Detlef Schuppan, Molecular and Translational Medicine, First Department of Medicine, Johannes Gutenberg University, Langenbeckstr.1, 55131 Mainz, Germany. Email: email@example.com
Non-alcoholic fatty liver disease (NAFLD) refers to a disease spectrum, ranging from mere hepatic steatosis to hepatic necroinflammation (NASH, non-alcoholic steatohepatitis). NASH often leads to fibrosis, which can progress to cirrhosis with a high risk of liver failure and hepatocellular carcinoma. The course of NAFLD is highly variable, and only a minority of patients (2–3%) progress to end-stage liver disease. However, due to a dramatic increase of the risk factors for NAFLD, that is obesity and insulin resistance/type 2 diabetes, that affect 15–30% and 7–15% of subjects, in most industrialized countries, respectively, NAFLD has become the most frequent liver disease and is even considered a pace setter of the metabolic syndrome. Sedentary lifestyle, modern Western nutrition, and genetic predispositions have been identified as major causes of NAFLD. These lead to liver injury via insulin resistance and an excess of free fatty acids in hepatocytes, resulting in oxidant stress and lipotoxicity that promote the activation of intracellular stress kinases and apoptosis or necroapoptosis (NASH). The damaged hepatocytes directly trigger inflammation and fibrogenesis, but can also lead to the emergence of fibrogenic progenitor cells. Moreover, NASH is linked to inflammation in peripheral adipose tissues that involves mainly macrophages and humoral factors, such as adipokines and cytokines. The most efficient treatment is by weight loss and exercise, but (adjunctive) pharmacological strategies are urgently needed. Here, we highlight the aspects of NAFLD epidemiology and pathophysiology that are beginning to lead to novel pharmacological approaches to address this growing health-care challenge.
The face of clinical hepatology is currently experiencing a major shift: away from (increasingly well-treatable viral) infections as prominent etiologies to non-alcoholic fatty liver disease (NAFLD). NAFLD consists of a disease spectrum that is associated and overlapping with obesity, dyslipidemia, cardiovascular disease, and insulin resistance/type 2 diabetes, that is features of the metabolic syndrome, a major cause of morbidity in developed and developing societies (Fig. 1). Ninety percent of NAFLD patients exhibit at least one of these risk factors, and one third exhibits three or more (Table 1). The exact numbers of patients with NAFLD can only be estimated due to the lack of reliable non-invasive markers and the need for histological definition of disease stage. In a recent study from the United States involving 400 volunteers at an army medical center with a mean age of 54.6 years and 45% obese subjects, the reported prevalence of NAFLD was 46%. Non-alcoholic steatohepatitis (NASH), that is histological necroinflammation, was diagnosed in 12% and twice as frequently in Hispanics versus Caucasians. Patients with NASH mostly (80%) exhibited a body mass index (BMI) > 30, had a mean alanine aminotransferase (ALT) of 50 U/L, and a higher quantitative insulin-sensitivity check index. Already in 2004, the Dallas Heart study that examined 2287 adults in a population-based setting showed a 31% prevalence of NAFLD, as confirmed by magnetic resonance imaging (MRI) in 31%. In this Texan cohort, the average age was 45 years, and the highest prevalence of hepatic steatosis was observed in the Hispanic cohort despite, on average, being 5 years younger than non-Hispanics. Likewise, in US populations with a BMI below 25, Hispanic origin and hypertension were significantly correlated with the presence of NAFLD on ultrasound. In Asian populations, despite a historically lower prevalence of metabolic risk factors and NAFLD, the incidence of NASH has increased dramatically, and the prevalence of the disease is growing rapidly.[6, 7] Additionally, a marked difference between rural and urban areas exists, indicating that lifestyle and education are contributing to NAFLD and NASH in Asia. However, the underlying mechanisms appear too complex. Even in a non-obese, non-affluent, rural population in India (n = 1991), with an average age of 35.5 years and a mean BMI of 19.6, the prevalence of NAFLD was 8.7%. In this study, the group with hepatic steatosis as determined by ultrasound and computed tomography scan exhibited a mean age of 39 and a mean BMI of 23, well below that of similar Western populations, perhaps due to a higher predisposition to accumulate visceral fat. Therefore, with the increasing prevalence of environmental risk factors of NAFLD in Asia recently and a comparable genetic predisposition, NAFLD is likely soon to rise to similar prevalence in most Asian countries as in the West despite a lower frequency of adiposity. In high-risk Western populations with diabetes and obesity, the prevalence of NAFLD can reach up to 75%,[10, 11] but the overall incidence of NASH is difficult to assess due to reliance on biopsy, especially in follow-up. A study from Hong Kong derived from a hospital cohort reported histological progression in 58% and fibrosis progression in 28% during a 3-year follow-up of patients at risk but with a low NAFLD activity score of < 3. In the absence of fibrosis or inflammation, the course of hepatic steatosis appears to be more benign. Thus, in a cohort of 144 patients with alcoholic and non-alcoholic fatty liver, regression as determined by ultrasound was observed in nearly every second case.[13, 14]
Table 1. Criteria for clinical diagnosis of the metabolic syndrome
Categorical cut points
Presence of any three risk factors constitutes the diagnosis.
HDL-c, high-density lipoprotein cholesterol.
Elevated waist circumference
≥ 94 cm men
≥ 80 cm women
Elevated triglycerides (or drug treatment for elevated triglycerides)
≥ 150 mg/dL (1.7 mmol/L)
Reduced HDL-C (or drug treatment for reduced HDL-C)
< 40 mg/dL (1.0 mmol/L) men
< 50 mg/dL (1.3 mmol/L) women
Elevated blood pressure (or antihypertensive drug treatment)
Systolic ≥ 130 and/or diastolic ≥ 85 mmHg
Elevated fasting glucose (or drug treatment)
≥ 100 mg/dL
Apart from a waxing and waning course of disease activity, which may in part depend on (minor) lifestyle changes, the factors that determine disease progression in individual patients remain poorly defined. A meta-analysis on 10 studies comprising 221 patients found that over a mean time of 5.3 years, 21% of patients improved, 41% had unchanged liver histology, and 38% showed fibrosis progression by at least one histological stage (out of four stages). The strongest predictor of NASH progression was the degree of necroinflammation on initial biopsy. Sedentary lifestyle and overnutrition feed into the genetic predisposition of the “thrifty phenotype” that is partly determined by race, gender, and epigenetic changes, as reflected by a positive family history of NAFLD and the metabolic syndrome.[16-18] Notably, advanced fibrosis is prominent in patients older than 45 years, and liver-related mortality is increased approximately ninefold in patients suffering from NASH. Moreover, NASH is a key contributor to mortality from cardiovascular disease independent of traditional risk factors, and advanced stages of NAFLD predict carotid intima-media thickness and carotid plaques. Exploration of the underlying pathomechanism has led to a concept in which a pro-inflammatory state, including increased levels of pro-inflammatory cytokines, decreased protective adipokines, enhanced oxidant stress, emanating from the expanding and inflamed adipose tissue of obese patients, induces the release of atherogenic and prothrombotic factors from the liver. Certain studies even suggest that the liver is the prime driver of adipose inflammation and atherogenesis.
The incidence of hepatocellular carcinoma (HCC) in NAFLD remains controversial, since the association of NASH with cryp togenic cirrhosis as cause of HCC is difficult to prove. Patients with NASH can develop HCC even in the absence of cirrhosis, influenced by risk factors that contribute to the development of HCC. A systematic review of epidemiology studies including a total of 35 cohort, case control, and cross-sectional studies, as well as case reports, reported a cumulative HCC mortality rate during a follow-up of up to 20 years in non-cirrhotic NASH below 3%. In cirrhotic NASH, the cumulative incidence ranged from 2.4% to 12.8% in 3–12 years. Overall, this is considerably lower compared with virus-associated HCCs. In Hepatitis B surface antigen-positive patients with compensated cirrhosis, the 5-year cumulative HCC risk reaches 15% in endemic areas.
Pathogenesis of NASH: Consequence of expanding labile fat stores
Since only a subset of patients with NAFLD progresses to NASH, lifestyle and genetic predisposition remains the best defined disease determinants. Recently, high dietary cholesterol, an activator of liver x receptor,[25, 26] was shown to negatively affect the balance between storage and oxidation of fatty acids. Thus, with excessive supply to the liver, either from de novo lipogenesis or from excess dietary fat, fatty acids are processed to non-triglyceride metabolites, including diacylglycerol (DG) and lysophosphatidyl choline, that drive lipotoxic injury of hepatocytes (Fig. 2).
The type of dietary fat contributes to the development of NASH, as shown in mice on a diet enriched in trans-saturated fats. Moreover, fructose, which depletes intracellular ATP, is transformed to lipid in the absence of insulin, thus increasing fat deposits and contributing to NAFLD and NASH, as also evidenced by the strong association of type 2 diabetes and NASH in individuals consuming high-fructose-containing soft drinks. The depletion of hepatic ATP favors mitochondrial dysfunction, generation of reactive oxygen species and the resultant inflammation, and enhances endoplasmic reticulum stress, with subsequent activation of the stress-related Jun N-terminal kinase (JNK) which promotes hepatocyte apoptosis, the hallmark of NASH. The amount of lipotoxic metabolites is influenced by peripheral lipolysis, hepatic de novo lipogenesis, and the oxidative disposal of triglycerides involving lysosomes and β-oxidation. Toxic lipid metabolites, rather than the amount of hepatic triglycerides, lead to the activation of stress kinases and activation of cell death receptor signaling pathways, and trigger organelle dysfunction, leading to progressive cell injury (Fig. 3). Other factors are activation of toll-like receptor 4 (TLR4) by intestinal bacterial lipopolysaccharide[33-36] and other pro-inflammatory signals produced by a pathological microbiota, which in most studies is dominated by firmicutes versus proteobacteriaceae and enterobacteriaceae, and favors a more effective energy harvest.[37-39] Extrahepatic sources of inflammation involve increased permeability of the gut and translocation of bacterial endotoxins, which fuel apoptotic injury and fibrogenesis. The transmission of an unfavorable gut microbiome in mice resulted in the development of NASH, while transplantation of a gut microbiome from lean patients to patients with obesity and type 2 diabetes improved insulin resistance.
Differences in the development of NASH have recently been linked to genetic susceptibility. The single nucleotide polymorphism (rs738409) in the human patatin-like phospholipase domain containing 3 gene (PNPLA3 or adiponutrin) results in a I148M variant and is a strong predictor of steatosis, inflammation, and fibrosis across different populations, being independent of body mass, insulin resistance, or serum lipid levels. The expression of PNPLA3 is regulated by nutrition: fasting inhibits, and high-carbohydrate diet feeding increases, PNPLA3 expression. In humans, PNPLA3 is predominantly expressed in liver, while in mice the strongest expression is observed in adipose tissue. PNPLA3 possesses triglyceride hydrolase and DG transacylase activity, and converts lysophosphatidic to phosphatidic acid form. By modulating lipid intermediates, dysfunctional PNPLA3 promotes the accumulation of lipotoxic substrates, which lead to lipoapoptosis and inflammation.
Treatment of NASH: New targets on the horizon
Assessment and biomarkers
The increasing prevalence of NASH has led to a great demand for medical therapy. However, no pharmacological therapy has been proven effective in long-term use. A major limitation in designing clinical trials in NASH has been the lack of appropriate non-invasive diagnostic tools that can be applied to stage and predict the course of the disease. Necroinflammation, hepatocellular ballooning, and the degree of fibrosis strongly predict the risk of disease progression, and are based on histology that itself confers high sampling variability. Risk scores that have been developed, including the NASH test or the NAFLD fibrosis score, are limited by their inaccuracy. Therefore, both for patient monitoring and clinical drug development, there is a yet unmet need for novel biomarkers that exactly differentiate disease stages. A novel class of diagnostic markers are circulating membrane microparticles that are released from activated immune cells. Thus, patients with histological NAFLD and NASH show a characteristic increase in macrophage and invariant natural killer T (iNKT) cell-derived microparticles, cells that are unique to NASH pathogenesis.
The primary approach to treat NAFLD involves elimination of the underlying risk factors. Maintaining a weight loss of 5–10% significantly improves histological severity, but frequently occurring subsequent weight gain leads to the recurrence of NASH. Even moderate physical exercise, such as treadmill walking, improves markers of apoptosis and insulin sensitivity in NAFLD. The dietary composition is also of great importance. A 2% increment in energy intake from trans fats resulted in a 0.77-cm waist gain over 9 years, and reduction of harmful trans fats improved histological features in a mouse model despite persistent obesity. Even if all these measure are effective, (adjunctive) pharmacological therapies will still be required for the majority of patients with NASH.
Inhibitors of caspases and inflammation
The severity of NASH and the risk of progression correlate with hepatocyte injury that often includes necroapoptosis, and the associated inflammation. Necroapoptosis involves cell death signaling pathways, which lead to the activation of caspases, cellular proteases that degrade structural proteins required for the cell survival. Inhibition of caspases has been proposed as a therapeutic approach in inflammation-associated disease. In mice on a methionine-choline-deficient (MCD) diet, a model of steatohepatitis that lacks features of the metabolic syndrome but displays features of hepatocyte lipoapoptosis characteristic of NASH, hepatocyte-specific deletion of caspase 8 ameliorated hepatic inflammation, oxidative stress, and liver injury. In mice with a mutation of the leptin receptor (db/db) and on the MCD diet, hepatocyte apoptosis and inflammation were suppressed by the pan-caspase inhibitor VX-166.
In a double-blind, randomized phase II study of 124 patients with NASH, GS-9450, an inhibitor of caspases 1, 8, and 9, reduced serum ALT and cytokeratin-18 fragments at 4 weeks of treatment. However, the compound was later withdrawn due to safety concerns in patients with chronic hepatitis C (http://www.gilead.com/pr_1414682). However, dampening necroapoptosis in active NASH remains an attractive target to reduce the amount of cell death and subsequent fibrosis, and prevent disease progression. On the other hand, increasing cellular viability during inflammation also raises concerns of malignancy, and antiapoptotic agents likely need to be given in a small therapeutic window.
Adenosine is a physiological modulator of tissue responses to injury, and regulates cell survival, immuno-inflammatory reactions, and tissue repair involving four adenosine receptors (A1, A2A, A2B, and A3) in an auto and paracrine fashion. In rats that are on the MCD diet, activation of the adenosine A2A receptor, which is expressed on inflammatory cells and hepatic stellate cells (HSCs), with the agonist CGS21680 reduced inflammatory cell activation, the subsequent JNK cascade in hepatocytes, and fibrosis, without affecting steatosis.[63, 64] Interestingly, differential effects of A1R and A2BR have been observed in the context of alcohol-induced hepatic lipogenesis. While expression of genes involved in fatty acid synthesis was prevented by blockade of A1R, decreased expression of genes involved in fatty acid metabolism was prevented by blockade of A2BR.[63, 65] Thus, depending on the cells and cellular receptors, adenosine can induce contrasting effects cellular injury, fibrosis, and steatosis. The complexity of adenosine signaling requires further testing of specific receptor agonists and antagonists.
Dipeptidylpeptidase-IV (DPP-IV) and glucagon-like peptide-1 (Glp-1)
The regulation of energy balance in peripheral tissues (including muscle, adipose, and hepatic tissue) involves the central and enteric nervous system, and is influenced by humoral factors that control appetite and physical activity. Signaling through satiety-inducing hormones and endocanabinoids is deregulated in NASH, contributing to adipose tissues expansion and hepatic inflammation. Glp-1 and gastric inhibitory polypeptide belong to the class of incretins, which are released from enterocytes in response to nutrient uptake. Especially, Glp-1 regulates postprandial insulin release, inhibits glycolytic glucagon, and suppresses appetite. Locally and in the blood, rapid degradation of Glp-1 is mediated by the membrane-anchored enzyme DPP-IV, which is expressed prominently on epithelia, endothelial cells, and lymphocytes. While indirect and direct Glp-1 agonists have been introduced in the treatment of diabetes, their potential in NASH is less clear. DPP-IV activity is increased in NASH, and the DDP-VI inhibitors, vildagliptin and linagliptin, improved hepatic steatosis, adipose tissue inflammation, and insulin sensitivity in obese and diabetic Zucker rats and in a high-fat diet model in mice.[70, 71]
The more protease-resistant, direct-acting Glp-1 agonists, exenatide and liraglutide, showed similar, if not better, therapeutic efficacy. Liraglutide corrected impaired fatty acid beta-oxidation in a rodent model with high dietary trans fats and fructose-enriched drinking water. In a high-fat model using wild-type C57Bl6 and ob/ob mice, the exenatide analogue AC3174 attenuated weight gain and mitigated elevations of ALT and hepatic triglycerides. Exenatide also reduced ER stress-related hepatocyte cell death and increased protective macroautophagy in response to treatment with saturated and unsaturated fatty acids. Thus, enhancement of incretin signaling showed modest to considerable improvements in vitro and in animal models of NASH. Since these drugs have shown safety in patients with type 2 diabetes, clinical studies in patents with NAFLD are warranted.
Peroxisome proliferator-activated receptor (PPAR) alpha and delta
PPARs belong to the class of nuclear receptors that regulate expression of genes involved in lipid and glucose homeostasis but also modulate (hepatic) inflammation and fibrosis. The different PPAR isoforms have been implicated in metabolic protection, such as improvement of fatty acid oxidation and breakdown (PPARα), storage and deposition of (harmful inert) triglycerides, anti-inflammatory signaling and improved insulin sensitivity (PPARγ), and improvement of metabolic performance by altering dyslipidemia and increasing fat oxidation in the musculature (PPARδ). In the NASH models of a high-fat or MCD diet, the pan-PPAR agonist bezafibrate, as well as a PPARδ/β (GW5051516) and a PPARα agonist (Wy14 643), improved hepatic steatosis and inflammation.[75, 76] Moreover, hepatocyte insulin resistance from inflammatory cytokines was improved by GW501516. The thiazolinediones rosiglitazone and pioglitazone, both PPARγ and PPARγ>α agonists, respectively, were successfully employed in clinical trials to improve histological features of NASH. However, an effect of fibrosis progression could not be demonstrated, and their use is limited by side effects, including significant peripheral weight gain and potentially worsening cardiovascular disease.[78, 79] In two randomized, placebo-controlled trials with a total of 53 patients in the verum groups, the combined PPARα/δ agonist GFT505 improved dyslipidemia and insulin resistance. Currently, larger studies to evaluate its potential efficacy in patients with NASH are ongoing (ClinicalTrials.gov Identifier: NCT01694849).
Inhibition of SGLT-2 in the kidney significantly improves hyperglycemia control by blocking glucose reabsorption, and thus increasing glucosuria. This improves insulin sensitivity and should decrease adipose tissue (and liver) inflammation, for example via improved chemokine, cytokine, incretin, and adipocytokine profiles. In KK-A(y) mice exhibiting spontaneous diabetes and fatty liver, treatment with sergliflozin etabonate improved glycemic control and hepatic steatosis. Future studies on the role of SGLT-2 inhibitors in NASH are warranted.
Peripheral cannabinoid 1 receptor (CB1R) agonists
Antagonists to the CB1R were successfully employed to improve the metabolic phenotype in NASH. The peripherally and centrally active rimonabant prevented diet-induced fatty liver and obesity, and decreased de novo fatty acid synthesis and the fibrogenic activation of hepatic stellate cells in models of acute and chronic liver injury.[82, 83] However, the considerable neuro-psychiatric side effects, including suicidal depression, have led to the withdrawal of rimonabant in 2008. Now, efforts are made to develop and investigate predominantly peripherally acting CB1R blockers for the treatment of NASH and obesity, since peripheral CB1R antagonism reduced hepatic lipogenesis and decreased peripheral lipolysis, promoted a favorable anti-inflammatory cytokine and adipokine profile, and led to reduced food intake with an associated reduction of body weight.
Farnesoid X receptor (FXR) agonists
Bile acids are secreted in response to food uptake, undergo enterohepatic circulation, and act as endogenous ligand to a class of nuclear hormone receptors that function as ligand-activated transcription factors to regulate numerous physiological processes. These receptors include the FXR, pregnane X receptor, and the constitutive androstane receptor. FXR, a receptor for hydrophobic bile acids, has been most widely studied. Beyond the regulation of bile acid synthesis, FXR improves insulin sensitivity and glucose uptake in adipose tissue, and the liver and the skeletal muscle, by regulating metabolic genes such as PEPCK, G6Pase, and FBP1. Moreover, FXR suppresses pro-inflammatory genes like interferon γ, tumor necrosis factor-α, and interleukin-6 by affecting NFkappaB transcriptional activity.[86, 87] However, this broad spectrum is likely to result in adverse side effects, and selective FXR agonists are required to mainly alter gene expression relevant to NASH and insulin resistance. In the MCD model of steatohepatitis, WAY-362450, a synthetic FXR ligand, protected against hepatic inflammation and fibrosis without inhibiting hepatic triglyceride accumulation. Currently, obeticholic acid, a semi-synthetic bile acid derivative, is tested in patients with biopsy-proven NASH (ClinicalTrials.gov Identifier: NCT01265498). An unwanted side effect of obeticholic acid is exacerbation of itching.
Probiotics: Altering the gut microbiome
A pro-inflammatory intestinal microbiome has been observed in mice and patients with NASH.[37-39, 41] In a model of genetic dyslipidemia using ApoE-deficient mice, supplementation of the probiotic VSL#3 that contains different lactobacilli and bifidobacteria improved insulin signaling in hepatocytes and ameliorated adipose tissue inflammation, and the supplementation of lactobacillus casei shirota protected mice from increased activation of TLR4 and hepatic steatosis induced by a high-fructose diet. In an open-label pilot study in 20 patients with biopsy-proven NASH, supplementation of a probiotic containing lactobacilli and bifidobacteria over 6 months improved hepatic steatosis, as determined by MRI and serum transaminases. Together with the human randomized controlled study on fecal transplantation of a healthy microbiota in patients with insulin resistance, these recent reports support the role of microbiota in the pathogenesis of insulin resistance and NASH, partly by reducing bacterial inflammatory triggers and nutrient extractions and modification. It also hints to a role of prebiotics, that is nutrients that favor the growth of certain bacterial species, that may likely play in the treatment of obesity and NASH.
NAFLD has become a global challenge to our health-care systems. Changes in lifestyle and nutrition have put large parts of the population at risk of developing NASH, cirrhosis, and liver cancer. In contrast to other liver diseases with emerging therapeutic options, and despite the benefit of lifestyle changes, NAFLD will remain a great health problem necessitating (adjunctive) pharmacological therapies. Moreover, given the unpredictable course of this common disease, improved non-invasive biomarkers are urgently needed to better assess NAFLD/NASH activity and fibrosis, and to speed up drug development. Currently, pharmacological interventions aiming to improve insulin sensitivity without producing side effects, such as peripheral weight gain, and to reduce hepatocellular injury by reducing lipotoxicity are being developed. Moreover, their combination with prebiotics and probiotics that favorably modulate nutrient extraction/metabolism and the intestinal microbiome is promising.
Conflict of interests
DS and JMS declare no conflicting interests. DS received funding from the NIH, European Union, the State of Rhino-Palatinate, the German Research Foundation, the German Ministry of Education and Research, and Boehringer-Ingelheim. JMS receives funding from the DFG and intramural funds of the University Medical Center Mainz.