Rilpivirine

Rilpivirine attenuates liver fibrosis through selective STAT1-mediated apoptosis in hepatic stellate cells
Alberto Martí-Rodrigo, 1 Fernando Alegre,1,2 Ángela B Moragrega,1 Francisco García-García,3 Pablo Martí-Rodrigo, 1 Anabel Fernández-Iglesias,4 Jordi Gracia-Sancho, 4,5 Nadezda Apostolova,1 Juan V Esplugues,1,2
Ana Blas-García 1

► Additional material is published online only. To view, please visit the journal online (http://dx.doi.org/10.1136/ gutjnl-2019-318372)
1Department of Pharmacology, Faculty of Medicine, University of Valencia-CIBERehd, Valencia, Spain
2FISABIO-Hospital Universitario Dr. Peset, Valencia, Spain 3Bioinformatics & Biostatistics Unit, Principe Felipe Research Center, Valencia, Spain
4Liver Vascular Biology Research Group, Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS)-CIBERehd, Barcelona, Spain
5Hepatology, Department of Biomedical Research, Inselspital, University of Bern, Bern, Switzerland

Correspondence to
Dr Ana Blas-García, Department of Pharmacology, Faculty of Medicine, University of Valencia- CIBERehd, Avenida Blasco Ibáñez, 15-17, Valencia 46010, Spain; [email protected]

JVE and AB-G are joint senior authors.

Received 25 January 2019
Revised 8 August 2019
Accepted 15 August 2019

© Author(s) (or their

ABSTRACT
Objective Liver fibrosis constitutes a major health problem worldwide due to its rapidly increasing prevalence and the lack of specific and effective treatments. Growing evidence suggests that signalling through cytokine-activated Janus kinase (JAK)-signal transducer and activator of transcription (STAT) pathways regulates liver fibrosis and regeneration. Rilpivirine (RPV) is a widely used anti-HIV drug not reported to produce hepatotoxicity. We aimed to describe the potential hepatoprotective effects of RPV in different models of chronic liver injury, focusing on JAK-STAT signalling regulation.
Design The effects of RPV on hepatic steatosis, inflammation and fibrogenesis were studied in a nutritional mouse model of non-alcoholic fatty liver disease, carbon tetrachloride-induced fibrosis and bile duct ligation-induced fibrosis. Primary human hepatic stellate cells (hHSC) and human cell lines LX-2 and Hep3B were used to investigate the underlying molecular mechanisms.
Results RPV exerted a clear anti-inflammatory and antifibrotic effect in all the in vivo models of liver injury employed, and enhanced STAT3-dependent proliferation in hepatocytes and apoptosis in HSC through selective STAT1 activation. These results were reproduced in vitro; RPV undermined STAT3 activation and triggered STAT1- mediated pathways and apoptosis in HSC. Interestingly, this selective pro-apoptotic effect completely disappeared when STAT1 was silenced. Conditioned medium experiments showed that HSC apoptosis activated STAT3 in hepatocytes in an interleukin-6- dependent mechanism.
Conclusion RPV ameliorates liver fibrosis through selective STAT1-dependent induction of apoptosis in HSC, which exert paracrinal effects in hepatocytes, thus promoting liver regeneration. RPV’s actions may represent an effective strategy to treat chronic liver diseases of different aetiologies and help identify novel therapeutic targets.

Unfortunately, there is no cure for this disease,

employer(s)) 2019. No
commercial re-use. See rights and permissions. Published by BMJ.
INTRODUCTION
Liver fibrosis, a serious health problem world- wide whose prevalence continues to rise, can be caused by several aetiologies, and is characterised by excessive deposition of extracellular matrix, loss of parenchymal structure and organ dysfunction.

which can progress to cirrhosis, hepatocarcinoma and even death.1 Therefore, there is great interest in developing new effective therapies, with drug repurposing or use of compounds with pleiotropic effects approved for other therapeutic uses repre- senting one of the main strategies. In this way, drugs whose toxicological profile and bioavailability have already been validated are analysed to determine their potential antifibrotic effects. Rilpivirine (RPV)

Martí-Rodrigo A, et al. Gut 2019;0:1–14. doi:10.1136/gutjnl-2019-318372 1

is an antiretroviral drug widely used to treat HIV infection; in addition to good efficacy and tolerability, it has proved to have a more favourable safety profile than other antiretroviral drugs, especially regarding liver toxicity and lipid profile.2–7
Chronic liver diseases are characterised by the presence of damaged hepatocytes and the activation and proliferation of hepatic stellate cells (HSC), the main contributors to hepatic fibrosis.8 Consequently, potential therapies for these pathologies should ideally possess pro-regenerative and antifibrotic proper- ties, for example, the ability to control hepatocyte proliferation and to inactivate activated HSC.
Increasing evidence suggests that signalling through Janus kinase-signal transducer and activator of transcription (JAK-STAT) pathways plays key roles in controlling chronic liver injury progression and liver regeneration.9 Interestingly, STAT1 and STAT3 activation takes place in both parenchymal and non-parenchymal cells, playing contrasting roles in funda- mental aspects of liver pathophysiology.10 In hepatocytes, STAT1 activation is a pro-apoptotic signal that enhances liver damage, whereas STAT3 activation protects against cell death and increases cellular proliferation.11 Conversely, STAT1 acti- vation in HSC limits their proliferation and fibrogenic activity, exerting beneficial effects within the liver, while STAT3 activa- tion is deleterious, as it promotes fibrogenesis.12 Importantly, hepatic STAT1 and STAT3 antagonise each other functionally, and mutually inhibit each other’s activation.10 13
We analysed the effects of RPV in different validated mouse models of chronic liver injury and in in vitro human cell culture systems (both cell lines and primary hepatic cells), focusing on the control of activated HSC and the regeneration of liver paren- chyma. We paid special attention to the regulation of STAT1-me- diated and STAT3-mediated signalling pathways, as well as to their involvement in RPV-induced effects on liver fibrosis. Finally, we analysed public data registries to find supporting clinical data.

METHODS
Cell culture and treatments
Human hepatoblastoma Hep3B cells (ATCC HB-8064) were cultured as described previously.14 Immortalised human HSC LX-2 (kindly provided by Dr Scott L. Friedman, Icahn School of Medicine at Mount Sinai, New York, USA) were grown in Dulbecco’s Modified Eagle Medium supplemented with 10% heat-inactivated fetal bovine serum and antibiotics (penicillin (50 U/mL) and streptomycin (50 mg/mL)). Primary human HSC (hHSC) were obtained from tissue (20 g) after surgical resection of tumour metastasis or liver transplantation, and were isolated and cultured as described elsewhere.15 Samples were collected and processed by the Liver Vascular Biology Research Group at Institut D’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS) and Hospital Clínic (Barcelona, Spain). Patients gave their informed consent in all cases.
Cells were treated with clinically relevant concentrations of RPV (Sequoia Research Products, 1–8 µM) for 24–72 hours; dimethyl sulfoxide (DMSO) was used as a control. The profi- brogenic stimulus transforming growth factor beta (TGF) 1 (2.5 ng/mL) and the STAT3 activator interleukin (IL)-6 (5 ng/mL) were employed as further controls.

Mouse models of chronic liver diseases
All animal procedures were performed in accordance with the University of Valencia’s guidelines for the care and use of labora- tory animals, and were approved by the local ethics committee.

Female C57BL/6J mice aged 9 weeks (Janvier Labs) were randomly assigned to different experimentation groups (10 mice per group). Animal dosage (oral) was calculated using the normalised interspecies allometric scaling factor established by the Food and Drug Administration (FDA) to achieve the equiv- alent to the maximum daily therapeutic dose of RPV (25 mg).16 At the end of the protocols, blood analysis, liver histology and molecular analysis were performed as described below.
► Non-alcoholic fatty liver disease (NAFLD) nutritional model: animals received a normal diet (ND) or high-fat diet (HFD) (Ssniff Labs; EF R/M D12331 mod.*/Surwit; 59% fat and 2% free cholesterol) for 12 weeks, and were administered either RPV (5 mg/kg/day, oral) or its vehicle (Veh, DMSO).
► Liver fibrosis models: carbon tetrachloride (CCl4) was admin- istered (0.5 mg/kg in corn oil; intraperitoneal) three times per week for 4 or 6 weeks. In the former protocol, animals were administered daily with RPV (5 mg/kg/day, oral) or its Veh during the duration of the protocol; in the latter, RPV or its Veh were administered during the last 2 weeks.
► Bile duct ligation (BDL) model: mice were randomly assigned to sham-operated or BDL groups. RPV (5 mg/kg/day, oral) or its Veh were administered for 15 days from the day of surgery (details in online supplementary material).

Protein extraction and western blot analysis
Protein expression was analysed in liver tissue and total cell protein extracts (online supplementary material).

RNA isolation and real-time RT-PCR
Total mRNA isolation and real-time RT-PCR were performed as described in online supplementary material.

Transcriptomic analysis
Whole-liver RNA from HFD-fed mice (treated with Veh or RPV, 12 weeks) was isolated using TriPure Isolation Reagent, as described in the previous section, and used for transcriptomic analysis (data deposited in NCBI Gene Expression Omnibus17 accession number GSE120484). Details supplied in online supplementary material.

Myeloperoxidase activity
Liver myeloperoxidase (MPO) activity was determined spec- trophotometrically in murine liver homogenates by measuring absorbance at 450 nm (details in online supplementary material).

Plasma determinations
Mice plasma samples were analysed in a laboratory specialised in veterinarian diagnosis (CEDIVET, Valencia, Spain), to quantify concentrations of different markers of liver injury.

Histological studies
Histological analysis was performed in 5 µM paraffin-embedded mouse liver samples in which we performed H&E staining, picro-Sirius Red staining, terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay and immunohistochem- istry (IHC) with specific antibodies (see online supplementary material).

Flow cytometry
Induction of apoptosis in LX-2 cells was measured by bivariate annexin V/propidium iodide analysis (Annexin V-FITC apop- tosis detection kit, Abcam), after 72 hours treatment (details in online supplementary data).

Secretome analysis
The composition of the conditioned medium from TGF-ac- tivated LX-2 cells untreated or treated with RPV (4 µM) for 72 hours was analysed using a commercial Human Cytokine Array kit (R&D, Abingdon, UK), following the manufacturer’s instructions. In key experiments, soluble neutralising mono- clonal antibodies against human IL-6 and interferon gamma (IFN) (InvivoGen, Toulouse, France) were employed (100 ng/ mL, 1 hour) to capture these cytokines in the conditioned medium of RPV-treated LX-2 cells (72 hours), to subsequently determine their role in the expression of pSTAT3 in Hep3B cells incubated with said secretomes (24 hours).

Clinical data
Raw clinical data were obtained from the Multicenter AIDS Cohort Study (MACS; website is located at http://aidscohort- study.org/) data set, which compiles data from blood tests of patients with HIV treated in four different university hospitals in the USA. Only data obtained from 2011 to 2014 for patients on RPV-containing or non-RPV-containing therapies were analysed.

Presentation of data and statistical analyses
All images display representative data for four to seven indepen- dent experiments (except for transcriptomic, hHSC and cyto- kine array studies, in which three independent experiments were performed). Data (mean±SD) were analysed using GraphPad Prism V.6.01 with a Student’s t-test (#p<0.05) or a one-way analysis of variance followed by a Bonferroni test (*p<0.05). A non-parametric Mann-Whitney U test was employed for clinical data analysis (mean±SEM, *p<0.05). RESULTS RPV ameliorates liver injury in a nutritional model of NAFLD We first validated our nutritional model by assessing lipid infil- tration and inflammatory responses—two important outcomes in NAFLD—in whole liver samples. As expected, 12-week administration of HFD induced a clear increment of hepatic lipid droplets (quantified by histological analysis; figure 1A) and a significant enhancement of the inflammatory response, analysed by measuring MPO activity (figure 1B) and the upregu- lation of several pro-inflammatory cytokines and NLRP3 inflam- masome components (figure 1C). Furthermore, phosphorylation of the nuclear factor B (NF-B) p65 subunit and cleavage of pro-caspase 1 (figure 1D) were significantly induced by this diet. Increased lipid deposition was confirmed by the augmented expression of perilipin 2 (PLIN-2) and peroxisome prolifera- tor-activated receptor gamma (PPAR)—two proteins involved in hepatic steatosis—in liver extracts from HFD-fed animals (figure 1D). Next, we evaluated these parameters in liver samples from RPV-treated mice fed the same diet. We observed a clear improvement of both conditions; histological analysis of liver sections demonstrated significant reductions in lipid droplet accumulation, with the HFD-produced increase being almost completely reversed (figure 1A). In this line, RPV altered the expression of several genes involved in lipid metabolism, upregu- lating the antisteatotic genes Adipor1 and Pnpla3 and decreasing the expression of the pro-steatotic gene Lepr (figure 1C). Addi- tionally, protein levels of PLIN-2 and PPAR dropped to levels similar to those observed in ND Veh mice (figure 1D). RPV also exerted anti-inflammatory actions, significantly diminishing MPO activity, NF-B and Casp1 activation and the HFD-induced upregulation of pro-inflammatory cytokines and inflammasome components (figure 1B–D). Interestingly, Il-6 levels were dimin- ished in HFD+Veh mice, and significantly increased in animals administered RPV. Besides enhanced lipid accumulation and inflammation, the livers of HFD-fed animals showed increased collagen deposition, quantified by Sirius Red staining (figure 1A), and overexpres- sion of numerous genes associated with fibrogenesis (figure 1C). Induction of this process was confirmed by western blot analysis of protein expression of collagen 1, vimentin and desmin, which showed a clear augmentation of the three proteins (figure 1D). Again, RPV produced an attenuation of all these effects, signifi- cantly reducing collagen deposition and gene and protein expres- sion of fibrogenic markers (figure 1A, C, D). RPV triggers hepatic STAT1-mediated and STAT3-mediated pathways in mice Transcriptomic analysis of livers from HFD-fed mice treated with RPV showed significant differences in biological functions associated with lipid metabolism, inflammation, the immune system, cell cycle and apoptosis when compared with Veh-ad- ministered HFD animals (online supplementary figure 1). Data analysis revealed a conjunction of enhanced pro-proliferative and antiproliferative signalling pathways that may be associated with differential responses exerted by liver cell populations (online supplementary figure 1C and online supplementary tables 5–9). To clarify this point, we analysed cell proliferation and apoptosis in liver sections using Ki67 staining and TUNEL assay, respec- tively. Livers from RPV-treated mice presented a lower number of proliferating non-parenchymal cells and rise in that of paren- chymal cells with respect to Veh-treated HFD mice (figure 2A). Conversely, apoptosis was clearly increased in non-parenchymal cells and diminished in hepatocytes (figure 2B). RPV-mediated alterations of cell proliferation were confirmed in whole-liver samples by qRT-PCR (figure 2C), with significant gene expres- sion of important markers. Interestingly, transcriptomic analysis also revealed a posi- tive regulation of the IL-6-mediated signalling pathway (online supplementary figure 1B), which pointed to a role for the IL-6- STAT3 axis in cell proliferation control. This led us to analyse the expression and activation of STAT3 in the different experi- mental groups; we observed a clear reduction in the phosphor- ylation of this factor in HFD mice, whereas administration of RPV normalised the pSTAT3/STAT3 ratio in whole-liver extracts (figure 2D). These results were expanded by IHC, which revealed that STAT3 was expressed mainly in hepatocytes, with an intense nuclear signal being detected in ND mice. STAT3 activation was decreased in HFD mice and restored to normal levels in animals treated with RPV (figure 2E). Importantly, this expression pattern was not reproduced with STAT1, which was expressed mostly in non-parenchymal cells; nevertheless, STAT1 activation was altered in HFD and HFD-RPV animals. Specifically, the STAT1 signal was almost abolished in HFD-fed mice, while RPV activated this transcription factor (figure 2F), suggesting a role for STAT1 in the regulation of this cell population. The hepatoprotective effects of RPV are reproduced in CCl4- induced fibrosis and BDL-induced fibrosis To detect if RPV-induced effects in the liver were associated with types of chronic liver disease other than fatty liver disease, we established a pharmacological model of liver fibrosis in wild- type mice, which were administered CCl4 for 6 weeks. In some groups, we also administered daily doses of RPV during the last 2 weeks of the protocol, once liver damage had been established. Figure 1 Chronic administration of rilpivirine (RPV) decreases lipid accumulation, collagen deposition and liver inflammation in a chronic model of non-alcoholic fatty liver disease. (A) Representative images and quantifications of lipid infiltration (H&E) and collagen deposition (by Sirius Red) of healthy (normal diet (ND)), steatotic (high-fat diet (HFD)+vehicle (Veh)) and RPV-treated steatotic (HFD+RPV) groups (n=7). Scale bar=0.1 mm. (B) Relative myeloperoxidase (MPO) enzyme activity in whole-liver extracts (n=5). (C) Heatmap representation of colour-coded expression levels (RT- PCR) of differentially expressed genes associated with lipid metabolism, liver inflammation and fibrogenesis from whole-liver samples. Data were normalised vs the housekeeping gene Actb (-actin) (n=7). (D) Representative western blot analysis images and histogram expressing quantification of proteins involved in (i) lipid metabolism, (ii) inflammation and (iii) fibrogenesis from whole-liver protein extracts (n=5). Data for RPV-treated ND mice are displayed in the images (D), but not in the graphs, as significant changes were not observed. Data (mean±SD) were analysed by a Student’s t-test: #p<0.05 vs ND group, *p<0.05 vs HFD+Veh group. As expected, CCl4 significantly increased the expression of fibrotic and inflammatory markers, as well as collagen deposi- tion and HSC proliferation (represented by vimentin and desmin expression), number of hepatic macrophages, MPO activity and activation of NF-B (figure 3A–C). In line with the actions observed in the nutritional model, RPV produced an attenuation of all these effects, significantly reducing collagen deposition and protein expression of fibrogenic markers. Furthermore, similar beneficial effects were observed in serum levels of parameters related to liver function, such as bilirubin and hepatic enzymes, which were all clearly enhanced in CCl4-injected mice and normalised in those co-administered with RPV (figure 3D). Analysis of proliferative and apoptotic cells in the different experimental groups revealed the same pattern of staining observed in the NAFLD model, as demonstrated by the quantifi- cation of both parameters in parenchymal and non-parenchymal cells. Hepatocytes were mainly apoptotic and did not prolif- erate in animals exposed to hepatic insult, while RPV admin- istration enhanced the number of proliferative hepatocytes and, very importantly, triggered apoptosis in non-parenchymal Figure 2 Chronic administration of rilpivirine (RPV) boosts hepatic regeneration, induces apoptosis of hepatic stellate cells and restores Janus kinase-signal transducer and activator of transcription (JAK-STAT)3 signalling in a mice model of non-alcoholic fatty liver disease. Representative images and cell count of (A) proliferating liver cells measured by Ki67 immunohistochemistry (IHC) and (B) apoptotic cells measured by TUNEL assay in normal diet (ND), high-fat diet (HFD+vehicle (Veh)) and HFD+RPV mice groups (n=5). (C) Heatmap representation of colour-coded expression levels (RT-PCR) of differentially expressed genes associated with cell cycle and JAK-STAT3 and JAK-STAT1 signalling pathways. Data were normalised vs the housekeeping gene Actb (-actin) (n=7). (D) Representative western blot analysis of STAT3 activation by phosphorylation (phospho-STAT3/total STAT3 ratio) in response to RPV in both ND and HFD groups. RPV-treated ND mice are displayed in the images, but not in the graphs, as significant changes were not observed (n=5). Representative images of (E) STAT3 and (F) STAT1 IHC determinations in mice liver tissue from ND, HFD+Veh and HFD+RPV mice groups (n=4). Scale bar=0.1 mm. Data (mean±SD) were analysed by a Student’s t-test: #p<0.05 vs ND group, *p<0.05 vs HFD+Veh group. cells (figure 4A, B). Similar to the previous experiments in the NAFLD model, we then analysed the capacity of RPV to effect the equilibrium between STAT3 and STAT1 signalling and, consequently, its potential to modulate both proliferation and apoptotic cell death in different cell subpopulations. Impor- tantly, all the regulatory and hepatoprotective actions of RPV previously observed in the nutritional model of NAFLD were reproduced in this model. As expected, CCl4-induced injury clearly undermined activation of STAT3 in hepatocytes in whole liver extracts (figure 4C) and liver sections (figure 4D) and of STAT1 in non-parenchymal cells (figure 4E). Administration of RPV normalised STAT3 and STAT1 expression in parenchymal and non-parenchymal cells, respectively (figure 4D, E). To explore the ability of RPV to prevent CCl4-induced injury, we performed further experiments in animals treated with RPV since the beginning of CCl4 administration, and again observed a clear attenuation of all the alterations generated by the injection of CCl4 (online supplementary figures 2 and 3). Importantly, beneficial RPV-induced effects were also evident in a murine model of obstructive cholestasis provoked by BDL. Although this model produced a severe fibrosis, animals administered RPV showed less collagen deposition, reduced expression of inflammatory and fibrogenic markers, improved liver function and a normalisation of STAT1 and STAT3 acti- vation (figure 5A–E). Furthermore, a selective induction of apoptosis in non-parenchymal cells was observed in RPV-treated mice, as shown by representative IHC images, which reveal an enhanced non-parenchymal signal of cleaved caspase-3 (figure 5F). RPV exerts a differential regulation of STAT1 and STAT3 in HSC and hepatocytes To better characterise the specific effects of RPV on hepato- cytes and HSC, we performed key experiments in human cell lines. MTT determinations revealed that direct treatment with RPV did not alter cellular viability in hepatocyte cell lines, with the highest concentration employed producing only a slight deterioration in HepG2 (online supplemen- tary figure 4A). Interestingly, RPV induced a significant and concentration-dependent cytotoxic effect in LX-2 cells (online supplementary figure 4B), both in normal conditions and following activation with the fibrogenic mediator TGF. This effect was confirmed by assessing cell proliferation/ Figure 3 Rilpivirine (RPV) decreases liver inflammation and fibrosis progression in a chronic model of carbon tetrachloride (CCl4)-induced liver injury. (A) Representative images and quantification of fibrosis progression (collagen deposition measured by Sirius Red staining and hepatic stellate cell activation by vimentin immunohistochemistry-IHC-) and macrophage infiltration (measured by F4/80 IHC) in healthy (vehicle (Veh)), fibrotic (CCl4+Veh) and RPV-treated fibrotic (CCl4+RPV) mice groups (n=7). Scale bar=0.1 mm. (B) Representative western blot analysis images and quantification of desmin expression and nuclear factor B activation (phospho-p65/total p65 ratio) in whole-liver samples. (C) Relative myeloperoxidase (MPO) enzyme activity in whole-liver extracts (n=5). (D) Serum bilirubin, aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels (n=6). Data (mean±SD) were analysed by a Student’s t-test: #p<0.05 vs Veh group, *p<0.05 vs CCl4+Veh group. survival in this same cell line using fluorescence microscopy (data not shown). Furthermore, light microscopy revealed activation of cell death processes in the LX-2 cell phenotype after 48 hours treatment (online supplementary figure 4C). We next studied the gene expression of several markers of HSC activation and fibrogenesis in the same experimental conditions. In non-stimulated cells, RPV exerted a discrete effect that was only evident at higher concentrations, except for SERPINE-1 expression, which was clearly reduced. However, the antifibrotic effect of RPV was perfectly observed in cells co-treated with TGF, in which a significant downregulation of all fibrogenic genes was expressed in a concentration-dependent fashion (online supplementary figure 4D). As expected, RPV also induced a concentration-dependent decrease in collagen protein expression in all the conditions of co-treatment with TGF (online supplementary figure 4E). Characterisation of pSTAT3 and pSTAT1 protein expression in HSC revealed that RPV induced an opposite effect. In both non-stimulated and stimulated conditions, and especially in the latter, RPV impeded STAT3 activation and promoted that of STAT1, inducing a Figure 4 Chronic administration of rilpivirine (RPV) enhances hepatic regeneration, induces apoptosis of hepatic stellate cells and restores Janus kinase-signal transducer and activator of transcription (JAK-STAT)3 signalling in chronic models of carbon tetrachloride (CCl4)-induced liver injury. Representative images and cell count of (A) proliferating liver cells measured by Ki67 immunohistochemistry (IHC) and (B) apoptotic cells measured by TUNEL assay in healthy (vehicle (Veh)), fibrotic (CCl4+Veh) and RPV-treated fibrotic (CCl4+RPV) mice groups (n=5). (C) Representative western blot analysis images of STAT3 activation by phosphorylation (phospho-STAT3/total STAT3 ratio) in response to RPV in whole-liver tissue (n=5). Representative images of (D) STAT3 and (E) STAT1 IHC determinations in liver tissue from Veh, CCl4+Veh and CCl4+RPV mice groups (n=4). Scale bar=0.1 mm. Data (mean±SD) were analysed by a Student’s t-test: #p<0.05 vs Veh group, *p<0.05 vs CCl4+Veh group. marked increase in pSTAT1 expression (online supplementary figure 4E), thus confirming the antifibrogenic properties exerted by RPV in this cell type. Further analysis of HSC death in vitro demonstrated that RPV induced a concentration-dependent pro-apoptotic effect in LX-2 cells after 72 hours treatment (online supplementary figure 4F). Interestingly, the effect was similar in both non-stimulated and stimulated cells. Moreover, compared with the positive control staurosporine (STS), RPV produced only a moderate pro-apoptotic effect, but this response was reproduced in all the experimental replicates. It is important to point out that this pro-apoptotic effect was time-dependent, as it was not detected when LX-2 cells were treated with RPV for shorter periods of time (24 and 48 hours incubation induced no changes; data not shown). Finally, we tested the effect of RPV in primary hHSC, which were treated for 72 hours in the same conditions as LX-2 cells, with intermediate concentrations of RPV (2–4 µM) and the same amount of TGF to stimulate their proliferation (2.5 ng/ mL). Conversely to LX-2 cells, hHSC did not exhibit relevant morphological alterations in response to RPV (data not shown). However, gene and protein expression of fibrogenic markers (figure 6A, B) and protein levels of pSTAT3 and pSTAT1 (figure 6B) reproduced perfectly the results obtained in LX-2 cells. Inhibition of STAT1 in HSC undermines the antifibrogenic effect of RPV In order to ascertain the role of STAT1 signalling in the func- tional alterations observed in HSC in response to RPV, we chemically inhibited STAT1 activation in hHSC and LX-2 cells by pre-incubating them with fludarabine (5 µM, 2 hours). This inhibitor clearly decreased STAT1 phosphorylation, but was not specific, since STAT3 phosphorylation was also substantially reduced. Interestingly, the RPV-induced decrease in Col1a1 protein expression was seen to disappear or increase when STAT1 was not activated (figure 6B, online supplementary figure 5). We then performed transient silencing of this transcription factor to clarify its involvement in the apoptotic effects of RPV; Figure 5 Rilpivirine (RPV) decreases inflammation and fibrosis progression, induces apoptosis of hepatic stellate cells and restores Janus kinase- signal transducer and activator of transcription (JAK-STAT)3 signalling in bile duct ligation (BDL)-induced liver injury. (A) Representative images and quantification of fibrosis progression (collagen deposition measured by Sirius Red staining and hepatic stellate cell activation measured by vimentin immunohistochemistry (IHC)) in healthy (sham), fibrotic (BDL+vehicle (Veh)) and RPV-treated fibrotic (BDL+RPV) mice groups (n=8). Scale bar=0.1 mm. (B) Representative western blot analysis images and quantification of Col1a1 and desmin expression and nuclear factor B activation (phospho-p65/total p65 ratio) in whole-liver samples. (n=8). (C) Serum bilirubin, aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels (n=7). Representative images of (D) STAT3, (E) STAT1 and (F) cleaved caspase-3 IHC determinations in liver tissue from sham, BDL+Veh and BDL+RPV mice groups (n=8). Scale bar=0.1 mm. Data (mean±SD) were analysed by a Student’s t-test: #p<0.05 vs sham group, *p<0.05 vs BDL+Veh group. in this case, STAT1 silencing did not alter STAT3 activation (figure 7A, B). The viability of transfected cells was also unaf- fected (data not shown), although cytometric analysis revealed a small, non-significant increase in the number of necrotic cells, both from siC and siSTAT1 (figure 7C). In line with our previous results, a RPV-mediated decrease in Col1a1 and pSTAT3 protein expression was not noted when STAT1 signalling was blocked (figure 7B). Importantly, the pro-apoptotic effect of RPV also disappeared once STAT1 signalling had been impaired, as shown by flow cytometry experiments (figure 7C). Hepatocyte STAT3 activation is selectively promoted by the secretome of RPV-induced apoptotic HSC As direct treatment with RPV failed to increase pSTAT3 expres- sion in Hep3B cells, further experiments were performed to combine RPV with IL-6, the endogenous activator of STAT3. Once again, RPV did not enhance expression of this protein; as displayed in figure 8A, STAT3 activation was altered only by IL-6 treatment. We then carried out conditioned medium exper- iments to explore the possibility of an active interplay between HSC and hepatocytes, induced by RPV. We hypothesised that STAT3 activation in hepatocytes was triggered in response to the secretome of apoptotic HSC. Thus, we collected the culture medium from HSC previously treated with RPV for 72 hours— and consequently rendered apoptotic—and cultured Hep3B cells with it. This treatment induced a concentration-dependent activation of STAT3 in Hep3B cells (figure 8B); despite not being very intense, the overexpression of pSTAT3 correlated closely with the discrete but constant pro-apoptotic effect of RPV in LX-2 cells. A significant increase in pSTAT3 was also observed in Hep3B cells treated with the conditioned medium of RPV-treated hHSC (data not shown). Figure 6 Rilpivirine (RPV) inactivates primary human hepatic stellate cells. (A) Relative mRNA expression levels of several fibrogenic markers of hepatic stellate cell (HSC) activation after 72 hours of treatment. Data were normalised vs the housekeeping gene ACTB (-actin) (n=3). (B) Representative western blot analysis images and quantification of Col1a1, phospho-STAT3 and phospho-STAT1 in human HSC pre-incubated with fludarabine (Flu, 2 hour) and treated for 72 hours (n=3). Data (mean±SD) were analysed by a Student’s t-test (#p<0.05 for transforming growth factor beta (TGF) vs vehicle (Veh)) or one-way analysis of variance followed by a Bonferroni test (*p<0.05 vs the respective Veh—with or without TGF). Additionally, we confirmed that STAT3 activation in hepato- cytes was dependent on the number of apoptotic LX-2 cells in response to RPV, as treatment of Hep3B cells with conditioned medium from LX-2 cells incubated for shorter periods of time (24–48 hours) did not induce any change in STAT3 activation (data not shown). It is highly relevant that STAT1 silencing in LX-2 cells abolished this overexpression of pSTAT3 in Hep3B cells, since STAT1 deletion in LX-2 cells blocked the pro-apop- totic effect of RPV (figure 8C). We also induced apoptosis in LX-2 cells with pro-apoptotic molecules whose cytotoxic mech- anisms do not involve JAK-STAT signalling, such as STS and etoposide; incubation with these conditioned media did not enhance pSTAT3 expression in hepatocytes (figure 8D). To further explore the molecules responsible for this activa- tion of STAT3 in hepatocytes, we employed a cytokine array to analyse the composition of the HSC secretome. We found that RPV treatment induced alterations in different cytokines, with the enhancement in IFN and IL-6 proving to be statis- tically significant (figure 8E, F, online supplementary figure 6). Finally, the use of soluble neutralising anti-IL-6 and anti- IFN antibodies revealed that IL-6 plays a main role in the HSC-hepatocyte crosstalk by inducing STAT3 phosphoryla- tion (figure 8G). RPV-administered patients with HIV show a better liver profile than patients treated with RPV-free antiretroviral therapies In order to identify clinical data supporting the beneficial effect of RPV, we performed a retrospective analysis of the MACS data set, comparing blood parameters associated with liver function or lipid and glucose metabolism between patients receiving RPV-containing vs RPV-free antiretroviral regimens. Impor- tantly, the former group exhibited better liver function than the latter, as shown by a significant reduction in liver enzymes, bilirubin, albumin and prothrombin time (figure 9). Similarly, RPV-treated patients displayed lower levels of total cholesterol and glycated haemoglobin (among other related parameters) Figure 7 Signal transducer and activator of transcription (STAT)1 signalling is crucial for rilpivirine (RPV)-induced apoptosis of hepatic stellate cells. (A) Relative mRNA expression levels of STAT1 in LX-2 cells transiently transfected with siC and siSTAT1, and a representative image of PCR bands. Data were normalised vs the housekeeping gene ACTB (-actin) (n=4). (B) Representative western blot analysis images and histograms expressing quantification of Col1a1, phospho-STAT3 and phospho-STAT1 in LX-2 cells transfected with siC and siSTAT1 after 48 hours treatment (n=4). (C) Flow cytometry assay of apoptosis induction in treated LX-2 cells transfected with siC and siSTAT1 (quantification of apoptotic cells—%—and representative cytometric plots) (n=4). Data (mean±SD) were analysed by a Student’s t-test (#p<0.05 for transforming growth factor beta (TGF) vs vehicle (Veh)) or one-way analysis of variance followed by a Bonferroni test (*p<0.05 vs the respective Veh—with or without TGF). clearly pointing to an improvement in lipid and glucose metabo- lism (online supplementary figure 7). DISCUSSION In this study, we demonstrate for the first time the hepatopro- tective effects of the antiretroviral drug RPV on chronic liver disease. Specifically, this compound exerts clear antifibrogenic effects in different mouse models of liver injury, pointing to a direct protective mechanism in injured livers, regardless of the disease’s aetiology. Importantly, RPV decreased severe liver inflammation and fibrosis progression when administered in combination with the toxic insult, and it reduced hepatic fibrosis and inflammation when administered after consolidation of fibrosis, thus highlighting the potential of this compound to Figure 8 Rilpivirine (RPV)-induced apoptosis in hepatic stellate cells promotes STAT3 activation in hepatocytes via interleukin (IL)-6 secretion. (A) Representative western blot analysis images and quantification of phospho-STAT3 in Hep3B cells treated for 24 hours (n=4). (B) Representative western blot analysis and quantification of phospho-STAT3 in Hep3B cells treated (16 hours) with conditioned medium from naïve and transforming growth factor beta (TGF)-activated LX-2 cells treated with RPV for 72 hours (n=4). (C) Representative western blot analysis and quantification of phospho-STAT3 in Hep3B cells treated (16 hours) with conditioned medium from siC and siSTAT1 LX-2 cells treated with RPV for 72 hours (n=4). (D) Representative western blot analysis of phospho-STAT3 in Hep3B cells treated (16 hours) with conditioned medium from LX-2 cells treated with staurosporine (0.5 and 1 µM) and etoposide (20 and 40 µM) for 24 hours (n=4). (E) Representative cytokine array performed with conditioned medium from TGF-activated LX-2 cells treated with or without RPV for 72 hours. (F) Quantification of alterations of cytokine content in conditioned medium from TGF-activated LX-2 cells treated with RPV for 72 hours (n=3). (G) Representative western blot analysis and quantification of phospho-STAT3 in Hep3B cells treated (16 hours) with conditioned medium from naïve and TGF-activated LX-2 cells treated with RPV for 72 hours and incubated with anti-IL-6 and/or anti-interferon gamma (anti-IFN) neutralising antibodies (1 hour) (n=3). Data (mean±SD) were analysed by a Student’s t-test (#p<0.05 for TGF vs vehicle (Veh)) or one-way analysis of variance followed by a Bonferroni test (*p<0.05 vs the respective Veh—with or without TGF). reverse fibrosis. Mechanistic studies revealed an interesting and selective pro-apoptotic effect of RPV in activated HSC through STAT1 signalling, which was not observed in hepatocytes, in accordance with previous reports of a lack of cytotoxic effects in parenchymal cells.18 19 Furthermore, we describe an inter- esting interplay between HSC and hepatocytes via JAK-STAT signalling, which leads to liver regeneration: RPV treatment also restored the number of proliferative hepatocytes through STAT3 activation, thereby re-establishing parenchymal homeostasis. Nevertheless, this activation was secondary to and dependent on the RPV-mediated apoptosis of HSC. RPV is commonly used as a component of anti-HIV therapies and, consequently, its safety profile in chronic treatments is well known. Although some mild-to-moderate biochemical abnormal- ities have been pinpointed as RPV-related adverse hepatic events, these alterations are soon normalised.20 Overall, clinical evidence Figure 9 Rilpivirine (RPV)-treated patients with HIV display a better liver function than patients on RPV-free antiretroviral regimens. Blood parameters associated with liver function and obtained from patients included in the Multicenter AIDS Cohort Study data set that were on active antiretroviral treatments from 2011 to 2014 (GGT, ALT, AST, Bilirubin, Albumin, Prot. time). Data (mean±SEM) were analysed by a non-parametric Mann-Whitney U test between RPV-free (n=930) and RPV-containing (n=55) groups (*p<0.05). ALT, alanine aminotransferase; AST, aspartate aminotransferase; Prot. time, prothrombin time; GGT, gamma glutamyl transpeptidase. supports the use of RPV in lifelong treatments, with some recent reports even suggesting its beneficial effects on several clinical parameters after switching from other non-RPV-containing regi- mens.21 22 However, its particular contribution to liver disease progression and its actions on different cell subpopulations under varying pathological hepatic conditions were unknown previously to the present study. One of the main reasons why the protective effect we describe here has been obscured until now is the fact that antiretroviral compounds are always administered in combined therapy, and, thus, their particular contribution is extremely difficult to determine in clinical practice. Significantly, the preliminary analysis of the MACS public data set we have carried out is in line with our preclinical results and affirms that RPV-treated patients with HIV display a better liver function than patients with HIV treated with antiretroviral regimens not containing RPV. Further research must be performed to deter- mine whether this improved liver outcome is directly related to RPV treatment, but we feel our data undoubtedly constitute an exciting starting point that supports the necessity of more specific clinical trials. All in all, our results may be important for the management of HIV-infected patients, as they endorse the clinical use of RPV in individuals who are specially suscep- tible to liver disease (eg, those with severe metabolic syndrome or HCV-HIV co-infected patients). Even though the highest concentrations of RPV employed in our in vitro experiments are supratherapeutic considering patients’ plasma levels, intrahe- patic concentrations can be even higher due to bioaccumulation of the compound in this organ.20 23 24 In addition to its antifibrogenic effects, it is important to note that our data describe a marked anti-adipogenic and anti-inflam- matory capacity of RPV in our animal models. While further experiments are needed to characterise in more depth the molecular mechanisms involved in this anti-adipogenic effect, we consider these actions extremely relevant and promising in the context of the NAFLD pandemic, as they may help identify new drug targets.25–27 Although JAK-STAT signalling has been widely debated as a promising target for the treatment of chronic liver disease,12 13 28 no molecules able to modulate the balance between STAT1 and STAT3 and to consequently ameliorate liver fibrosis have been described to date. In this study, we identify a hepatoprotective compound capable of inducing a differential effect in hepato- cytes and HSC via JAK-STAT signalling, leading to liver regen- eration. RPV-mediated hepatoprotection primarily stems from its ability to directly induce apoptosis in HSC through STAT1 activation. Secondary to this effect, we have also observed an intense proliferative response in liver parenchyma driven by the restoration of STAT3 signalling. In light of this combined effect, we believe that RPV-mediated effects in the liver constitute a regenerative response, and not only a mere consequence of the discontinuation of injury progression produced by HSC inacti- vation; thus, the compound may also have potential as a novel therapeutic agent. RPV-induced actions show certain interesting and unexpected characteristics that make it unique and may support its clinical application in liver diseases. Importantly, this non-nucleoside reverse transcriptase inhibitor (NNRTI) significantly improves liver function when administered to severely damaged tissue, even when the insult is still present. Moreover, RPV differentially regulates JAK-STAT signalling pathways in both parenchymal and non-parenchymal cells, eventually triggering a regenera- tive response. Considering the limited number of drugs that are effective in improving liver function when the organ is chron- ically damaged, the therapeutic potential of RPV is exciting.29–33 In our study, STAT1 signalling proved to be fundamental in the hepatoprotective actions of RPV, suggesting it is a novel and promising target worthy of further investigation in liver fibrosis and in other fibrotic disorders. Other drugs have previously been reported to induce beneficial effects in liver diseases through direct effects on STAT3 regulation. Specifically, sorafenib amelio- rates liver fibrosis through direct STAT3 inhibition in HSC,34 and through STAT3 activation in hepatocytes by means of the IL-6 released by local Kupffer cells.35 However, the clinical use of sorafenib as an antifibrotic agent has important limitations, as it is an extremely toxic compound, with a narrow therapeutic window. Additionally, RPV-induced and sorafenib-induced hepatoprotective mechanisms seem to be complementary; thus, it would be interesting to analyse their synergic effects. Finally, we have also demonstrated that IL-6 secreted by RPV-induced apoptotic HSC is responsible for the specific activation of STAT3 in hepatocytes, in a process dependent on STAT1 activation. These data could help to guide novel ther- apeutic approaches to liver diseases, especially considering previous reports on the role of this cytokine in liver regenera- tion.35 36 In conclusion, this study demonstrates, for the first time, a selective regulation of STAT1 in HSC by an FDA-approved drug—RPV—that prevents liver fibrosis and inflammation and enhances liver regeneration in different experimental models of chronic liver injury. Our results suggest RPV is especially indi- cated in HIV-infected patients with particular susceptibility to certain liver diseases, an idea that is supported by the clinical data analysed. Moreover, it may have great clinical relevance as a key factor in improving pharmacological therapy for hepatic diseases with a fibrotic component, serving as the basis for discovering novel targets for clinical intervention, as a new ther- apeutic strategy itself, and improving our current understanding of the molecular mechanisms involved in liver fibrosis. Acknowledgements The authors would like to thank Brian Normanly for his English language editing and help with the preparation of the manuscript, María Miralles for her assistance during the in vivo experiments, Professor Scott Friedman of the Icahn School of Medicine at Mount Sinai for kindly providing LX-2 cells and Vicente Arnau for his technical support. Clinical data in this manuscript were collected by the Multicenter AIDS Cohort Study (MACS) with centres at Baltimore (U01-AI35042): The Johns Hopkins University Bloomberg School of Public Health: Joseph B. Margolick (PI), Todd Brown (PI), Jay Bream, Adrian Dobs, Michelle Estrella, W. David Hardy, Lisette Johnson-Hill, Sean Leng, Anne Monroe, Cynthia Munro, Michael W. Plankey, Wendy Post, Ned Sacktor, Jennifer Schrack, Chloe Thio; Chicago (U01-AI35039): Feinberg School of Medicine, Northwestern University and Cook County Bureau of Health Services: Steven M. Wolinsky (PI), Sheila Badri, Dana Gabuzda, Frank J. Palella, Jr., Sudhir Penugonda, John P. Phair, Susheel Reddy, Matthew Stephens, Linda Teplin; Los Angeles (U01-AI35040): University of California, UCLA Schools of Public Health and Medicine: Roger Detels (PI), Otoniel Martínez-Maza (PI), Otto Yang (Co-PI), Peter Anton, Robert Bolan, Elizabeth Breen, Anthony Butch, Shehnaz Hussain, Beth Jamieson, John Oishi, Harry Vinters, Dorothy Wiley, Mallory Witt, Stephen Young, Zuo Feng Zhang; Pittsburgh (U01-AI35041): University of Pittsburgh, Graduate School of Public Health: Charles R. Rinaldo (PI), Lawrence A. Kingsley (PI), Jeremy J. Martinson (PI), James T. Becker, Phalguni Gupta, Kenneth Ho, Susan Koletar, John W. Mellors, Anthony J. Silvestre, Ronald D. Stall; Data Coordinating Center (UM1-AI35043): The Johns Hopkins University Bloomberg School of Public Health: Lisa P. Jacobson (PI), Gypsyamber D’Souza (PI), Alison Abraham, Keri Althoff, Michael Collaco, Priya Duggal, Sabina Haberlen, Eithne Keelaghan, Heather McKay, Alvaro Muñoz, Derek Ng, Anne Rostich, Eric C. Seaberg, Sol Su, Pamela Surkan, Nicholas Wada. Institute of Allergy and Infectious Diseases: Robin E. Huebner; National Cancer Institute: Geraldina Dominguez. Contributors AB-G, AM-R and JVE designed the experiments and analysed the data; AB-G, AM-R, FA, ÁBM, FG-G, PM-R and AF-I performed the experiments; AB-G, JG-S, NA and JVE wrote the manuscript. Funding This work was supported by grants PI14/0312 (from the Fund for Health Research—FIS, co-funded by the European Regional Development Fund of the European Union—’A way to build Europe’), CIBER CB06/04/0071 and EHD19PI03 (both from Instituto de Salud Carlos III, Ministerio de Economia y Competitividad) and by grants SAF2015-67678-R, RTI2018-096748-B-I00 (from Ministerio de Economia y Competitividad, co-funded by the European Regional Development Fund of the European Union) and PROMETEO2018/141 (from Conselleria d’Educació, Investigació, Cultura i Esport, Generalitat Valenciana). AM-R and ÁBM are recipients of Predoctoral Trainee Research Grants (FPU13/00151 and FPU16/05896, respectively; Ministerio de Educación, Cultura y Deporte); AM-R also received a local grant from Fundación Juan Esplugues. MACS is funded primarily by the National Institute of Allergy and Infectious Diseases (NIAID), with additional co-funding from the National Cancer Institute (NCI), the National Institute on Drug Abuse (NIDA) and the National Institute of Mental Health (NIMH). Targeted supplemental funding for specific projects was also provided by the National Heart, Lung and Blood Institute (NHLBI), and the National Institute on Deafness and Communication Disorders (NIDCD). MACS data collection is also supported by UL1-TR001079 (JHU ICTR) from the National Center for Advancing Translational Sciences (NCATS), a component of the National Institutes of Health (NIH), and NIH Roadmap for Medical Research. Disclaimer The contents of this publication are solely the responsibility of the authors and do not represent the official views of the National Institutes of Health (NIH), Johns Hopkins ICTR or NCATS. Competing interests None declared. Patient consent for publication Not required. Ethics approval The protocol was approved by the Ethics Committee of Hospital Clínic (Barcelona, Spain) (HCB/2015/0624). Provenance and peer review Not commissioned; externally peer reviewed. Data availability statement Data are available in a public, open-access repository. All data relevant to the study are included in the article or uploaded as supplementary information. REFERENCEs 1 Lee YA, Wallace MC, Friedman SL. Pathobiology of liver fibrosis: a translational success story. Gut 2015;64:830–41. 2 Merli M, Galli L, Marinaro L, et al. Pharmacokinetics of dolutegravir and rilpivirine in combination with simeprevir and sofosbuvir in HIV/hepatitis C virus-coinfected patients with liver cirrhosis. J Antimicrob Chemother 2017;72:812–5. 3 Thamrongwonglert P, Chetchotisakd P, Anunnatsiri S, et al. Improvement of lipid profiles when switching from efavirenz to rilpivirine in HIV-infected patients with dyslipidemia. HIV Clin Trials 2016;17:12–16. 4 Tebas P, Sension M, Arribas J, et al. Lipid levels and changes in body fat distribution in treatment-naive, HIV-1-infected adults treated with rilpivirine or efavirenz for 96 weeks in the ECHO and THRIVE trials. Clin Infect Dis 2014;59:425–34. 5 Gagliardini R, Bandera A, Zaccarelli M, et al. 3-Year efficacy and durability of simplification to single tablet regimens: a comparison between co-formulated efavirenz/emtricitabine/tenofovir and rilpivirine/emtricitabine/tenofovir. Antivir Ther 2017;23:139–48. 6 Neukam K, Espinosa N, Collado A, et al. hEPAtic safety of Rilpivirine/Emtricitabine/ Tenofovir disoproxil fumarate fixed-dose Single-Tablet regimen in HIV-infected patients with active hepatitis C virus infection: the hEPAtic study. PLoS One 2016;11:e0155842. 7 Wilkins EL, Cohen CJ, Trottier B, et al. Patient-reported outcomes in the single-tablet regimen (StAR) trial of rilpivirine/emtricitabine/tenofovir disoproxil fumarate versus efavirenz/emtricitabine/tenofovir disoproxil fumarate in antiretroviral treatment- naive adults infected with HIV-1 through 48 weeks of treatment. AIDS Care 2016;28:401–8. 8 Friedman SL. Mechanisms of hepatic fibrogenesis. Gastroenterology 2008;134:1655–69. 9 Zhang H, Chen F, Fan X, et al. Quantitative proteomic analysis on activated hepatic stellate cells reversion reveal STAT1 as a key regulator between liver fibrosis and recovery. Sci Rep 2017;7:44910. 10 Gao B, Wang H, Lafdil F, et al. STAT proteins - key regulators of anti-viral responses, inflammation, and tumorigenesis in the liver. J Hepatol 2012;57:430–41. 11 Mair M, Zollner G, Schneller D, et al. Signal transducer and activator of transcription 3 protects from liver injury and fibrosis in a mouse model of sclerosing cholangitis. Gastroenterology 2010;138:2499–508. 12 Xiang D-M, Sun W, Ning B-F, et al. The HLF/IL-6/STAT3 feedforward circuit drives hepatic stellate cell activation to promote liver fibrosis. Gut 2018;67:1704–15. 13 Park O, Wang H, Weng H, et al. In vivo consequences of liver-specific interleukin-22 expression in mice: implications for human liver disease progression. Hepatology 2011;54:252–61. 14 Blas-García A, Apostolova N, Ballesteros D, et al. Inhibition of mitochondrial function by efavirenz increases lipid content in hepatic cells. Hepatology 2010;52:115–25. 15 de Mesquita FC, Guixé-Muntet S, Fernández-Iglesias A, et al. Liraglutide improves liver microvascular dysfunction in cirrhosis: evidence from translational studies. Sci Rep 2017;7:3255. 16 Reagan-Shaw S, Nihal M, Ahmad N. Dose translation from animal to human studies revisited. FASEB J 2008;22:659–61. 17 Edgar R, Domrachev M, Lash AE. Gene expression Omnibus: NCBI gene expression and hybridization array data Repository. Nucleic Acids Res 2002;30:207–10. 18 Sharma D, Lau AJ, Sherman MA, et al. Agonism of human pregnane X receptor by rilpivirine and etravirine: comparison with first generation non-nucleoside reverse transcriptase inhibitors. Biochem Pharmacol 2013;85:1700–11. 19 Sharma D, Lau AJ, Sherman MA, et al. Differential activation of human constitutive androstane receptor and its SV23 and SV24 splice variants by rilpivirine and etravirine. Br J Pharmacol 2015;172:1263–76. 20 Domingo P, Lozano F. [Management of antiretroviral drug toxicity]. Enferm Infecc Microbiol Clin 2011;29:535–44. 21 Hagins D, Orkin C, Daar ES, et al. Switching to coformulated rilpivirine (RPV), emtricitabine (FTC) and tenofovir alafenamide from either RPV, FTC and tenofovir disoproxil fumarate (TDF) or efavirenz, FTC and TDF: 96-week results from two randomized clinical trials. HIV Med 2018;19:724–33. 22 Taramasso L, Tatarelli P, Ricci E, et al. Improvement of lipid profile after switching from efavirenz or ritonavir-boosted protease inhibitors to rilpivirine or once-daily integrase inhibitors: results from a large observational cohort study (SCOLTA). BMC Infect Dis 2018;18:357. 23 Thompson MA, Aberg JA, Hoy JF, et al. Antiretroviral treatment of adult HIV infection. JAMA 2012;308:387–402. 24 Casado JL. Liver toxicity in HIV-infected patients receiving novel second-generation nonnucleoside reverse transcriptase inhibitors etravirine and rilpivirine. AIDS Rev 2013;15:139–45.
25 Friedman SL, Neuschwander-Tetri BA, Rinella M, et al. Mechanisms of NAFLD development and therapeutic strategies. Nat Med 2018;24:908–22.
26 Younossi Z, Tacke F, Arrese M, et al. Global perspectives on non-alcoholic fatty liver disease and non-alcoholic steatohepatitis. Hepatology 2018.

27 Tsuchida T, Lee YA, Fujiwara N, et al. A simple diet- and chemical-induced murine NASH model with rapid progression of steatohepatitis, fibrosis and liver cancer. J Hepatol 2018;69:385–95.
28 Mair M, Blaas L, Österreicher CH, et al. JAK-STAT signaling in hepatic fibrosis. Front Biosci 2011;16:2794–811.
29 Ambade A, Lowe P, Kodys K, et al. Pharmacological inhibition of CCR2/5 signaling prevents and reverses alcoholinduced liver damage, steatosis and inflammation in mice. Hepatology 2018.
30 Krenkel O, Puengel T, Govaere O, et al. Therapeutic inhibition of inflammatory monocyte recruitment reduces steatohepatitis and liver fibrosis. Hepatology 2018;67:1270–83.
31 Lefebvre E, Moyle G, Reshef R, et al. Antifibrotic effects of the dual CCR2/CCR5 antagonist Cenicriviroc in animal models of liver and kidney fibrosis. PLoS One 2016;11:e0158156.
32 Tripathi DM, Vilaseca M, Lafoz E, et al. Simvastatin prevents progression of acute on chronic liver failure in rats with cirrhosis and portal hypertension. Gastroenterology 2018;155:1564–77.
33 Gracia-Sancho J, García-Calderó H, Hide D, et al. Simvastatin maintains function and viability of steatotic rat livers procured for transplantation. J Hepatol 2013;58:1140–6.
34 Su T-H, Shiau C-W, Jao P, et al. Sorafenib and its derivative SC-1 exhibit antifibrotic effects through signal transducer and activator of transcription 3 inhibition. Proc Natl Acad Sci U S A 2015;112:7243–8.
35 Deng Y-R, Ma H-D, Tsuneyama K, et al. STAT3-mediated attenuation of CCl4- induced mouse liver fibrosis by the protein kinase inhibitor sorafenib. J Autoimmun 2013;46:25–34.
36 Yu H, Lee H, Herrmann A, et al. Revisiting STAT3 signalling in cancer: new and unexpected biological functions. Nat Rev Cancer 2014;14:736–46.