Combined obeticholic acid and elafibranor treatment promotes additive liver histological improvements in a diet-induced ob/ob mouse model of biopsy-confirmed NASH

Obeticholic acid (OCA) and elafibranor (ELA) are selective and potent agonists for the farnesoid X receptor (FXR) and dual peroxisome proliferator-activated receptor α/δ (PPAR-α/δ), respectively. Both agents have demonstrated clinical efficacy in nonalcoholic steatohepatitis (NASH). The present study used OCA and ELA to compare the effects of mono- and combination therapies on metabolic and histological endpoints in Lepob/ob mice with established diet-induced and biopsy-confirmed NASH (ob/ob-NASH). ob/ob-NASH mice were fed the AMLN diet high in trans-fat, fructose and cholesterol for 15 weeks, whereafter they received vehicle, OCA (30 mg/kg, PO, QD), ELA (3, 10 mg/kg, PO, QD), or combinations (OCA + ELA) for eight weeks. Within-subject comparisons were performed on histomorphometric changes, including fractional area of liver fat, galectin-3 and Col1a1. OCA and ELA monotherapies improved all quantitative histopathological parameters and OCA + ELA combinations exerted additive effects on metabolic and histological endpoints. In agreement with their different molecular mechanisms of action, OCA and ELA monotherapies elicited distinct hepatic gene expression profiles and their combination led to profound transcriptome changes associated with further improvements in lipid handling and insulin signaling, suppression of immune responses and reduced extracellular matrix formation. In conclusion, these findings provide preclinical proof-of-concept for combined FXR and PPAR-α/δ agonist-based therapies in NASH.


Materials and Methods
Animals. The Danish Animal Experiments Inspectorate approved all experiments which were conducted using internationally accepted principles for the use of laboratory animals under the personal license #2013-15-2934-00784. B6.V-Lep ob /JRj (ob/ob) mice (6 weeks old) were from Janvier Labs (Le Genest Saint Isle, France) and housed in a controlled environment (12 h light/dark cycle, lights on at 3 AM, 21 ± 2 °C, humidity 50 ± 10%). Each animal was identified by an implantable subcutaneous microchip (PetID Microchip, E-vet, Haderslev, Denmark). Mice had ad libitum access to tap water and a diet high in fat (40%, containing 18% trans-fat; 40% carbohydrates, 20% fructose) and 2% cholesterol (AMLN diet; D09100301, Research Diets, New Brunswick, NJ) 34,35 for 15 weeks prior to treatment start and during drug treatment. All ob/ob animals underwent liver biopsy prior to treatment (see below), whereupon they were single-housed throughout the remainder of the study. An outline of the study design is shown in Supplementary Figure S1.
Baseline liver biopsy. The biopsy procedure was applied to all mice approximately three weeks before completion of the dieting period, as detailed previously 34 . In brief, mice were pretreated with enrofloxazin (Baytril ® , 5 mg/mL, 1 mL/kg; Bayer, Leverkusen, Germany) one day prior to biopsy. On the surgery day, mice were anesthetized with isoflurane (2-3%, in 100% oxygen), a small abdominal incision in the midline was made, and the left lateral lobe of the liver was exposed. A cone-shaped wedge of liver tissue (50-100 mg) was excised from the distal part of the lobe. The cut surface of the liver was closed by electrosurgical bipolar coagulation using an electrosurgical unit (ERBE VIO 100 C, ERBE, Marietta, GA). The liver was returned to the abdominal cavity, the abdominal wall was sutured and skin stapled. Carprofen (Rimadyl ® , 5 mg/mL, 0.01 mL/10 g; Pfizer, NY) and enrofloxazin (5 mg/mL, 1 mL/kg, i.p.) were administered at the time of surgery and at post-operative day one and two. Animals were single-housed after the procedure and recovered for three weeks prior to drug treatment. were freshly dissolved in 0.5% carboxymethyl cellulose and orally administered in a dosing volume of 2.5 ml/kg. Animals were stratified (n = 9-12 per group) based on mean fibrosis as assessed by collagen 1a1 (Col1a1) immunostaining (primary) and body weight (secondary). After 15 weeks on AMLN diet, mice were treated once daily for 8 weeks with vehicle, OCA (30 mg/kg), ELA (3 or 10 mg/kg), OCA (30 mg/kg) + ELA (3 mg/kg), OCA (30 mg/ kg) + ELA (10 mg/kg). To control for the two simultaneous drug injections, a vehicle injection was administered after monotherapy. Food intake (24 h) was measured once weekly. A terminal blood sample was collected from the tail vein in non-fasted mice and used for plasma biochemistry. Animals were sacrificed by cardiac puncture under isoflurane anesthesia. Liver samples were processed as described below.
Body weight and body composition analysis. Body weight was monitored daily during the intervention period. Whole-body fat and lean mass was analyzed in week 7 of the treatment period by non-invasive Plasma biochemistry. Terminal plasma concentrations of total triglycerides (TG), total cholesterol (TC), alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were determined as described previously 34 .
Liver histology and digital image analysis. Baseline liver biopsy and terminal samples (both from the left lateral lobe) were fixed overnight in 4% paraformaldehyde. Liver tissue was paraffin-embedded and sectioned (3 µm thickness). Sections were stained with hematoxylin-eosin (HE), anti-galectin-3 (cat. 125402, Biolegend, San Diego, CA), or anti-type I collagen (Col1a1, cat. 1310-01, Southern Biotech, Birmingham, AL) using standard procedures 34 , and quantitative histomorphometry was applied using a digital imaging software (Visiomorph ® , Visiopharm, Hørsholm, Denmark). The fractional area of liver fat (macrosteatosis) was determined on HE-stained sections and expressed relative to total sectional area. The fractional area of galectin-3 and Col1a1 immunostaining was expressed relative to total parenchymal area by subtracting corresponding fat area determined on adjacent HE-stained sections. All histological assessments were performed by histologists blinded to the experimental groups.
RNA sequencing. Liver transcriptome analysis was performed by RNA sequencing on RNA extracts from terminal liver samples (15 mg fresh tissue), as described in detail elsewhere 34  Functional annotation of differentially expressed genes. Candidate NAFLD-and fibrosis-associated pathways were used to annotate genes involved in disease progression (Supplementary Table S1). A gene set analysis was conducted with the R package PIANO version 1.18.1 using the Stouffer method, and p-values were corrected for multiple testing using the Benjamini-Hochberg method (FDR < 0.05). The Reactome pathway database was retrieved and used for gene annotation enrichment analysis.
Statistical analyses. Except from RNA sequencing, data were analyzed using GraphPad Prism v7.03 software (GraphPad, La Jolla, CA). All results are shown as mean ± standard error of mean (S.E.M.). A two-way ANOVA with Tukey's multiple comparisons test was performed for body weight and quantitative histological analyses. A one-way ANOVA with Dunnett's post-hoc test was used for all other parameters. A p-value < 0.05 was considered statistically significant.
www.nature.com/scientificreports www.nature.com/scientificreports/ Individual reductions in Col1a1 fractional area closely correlated to corresponding improvements in steatosis and galectin-3%-area (Fig. 3). Representative photomicrographs on post-biopsy Col1a1 immunostaining are shown in Fig. 4. In addition to quantitative histology, liver histomorphology was evaluated in pre-vs. post-treatment liver biopsies, as outlined by Kleiner et al. 38 . Individual histopathology scores are indicated in Supplementary Figs S3 and S4. At baseline, AMLN ob/ob-NASH mice showed severe steatosis (score 3), mild-to-moderate lobular inflammation (score 1-2) and a relatively low rate of hepatocyte ballooning. Baseline fibrosis stage was mild to moderate (F1-F2) in all treatment groups. Whereas NAFLD activity scores (NAS) were unchanged or slightly increased over the 8-week dosing period in AMLN ob/ob-NASH vehicle control mice, treatment with OCA + ELA 3 mg/kg and OCA + ELA 10 mg/kg resulted in significant reductions in NAS, mainly due to reduced steatosis and inflammation scores. Drug treatments did not significantly change fibrosis scores compared to vehicle controls.

Discussion
The present study characterized the combined effect of eight weeks of OCA and ELA treatment in an AMLN diet-induced obese ob/ob mouse model of biopsy-confirmed NASH. Monotherapy with OCA or ELA improved all liver quantitative histopathological parameters and drug treatment combinations exerted significantly greater effects than either monotherapy on both liver histopathology and associated liver transcriptome changes.
As in NASH patients, mouse models of high-fat/carbohydrate diet-induced NASH show heterogeneous baseline disease state severity and rates of progression 12,20,22,34,39 . To account for the inherent variability in histopathological parameters, AMLN ob/ob-NASH mice were stratified and randomized to treatment based on biopsy-confirmed liver histopathology which also enabled assessment of within-subject treatment responses at study termination. Consistent with previous findings in the model, quantitative histomorphometry on liver (B) Relative gene expression levels (z-scores) of differentially expressed candidate genes associated with NASH and fibrosis (see Supplementary Table S1 for complete list of candidate genes); (C) Expression levels (RPKM values) of genes represented in most significantly regulated pathways associated with lipid metabolism/FXR signaling, monocyte recruitment/inflammation signaling, and stellate cell activation/extracellular matrix (ECM) organization. *p < 0.05, **p < 0.01, ***p < 0.001 (vs. vehicle); # p < 0.05, ## p < 0.01, ### p < 0.01 (vs. OCA alone); ¤ p < 0.01, ¤¤ p < 0.01, ¤¤¤ p < 0.001 (vs. Elafibranor alone). www.nature.com/scientificreports www.nature.com/scientificreports/ pre-biopsies indicated marked liver fat accumulation, significant inflammation and collagen deposition in AMLN ob/ob-NASH mice 34,35,39,40 . Compared to baseline, ob/ob-NASH control mice showed significantly increased terminal fractional area of Col1a1, being in agreement with a recent report using this model 21 . Similar to previous www.nature.com/scientificreports www.nature.com/scientificreports/ studies in AMLN ob/ob-NASH mice 20,21 , galectin-3 immunohistochemistry was applied for quantitative assessment of hepatic inflammation. Galectin-3 (also known as Mac-2) is expressed in various immunocompetent and inflammatory cells, in particular activated macrophages, but also in eosinophils, mast cells and activated lymphocyte profiles 41,42 . In addition to a prominent role in inflammation, galectin-3 secretion from immune cells stimulates proliferation and activation of myofibroblasts, including hepatic stellate cells (HSCs) 42,43 . It should be noted that activated HSCs also express galectin-3, as demonstrated in cultured primary HSCs [44][45][46][47] as well as in chemotoxin-and surgically-induced rodent models of severe fibrotic liver injury 45,46 . Although it is not resolved if activated HCSs could potentially contribute to galectin-3 expression in NASH models with less advanced inflammation and fibrosis, liver samples from AMLN ob/ob-NASH mice show highly different distribution of α-SMA and galectin-3 immunostaining (data not shown), which argues for HSC-derived galectin-3 expression being neglectable in this model.
OCA and ELA were probed for individual and combined drug effects on quantitative liver histopathology, plasma biochemistry, body weight and body composition in AMLN ob/ob-NASH mice. Individual drug doses were selected based on previous studies characterizing effects of OCA and ELA monotherapy in AMLN ob/ ob-NASH mice 20,21,40 . OCA treatment resulted in significant quantitative reductions in liver fat, inflammation and collagen deposition. The marked improvement in steatosis was reflected by reduced hepatomegaly in OCA-treated mice. OCA was weight neutral and did not influence body weight and whole-body fat mass in AMLN ob/ob-NASH mice, indicating that OCA-induced improvements in liver histopathology were not dependent on changes in adiposity. These findings corroborate recently reported histological and metabolic effects of OCA treatment in AMLN ob/ob-NASH mice 20,21,40 as well as in a comparable AMLN diet-induced obese (DIO) NASH model in C57BL/6 J mice (AMLN DIO-NASH) 20,23 . Also, our data are in line with previous data on metabolic, anti-inflammatory and anti-fibrotic effects of FXR agonist treatment, including OCA, in other rodent models of obesity with features of simple steatosis/low-grade inflammation 22,48 and NASH 49 , as well as in nutrient-deficient dietary [50][51][52] and surgery-based models of NASH 51 . It should be noted that others have reported concomitant reductions in body weight and white adipose tissue mass following administration of OCA in high-caloric diet-fed mice with simple steatosis 25 , suggesting model phenotype-specific effects of OCA on adipose tissue metabolism. Several genes involved in triglyceride and cholesterol metabolism are major transcriptional targets of FXRs and act in concert to regulate hepatocyte lipid clearance 53,54 , and suppression of carbohydrate-responsive gene expression and improved hepatic insulin resistance may contribute to FXR-mediated lowering of hepatic lipogenesis 55 . FXRs activation also exerts immunosuppressive actions in various immune cell populations, including monocytes and macrophages [56][57][58][59] , and inhibited pro-fibrotic activity of hepatic stellate cells 14 .
Our data also confirm and extend upon findings of ELA-induced reductions in body weight, liver histomorphometry and plasma biochemical markers in AMLN ob/ob-NASH mice 20 . Interestingly, ELA treatment reduced weight gain without influencing hepatomegaly in AMLN ob/ob-NASH mice. Unlike humans, stimulated PPAR-α and PPAR-δ function has been associated with weight loss and appetite suppression in diet-induced obese mice 60,61 . In addition, PPAR-α agonist-mediated hepatocyte peroxisome proliferation can lead to rodent-specific increases in liver mass 62,63 , which is therefore likely to prevent ELA from improving hepatomegaly in rodent models of NASH 20,25 . Compared to OCA, the effect of ELA treatment has been less characterized in animal models of NASH. Previous reports have demonstrated that ELA ameliorates steatosis, inflammation, and fibrosis in AMLN DIO-NASH mice 20 , improves liver metabolic parameters and reverses development of fibrosis in nutrient-deficient and hepatotoxin-induced models of NASH 25 . The significant reduction of steatosis in ELA-treated AMLN ob/ob-NASH mice is consistent with the diverse actions of PPAR-α on lipid metabolic signaling pathways. The major function of PPAR-α is to facilitate hepatic fatty acid utilization by transcriptional upregulation of rate-limiting peroxisomal and mitochondrial enzymes controlling fatty acid transport and β-oxidation 64 , mitochondrial ketogenesis 65 as well as lipolysis 66 . Also, selective PPAR-α agonists stimulate hepatic clearance of triglyceride-rich lipoproteins 67 . Because rodent liver PPAR-α expression is mainly confined to hepatocytes 68,69 , the anti-inflammatory and anti-fibrotic effects of PPAR-α agonists are considered subsequent to direct actions on liver parenchymal cells preventing the release of lipotoxic and pro-inflammatory mediators 70 . Interestingly, the beneficial effects of ELA on steatohepatitis are preserved in high-fat diet-fed double transgenic hApoE2/ PPAR-α knock-out mice 25 , indirectly supporting the concept that PPAR-δ (also termed PPAR-β/δ) contributes to the anti-NASH effects of ELA. PPAR-δ agonist-mediated actions on lipid metabolism is reportedly dependent on co-operative PPAR-α activity 71 , and may indirectly reflect suppression of hepatic gluconeogenesis and hepatic glucose output 72,73 . Whereas PPAR-δ regulates Kupffer cell polarization towards an M2 anti-inflammatory phenotype 74 , functional implications of PPAR-δ expression during hepatic stellate cell activation remain to be clarified 75 .
We have previously reported that AMLN ob/ob-NASH and DIO-NASH mice display hypercholesterolemia, but not hypertriglyceridemia, concurrent with marked hepatic accumulation of triglycerides and cholesterol 20,34,39,76,77 . This is also a characteristic of other cholesterol-enriched Western diet-based mouse models of NASH 76,77 , and it is speculated that increased dietary cholesterol intake may impair hepatocyte triglyceride secretion by modifying cholesterol ester and lipoprotein synthesis 77,78 . As OCA and ELA mono-treatment, as well as combinations, reduced plasma total triglyceride and cholesterol levels while also upregulating hepatic genes linked to cholesterol biosynthesis, this further points to complex regulation of cholesterol metabolism in AMLN ob/ob-NASH mice. The lack of hypertriglyceridemia in AMLN ob/ob-NASH mice contrasts the dyslipidemia profile in NAFLD/NASH patients which is characterized by elevated triglyceride and low-density lipoprotein cholesterol levels as well as decreased high-density lipoprotein cholesterol concentrations 79,80 . Although we did not specifically determine the plasma lipoprotein cholesterol profile in the present study, it should be emphasized that DIO mice (and rats) are generally resistant to develop human-like atherogenic dyslipidemia 81,82 , which should be taken into account when interpreting changes in plasma lipid profiles in these models, including AMLN ob/ ob-NASH mice.
www.nature.com/scientificreports www.nature.com/scientificreports/ Notably, OCA and ELA combinations exerted additive therapeutic effects on liver histology in AMLN ob/ ob-NASH mice. Compared to baseline, co-administration of OCA and ELA led to marked improvements in the proportionate area of liver fat and galectin-3. The histomorphometric analyses indicated that OCA markedly enhanced the anti-fibrotic efficacy of both doses of ELA. Accordingly, combined treatment with OCA and ELA resulted in almost complete prevention of progressive hepatic Col1a1 deposition. The individual reduction in the proportionate area of Col1a1 was closely correlated with improvements in steatosis and inflammation, illustrating a highly consistent within-subject effect on all three histological endpoints. Also, AMLN ob/ob-NASH mice receiving both OCA and ELA treatment attained further weight loss as compared to that achieved by ELA alone. Whether the magnitude of weight loss contributed to promote further benefits on liver histopathology must await further studies using additional control conditions, e.g. weight-matched/calorie-restricted animals.
As discussed above, the enhanced liver histological outcome of combined OCA and ELA treatment is consistent with both receptor families being master transcriptional regulators of a broad group of metabolic enzymes and signaling molecules. Accordingly, our full-scale mapping of hepatic gene expression indicated that OCA and ELA monotherapies elicited distinct hepatic expression signatures in ob/ob-NASH mice and their combination led to profound changes in the liver transcriptome. The extent of transcriptional changes argues for widespread alterations in hepatic molecular signaling conferred by recruitment of complementary FXR and PPAR-α/δ associated mechanisms of action. Importantly, the liver transcriptome signature in ob/ob-NASH mice receiving combined OCA and ELA treatment supports the histological outcomes by indices of improved lipid handling and insulin signaling with concurrent attenuation of both immune and pro-fibrotic gene expression patterns. Interestingly, OCA and ELA co-administration resulted in a marked potentiation of their individual suppressive effects on a variety of pathways regulating the activity of the innate and adaptive immune system, suggesting that regression of steatohepatitis played an integral role in preventing further collagen deposition in AMLN ob/ob-NASH mice.
Histopathology scores partially correlated to the quantitative histological data, however, confirmed significantly reduced steatosis and inflammation in AMLN ob/ob-NASH mice receiving OCA monotherapy and the further improvements in these parameters following combined OCA and ELA treatment. The disparity likely reflects that histopathological disease scoring is based on semiquantitative morphological criteria and imaging-based quantitative histomorphometry may therefore better capture individual differences in treatment responses. The lack of effect of OCA on fibrosis scores is consistent with an 8-week dosing study in AMLN ob/ob-NASH mice 20 . Because 16 weeks of treatment with OCA has recently been reported effective in reducing both Col1a1 deposition and fibrosis scores in AMLN ob/ob-NASH mice 21 , this suggests that extended treatment periods may be required to observe parallel reductions in both parameters. Also, compared to the relative low doses of ELA (3-10 mg/kg/ day) employed in the present study, a higher dose (30 mg/kg/day) has been reported to improve fibrosis scores in AMLN ob/ob-NASH mice 20 .
In conclusion, monotherapy with OCA and ELA improved histomorphometric indices of steatosis, inflammation and fibrosis in an ob/ob mouse model of biopsy-confirmed NASH. When administered in combination, OCA and ELA promoted additive metabolic and histological effects. Liver transcriptome changes indicated that OCA and ELA were complementary in targeting hepatic molecular mechanisms facilitating further attenuation of steatohepatitis and fibrogenesis in AMLN ob/ob-NASH mice. These findings provide preclinical proof-of-concept for combined FXR and PPAR-α/δ agonist-based therapies in NASH.

Data Availability
All data generated or analyzed during this study are included in this published article (and its Supplementary Information files).