Carvedilol impairs bile acid homeostasis in mice: implication for nonalcoholic steatohepatitis

Abstract

Carvedilol is a widely used beta-adrenoreceptor antagonist for multiple cardiovascular indications; however, it may induce cholestasis in patients, but the mechanism for this effect is unclear. Carvedilol also prevents the development of various forms of experimental liver injury, but its effect on nonalcoholic steatohepatitis (NASH) is largely unknown. In this study, we determined the effect of carvedilol (10 mg/kg/day p.o.) on bile formation and bile acid (BA) turnover in male C57BL/6 mice consuming either a chow diet or a western-type NASH-inducing diet. BAs were profiled by liquid chromatography-mass spectrometry and BA-related enzymes, transporters, and regulators were evaluated by western blot analysis and qRT-PCR. In chow diet-fed mice, carvedilol increased plasma concentrations of BAs resulting from reduced BA uptake to hepatocytes via Ntcp transporter downregulation. Inhibition of the β-adrenoreceptor-cAMP-Epac1-Ntcp pathway by carvedilol may be the post-transcriptional mechanism underlying this effect. In contrast, carvedilol did not worsen the deterioration of BA homeostasis accompanying NASH; however, it shifted the spectra of BAs toward more hydrophilic and less toxic α-muricholic and hyocholic acids. This positive effect of carvedilol was associated with a significant attenuation of liver steatosis, inflammation, and fibrosis in NASH mice. In conclusion, our results indicate that carvedilol may increase BAs in plasma by modifying their liver transport. In addition, carvedilol provided significant hepatoprotection in a NASH murine model without worsening BA accumulation. These data suggest beneficial effects of carvedilol in patients at high risk for developing NASH.

 

Bile acids (BAs) are amphipathic steroid molecules that are synthesized in the liver from cholesterol and subsequently secreted into the bile as its main component. Bile transports BAs into the small intestine, where they emulsify dietary fat to enhance its absorption. Upon reaching the terminal ileum, BAs are almost completely absorbed into the portal circulation and effectively extracted into hepatocytes. BAs regulate lipid, glucose, and energy homeostasis through the activation of the farnesoid X receptor (FXR) and Takeda G-protein receptor 5 (TGR5, also known as Gpbar-1 for G-protein coupled bile acid receptor-1) (Arab et al., 2017). Individual BAs exhibit varying potency and efficacy on these receptors as agonists or even antagonists (Liu et al., 2014; Song et al., 2015). In addition, individual BAs show different toxicities in organs when accumulating during cholestasis. Consequently, it is not surprising that many drugs may cause liver injury by altering BA homeostasis. Therefore, it is important to characterize the changes in BA turnover for all drugs that are administered for liver disorders.

Carvedilol is a nonselective β-adrenoreceptor antagonist with α₁-adrenoreceptor blocking activity. It is widely used for the treatment of cardiovascular diseases (CVD). Carvedilol is also recommended for portal hypertension therapy in patients with liver cirrhosis (Villanueva et al., 2019). Compared with other β-adrenoreceptor antagonists, carvedilol does not impair glycemia, insulin sensitivity, or serum lipid profiles in type 2 diabetes mellitus patients (Nardotto et al., 2017). It also exhibits antifibrotic activity in animal models of liver injury induced by ethanol (Hakucho et al., 2014), carbon tetrachloride (Ling et al., 2019), or bile duct ligation (Tian et al., 2017). In addition, amelioration of liver impairment by carvedilol was described in a rat model of liver steatosis (Soliman et al., 2019); however, the efficacy of carvedilol treatment in nonalcoholic steatohepatitis (NASH), which is a more severe form of fatty liver, remains unclear. In contrast, treatment with propranolol, a β-adrenoreceptor antagonist, worsened liver injury in a mouse NASH model (McKee et al., 2013). Carvedilol may also induce cholestatic liver injury in patients (Rua et al., 2019) through an unknown mechanism. Recently, we observed BA retention in the plasma of estrogen-treated mice following treatment with labetalol, another α/β adrenoreceptor antagonist (Cristina Igreja Sá et al., 2023). Thus, the effect of carvedilol on BA homeostasis under physiological conditions and during liver disorders require further study.

Nonalcoholic fatty liver disease (NAFLD) represents the most common liver pathology worldwide and the incidence is increasing (Younossi et al., 2018). The clinicopathological features range from simple steatosis to NASH. The latter form can transition into irreversible damage, such as cirrhosis or cancer (Sheka et al., 2020). Moreover, 30%–40% of patients with NAFLD can develop NASH, which is characterized by necroinflammatory changes and varying degrees of liver fibrosis (Ekstedt et al., 2006). Unfortunately, NASH patients are at high risk of CVD, which is a major cause of death in these patients. Interestingly, drugs used to treat CVD comorbidities, such as statins or antidiabetics, may also influence NASH (Kothari et al., 2019). Therefore, understanding the basis for the positive effects is important for guiding therapeutic decisions.

The role of adrenergic receptors and their blockade during BA turnover regulation is not well understood. Only one recent study reported changes in hepatic BA transporters (Mayati et al., 2017) using primary human hepatocytes and HepaRG cells. The results showed that epinephrine repressed the expression of sinusoidal influx transporter mRNA, such as the sodium taurocholate cotransporting polypeptide (NTCP), organic anion transporting polypeptides OATP1B1 and OATP2B1, and the efflux transporters, multidrug resistance-associated proteins MRP2 and MRP3 as well as the bile salt exporting pump (BSEP). Similar changes were detected with the selective β₂-adrenoreceptor agonist fenoterol, the adenylate cyclase activator forskolin, and the cAMP (cyclic adenosine monophosphate) analog 8-bromo-cAMP. Importantly, β₂-adrenoreceptor antagonists, but not α-adrenoreceptor antagonists, suppressed the epinephrine-mediated repression of the BA transporters (Mayati et al., 2017), which indicates a β-adrenoreceptor-cAMP-dependent effect. These results are in contrast, however, with the inducing effect of cAMP, an important second messenger released following β-adrenoreceptor stimulation, on BA biliary secretion by stabilizing BA transporters at the canalicular and basolateral membranes (Mayer et al., 2019). Carvedilol has not been studied in this context yet. Therefore, we determined the effect of carvedilol on the liver and the enteric mechanisms responsible for BA homeostasis in chow diet-fed and NASH-induced mice.

Materials and methods

 

 

Animal study

All animal experiments were performed in compliance with the EU Directive 2010/63/EU and all animals received care according to the guidelines established by the Animal-welfare Body of the Faculty of Medicine in Hradec Kralove. The project was approved by the Animal-welfare Body of the Ministry of Education, Youth, and Sports (approval No. MSMT-11766/2018-2). Male mice were selected because of their higher susceptibility to diet-induced NASH (Ganz et al., 2014). Six-week-old male C57BL/6 mice were obtained from Velaz (Prague, Czech Republic) and housed individually in ventilated cages at a constant temperature of 23°C ± 1°C and humidity of 55% ± 10% under a 12-h light/dark cycle. The mice had free access to food and water. After 1-week of acclimation period, the animals were randomly divided into 2 groups (16 mice each) as follows: (1) mice fed a chow diet (C; PicoLab RD 20, LabDiet) and tap water or (2) mice fed a diet high in saturated fat, fructose, and cholesterol (FFC or F) consisting of an AIN-76A western diet (TestDiet, 40% calories from fat, 0.2% cholesterol), and glucose (18.1 g/l) with fructose (24 g/l) in water. At week 21 of the diets, the mice were randomly assigned to 4 experimental groups (8 mice each): (1) control group on chow diet with oral vehicle (CV), (2) control group receiving 10 mg/kg carvedilol (CC), (3) FFC diet with vehicle (FV), and (4) FFC diet administered with 10 mg/kg carvedilol (FC). Carvedilol or vehicle (1% methylcellulose) were administered once daily by oral gavage for 3 weeks together with the diets. After the final administration of carvedilol/vehicle, the animals were immediately placed into metabolic cages and stool was collected for 24 h. The mice were anesthetized with pentobarbital (50 mg/kg, i.p.) and bile was collected for 45 min through the cannulated gallbladder between 9:00 and 11:00 am to synchronize with BA production. At the end of the experiment, a blood sample was collected and the animals were sacrificed with an overdose of anesthesia. Plasma samples were obtained from whole blood by centrifugation at 2000 × g for 5 min at 4°C. The liver and intestine were excised and immediately frozen in liquid nitrogen. All samples were stored at −80°C until analysis. In the present study, the dose and timing of carvedilol were selected based on previous studies (Gao et al., 2017; Wu et al., 2019) upon consideration of the pharmacokinetics (Abdullah Shamim et al., 2022) and analysis of drug tolerance.

Analytical methods

Liver enzymes were analyzed in plasma using a commercial Preventive Care Profile Plus test with a Vetscan 2 device (Abaxis) according to the manufacturer’s instructions. Triglyceride levels in plasma and liver were measured using commercial kits (Erba Lachema s.r.o., Brno, Czech Republic) as previously described (Igreja Sá et al., 2020). The concentration of BAs in plasma, bile, liver, small intestine, and feces was measured using a liquid chromatography-mass spectrometry (LC-MS) method as described previously (Uher et al., 2023). A list of all quantified BAs are provided in the Supplementary data. For sample preparation, specific homogenization was first necessary for liver, small intestine, and stool. A piece of liver (50 mg) was homogenized in 1 ml of 75% ethanol solution using an Ultra-turrax dispenser (IKA-Werke GmbH, Germany). To determine the content of BA in the small intestine, we first pulverized whole duodenum and jejunum together in liquid nitrogen and a 50 mg sample was added to 1 ml of 75% ethanol. Similarly, whole stool collected over 24 h in metabolic cages (housed individually) was dried, weighed, mixed, and pulverized. A 50 mg sample was added to 1 ml of 75% ethanol. The BA content was calculated per dry weight. Dissolution of BA from stool, liver, and small intestine samples using ethanol was performed by shaking the mixtures at 50°C for 2 h. The mixture was centrifuged at 15 000 × g for 10 min and the supernatant was collected.

Then, 300 µl of 0.1% (v/v) formic acid in acetonitrile containing deuterated internal standards was added to a volume of 100 µl of plasma or supernatant of the liver, or small intestine homogenates. Bile samples were processed by a slight modification because of the high BA concentrations. Two sets of bile samples were prepared. The first set was 10-fold diluted, and the second set was 100-fold diluted with 50% (v/v) methanol. The feces were analyzed in the same way as bile samples, except that feces homogenate was diluted only 2-fold with 50% (v/v) methanol. Aliquots of the diluted samples in a 100 µl volume were taken and 900 µl of 0.1% (v/v) formic acid in acetonitrile containing deuterated internal standards were added. The samples were vortexed for 60 s at 1500 RPM and left at −20°C (for 10 min). Any precipitate was removed by centrifugation at 13 000 × g(5 min, 20°C). The supernatants (100 µl) were transferred to clean sample tubes and evaporated to dryness under reduced pressure at a temperature of 60°C (Eppendorf Concentrator 5301). The dried samples were reconstituted in 100 µl of 50% acetonitrile, vortexed, centrifuged, and filtered through a 96-well filter plate (AcroPrep Advance, wwPTFE). Two microliters of sample were injected onto the column. Analyzed BAs were ursodeoxycholic acid (UDCA), hyoxycholic acid (HCA), hyodeoxycholic acid (HDCA), chenodeoxycholic acid (CDCA), deoxycholic acid (DCA), α/β muricholic acid (α/β MCA), cholic acid (CA), lithocholic acid (LCA), and their taurine- or glycine conjugates (TLCA, TUDCA, THDCA, TCDCA, TDCA, TMCA, TCA, and THCA). Corresponding deuterated standards were used for 14 of 23 BAs, including CA (CA-D5), UDCA (UDCA-D4), CDCA (CDCA-D4), DCA (DCA-D4), GCA (GCA-D4), GCDCA (GCDCA-D4), GDCA (GDCA-D4), TUDCA (TUDCA-D4), GUDCA (GUDCA-D4), GLCA (GLCA-D4), TCDCA (TCDCA-D4), TDCA (TDCA-D4), TCA (TCA-D4), and TLCA (TLCA-D4). For the remaining BAs these deuterated standards were used: HDCA, αMCA (CA-D5); βMCA, HCA (UDCA-D4); GHDCA, THDCA (GUDCA-D4); GHCA, TMCA, and THCA (TUDCA-D4). The sources of the internal standards are provided in the supplementary files. Due to issues with sufficient separation of TαMCA and TβMCA isomers, their concentration is presented as a sum.

The concentrations of individual BAs were summed to calculate the concentration of conjugated, unconjugated and total BAs. Primary BAs: (T)CA, (T)CDCA, (T)α/βMCA; secondary BA: (T)DCA, (T)LCA, (T)UDCA, (T)HCA, (T)HDCA; 12α-OH BAs: (T)CA and (T)DCA; and non-12α-OH BAs refers to all of the remaining BAs. The liver content of BAs and their biliary excretion was calculated per gram of the liver. The BA pool was calculated as the sum of the total BAs circulating in the enterohepatic circulation, including the liver, small intestine, and biliary system (Chiang, 2017). The BA content in the biliary system was estimated from BA biliary secretion over 20 min, which corresponds with the average volume of the gallbladder.

Quantification of gene and protein expression in the in vivo samples

Gene expression was semiquantified using quantitative reverse transcription real time PCR (qRT-PCR) with a Quantstudio 7 HT Fast Real Time PCR System using predesigned TaqMan Gene Expression Assays (Supplementary Table 1) and TaqMan Fast Universal PCR Master Mix (ThermoFisher Scientific, Waltham, Massachusetts). Relative gene expression was calculated by the standard 2^−ΔΔCt method using glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as a housekeeping gene. Protein isolation and western blot analyses were done as previously described (Igreja Sá et al., 2020). Briefly, proteins were separated by SDS-PAGE on 5%–10% polyacrylamide gels, transferred to a PVDF membrane (Millipore, New York, New York) and incubated with the appropriate primary and secondary antibodies (Supplementary Table 2). Visualization of the bands was performed using Evolution-capt software (Fusion Solo 6S Edge, Vilber Lourmat SAS, France) and quantified with ImageLab imaging software version 6.0.1 (Bio-Rad). The expression of proteins was normalized to the total content of protein visualized on the blotting membrane using stain-free technology (Gürtler et al., 2013).

Gut microbiota analysis

Total DNA was isolated from stool (24 h collection) using the QIAamp Fast DNA Stool Mini Kit (QIAGEN, cat. no. 51604) following the manufacturer’s protocol. DNA yield and integrity was confirmed using a NanoDrop ND-1000 spectrophotometer. Quantitative PCR was performed using a Quantstudio 7 HT Fast Real Time PCR System with Sybr Select Master Mix (ThermoFisher Scientific). The 16S rRNA gene-targeted group-specific primers were used to quantify bacteria phyla from Firmicutes, Bacteroidetes, Actinobacteria, and Preotobacteria. The primer sequences are listed in Supplementary Table 3. A standard curve of Universal Eubacteria 16S rRNA gene was used to measure bacterial abundance.

Bile salt hydrolase enzyme activity assay

Faecal samples were prepared by mixing 50 mg of cecum content with 1 ml of PBS containing protease inhibitor cocktail (Roche Diagnostics GmbH, Mannheim, GE), and DTT (1 mM final concentration, Merck, no. 43815) to minimize enzyme oxidation. Total protein was quantified using the bicinchoninic acid assay according to the manufacturer’s instructions (PIERCE, Rockford) (Mullish et al., 2018). A 500 µg sample of cecal content protein was incubated at 37°C in 200 µl of PBS (pH 5.8, final concentration of 0.02 mM). The reaction was started by the addition of taurocholic acid (TCA, Merck, no. T4009) at a final concentration of 5 µM. The reaction was stopped after 1 min by adding 600 µl of 100% ethanol, followed by a 1-h incubation at 60°C to dissolve BAs. The sample was then analyzed by LC-MS for concentration of BAs as described above. Hydrolysis efficacy was determined by the percentage of conjugated TCA reduction in the cecal content homogenate, which was normalized to TCA concentration at the start of reaction (0 min) designated as 100%.

Histopathological examination of the liver

Liver tissue was harvested, fixed with formalin, and embedded in paraffin. Tissue sections (4 µm) were cut and placed onto glass slides. Hematoxylin and eosin staining was used to grade liver injury. The degree of fibrosis was classified using Sirius red (SR) and the liver tissue was stained with Direct Red 80 and Fast-Green FCF (color index 42053, Sigma). For immunohistochemistry, liver slices were deparaffinized, hydrated, and incubated with anti-Mac-2 (Galectin 3, eBioscience no. 14-5301-82, 1:250) and anti-NTCP (Bioss no. bs-1958R, 1:1000) antibodies overnight at 4°C. For antibody detection, we used the Vectastain ABC kit (no. PK-4001, no. PK-6102, Vector Laboratories), diaminobenzidine as a substrate, and hematoxylin as a counterstain. Samples were quantified using 20 random segments in a blind manner using Nikon Eclipse TE300 polarized light (SR) and microscope software.

Human hepatic stellate cells

The human hepatic stellate cell (HSC), stellate cell medium with all required supplements, and poly-l-lysine were purchased from ScienCell Research Laboratories (Carlsbad, California). Cells from the third and fourth passage were used for all experiments. Human hepatic stellate cells were grown in poly-l-lysine coated flasks in stellate cell medium supplemented with 2% fetal bovine serum (FBS), 100 IU/ml penicillin and 100 μg/ml streptomycin in a humidified atmosphere containing 5% CO₂ at 37°C. After reaching confluence, the cells were cultured for 24 h under the following conditions: (1) medium only (control cells), (2) 3 ng/ml transforming growth factor beta 1 (TGF-β1, Sigma-Aldrich) to activate a fibrotic phenotype, and (3) 3 ng/ml TGF-β1 plus 10 µM carvedilol (Sigma-Aldrich). After 24 h, total RNA was extracted using the High Pure RNA Isolation Kit (Roche) and reverse-transcribed into cDNA using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) according to the manufacturer’s instructions. Real time PCR was done using a Quantstudio 7 HT Fast Real Time PCR System (Applied Biosystems) with predesigned TaqMan Gene Expression Assays (Supplementary Table 1) and TaqMan Fast Universal PCR Master Mix. Data were normalized to the 18S rRNA gene and the results were analyzed by the 2^−ΔΔCt method. Three independent experiments were performed.

HepaRG incubation

Cryopreserved HepaRG (GIBCO) cells and media were purchased from Life Technologies (Carlsbad, California). The HepaRG cell line was cultured and differentiated as described previously (Prasnicka et al., 2017). After 72 h of stabilization in dimethyl sulfoxide (DMSO 0.1%), cells were treated with 10 µM carvedilol (CARV) ±1 or 10 µM salbutamol (SAL) for 48 h. Total RNA was isolated using TRIZOL, and reverse transcription was performed with the High Capacity cDNA Reverse Transcription Kit. Quantitative PCR was performed using TaqMan Fast Advanced Master Mix with TaqMan probes for Slc10a1 (Assay ID: Hs00161820_m1), Abcb11 (Assay ID: Hs00184824_m1), and B2M (Assay ID: Hs00187842_m1) (ThermoFisher Scientific). The 2^−ΔΔCt method was used for gene expression quantification with normalization to GAPDH and B2M average gene expression. Another set of HepaRG cells were treated either with vehicle (0.1% DMSO), 10 µM CARV, 1 µM SAL, the combination of both, or 10 µM obeticholic acid (OCA) for 48 h. The cells were lysed in RIPA buffer containing protease and phosphatase inhibitors. The expression of the NTCP protein was assessed by western blot analysis. Protein levels were normalized by stain-free imaging technology.

Luciferase gene reporter assays in murine cell lines

The murine enteroendocrine cell line GLUTag was kindly provided by Dr Colette Roche from the Centre de Recherche en Cancérologie de Lyon (INSERM U1052, Lyon, France) with the permission of Dr Daniel J. Drucker (Lunenfeld Tanenbaum Research Institute Mt. Sinai Hospital, Toronto). GLUTag cells were cultured in DMEM (GlutaMAX, ThermoFisher Scientific ) supplemented with 10% FBS (HyClone, ThermoFisher Scientific). The AML-12 cell line (ATCC CRL-2254) was established from normal hepatocytes of a CD1 mouse (line MT42) and is transgenic for human TGFα. AML-12 cells were cultured in DMEM: F12 medium (ThermoFisher Scientific) supplemented with 10% FBS, 10 μg/ml insulin, 5.5 μg/ml transferrin, 5 ng/ml selenium, and 40 ng/ml dexamethasone. Chenodeoxycholic acid (CDCA, Cat. No. 700198P) and GW4064 (Cat. No. G5172) were purchased from Merck, Darmstadt, Germany. Obeticholic acid (Synonyms: OCA; INT-747; 6-ECDCA; 6-ethylchenodeoxycholic acid; Cat. No HY-12222) was purchased from MCE MedChem Express (New Jersey).

Transient transfection gene reporter assays were done using Lipofectamine 3000 (ThermoFisher Scientific) as previously described (Skoda et al., 2020; Stefela, 2021). The GLUTag and AML-12 cells were seeded at a density of 40 000 cells/cm² on 48-well plates and transfected after 24 h with pFxrE-luc or pShp-luc luciferase reporter constructs (150 ng/well) together with expression vectors for murine Fxr (100 ng/well), Rxrα (100 ng/well), and the RL-TK vector for Renilla luciferase normalization (30 ng/well).

The murine Fxr expression vector for the murine Nr1h4 gene cDNA cloned into the pcDNA3.1+C-(K)-DYK vector was purchased from GenScript. The pFXRE-luc construct containing 3 copies of FXR responsive elements, the pSHP-luc construct with 2080 bp of the human NR0B2 gene promoter, and the RXRα expression vector are described in our previous report (Stefela et al., 2020). The pRL-TK Renilla expression vector was used for transfection normalization (Promega). Cells were treated with OCA, GW4064, CDCA (all FXR agonists), and CARV at a 10 µM concentration for 24 h. The cells were lysed with Passive lysis buffer (Promega) and both firefly luciferase and Renilla luciferase activities were measured using the Dual reporter assay (Promega, Hercules) according to the manufacturer’s protocol.

For RT-PCR experiments, GLUTag and AML-12 cells were seeded into 12-well plates 1 day prior to treatment. Cells were treated with the prototype FXR/Fxr agonist 10 µM OCA or 10 µM CARV for 24 h. RNA from GLUTag and AML-12 cells was isolated using TRI-reagent (Sigma-Aldrich, now Merck) and cDNA was synthesized using the Thermo Scientific RevertAid RT Kit (ThermoFisher Scientific). Real time PCR was performed using the Quant Studio 6 instrument with the Fast Advanced Master Mix (ThermoFisher Scientific) according to the MIQE guidelines. TaqMan probes were purchased from ThermoFisher Scientific, which included Abcb11 (Bsep, Mm00445168_m1), Nr0b2 (Shp, Mm00442278_m1), Nr1h4 (Fxr, Mm00436425_m1), Ostβ (Slc51b, Mm01175040_m1), and Fgf15 (Mm00433278_m1). Data were normalized to β₂ microglobulin (B2M, Mm00437762_m1) as a reference gene, analyzed using the ^ΔΔCt method, and presented as fold difference relative to the control (vehicle treated) sample.

Statistical analysis

All statistical analyses were done using GraphPad Prism 8 statistical software (San Diego). The normal distribution of the data was tested using the Shapiro-Wilk test. Data are presented as the median with boxes and whiskers representing the interquartile range and 5th to 95th percentiles, respectively. Statistical significance (p < .05) was determined using either a 1-way ANOVA with Holm-Sidak’s post hoc test for data with Gaussian distribution or the Kruskal-Wallis 1-way analysis with Dunn’s post hoc test if at least one mouse group demonstrated a non-normal distribution for the evaluated parameter.

Results

 

Bile acid concentrations in plasma

Analysis of BA plasma concentrations in vehicle-treated chow mice revealed a dominant proportion of non-12α-hydroxylated, primary, and conjugated BAs (Figure 1A;Supplementary Table 4) as determined by summing the corresponding BAs (12α-hydroxylated; non-12α-hydroxylated, primary, secondary, taurine-conjugated, and unconjugated) and calculating their mutual ratios (12α-OH/non-12α-OH; primary/secondary; conjugated/unconjugated). Carvedilol administration to chow mice resulted in a significant increase (188%) in total BA plasma concentrations (Figure 1A). Compared with the chow diet-fed animals, total plasma BA concentrations were significantly increased by 149% in vehicle treated NASH mice due to elevated plasma levels of 12α-hydroxylated (248%), primary (150%), secondary (91%), and conjugated BAs (250%) (Figure 1A), The 12α-hydroxylated/non-12α-hydroxylated (48%) and conjugated/unconjugated BA (242%) ratios were concomitantly raised in mice fed an FFC diet (Figure 1A). Carvedilol exerted no significant effect on net plasma BA concentrations in NASH animals compared with untreated NASH mice (Figure 2A). Thus, carvedilol appears to elevate plasma BA concentrations in the chow controls, but this detrimental effect is absent in NASH animals.

Figure 1.

Bile acids in the plasma and hepatobiliary system. Plasma concentrations (A), liver content (B), and biliary secretions calculated per gram of the liver (C) of the main BA groups and their ratios were analyzed in mice treated with either vehicle or carvedilol (10 mg/kg, daily p.o. for 3 weeks) at the end of a 24-week feeding cycle with either a chow diet or a diet rich in saturated fat, fructose, and cholesterol (FFC). Animal labeling: CV (chow diet with the vehicle), CC (chow diet with carvedilol), FV (FFC diet with the vehicle), and FC (FFC diet with carvedilol). BA groups: sum of all BAs (∑-BAs), 12α-hydroxylated BAs (12α-OHs), non-12α-hydroxylated BAs (non-12α-OHs), primary BAs (1° BAs), secondary BAs (2° BAs), taurine-conjugated BAs (T-BAs), and unconjugated BAs (U-BAs). Statistical significance was determined using either a 1-way ANOVA with Holm-Sidak’s post hoc test for data with Gaussian distribution or the Kruskal-Wallis 1-way analysis with Dunn’s post hoc test if at least one mouse group demonstrated a non-normal distribution for the evaluated parameter. Data are presented as medians with boxes and whiskers representing the interquartile range and 5th to 95th percentiles, respectively (n = 8). Asterisks (*) denote statistical differences from the vehicle chow diet CV group (p < .05). Dagger (†) denotes statistical differences from the vehicle FFC diet FC group (p < .05).

Figure 2.

Individual bile acids in the plasma and hepatobiliary system. Plasma concentrations (A), liver content (B), and biliary secretion (C) of individual BAs were analyzed in mice treated with either vehicle or carvedilol (10 mg/kg, daily p.o. for 3 weeks) at the end of a 24-week feeding cycle with either a chow diet or a diet rich in saturated fat, fructose, and cholesterol (FFC). Animal labeling: CV (chow diet with the vehicle), CC (chow diet with carvedilol), FV (FFC diet with the vehicle), and FC (FFC diet with carvedilol). Statistical significance was determined using either a 1-way ANOVA with Holm-Sidak’s post hoc test for data with Gaussian distribution or the Kruskal-Wallis 1-way analysis with Dunn’s post hoc test if at least one mouse group demonstrated a non-normal distribution for the evaluated parameter. Data are presented as medians with boxes and whiskers representing the interquartile range and 5th to 95th percentiles, respectively (n = 8). Asterisks (*) denote statistical differences from the vehicle chow diet CV group (p < .05). Dagger (†) denotes statistical differences from the vehicle FFC diet FC group (p < .05).

The spectra of BAs in the liver

To determine whether carvedilol modifies the uptake of BAs into hepatocytes, we measured BA content in the liver (Figure 1B;Supplementary Table 5). Dominant BAs were non-12α-hydroxylated, primary, and conjugated BAs. On the other hand, the proportion of conjugated BAs was markedly higher compared with that in plasma (21-fold), which indicates robust intracellular conjugation of BAs with taurine. Carvedilol did not alter BA content in the liver of chow diet-fed animals. Similarly, the liver BA content calculated per gram of liver weight was unchanged in NASH mice. The only exception was increased liver content of secondary BA (64%) in the FFC-fed group.

Biliary secretion of BAs

The biliary secretion rate was determined by analyzing BAs in the bile collected over 45 min after cannulation of the gallbladder (Figure 1C;Supplementary Table 6). Unlike plasma and the liver, equal quantities for 12α-hydroxylated and non-12α-hydroxylated BAs were detected in chow diet-fed mice. Primary BAs dominated over secondary BAs and the ratio of conjugated to unconjugated BAs was more than 100, which was consistent with dominant biliary secretion of conjugated BAs. Carvedilol significantly increased the biliary secretion of secondary BA by 89% in the chow controls (Figure 1C). In contrast, net biliary BA secretion was significantly decreased in untreated FFC mice by 70% compared with chow diet-fed mice (Figure 1C), because of significantly decreased levels of 12α-hydroxylated (64%), non-12α-hydroxylated (74%), primary (71%), conjugated (70%), and unconjugated BAs (95%). Reduction of unconjugated BAs significantly increased the conjugated/nonconjugated BA ratio by 655% (Figure 1C) in NASH mice. Carvedilol treatment did not change the biliary secretion of net BAs in NASH mice.

Individual BAs in plasma, liver, and bile

The most abundant BAs in plasma were TMCA, TCA, and their unconjugated forms α/βMCA (Figure 2A;Supplementary Table 7). Carvedilol administration to chow mice caused significant increase of TCA (363%), and THDCA (323%) (Figure 2A). Other BAs also tended to increase. FFC diet led to significant increase in plasma concentrations of TMCA (205%), TCA (402%), TCDCA (289%), TUDCA (285%), and TDCA (242%) in comparison with chow diet-fed CV group (Figure 2A). Carvedilol significantly increased αMCA concentration (432%) in plasma of NASH mice.

The most abundant BAs in the liver were TMCA and TCA (Figure 2B;Supplementary Table 8). Carvedilol did not alter BA content in the liver of chow diet-fed animals. FFC diet raised the liver content of TDCA (150%) compared with chow mice. Carvedilol treatment increased the presence of αMCA in the liver of NASH mice by 173% (Figure 2B). The TMCA and TCA were also the major BAs in the bile (Figure 2C;Supplementary Table 9). The biliary secretion of BAs was not modified by carvedilol in chow diet-fed mice. In contrast, onset of NASH significantly decreased biliary secretion of αMCA (87%), βMCA (93%), CA (99%), TUDCA (65%), THDCA (80%), TCDCA (56%), TMCA (76%), THCA (88%), and TCA (66%) relative to chow diet-fed mice (Figure 2C). A significant increase of THCA (411%) was detected in carvedilol-administered FFC-fed mice compared with untreated NASH group (Figure 2C).

Bile production and biliary secretion of glutathione

To further analyze bile formation, we measured the net bile flow and biliary secretion of glutathione, which is a major component of BA-independent bile secretion (Figure 3). Both of these parameters were not modified by carvedilol in mice receiving a chow diet (Figs. 3A and 3B); however, they were significantly decreased in the vehicle treated NASH group by 69% and 74%, respectively, compared with the vehicle treated chow controls. Carvedilol significantly increased bile flow (49%) and total glutathione biliary secretion (49%) in NASH mice (Figs. 3A and 3B). Interestingly, the liver content of reduced glutathione (GSH), a major liver form of glutathione, was significantly reduced by carvedilol by 25% in the chow controls and significantly increased in the NASH groups by 29% compared with the untreated chow diet-fed animals (Figure 3C). The oxidized glutathione (GSSG) content in the liver was unchanged by carvedilol or NASH (Figure 3D). Therefore, the GSH to GSSG ratio was significantly decreased by 17% in the carvedilol-treated controls receiving a chow diet and increased by 34% in the untreated NASH mice (Figure 3E). These results indicate that carvedilol significantly influences bile flow or glutathione-dependent bile formation.

Figure 3.

Bile flow and hepatobiliary disposition of glutathione. Bile flow (A), biliary secretion of total glutathione (B) and liver content of reduced GSH glutathione (C), oxidized GSSG glutathione (D), and the liver ratio of both (E) were analyzed in mice treated with either vehicle or carvedilol (10 mg/kg, daily p.o. for 3 weeks) at the end of a 24-week feeding cycle with either a chow diet or a diet rich in saturated fat, fructose, and cholesterol (FFC). Animal labeling: CV (chow diet with the vehicle), CC (chow diet with carvedilol), FV (FFC diet with the vehicle), and FC (FFC diet with carvedilol). Statistical significance was determined using either a 1-way ANOVA with Holm-Sidak’s post hoc test for data with Gaussian distribution or the Kruskal-Wallis 1-way analysis with Dunn’s post hoc test if at least one mouse group demonstrated a non-normal distribution for the evaluated parameter. Data are presented as medians with boxes and whiskers representing the interquartile range and 5th to 95th percentiles, respectively (n = 8). Asterisks (*) denote statistical differences from the vehicle chow diet CV group (*p < .05, **p < .01, ***p < .001). Dagger (†) denotes statistical differences from the vehicle FFC diet FC group (p < .05).

Protein expression of BA transporters in the liver

To identify mechanisms associated with the observed changes in BA transport across hepatocytes and their intracellular synthesis, we analyzed the hepatic protein expression of enzymes and transporters involved in BA homeostasis (Figure 4A). Carvedilol treatment of chow diet-fed mice significantly reduced liver Mrp2 protein expression (60%), the major apical transporter for biliary secretion of organic anions, such as BA conjugates; Mrp4 (56%), an essential basolateral transporter for BA efflux from hepatocytes to plasma; Bsep (25%), the major transporter for the secretion of BAs from hepatocytes into the bile; and Ntcp (58%), a basic basolateral transporter for BA uptake from sinusoids to hepatocytes (Figure 4A). In contrast, carvedilol significantly increased the expression of Cyp7a1 (33%), the rate-limiting step in BA synthesis. The FFC diet significantly down-regulated the Mrp4 (44%), Bsep (53%), and Ntcp (76%) transporters as well as the enzymes required for liver BA synthesis, such as Cyp7a1 (33%), Cyp8b1 (38%), Cyp27a1 (53%), and Cyp2c70 (65%), except Cyp7b1, which was significantly increased (28%). Interestingly, in FFC mice, carvedilol significantly altered only the Cyp2c70 protein, which was significantly induced (106%) (Figure 4A). We confirmed reduced expression of the Ntcp protein using immunohistochemical staining in liver sections of chow mice (Figure 4B).

Figure 4.

Effect of carvedilol on hepatic protein of bile acid-related molecules. Analyses were performed in the liver of mice that were treated with either vehicle or carvedilol (10 mg/kg, daily p.o. for 3 weeks) at the end of a 24-week feeding cycle with either a chow diet or a diet rich in saturated fat, fructose, and cholesterol (FFC). The protein expression of important bile acid synthetizing enzymes and transporters in the liver was analyzed by western blotting (A). Protein levels were normalized by stain-free imaging technology. Immunohistochemical analysis of Ntcp on paraffin-embedded mouse liver tissue (B). Scale bar 100 µm. The Epac1 protein was analyzed by western blotting (C). Animal labeling: CV (chow diet with the vehicle), CC (chow diet with carvedilol), FV (FFC diet with the vehicle), and FC (FFC diet with carvedilol). Statistical significance was determined using either a 1-way ANOVA with Holm-Sidak’s post hoc test for data with Gaussian distribution or the Kruskal-Wallis 1-way analysis with Dunn’s post hoc test if at least one mouse group demonstrated a non-normal distribution for the evaluated parameter. Data are presented as medians with boxes and whiskers representing the interquartile range and 5th to 95th percentiles, respectively (n = 8). Asterisks (*) denote statistical differences from the vehicle chow diet CV group (*p < .05, ***p < .001). Dagger (†) denotes statistical differences from the vehicle FFC diet FC group (p < .05).

A major pathway known for regulating transporters such as Bsep and Ntcp is cAMP, which also acts as a second messenger for β-receptors blocked by carvedilol. Therefore, we measured the expression of Epac1 (exchange protein directly activated by cAMP 1), a cAMP sensor regulating BA transporters membrane trafficking (Figure 4C). Compared with chow diet-fed mice, Epac1 protein was significantly decreased by carvedilol (70%) and the FFC diet (73%). In contrast, carvedilol did not further affect Epac1 in FFC-fed mice.

Gene expression of BA transporters in the liver

To further assess of changes in protein content, we quantified the mRNA of liver BA transporters and enzymes (Figure 5A). Carvedilol did not alter the mRNA expression of the target genes in the chow diet-fed group or FFC-fed mice. The only exception was a significant induction of Cyp2c70 (37%) in the chow controls. These discrepancies between protein and mRNA content suggests post-transcriptional regulation of BA transporters by carvedilol. In contrast, a significant reduction caused by the FFC diet was observed for Abcb11 (Bsep) (42%), Slc10a1 (Ntcp) (35%), Cyp7a1 (66%), Cyp8b1 (57%), Cyp27a1 (37%), and Cyp2c70 (36%) mRNA, which was consistent with their corresponding protein expression, indicating transcriptional regulation. Slc51b mRNA was significantly induced in NASH mice; however, discrepancies in Abcc2 (Mrp2) (16% reduction), Abcc4 (Mrp4) (28% induction), and Cyp7b1 (88% reduction) also suggests the contribution of a translational mechanism in NASH. Therefore, the expression of SLC10A1 (NTCP) (Figure 5B), and ABCB11 (BSEP) (Figure 5C) mRNA was measured in HepaRG cells in the presence of carvedilol and/or salbutamol, a selective β₂-adrenoreceptor agonist. Carvedilol alone had no effect on either transporter, but it significantly decreased the expression of SLC10A1 (34%) and ABCB11 (15%) transporter mRNA when combined with salbutamol. Salbutamol at 1 µM concentration significantly increased SLC10A1 mRNA expression (65%). Western blot analyses confirmed the significant downregulation of NTCP by carvedilol in salbutamol-treated HepaRG cells (35%) (Figure 5D).

Figure 5.

Effect of carvedilol on the expression of bile acid-related molecules in mouse livers and HepaRG cells. Analyses were performed in the liver of mice treated with either vehicle or carvedilol (10 mg/kg, daily p.o. for 3 weeks) at the end of a 24-week feeding cycle with either a chow diet or a diet rich in saturated fat, fructose, and cholesterol (FFC). The mRNA gene expression of liver BA-related molecules were measured by quantitative reverse transcription polymerase chain reaction in mouse livers (A). Animal labeling: CV (chow diet with the vehicle), CC (chow diet with carvedilol), FV (FFC diet with the vehicle), and FC (FFC diet with carvedilol). Statistical significance was determined using either a 1-way ANOVA with Holm-Sidak’s post hoc test for data with Gaussian distribution or the Kruskal-Wallis 1-way analysis with Dunn’s post hoc test if at least one mouse group demonstrated a non-normal distribution for the evaluated parameter. Data are presented as medians with boxes and whiskers representing the interquartile range and 5th to 95th percentiles, respectively (n = 8). Asterisks (*) denote statistical differences from the vehicle chow diet CV group (p < .05). Dagger (†) denotes statistical differences from the vehicle FFC diet FC group (p < .05). The expressions of SLC10A1 (NTCP) (B) and ABCB11 (BSEP) transporters (C) mRNA and NTCP protein expression (D) were also evaluated (Kruskal-Wallis 1-way analysis with Dunn’s post hoc test) in HepaRG cells after incubation with salbutamol (β₂ adrenoreceptor agonist, 1 µM or 10 µM) and carvedilol (10 µM). Values are presented as medians, with boxes and whiskers representing the interquartile range and 5th to 95th percentiles of 3 independent experiments. *p < .05, **p < .01 versus vehicle-treated controls, ^†p < .05 versus salbutamol 1 µM treated cells.

Intestinal content and fecal excretion of BAs

The small intestine is the principal reservoir for BAs in the organism. The LC-MS analysis of BA spectra revealed a primary proportion of non-12α-hydroxylated, primary, and conjugated BAs in the chow controls (Figure 6A;Supplementary Table 10). Carvedilol treatment had no effect on total BA content in the small intestine of the chow diet-fed mice (Figure 6A). The FFC diet significantly increased the overall small intestinal BA content (80%) (Figure 6A). Compared with CV animals, the FFC diet significantly increased 12α-hydroxylated (118%), primary (79%), secondary (224%), conjugated (170%), and reduced unconjugated (98%) BAs followed by significantly reduced primary/secondary (41%) and increased conjugated/unconjugated (207-fold) BA ratios (Figure 6A). The intestinal content of BA groups and ratios in FFC mice was not influenced by carvedilol compared with untreated NASH group (Figure 6A).

Figure 6.

Effect of carvedilol on enteric homeostasis of bile acids. Small intestine BA content (A), BA fecal excretion (B), DNA content of Firmicutes phylum bacteria in feces (C), total BA pool size in the organism (D), and the ileum mRNA expression of bile acid (BA)-related molecules (E) were analyzed in mice that were treated with either vehicle or carvedilol (10 mg/kg, daily p.o. for 3 weeks) at the end of a 24-week feeding cycle with either a chow diet or a diet rich in saturated fat, fructose, and cholesterol (FFC). Animal labeling: CV (chow diet with the vehicle), CC (chow diet with carvedilol), FV (FFC diet with the vehicle), and FC (FFC diet with carvedilol). BA groups: sum of all BAs (∑-BAs), 12α-hydroxylated BAs (12α-OHs), non-12α-hydroxylated BAs (non-12α-OHs), primary BAs (1° BAs), secondary BAs (2° BAs), taurine-conjugated BAs (T-BAs), and unconjugated BAs (U-BAs). Statistical significance was determined using either a 1-way ANOVA with Holm-Sidak’s post hoc test for data with Gaussian distribution or the Kruskal-Wallis 1-way analysis with Dunn’s post hoc test if at least one mouse group demonstrated a non-normal distribution for the evaluated parameter. Data are presented as medians with boxes and whiskers representing the interquartile range and 5th to 95th percentiles, respectively (n = 8). Asterisks (*) denote statistical differences from the vehicle chow diet CV group (p < .05). Daggers (†) denote statistical differences from the vehicle FFC diet FC group (p < .05).

Compared with other matrices, fecal excretion of 12α-hydroxylated, non-12α-hydroxylated, primary, and secondary BAs was proportional in chow diet-fed mice with mutual ratios of approximately 1 (Figure 6B;Supplementary Table 11). The predominant BAs were unconjugated species (Figure 6B). The fecal excretion of total BAs, their major groups, and mutual ratios were not modified by carvedilol in both treated groups, except for the significantly reduced primary/secondary BAs ratio (43%) in the NASH group (Figure 6B). In contrast, an FFC diet significantly increased the fecal output of total BAs (227%), because of increased 12α-hydroxylated (350%), non-12α-hydroxylated (202%), primary (264%), secondary (266%), and unconjugated (276%) BAs compared with controls on a chow diet (Figure 6B). The ratios of 12α-hydroxylated/non-12α-hydroxylated BAs and primary/secondary BAs remained unchanged between the chow diet-fed and FFC groups. The ratio of conjugated to unconjugated BAs was decreased (59%) in feces of FFC-fed animals, suggesting an extensive deconjugation by bacteria. Therefore, we measured the relative 16S DNA abundance of essential bacterial phyla in the feces (Figures 6C;Supplementary Figure 1A). The FFC diet significantly increased the abundance of Firmicutes (30%) (Figure 6C) compared with the chow diet-fed group that was significantly decreased (35%) in carvedilol-treated NASH mice. Other bacterial phyla were not influenced by either an FFC diet or carvedilol administration (Supplementary Figure 1A). These findings were consistent with unchanged BSH activity by either carvedilol or NASH (Supplementary Figure 1B).

The BA content in the biliary tract (Supplementary Figure 1C and Table 12) was calculated from their biliary secretion. Carvedilol significantly increased secondary BA content (73%) in chow mice, without a significant modulation of any specific BA (Supplementary Figure 1D and Table 13). The FFC diet significantly decreased unconjugated BAs (88%) with concomitant increase in the conjugated to unconjugated BA ratio (655%) (Supplementary Figure 1C) compared with chow mice because of a reduced content of α/βMCA (68%/83%), CA (97%), and THCA (72%). Carvedilol treatment of FFC mice significantly reduced this ratio (56%). Carvedilol did not modify the BA pool size (Figure 6D), but total BA content was significantly increased in NASH mice (59%) compared with the chow controls.

To identify the underlying mechanisms for the observed changes in fecal BA excretion, we measured the expression of essential BA transporters and regulators in the ileum (Figure 6E). Carvedilol administration to chow mice significantly increased the mRNA of the apical sodium-dependent bile salt transporter Slc10a2 (Asbt) (79%), the major uptake transporter for the reabsorption of BA; as well as other target genes for Fxr such as the ileal BA-binding protein (Fabp6, protein Ibabp) (142%) and organic solute transporter alpha and beta (Slc51a/b, protein Ostα/β) (288% and 232%, respectively), major transporters for BA efflux from enterocytes to the interstitium. Untreated NASH mice exhibited a significant reduction of Slc10a2 (75%) and the induction of Fgf15 (606%) compared with the chow controls. Administration of carvedilol to NASH mice significantly increased the expression of Slc10a2 (229%), Nr0b2 (encoding Shp protein—small heterodimer partner) (12.8-fold), Fgf15 (110%), Fabp6 (297%), and Slc51a/b (359%/161%, respectively) mRNA compared with untreated FFC-fed animals. These results indicate a significant induction of these Fxr-target genes in the ileum of carvedilol-treated lean and NASH mice.

The predominate BAs in the small intestine of chow diet-fed mice were TMCA and TCA and their unconjugated forms (Figure 7A;Supplementary Table 14). Contents of individual BAs were not modified by carvedilol in the chow controls. In contrast, mice with NASH showed increased intestinal abundance of TLCA (75%), TUDCA (411%), TCDCA (162%), TDCA (394%), TMCA (117%), and TCA (270%) (Figure 7A). The administration of carvedilol to FFC mice significantly increased the intestinal content of TUDCA (44%) and THCA (179%) (Figure 7A). The most abundant BAs in feces were DCA and βMCA (Figure 7B;Supplementary Table 15). The spectra of individual BAs were shifted in NASH mice with a significantly enhanced fecal excretion of UDCA (518%), DCA (314%), βMCA (417%), and CA (1227%) (Figure 7B). Carvedilol significantly increased the fecal excretion of HCA (405%) in NASH mice compared with vehicle administered FFC-fed mice (Figure 7B).

Figure 7.

Effect of carvedilol on individual bile acids in the small intestine and feces. Small intestinal BA content (A), and BA fecal excretion (B) were analyzed in mice treated with either vehicle or carvedilol (10 mg/kg, daily p.o. for 3 weeks) at the end of a 24-week feeding cycle with either a chow diet or a diet rich in saturated fat, fructose, and cholesterol (FFC). Animal labeling: CV (chow diet with the vehicle), CC (chow diet with carvedilol), FV (FFC diet with the vehicle), and FC (FFC diet with carvedilol). Statistical significance was determined using either a 1-way ANOVA with Holm-Sidak’s post hoc test for data with Gaussian distribution or the Kruskal-Wallis 1-way analysis with Dunn’s post hoc test if at least one mouse group demonstrated a non-normal distribution for the evaluated parameter. Data are presented as medians with boxes and whiskers representing the interquartile range and 5th to 95th percentiles, respectively (n = 8). Asterisks (*) denote statistical differences from the vehicle chow diet CV group (p < .05). Daggers (†) denote statistical differences from the vehicle FFC diet FC group (p < .05).

Carvedilol is not an Fxr agonist

The differential effect of carvedilol on the expression of liver and intestinal BA-related genes, which are regulated by the Fxr, prompted us to analyze whether carvedilol may be an Fxr agonist. We found that carvedilol significantly decreased Nr0b2 (Shp) mRNA expression (55%), but induced Nr1h4 (Fxr) mRNA expression (36%) in mouse enteroendocrine GLUTag cells, a model for Fxr activation in enterocytes (Figure 8A). However, carvedilol showed no effect on the expression of another Fxr target, Fgf15 and it did not significantly regulate the expression of Fxr target genes, such as Nr0b2, Abcb11, or Slc51b in the murine AML-12 hepatocyte cell line (Figure 8B). Obeticholic acid, a known agonist of Fxr significantly stimulated mRNA of its downstream genes, such as Fgf15 (390-fold), Nr0b2 (81-fold), Abcb11 (6-fold), and Slc51b (62-fold) in both cell types (Figs. 8A and 8B).

Figure 8.

The effect of carvedilol on murine Fxr (Fxr) activation and regulation of Fxr target genes. Regulation of Fxr target genes by carvedilol in murine enterocyte GLUTag (A) and hepatocyte AML-12 cells (B) treated with the prototype FXR/Fxr agonist obeticholic acid (OCA, 10 µM) or carvedilol (CARV, 10 µM) for 24 h. Data were normalized to the expression β2 microglobulin mRNA and are expressed as relative mRNA expression to that of the control (vehicle-treated) cells. Luciferase reporter gene assays using pFXRE-luc or pSHP-luc constructs in combination with the murine Fxr expression vector were performed on transiently transfected GLUTag cells (C). Cells were treated with obeticholic acid (OCA), GW4064, chenodeoxycholic acid (CDCA) (all FXR agonists), and carvedilol (CARV) at a 10 µM concentration for 24 h. Data were normalized to Renilla luciferase activity and expressed as fold-activation relative to the control (vehicle-treated) cells. Kruskal-Wallis 1-way test with Dunn’s post hoc test was used for statistical analysis. Values are presented as medians with boxes and whiskers representing the interquartile range and 5th to 95th percentiles of 3 independent experiments (n = 4). *p < .05, **p < .01, ***p < .001 versus vehicle-treated controls (Ctrl).

In the transient transfection assays using the FXRE-driven or SHP gene promoter-driven luciferase constructs with the murine Fxr expression vector, we confirmed that known Fxr agonists, such as OCA, GW4064, or CDCA, exhibited the capacity to activate murine Fxr (Figure 8C). However, carvedilol did not activate murine Fxr (Figure 8C). Thus, the absence of a carvedilol effect on Fxr in most assays and the limited contradictory results in enterocytes indicate that carvedilol is not an agonist of the Fxr receptor.

Carvedilol alleviates NASH

To further analyze the reason for discrepancies between increased BA plasma concentrations after carvedilol in chow diet-fed animals and the absence of an effect in NASH animals, we evaluated the effect of carvedilol on NASH pathology (Figure 9). A 24-week FFC diet in mice significantly increased the body weights of mice by 52% (Figure 9A), indicating the reproduction of metabolic syndrome. Treatment with carvedilol did not change the body weights of these animals. The FFC diet also significantly increased liver weight by 153% (Figure 9B) and carvedilol significantly reduced this increase by 15%. Thus, liver-to-body weight ratios were significantly increased by 59% in vehicle-administered FFC mice (Figure 9C), which was significantly ameliorated by carvedilol (13%). Hematoxylin & eosin staining of liver sections (Figure 9D) revealed that hepatomegaly in FFC-fed mice resulted from extensive steatosis, which was significantly reduced by carvedilol treatment (16%). Increased fat accumulation in the liver of FFC mice was further confirmed by a 5-fold increase in triglyceride content (Figure 9E) compared with the chow diet group. The administration of carvedilol significantly decreased the abundance of triglycerides in steatotic livers by 12%. The analysis of genes associated with fatty acid (FA) liver homeostasis indicated that the mechanism of carvedilol-mediated amelioration of liver steatosis in NASH animals may be a significant mRNA repression and protein downregulation of stearoyl-CoA desaturase-1 (Scd1) (27% and 57%, respectively), a key enzyme in monounsaturated FA synthesis (Supplementary Figure 2). Liver impairment by persistent NASH was further verified by a significant increase in plasma activities of alanine aminotransferase (ALT) (533%), alkaline phosphatase (ALP) (56%), and aspartate aminotransferase (AST) (171%) (Figure 9F), indicating combined hepatocellular and cholestatic liver impairment. Compared with untreated FFC-fed mice, ALT (36%) and ALP (32%) were significantly reduced in carvedilol-treated NASH animals. These findings indicate a significant protective effect of carvedilol on liver steatosis caused by an FFC diet.

Figure 9.

Carvedilol reduced liver injury induced in mice by an FFC diet. Body weight (A), liver weight (B), liver to body weight ratio (C), steatosis in paraffine-embedded hematoxylin- and eosin-stained livers (scale bar 50 µm) (D), liver triglyceride content (E), and liver enzymes in plasma (F) were analyzed in mice that were treated with either vehicle or carvedilol (10 mg/kg, daily p.o. for 3 weeks) at the end of a 24-week feeding cycle with either a chow diet or a diet rich in saturated fat, fructose, and cholesterol (FFC). Animal labeling: CV (chow diet with the vehicle), CC (chow diet with carvedilol), FV (FFC diet with the vehicle), and FC (FFC diet with carvedilol). Statistical significance was determined using either a 1-way ANOVA with Holm-Sidak’s post hoc test for data with Gaussian distribution or the Kruskal-Wallis 1-way analysis with Dunn’s post hoc test if at least one mouse group demonstrated a non-normal distribution for the evaluated parameter. A Mann-Whitney test was used to compare steatosis (D). Data are presented as medians with boxes and whiskers representing the interquartile range and 5th to 95th percentiles, respectively (n = 8). Asterisks (*) denote statistical differences from the vehicle chow diet CV group (**p < .01, ***p < .001). Daggers (†) denote statistical differences from the vehicle FFC diet FC group (^†p < .05, ^††p < .01). Abbreviations: ALT, alanine aminotransferase; ALP, alkaline phosphatase; AST, aspartate aminotransferase.

BA homeostasis is markedly impaired by ongoing liver inflammation and fibrosis. To further understand these processes, we analyzed several crucial parameters (Figure 10). Sirius red staining confirmed liver fibrosis in FFC-fed mice, which was significantly attenuated by carvedilol treatment (55%) (Figure 10A). Hepatic macrophage infiltration was examined by immunohistochemistry for macrophage galactose-specific lectin (Mac-2), a marker of macrophages engaged in phagocytosis (Figure 10B). An FFC diet caused a significant increase in the Mac-2-immunopositive area in NASH animals (215%), which was decreased by carvedilol treatment (43%) (Figure 10B). The FFC diet significantly increased the expression of pro-inflammatory and fibrotic genes such as Tgfβ1 (92%), Acta2 (αSMA) (95%), Col1a1 (1321%), Ccl2 (445%), and Vcam1 (178%) (Figure 10C). The protective effect of carvedilol resulted in decreased hepatic expression of Tgfβ1 (11%), Col1a1 (66%), Ccl2 (53%), and Vcam1 (30%) mRNA when compared with the untreated NASH group. In addition, the expression of fibrotic markers was analyzed in cultured primary human HSCs following TGF-β1 exposure (Figure 10D). The addition of carvedilol to the culture media reduced the expression of ACTA2 (69%) and COL1A1 (43%) in TGF-β1-treated HSCs. These results confirm a beneficial effect of carvedilol on hepatic inflammation and fibrosis in NASH.

Figure 10.

Carvedilol ameliorates liver inflammation and fibrosis induced in mice by FFC diet. Livers were harvested to analyze images (scale bar 100 µm) of Sirius red staining (A), Mac-2 immunostaining (B), and the mRNA expression of pro-inflammatory and fibrotic genes (C) in mice that were treated with either vehicle or carvedilol (10 mg/kg, daily p.o. for 3 weeks) at the end of a 24-week feeding cycle with either a chow diet or a diet rich in saturated fat, fructose, and cholesterol (FFC). Animal labeling: CV (chow diet with the vehicle), CC (chow diet with carvedilol), FV (FFC diet with the vehicle), and FC (FFC diet with carvedilol). The antifibrotic effect was also evaluated in human hepatic stellate cells (HSC) cultured alone or with 3 ng/ml transforming growth factor beta 1 (TGF-β1) or 3 ng/ml TGF-β1 and 10 µM carvedilol (CARV) (D). Statistical significance was determined using either a 1-way ANOVA with Holm-Sidak’s post hoc test for data with Gaussian distribution or the Kruskal-Wallis 1-way analysis with Dunn’s post hoc test if at least one mouse group demonstrated a non-normal distribution for the evaluated parameter. Kruskal-Wallis 1-way test was used for the analysis of data from HCS cells. Data are presented as medians with boxes and whiskers representing the interquartile range and 5th to 95th percentiles, respectively (n = 8 for mice, n = 3 for cells). *p < .05, **p < .01, ***p < .001 versus vehicle-administered chow diet-treated mice (in cells: TGF-β1 vs vehicle-treated cells); ^†p < .05, ^††p < .01, ^†††p < .001 carvedilol-treated versus vehicle-treated NASH mice (in cells: TGF-β1 + CARV vs TGF-β1-treated cells).

Discussion

The effect of carvedilol, a nonselective β-adrenoreceptor antagonist, on BA turnover has not been studied comprehensively. A major finding reported here is that carvedilol increases BA plasma concentrations in chow diet-fed mice with intact livers. Unchanged BA ratios in all matrices analyzed suggest that increased plasma BA levels are not primarily caused by final steps in neutral or acidic liver BA synthesis or intestinal bacterial metabolism. These findings are consistent with the unaltered expression of end-phase BA synthetic enzymes in the liver. Similarly, unaltered BA biliary secretion, small intestinal content, and fecal excretion indicate that BA reabsorption in the ileum is not changed by carvedilol in chow diet-fed mice. Instead, proportional changes in the BA spectra were observed across matrices without a significant shift toward specific modification of Cyp7a1, the rate-limiting enzyme for BA synthesis and BA transporters. In fact, we detected significant post-transcriptional downregulation of Ntcp, the major uptake transporter of BAs at the basolateral membrane of hepatocytes, and upregulation of Cyp7a1. Downregulation of the BA basolateral efflux Mrp4 transporter suggests that the observed increase in plasma BA concentration was the primary consequence of their reduced uptake from portal blood to hepatocytes by Ntcp.

The biliary secretion of BAs was not reduced despite the downregulation of Bsep, the rate-limiting transporter for BA biliary secretion. It is possible that decreased efflux of BAs from hepatocytes to plasma via down-regulated Mrp4 together with increased BA synthesis by induced Cyp7a1 compensated reduced BA uptake via Ntcp and maintain sufficient BA subapical membrane concentrations in hepatocytes to overcome reduced Bsep expression and preserve BA biliary secretion. Importantly, as a second key mechanism for bile formation, the biliary secretion of glutathione was also unchanged by carvedilol treatment in chow diet-fed mice which contributed to unchanged bile flow. The major implication of these results is that, for the first time, they present a mechanism of cholestatic liver injury occasionally described during carvedilol therapy (Rua et al., 2019). Increased systemic concentrations of BAs may also explain pruritus detected in these patients. Therefore, the measurement of plasma BA levels may be recommended in patients with carvedilol-induced liver injury.

The decreased expression of Bsep, Ntcp, and Mrp4 protein is consistent with precious studies demonstrating that cAMP, a second messenger produced by activated β-adrenoreceptors, increases hepatocyte transcellular BA transport by upregulating Ntcp in the basolateral membrane (Mukhopadhayay et al., 1997) and Bsep in the canalicular membrane (Misra et al., 2003). Molecules that increase cAMP, such as glucagon, salbutamol, or UDCA, subsequently prevent ethinylestradiol-induced cholestasis in rats (Li et al., 2016; Zucchetti et al., 2013). Interestingly, cAMP increases the trafficking of BA transporters to the plasma membrane through a protein kinase A-independent mechanism involving activation of the cAMP-Epac1 pathway (Zucchetti et al., 2011). The activation of cAMP-Epac1 also induces Mrp4 expression (Bröderdorf et al., 2014). Consistent with this concept of Ntcp, Bsep, and Mrp4 regulation, we observed the concomitant post-transcriptional repression of these molecules with a reduction of the cAMP sensor, Epac1. This suggests that carvedilol reduces Ntcp and Bsep proteins in chow diet-fed mice directly through β₂-adrenoreceptor-mediated blockade of the cAMP-Epac1 pathway. In contrast, Mayati et al. (2017) found repressed Ntcp and Bsep mRNA expression by activation of β₂-adrenoreceptor-mediated cAMP production in primary human hepatocytes and HepaRG cells. However, we did not observe this effect in mice and were unsuccessful in reproducing the repression of ABCB11 (BSEP protein) and SLC10A1 (NTCP protein) by β₂-adrenoreceptor agonist salbutamol in HepaRG cells. Therefore, we propose that cAMP-dependent post-transcriptional modulation of liver BA transporters is the major mechanism of increased plasma BAs concentrations in carvedilol-treated mice on a chow diet.

Carvedilol treatment in NASH mice did not change net plasma concentrations, liver content, biliary secretion, small intestinal content, or fecal excretion of BAs. However, carvedilol modified BA composition with increased concentrations of hydrophilic αMCA in the plasma and liver, and elevated (T)HCA in the bile, small intestine, and stool. These changes are consistent with augmented αMCA synthesis via Cyp2c70 induced in the liver (de Boer et al., 2020). The expression of Cyp2c70 mRNA is maintained by TGR5 as mice devoid of this BA receptor lack Cyp2c70 mRNA (Carino et al., 2021). However, the upregulation of Cyp2c70 by carvedilol in FFC-fed animals occurred post-transcriptionally. Therefore, the possibility of direct regulation of Cyp2c70 or Cyp7a1 by β-adrenoreceptors should be further analyzed. Additionally, αMCA was not detected in the stool as it is converted into βMCA by C7-epimerization mediated also by Cyp2c70 in the liver and into HCA by C6-epimerization mediated by bacteria in the intestine (de Boer et al., 2020). Increased (T)HCA in the bile, small intestine, and stool and reduced primary/secondary BAs ratio in the stool suggest that carvedilol increased the synthesis of secondary BAs by altering intestinal microbiome.

Indeed, expansion of the Bacilli class of the Firmicutes phylum was detected in β₁/β₂ adrenoreceptor knockout mice (Bartley et al., 2018). Firmicutes phylum is involved in the 7α-dehydroxylation of primary into secondary BAs. Most studies reported an increased Firmicutes-to-Bacteroidetes ratio in NASH (Albhaisi et al., 2020), which is consistent with our microbiome analysis and an increased proportion of secondary BAs in the intestine of NASH mice. Carvedilol in NASH mice shifted the spectra of intestinal bacteria toward a physiological status by reducing Firmicutes phylum. However, HCA synthesis was not inhibited by this change. An important implication of this finding is that HCA accumulation may exert a significant metabolic effect by activating the TGR5 receptor (Zheng et al., 2021). Interestingly, we also observed induction of Fxr-regulated genes in ileal enterocytes of the carvedilol-treated NASH group. A detailed in vitro analysis revealed the absence of a consistent stimulating effect of carvedilol on murine Fxr. The induction of Fxr target genes in the ilea could be therefore indirect, related to the reduced reabsorption of Fxr antagonistic BAs such as THCA (Zheng et al., 2021), as demonstrated by increased content in the feces of carvedilol treated NASH mice, which may shift the enterocyte concentrations in favor of Fxr agonistic BAs.

Carvedilol treatment of NASH animals did not alter the expression of Ntcp, Bsep, or Mrp4. These transporters were down-regulated by an FFC diet as previously described (Lastuvkova et al., 2021) and discrepancies between the mRNA and protein levels suggest a role for post-transcriptional regulation. Consistently, activated cAMP-dependent Epac1 protein was markedly decreased by NASH, whereas carvedilol did not aggravate this decrease. Therefore, we examined the effect of carvedilol on NASH pathology and a significant protective effect was observed. The amelioration of pro-cholestatic liver inflammation and fibrosis by carvedilol was demonstrated in rodent models of liver injury induced by ethanol (Araujo Junior et al., 2016), carbon tetrachloride (Abdel-Kawy, 2021), or bile duct ligation (Tian et al., 2017). These effects of carvedilol were attributed to the inhibition of IL-1β/TNF-α production and decreased NF-κB activation in hepatocytes, which in turn, prevented neutrophil infiltration and activation of Kupffer cells (Araujo Junior et al., 2016). In the present study, carvedilol reduced hepatic Tgf-β signaling consistent with its ability to inhibit the activation of hepatic stellate cells (Ling et al., 2019; Meng et al., 2018). Hepatic stellate cells express α/β-adrenoreceptors and respond by activation to increased sympathetic activity observed during NAFLD (Sigala et al., 2013). Thus, inhibiting this activation by carvedilol may represent a major mechanism of NASH-attenuation by this drug.

Consistent with our findings, Soliman et al. (2019) recently showed that carvedilol reduced hepatic α-smooth muscle actin immunopositivity (an indicator of stellate cells activation), malondialdehyde (oxidative stress marker), and triglyceride content in rats with diet-induced liver steatosis. We successfully reproduced the reducing effect of carvedilol on the excessive cumulation of triglycerides in the liver of NASH mice. A detail molecular analysis revealed that this occurs through repression of stearoyl desaturase 1, an important enzyme for monounsaturated FA synthesis (MUFA). A reduction in Scd1 activity in rodents by either genetic deficiency (Miyazaki et al., 2009) or treatment with inhibitor aramchol (Fernández-Ramos et al., 2020) significantly decreased liver triglyceride content (Dobrzyn et al., 2004). It appears that reduced hepatic MUFA attenuates the insulin-activated FA synthesis. Consistently, Shi et al. (2021) recently showed an induction of hepatic steatosis in mice using a β₂-receptor agonist via activation of lipogenesis. Taken together, the beneficial effects of carvedilol in NASH prevail and prevent further deterioration of BA homeostasis induced by NASH. Our results suggest the additive hepatoprotective potential of carvedilol when used as a part of cardiovascular regimens in patients with metabolic syndrome who are at high risk for NASH.

In conclusion, our findings show that carvedilol increases plasma BA concentrations in chow diet-fed mice by reducing uptake into hepatocytes through Ntcp downregulation. Carvedilol-induced inhibition of the β-adrenoreceptor-cAMP-Epac1 pathway may represent the post-transcriptional mechanism for this effect. Importantly, carvedilol did not augment plasma BA concentrations elevated by NASH. Instead, it enhanced the synthesis of the more hydrophilic muricholic acid and its secondary metabolite, hyocholic acid. This positive effect of carvedilol was associated with the significant amelioration of liver steatosis and fibrosis in NASH mice. Our results suggest beneficial effects of carvedilol in patients at high risk for NASH development.

Supplementary data

Supplementary data are available at Toxicological Sciences online.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The project was supported by grants SVV 260543/2020, GAUK 3462/18, GACR 22-05167S, and Projects INOMED CZ.02.1.01/0.0/0.0/18_069/0010046, and ESF-Project “International mobility of RTAS ChU No. CZ.02.1.01/0.0/0.0/17_048/0007421”. P.H. received support from the National Institute of Diabetes and Digestive and Kidney Diseases under Award Number R01DK130884.

Author contributions

Hana Lastuvkova: Methodology, Investigation, Writing—original draft. Zuzana Nova: Methodology, Software. Milos Hroch: Investigation, Methodology. Fatemeh Alaei Faradonbeh: Methodology. Jolana Schreiberova: Investigation, Methodology. Jaroslav Mokry: Validation, Visualization. Hana Faistova: Methodology, Validation, Visualization. Alzbeta Stefela: Investigation, Methodology. Jan Dusek: Investigation, Methodology. Otto Kucera: Investigation, Methodology. Radomir Hyspler: Methodology. Ester Dohnalkova: Investigation. Rachel L. Bayer: Methodology. Petra Hirsova: Validation, Writing—review & editing. Petr Pavek: Methodology, Resources, Validation. Stanislav Micuda: Supervision, Project administration, Resources, Funding acquisition, Writing—review & editing.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials. Individual values are available upon request.