BAR502, a drug activating two receptors, encourages white fat to become heat-producing and improves fatty liver and scarring

Design and synthesis of BAR502 (6α-ethyl-3α, 7α-dihydroxy-24-nor-5β-cholan-23-ol) has been described previously20. The agent was dissolved daily in drinking water containing 1% methyl cellulose before its administration by gavage (100 μl).

C57BL6 mice 24 weeks old were fed a high fat diet containing 60% kj fat (ssniff ^® EF acc. D12492 (I) mod.) and fructose in drinking water (42 g/l) or normal diet (6 mice) for 18 weeks11,12,13. After 10 weeks of HFD, mice were randomized to receive HFD alone (9 mice) or HFD plus BAR502 (15 mg/kg/day) body weight by gavage (9 mice) for 8 weeks. HFD mice were administered 100 μl of 1% methyl cellulose in drinking water. Mice were housed under controlled temperatures (22 °C) and photoperiods (12:12-hour light/dark cycle), allowed unrestricted access to standard mouse chow and tap water and allowed to acclimate to these conditions for at least 5 days before inclusion in an experiment. The care and use of the animals were approved by the Institutional Animal Care and Use Committee of the University of Perugia and were in accordance to European guidelines for care of experimental animals. Protocols were approved by the Italian Minister of Health and Istituto Superiore di Sanità (Italy) and were in agreement with the European guideline for use of experimental animals (permission n. 41/2014-B released to AD). The general health of the animals was monitored daily by the Veterinarian in the animal facility. The study protocol caused minor suffering, however, animals that lost more than 25% of the initial body weight were euthanized. At the day of sacrifice mice were deeply anesthetized with a mixture of tiletamine hypochoride and zolazepam hypocloride/xylazine at a dose of 50/5 mg/Kg. Food intake was estimated as the difference of weight between the offered and the remnant amount of food at 7-days intervals. The food was provided as pressed pellets and the residual spillage was not considered.

Liver fibrosis was induced by carbon tetrachloride (CCl₄) administration according to a previously published method21. For this purpose, C57BL6 mice were administered i.p. 500 μL/Kg body weight of CCl₄ in an equal volume of paraffin oil twice a week for 8 weeks. CCl₄ mice were randomized to receive BAR502 (15 mg/Kg daily by gavage) or vehicle (distilled water).

For histological examination, portions of the right and left liver lobes and epididymal white adipose tissue (epWAT) were fixed in 10% formalin, embedded in paraffin, sectioned and stained with Sirius red and Hematoxylin/Eosin (H&E), for morphometric analysis. NASH severity was scored in H&E-stained cross sections using an adapted grading system of human NASH as described previously6,22. The level of macrovesicular and microvesicular steatosis was determined at 40 × magnification relative to the total liver area analyzed and expressed as a percentage. Inflammation was scored by counting the number of aggregates of inflammatory cells per field using a 100 × magnification. The average of five random fields was taken. Relative values against the average of the HFC control group were calculated. Hepatic fibrosis was identified using Sirius Red stained slides and evaluated using an adapted grading system of human NASH as described previously. The presence of pathological collagen deposition was scored as either absent (0), observed within perisinusoidal/perivenular or periportal area (1), within both perisinusoidal and periportal areas (2), bridging fibrosis (3) or cirrhosis (4). The sum of the scores (degree of steatosis, hepatocyte ballooning, lobular inflammation, and portal inflammation) was considered as the total pathology score.

AST, triglyceride, total- and HDL- cholesterol, albumin and fasting glucose concentrations were quantified using an automated clinical chemistry analyzer (Cobas, Roche) with enzymatic methods. Plasma insulin levels were measured by ELISA assays according to the manufacturer’s instructions. The homeostatic model assessment (HOMA) of IR and insulin sensitivity index were calculated by the concentrations of plasma glucose and insulin.

For determination of total triglyceride and cholesterol content, liver samples from left hepatic lobules of ∼100 mg of liver were homogenized with 1 ml of T-PER (Pierce). The homogenates were used for protein concentration analysis (Bradford assay, Bio-rad, Milan, Italy), and 100 μl of tissue extracts added to 1.6 ml CHCl₃: methanol (2:1) for 16 h at 4 °C, after which 200 μL of 0.6% NaCl was added and the solution centrifuged at 2,000 g for 20 min. The organic layer was removed and dried by Speed Vac System (HETO-Holten, Waltham, MA). The resulting pellet was dissolved in 100 μL phosphate buffered saline containing 1% Triton X-100 and triglyceride, cholesterol, and FFA content was measured by specific enzymatic reagents.

After 9, 13 and 18 weeks of HFD administration, the mice were fasted overnight and orally administered glucose (1.5 g/kg body weight) for OGTT or fasted for 4 h and intraperitoneally injected insulin (0.35 unit/kg body weight) for ITT. The blood glucose concentrations were measured at 0, 15, 30, 60, 90, and 120 min after feeding or injection using a portable glucose meter (Accu-Check Go, Roche). Plasma insulin levels were measured by Mercodia Ultrasensitive Mouse Insulin ELISA assays according to the manufacturer’s instructions.

Total RNA was extracted from liver, epWAT, BAT, gastrocnemius and intestine (4 mice/group) using TRIzol reagent (Invitrogen) and reverse transcribed with SuperScript II Reverse Transcriptase (Invitrogen) following the manufacturer’s instructions. A total of 10 ng of cDNA was pipetted into each well of a 96-well Custom RT² Profiler PCR ArrayCAPM13349, RT² Profiler PCR Array, Qiagen (http: www.qiagen.com/it/products/catalog/assay-technologies/real-time-pcr-and-rt-pcr-reagents/rt2-qpcr-primer-assays/l↗) and amplified following the manufacturer’s instructions in StepOnePlus instrument (Applied Biosystems). This RT² PCR array is designed to assess a total of 75 genes:30 genes for liver, 6 for gastrocnemius, 6 for intestine, 22 for epWAT and 11 for BAT; 2 housekeeping genes were used for each tissue for data normalization. The RT² Profiler PCR Array also includes control elements for: Genomic DNA contamination detection, RNA sample quality and General PCR performance. Data analysis is based on the ΔΔC_T method with normalization of the raw data to either housekeeping genes.

RNA extracted from liver, epWAT, BAT, gastrocnemius and intestine was subjected to reverse transcription; for Real Time PCR 10 ng cDNA were amplified in a 20 μL solution containing 200 nM of each primer and 10 μL of 2X SYBR FAST Universal ready mix (Invitrogen). All reactions were performed in triplicate, and the thermal cycling conditions were as follows: 10 min at 95 °C, followed by 40 cycles of 95 °C for 10 sec, 55 °C for 10 s and 60 °C for 30 sec in StepOnePlus instrument (Applied Biosystems). The relative mRNA expression was calculated and expressed as 2^−(ΔΔCt). Expression of the respective genes was normalized to both B2M and GAPDH mRNA as an internal controls. PCR primer sequences were:mouse-GAPDH: ctgagtatgtcgtggagtctac and gttggtggtgcaggatgcattg; mouse-B2M: ctttctggtgcttgtctcactg and ttcagcatttggatttcaatgt; mouse-ATG5: gcatcaagttcagctcttcctt and gcatcagcttctttcatacac; mouse-ATG7: tttctgtcacggttcgataatg and gtgaatccttctcgctcgtact; mouse-ATG12: gctgaaggctgtaggagacact and gagttccaacttcttggtctgg; mouse-LC3a: ctgtaaggaggtgcagcagat and cttgaccaactcgctcatgtta; mouse-adiponectin: gacaggagatcttggaatgaca and gaatgggtacattgggaacagt; mouse-UCP1: ctcactcaggattgtgcctctac and tctgaccttcacgacctctgta; mouse-PPARα: cagaggtccgattcttccac and gatcagcatcccgtctttgt; mouse-Rantes: ctgctgctttgcctacctct and tccttcgagtgacaaacacg ; mouse- IL6: cttcacaagtcggaggctta and ttctgcaagtgcatcatcgt; mouse PPARγ: gccagtttcgatccgtagaa and aatccttggccctctgagat ; mouse aP2: aaacaccgagatttccttcaaa and tcacgcctttcataacacattc; mouse-UCP3: ccaacatcacaagaaatgc and tacaaacatcatcacgttcc.

Immunohistochemistry on white adipose tissue and on brown adipose tissue (used as positive control) were performed. Formalin-fixed paraffin-embedded sections (7 μm thick) were deparaffinized and then antigen retrieval was performed by treatment with protease. Sections were then washed in PBS. and treated with hydrogen peroxide (3%) for 5 minutes to block endogenous peroxidase. Non specific binding was blocked by treatment with BSA (Bovine Serum Albumin) 8% plus Triton 0,2% in PBS for 30 minutes. Anti UCP-1 (Abcam 1:500) in PBS (2% BSA +0,05% Triton) was incubated overnight at +4 °C. After washing, anti-rabbit IgG (1:200 Vector) in PBS was incubated for 1 hour, section were washed and incubated with ABC (Vectastain ABC Kit elite) and then, after washing with DAB peroxidase substrate kit (Vector), according to the manufacturer instructions. Sections were counterstained with haematoxylin, dehydrated and coverslips mounted with mounting medium.

3T3-L1 preadipocytes were plated at 37.5 × 10³ cells/well in 24-well plates. Preadipocytes were grown to confluence in DMEM and 10% FCS23. Two days after reaching confluence (day 0), the cells were stimulated with differentiation mix (Mix) alone (DMEM, 10% FCS, 517 μM 3-isobutyl-1-methylxanthine, 1 μM dexamethasone, and 172 nM insulin). At day 2, Mix was replaced by insulin medium (DMEM, 10% FBS, and 172 nM insulin) alone. At day 4 cells were stimulated 24 hours with 10 μM BAR 502 or 10 μM BAR501.

Total lysates were prepared by solubilization of 3T3L1 cells in E1A lysis buffer (250 mM NaCl, 50 mM Hepes pH 7.0, 0.1% NP40, 50 mM EDTA) containing phosphatase and protease inhibitors. Protein extracts were electrophoresed, blotted to nitrocellulose membrane and then incubated with primary antibodies against UCP-1 (Abcam) and Tubulin (Sigma). Primary antibodies were detected with the horseradish peroxidase (HRP)-labeled secondary antibodies. Proteins were visualized by SuperSignal West Dura Extended Duration Substrate (Thermo Fisher Scientific Inc.) according to the manufacturer’s instructions.

Bile acids pools were measured by liquid chromatography-tandem mass spectrometry (MS/MS) analysis, using chromatographic conditions as described elsewhere24. The stock solutions of the individual tauro-conjugated and un-conjugated bile acids were prepared separately in methanol at a concentration of 1 mg/mL. All stock solutions were stored at −20 °C. Calibration standards were prepared by combining appropriate volumes of each bile acid stock solution and methanol. The calibration range was from 10 nM to 100 μM of each bile acid in the final solution. Mice serum sample aliquots of 20 μL were mixed with 80 μL of CH₃OH, shaken continuously, vortexed and, after centrifugation at 16000 g for 10 min, the clear supernatant was transferred to a new vial, snap frozen and lyophilized. The sample was then re-dissolved in 40% water: 60% MeOH with 0.1% formic acid and ammonium acetate 5 mM. A bile acids extraction yield of 95% has been measured using bile acids standards addition in plasma sample before and after deproteinization procedure.

For gallbladder and fecal samples, 10 mg of lyophilized gallbladder or feces was manually pestled using a mortar and dissolved in 1 mL CH₃OH. After centrifugation at 16000 g for 10 min, 500 μL of supernatants were lyophilized and reconstituted in 100 μl of 40% water: 60% MeOH with 0.1% formic acid and ammonium acetate 5 mM.

For LC–MS/MS analysis, chromatographic separation was carried out on the HPLC–MS system Q-TRAP 6500 LC-MS/MS System from AB Sciex equipped with Shimadzu LC-20A LC and Auto Sampler system. The mixture was separated on a Synergi Fusion RP 4 μ 80°A from Phenomenex (150 × 2.00 mm). Tauro-conjugated and non-conjugated bile acids were separated at a flow rate of 200 μl/min using a methanol–aqueous ammonium acetate (NH₄OAc) gradient. Mobile phase A was water containing 5 mM ammonium acetate and 0.1% formic acid, mobile phase B was methanol, containing ammonium acetate at 5 mM and 0.1% formic acid. The gradient started at 65% B and increased to 85% B in 23 min, kept at 85% B for 5 min then decreased to 65% B in 1 min and kept at 65% B for 10 min. ESI was performed in negative ion mode, the ion source temperature was set at 280 °C. The tune page parameters were automatically optimized injecting taurocholic acid at 1 μM as standard.

All of the data are shown as the means ± SEM. Difference among group means were estimated using one-way ANOVA followed by Tukey’s post hoc test by GraphPad Prism 5.0 software. Significance was set up at p < 0.05.

To induce NASH-like features, C57BL6 mice were fed a HFD and fructose (collectively HFD) for 18 weeks. HFD mice gained significantly more weight than naïve mice starting from week 2. After 4 weeks of HFD food intake in this group started to decrease and the trend was maintained throughout the study (Fig. 1A,B). After 9 weeks, mice feed the HFD were significantly more heavy (Fig. 1A) than naïve mice (≈30% increase of body weight compared to 10%, P < 0.01), had higher glucose plasma levels and were insulin resistant as demonstrated by glucose response to OGTT and ITT (Fig. 1B, upper and lower panel; P < 0.05 versus naïve mice).

HFD, insulin-resistant, mice were than randomized into two groups: HFD alone or HFD in combination with BAR502, 15 mg/kg/day, and followed for additionally 9 weeks (see above for the protocol). At the end of the study, HFD mice developed biochemical and histopathology features of NASH, including elevated AST, cholesterol and HDL plasma levels (Fig. 1C; P < 0.01), while plasma triacylglycerol levels were similar to that of naïve mice.

While the liver weight and the percent of liver weight to the body weight did not changed in this dietary model in comparison with naïve mice of the same age (Fig. 1F), the histopathology analysis, the liver of HFD mice showed evidence of NASH-like features, including severe steatohepatitis with extensive lobular fat deposition, diffuse hepatocytes ballooning and microvescicolar steatosis (Fig. 1D, middle panel and inset). Mice on HFD had significantly higher steatosis scores and increased liver content of triacylglycerol and cholesterol than control mice (Fig. 1G, P < 0.01 versus naïve mice). Additionally, HFD mice developed moderate liver fibrosis as measured by morphometric analysis of fibrotic areas in liver sections immune-stained with Sirius red and by assessment of the expression of αSMA and α1-Collagen, two pro-fibrogenetic genes (Fig. 1H and I; P < 0.01 versus naïve). Finally, feeding a HFD caused an influx of inflammatory cells as assessed by measuring the inflammatory score at histopathology and expression of F4/80 (a marker for macrophages) along with IL-1β, IL-6, MCP-1, RANTES and TNFα mRNA (Fig. 1L, P < 0.05 versus naïve).

As shown in Fig. 2, eating of a HFD increased the expression of genes encoding for proteins involved in regulating cholesterol and triacylglycerol metabolism: Srepb1c, Fas ApoC2, Cpt1, CD36 and Pparα and γ (Fig. 2A–C), while the expression of other nuclear receptors such as Lxr, Fxr and Shp was unchanged.

These NASH-like features were reversed by feeding mice with BAR502. As detailed above, treatment with BAR502 was initiated after 9 week of HFD, i.e. at a time point at which mice had reach a stable body weight and were insulin resistant. Thus, from week 9 to week 18 mice administered the HFD alone gained an additional 3% of body weight, while treating mice with BAR502 resulted in ≈10% decrease of body weight (P < 0.05 versus HFD). These change associates with robust remodeling of plasma lipid metabolism profile, since BAR502 increased HDL (Fig. 1C). Administering HFD mice with BAR502 reversed liver lipid accumulation resulting in a robust reduction of liver steatosis score and along with reduction of liver lipid content (Fig. 1D–G; P < 0.05 versus HFD). In addition, BAR502 reduced liver fibrosis scores by 70% and expression of pro-fibrogenetic genes (αSMA and α1-collagen), as well as the inflammatory score and liver expression of F4/80, IL-1β, IL-6, MCP-1, RANTES and TNFα mRNA (Fig. 1H–L).

These phenotypic changes associate with remodeling of the liver expression of genes involved in lipid metabolism. As illustrated in Fig. 2A–C, BAR502 effectively reduced the relative expression of Fas and Srepb1c mRNA (n = 5, p < 0.05). Quantitation of Cyp7α1 and Bsep mRNA levels confirmed that BAR502 in vivo behaves as full FXR agonist as it reduced Cyp7α1 while induced Bsep and Shp (Fig. 2, n = 5, p < 0.05).

Amelioration of hepatic steatosis by BAR502 was also linked to the modulation of uptake and exportation of cholesterol from the liver, as demonstrated by the investigation of genes involved in cholesterol metabolism. Indeed, BAR502 significantly reduced the mRNA level of CD36 and upregulated the expression of ABCG5, a membrane transporter involved in cholesterol efflux from hepatocytes (Fig. 2; n = 5, p < 0.05). BAR502 reduced Pparγ mRNA (Fig. 2C; n = 5, p < 0.05).

The above mentioned changes demonstrate that feeding mice with BAR502 reshapes liver lipid partitioning in this dietary model of NASH. Because the WAT is the major storage tissue for fat, we have then examined whether BAR502 modulates fat depots in the model. Eating a HFD for 18 weeks increased of weight of epWAT by 3-fold (P < 0.01) while the BAT weight remained unchanged (Fig. 3C and H). BAR502 caused a further increase in epWAT weight but it also increased the amount of BAT (Fig. 3A–C and H; P < 0.05 versus HFD). These changes associate with the browning of epWAT as assessed by morphometry and gene array (3B). Exposure to the high fat diet increased the mRNA levels of Fas, leptin and adiponectin (Fig. 3D–F, n = 5, p < 0.05). In contrast, exposure to BAR502 reduced the size of adipocytes and effectively induced the expression of Ucp1, a marker of brite/beige trans-differentiation, as assessed by RT-PCR and immune-histochemistry (Fig. 3B and E) and expression of Cited1 mRNA (Fig. 3E, n = 5, p < 0.05). As illustrated in Fig. 3G, the BAT expresses very high levels of UCP1 at immune-histochemistry and treating mice with BAR502 increased the weight of BAT (Fig. 3H), but had no effect on the level of major genes involved in lipid metabolism in this tissue, including relative expression of Ucp1 mRNA (Fig. 3I, n = 5, p < 0.05).

Administering HFD mice with BAR502 also resulted in a 2–6 folds induction of Ucp1, 2 and 3 mRNA in muscles, further highlighting that BAR502 has the ability to redirect the metabolism of lipid in the whole body (Fig. 4). Additionally, BAR502 increased the expression of Fgf-15 and Shp (two FXR target genes), and Glp1 (a GPBAR1 target gene) in the intestine (P < 0.05 versus HFD) demonstrating that this agent acts as a dual FXR and GPBAR1 ligand in vivo.

Treating mice with BAR502 increased insulin sensitivity as measured by assessing the OGTT and ITT after 4 weeks administration, as shown in Fig. 5A (P < 0.05 versus HFD). The beneficial effect of BAR502 on OGTT was maintained up to the end of the study (data not shown). However, BAR502 had a minor impact on insulin plasma levels measured at the end of the study (Fig. 5B, n = 9) and liver expression of genes involved in glucose metabolism (Fig. 5C, n = 5; *P < 0.05 versus naïve; #P < 0.05 versus HFD). This could partially due to fructose feeding that was continued through the whole duration of the study including the day of sacrifice.

As shown in Fig. 6A and B, feeding mice a HFD increased the fecal concentrations of nonconjugated bile acids, CA, HCA, UDCA and DCA (n = 5, p < 0.05). Moreover, the HFD fat diet significantly increased the fecal concentration of tauro(T)-conjugated bile acids (TβMCA, TCA, THCA, TCDCA and TDCA) (n = 5, p < 0.05). Administering mice with BAR502 attenuated these changes and significantly reduced the fecal concentrations of all T-conjugated bile acids (n = 5, p < 0.05) as well as that of primary bile acids, CA and CDCA, but increased the excretion of HCA, UDCA and LCA (n = 5, p < 0.05).

The ability of BAR502 to induce the browning of WAT, was confirmed in vitro. For this purpose 3T3-L1 pre-adipocytes were differentiated into mature white adipocytes for 4 days and then exposed for 24–48 hours to BAR502, 10 μM as shown in Fig. 7A. The relative mRNA expression of adipogenic genes (i.e. adiponectin, PPARα, PPARγ and FABP4) and genes involved in brite differentiation (i.e. UCP1) and autophagic genes (i.e. ATG5, ATG7, ATG12 and LC3α) was measured at the end of the study (Fig. 7B). Despite the fact that exposure to BAR502 did not reverted expression of markers of adipocytes differentiation (adiponectin, PPARα, PPARγ and FABP4 (AP2), it increased the UCP1 mRNA and protein (Fig. 7B and C) and ATG7 and LC3α mRNAs (Fig. 7B; n = 3, p < 0.05).

Because fibrosis is an essential therapeutic target in liver disorders, and BAR502 reverses liver fibrosis in HFD model, we have then searched whether this compound reverts liver fibrosis in a model of severe extracellular matrix (ECM) remodeling induced by treating mice with CCl₄. As illustrated in Fig. 8, treating mice with CCl₄ resulted in severe liver injury as measured by assessing AST and bilirubin plasma levels, and caused a parenchymal cell extinction as indicated by the marked reduction of albumin plasma levels and increased portal pressure (Fig. 8A and B). These biochemical features associate with development of severe fibrosis as demonstrated by morphometric analysis of liver samples stained with Sirius red. The fibrotic area increased approximately 8-folds in response to CCL₄ administration. Supporting these phenotypic changes, the liver expression of pro-fibrogenetic genes, including α1-Collagen and αSMA was increased by 10–25 folds in response to CCl₄. Treating mice with BAR502 reversed this pattern and protected against liver fibrosis development. BAR502 reduced AST and bilirubin plasma levels, increased albumin plasma levels and reduced ECM deposition (Fig. 8), as measured by assessing the fibrosis score and expression of pro-fibrogenetic genes.

Exposure to BAR502 also reduced portal pressure (Fig. 8A). This effect on portal pressure was linked to reduction of ECM deposition but also involved reshaping of genes such as Cse, Cbs, Et-1 and caveolin 1 which are involved in generation of endogenous vasodilators. Finally, confirming the fact that this agent is a FXR ligand, treating CCl₄-mice with BAR502 resulted in a profound induction of Shp mRNA in the liver (Fig. 8H), and reduced the synthesis of primary acids (CA and CDCA) as shown in Fig. 9 (# denotes P < 0.05 versus CCl₄ alone).