Non-alcoholic fatty liver disease--a chronic disease of the 21^st century

Abstract

Non-alcoholic fatty liver disease (NAFLD) comprises a spectrum of metabolic states ranging from simple steatosis to inflammation with associated fibrosis to cirrhosis. Though accumulation of hepatic fat is not associated with a significant increase in mortality rates, hepatic inflammation is, as this augments the risk of terminal liver disease, i.e., cirrhosis, hepatic decompensation (liver failure) and/or hepatocellular carcinoma. Disease progression is usually slow, over a decade or more and, for the most part, remains asymptomatic. Recent estimates suggest that the global prevalence of NAFLD is high, about one in four. In most cases, NAFLD overlaps with overweight, obesity, cardiovascular disease and the metabolic syndrome with numerous contributing parameters including a dysregulation of adipose tissue, insulin resistance, type 2 diabetes, changes in the gut microbiome, neuronal and hormonal dysregulation and metabolic stress. NAFLD is diagnosed incidentally, despite its high prevalence. Non-invasive imaging techniques have emerged, making it possible to determine degree of steatosis as well asfibrosis. Despite this, the benefit of routine diagnostics remains uncertain. A better understanding of the (molecular) pathogenesis of NAFLD is needed combined with long-term studies where benefits of treatment can be assessed to determine cost-benefit ratios. This review summarizes the current state of knowledge and possible areas of treatment.

Keywords

non-alcoholic fatty liver disease / pathogenesis / molecular mechanism / diagnostics / biomarker

Introduction

Overweight and obesity are hallmarks of the metabolic syndrome and affect a large proportion of the general population. Of associated disease states, type 2 diabetes (T2D) and cardiovascular disease (CVD) show a marked increase in mortality rates. Non-alcoholic fatty liver disease (NAFLD), fat accumulation in the liver caused by factors other than alcohol, is a common manifestation of the metabolic syndrome yielding hypertriglyceridemia and abnormal hepatic fat accumulation presented either as simple steatosis (non-alcoholic fatty liver (NAFL)) or non-alcoholic steatohepatitis (NASH), the latter usually in conjunction with fibrosis. Patients with NASH/fibrosis are at risk of developing terminal liver disease through progressing to cirrhosis and hepatic decompensation requiring liver transplantation. Alternatively, NASH and fibrosis may progress to hepatocellular carcinoma (HCC). The typical time-frame of disease progression is a decade or more, i.e. NAFLD is best viewed as a chronic disease. During progression, NAFLD often remains asymptomatic and is usually diagnosed incidentally.

Epidemiology of NAFLD and overlap with obesity, T2D and CVD

The prevalence of NAFL in the global population is around 25% and in the presence of obesity, as high as 51%. The overall global prevalence of NASH is estimated to be between 1.5%-6.45%^[1]. There exists a significant overlap between NAFLD and NASH with obesity, 51% and 82%, respectively, and CVD with cardiac-related deaths as being one of the most common outcomes for NAFLD patients^[2-^3]. Indeed, the global prevalence of hyperlipidemia/dyslipidemia among NAFLD and NASH patients is estimated to be 70% whereas hypertriglyceridemia in NAFLD and NASH patients is estimated to be 40.7% and 83.3%, respectively^[1]. In addition, hypertension has been diagnosed in 39.3% and 68% of NAFLD and NASH patients, respectively^[1]. For patients with T2D, NAFLD was reported in 76% of the cases studied using acohort selected for normal rather than elevated plasma levels of aminotransferases^[4]. About 56% of the patients of this cohort had NASH underscoring the apparent link between NAFLD/NASH and T2D as well as the limitations of using aminotransferases as the only means of diagnosing NAFLD (see below). In one study, the pooled overall diabetes prevalence in NAFLD and NASH patients were 22.5% and 43.6%, respectively^[1]. Fibrosis progression in conjunction with NASH is seen in about half of NASH patients placing these at risk of developing terminal liver disease including HCC which has one- and three-year survival rates of 36% and 17%, respectively (for discussion on NAFLD and links to HCC, please see^[5-^6]). Resection, ablation and/or transplantation improves survival rates (one- and three-year survival rates of 70% and 55%, respectively)^[7] for patients diagnosed with HCC. In the case of hepatic decompensation, liver transplantation is the only viable option. Given the prevalence of NAFLD and NASH in the general population, the availability of suitable organ grafts becomes increasingly restricted necessitating research and method developments in the areas of ex-vivo reconditioning (e.g. defatting) of donor livers prior to transplantationas well as improving assessment of border-line grafts (extended criteria).Such improvements are feasible given that patients have already been transplanted with livers kept at 37°C ex vivo using a normothermic perfusion system^[8].

NAFLD-NASH diagnostics

NAFLD is diagnosed as hepatic fat deposition (in cytoplasmic lipid droplets, CLDs) in more than 7% of the hepatocytes (as deduced by histochemistry). This usually occurs in the centrilobular zone 3 localized around central veins. Associated ballooning of hepatocytes signifies NASH, which is usually cryptogenic in nature. Whereas it is commonly agreed that NAFLD without NASH has slow or negligible histological progression, patients with NASH may exhibit progression to terminal liver disease (cirrhosis, HCC or decompensation). Apart from hepatic ballooning, NASH is often associated with an invasion of leukocytes and differentiation of hepatic stellate cells, a cell type that represents 5-8% of liver cells and store vitamin A deposited in CLDs. In NASH, activated hepatic stellate cells lose their vitamin A deposits, migrate to sites of damage (e.g. dead cells) and produce collagen, i.e., producing fibrinous tracks (revealed by tri-chrome stain). NASH is often seen together with fibrosis and, at some point, becomes irreversible with subsequent progressing to cirrhosis. All this takes time, sometimes several years to a decade. Meanwhile, liver conditions may remain asymptomatic. As such, NAFLD and NASH are usually diagnosed incidentally or in conjunction with one or more co-morbidity (e.g. obesity). As the prevalence of NAFLD is on the increase^[5-^6] and overlaps with the metabolic syndrome, clinical practice guidelines now recommend that patients with obesity and/or T2D should be examined for NAFLD^[9]. High triglyceride levels in combination with low serum HDL are also known to be common in patients with NAFLD and frequent (50%) in patients with dyslipidemia attending lipid clinics^[10]. Taken together and given the observed prevalence of NAFLD, an argument for routine NAFLD diagnostics seems obvious. Such an argument, however, is countered by a high cost/benefit ratio coupled with an increased risk for the patient.

The diagnosis and staging of NAFLD has for a long time focused on histology-based evaluation of liver biopsies taken from liver grafts or patients suspected of having liver damage. This approach, especially with respect to staging of NASH, remains the gold-standard for diagnosis. Both percutaneous and transjugular liver biopsies, however, have associated complications and though difficult to pinpoint exactly, a non-negligible mortality rate. Complications include hemorrhage (0.35%-0.5%), puncture of other vicera (0.01%-0.1%) and moderate to severe pain (1.5%-3%) (see Guidelines on the use of Liver Biopsy in Clinical Practice by A. Grant, J. Neuberger, C. Day and S. Saxseena available at www.bsg.org.uk). A biopsy represents only a 1/50€000^th of the liver and is therefore prone to significant sample error, i.e., that the biopsy does not accurately capture the disease state as steatosis, hepatitis and fibrosis are manifested unevenly^[11-^12]. There is also an apparent subjectivity in pathology-based assessment of NAFLD^[13]. As such, liver biopsy is not considered suitable as part of routine diagnostics and is recommended only as a means of in-depth assessment of disease severity^[9].

Non-invasive diagnostics center around serum markers and imaging. AST-to-platelet ratio index (APRI) has an AUROC of 0.842 as determined in patients with hepatitis C virus^[14]. BMI, hyperglycemia, platelet count and albumin are factors that are added to the AST/SLT ratio and is referred to as the NAFLD fibrosis score calculated according to published formula (www.nafldscore.com) (see also hepatic steatosis index, HIS, for comparison^[15]). The NAFLD fibrosis score has an AUROC of 0.85 predicting advanced fibrosis (bridging fibrosis and cirrhosis) with a 90% sensitivity and 60% specificity to exclude (at a score of<-1.455) and a 67% sensitivity and 97% specificity to identify advanced fibrosis (at a score of>0.676). Another set, the enhanced liver fibrosis panel, examines the plasma levels of two matrix-turnover proteins (PIIINP and TIMP-1) and hyaluronic acid and has a AUROC of 0.90 with 80% sensitivity and 90% specificity^[16]. Additional promising serum markers include FIB-4, KRT18, ALP (alkaline phosphatase) and bilirubin (a good indicator of more severe liver damage).

Magnetic resonance imaging (MRI) and elastrography (MRE) have been demonstrated to be highly accurate in detecting hepatic steatosis and fibrosis (for review, see^[17]). Advanced fibrosis can also be detected through transient elastography (Fibroscan)^[18-^20] and grade of steatosis through ultrasonography^[21] Xenon-133 liver scan^[22]. As shown in a recent longitudinal study^[23]monitoring liver fibrosis caused by viral infection or in conjunction with NAFLD/NASH, serum markers as well as imaging (e.g. Fibroscan) are well suited to predict the clinical outcome of NASH, a conclusion supported in a related study^[24] (for recent reviews-see^[25-^26]). It is therefore possible to consider the introduction of one or more non-invasive diagnostic modalities as part of routine diagnostics. Challenges and limitations nevertheless remain. Diagnosis and staging of NASH are difficult without biopsy and can only be inferred through fibrosis. As NAFLD and NASH overlap with the metabolic, imaging-based diagnostics is limited due to the white adipose tissue. Also, imaging instruments (e.g. MRI, MRE, Fibroscan) are expensive, require highly qualified personnel and as such, are restricted to specialized and academic centers. Development of new metabolic (bio)markers are therefore needed to complement existing ones to better stage NAFLD from NAFL through NASH with increasing fibrosis to cirrhosis.

Current development of improved NALD/NASH biomarkers

Animal model systems exist for both NAFLD and NASH providing a homogenous genetic background and a controlled environment. A rich diet combined with fructose is often used to induce NAFLD and NASH in rodents, either in wildtype or genetically modified animals (e.g. ob/ob or db/db mice deficient in leptin production or leptin binding, respectively. Such animal model systems display many of the hallmarks of liver disease (for a recent review, see^[27]) though with some limitations (e.g. in the induction of disease and an insufficient histopathological representation compared to human NAFLD/NASH (for review, see^[28]). Given the limited success of pharma industry using rodent model systems in drug development, many have questioned the usefulness of rodent model systems and instead, increasingly promote the use of human material^[29-^33]. This includes the use of embryonic stem cells, induced pluripotent stem cells and adult stem cells that can either be used to generate differentiated functional cells or organoids; self-organizing multicellular structures containing multiple cell types that mimic organ structure and function. Alternatively, material can be obtained from patients and organ donors including body fluids (e.g. blood, urine), tissue and whole organs. Biomarker research using human-derived material, however, has its own limitations. Restricted availability, large variability and ranging quality of human-derived material present significant problems as do alignment of research data with patient/donor data, the latter being safeguarded to ensure patient and donor confidentiality. In our case, human research material has been obtained from donor livers and from patients undergoing liver surgery. As expected, this material displays great variability but importantly, provides snap-shots of the NAFLD/NASH disease state. Such snap-shots combine genetic and environmental factors that are difficult to mimic in an animal model system. With respect to sample variability, we found that this can be beneficial in that it has revealed new and unexpected insights and enabled hypothesis-driven research (e.g. delineating pathways in a disease context). Despite limited availability and large variability, the use of human material for NAFLD/NASH research has proven valuable (manuscript in preparation, Nilsson, Metrakos et al.).

With respect to alignment of research data with patient/donor data, were centlydeveloped a framework for incidental research findings such that these can get back to the physician treating the patient (if deemed significant by an independent board)^[34]. In this framework, patient confidentiality is guaranteed through coded keys held by a third party. An important component of this framework is also informed consent. Using a similar approach, it should be possible to design legal and ethical protocols to enable research data to be more fully aligned with patient data. For this to work, emphasis needs to be put on policy development in the context of an integrated biobank framework including access and review board dealing with requests for research material and associated information. Equally important is the standardization of sample procurement, handling and long-term storage. Once in place, the prospect of finding new biomarkers will likely improve. Below follows a summary of existing markers ranging from genetic, epigenetic, protein and metabolic markers.

Genome wide association (GWAS) studies of NAFLD and NASH have already yielded genetic markers (variants). Of the more wide-spread ones, PNPLA3^I148M (rs738409) shows the strongest correlation with hepatic steatosis and liver fibrosis^[35-^40]. A somewhat weaker yet strong association with hepatic steatosis and liver fibrosis is also seen with TM6SF2^E167K (rs58542926)^[41], a protein associated with lipidation of secreted VLDL particles^[42]. Even though carriers of the PNPLA3^I148M and the TM6SF2^E167K variants have higher liver fat content and increased risk of developing NASH, genotyping is not recommended routinely^[9]. Other genetic variants include ApoCIII; FDFT1, a farnesyl transferase; NCAN, a chondroitin sulfate proteoglycan; PPP1R3B, a protein phosphatase; GCKR, a regulatory protein of glucose kinase; and LYPLAL1, an enzyme with unknown substrate specificity^[37,⁴³-^44]. A number of additional variants have been identified awaiting independent verifications (reviewed in^[45]).

Epigenetic regulation and modifications are also thought to contribute to NAFLD and NASH. This includes micro RNAs where miR-122, -192 and -375 correlatewell with both NAFLD and NASH.Thesecan be used to distinguish NASH from simple steatosis with similar predictive values as KRT18, ALTor AST and in addition, offer targets for intervention^[46]. Other types of epigenetic changes include histone modifications and DNA methylations. So far, only a few human studies have been conducted. Animal studies demonstrate changes in chromatin structure via histone modifications^[47], aberrant histone trimethylation of lipid catabolism and PPARα genes^[48], changes in DNA methylation upon depletion of methyl donors and post translational modifications of transcription factors (for reviews, see^[49-^50]).

The predictive value of the above listed markers remains uncertain as different studies show great variability in the assessment of disease state and progression (reviewed in^[49]). As such, the cost/benefit ratio remains high precluding the use of these in (routine) diagnostics.

Molecular pathophysiology of NAFLD and NASH

Most agree that NAFLD can be viewed as the hepatic expression of the metabolic syndrome. As NAFLD and NASH are chronic diseases, it is difficult to discern how either feeds into different co-morbidities associated with the metabolic syndrome or, how NAFLD and NASH might develop as a consequence of the metabolic syndrome. It is clear that significant overlap exists between NAFLD/NASH and obesity, pre-diabetes, hypertension, T2D as well as CVD. A long favored hypothesis explaining how NAFL progresses into NASH was the “two-hit” hypothesis whereby hepatic steatosis is the consequence of overweight/obesity (constituting the first hit). Progression of NAFL to NASH is then induced by an additional assault on the liver (e.g. lipotoxicity through circulating free fatty acids such as palmitate). This hypothesis is now considered obsolete and is replaced by the multiple hit hypothesis^[51]. Lipotoxicity, ER stress, hormonal and cytokine secretion from adipose tissue, changes in gut microbiota, medication, genetic, epigenetic factors and insulin resistance all contribute to NAFLD progression^[51]. For example, insulin levels in patients afflicted by T2D range from very low (due to pancreatic beta cell failure to produce insulin) to higher than normal levels (due to insulin resistance). In those with low insulin production, hormone-sensitive lipase becomes constituently activated resulting in TG breakdown and release of free fatty acids (FFAs) from adipose tissue. These are then taken up by the liver and either re-released via VLDL or stored in hepatic CLDs. Increased levels of circulating TGs in turn contribute to the development of CVD as most if not all VLDL secreted in conjunction with NAFLD are in the form of VLDL₁ containing ApoC3 (ApoCIII) (for reviews, see^[52-^53]).

Loss of function polymorphism of ApoC3 results in lower risk of CVD and lower amounts of circulating TGs. The exact mechanistic reasons remain unclear although it is known that ApoC3 inhibits both lipoprotein lipase (LPL) and hepatic lipase-mediated lipolysis and that this protein facilitates VLDL₁ assembly and secretion^[54-^55]. The activity of ApoCIII is governed by O-linked glycosylation as mutations in polypeptide N-acetylgalactosaminyltransferase 2 (GALNT2) results in a modulationof ApoCIII function decreasing its ability to inhibit LPL-mediated lipolysis of VLDL-bound TG^[56]. If this affects hepatic steatosis and fat accumulation in other tissues remains to be tested. The ability to clear lipids via secreted VLDL particles is one of the more important routes whereby the hepatocyte can offset uptake of circulating FFAs alleviating overall steatosis. Other means are through upregulation of β-oxidation or unconventional secretion. A candidate pathway for unconventional secretion has been highlighted in differentiated 3T3-L1 adipocytes with respect to the secretion of aP2 (FABP4) which contributes to increased liver glucose secretion and consequently, hyperglycemia and T2D^[57]. This pathway involves inclusion into multi-vesicular bodies, an endo/lysosomal compartment essential for antigen presentation. Similarly, a portion of adiponectin appears to follow the same route as a P2. Recent work also shows that some forms of autophagy are intimately coupled with unconventional secretion (reviewed in^[58]).Whether or not this enables free fatty acids or triglycerides to be secreted remains to be determined.

The ability to clear hepatic triglycerides through an upregulation of autophagy has been demonstrated in several studies (for a recent review, see^[59]). Caffeine ameliorates the symptoms of both NAFLD and NASH (for recent reviews, see^[60-^61]). It is estimated that a daily consumption of 2 cups of brewed regular coffee is sufficient to stimulate clearance of hepatic triglycerides through lipophagy coupled toβ-oxidation^[62-^63]. Caffeine-induced autophagy does not appear restricted to hepatic tissue and is also seen elsewhere including skeletal and neuronal tissue. Dietary lipids such as omega-3 fatty acids are also known to lower circulating triglyceride levels through decreased VLDL secretion^[64] coupled with induced hepatic autophagy^[65] and presumably, increased β-oxidation. In mice, this leads to improvement in NAFL, NASH as well as fibrosis^[66]. Other dietary-based management strategies also exist (e.g. dietary polyphenols^[67]) and when viewed together, offer possible avenues for treatment. At this stage, however, no clinical studies have been performed to show a clear reduction in the rate of disease progression from NAFL to NASH, fibrosis into a terminal liver disease stage.

The clearance of hepatic lipids through an upregulation of autophagy (i.e. lipophagy), increased β-oxidation, VLDL secretion and/or unconventional secretion must exceed the rate whereby neutral lipids are taken up or synthesized. This is far from simple to achieve. First, multiple lines of evidence suggest that CLDs may serve as obligatory intermediates in autophagy-related processes^[68-^69]. In other words, that CLDs form to supply the forming autophagosome with lipid material (e.g. phospholipids). Lipid content is also intimately intertwined with ER stress such that lipid composition effects the unfolded protein response (for review, see^[70]). Also, that reactive oxygen and nitrogen species causes ER stress by modifying proteins and lipids. This, in turn, promotes CLD formation.

Another promotor of CLD formation is uric acid linking renal impairment to the development of NAFLD^[71]. Addition of uric acid to hepatoma cells resulted in the induction of ER stress and activation of SREBP-1c, an ER bound transcription factor that upon activation, is proteolytically cleaved and translocates to the nucleus to promote transcription of genes involved in fatty acid synthesis (e.g. ACC1, FAS and SCD1). This mechanistic framework, linking ER stress to CLD formation through SREBP-1c, extends to autophagy through MTORC1 and LPIN1. During growth and high nutrient conditions, the MTORC1 complex is active inhibiting autophagy while promoting protein synthesis and the activity of transcription factors such as SREBP-1c. LPIN1, a phosphatidic acid (PA) phosphatase converting PA to diacylglycerol (DAG) becomes phosphorylated by the active MTORC1 preventing its entry into the nucleus^[72]. This prevents LPIN1 to inhibit the nuclear activity of SREBP-1c.

LPIN1 is also required for autophagic flux through the activation of the PKD-VPS34 signaling pathway^[73]. Again, its ability to convert DAG to PA is inhibited by MTORC1-dependent phosphorylation ensuring multiple blocks of autophagy during growth and high nutrient conditions. Activation of MTORC1, however, is also linked to ER stress causing conflicting signaling cascades, on the one hand promoting lipogenesis and on the other hand, preventing autophagy and CLD clearance through the inactivation of LPIN1 (see for example, work in Zebra fish^[74]). Such dysregulation might have a direct relevance to the progression of NAFLD and offers direct avenues for intervention through drugs that inhibit MTORC1 (e.g. rapamycin derivatives).

With respect to T2D and pre-states of T2D, the hallmarks of liver insulin resistance are unabated gluconeogenesis and unsuppressed lipogenesis. Gluconeogenesis and lipogenesis are respectively regulated through the insulin/IRS2 and the insulin/IRS1 pathways. Under normal conditions, insulin action through the IRS2 pathway leads to phosphorylation and nuclear exclusion of FOXO1 with dampened expression of gluconeogenesis genes such as G6Pase and PEPCK but also MTP and apoC-III that are both essential for VLDL₁ assembly. Sustained hyperinsulinemia and hyperglycemia with compromised insulin/IRS2/FOXO1 signaling yields high expression of MTP and apoC-III with an overproduction of TG-VLDL₁^[52,⁷⁵-^79] yet causes of compromised insulin/IRS2/FOXO1 signaling remain unclear. It is also possible that hepatic accumulation of CLDs directly impacts glucose intolerance and insulin resistance.

Disease management and possible treatment of patients with NAFLD

Lifestyle modifications including dietary changes, restricted calorie intake and a minimum of weekly exercise have been shown to improve NAFL as deduced by ultrasound or MRI (summarized in^[16]) as well as NASH^[80].Changes in lifestyle is therefore recommended to reduce NAFL and to improve NASH^[16]. Use of metformin show only limited improvements in NAFL but has no significant impact on liver histology and is therefore not recommended in the treatment of adults with NASH^[16]. In contrast, pioglitazones have been shown to improve NASH but should be restricted to non-diabetic patients. Also, long-term safety of pioglitazones is under considerable debate and is either not recommended or severely restricted in use. This is due to increased risks in coronary events^[81]. Vitamin E administered daily improves liver histology in non-diabetic patients with NASH and is recommended as a first line treatment^[16] and other dietary supplements can be considered including those mentioned above though at present, may be premature awaiting clinical studies.

Concluding remarks

The metabolic dysregulation leading to NAFLD and NASH and consequent dysregulations caused by NAFLD and NASH impacting other disease states (e.g. T2D and CVD) are progressive in nature each feeding into the other. As NAFLD and NASH are increasing in prevalence and already affecting large part of the population, it is important to understand that the consequences of NAFLD and NASH will place a huge burden on healthcare. A first step should therefore be to educate primary care physicians to recognize risk factors associated with NAFLD and in particular, NASH with fibrosis, such that the appropriate investigation can be carried out. A second step should be to implement serum-based diagnostics and imaging modalities as part of routine diagnostics. It should be stated that it is presently difficult to design prospective studies; to recruit relevant cohorts that are representative of the population. This difficulty is compounded by granting agencies still questioning the relevance of NAFLD as a metabolic disease despite its documented impact, prevalence and increased mortality rate (in the case of NASH, fibrosis and cirrhosis). Large NAFLD and NASH consortia have nevertheless been established both in the US and the EU that all focus on a better understanding of disease, improved diagnostics and enhanced treatment strategies.

References

[1] Younossi ZM, Koenig AB, Abdelatif D, Global epidemiology of non-alcoholic fatty liver disease-meta-analytic assessment of prevalence, incidence and outcomes[J]. Hepatology, 2016, 64(1):73–84.

[2] Ekstedt M, Hagstrom H, Nasr P, Fibrosis stage is the strongest predictor for disease-specific mortality in NAFLD after up to 33 years of follow-up[J]. Hepatology, 2015, 61(5): 1547–1554.

[3] Stepanova M, Rafiq N, Makhlouf H, Predictors of all-cause mortality and liver-related mortality in patients with non-alcoholic fatty liver disease (NAFLD)[J]. Dig Dis Sci, 2013, 58(10): 3017–3023.

[4] Portillo-Sanchez P, Bril F, Maximos M, High prevalence of nonalcoholic fatty liver disease in patients with type 2 diabetes mellitus and normal plasma aminotransferase levels[J]. J Clin Endocrinol Metab, 2015, 100(6): 2231–2238.

[5] Page JM, Harrison SA. NASH and HCC[J]. Clin Liver Dis, 2009, 13(4): 631–647.

[6] Starley BQ, Calcagno CJ, Harrison SA. Nonalcoholic fatty liver disease and hepatocellular carcinoma: a weighty connection[J]. Hepatology, 2010, 51(5): 1820–1832.

[7] Davis GL, Dempster J, Meler JD, Hepatocellular carcinoma: management of an increasingly common problem[J]. Proc Bayl Univ Med Cent, 2008, 21: 266–280.

[8] Ravikumar R, Jassem W, Mergental H, Liver Transplantation After Ex Vivo Normothermic Machine Preservation: A Phase 1 (First-in-Man) Clinical Trial[J]. Am J Transplant, 2016, 16(6): 1779–1787.

[9] European Association for the Study of the Liver. Electronic address eee, European Association for the Study of D, European Association for the Study of O. EASL-EASD-EASO Clinical Practice Guidelines for the management of non-alcoholic fatty liver disease[J]. J Hepatol, 2016, 64(6): 1388–1402

[10] Assy N, Kaita K, Mymin D, Fatty infiltration of liver in hyperlipidemic patients[J]. Dig Dis Sci, 2000, 45(10): 1929–1934

[11] Ratziu V, Charlotte F, Heurtier A, Sampling variability of liver biopsy in nonalcoholic fatty liver disease[J]. Gastroenterology, 2005, 128(7): 1898–1906.

[12] Janiec DJ, Jacobson ER, Freeth A, Histologic variation of grade and stage of non-alcoholic fatty liver disease in liver biopsies[J]. Obes Surg, 2005, 15(4): 497–501.

[13] Jung ES, Lee K, Yu E, Interobserver Agreement on Pathologic Features of Liver Biopsy Tissue in Patients with Nonalcoholic Fatty Liver Disease[J]. J Pathol Transl Med, 2016, 50(3): 190–196.

[14] Verlinden W, Bourgeois S, De Maeyer M, Validation of APRI and FIB-4 score in an Antwerp cohort of chronic hepatitis C patients[J]. Acta Gastroenterol Belg, 2015, 78: 373–380.

[15] Lee JH, Kim D, Kim HJ, Hepatic steatosis index: a simple screening tool reflecting nonalcoholic fatty liver disease[J]. Dig Liver Dis, 2010, 42(7): 503–508.

[16] Chalasani N, Younossi Z, Lavine JE, The diagnosis and management of non-alcoholic fatty liver disease: practice Guideline by the American Association for the Study of Liver Diseases, American College of Gastroenterology, and the American Gastroenterological Association[J]. Hepatology, 2012, 55(6): 2005–2023.

[17] Dulai PS, Sirlin CB, Loomba R. MRI and MRE for non-invasive quantitative assessment of hepatic steatosis and fibrosis in NAFLD and NASH: Clinical trials to clinical practice[J]. J Hepatol, 2016, 65(5): 1006–1016.

[18] Kwok R, Tse YK, Wong GL, Systematic review with meta-analysis: non-invasive assessment of non-alcoholic fatty liver disease–the role of transient elastography and plasma cytokeratin-18 fragments[J]. Aliment Pharmacol Ther, 2014, 39(3): 254–269.

[19] Sandrin L, Fourquet B, Hasquenoph JM, Transient elastography: a new noninvasive method for assessment of hepatic fibrosis[J]. Ultrasound Med Biol, 2003, 29(12): 1705–1713

[20] Sebastiani G, Tempesta D, Fattovich G, Prediction of oesophageal varices in hepatic cirrhosis by simple serum non-invasive markers: Results of a multicenter, large-scale study[J]. J Hepatol, 2010, 53(4): 630–638.

[21] Lee SS, Park SH, Kim HJ, Non-invasive assessment of hepatic steatosis: prospective comparison of the accuracy of imaging examinations[J]. J Hepatol, 2010, 52(4): 579–585.

[22] Al-Busafi SA, Ghali P, Wong P, The utility of Xenon-133 liver scan in the diagnosis and management of nonalcoholic fatty liver disease[J]. Can J Gastroenterol, 2012, 26(3): 155–159.

[23] Sebastiani G, Alshaalan R, Wong P, Prognostic value of non-invasive fibrosis and steatosis tools, hepatic venous pressure gradient (HVPG) and histology in nonalcoholic steatohepatitis[J]. PLoS One, 2015, 10(6): e0128774

[24] Angulo P, Bugianesi E, Bjornsson ES, Simple noninvasive systems predict long-term outcomes of patients with nonalcoholic fatty liver disease[J]. Gastroenterology, 2013;145:782–9 e4.

[25] Hannah WN Jr, Harrison SA. Noninvasive imaging methods to determine severity of nonalcoholic fatty liver disease and nonalcoholic steatohepatitis[J]. Hepatology, 2016;64(6):2234–2243.

[26] Bedossa P, Patel K.Biopsy and noninvasive methods to assess progression of nonalcoholic fatty liver disease[J]. Gastroenterology. 2016;150:1811–22 e4.

[27] Ibrahim SH, Hirsova P, Malhi H, Animal models of nonalcoholic steatohepatitis: eat, delete, and inflame[J]. Dig Dis Sci, 2016, 61(5): 1325–1336

[28] Sanches SC, Ramalho LN, Augusto MJ, . Nonalcoholic steatohepatitis: a search for factual animal models[J]. biomed research international. 2015;2015:574832.

[29] Collins FS. Reengineering translational science: the time is right[J]. Sci Transl Med, 2011, 3(90): 90cm17

[30] Sartipy P, Bjorquist P. Concise review: Human pluripotent stem cell-based models for cardiac and hepatic toxicity assessment[J]. Stem Cells, 2011, 29(5): 744–748.

[31] Singh M, Ferrara N. Modeling and predicting clinical efficacy for drugs targeting the tumor milieu[J]. Nat Biotechnol, 2012, 30(7): 648–657.

[32] Raven K. Rodent models of sepsis found shockingly lacking[J]. Nat Med, 2012, 18(7): 998.

[33] Mullane K, Williams M. Translational semantics and infrastructure: another search for the emperor’s new clothes[J]? Drug Discov Today, 2012, 17(9-10): 459–468.

[34] Thorogood A, Joly Y, Knoppers BM, An implementation framework for the feedback of individual research results and incidental findings in research[J]. BMC Med Ethics, 2014, 15(1): 88.

[35] Romeo S, Kozlitina J, Xing C, Genetic variation in PNPLA3 confers susceptibility to nonalcoholic fatty liver disease[J]. Nat Genet, 2008, 40(12): 1461–1465.

[36] Sookoian S, Pirola CJ. Meta-analysis of the influence of I148M variant of patatin-like phospholipase domain containing 3 gene (PNPLA3) on the susceptibility and histological severity of nonalcoholic fatty liver disease[J]. Hepatology, 2011, 53(6): 1883–1894.

[37] Speliotes EK, Butler JL, Palmer CD, PNPLA3 variants specifically confer increased risk for histologic nonalcoholic fatty liver disease but not metabolic disease[J]. Hepatology, 2010, 52(3): 904–912

[38] Rotman Y, Koh C, Zmuda JM, The association of genetic variability in patatin-like phospholipase domain-containing protein 3 (PNPLA3) with histological severity of nonalcoholic fatty liver disease[J]. Hepatology, 2010, 52(3): 894–903

[39] Valenti L, Alisi A, Galmozzi E, I148M patatin-like phospholipase domain-containing 3 gene variant and severity of pediatric nonalcoholic fatty liver disease[J]. Hepatology, 2010, 52(4): 1274–1280

[40] Valenti L, Al-Serri A, Daly AK, Homozygosity for the patatin-like phospholipase-3/adiponutrin I148M polymorphism influences liver fibrosis in patients with nonalcoholic fatty liver disease[J]. Hepatology, 2010, 51(4): 1209–1217

[41] Kozlitina J, Smagris E, Stender S, Exome-wide association study identifies a TM6SF2 variant that confers susceptibility to nonalcoholic fatty liver disease[J]. Nat Genet, 2014, 46(4): 352–356

[42] Smagris E, Gilyard S, BasuRay S, Inactivation of Tm6sf2, a gene defective in fatty liver disease, impairs lipidation but not secretion of very low density lipoproteins[J]. J Biol Chem, 2016, 291(20): 10659–10676

[43] Petersen KF, Dufour S, Hariri A, Apolipoprotein C3 gene variants in nonalcoholic fatty liver disease[J]. N Engl J Med, 2010, 362(12): 1082–1089

[44] Burger M, Zimmermann TJ, Kondoh Y, Crystal structure of the predicted phospholipase LYPLAL1 reveals unexpected functional plasticity despite close relationship to acyl protein thioesterases[J]. J Lipid Res, 2012, 53(1): 43–50

[45] Wood KL, Miller MH, Dillon JF. Systematic review of genetic association studies involving histologically confirmed non-alcoholic fatty liver disease[J]. BMJ Open Gastroenterol, 2015, 2(1): e000019

[46] Pirola CJ, Fernandez Gianotti T, Castano GO, Circulating microRNA signature in non-alcoholic fatty liver disease: from serum non-coding RNAs to liver histology and disease pathogenesis[J]. Gut, 2015, 64(5): 800–812

[47] Aagaard-Tillery KM, Grove K, Bishop J, Developmental origins of disease and determinants of chromatin structure: maternal diet modifies the primate fetal epigenome[J]. J Mol Endocrinol, 2008, 41(2): 91–102

[48] Jun HJ, Kim J, Hoang MH, Hepatic lipid accumulation alters global histone h3 lysine 9 and 4 trimethylation in the peroxisome proliferator-activated receptor alpha network[J]. PLoS One, 2012, 7(9): e44345

[49] Eslam M, George J. Genetic and epigenetic mechanisms of NASH[J]. Hepatol Int, 2016, 10(3): 394–406

[50] Gallego-Durán R, Romero-Gomez M. Epigenetic mechanisms in non-alcoholic fatty liver disease: An emerging field[J]. World J Hepatol, 2015, 7(24): 2497–2502.

[51] Buzzetti E, Pinzani M, Tsochatzis EA. The multiple-hit pathogenesis of non-alcoholic fatty liver disease (NAFLD). Metabolism, 2016, 65(8): 1038–1048.

[52] Jiang ZG, Robson SC, Yao Z. Lipoprotein metabolism in nonalcoholic fatty liver disease[J]. J Biomed Res, 2013, 27(1): 1–13

[53] Bernelot Moens SJ, van Capelleveen JC, Stroes ES. Inhibition of ApoCIII: the next PCSK9[J]? Curr Opin Lipidol, 2014, 25(6): 418–422.

[54] McConathy WJ, Gesquiere JC, Bass H, Inhibition of lipoprotein lipase activity by synthetic peptides of apolipoprotein C–III[J]. J Lipid Res, 1992, 33: 995–1003.

[55] Qin W, Sundaram M, Wang Y, Missense mutation in APOC3 within the C-terminal lipid binding domain of human ApoC-III results in impaired assembly and secretion of triacylglycerol-rich very low density lipoproteins: evidence that ApoC-III plays a major role in the formation of lipid precursors within the microsomal lumen[J]. J Biol Chem, 2011, 286(31): 27769–27780.

[56] Holleboom AG, Karlsson H, Lin RS, Heterozygosity for a loss-of-function mutation in GALNT2 improves plasma triglyceride clearance in man[J]. Cell Metab, 2011, 14(6): 811–818

[57] Ertunc ME, Sikkeland J, Fenaroli F, Secretion of fatty acid binding protein aP2 from adipocytes through a nonclassical pathway in response to adipocyte lipase activity[J]. J Lipid Res, 2015, 56(2): 423–434.

[58] Deretic V, Jiang S, Dupont N. Autophagy intersections with conventional and unconventional secretion in tissue development, remodeling and inflammation[J]. Trends Cell Biol, 2012, 22(8): 397–406.

[59] Mao Y, Yu F, Wang J, Autophagy: a new target for nonalcoholic fatty liver disease therapy[J]. Hepat Med, 2016, 8: 27–37.

[60] Kennedy OJ, Roderick P, Poole R, Coffee, caffeine and non-alcoholic fatty liver disease[J]? Therap Adv Gastroenterol, 2016, 9(3): 417–418.

[61] Marventano S, Salomone F, Godos J, Coffee and tea consumption in relation with non-alcoholic fatty liver and metabolic syndrome: A systematic review and meta-analysis of observational studies[J]. Clin Nutr, 2016, 35(6): 1269–1281.

[62] Sinha RA, Farah BL, Singh BK, Caffeine stimulates hepatic lipid metabolism by the autophagy-lysosomal pathway in mice[J]. Hepatology, 2014, 59(4): 1366–1380.

[63] Ding WX. Drinking coffee burns hepatic fat by inducing lipophagy coupled with mitochondrial beta-oxidation[J]. Hepatology, 2014, 59(4): 1235–1238.

[64] Wong AT, Chan DC, Ooi EM, Omega-3 fatty acid ethyl ester supplementation decreases very-low-density lipoprotein triacylglycerol secretion in obese men[J]. Clin Sci, 2013, 125(1): 45–51.

[65] Chen Y, Xu C, Yan T, . omega-3 Fatty acids reverse lipotoxity through induction of autophagy in nonalcoholic fatty liver disease[J]. Nutrition, 2015;31:1423–9 e2.

[66] Kim JK, Lee KS, Lee DK, Omega-3 polyunsaturated fatty acid and ursodeoxycholic acid have an additive effect in attenuating diet-induced nonalcoholic steatohepatitis in mice[J]. Exp Mol Med, 2014, 46(12): e127

[67] Parafati M, Lascala A, Morittu VM, Bergamot polyphenol fraction prevents nonalcoholic fatty liver disease via stimulation of lipophagy in cafeteria diet-induced rat model of metabolic syndrome[J]. J Nutr Biochem, 2015, 26(9): 938–948.

[68] Shibata M, Yoshimura K, Furuya N, The MAP1–LC3 conjugation system is involved in lipid droplet formation[J]. Biochem Biophys Res Commun, 2009, 382(2): 419–423.

[69] Dupont N, Chauhan S, Arko-Mensah J, Neutral lipid stores and lipase PNPLA5 contribute to autophagosome biogenesis[J]. Curr Biol, 2014, 24(6): 609–620.

[70] Volmer R, Ron D. Lipid-dependent regulation of the unfolded protein response[J]. Curr Opin Cell Biol, 2015, 33: 67–73.

[71] Choi YJ, Shin HS, Choi HS, Uric acid induces fat accumulation via generation of endoplasmic reticulum stress and SREBP-1c activation in hepatocytes[J]. Lab Invest, 2014, 94(10): 1114–1125.

[72] Peterson TR, Sengupta SS, Harris TE, mTOR complex 1 regulates lipin 1 localization to control the SREBP pathway[J]. Cell, 2011, 146(3): 408–420.

[73] Zhang P, Verity MA, Reue K. Lipin-1 regulates autophagy clearance and intersects with statin drug effects in skeletal muscle[J]. Cell Metab, 2014, 20(2): 267–279.

[74] Sapp V, Gaffney L, EauClaire SF, EauClaire SF, Matthews RP. Fructose leads to hepatic steatosis in zebrafish that is reversed by mechanistic target of rapamycin (mTOR) inhibition[J]. Hepatology, 2014, 60(5): 1581–1592.

[75] Altomonte J, Cong L, Harbaran S, Foxo1 mediates insulin action on apoC-III and triglyceride metabolism[J]. J Clin Invest, 2004, 114(10): 1493–1503

[76] Ferré P, Foufelle F. Hepatic steatosis: a role for de novo lipogenesis and the transcription factor SREBP-1c[J]. Diabetes Obes Metab, 2010, 12(Suppl 2): 83–92.

[77] Kamagate A, Qu S, Perdomo G, FoxO1 mediates insulin-dependent regulation of hepatic VLDL production in mice[J]. J Clin Invest, 2008, 118: 2347–2364.

[78] Wu X, Chen K, Williams KJ. The role of pathway-selective insulin resistance and responsiveness in diabetic dyslipoproteinemia[J]. Curr Opin Lipidol, 2012, 23(4): 334–344.

[79] Yao Z, Wang Y. Apolipoprotein C–III and hepatic triglyceride-rich lipoprotein production[J]. Curr Opin Lipidol, 2012, 23(3): 206–212.

[80] Promrat K, Kleiner DE, Niemeier HM, Randomized controlled trial testing the effects of weight loss on nonalcoholic steatohepatitis[J]. Hepatology, 2010, 51(1): 121–129.

[81] Lincoff AM, Wolski K, Nicholls SJ, Pioglitazone and risk of cardiovascular events in patients with type 2 diabetes mellitus: a meta-analysis of randomized trials[J]. JAMA, 2007, 298(10): 1180–1188.

RIGHTS & PERMISSIONS

2017 by the Journal of Biomedical Research.