the future treatment of chronic liver disease.[56]


The authors thank David Shapiro and Luciano Adorini (Intercept Pharmaceuticals) for their kind provision of the obeticholic acid compound.