Inflammatory and fibrotic mechanisms in NAFLD - implications for new treatment strategies

Abstract

Non-alcoholic fatty liver disease is comprised of either simple steatosis (non-alcoholic fatty liver, (NAFL)) or a more advanced inflammatory and fibrogenic stage (non-alcoholic steatohepatitis (NASH). NASH affects a growing proportion of the global adult and pediatric population, leading to rising rates of liver fibrosis and hepatocellular carcinoma.

NASH is a multifactorial disease that is part of a systemic metabolic disorder. Here, we provide an overview of the metabolic underpinnings of NASH pathogenesis and established drivers of inflammation and fibrosis. Clarification of underlying fibrogenic and inflammatory mechanisms will advance the development of novel treatment strategies as there are no approved therapies at present. We discuss emerging experimental approaches and potential novel investigational strategies derived from animal models including the inflammasome, epigenetic reprogramming, Hippo signaling, Notch signaling, engineered T cells to remove fibrogenic HSCs, and HSC-specific targeting therapies. Recently completed and ongoing clinical trials and antifibrotics are discussed, illuminating the growing expectation that one or more therapies will yield clinical benefit in NASH in the coming years.

Graphical Abstract

Introduction

Non-alcoholic fatty liver disease (NAFLD) affects up to 25% of the global adult population linked to the steeply increasing epidemic prevalence of obesity worldwide and affecting an increasing number of children as well. Current environmental drivers of NAFLD include highly processed foods rich in fat and sugar made worse during the COVID-19 pandemic due to lockdowns and shelter-at-home measures, as well as increased alcohol consumption in many countries [1].

NAFLD is comprised of either simple steatosis (non-alcoholic fatty liver, (NAFL)) or a more advanced inflammatory and fibrogenic stage (non-alcoholic steatohepatitis (NASH). NAFLD progresses in up to 25% of all patients to NASH, which leads to fibrosis and complications of end-stage liver disease including cirrhosis and hepatocellular carcinoma (HCC). In fact, NAFLD is the fastest growing cause for HCC [2] and indication for liver transplantation [3]. The impending public health burden to NASH has provoked explosive interest in the disease by clinicians, the pharmaceutical industry and scientists alike, and has rapidly expanded the pipeline for potential therapies, yet no drugs are approved to date for this disease, despite promising data in many animal studies.

Of the pathologic features of NASH fibrosis most directly impacts long-term outcome and mortality. However, definitive assessment of fibrosis still requires liver biopsy as the gold standard, while intense efforts to develop non-invasive diagnostics to supplant biopsy are rapidly evolving

Recent advances emphasize that NAFLD is the liver’s manifestation of a systemic disease known as the ‘metabolic syndrome’. Thus, weight loss with implementation of healthy lifestyle modifications (diet and physical exercise) are recognized as the most effective and cost-effective methods to rectify not only NAFLD but also other co-morbidities of the metabolic syndrome including hypertension, insulin resistance, and cardiovascular risk, as well as obesity-associated cancers. However, weight loss and life- style modifications are difficult to sustain, which has led to a growing focus on pharmacotherapies to treat NASH.

Complications of NASH include not only advanced fibrosis and cirrhosis but also HCC, or primary liver cancer. Almost all liver cancers develop on a fibrotic background, yet in contrast to other diseases such as chronic HBV or HCV, where > 90% have cirrhosis, in NASH up to 1/3 of all HCCs develop in non-cirrhotic patients [4]; thus, tumors tend to be larger and more advanced, reducing the hope of curative treatments. For HCC, public health challenges include the screening and identification of patients at risk and in need of surveillance for HCC. Importantly, it is essential to identify patients with fibrosis since this important feature drives outcomes and determines long-term outcome [5,6]. As a result, clinical trials for NASH therapies are increasingly focused on the endpoint of improving fibrosis without worsening of NASH, or ‘NASH resolution without worsening of fibrosis’. Accordingly, many therapeutic strategies target fibrogenesis or activation pathways of fibrogenic cells (hepatic stellate cells) and mitigation of hepatic inflammation, which drives fibrosis. Nonetheless, systemic approaches and extrahepatic targets may also contribute to disease improvement, for example targeting the gut microbiome, adipose tissue, as well as the immune system, among others. In this article, we review liver-specific as well as systemic disease mechanisms contributing to inflammation and fibrotic mechanisms in NAFLD. While many exciting drugs are in the NASH study pipeline - reviewed in this and future issues of this journal - here, we focus on antifibrotic therapeutics currently under investigation.

Pathogenesis of NAFLD - inflammatory and fibrogenic mechanisms

NAFLD is defined as steatosis of more than 5% of hepatocytes in the absence of excess alcohol consumption, while excluding other underlying liver diseases. Early studies in NASH proposed a two-hit model with steatosis as the ‘first hit’ and suggested that a ‘second hit’, for example oxidative stress, provoked disease onset. More recently, however, more nuanced models of pathogenesis emerged. For example, Marchesini et al associated NAFLD with insulin resistance [7] as a feature of the metabolic syndrome [8]. Now NASH is viewed as a metabolic stress response of the liver from an overabundance of fatty acids and carbohydrates. Overnutrition from high-caloric foods, coupled with low energy expenditure due to a sedentary lifestyle, promotes adipose tissue hypertrophy to store extra calories, which can provoke adipose tissue hypoxia owing to the greatly expanded adipocyte mass. Unlike other tissues, relative hypoxia is insufficient to lead to vascularization but instead results in adipose tissue fibrosis and inflammation, which further decrease adipose tissue function and increase the circulating levels of glucose and lipids [9]. Apart from adipose tissue, fatty acids are also derived from the liver through de novo lipogenesis, especially from high carbohydrate content, especially fructose, in sugared beverages.

Lipid substrates are metabolized by beta oxidation in the mitochondria or peroxisomes, or transiently stored in hepatocytes as inert triglycerides before being secreted within VLDLs. Impairment in either hepatic downstream pathways or overabundance of available substrates leads to generation of lipotoxic lipid species which result in ER stress, mitochondrial dysfunction, and oxidative stress, ultimately leading to hepatocyte injury, genomic instability, inflammation, hepatic stellate cell activation and fibrosis complicated by hepatocellular cancer. Factors triggering hepatic inflammation and NASH originate both from inside the liver (lipotoxicity, apoptosis, mitochondrial dysfunction, and ER stress, innate immune response) as well as in extrahepatic sites. A ‘Western diet’ can induce gut dysbiosis leading to altered gut microbiome composition, destabilization of gut barrier function and translocation of bacteria and bacterial products, which then amplify inflammation, stimulate ethanol production by the gut, and alter bile acid signaling. Insulin resistance, which impairs glucose uptake in extrahepatic tissues (adipose tissue, skeletal muscle) leads to increased release of fatty acids, thereby enhancing hepatic de novo lipogenesis of glucose to fatty acids. In addition, genetic polymorphisms, and other factors (e.g., smoking, fructose, sleep apnea) further contribute to this multifactorial disease. Some of these factors are independent of obesity or insulin resistance, as NASH can occur in lean patients in 10–20% of those with the disease. It is likely that there is significant variability in which disease drivers lead to NASH across the affected population, and indeed NASH may reflect the end-result of different inputs and is more than one disease.

Intrahepatic inflammatory and fibrogenic pathways

ER Stress.

The role of ER stress in the pathogenesis of NAFLD is well established [10,11]. The endoplasmic reticulum (ER) is central to lipid metabolism, as many proteins regulating lipid metabolism are located in the ER, including the transcription factors regulating hepatic de novo lipogenesis, enzymes involved in triglyceride synthesis, and other proteins involved in lipoprotein assembly (e.g., VLDLs). ER homeostasis is thus essential to maintain intrahepatic lipid metabolism, membrane lipid composition and extrahepatic plasma lipid homeostasis. Chronic overabundance of saturated fatty acids, such as palmitate, lead to ER stress, which activates the unfolded protein response (UPR) to restore ER function. Moreover, the UPR may activate the master lipogenic regulator sterol regulatory element-binding protein (SREBP)-1c and SREB2 [12] and thus stimulate key regulatory genes of hepatic de novo lipogenesis including fatty acid synthase (FAS), stearoyl CoA desaturase 1 (SCD-1), diacylglycerol acyltransferase (DGAT) and acetyl-CoA carboxylase (ACC). The UPR also decreases peroxisome proliferator-activated receptor (PPAR)g and FXR [13], as well as increasing expression of VLDL receptor. Together, these effects further enhance hepatic steatosis, hypertriglyceridemia, and generation of lipotoxic metabolites, thus worsening hepatocellular injury and apoptosis [10]. If ER stress is not alleviated, excessive or prolonged UPR activation can lead to inflammation, inflammasome activation and hepatocyte death.

Some compounds aim to reduce ER stress by increasing protein folding capacity in the ER and have shown hepatoprotective effects in preclinical models. For example, 4-phenyl butyric acid (PBA), a short chain fatty acid derivative decreases steatosis in a genetic mouse model, ob/ob mice [14]. Although 4-PBA is approved for the treatment of urea cycle disorders, its efficacy in NASH needs to be further explored. The taurine-conjugated ursodeoxycholic acid (TUDCA) is approved for primary biliary cirrhosis, can improve steatosis in mice [14] and is hepatoprotective in rats [15]. Further analyses are required to assess their efficacy and applicability in NAFLD. Similarly, inhibitors of the UPR such as STF-083010 are well characterized to inhibit distinct UPR signaling pathways. STF-083010 inhibits IRE1a and can protect BI-1-deficient mice from HFD-induced NASH [16].

Activation of apoptotic pathways.

Prolonged ER stress may result in hepatocellular apoptosis (programmed cell death), which in turn may amplify liver inflammation by activating fibrogenic hepatic stellate cells (HSCs) and Kupffer cells. HSCs can engulf apoptotic bodies, which enhances fibrogenesis. Kupffer cells phagocytose apoptotic bodies and release TNFαcontributing to inflammation. Apoptotic signaling pathways such as TRAIL (TNFSF10) are pro-inflammatory by releasing extracellular vesicles and chemokines that stimulate immune cells [17]. In preclinical models and NAFLD patients, apoptosis signal regulating kinase 1 (ASK1) is highly activated, leading to activation of JNK and p38 MAPK pathways, which results in cell death and inflammation. ASK1 may also be activated by as an intermediary of the IRE1a/JNK pathway of the UPR. Therapeutic approaches targeting these pathways have included caspase inhibitors (emricasan) and ASK1 inhibitors (selonsertib). ASK1 is expressed by HSCs and inflammatory cells, and thus its inhibition was expected to block inflammation and fibrosis. Indeed, the ASK1 inhibitor selonsertib improved inflammation and decreased fibrosis in a short clinical trial of 24 weeks [18]. However, in two phase 3 trials that assessed fibrosis improvement in patients with bridging fibrosis (F3, Stellar-3) or compensated cirrhotic patients (F4, Stellar-4), respectively, the drug was ineffective, and the trials were terminated [19]. Some have argued that ASK1 inhibition remains an attractive target, but more potent molecules are required. Similar to the effects of selonsertib, a caspase inhibitor, emricasan, was not effective in patients with NASH fibrosis [20] or cirrhosis [21]. Results of emricasan (ENCORE-LF) assessing event-free survival have been posted yet (NCT03205345).

The unfolded protein response (UPR) may also decrease hepatic PPARα signaling [13]. The nuclear receptors, PPARα, γ, δ are key regulators in lipid metabolism: PPARγ, a key regulator in adipocyte differentiation, PPARα promotes hepatic fatty acid oxidation [22] and PPARδ decreases DNL and has anti-inflammatory effects [23]. Thus, PPAR agonists may alleviate lipid-related toxicity, promote hepatic regeneration and mitigate hepatocyte stress. Dual PPARα/δ agonists are effective in murine models of NASH, yet elafibranor a dual PPARα/δ agonist, failed in the phase III RESOLVE-IT trial. Elafibranor did not show significant advantage over placebo with regards to NASH resolution without worsening of fibrosis with a response rate of (19.2 vs. 14.7%). Similarly, it did reduce fibrosis (24% vs. 22.4%). Another drug, Lanifibranor, a PPARα/ δ/γ agonist, has improved inflammation and fibrosis in a phase 2b 6-month trial. Yet other trials targeting activation of PPARs are ongoing. These include a PPAR α/γ agonist Saroglitazaar (NCT03061721), and a PPAR α/γ/δ agonist, Lanifibranor (NCT03008070) [24–26]. Combination therapies of PPARγ agonist with a GLP1 agonist (NCT03950505) or SGLT1 inhibitor is also under evaluation.

Inhibition of hepatic de novo lipogenesis.

Several therapies are in clinical trials to inhibit hepatic de novo lipogenesis by directly targeting these key enzymes. Arachidyl-amido cholanoic acid (Aramchol) is a fatty acid-bile acid conjugate that inhibits SCD-1 in hepatocytes, decreasing MUFA and triglyceride synthesis. In hepatic stellate cells, aramchol additionally induces PPARγ and thus decreases collagen and αSMA production [27]. Additional mechanisms of action may include activation of AMPK and inhibition of mTORC1, thus activating beta oxidation and oxidative phosphorylation [28]. In preclinical models of NASH, Aramchol significantly decreases hepatic steatosis in preclinical models [29] and in a phase 2 trial [30]. Another phase 2b trial showed significant improvement in NASH without worsening of fibrosis, as well as a trend towards fibrosis improvement (ARREST, NCT02279524). A phase 3 trial is currently ongoing (ARMOR, NCT0410432). While phase 2 clinical trials assessing inhibitors to DGAT1 (LCQ908, NCT01811472) and ACC inhibitor PF-05221304 (Clesacostat, NCT03248882) were not successful, the ACC1/2 inhibitor Firsocostat reduced hepatic steatosis and improved serum markers of fibrosis and liver biochemistry [31]. When monotherapies appear not sufficiently effective, combination (dual or triple therapies) may have greater efficacy, and combination trials are ongoing. The nuclear hormone receptor LXRα antagonist Oltipraz may also activate AMPK and decrease SREBP-1c expression thus reducing fatty acid synthesis and increasing fatty acid oxidation [32]. A phase 2 study showed decreased hepatic steatosis in dose-dependent manner [33], and a phase 3 trial is ongoing (NCT04142749).

Autophagy.

Emerging evidence links macro-autophagy to NASH [34,35]. Autophagy may be a protective mechanism during ER stress, which can either inhibit or enhance autophagy [36]. Autophagic flux may also be impaired with aging [37] and features of the metabolic syndrome, including hyperlipidemia with altered membrane lipid composition and insulin resistance, are associated with down-regulation of autophagy-related genes and aging-related decreased chaperone mediated autophagy [38]. Lipid droplets are degraded by autophagy (lipophagy) [39] and hepatic impairment of autophagy leads to steatosis, fibrosis, and tumorigenesis via Yes-associated protein [40], HMGB1 [41] and NRF2 [42]. Autophagy is also inversely regulated by the AKT/mTOR pathway. Overnutrition leads to AKT/mTOR activation which then may lead to decreased autophagic flux and Yap activation as well [43]. The Yap paralog TAZ is over-expressed in dietary murine models of NAFLD and in patients with NAFLD, and it is critical in driving liver pathogenesis [44]; however, mechanisms leading to TAZ upregulation are unclear. While targeting the YAP/TAZ pathway is difficult due to its ubiquitous expression and tissue- specific functions (e.g., in stem cells), modulators of autophagy may offer therapeutic promise. Metformin, a widely used and pleiotrophic drug that enhances autophagy has been in clinical trials for NASH (NCT00303537).

The Innate Immune System.

Hepatocellular stress and injury lead to release of intracellular components including RNA, DNA, ATP, and proteins (e.g., heat shock proteins) that are usually not found outside of healthy cells. These signals function as danger-associated molecular patterns (DAMPs) that activate pattern recognition receptors (PPRs), which are expressed on Kupffer cells, stellate cells, lymphocytes, and dendritic cells [45–47]. PPRs may also be activated by pathogen-associated molecular patterns (PAMPs), which are bacterial or viral components reaching the liver due to NASH-associated alterations in gut microbiome and intestinal barrier function (see more below). Among PPRs, TLR2, TLR4 and TLR9 are increased in NASH liver samples and amplify liver inflammation. TLR4 inhibition showed promise in murine models of NASH; however, a clinical phase 2 trial with TLR4 inhibitor JKB-121 yield disappointing results (NCT02442687), with improved response rates in the placebo group instead of the treatment group [48].

Inflammasome.

(NOD-, LRR- and pyrin domain-containing protein 3) is an intracellular sensor that can be primed and activated by a wide variety of triggers including PAMPs, DAMPs, TNFa and cholesterol crystals [49]. NLPR3, together with ASC and procaspase 1, forms the NLRP3 inflammasome complex that releases proinflammatory cytokines IL-1b and IL-18 with pyroptotic cell death, further driving inflammation and fibrosis [50]. NLRP3 is increased in human NASH livers and its inhibition attenuates inflammation and fibrosis in a murine model of NASH [51]. NLRP3 is not only expressed in immune cells, but also hepatocytes, hepatic stellate cells and macrophages. While most cells require a two-step process, with priming and activation of the inflammasome, monocyte-derived macrophages may be rapidly activated in a one-step process by simultaneous TLR and NLRP3 stimulation [52]. The diarylsulfonylurea compound MCC950 is a potent and specific NLRP3 inhibitor that is effective in inhibiting fibrosis in a murine NASH model [51] and in other immune organ disease models including atherosclerosis [53], diabetes [54], and colitis [55]. Unfortunately, a phase 2 clinical trial of MCC950 for rheumatoid arthritis was suspended due to hepatic toxicity [56]. Other inhibitors of the inflammasome (SGM-1019) did not improve hepatic inflammation, and a phase 2b clinical trial was terminated (NCT03676231). It is possible that more specific, less hepatotoxic compounds may be more successful in the future.

IL-17A.

This interleukin contributes to NAFLD pathogenesis [57]. High-fat diet-fed mice exhibit increases levels of the unconventional prefoldin RPB5 interactor (URI), thereby inducing DNA damage and leading to Th17 infiltration with subsequent IL-17A release; this in turn promotes hepatic inflammation, NASH and HCC [58]. Notably, hepatic URI expression leads to neutrophil infiltration in white adipose tissue and thus indirectly provokes insulin resistance. Conditional deletion of URI or inhibition of IL-17A signaling by digoxin has led to marked improvement. Digoxin, a RORγt inhibitor can decrease circulating Th17 cells. However, digoxin may be beneficial in NASH through other pathways as well. For example, digoxin alleviates steatohepatitis by suppression of the Pyruvate Kinase M2 (PKM2)-mediated transactivation of HIF-1α, thus reducing oxidative stress to decrease sterile inflammation in murine models of ASH and NASH [59,60]. Digoxin is currently being tested in an open label phase 2 trial (NCT04216693). Indeed, repurposing studies may uncover more small molecules for novel applications [61–63].

BET inhibitors.

Epigenetic reprogramming modulates NASH pathogenesis in animal models [64,65]. Chromatin readers bind to modified histones to regulate gene expression. Bromodomain and extra terminal (BET) family members consist of BRD2–4 and contain two bromodomain motifs (BD1 and BD2) to bind to acetylated histones.

While modulation of epigenetic programming has been an attractive target in cancer due to is potential modulation of gene expression, clinical trials of BRD4 inhibition in solid cancers have largely been terminated due to lack of efficacy. While greatest focus has been on BRD4, more specific inhibitors revealed immunomodulatory functions for BD1/BD2 and inflammatory functions for BD2 that are independent of BRD4. In fact, BD2 specific inhibitors decrease not only steatosis, lobular inflammation and hepatic fibrosisl, but also inflammatory arthritis, and colitis in preclinical models [64]. In a DEN/CDAHFD model, a small molecule inhibitor to BRD4, JQ1, can decrease fibrosis, inflammation, tumor formation and interestingly, also suppresses pathways related to oxidative phosphorylation, bile acid and fatty acid metabolism [65]. A broadly active BET family inhibitor (I-BET151) has improved NASH and fibrosis in a mouse NASH model [66]. One study has demonstrated that BRD4 inhibitors and FXR agonists should not be combined due to antagonistic functions [67]. Targeted epigenetic modulation of stellate cells via synaptophysin has been efficient to inhibit liver fibrosis in a proof- of-concept study using a histone methyltransferase inhibitor, 3-deazaneplanocin A, in a mouse model thus avoiding unwanted systemic effects [68].

Liver macrophages

Tissue resident macrophages (Kupffer cells) and circulating macrophages recruited from the bone marrow are a central part of the liver’s innate immune system, and are drivers of NAFLD pathogenesis (reviewed in detail by [69,70]). Activated Kupffer cells secrete TNF-α, IL-1β and CCL2/CCL5 to recruit proinflammatory monocyte-derived macrophages.

In murine models of NASH and in human liver samples, the C-C chemokine receptor 2 (CCR2) is highly expressed on monocyte-derived proinflammatory macrophages and positively correlated with NAFLD severity in obese patients. While the dual CCR2/CCR5 inhibitor CVC, which also targets stellate cells, showed exceptional results in rodent models of NASH [71] a phase 3 clinical trial (AURORA) was terminated in December 2020 due to lack of efficacy.

Next-generation techniques such as single cell RNAseq of macrophages in NASH have demonstrated epigenetic reprogramming of macrophage populations and have clarified the dynamics of infiltrating and tissue resident macrophages [72]. Interestingly, in humans a hypomorphic variant of MERTK, a macrophage c-mer tyrosine kinase receptor, conferred protection against liver fibrosis. In murine models of NASH, MerTK promotes liver fibrosis via an ERK-TGFb1 pathway, thus providing a plausible mechanism for macrophage MERTK as a genetic risk factor for NASH fibrosis [73]. Exciting new insights into macrophage heterogeneity and underlying mechanisms portend a fertile path for the discovery of novel therapeutic targets for NASH.

Fibrosis, the Role of Hepatic Stellate cells, and Antifibrotic Therapies for NASH

Liver fibrosis is the accumulation of excessive extracellular matrix proteins, leading to deposition of scar tissue that may distort the liver’s architecture over time. Hepatic stellate cells (HSCs) are the primary non-parenchymal liver cell population, contributing to 80–95% of fibrogenic myofibroblasts in chronic liver diseases. Apart from fibrogenesis, HSCs also contribute to diverse functions including immunity and inflammation, energy and nutrient homeostasis and immunometabolism. HSCs are the main storage site for vitamin A in the body and are readily recognizable by their autofluorescent perinuclear lipid droplets. They are located at the interface of hepatocytes and liver sinusoidal endothelial cells.

Upon liver injury, HSCs activate to become fibrogenic myofibroblasts, which, as noted, contribute most of the extracellular matrix proteins (collagens, proteoglycans, etc) of fibrotic and cirrhotic livers. Circulating fibroblasts or scar-associated fibroblasts, mesothelial fibroblasts, and portal myofibroblasts, while significant contributors in cholestatic and biliary liver diseases play minor roles in NASH related fibrosis. Recent single cell RNA sequencing analyses have uncovered an unprecedented level of clarity about HSCs and reveal an unexpected level of their heterogeneity across the liver lobule, with zonal distribution of HSCs [74], similar to the zonality of hepatocytes that has been recognized for decades [75]. Regions of HSC activation overlap with areas of injury, and while periportal HSCs proliferate upon pericentral injury, they may not transdifferentiate into collagen-producing HSCs [74]. HSCs also display unexpected heterogeneity with respect to contributions to hepatic regeneration [76]. They also support sinusoidal endothelial cells in a paracrine manner, further amplifying their contributions to liver injury and regeneration [77–79].

The availability of HSC lines and murine models to specifically target and modulate HSC genes has identified a plethora of pathways that are critical in HSC activation, which are reviewed in detail elsewhere [80–82].

Emerging antifibrotic strategies.

The best characterized profibrogenic and proliferative cytokines include transforming growth factor β (TGFβ), platelet derived growth factor (PDGF), vascular endothelial growth factor (VEGF) and connective tissue growth factor (CTGF). While broadly inhibiting these pathways has some appeal, these molecules have multiple homeostatic functions that require that they be antagonized specifically in HSCs. For example, TGFβ has strong tumor suppressor functions and thus systemic inhibition of TGFβ signaling will induce a tumor-promoting tissue micro-environment.

TGFβ or TGFβR inhibitors including antibodies, ligand traps and small molecules are under development to target cancer-associated fibroblasts and reduce immune evasion of cancer cells. The TβRI kinase inhibitor galunisertib is currently in phase 2 clinical trials for advanced HCC (NCT01373164) [83]. While the drug improved overall survival, it exhibited cardiac toxicity. Although TGFβ inhibitors have not been tested in preclinical models of NASH, cell type-specific delivery via novel chemically engineered techniques targeting stellate cells could salvage many compounds that otherwise have too many off-target effects.

HSC-specific targeting is currently a feature of a novel therapy to deliver a siRNA to heat shock protein 47 (ND L02-s0201/BMS986263) using liposomes conjugated to Vitamin A. By inhibiting this heat shock protein that regulates collagen folding, abnormally folded collagen accumulates in HSCs, leading to their death and clearance from injured tissue. Other approaches to target HSCs focus on their expression of the PDGFRb- or EGF-receptors [84], synaptophysin [68] or other specifically modified liposomes, lipid nanoparticles or conjugates [85].

Hippo pathway

The Hippo pathway is an emerging critical regulator of HSC activation. Hippo effector Yes-associated protein (YAP) is expressed in HSCs in healthy livers and is essential for cellular activation [86–88]. Hepatocytic Yap/Taz may lead to overexpression of downstream targets CTGF and CYR61 (also known as CCN1), which leads to HSC activation and promotes liver fibrosis [87] [89]. CYR61 is positively correlated with disease severity in patients with high grade NASH. The Taz downstream target, Indian hedgehog, is increased in NASH but not in plain steatosis [44], and deletion of IHH decreases fibrosis and HCC in a murine NASH model of DEN/High fat diet [90].

While still experimental, targeted delivery of therapies might offer alternative approaches. Integrin αvβ3, a surface receptor that is overexpressed in activated HSCs, has been used as a target for liposomes containing vismodegib, a selective hedgehog pathway inhibitor that binds to Smo and inhibits activation of downstream Hh target genes [91]. Encouragingly, it attenuated liver fibrosis in preclinical models of liver fibrosis [92].

Notch signaling

Studies indicate Notch activation in murine and clinical models of NASH [93]. Notch activation induces liver fibrosis in NASH via Sox9-dependent SPP1/Osteopontin release, which activates hepatic stellate cells driving fibrosis [93]. Providing proof of concept, a Notch inhibitor (Nicastrin antinsense oligonucleotide) reduces fibrosis in the NASH-diet fed mice. Of note, Notch signaling may be activated through upregulation of the Yap target genes Notch2 and Jag1, a Notch ligand ([94,95] although other factors might contribute as well [96].

Engineered T cells to remove fibrogenic HSCs.

Because the amount of fibrosis will determine clinical outcomes, their clearance by attack from specific T cells is appealing. Indeed, in a murine model of cardiac fibrosis, clearing fibrogenic cells that express fibroblast activation protein (FAP) using FAP-specific engineered CAR T (chimeric antigen receptor T) cells can reduce fibrosis and restore cardiac function [97]. Similarly in liver, targeting senescent HSCs by adoptive transfer of CAR T cells that are specific for the senescence marker urokinase-type plasminogen activator receptor (uPAR), can decrease hepatic fibrosis and injury [98]. These approaches, while currently restricted to mouse models, nonetheless offer exciting new possibilities for therapy of fibrosis in humans, provided that they selective clear only fibrogenic HSCs and do not undermine the cell’s roles in maintaining liver homeostasis.

Intestine-Derived Fibrogenic and Inflammatory Agents

The gut barrier is often compromised in NAFLD patients even before there is advanced liver disease, but its underlying mechanisms are not known. Increased intestinal permeability leads to translocation of bacteria or bacterial components including lipopolysaccharides, PAMPs, or DAMPs, which activate inflammasome and TLR signaling. Several mechanisms potentially link the gut microbiome to hepatic inflammatory and fibrogenic signaling, as reviewed below.

Proinflammatory bacterial products.

Translocation of bacteria or bacterial components are transported via the portal system to the liver where they can activate the hepatocyte inflammasome, resulting in release of proinflammatory cytokines including IL1β, IL18; these signals provoke hepatocyte apoptosis via caspase-1 activation. A Western diet can also directly trigger systemic inflammation that induces reprogramming of myeloid cells which is NLRP3-dependent [99]. Lipopolysaccharides activate TLR signaling and lead to inflammation. Trials with IMM-124E, purified hyperimmune bovine colostrum which binds LPS within the intestine, had no effect in NASH (NCT02316717).

Altered abundance of bacterial species in the gut microbiome may lead to production of bacterial products that enhance inflammation and promote NASH. NASH-associated gut bacteria may produce N,N,N-trimethyl-5-aminovaleric acid (TMAVA), which can impair lipid metabolism and exacerbate steatosis [100].

Endogenous bacterial alcohol production.

Microbial fermentation of carbohydrates in the gut may generate ethanol that elevates blood ethanol levels, as reported in pediatric and adult NASH patients [101–103]. High levels of ethanol producing K. pneumoniae have been described in a patient with NASH, leading to an ‘auto brewery syndrome’ that provokes liver injury through increased oxidative stress and production of acetaldehyde by metabolism of ethanol [104]. Weight loss or antibiotic treatments may significantly decrease endogenous ethanol production [104]. Changing the existing microbiome by fecal microbiome transplantation (FMT) has been studied in patients with metabolic syndrome, but not yet in patients with NASH [105]. Autologous FMT from a lean donor has led to improvements in insulin sensitivity in some studies [106,107], while it had no effect in others [108,109]. While still experimental, approaches to “drug the bug” – genetic alteration of bugs, or “bugs as drugs” – bugs as probiotics, might offer therapeutics in the future as the role of the microbiome in driving NASH pathogenesis is further clarified [110,111].

Short-chain fatty acids (SCFAs).

Not all bacterial products are detrimental. Short chain fatty acids (SCFA) including acetate, propionate and butyrate are produced by bacterial fermentation of dietary fibers and resistant starch that may have beneficial effects on energy metabolism, immune system, and the adipose tissue [110,112,113]. In the colon, butyrate has anti-inflammatory effects by decreasing pro-inflammatory cytokines, inhibition of NF-κB and expansion of Tregulatory cells [114] thus enhancing mucosal immunity [115] and strengthening the intestinal barrier by inducing mucins and tight junction proteins [116]. Microbiota-derived butyrate prevents dysbiosis by activating PPARγ in colonocytes and enhancing oxygen consumption via beta-oxidation, which in turn maintains low intraluminal oxygen and nitrate levels thus ultimately preventing expansion of pathogenic bacteria [117]. SCFAs also have systemic functions. They are transported across the gut epithelial barrier via monocarboxlate transporter 2 (MCT-1) and sodium-coupled monocarboxylate transporter 1 (SMCT-1) [118,119] into portal blood to reach the liver and circulation. SCFAs are signaling molecules with differing receptor affinity for orphan G protein-coupled receptor 41 (GPR41) and GPR43, which are expressed on adipocytes, immune cells and enteroendocrine cells and the bone marrow, among others [116,120,121]. SCFAs thus decrease fat accumulation by modulating insulin sensitivity and enhancing energy expenditure in the liver and adipose tissue [122,123]. Butyrate can alter gene expression by inhibiting histone deactetylase (HDAC), thereby enhancing the accessibility of the chromatin and altering the synthesis of cholesterol, which is the building block of bile acids [124]. In a dietary murine NASH model, butyrate reduced steatosis and inflammation by restoring expression of glucagon-receptor 1 (GLP-1R) which are otherwise downregulated in murine and human NASH [125]. Acetate exposure may also lead hepatic CXCR6+CD8 T cells to become auto aggressive via distinct mechanisms and ultimately mediate damage to hepatocytes in a dietary model of NASH [126]. Overall, SCFA may modulate inflammatory signaling in murine models of NASH [127,128] although studies assessing fecal SCFA in patients with NASH are inconsistent [129]. Higher fecal SCFAs could reflect increased production of SCFAs or impaired intestinal SCFA uptake. While SCFAs have overlapping actions, identification of its individual components and their effects could provide some clarity.

Clinical trials have sought to mimic the beneficial effects of SCFAs by their supplementation. In a randomized clinical trial of overweight patients, inulin-propionate ester given for 24 weeks resulted in significantly reduced weight gain, intra-abdominal adipose and hepatic lipid content compared to insulin-control, without adverse events [130]. Oral butyrate also improved glucose metabolism in lean patients but not those with metabolic syndrome [131]. Effects of SCFAs in obesity are inconsistent, which may be due to obesity-associated changes in glycolipid metabolism (see discussion in [116]). Although no adverse events have been reported following use of SCFAs, there is currently only one active clinical trial of butyrate in NAFLD (NCT04465032).

Bile acid signaling

Primary bile acids are synthesized by hepatocytes and secreted across the apical membrane into the biliary tract, from which they are released into the duodenum upon food intake. In the intestine, primary bile acids are metabolized to secondary bile acids by the gut microbiota. Apart from their role in digestion and resorption of fats, bile acids control bacterial overgrowth and signal through specialized receptors. Several intracellular nuclear hormone receptors (FXR, PXR, CAR) and one cell surface receptor TGR5 (GPBAR1), a transmembrane G-coupled bile acid receptor, function as bile acid receptors. Only FXR and TGR5 are primarily dedicated to bile acid signaling whereas PXR and CAR are more promiscuous and also bind to other xenobiotics. Bile acids are reabsorbed in the ileum and undergo enterohepatic circulation to return to the liver. All tissues associated with the enterohepatic circulation, including enterocytes, express FXR (hepatocytes, sinusoidal endothelial cells, stellate cells), but so do other organs including kidney, adrenal glands, adipose tissue, immune cells (macrophages), and to lower extent thyroid, lung and skin.

Upon bile acid uptake by enterocytes, FXR is activated and induces expression of fibroblast growth factor 19 (FGF19, Fgf15 in mice). In other cells and organs, both bile acid mediated FXR transcriptional activation, as well as its downstream effector FGF19, have common as well as distinct functions through their modulation of key genes in lipid and glucose metabolism, inflammatory signaling and bile acid synthesis. In hepatocytes, FXR inhibits CYP7A1 to limit bile acid synthesis, and decreases BA uptake transporters while upregulating basolateral BA exporters in obstructive cholestasis. FXR/FGF19 signaling decreases triglyceride, cholesterol, and glucose metabolism. It also reduces proinflammatory cytokine production (TNFα, IL-6, IFNγ) via modulation of NF-κB transcriptional activity. FXR may also have anti-tumorigenic effects, as FXR knockout mice spontaneously develop HCC when they age [132,133]. In turn, FXR agonists may reduce tumor burden in preclinical models of CCC and NASH dependent HCC [134–137].

FGF19 has direct mitogenic effects that could provoke HCCs. Therefore, engineered FGF19 agonists have been developed that preserve its beneficial metabolic effects while removing the mitogenic activity [138]. Aldafermin (NGM282) is one such engineered analog of FGF19, which does not have mitogenic activity in rodents, and has led to decreased serum bile acid levels, AST, hepatic steatosis and fibrosis in a small 24 week phase 2 clinical trial, although not all effects were statistically significant [139]; longer, larger trials will clarify Adafermin’s impact on NASH. The FXR agonist obeticholic acid has reduced fibrosis in a large phase 3 clinical trial of non-cirrhotic NASH patients (REGENERATE – NCT02548351). While FDA approval was anticipated, raised safety concerns has provoked a re-evaluation of the drug’s activity and safety this is currently underway. There is also a separate ongoing phase 3 trial of obeticholic acid in cirrhotic NASH patients (REVERSE – NCT03439254).

TGR5 / G protein bile acid receptor-1 (GPBAR1), which is expressed in hepatic non-parenchymal cells (KC, HSC, LSEC) and cholangiocytes but not hepatocytes, leads has anti-inflammatory effects, alleviates portal hypertension and features of NASH, and modulates liver microcirculation. TGR5 leads to release of GLP-1 from L-cells in the intestine. GLP1 enhances insulin release from beta-cells and may increase energy expenditure and satiety leading to weight loss [140].

The glucagon-like peptide-1 receptor (GLP-1R) agonist Semaglutide promotes insulin release from pancreatic beta-cells and was tested in a 72-week double-blind phase 2 trial in 320 patients with biopsy confirmed NASH and fibrosis [141]. Semaglutide significantly improved inflammatory activity of NASH, however there was no improvement in fibrosis. Gastrointestinal adverse events and weight loss were greatest in treatment groups receiving higher doses. Also, a smaller number of patients had attenuated progression of fibrosis compared to the placebo group (5% vs. 19%); thus, while it may not improve fibrosis per se Semaglutide might inhibit fibrosis progression. As the liver does not express GLP-1 receptors, its effects are likely mediated through extrahepatic mechanisms, with part of its efficacy attributable to weight loss induced by the drug.

For selected patients, bariatric surgery might be a suitable approach to improve NASH, even though liver disease is very rarely the primary indication for these procedures. Roux-en-Y bypass increases serum bile acid concentrations and may account for improved lipid and glucose metabolism [142]. Similarly, bariatric surgery increases bile flow and fasting serum bile acids and FGF19 in patients [143]. The mechanisms underlying bariatric surgery’s benefit towards insulin sensitivity remain unclear. A recent study indicated that bariatric surgery improves type 2 diabetes long term better than drug therapy [144], and improves long-term resolution of NASH and regression of fibrosis [145] while decreasing obesity-related cancers including NASH-HCC [146]. It remains controversial whether improvements following bariatric surgery are solely attributable to weight loss [147] or if processes apart from weight loss also contribute to metabolic improvements [142,143,148]. Changes in circulating bile acids or branched amino-acids, changes in the microbiome, higher production of the hormonea ghrelin or neuropeptide Y, and greater release of gastrointestinal hormones (glucagon, secretin, gastrointestinal peptide) have been implicated [149,150]. Clarification of underlying mechanisms that account for bariatric surgery’s benefits might offer novel strategic approaches in the treatment of NASH and other features of the metabolic syndrome.

Thryoid hormone receptor beta mimetics.

Thyromimetics have beneficial systemic effects such as increased energy expenditure and enhanced fat and cholesterol metabolism, which make them attractive prospects to treat NAFLD. Early studies indicated that thyroid hormone agonists decrease hyperlipidemia in murine models and reduce hepatic steatosis in a high-fat, high-cholesterol diet model and in the leptin-deficient ob/ob mouse, but they impair glucose sensitivity [151,152]. Other adverse effects associated with thryomimetics may include tachycardia, cardiac hypertrophy, muscle wasting and osteopenia [153].

Because of these unwanted activities of non-specific thyromimetics, Finan et al took a new approach by chemically engineering a hybrid glucagon-T3 molecule, taking advantage of the tissue-specific expression of thyroid hormone receptor β [154]. Thyroid hormone receptors have two isoforms, THR α and THR β. THR β is mainly expressed in liver and brain. Hybridization of a peptidase resistant glucagon with thyroid hormone T3 (glucagon-T3) allowed for targeting of glucagon and T3 specifically to the liver [154], synergizing the advantages of each compound while decreasing negative systemic effects. Glucagon enhances hepatic beta oxidation of fatty acids while inhibiting lipogenesis and increases energy expenditure [155]. Glucagon-T3 thus improves energy expenditure, decreases hepatic steatosis, and improves insulin resistance and hyperlipidemia. This compound improved NASH and atherosclerosis in preclinical models, thus targeting the leading cause of death, cardiovascular disease, in NASH patients. A phase 2 randomized clinical trial (RCT) of MGL-3196 led to a significant decrease in hepatic steatosis as assessed by MRI proton density fat fraction (MRI-PDFF) at 12 weeks [156]. An open-label extension study of MGL-3196 of 31 patients from the initial study further demonstrated improved dyslipidemia and hepatic steatosis. Fibrosis assessed by non-invasive means has demonstrated decreased liver stiffness by elastography and N-terminal III collagen pro-peptide (Pro-C3) [157]. A phase 3 clinical trial NASH study is ongoing (MAESTRO-NASH, NCT03900429) and will assess NASH activity (NAS), LDL-C levels and fibrosis stage by liver biopsy. Another thyromimetic VK2809 is currently in phase 2 clinical trial for patients with biopsy proven NASH (NCT04173065).

Summary

NASH is a multifactorial disease that is part of a systemic metabolic disorder. It is unknown how much each of the many defects associated with the disease contribute to its progression. Thus, better characterization of disease drivers in each patient is necessary to contemplate a personalized approach for therapy. Stratification of patients in clinical trials according to these individual disease drivers might enrich for patients who will benefit from a specific intervention, thereby shortening trials and yielding clearer results.

Similarly, due to its complex etiology, targeting one pathogenic pathway among the many that cause NASH might not be sufficient, and combination therapies may be necessary as even those drugs that seem effective only benefit a minority of patients. Because weight loss and lifestyle changes are difficult to achieve, distinguishing between those who are obese but metabolic healthy from obese, metabolic unhealthy individuals, which has been attributed to adipose tissue dysfunction [9,158]), may help enrich trials for patients who will need pharmacotherapy [158,159].

A growing number of experimental drugs are in the NASH drug development pipeline (see recent comprehensive overview in [160]). While there have been some setbacks with advanced clinical trials failing recently, we remain optimistic that effective therapies will be established, to be followed by incremental successes that further refine, enhance and perhaps individualize therapies. Until then, primary prevention by promoting healthier lifestyle and diets in both children and adults alike will surely promote a sustained improvement in public health.

Figure 1. Pathogenesis of NAFLD – inflammatory and fibrogenic signaling events.

NASH is the hepatic manifestation of the metabolic syndrome. Adipose tissue inflammation and insulin resistance leads to hepatic overload with fatty acids and glucose driving hepatic steatosis, ER stress and activation of the unfolded protein response leading to inflammasome activation and cell death. Adipokines and gut microbiome derived bacterial products modulate hepatic inflammation and nuclear receptor signaling (FXR, LXR, PXR and Vitamin D receptor) but may also directly activate hepatic stellate cells. Hepatocyte injury signals lead to activation of hepatic stellate cells and inflammatory cells converging on hepatic fibrogenesis. Emerging pathways and intracellular signaling molecules may enable novel strategies for therapeutic intervention as are new techniques such as engineered T cells to kill activated stellate cells or HSC-targeted chemically engineered antifibrotic drugs. HSC, hepatic stellate cells. Modified from [161]

Table 1.

Overview of selected drugs and compounds tested in clinical trials for NASH. Modified from [162,163].

Agent (drug)	Trial name	Mechanism	Inclusion criteria	Primary outcome	NCT number	Status
Phase 3
Aramchol	ARMOR	SCD-1 inhibitor	NASH stage 1–3 fibrosis	NASH resolution, fibrosis improvement	NCT04104321	Active. No results posted yet. Estimated completion June 2027
Belapectin (GR-MD-02)	NASH-RX	FGF19 analogue	NASH cirrhosis, evidence of portal hypertension	New esophageal varices	NCT04365868	Active. No results posted yet. Estimated completion December 2023
Cenicriviroc	Aurora	CCR2/CCR5 antagonist	NASH + F2/F3 fibrosis	Improvement in fibrosis without worsening NASH	NCT03028740	Completed (3/9/2021). No results posted yet.
Dapaglifloxin (DEAN)	DEAN	SGLT2 inhibitor	NASH and T2DM (HbA1c <9.5%)	Histological improvement, NASH resolution	NCT03723252	Active. No results posted. Estimated completion November 2021
Elafibranor	RESOLVE-IT	PPAR a/δ ligand	NASH + stage 1–3 fibrosis	Resolution of nash without worsening of fibrosis or Improvement in fibrosis without worsening NASH	NCT02704403	Failed. No significant improvement in primary or secondary endpoints in interim analysis.
Obeticholic acid	REVERSE	FXR ligand	NASH + stage 4 fibrosis	Improvement in fibrosis without worsening of NASH.	NCT03439254	Active, not recruiting (last update 04/26/2021)
Obeticholic acid	REGENERATE	FXR ligand	NASH with stage 2 or 3 fibrosis or stage 1 fibrosis with risk for progression	Resolution of nash without worsening of fibrosis or Improvement in fibrosis without worsening NASH	NCT02548351	Active, not recruiting (last update 04/26/2021)
Oltipraz		Liver X receptor α inhibitor	NAFLD (liver fat>20% on MRS)	Improvement in steatosis	NCT04142749	Active. No results posted. Estimated completion October 2021
Resmeritom (MGL3196)	MAESTRO-NASH	THR-β agonist	NASH stage 2–3 fibrosis.	NASH resolution, long term outcome. Improvement in fibrosis	NCT03900429	Active. No results posted yet. Estimated completion March 2024
Selonsertib	STELLAR-3	ASK-1 inhibitor	NASH on biopsy with stage 3 fibrosis	Resolution of NASH without worsening of fibrosis or Improvement in fibrosis without worsening NASH	NCT03053050	Failed [19]
Selonsertib	STELLAR-4	ASK-1 inhibitor	NASH on biopsy with stage 4 fibrosis but no ascites, HE or variceal bleeding.	Resolution of NASH without worsening of fibrosis or Improvement in fibrosis without worsening NASH	NCT03053063	Failed [19]
Agent	Trial name	Mechanism	Inclusion criteria	Primary outcome	NCT number	Status
Phase 2
Aldafermin (NGM282)		FGF19 analog	NASH on biopsy	Improvement in liver fat by MRI	NCT02443116	Significant reduction in liver fat. OLE study: improved NAS score and fibrosis after 12 weeks. [164,165]
Aramchol	ARREST	SCD-1 inhibitor	NASH on biopsy	Improvement in steatosis by NMRS	NCT02279524	Completed. No results posted.
Cilofexor		FXR agonist	NASH without cirrhosis	Overall safety	NCT02854605	Significant improvement in steatosis, liver biochemistry, serum bile acids. [166]
Emricasan	ENCORE-NF	Caspase inhibitor	NASH on biopsy with stage2 or 3 fibrosis or stage 1 fibrosis with risks for progression	Improvement in fibrosis without worsening NASH	NCT02686762	Failed. No significant improvement in endpoints. Possibly more fibrosis and hepatocyte ballooning. [20]
Emricasan	ENCORE-PH	Caspase inhibitor	NASH cirrhosis and HVPG >12 mmHg	Improvement in HVPG	NCT02960204	Failed. No significant improvement in endpoints. [21]
Emricasan	ENCORE-LF	Caspase inhibitor	NASH cirrhosis with history of variceal bleeding and/or noderate or severe ascites	Event free survival	NCT03205345	Completed. No results posted.
Firsocostat (GS-0976)		ACC inhibitor	NASH Fibrosis F1–3	Safety profile / adverse events	NCT02856555	Reduction of steatosis, selected markers of fibrosis, liver biochemistry [31]
GR-MD-02	NASH-CX	Galectin 3 inhibitor	NASH cirrhosis and HVPG >6 mmHg	Improvement in HVPG	NCT02462967	Failed. No significant reduction in fibrosis or HVPG. [167]
IMM-124E		Anti-LPS	NASH on biopsy	Improvement in liver fat MRI	NCT02316717	Failed. No significant improvement in clinical endpoints.
Lanifibranor	NATIVE	Pan-PPAR agonist	NASH without cirrhosis	Improvement in SAF activity score without worsening of CRN score (Fibrosis)	NCT03008070 [168]	SAF score significantly improved in higher dosage acc. to posted results on clinicaltrials.gov (4/21/2021).
MSDC 0602K	EMMINENCE	mTOT inhibitor	NASH on biopsy with stage 1–3 fibrosis	NAS reduction > points without worsening fibrosis	NCT02784444	Failed. No significant differences in primary or secondary endpoints [169]
Pegbelfermin	FALCON1	Pegylated FGF21	NASH fibrosis (F3)	Improvement NASH without worsening of fibrosis. Fibrosis improvement without worsening of NASH	NCT03486899	Active. No results posted. Estimated completion December 2021.
Pegbelfermin	FALCON2	Pegylated FGF21	NASH cirrhosis	Improvement in fibrosis without worsening of NASH	NCT03486912	Active. No results posted. Estimated completion October 2021.
Clesacostat (PF-05221304)		ACC inhibitor	NAFLD and PDFF >8%	Improvement in liver fat by MRI	NCT03248882	Not active (10/2020)
SAR425899	RESTORE	GLP-1 and glucagon receptor agonist	NASH on biopsy and type 2 diabetes	Resolution of ballooning without worsening of fibrosis	NCT03437720	Withdrawn
Semaglutide		GLP-1 mimetic	NASH on biopsy with stage 1–3 fibrosis	Resolution of NASH without worsening of fibrosis	NCT02970942	Significant improved NASH, but no improvement in fibrosis. Phase 3 study planned for 2021[141]
Tropifexor	FLIGHT-FXR	FXR agonist	NASH on biopsy with stage 1–3 fibrosis or clinical risks for NASH	Improvement in ALT and liver fat by MRI	NCT02855164	Completed (04/06/2021). No results posted.
Tropifexor + cenicriviroc	TANDEM	FXR agonist + CCR2/5 antagonist	NASH on biopsy stage 2–3	Adverse events. Resolution of NASH, improvment in fibrosis	NCT03517540	Completed (10/13/2020). No results posted.[170]
VK2809	VOYAGE	THR-b agonist	NASH fibrosis (F1–3), NAS score >4, MRI-PDFF >8%	Improvement in steatosis by MRI-PDFF	NCT04173065	Active. Estimated completion November 2021.
Volixibat (SHP626)		ASBT inhibitor	NASH on biopsy with stage 0–3 fibrosis, >5% liver fat by MRI	NASH reduction >2 points without worsening of fibrosis	NCT02787304	Interim analysis: no significant improvement. Study terminated. [171]

Abbreviations:

NAFLD

non-alcoholic fatty liver

NASH

nonalcoholic steatohepatitis

HCC

hepatocellular carcinoma

HSC

hepatic stellate cell