GLP-1 is an antihyperglycemic hormone that induces pancreatic β cells to secrete insulin. Trevaskis et al (29) demonstrated that GLP-1 receptor agonists significantly improve metabolism and biochemical and histopathological indicators in NASH mice. Bernsmeier et al (30) reported that patients with NAFLD exhibited IR and insufficient GLP-1 secretion. Furthermore, Chellali et al (31) demonstrated that patients with NAFLD, T2D or NAFLD+T2D generally exhibited IR and that compared with controls, plasma GLP-1 levels were obviously lower in patients with NAFLD, T2D or NAFLD+T2D. GLP-1 receptor agonists have been approved for T2D treatment and an increasing number of previous studies have demonstrated that they are also effective for NAFLD (32-35).

FGF21 is a member of the fibroblast growth factor family and is a hormone-like endocrine factor secreted by the liver and serves a critical role in maintaining the homeostasis of glycolipid metabolism (36). Li et al (37) reported that serum FGF21 levels, such as the fatty liver index (38), which is an already known score associated with steatosis presence, have specificity for the early diagnosis of NAFLD. FGF21 is primarily produced by the liver and specifically acts on liver, fat and islet cells, FGF21 can reduce body weight, whole-body fat mass and liver triglyceride content, increases fat utilization and energy expenditure, and improves glucose tolerance and insulin sensitivity in the liver and adipose tissue (39). FGF21 is independent of insulin and regulates blood glucose and blood lipid (39). FGF21 can reduce glucose levels without triggering hypoglycemia, and increase the glucose uptake in brown fat, the browning of white fat and the increase in overall energy consumption. As a metabolic regulator, FGF21 significantly reduces the expression of key genes involved in lipid metabolism, including stearoyl-coenzyme A desaturase 1 and glucokinase, reduces fat syntheses, upregulates the expression of key genes involved in lipolysis, including leptin receptor and insulin-like growth factor binding protein 2 and promotes fat decomposition and energy metabolism, thereby reducing the accumulation of fat in liver tissue and restoring normal liver function (40).

IR is associated with the severity of liver fibrosis, the main determinant of NAFLD prognosis (41,42). LOXL2 is an enzyme that promotes the network of collagen fibers of the extracellular matrix (43). LOXL2 expression is strongly correlated with the amount of steatosis and severity of fibrosis. Hepatic LOXL2 upregulation was specifically detected in patients with NAFLD and T2D progressing to advanced fibrosis (22). Ikenaga et al (44) reported that LOXL2 expression was virtually absent from healthy livers while strongly induced in fibrotic livers and that LOXL2 inhibition promoted fibrosis reversal. Therefore, LOXL2 antibodies are a potential treatment for liver fibrosis. A monoclonal antibody against LOXL2, simtuzumab, has been developed and is currently in phase II trials (45).

Numerous previous studies have demonstrated that various inflammatory cytokines serve key roles in the progression of steatosis to NASH (53.54). Zahran et al (55) reported that tumor necrosis factor (TNF)-α levels were elevated in a patient population with NASH compared with controls. Furthermore, Bocsan et al (56) demonstrated that plasma interleukin (IL)-6 and TNF-α levels were significantly higher in patients with NASH compared with controls. TNF-α is produced by hepatocytes and infiltrating immune cells. Moreover, it stimulates liver steatosis, which in turn increases serum TG levels (57). Additionally, TNF-α activates Kupffer cells, which in turn stimulate liver fibrosis during NAFLD progression (58). IL-6 is primarily secreted by adipose tissue and, as an important mediator of fatty acid metabolism, serves a paradoxical role in the liver: While the hepatic IL-6 signaling pathway has a protective effect against the development of liver steatosis, it may also promote liver inflammation (59).

PNPLA3 is a secretory adipose factor that is widely expressed in human tissues, particularly in liver and adipose tissues (60). In 2008, Romeo et al (61) conducted a genome-wide association analysis to screen for pathogenic genes in NAFLD. A group of nonsynonymous sequence variations including Hispanic, African-American and European-American participants (n=9.229) were selected and polymorphic PNPLA3 rs738409 in liver tissue was identified in order to identify genetic variants contributing to differences in hepatic fat content. Increased concentrations of fat and liver inflammation were highly associated with Homeostatic Model Assessment of IR and the mechanism of liver damage may be associated with intrahepatic lipid deposition. Additionally, Wang et al (62) studied 768 patients from the Han Chinese population and confirmed the association between PNPLA3 rs738409 and an increased risk for NAFLD in this population, emphasizing its effects on NAFLD among multiple ethnic groups. Several previous studies in a meta-analysis demonstrated that genetic variation in PNPLA3 increases the risk of NAFLD in European and Chinese populations (63). Numerous previous studies have reported that PNPLA3 gene polymorphism is associated with an increase in liver fat content and liver inflammation (63-65). As an enzyme with a bidirectional ability to regulate fat metabolism, PNPLA3 promotes lipid synthesis and may be involved in ectopic lipid deposition and IR, the two key factors in the pathogenesis of NAFLD; therefore, PNPLA3 leads to the pathogenesis of NAFLD (66).

PPAR ligand-activated transcription factors control different aspects of lipid catabolism. The PPAR subtype PPAR-α is highly expressed in the liver, where it promotes genes that regulate the β-oxidation of fatty acids. Thus, inducing PPAR-α-regulated lipid catabolic pathway genes may enhance NAFLD. Another subtype, PPAR-γ, is a nuclear receptor that regulates insulin sensitivity and lipid metabolism, and a molecular target for insulin sensitizing agents. PPAR-γ is expressed in hepatic stellate cells (HSCs) to influence perilipin expression, which regulates lipid droplet metabolism. The activation of HSCs largely contributes to the fibrogenesis process (67).

SREBPs control cholesterol, TGs and fatty acids. SREBP-1c is a subtype predominantly localized in the adult liver and is required for simple steatosis during fasting. Overexpressed SREBP-1c leads to excessive TG accumulation in the liver and ultimately to NAFLD. Thus, the inhibition of SREBP-1c is a preventive approach (68). PPARα and SREBP-1 are key determinants of NAFLD (69).

SREBP-2 is an important nuclear transcription factor that regulates cellular cholesterol homeostasis. Its activation promotes the accumulation of hepatocyte cholesterol by facilitating cholesterol synthesis and uptake and suppressing reverse cholesterol transport (70). Additionally, SREBP-2 predicts NASH and widely impairs pancreatic β cell function, tissue insulin sensitivity and adipokine and lipoprotein responses to fat uptake. Thus, elevated SREBP-2 increases the prevalence of NAFLD and cardiovascular diseases. Abnormal SREBP-2 activation can be reversed by depleting cholesterol in adipocytes (71).

Adiponectin, an insulin-sensitizing adipocytokine, is an adipocyte-derived anti-inflammatory mediator that acts via two receptors that elicit AMP kinase signaling (72). Adipose tissue is a major site of endogenous adiponectin production. Previous research has demonstrated that plasma levels of adiponectin are significantly diminished in visceral obesity and states of IR such as NASH, T2D and atherosclerosis (73,74). Adiponectin and its agonists may represent potential alternative therapeutic approaches for NAFLD treatment.

Leptin is a hormone secreted mainly by the adipocytes in white adipose tissue and is a product of the obese gene (75). Perfield et al (76) reported that de novo lipogenesis was induced and that hepatic mitochondria were reduced and inhibited in leptin-deficient NAFLD mice. Zelber-Sagi et al (77) demonstrated that leptin levels were significantly higher in patients with NAFLD compared with patients without NAFLD at 7 years of follow-up. Furthermore, Polyzos et al (78) revealed that serum leptin levels in patients with NAFLD were higher compared with controls and that serum leptin levels were associated with increased severity of NAFLD. However, while a lack of leptin may lead to liver steatosis, which can be reversed by leptin replacements (79), excessive leptin may cause liver inflammation and fibrosis.

As members of the metabolic nuclear receptor superfamily, LXRs influence the regulation of cholesterol absorption, conveyance, transportation and excretion (80). Research has demonstrated that LXR activation regulates cholesterol homeostasis, induces anti-inflammatory effects and increases insulin sensitivity (81). Inhibiting liver LXR transcriptional activity in NAFLD effectively reduces liver steatosis, inflammation and fibrosis. However, research has found that LXR is highly expressed in the livers of patients with NAFLD and increases with the severity of NAFLD. LXR expression is directly proportional to liver lipid deposition (82). However, the uncertainty about whether LXR activation or inhibition is beneficial for NAFLD treatment has hindered the development of LXR-related drugs (83).

Plin2, also known as adipophilin or adipose differentiation-related protein, is commonly reported on the surface of lipid droplets in most tissues and hepatic lipids effectively induce Plin2 expression. Plin2 inhibition increases insulin sensitivity and reduces fatty liver (84). While perilipin 2-antisense oligonucleotide (Plin2-ASO) treatment decreases steatosis, it also promotes proliferation and fibrosis. Furthermore, Plin2-ASO inhibits the expression of genes that participate in TG biosynthesis and utilization, including diacylglycerol O-acyltransferase 2, acetyl coenzyme A carboxylase and fatty acid oxidation carnitine palmitoyltransferase 1 by suppressing transcription factors. Additionally, Plin2-ASO therapy significantly decreases the expression of genes related to cholesterol and steroid metabolism, reduces hepatic VLDL secretion and increase insulin sensitivity (85).

Choline is an important phospholipid component present in cell membranes and is ingested exogenously and synthesized endogenously. Choline is one of the methyl donors for the synthesis of phosphatidylcholine in the liver, which is required for the synthesis and secretion of VLDL. Therefore, choline serves a vital role in liver lipid transport (86). Choline has a negative correlation with NAFLD development by affecting intracellular lipid metabolism and muscle membrane lipid composition (87). Dietary choline is metabolized by intestinal microbiota and converted to trimethylamine, resulting in choline deficiency in the liver. Choline deficiency hinders the synthesis and secretion of VLDL, leads to the accumulation of TG in the liver and promotes the onset of NAFLD. High-fat diets lead to the formation of intestinal microbiota that convert dietary choline into methylamines, reducing circulating plasma levels of phosphatidylcholine to produce similar effects of choline-deficient diets and causing NASH (88). Choline-deficient diets have been used to establish animal models of NASH (89).

BCAAs are amino acids with nonlinear aliphatic side-chains and include the essential amino acids leucine, valine and isoleucine. Cheng et al (92) reported increased serum BCAAs in patients with NAFLD. Lake et al (91) demonstrated that during the progression of steatosis to NASH, serum leucine, isoleucine and valine levels were significantly increased. This increase is associated with liver fat accumulation during the early stages of NAFLD. BCAAs mediate numerous biological functions in NAFLD. Firstly, BCAAs regulate glucose metabolism by inducing insulin signaling (93,94). Secondly, they increase lipolysis and hyperlipidemia, which cause hepatic lipotoxicity that leads to inhibited autophagy in hepatocytes (95). Thirdly, BCAAs are beneficial for treating cirrhosis by decreasing portal vein pressure and increasing mean artery pressure (96). Finally, BCAAs induce higher branched-chain α-keto acid dehydrogenase E1α expression levels in skeletal muscles, which facilitates BCAA metabolism in muscle tissue (97).

Leucine supplementation has been demonstrated to activate mammalian target of rapamycin, which mediates cellular proliferation and protein synthesis, and increases insulin sensitivity. However, while leucine alleviates T2D, it also exacerbates NAFLD, making it difficult to use in NAFLD treatment (98).

AAAs are a class of α-amino acids that have aromatic rings and include tryptophan, phenylalanine and tyrosine. Celinski et al (99) demonstrated that tryptophan substantially attenuated the levels of proinflammatory cytokines and improved certain parameters (γ-glutamyl transferase, triglyceride and low-density lipoprotein) of lipid metabolism in patients with NAFLD. Chen et al (100) reported that serum phenylalanine concentration was two-fold higher in patients with NAFLD compared with controls. Furthermore, Jin et al (101) investigated the metabolic pathways that were altered in association with hepatic steatosis in adolescents and demonstrated that plasma tyrosine levels were positively correlated with the severity of liver steatosis. Phenylalanine is catalyzed to tyrosine by phenylalanine hydroxylase in the liver, where it is further metabolized (102). Presently, the mechanism of AAA metabolic disorders in NAFLD remains unclear. A possible explanation may be that tyrosine enters the ketogenic pathway and is directly degraded to acetyl coenzyme A via ketogenic action (101). Therefore, increased phenylalanine and tyrosine intake may contribute to fat deposition in the liver.

Methionine is an essential sulfur-containing amino acid that is mainly metabolized in the liver (103). Homocysteine is a sulfhydryl-containing amino acid produced and catabolized primarily in the liver and increased homocysteine concentrations in the event of liver injury. Alternatively, elevated homocysteine may conversely promote the progression of liver damage (104). Both methionine and homocysteine are associated with NAFLD. Methionine metabolism impacts the methylation process that leads to the production of glutathione and methylarginines, and regulates homocysteine concentrations. However, methionine deficiency leads to impaired methyltransferase and antioxidative reactions, and elevated homocysteine (105). Increased homocysteine causes hyperhomocysteinemia, which alters intracellular lipid metabolism to facilitate fat accumulation (106).

Arginine has antioxidant effects on NASH by inhibiting cytochrome P450 family 2 subfamily E member 1 activity and by decreasing TNF-α and antioxidant enzyme concentrations (107). Dogru et al (108) reported that increased circulating asymmetric dimethylarginine may be used as an early marker for NAFLD. Additionally, Voloshin et al (109) synthesized a novel chenodeoxycholic acid-arginine ethyl ester conjugate that may be used to treat NAFLD. The conjugate reduces liver steatosis by decreasing cholesterol and blood glucose.

Aminotransferases are barometers of liver health as they transport excessive amino acids into the circulation to manage liver metabolic derangement associated with severe gluconeogenesis and IR (90). Aspartate aminotransferase, alanine transaminase and γ-glutamyltransferase are generally elevated in NAFLD and are positively associated with CIC-2 expression in the liver (24,110).

Nuclear receptors (NRs) are transcriptional factors sensitive to environmental or hormonal signals (115). The expression of NRs, such as FXR, have a negative correlation with NAFLD severity at the transcriptional level (116). Schiöth et al (117) reported that additional NAFLD-induced modulating effects on type 2 diabetes were associated with altered composition and concentrations of bile acids, which have a continuous effect on FXR-dependent regulatory pathways.

FXRs are abundantly expressed in the liver and their upregulation controls glucose, lipid, cholesterol and bile acid homeostasis and inflammation to reduce NAFLD. Firstly, FXRs inhibit hepatic gluconeogenesis and peripheral insulin sensitivity to reduce plasma glucose concentrations. Secondly, FXRs suppress bile acid synthesis by inhibiting cholesterol 7 α-hydroxylase (CYP7A1), the rate-limiting enzyme for the synthesis of bile acids from cholesterol (118). The interaction between bile acids and FXRs leads to a reduction in triglyceride concentrations in plasma and, as a consequence, inhibits hepatic lipid deposition (119).

NPC1L1, which reabsorbs cholesterol from bile and transports it to the liver, is expressed on the bile duct membrane and is an aggravating factor of NAFLD (120). NAFLD often exhibits oxidative stress, which leads to increased free-radical activity and lipid peroxidation by increasing reactive oxygen species (ROS). ROS facilitate the peroxidation of lipids, induction of cytokines, chemoattraction of inflammatory cells and HSC activation, and ultimately cause fibrosis with extracellular matrix deposition (1). NPC1L1 resists oxidative stress and endoplasmic reticulum stress (ERS) by transporting free cholesterol in hepatic cells and modulating intracellular cholesterol content, particularly in mitochondria and the endoplasmic reticulum (121). Inhibiting NPC1L1 eventually reduces cholesterol uptake and induces the recovery of key molecules involved in insulin signaling, including Akt phosphorylation and gluconeogenic enzymes (122).

Bile acid transporters are associated with the progression of NAFLD (123). Na⁺-taurocholate cotransporting polypeptide (NTCP), an influx bile acid transporter, participates in the reabsorption of bile acids (124). As a bile salt export pump, NTCP is an effluent bile acid transporter that exports bile acids. Thus, the upregulation of NTCP causes excessive bile acid concentrations in hepatocytes, which ultimately leads to cell injury (116). Therefore, blocking apical sodium-dependent bile acid transporter function improves NAFLD (125).

Intestinal ecological disruption in NAFLD has been fully documented (126-128). Intestinal flora mediates the detoxification of bile acids and the conversion of primary bile acid to secondary bile acid species, including lithocholic acid (LCA) and deoxycholic acid. Therefore, inhibiting intestinal microbiota increases hepatic binding to secondary bile acids, which leads to the increased burden of the secondary bile acid system and possible toxic effects. Vitamin D attenuates the toxic effects of LCA and may be a biomarker for the noninvasive diagnosis of NAFLD progression (126-128).

Trimethylamine-N-oxide (TMAO) is a gut-flora-dependent metabolite of choline. Chen et al (129) demonstrated that the increased risk of fatty liver disease may be caused by TMAO due to its effect of decreasing the total bile acid levels via two pathways: By decreasing the synthesis of bile acids via the inhibition of key enzyme CYP7A1 or by limiting the enterohepatic circulation of bile acids between the liver and intestines by decreasing organic anion transporter and multidrug resistance protein family protein expression (130). Therefore, TMAO may possibly affect hepatic TG levels and reverse the direction of cholesterol transport and glucose and energy homeostasis by altering the synthesis and transport of bile acids, indicating that it is a potential risk factor for fatty liver disease (129).

Ferritin is the primary iron-storage protein in the majority tissues such as in the liver and brain, and increased expression of ferritin indicates iron overload in tissues and circulation. Patients with high ferritin have more severe NASH and advanced fibrosis compared with patients with normal ferritin (140,141). Ferritin can be decreased by restricted calorie and iron intake (142). Serum ferritin concentrations are commonly elevated in patients with NAFLD due to increased iron stores, fat content, oxidative stress, ERS, systemic inflammation and genetic background (143,144).

Hepcidin is the main iron-regulating peptide (145). Decreased hepcidin expression is frequently observed in patients with NAFLD and is considered to be a cause of hepatic iron overload (146). Furthermore, hepcidin suppresses iron influx from diets and efflux from macrophages by inhibiting ferroportin, determining oxidative stress and driving transformation of macrophages into foam cells (147).