Clinical Science

Review article

Metabolic disturbances in non-alcoholic fatty liver disease

Christopher D. Byrne, Rasaq Olufadi, Kimberley D. Bruce, Felino R. Cagampang, Mohamed H. Ahmed


NAFLD (non-alcoholic fatty liver disease) refers to a wide spectrum of liver damage, ranging from simple steatosis to NASH (non-alcoholic steatohepatitis), advanced fibrosis and cirrhosis. NAFLD is strongly associated with insulin resistance and is defined by accumulation of liver fat >5% per liver weight in the presence of <10 g of daily alcohol consumption. The exact prevalence of NAFLD is uncertain because of the absence of simple non-invasive diagnostic tests to facilitate an estimate of prevalence. In certain subgroups of patients, such as those with Type 2 diabetes, the prevalence of NAFLD, defined by ultrasound, may be as high as 70%. NASH is an important subgroup within the spectrum of NAFLD that progresses over time with worsening fibrosis and cirrhosis, and is associated with increased risk for cardiovascular disease. It is, therefore, important to understand the pathogenesis of NASH and, in particular, to develop strategies for interventions to treat this condition. Currently, the ‘gold standard’ for the diagnosis of NASH is liver biopsy, and the need to undertake a biopsy has impeded research in subjects in this field. Limited results suggest that the prevalence of NASH could be as high as 11% in the general population, suggesting there is a worsening future public health problem in this field of medicine. With a burgeoning epidemic of diabetes in an aging population, it is likely that the prevalence of NASH will continue to increase over time as both factors are important risk factors for liver fibrosis. The purpose of this review is to: (i) briefly discuss the epidemiology of NAFLD to describe the magnitude of the future potential public health problem; and (ii) to discuss extra- and intra-hepatic mechanisms contributing to the pathogenesis of NAFLD, a better understanding of which may help in the development of novel treatments for this condition.

  • diabetes
  • inflammation
  • mitochondrion
  • non-alcoholic fatty liver disease (NAFLD)
  • non-alcoholic steatohepatitis (NASH)
  • obesity


NAFLD (non-alcoholic fatty liver disease) refers to a wide spectrum of liver damage, ranging from simple steatosis to steatohepatitis, advanced fibrosis and cirrhosis. NAFLD is strongly associated with insulin resistance and is defined by accumulation of liver fat >5% per liver weight in the presence of <10 g of daily alcohol consumption. The characteristic histology of NAFLD resembles that of alcohol-induced liver injury, but occurs in people who consume minimal alcohol. NAFLD is regarded as the most common cause of increased liver enzymes in the U.S.A., associated with Type 2 diabetes, obesity and hyperlipidaemia [1]. The reported prevalence of obesity with NAFLD varies between 30 and 100%, whereas the prevalence of NAFLD within Type 2 diabetes varies between 10 and 75%. In routine clinical practice, most cases of fatty liver disease are attributable to alcohol excess; however, fatty liver disease can also occur in association with a wide range of toxins, drugs and diseases, such as morbid obesity, cachexia, Type 2 diabetes, hyperlipidaemia and after jejunoileal bypass surgery. As important risk factors for NAFLD, such as obesity and Type 2 diabetes, are increasing in prevalence it is likely that there will be a marked increase in NAFLD with important consequences for healthcare providers.

NAFLD is rapidly becoming an important public health problem. Undiagnosed, this condition may progress silently and results in cirrhosis, portal hypertension and liver-related death in early adulthood. Interestingly, NAFLD is a hepatic component of the metabolic syndrome and is an independent risk factor for CVD (cardiovascular disease). Importantly, NAFLD is associated with an increased risk of all-cause death and predicts future CVD events, independently of age, gender, LDL (low-density lipoprotein)-cholesterol, smoking and the cluster of features of the metabolic syndrome. Currently, there are no biochemical markers for NAFLD, and an increase (or decrease) in ALT (alanine aminotransferase) is often used as a biochemical marker to monitor progression (or amelioration) of NAFLD (despite the fact that ALT concentrations can be misleading and do not reflect the severity or the outcome). Mass screening for significant liver injury in patients with NAFLD will be an important medical challenge in the years to come because of the epidemics of obesity and diabetes.


NAFLD is very common and occurs in individuals of all ages and ethnic groups. Adjusted prevalence of NAFLD using MRS (magnetic resonance spectroscopy) in 2287 individuals (U.S.A.) was found to be approx. 34%, and 90% of these cases of NAFLD were attributed to non-alcoholic causes [2]. Furthermore, the Dallas Heart Study, which was a population-based cohort study performed in an ethnically diverse community in U.S.A. using MRS, reported that the prevalence of NAFLD is approx. 33.6% [3]. On the other hand, the use of ultrasound in the Dionysos nutrition and liver disease study in Italy found that the prevalence of NAFLD was approx. 25% and was associated with most features of the metabolic syndrome [4]. Findings from the same study showed that the prevalence of steatosis was increased in heavy drinkers {46.4% [95% CI (confidence interval), 34–59%]} and obese persons [B75">75.8% (95% CI, 63–85%)] compared with controls [16.4% (95% CI, 8–25%)]. Steatosis was found in 94.5% (95% CI, 85–99%) of obese heavy drinkers. Compared with controls, the risk for steatosis was 2.8-fold higher (95% CI, 1.4–7.1-fold) in heavy drinkers, 4.6-fold (95% CI, 2.5–11.0-fold) in obese people, and 5.8-fold (95% CI, 3.2–12.3-fold) in people who were obese and drank heavily. In heavy drinkers, obesity increased the risk of steatosis 2-fold (95% CI, 1.5–3.0-fold; P<0.001), but heavy drinking was associated with only a 1.3-fold (95% CI, 1.02–1.6-fold) increase in risk in obese persons (P=0.0053). It was concluded that steatosis is frequently encountered in healthy persons and is usually present in obese persons who drink more than 60 g of alcohol per day. Steatosis is more strongly associated with obesity than with heavy drinking, suggesting a greater role of overweight than alcohol consumption in accumulation of fat in the liver [4]. It is likely that there is an additive effect of obesity and alcohol consumption to worsen the NAFLD phenotype. Moreover, these findings raise important questions about the safety of any alcohol consumption in obese people who are insulin-resistant because obesity and alcohol are both likely to contribute to the NAFLD phenotype. Importantly, the prevalence of NAFLD as determined by ultrasound increased from 16.4% among individuals with normal BMI (body mass index) to 75.8% among obese people. The prevalence of NAFLD is even higher with morbid obesity and among morbidly obese undergoing bariatric surgery the prevalence may be as high as 96% [5]. Interestingly, the prevalence of NAFLD in the Japanese population rises with increasing degrees of hyperglycaemia, being approx. 27% in people with normal fasting glucose levels, increasing to 43% among those with impaired fasting glycaemia, and 62% among newly diagnosed diabetes [6]. On the other hand, in non-diabetic and non-obese individuals, the OR (odds ratio) for metabolic disorders (insulin resistance, hypertriacylglycerolaemia and hyperuricaemia) in subjects with NAFLD, compared with those without NAFLD in the normal-weight group, was higher than that in the overweight group. Multiple logistic regression analysis showed that gender, waist circumference, TAG [triacylglycerol (triglyceride)] level and insulin resistance were independently associated with NAFLD in the normal-weight group [7].

Among the Asian population (known to have high visceral fat), the prevalence of NAFLD using ultrasound has been estimated to vary between 5 and 40% [8]. In Japan, it has been estimated that the prevalence of increased ALT attributable to NAFLD has increased from less than 10% in 1984 to 25% in 2001 [9]. In China, the prevalence of NAFLD has been estimated to vary from 5 to 24%, and this variation may be attributed to lifestyle differences between rural and urban populations [10,11]. An increase in the prevalence of obesity and diabetes may also have resulted in an increase in the prevalence of NAFLD in the Middle East. However, in a study by el-Hassan et al. [12] in 1992, the prevalence of NAFLD was approx. 10%, whereas in contrast a small study in diabetic Saudi individuals has suggested the prevalence of NAFLD is approx. 55% [13].

Children may also develop NAFLD. In a retrospective review of 742 children between the ages of 2 and 19 years who had an autopsy performed from 1993 to 2003, fatty liver was present in 13% of subjects [14]. In a small study of 44 obese children, aged 6–16 years, with a BMI above the 97th centile, hepatic fat content was measured by phase-contrast MRI (magnetic resonance imaging) and an increased hepatic fat fraction was identified in 14 subjects (31.8%) [15].

There are few studies of the natural history of NAFLD; however, recent findings strongly suggest NAFLD is not a harmless condition and depends critically on disease stage. A Danish study (215 patients with biopsy-proven NAFLD) concluded that NAFLD has a benign course without excess mortality [16]. This view was endorsed further by 10-year follow-up data from the Dionyosis study [17]. However, a recent large study by Adams et al. [18] in 420 patients with NAFLD, between 1980 and 2000, concluded that the overall death rate is higher than expected [standardized mortality ratio, 1.34 (95% CI, 1.003–1.76); P=0.03], and this increase in mortality was associated with age, impaired fasting glucose/diabetes and cirrhosis. Importantly, Adams et al. [18] confirmed the early observation of both the Danish and Dionyosis studies that simple steatosis had benign natural history. Generally speaking, the natural history of NASH (non-alcoholic steatohepatitis) is associated with the possibility of progression to cirrhosis. Progression of liver fibrosis was found in one-third of 106 NASH patients 4.3 years after the first liver biopsy, and obesity and BMI were the only associated factors with such progression [19]. Furthermore, in a large study of the natural history of the disease, follow-up biopsies were performed after 12–20 years in 772 patients with NAFLD [20]. Of 132 patients for whom data were complete, 4% of subjects with fatty liver progressed to cirrhosis compared with 22% of patients with NASH and fibrosis on the first biopsy [20]. All liver-related morbidity and mortality were also noted in the later group with NASH, rather than simple hepatic steatosis. Thus the evidence, albeit rather limited, suggest that NASH is not a harmless condition. Importantly, Adams et al. [18] showed that the 8% liver-related mortality in biopsy-proven NASH is similar to the 10% mortality reported by Matteoni and co-workers [20], and there is general agreement that once cirrhosis develops in patients with NAFLD the prognosis is poor [1823].


At present, a liver biopsy remains the only reliable way to diagnose NAFLD and establish the presence of fibrosis. The sampling variability has the potential to alter significantly the diagnosis and staging of NAFLD. It is not practical to offer liver biopsy as a test for the diagnosis of NAFLD, as the prevalence of the condition would overwhelm service provision. Consequently non-invasive markers of NAFLD are urgently needed that can be used for diagnosis and monitoring responses to therapy. Table 1 shows a summary of biochemical markers and the potential advantages and disadvantages of different tests for identification of NAFLD. As can been seen from Table 1, no single biochemical test has, to date, shown sufficient sensitivity and specificity to be used in clinical practice for the diagnosis or monitoring of NAFLD.

View this table:
Table 1 Summary of the use of simple biochemical markers in NAFLD

CDT, carbohydrate-deficient transferrin; CK-18, cytokeratin 18; hsCRP, high-sensitivity C-reactive protein; oxLDL, oxidized LDL; RT–PCR, reverse transcriptase–PCR; TBARS, thiobarbituric acid-reacting substances.

It is beyond the scope of this review to discuss biomarkers for NAFLD in detail; however, it is important to realize that many studies investigating the aetiology and pathogenesis and investigating potential treatments for NAFLD have inevitably not been able to use liver biopsy, but have used proxy markers for the disease. This unfortunate problem has made it very difficult to establish which risk factors are aetiologically linked to the different components of the liver disease and which treatments improve liver fat, or liver inflammation or liver fibrosis within the spectrum of NAFLD. To date, these difficulties have impeded progress in understanding the aetiology of NAFLD in humans and have hampered progress in assessing new treatments for the condition. Despite these problems, progress is being made in developing non-invasive diagnostic markers for investigating aetiology and for testing new treatments for NAFLD. Recently, the ELF score (European Liver Fibrosis score) has been tested in a small subgroup of patients with NAFLD recruited with liver fibrosis [24]. The ELF score comprises three simple biochemical measurements fitted into a proprietary algorithm and shows promise with good sensitivity and specificity for NASH with fibrosis, although further research is needed. Another simpler algorithm has recently been developed and published in full utilizing simple anthropometric measurements and biochemical tests [25]. This test is also showing considerable promise with good sensitivity and specificity for NASH. A concentrated research initiative is now needed to address the pressing need of sensitive specific tests for NAFLD. The challenge remains to establish biomarkers that are simple, reproducible, inexpensive and with high sensitivity and specificity for NAFLD and can differentiate simple steatosis from NASH. Identification of simple inexpensive biomarkers would facilitate achieving reliable estimates of prevalence worldwide and would provide diagnostic tools for the monitoring of responses to therapeutic interventions.

Ultrasound, CT (computer tomography) scanning and MRI have all been used in the diagnosis of NAFLD. Ultrasound has a sensitivity of 89% and specificity of 77% and is commonly used in a clinical practice setting. Quantitative assessment of fatty infiltration is best achieved with MRI [1].


Recent findings now suggests that NAFLD may also be linked to increased CVD risk in Type 2 diabetes. CVD risk varies markedly between individuals with Type 2 diabetes and we have shown that, using traditional algorithms for risk prediction, CVD events may be underestimated by approximately one-third in Type 2 diabetes [26]. We have also shown that there is a >4-fold and statistically significant increase in risk of CHD (coronary heart disease) death if patients have all five features of the metabolic syndrome [central obesity, high TAG, low HDL (high-density lipoprotein)-cholesterol, high BP (blood pressure) and insulin resistance] compared with having only one feature (e.g. hyperglycaemia in isolation) [27]. Importantly, given the heterogeneity of cardiovascular risk in people with Type 2 diabetes, we have shown that patients with NAFLD and obesity are markedly more insulin-resistant in muscle and fat than those subjects with obesity alone [28].

Prospective studies have reported associations between increased liver enzymes [particularly the serum GGT (γ-glutamyltransferase) level as surrogate markers of NAFLD) [29,30] and the occurrence of CVD events in both non-diabetic subjects and Type 2 diabetic patients. In a study of 14874 middle-aged Finnish men and women, a small increase in GGT levels were independently associated with an increased risk of ischaemic stroke in both sexes [31]. Among 7613 middle-aged British men followed for 11.5 years, increased GGT levels were independently associated with a significant increase in mortality from all causes and from CHD [32]. Recent epidemiological studies have suggested the increase in ALT may be linked to increase in risk of CVD. The Hoorn Study is a population-based cohort of Caucasian men and women aged 50–75 years at baseline. The 10-year risk of all-cause mortality, fatal and non-fatal CVD and CHD events in relation to ALT baseline was assessed in 1439 subjects [33]. Subjects with prevalent CVD/CHD and missing data were excluded from these analyses. When compared with the first tertile, the age- and gender-adjusted hazard ratios (95% CIs) for all-cause mortality, CVD events and CHD events were 1.30 (0.92–1.83), 1.40 (1.09–1.81) and 2.04 (1.35–3.10) respectively, for subjects in the upper tertile of ALT. After adjustment for components of the metabolic syndrome and traditional risk factors, the association between ALT and CHD events remained significant for subjects in the third tertile relative to those in the first tertile, with a hazard ratio (95% CI) of 1.88 (1.21–2.92). Thus the Hoorn Study shows that ALT predicts cardiovascular events independently of traditional risk factors and the features of the metabolic syndrome [33], suggesting that NAFLD is associated with CHD independently of other features of the metabolic syndrome.

Interestingly, another study from the U.S.A. has examined the association between elevated serum ALT activity and the 10-year risk of CHD as estimated using the FRS (Framingham risk score) [34]. Among participants without viral hepatitis or excessive alcohol consumption, those with increased ALT activity [>43 IU (international units)/l] (n=267) had a higher FRS than those with normal ALT activity (n=7259), both among men [mean difference in FRS, 0.25 (95% CI, 0.07–0.4); hazard ratio for CHD, 1.28 (95% CI, 1.07–1.5)] and women [mean difference in FRS, 0.76 (95% CI, 0.4–1.1); hazard ratio for CHD, 2.14 (95% CI, 1.5–3.0)]. The ALT threshold for an increased risk of CHD was higher in men (>43 IU/l) than in women (>30 IU/l) [34]. Recently, the FIBAR (Firenze Bagno a Ripoli) study concluded that increased GGT or AST (aspartate aminotransferase) is an independent predictor of CVD. An increase in GGT levels above the reference range, or also in the upper reference range, was also an independent predictor of incident diabetes [35].

In a recent study by Targher et al. [36] (n=2839) NAFLD and CVD were the main outcome measures in Type 2 diabetic patients. The authors showed that the unadjusted prevalence of NAFLD was 69.5% among participants, and NAFLD was the most common cause (81.5%) of hepatic steatosis detected by ultrasound. The prevalence of NAFLD increased with age (65.4% among participants aged 40–59 years, and 74.6% among those aged ≥60 years; P<0.001) and the age-adjusted prevalence of NAFLD was 71.1% in men and 68% in women. NAFLD patients had a higher age (P<0.001) and gender-adjusted prevalence of coronary (26.6 compared with 18.3%), cerebrovascular (20.0 compared with 13.3%) and peripheral (15.4 compared with 10.0%) vascular disease than their counterparts without NAFLD. In logistic regression analysis, NAFLD was associated with prevalent CVD, independent of classical risk factors, glycaemic control, medication and the features of the metabolic syndrome [36].

The Valpolicella Heart Diabetes Study is a prospective nested case-control study in 2103 Type 2 diabetic patients who were free of diagnosed CVD at baseline. During 5 years of follow-up, 248 participants (case subjects) subsequently developed non-fatal CHD (myocardial infarction and coronary revascularization procedures), ischaemic stroke or cardiovascular death. After adjustment for age, gender, smoking history, diabetes duration, HbA1c (glycated haemoglobin), LDL-cholesterol, liver enzymes and use of medication, the presence of NAFLD was significantly associated with an increase in CVD risk [OR, 1.84 (95% CI, 1.4–2.1); P<0.001]. Additional adjustment for the metabolic syndrome (as defined by National Cholesterol Education Program Adult Treatment Panel III criteria) appreciably attenuated, but did not abolish, this association [OR, 1.53 (95% CI, 1.1–1.7); P=0.02] [37]. Thus the above two large studies have demonstrated that NAFLD is associated with an increased risk of future CVD events among Type 2 diabetic individuals and, importantly, this association was independent of classical risk factors, liver enzymes and the metabolic syndrome.

Studies have shown a link between NAFLD and increased carotid IMT (intima-media thickness). Fracanzani et al. [38] concluded (in a series of normal and NAFLD subjects) that independent risk predictors of increased IMT were the presence of hepatic steatosis (OR, 6.9), age (OR, 6.0) and increased systolic BP (OR, 2.3). More interestingly, Targher et al. [39] suggested that the severity of liver histopathology among 85 NAFLD patients was strongly associated with early carotid atherosclerosis, independent of age, gender, BMI, smoking, LDL-cholesterol, insulin resistance and the presence of the metabolic syndrome. In addition, a large population study has shown that NAFLD is associated with an increase in IMT [40]. On the other hand, in 100 dietcontrolled Type 2 diabetic individuals, the significant increase in carotid IMT in the presence of NAFLD has largely been explained by insulin resistance. Currently, it is not clear how NAFLD leads to an increase in risk of CVD. One explanation may be the fact that NAFLD appears to be associated with all metabolic risk factors that are features of the metabolic syndrome. In addition, NAFLD may accelerate the progression of dyslipidaemia, insulin resistance, atherosclerosis, endothelial dysfunction, inflammation and oxidative stress. Interestingly, NAFLD may also be associated with a detrimental effect on other organs that may have a direct or indirect influence on CVD, or organs that may accelerate presentation of CVD, at least in people with Type 2 diabetes. For example, NAFLD patients with Type 2 diabetes had higher (P<0.001) age- and gender-adjusted prevalence rates of both non-proliferative (39 compared with 34%) and proliferative/laser-treated (11 compared with 5%) retinopathy, and CKD (chronic kidney disease) (15 compared with 9%) than counterparts with Type 2 diabetes but without NAFLD. In these subjects, using logistic regression analysis, NAFLD was shown to be associated with increased rates of CKD [OR, 1.87 (95% CI, 1.3–4.1); P=0.020] and proliferative/laser-treated retinopathy [OR, 1.75 (95% CI, 1.1–3.7); P=0.031] independently of age, gender, BMI, waist circumference, hypertension, diabetes duration, HbA1c, lipids, smoking status and medication use [41]. Furthermore, in that study, NAFLD was associated with an increased incidence of CKD [42], independent of gender, age, BMI, waist circumference, BP, smoking, diabetes duration, HbA1c, lipids, baseline estimated GFR (glomerular filtration rate), microalbuminuria and medication (hypoglycaemic, lipid-lowering, antihypertensive or antiplatelet drugs). Interestingly, NAFLD has been shown to be associated with the development of CKD in Korean individuals [crude relative risk, 2.18 (95% CI, 1.75–2.71], and this relationship remained significant even after adjustment for age, GFR, TAG and HDL-cholesterol [aRR (adjusted relative risk), 1.55 (95% CI, 1.23–1.95)] [43]. The association between NAFLD and incident CKD was evident in the NAFLD group with elevated serum GGT [aRR, 2.31 (95% CI, 1.53–3.50)], even after adjustment for age, GFR, TAG and HDL-cholesterol, but not in the NAFLD group without elevated GGT [aRR, 1.09 (95% CI, 0.79–1.50); P=0.008 for interaction)] [43].


From the evidence presented above, it is important to understand the pathogenesis of NASH because of the risk of progression to more severe end-stage liver disease. It is also important to understand the relationship between NASH and CVD. With respect to the latter, it is plausible that factors outwith the liver are contributing to NASH and also to CVD. Changes in the liver associated with NASH may then compound the problem further and amplify risk for CVD. For these reasons, we will discuss the contribution of factors outwith the liver to the pathogenesis of NAFLD and we will discuss mechanisms within the liver contributing to NAFLD.

There is a strong link between insulin resistance and excessive deposition of TAG in hepatocytes, which is the hallmark for diagnosis of NAFLD [4446]. The excessive/ectopic fat depositions in the liver could be due to increased fatty acid delivery from adipose tissue, increased synthesis of fatty acid via the de novo pathway, increased dietary fat, decreased mitochondrial β-oxidation, decreased clearance of VLDL (very-LDL) particles or these factors in combination. It is still a matter of debate whether insulin resistance causes NAFLD or whether excessive accumulation of TAG, or precursors on the synthetic pathway, precede and promote insulin resistance [47].

Extrahepatic mechanisms contributing to the pathogenesis of NAFLD

Lipolysis and NEFAs [non-esterified fatty acids (‘free fatty acids’)]

There is evidence that increased delivery of NEFAs to the liver from the peripheral (adipose) tissue is fundamental to the development of NAFLD. In support of this notion, studies in humans and rodents have shown that increased NEFA delivery from adipose tissue is a significant source of fat accumulation in hepatocytes [48]. Some investigators [49] have reported that approx. 60% of fat deposited in hepatocytes is generated from adipose tissue sources. In insulin-resistant subjects, there is a failure of insulin-mediated suppression of HSL (hormone-sensitive lipase), resulting in uncontrolled lipolysis in the adipose tissue [48,50] (Figures 1A and 1B). HSL-knockout mice have decreased plasma NEFA and TAG concentrations, and have increased hepatic insulin sensitivity [51,52]. Peripheral (subcutaneous) fat constitutes a major proportion of the fat mass; however, it is not an absolute requirement for developing fatty liver. Patients with lipodystrophy, who are also insulin-resistant with no peripheral and intra-abdominal fat, develop fatty liver and severe hepatic insulin resistance [53]. In this situation, the stimulus for fatty liver may be increased NEFA flux to the liver stimulating lipogenesis, combined with inadequate compensatory fat oxidation, leading to fat accumulation over time. Furthermore, there is some contribution to NEFA flux to the liver from intra-abdominal fat in insulin-resistant subjects [54]. Intra-abdominal fat is strongly associated with reduced insulin sensitivity even in lean subjects [55], and has been shown to correlate positively with hepatic fat and hepatic insulin resistance [5658].

Figure 1 Interactions between liver, muscle and adipose tissue in the control of VLDL secretion

(A) Normal state; (B) insulin-resistant state, showing an increased flux of NEFAs from adipose tissue to the liver. FFA, NEFA; LPL, lipoprotein lipase.

Dietary macro- and micro-nutrients

Saturated fatty acids

The role of diet in the pathogenesis of NAFLD has been investigated in humans and animal models [59,60]. Subjects on a high-fat diet develop fatty liver. Subjects on a low-fat/high-carbohydrate diet also develop fatty liver via increased de novo fatty acid synthesis [61]. In addition to the above effect, high dietary fat intake is associated with obesity and insulin resistance. As a consequence of insulin resistance, there is impaired suppression of lipolysis by insulin, leading to increased NEFA delivery to the liver [62,63]. There is also reduced glucose uptake in the fed state by adipose tissue and skeletal muscle, resulting in hyperglycaemia and diversion of glucose to the hepatic de novo pathway [64].

An increase in hepatic fatty acid is capable of exacerbating hepatic insulin resistance at the insulin receptor level [65]. The mechanism has not been fully elucidated but is thought to be due to translocation of the PKCδ (protein kinase Cδ) isoform from the cytosol to the membrane compartment, leading to impaired IRS (insulin receptor substrate)/PI3K (phosphoinositide 3-kinase) activation [55]. High dietary saturated fatty acids have been shown to be associated with insulin resistance, NAFLD and CVD [66]. In rats, high dietary saturated fatty acids have been shown to promote ER (endoplasmic reticulum) dysfunction and hepatocyte injury [67]. In another study in humans [68], dietary consumption of saturated fat was reported to be higher in overweight patients with NASH compared with age- and BMI-matched control subjects (14 compared with 10% respectively).

n−3 PUFAs (polyunsaturated fatty acids)

There is increasing interest in the role of n−3 PUFAs in NAFLD. The benefits of DHA (docosahexaenoic acid) and EPA (eicosapentaenoic acid) following the high consumption of fish oil have been shown in insulin-resistant animal models [66], including (i) decreased plasma NEFA, TAG, glucose and insulin concentrations; (ii) decreased de novo lipogenesis, VLDL export and TAG concentrations; (iii) increased utilization and storage of glucose by skeletal muscle; (iv) increased insulin-mediated glucose uptake by adipose tissue; and (v) decreased adipocyte size and visceral fat content. Other potential benefits of PUFAs have also been reported in other animal experiments and these include: (i) decreased Kupffer cell activation [69]; (ii) decreased NF-κB (nuclear factor κB) macrophage activation [7072]; and (iii) decreased gene expression and production of inflammatory cytokines [7073]. Similar findings have also been reported in humans. In a small study of patients with NAFLD [74], there was decreased TAG and glucose concentrations, liver enzyme activities and hepatic steatosis following consumption of 1g of fish oil for 12 months. Another study [75] showed similar results and decreased TNF-α (tumour necrosis factor–α) concentrations after 2 g of daily supplementation with fish oil for 6 months. The authors [75] suggested that consumption of n−3 PUFAs could ameliorate NAFLD by improving steatosis, inflammation and hepatocyte injury.

There is now some evidence that dietary n−3 fatty acids have a role in the pathogenesis of NAFLD and could be a potential therapeutic target [76]. For example, there is an association between PUFA deficiency and the development of hepatic steatosis in animal models [66]. SREBP-1c (sterol-regulatory-element binding-protein-1c) is a key transcription factor for de novo lipogenesis (see below) and is negatively regulated by n−3 PUFAs [77]. Consequently, the activities of key enzymes for fatty acid and TAG synthesis are down-regulated with decreased hepatic fat deposition. PUFAs also negatively regulate the activity of a glucose-responsive transcription factor [ChREBP (carbohydrate-responsive-element-binding protein); see below] by increasing ChREBP mRNA decay and disrupting translocation of ChREBP from the cytosol to the nucleus [78]. As a result of these effects, it is plausible that high concentrations of dietary n−3 PUFA supplementation could be a promising treatment for the management of NAFLD, and randomized controlled trials in this area are urgently needed.

Carbohydrate and protein

Excessive consumption of simple sugars has been reported to promote the development of NAFLD. In one report, there was a 2–3-fold increase in de novo lipogenesis in both lean and obese subjects following a 4 day overfeeding with either a glucose or sucrose drink [79]. Similarly, high dietary intake of fructose also increases de novo lipogenesis in both animal models and humans [80,81].

The precise role of dietary protein in the pathogenesis of NAFLD is uncertain. There is a need for further investigation as whether there is a causal relationship between excess dietary protein and NAFLD.


The relationship between micronutrients and NAFLD is now attracting considerable interest. PC (phosphatidylcholine), the most abundant phospholipid in the mammalian cell, is formed from dietary-dependent and -independent pathways. In the presence of CCT (CTP:phosphocholine cytidylyltransferase) and CPT (CDP-choline:1,2 diacylglycerol choline phosphotransferase) enzymes, PC is formed from dietary choline and DAG (diacylglycerol). In the dietary-independent pathway, PC is formed from SAM (S-adenosylmethionine) and PE (phosphatidylethanolamine) in the presence of PEMT (phosphatidylethanolamine N-methyltransferase), suggesting that a supply of methyl donors is important for the synthesis of PC (Figure 2). There is some evidence that PC, from these different pathways, has a distinct molecular composition with different physiological functions dependent on the biochemical pathway leading to PC generation [82,83]. PC from the dietary-independent pathway consists mainly of long-chain PUFAs, whereas PC of dietary origin consists of medium-chain saturated fatty acids [82]. PEMT activity has been shown to be important in the mobilization of essential fatty acids from liver [83].

Figure 2 Potential control of dietary methyl donor regulation of gene expression relevant to NAFLD

Dietary-independent pathway synthesis of PC is formed from SAM and PE in the presence of PEMT, suggesting that a supply of methyl donors is important for the synthesis of PC.

In animal models, PC generated from the PE pathway has been shown to be essential for the incorporation of neutral lipids into VLDL particles [84], and for the normal concentrations and structural composition of plasma VLDL [85,86]. Additionally, it has been shown that PEMT-knockout mice have gender-specific differences in the plasma concentrations of PC, cholesterol esters and TAGs [85]. In one study [87], PEMT-knockout mice were fed an MCD (methionine- and choline-deficient) diet resulting in significantly decreased hepatic PC concentrations, severe liver damage and death. Animal models (mice) that are fed an MCD diet develop NASH and this has provided a good model for studying the liver in NASH, although the mice tend to be thin and are insulin-sensitive.

Dietary deficiency of methionine in animal models is associated with reduced glutathione deficiency, with secondary free-radical-induced hepatocyte damage and inflammation [83]. As PC is essential for the synthesis of VLDL particles, PC deficiency is associated with hepatic fat deposition secondary to the impaired export of VLDL particles. PEMT-knockout mice had a significant increase in hepatic TAG and cholesterol ester concentrations [83], and there was almost a total reversal of these effects following dietary choline supplementation, perhaps suggesting a role for methyl donors in the generation of PC and NASH.

Intrahepatic mechanisms

De novo lipogenesis

Role of key enzymes

Under basal conditions, the contribution to hepatic fat from the de novo pathway is less than 5% [88]; however, under pathological conditions, hepatic de novo lipogenesis also has been reported to be a significant source of fat deposited in the liver (approx. 30%) [47,48]. In the de novo pathway, dietary glucose is converted into acetyl-CoA by L-PK (liver-specific pyruvate kinase). Acetyl-CoA is then converted into malonyl-CoA and, subsequently, to fatty acids by actions of ACC (acetyl-CoA carboxylase) and FAS (fatty acid synthase) respectively.


The activity of ACC is regulated by the cell energy status, insulin and the availability of NADPH. ACC is inactivated by phosphorylation via AMPK (AMP-activated kinase), resulting in decreased conversion of acetyl-CoA into malonyl-CoA, and PP2A (protein phosphatase 2A; an insulin-regulated enzyme) reverses this action by dephosphorylation of ACC. When cell energy is low with a high AMP/ATP ratio, ACC is kept in its inactive phosphorylated form and, consequently, acetyl-CoA is channelled to β-oxidation and ketogenesis for energy production. ACC is also under hormonal control by insulin and glucagon. Insulin stimulates PP2A, which dephosphorylates ACC thereby activating this enzyme and promoting lipogenesis.

There are two isoforms of ACC namely ACC1 and ACC2. ACC1, which is in the cytosol, is highly expressed in liver and adipose tissue. ACC2, the mitochondrial isoform, is expressed mainly in muscle and, to a lesser extent, liver. ACC2 is involved in the negative regulation of mitochondrial β-oxidation, hence increased β-oxidation seen in ACC2-knockout animals. This is due to decreased malonyl-CoA, which derepresses CPT-1 (carnitine palmitoyltransferase-1) activity [89]. Fatty acid synthesis takes place in the cytosol and, therefore, only the ACC1 isoform is important in the de novo pathway for fatty acid synthesis. This occurs because there is no access to malonyl-CoA produced in the mitochondria by ACC2 [90,91]. In a LACC1KO (liver-specific ACC1-knockout) model, Mao et al. [92] showed a 70% reduction in hepatic malonyl-CoA relative to control, a 50% reduction in de novo lipogenesis and a 40% decrease in TAG concentrations. Paradoxically, this mouse model was not protected from high-fat high-calorie diet-induced obesity, fatty liver and insulin resistance, despite a significant decrease in fat deposition. This effect may be due to increased compensatory activity of liver-specific ACC2 [93]. In contrast, in vitro ASO (antisense oligonucleotide) inhibition of both ACC1 and ACC2 led to a significant decrease in hepatic malonyl-CoA concentrations, decreased hepatic lipids and increased hepatic insulin sensitivity [94], suggesting that targeting ACC activity could improve insulin resistance and ameliorate fatty liver.


FAS is the last key enzyme in de novo fatty acid synthesis and a total lack of FAS activity due to gene deletion is not compatible with intra-uterine life. Although this is a key lipogenic enzyme, FASKOL (liver-specific FAS knockout) mice were also not protected from developing hepatic steatosis. There was a 3-fold increase in malonyl-CoA, which would presumably inhibit CPT-1 and decrease mitochondrial β-oxidation, leading to increased hepatic NEFAs [95]. Increasing liver fatty acids would presumably then lead to TAG synthesis and fat accumulation, worsening the NAFLD phenotype, rather than ameliorating the phenotype.

SCD-1 (stearoyl-CoA desaturase-1):

MUFAs (mono-unsaturated fatty acids) are the major components of TAGs, cholesterol esters and membrane phospholipids. The synthesis of MUFAs is catalysed by SCD-1. Therefore modulating the activity of this enzyme may significantly decrease TAG synthesis. In a previous study [96], SCD-1-knockout mice had decreased de novo lipogenesis with increased mitochondrial fat oxidation. These mice were protected from diet-induced obesity, fatty liver and insulin resistance when fed a high-fat and high-calorie diet. ASO inhibition of SCD-1 in liver and adipose tissue also showed similar results with the SCD-1-knockout mice [97]. In another study [98], liver-specific SCD-1-knockout mice were shown to have decreased gene expression for key lipogenic enzymes (ACC and FAS), with decreased de novo lipogenesis, and were also protected from diet-induced hepatic steatosis. Additionally, the reduction in SCD-1 activity had a marked decrease in the nuclear content of two key transcription factors for lipogenesis SREBP-1c and ChREBP (see text below).

DGAT-2 (diacylglycerol acyltransferase-2):

There is now increasing interest in the role of DGAT-2, the enzyme that catalyses the final step in TAG synthesis. Theoretically, inhibition of DGAT-2 should prevent the synthesis of TAG and NAFLD, although this appears not completely true as, in a recent experiment in db/db mice [99], inhibition of DGAT-2 with ASOs produced surprising results. Feeding mice an MCD diet caused the development of a fatty liver after 4 weeks. Although there was marked improvement in hepatic steatosis following inhibition of DGAT-2 with ASO, paradoxically, there was worsening/progressive inflammation, hepatocyte injury and fibrosis in these mice. Histological sections of the liver in the DGAT-2 arm revealed marked inflammatory changes and hepatocyte necrosis. Additionally, there was worsening liver fibrosis with raised mRNA for most of the markers of fibrosis and hepatic stellate cell activation. Although an MCD diet has been shown to increase the expression of fibrosis markers [TGF-β1 (transforming growth factor-β1), α-SMA (α-smooth muscle actin), TIMP-1 (tissue inhibitor of metalloproteinases-1) and collagen], there was a failure of inhibition of most of these markers after ASO inhibition of DGAT-2. The plausible explanation for the progressive inflammation and fibrosis may be due to increased hepatic NEFAs and the generation of hepatotoxic free radicals [99]. Mice in the treatment arm had significantly higher hepatic fatty acids. Although an MCD diet inhibits mitochondrial fat oxidation, there was amplification of this effect in the ASO-inhibition mice with a diversion of fatty acids to microsomal and peroxisomal fat oxidation. Consequently, in response to increased microsomal fat oxidation, there was increased microsomal enzyme [Cyp2E1 cytochrome P450 2E1)] activity with increased (microsomal) production of ROS (reactive oxygen species). Interestingly, there was improved systemic insulin resistance after ASO inhibition of DGAT-2. There was a reduction in weight, and reduced serum NEFAs, glucose and insulin concentrations. There was also reduced TNF-α and increased serum adiponectin concentrations. In view of this evidence, the authors [99] postulated that TAG itself may not be harmful, but could instead protect the liver from lipid toxicity and hepatotoxic free-radical damage by buffering hepatic fatty acids into the synthesis of TAG. Further research in this area is clearly needed to elucidate which molecule(s) on the TAG synthetic pathway are able to trigger inflammatory pathways.

Role of key transcription factors

Evidence is now emerging that SREBP-1c is a positive transcription factor for ACC and FAS genes [100], and we have suggested that abnormalities in SREBP-1c function may play an important pathogenetic role in contributing to the NAFLD phenotype (see Figure 3) [101]. Overexpression of SREBP-1c in adult rats causes a 26-fold increase in fatty acid synthesis and fat deposition. Apart from modulating the key lipogenic enzymes (ACC and FAS), SREBP-1c also stimulates gene expression for fatty acid elongation and TAG synthesis. However, SREBP-1c is by itself insufficient for full transcription activity of fatty acid synthesis. In a knockout experiment, only a 50% reduction in fatty acid synthesis was reported following SREBP-1c knockout [102]. Current evidence indicates that the effect of insulin on lipogenesis is mediated via SREBP-1c activity [103,104].

Figure 3 Effect of dietary fatty acid intake and de novo fatty acid synthesis in regulating lipid synthesis

Potential interactions are shown between hepatic lipid metabolism, DNA methylation and gene expression, inflammation, VLDL secretion and hepatic triacylglycerol accumulation in the pathogenesis of NAFLD. FACoA, fatty acyl-CoA; UCP-2, uncoupling protein-2.

In an insulin-resistance state with hyperinsulinaemia, the transcriptional activity of SREBP-1c is up-regulated and both insulin and SREBP-1c synergistically stimulate key lipogenic genes, thereby promoting de novo fatty acid synthesis. Insulin-resistant ob/ob mice have increased concentrations of SREBP-1c and also develop spontaneous fatty liver. Increased SREPB-1c in these mice increases lipogenic gene expression, increases fatty acid deposition and accelerates TAG deposition in hepatocytes [105,106]. Administration of leptin, which opposes the action of SREBP-1c, reverses these metabolic derangements [107]. In contrast, in the fasting state, when insulin production is low but glucagon concentrations are increased, the quantity of SREBP-1c in liver is decreased, although this effect is reversed by re-feeding [108,109].

Forkhead transcription factors:

Decreased mitochondrial fatty acid oxidation is also contributory to excessive fat deposition in the liver in insulin-resistant subjects. A plausible mechanism for this effect could be mediated via phosphorylation and dephosphorylation of the transcription factor Foxa2 (forkhead box a2) [110].

Under physiological conditions, active (dephosphorylated) Foxa2 promotes mitochondrial fatty acid oxidation, whereas insulin reverses this effect via phosphorylation by IRS-1 and IRS-2. Paradoxically, in insulin-resistance, Foxa2 is still predominantly in its inactive form because of residual sensitivity to insulin in the liver resulting in decreased β-oxidation and increased hepatic fat deposition.

INSIG-1 (insulin induced gene-1), insulin signalling, PTEN (phosphatase and tensin homologue deleted on chromosome 10) and mTOR (mammalian target of rapamycin):

There is now some interest on the role of INSIG-1 and fatty liver. INSIG -1 is highly expressed in hepatocytes and adipocytes, and has been suggested to act as the ‘brake’ for lipogenesis in fat and limiting TAG deposition in hepatocytes [110a]. The gene may also play a modulatory role in preadipocyte diffentiation. It is possible that the benefits of PPAR-γ (peroxisome-proliferator-activated receptor-γ) agonists in NASH are due to the regulation of INSIG-1 by these drugs.

Under physiological circumstances, the phosphorylation of IRS, PI3K, Akt and the downstream proteins are essential for the action of insulin in insulin-sensitive organs. However uncontrolled phosphorylation would lead to amplification of insulin action with insulin hypersensitivity, whereas a lack of phosphorylation of the insulin substrate proteins could cause insulin resistance. There are key negative regulatory proteins (phosphatases) that physiologically terminate the action of insulin by dephosphorylation, and constitutive action of these proteins could lead to decreased insulin sensitivity. Other insulin signalling counter-regulatory mechanisms include: (i) serine or threonine phosphorylation of the insulin receptor; (ii) activation of tyrosine phosphatases; (iii) inhibition of insulin receptor and IRSs by the SOCS (suppressor of cytokine signalling) family; and (iv) activation of phosphoinositide phosphatases [PTEN, SHIP-2 (SH2-containing inositol phosphatase-2) and myotubularin] [111].

SHIP-2 is an insulin-signal-regulatory phosphatase, but the absence or reduced expression and activity of this enzyme has not been reported to be associated with fatty liver [112114].

PTEN is a tumour-suppressor protein, but with phosphatase activity, and has a regulatory effect on the insulin signalling pathways [111,114,115]. The substrate proteins for this enzyme are PtdIns(3,4)P2 and PtdIns(3,4,5)P3. As a phosphatase, the physiological function of PTEN is to dephosphorylate the second messengers generated by the activation of PI3K, thereby down-regulating or terminating insulin signalling downstream of PI3K [111]. Overexpression of PTEN has been shown to have inhibitory effects on insulin signalling, including decreased Akt activation and GLUT4 (glucose transporter 4) translocation to the cell membrane [116,117]. Overexpression of PTEN in muscle from obese fa/fa Zucker rats had been shown to contribute to muscle insulin resistance in these animals [118]. In contrast, down-regulation of PTEN has the opposite effect, with increased glucose uptake in fat and muscle in response to insulin [119]. In mice, liver-specific deletion of PTEN has been shown to increase insulin sensitivity, but paradoxically causes fatty liver disease and hepatocellular cancer [120,121]. The mechanism for this paradox is yet to be clarified, but there has been a number of hypotheses suggested, some of which highlight the lack of negative regulation on the insulin signalling pathways by PTEN. In PTEN-knockout mice, it has been shown that there was increased synthesis and storage of TAG in hepatocytes due to the up-regulation of PI3K/Akt activity [119,121]. As a consequence of the lack of PTEN activity, there is increased hepatocyte fatty acid uptake, increased fatty acid synthesis and increased esterification of fatty acids to TAG. Taken together, these findings suggest that decreased PTEN activity would lead to excessive fat deposition in the liver.

There is evidence that reduced or absent expression of liver-specific PTEN is associated with hepatic steatosis, inflammation, fibrosis and neoplasm [115]. In one study, PTEN-deficient mice were shown to have biochemical and histological evidence, including fibrosis of NASH, present at 40 weeks age [115]. The mechanism for NASH in this animal model has been reported to be due to increased expression of the PPAR-γ, SREBP-1c and downstream genes, including Akt and Foxo1, resulting in increased lipogenesis, inflammation and fibrosis [115]. The reason for the increase in fat deposition in the liver could be partly due to the increased expression of SREBP-1c. As SREBP-1c is a key transcription factor for lipogenesis with increased ACC, FAS and SCD-1 enzyme activities, all of these act synergistically to promote fatty acid synthesis.

As a consequence of increased PPAR-γ expression, there is also a secondary induction of key enzymes involved in mitochondrial β-oxidation. As a result, the increase in fat oxidation with a marked increase in the generation of oxidative free radicals leads to inflammation and fibrosis via activation of the NF-κB pathway [122,123]. For example, there is a 7-fold increase in the hepatic concentration of H2O2 in PTEN-deficient mice compared with the wild type [115]. There is also evidence of transcriptional down-regulation of PTEN in the liver with decreased enzyme protein levels in fatty liver disease, and this has been shown in the liver of obese insulin-resistant ZDF (Zucker diabetic fatty) rats with diabetes and hepatic steatosis [114]. This raises the question of cause or effect. Furthermore, only unsaturated NEFAs have been shown to decrease the expression of PTEN in hepatocytes, suggesting that this is a specific effect of unsaturated fatty acids. For example, the down-regulation of PTEN expression by the unsaturated fatty acid oleic acid is via the activation of mTOR and/or NF-κB pathways. Similar findings have been reported in humans in a small study of morbidly obese subjects (n=20), where there was a down-regulation of PTEN expression with a biopsy-proven decrease in immunohisto-chemistry staining for PTEN in subjects with hepatocytes steatosis of >80% [114].

In insulin-resistant subjects, there is an increase in plasma NEFA concentrations with an increase in delivery and uptake of NEFAs by hepatocytes. This results in the activation of mTOR, which in turn leads to activation of NF-κB [114]. As both mTOR and NF-κB exist as a complex, there is a subsequent dissociation and then translocation of NF-κB into the nucleus, where it down-regulates the expression of PTEN at the level of transcription (Figure 4). The precise mechanism of how NF-κB transcriptionally down-regulates PTEN is not fully understood, but could be through the sequestration of the CBP [CREB (cAMP-response-element-binding protein)-binding protein)]/p300, the transcriptional activator for PTEN [124].

Figure 4 Interaction between NEFAs, mTOR and NF-κB

Schematic diagram showing the effects of increased availability of fatty acids in hepatocytes resulting in activation of mTOR. The activation of mTOR leads to activation, dissociation and then translocation of NF-κB from the cytosol to the nucleus, where it regulates PTEN at the level of transcription. As a consequence, there is amplification of insulin signalling with increased hepatic lipogenesis. *Indicates activation of the protein. FFA, NEFA.

mTOR is a downstream target of the PI3K signalling pathway, but the activation of this protein down-regulates insulin signalling in insulin-responsive tissue [125,126]. The mechanism is through serine phosphorylation of IRS-1 [127]. The regulation of the mTOR signalling pathway is complex. This pathway is activated by hormonal- and nutrient-sensing factors, whereas inhibitory regulation is through AMPK activation. Insulin via IRS activates PI3K, which, in turn, activates Akt. Akt then activates mTOR via phosphorylation. The activation of mTOR leads to phosphorylation and activation of down-stream proteins, including p70S6 kinase and S6 ribosomal protein. As a consequence of S6 ribosomal protein phosphorylation, serine phosphorylation of IRS-1 occurs at Ser636/Ser639 [128]. Because of the serine phosphorylation of IRS-1, this key insulin substrate protein becomes unresponsive to insulin signals and, instead, undergoes rapid degradation and there is a decreased interaction between the serine-phosphorylated IRS-1, the insulin receptor (upstream) and PI3K [129], resulting in decreased insulin action (Figure 5).

Figure 5 Regulation of the mTOR signalling pathway by nutrients and insulin

Schematic diagram showing the up-regulation of mTOR by insulin and nutrient-sensing mechanisms and the down-regulation of mTOR by AMPK activation. Activation of mTOR leads to activation of the downstream proteins including S6 ribosomal protein resulting in serine phosphorylation of IRS-1 and consequent decreased insulin action.

The mTOR signalling pathway is also regulated via a nutrient-sensing mechanism. An increase in plasma leucine concentrations is a potent activator of the mTOR pathway through class III PI3K activation (class I PI3K being the substrate protein of the insulin/IRS pathway) [130]. A high-fat diet is associated with obesity, hepatic fat deposition, hepatic and whole-body insulin resistance and, in an animal experiment, the mTOR signalling pathway has been shown to be up-regulated in mice fed a high-fat diet compared with controls [131]. As insulin activates the mTOR pathway, we can assume that up-regulation of this pathway may be due to increased fasting insulin concentrations associated with high dietary fat and insulin resistance. Under normal conditions, this action of insulin would be mediated via tyrosine phosphorylation of the insulin receptor; however, in this animal experiment, there was no difference in tyrosine phosphorylation of the insulin receptor or the downstream target proteins in high-fat-fed mice and the control group. The authors [131] hypothesized that the activation of the mTOR pathway was through a poorly understood insulin/IRS-independent pathway, probably due to acute fat deposition in the liver. However, in the same report [131], the investigators were not able to show that the up-regulation of the mTOR pathway was due to acute hepatic steatosis. Pharmacological induction of acute hepatic steatosis in mice showed a 5-fold increase in hepatic TAG content, but failed to show a significant increase in the activation of mTOR and the downstream proteins, leading to the authors concluding that acute hepatic steatosis is, by itself, insufficient for sustained activation of the mTOR signalling pathway [131].

Increased fatty acid delivery has been shown to up-regulate the mTOR pathway, thereby exacerbating hepatic insulin resistance. In a recent study [132], in the absence of insulin, palmitic-acid-cultured hepatocytes had activation of the mTOR pathway with a 4-fold increase in p70S6 kinase activity. In the presence of insulin, there was an amplification of this effect. The antidiabetic agent metformin, via activation of AMPK, was shown to reverse this effect. In insulin-resistant subjects with increased fatty acid delivery to the liver and hyperinsulinaemia, we postulate that worsening insulin resistance may be due to sustained or enhanced activation of the mTOR pathway.


The gene for L-pyruvate kinase, a key component of the de novo pathway for fatty acid synthesis, is not regulated by SREBP-1c [133], but exclusively by glucose [134].

A glucose-responsive transcription factor, ChREBP, has been described and thus illustrates how glucose affects gene transcription [135]. ChREBP is controlled by glucose and regulates the translocation of this protein from cytosol to the nucleus [136]. In the nucleus, glucose promotes the binding of this protein to the carbohydrate-response element in the promoter region of target genes [137].

As part of deranged insulin signalling in obese insulin-resistant subjects, there is decreased GLUT4 translocation to the cell membrane in muscle and adipose tissue. The investigators suggested that fatty acyl-CoA and other NEFA derivatives and TNF-α impair the activation of IRS and PI3K [138,139]. Consequently, there is activation of JNK (c-Jun N-terminal kinase) pathways, leading to serine instead of tyrosine phosphorylation and insulin resistance [138,140]. ROS and TNF-α are also capable of activating JNK. The overall effect is decreased glucose uptake with persistent hyperglycaemia without or with hyperinsulinaemia. This will exacerbate de novo hepatic lipid synthesis via activation of SREBP and ChREBP being driven by insulin and glucose respectively.

LXR-α (liver X receptor-α):

ChREBP is also regulated by LXR-α, and important ligands for LXR-α are oxysterol, insulin and glucose. LXR-α is a member of the nuclear receptor family that plays an important role in lipogenesis. LXR-α exerts transcriptional control on SREBP-1c and, therefore, indirectly on ACC and FAS [141143]. ob/ob mice that are insulin-resistant develop fatty liver, although this is unlikely to be due to insulin-mediated effects on ACC and FAS. The explanation for this paradox may be that not all insulin-sensitive metabolic pathways are uniformly affected by insulin resistance. A plausible mechanism is via glucose control. The genes for the gluconeogenic enzymes PEPCK (phosphoenolpyruvate carboxykinase) and G6Ptase (glucose-6-phosphate translocase) are up-regulated in the liver of insulin-resistant ob/ob mice [144]. High hepatic glucose output is a consequence of increased gluconeogenesis. We may then postulate that the enhanced transcriptional activities of the glycolytic and lipogenic genes are regulated by the co-ordinated activity of ChREBP, LXR-α and SREBP-1c.

Inflammation, oxidative stress, fibrosis and ER stress

Fat is now considered a metabolically active endocrine organ producing pro-inflammatory cytokines, including TNF-α, IL (interleukin)-6 and IL-8, and there is evidence to support the activation of other inflammatory pathways, oxidative stress and the de novo pathway by TNF-α [145147]. It has recently been hypothesized that NAFLD is a consequence of a ‘two hit’ insult. As a consequence of insulin resistance, there is excessive accumulation of TAG in hepatocytes, which is the ‘first hit’. Oxidative stress from β-oxidation, increased expression of inflammatory cytokines by NF-κB-dependent pathways and adipocytokines are potential synergistic factors acting in concert as the ‘second hit’, resulting in hepatocyte injury, inflammation and fibrosis.

In contrast, all subjects with NAFLD have insulin resistance, even in lean subjects and those with normal glucose tolerance [59,148,149]. The core metabolic derangement is due to insulin resistance resulting in increased lipolysis in adipose tissue, increased NEFA uptake by hepatocytes and increased TAG synthesis in the liver. Mitochondrial fat oxidation and export of VLDL particles are not able to match TAG synthesis, leading to net fat deposition in the hepatocytes. As a consequence of abnormal fat accumulation in the hepatocytes, there is marked derangement in the insulin signalling pathways in the liver via activation of NF-κB, IKK-β [IκB (inhibitor of NF-κB) kinase-β], atypical PKC and JNK pathways by NEFAs [150,151].

There is evidence that increased mitochondrial fat oxidation, a consequence of hepatic insulin resistance, is linked to hepatic injury, inflammation and fibrosis. This is because NEFAs and their metabolites are ligands for PPAR-α. This transcription factor regulates a number of genes, including those involved in mitochondrial, peroxisomal and microsomal fat oxidation, and there is now evidence to support the up-regulation of these genes in NAFLD.

As a result of increased fat acid oxidation in the mitochondria and peroxisome, there is increased generation of hepatotoxic oxygen free radicals. Because of oxidative stress, there is lipid peroxidation and severe mitochondrial dysfunction with structural and functional abnormalities. There is ATP depletion due to the failure to generate ATP via oxidative phosphorylation [152,153], and this could be due to increased expression of UCP-2 (uncoupling protein-2) by PPAR-α activation [154,155]. Patients with NASH also have structural mitochondrial abnormalities [156160], and, under-expression of some genes necessary for normal mitochondrial function have also been reported in these patients [161].

Inflammation is the link between obesity and insulin resistance, and may have an important role in the pathogenesis of hepatic and systemic insulin resistance. For example, serine phosphorylation of IRS-1 by inflammatory cytokines is a mechanism mediating insulin resistance in obese subjects with chronic low-grade inflammation. In support of an association between insulin resistance and inflammatory cytokines, there is increased production of TNF-α in obese animal models and in obese insulin-resistant subjects [162]. In contrast, knockout mice for either TNF-α or the TNF-α receptor were reported to be insulin-sensitive [162]. In obese subjects with previously increased TNF-α expression, decreased TNF-α expression following weight loss and exercise was observed [163]. The molecular mechanism responsible for TNF-α-induced insulin resistance is most probably serine phosphorylation of IRS-1, which has an inhibitory effect on the downstream propagation of insulin signals [163,164].

Another mechanism for TNF-α-induced insulin resistance is through the activation of the IKK-β pathway [165]. Overexpression of IKK-β is associated with reduced insulin signalling in cultured cells [165]. In contrast, liver-specific deletion of IKK-β resulted in improved hepatic insulin sensitivity even on a high-fat diet [165]. As a result of the activation of IKK-β, there is downstream activation of the NF-κB pathway with worsening inflammation and hepatic insulin resistance. At the level of transcription, the activation of NF-κB also has an inhibitory effect on PPAR-γ agonism.

However, it is uncertain whether inflammatory cytokines cause NAFLD or occur as a consequence of the developing liver condition and NF-κB activation. In a previous review [166], increased inflammatory cytokine production was reasoned to be a consequence of IKK-β and NF-κB activation and, in support of this reasoning in NAFLD, increased hepatic NF-κB activity in high-fat-fed mice was associated with increased inflammatory cytokine expression, with attenuation of this effect with specific inhibition of NF-κB activity [150]. Additionally, activation of the upstream protein IKK-β was also associated with inflammation, hepatic and systemic insulin resistance, and there was also a reversal of this effect using a hepatic-specific inhibitor of NF-κB. In NAFLD, it is plausible that hepatic NEFAs have a toxic effect on hepatocyte lysosomal membranes with the consequent release of proteolytic enzymes that directly activate the IKK-β and NF-κB pathways with increased TNF-α expression [167].

JNK is a member of the MAPK (mitogen-activated protein kinase) family that regulates cell function via the control of AP-1 (activated protein-1), and there is a strong association between increased JNK activity and insulin resistance. In obesity with insulin resistance, increased NEFAs and inflammatory cytokines are accompanied by increased JNK activity in insulin-sensitive tissue, thereby causing insulin resistance [162]. Although a knockout model for JNK1 did not develop insulin resistance [168], a possible role for the JNK family of MAPKs in NAFLD is presently uncertain.

The ER is a cellular organelle for synthesis, storage and transport of proteins. Because of the presence of some chaperone proteins, the ER ensures proper folding of proteins. IRE-1 (inositol-requiring enzyme-1), PERK [PKR (double-stranded-RNA-dependent protein kinase)-like ER kinase), ATF-6 (activating factor-6) and XBP-1 (X-box-binding protein-1) are subcellular proteins necessary for ER function. Although these protein are important for normal ER function, abnormal activation may occur as a consequence of a ‘stressed’ environment, for example hypoxia, leading to inflammation and insulin resistance [169]. Evidence is beginning to emerge for the pathogenic role of hepatic ER ‘stress’ in inflammation and insulin resistance. For example, ER ‘stress’ has been shown to directly activate the IKK-β/NF-κB and the JNK/AP-1 pathways via interactions with IRE-1 and PERK, thereby inducing or exacerbating inflammation and insulin resistance [162,166] (Figure 6). Hepatic ER ‘stress’ has also been shown to occur with high-fat feeding and in a genetically induced obese animal model [170].

Figure 6 Interaction between inflammatory cytokines, ER stress, JNK, NF-κB pathways and hepatic inflammation

Schematic diagram showing activation of JNK/AP-1 and IKK-β/NF-κB pathways by ER stress and inflammatory cytokines resulting in exacerbated inflammation.

Because of the modulatory effect of endogenous bile acids and their derivatives on ER function, the therapeutic potential of these pharmacological agents have been explored in both humans and animal models. Treatment of NASH in insulin-resistant subjects with UDCA (ursodeoxycholic acid) has so far been disappointing [168], although, in obese mice, these authors have shown improved hepatic and whole-body insulin sensitivity with UDCA treatment.

NEMO (NF-κB essential modulator/IKK-γ) is the regulatory component of the IκB complex. The two catalytic components of this complex are IKK-1/IKK-α and IKK-2/IKK-β [171]. NF-κB has an important cellular function in the regulation of immune and inflammatory responses, and is maintained in its inactive form in the cytosol by binding inhibitory proteins [171]. Although the physiological function of NF-κB in adult liver is not known, deletion of either the catalytic component (IKK-2) or the regulatory component (NEMO) is not compatible with intra-uterine life [172].

Hepatocellular carcinoma is a well-known complication of NAFLD and the molecular mechanism in this group of patients in beginning to be better understood. There is evidence to support the tumour-suppressive effect of NEMO in hepatocytes. In one animal experiment [171], NEMO-knockout mice developed spontaneous hepatocellular carcinoma at 12 months of life. In addition, there was histological evidence of pre-malignant lesions in the liver of these mice by 9 months. In the same study, there was biochemical and histological evidence of NASH prior to the development of cancer in this mouse model. In contrast, specific deletion of IKK-2 (a catalytic component) was not associated with chronic liver disease or spontaneous liver cancer [171]. These findings led the authors to suggest that NEMO-dependent NF-κB activation has a physiological role in the prevention of spontaneous hepatic cancer.

LPS (lipopolysaccharide) is a potent inducing agent of TNF-α and other inflammatory cytokines [173,174], and bacterial antigens entering between tight junctions in the intestine could have an impact on liver damage in the pathogenesis of NAFLD. There is a failure in NF-κB activation following injection of LPS in the liver of NEMO-knockout mice, and the administration of TNF-α in these mice also showed a similar effect [171]. As a consequence of the absence of NF-κB activation, these mice developed severe liver failure due to exposure to LPS and increased cytokine drive. Thus LPS-induced activation of TNF-α could contribute to liver damage in the pathogenesis of NAFLD, particularly if NEMO-dependent NF-κB activation is inadequate to protect the liver from damage exacerbated through activation of the JNK pathway. It is possible that oxidative stress may be a contributory factor to hepatocyte necrosis observed in NEMO-knockout mice. In an animal study [171], NEMO-knockout mice were treated with a dietary antioxidant supplement, BHA (butylated hydroxyanisole), and were shown to have biochemical and histological improvements in chronic liver disease. There was also reduced hepatic JNK activity in the antioxidant group, leading the authors to postulate that antioxidant therapy could ameliorate or normalize chronic liver disease in this NEMO-knockout model.


There is emerging interest in the role of adipocytokines in the pathogenesis of NAFLD. Adiponectin has been shown to decrease de novo fatty acid synthesis but enhance fat oxidation. Stimulation of adiponectin receptors, which are expressed in hepatocytes and myocytes, leads to PPAR-α and AMPK activation [175]. As a result, adiponectin increases β-oxidation but decreases TAG synthesis. Adiponectin also has direct anti-inflammatory effects by decreasing hepatic TNF-α production. In support of a pathogenetic role for adiponectin in affecting inflammation in NAFLD, patients with NASH have decreased expression of adiponectin receptors in the liver and decreased serum adiponectin concentrations compared with patients with simple steatosis [176,177].


The psychotropic actions of the plant Cannabis sativa were first documented approx. 4000 years ago. The endocannabinoid system contributes to the physiological regulation of energy balance, food intake, and lipid and glucose metabolism through both central and peripheral effects. Cannabinoid receptors, named CB1 receptor and CB2 receptor, participate in the physiological modulation of many central and peripheral functions. The CB2 receptor is mainly expressed in immune cells, whereas the CB1 receptor is the most abundant G-protein-coupled receptor expressed in the brain. The CB1 receptor is expressed in the hypothalamus and the pituitary gland, and its activation is known to modulate all the endocrine–hypothalamic–peripheral endocrine axes. Recent exciting new results suggest that inhibition of the endocannabinoid system might be beneficial in the treatment of the NAFLD. Interestingly, a new class of anti-obesity medication, such as rimonbant (a CB1 receptor antagonist), was shown to be beneficial in an animal model of hepatic steatosis [178]. Treatment of obese (fa/fa) rats with rimonabant (30 mg/kg of body weight) daily for 8 weeks abolished hepatic steatosis, decreased hepatomegaly, decreased ALT and GGT levels, decreased dyslipidaemia and, importantly, increased the level of the anti-inflammatory hormone adiponectin [178]. At the hepatic cell level, activation of hepatic CB1 receptors by stellate cell endocannabinoid 2-AG (2-arachidonoylglycerol) is responsible for ethanol-induced steatosis by increasing lipogenesis and decreasing fatty acid oxidation [179]. Recently, rimonabant has been shown to be effective in reducing weight [180], and randomized clinical trials are now needed to establish whether rimonabant produces therapeutic benefit in the treatment of NAFLD.

Mitochondrial dysfunction and ROS

The mitochondrial genome synthesizes 13 respiratory chain polypeptides and is very sensitive to oxidative damage, probably because of (i) the proximity of mtDNA (mitochondrial DNA) to the inner mitochondrial membrane (a source of ROS), (ii) the absence of protective histones; and (iii) the lack of effective DNA repair mechanisms.

There is evidence that increased production of ROS, pro-inflammatory cytokines and hepatocyte necrosis secondary to mitochondrial dysfunction is important in the pathogenesis of NAFLD. Although the sequence of pathophysiological events are yet to be clarified, it is plausible that oxidative stress is an early event in NAFLD and could provide the ‘trigger’ for inflammation and fibrosis. Oxidative stress is undoubtedly a major feature of NASH, and potential sources of ROS are hepatocyte parenchymal cells, chronic inflammatory cells within the hepatocytes and, in obese subjects, adipose tissue with infiltrating activated macrophages [181,182].

There is some evidence that increased generation of ROS could enhance progression from simple steatosis to steatohepatitis and fibrosis [183]. The potential mechanisms are through lipid peroxidation, induction of pro-inflammatory cytokines and induction of FAS ligand. The plasma and mitochondrial membranes are vulnerable to free radical damage, and ROS-induced lipid peroxidation of membranes can cause cell necrosis or apoptosis. Lipid peroxidation can also cause immunological dysfunction, which could lead to hepatic fibrogenesis. It is possible that, as a result of lipid peroxidation, there is release of MDA (malondialdehyde) and 4-HNE (4-hydroxynonenal), which bind hepatocyte proteins forming new antigen(s) and, thereby, provoking a harmful immunological response. An immunological response could induce neutrophil chemotaxis and activate hepatic stellate cells to promote collagen synthesis and deposition [184,185].

Oxidative stress (ROS) has also been shown to increase production of pro-inflammatory cytokines [TNF-α, TGF-α, IL-6 and IL-8] via the activation of NF-κB [186]. NEFAs accumulate in the hepatocytes in NAFLD and are also capable of activating the NF-κB pro-inflammatory pathway [166,187].

In addition to the NF-κB-dependent pathways, Kupffer cells (liver-specific activated macrophages) are a source of inflammatory cytokines and, in an obese individual, adipose tissue with infiltrating macrophages is also a source of pro-inflammatory cytokines. Because of the increased production of pro-inflammatory cytokines in NAFLD, these cytokines (e.g. TNF-α) could exacerbate systemic and hepatic insulin resistance with worsening inflammation and fibrosis. Inflammatory cytokines cause impaired insulin signalling, cell damage, neutrophil chemotaxis, hepatic stellate cell activation and apoptosis [188190], and there is increased expression of TNF-α and its receptor in liver and adipose tissue in obese individuals with NASH, which correlates with the degree of fibrosis [191].

Under physiological conditions, ketogenesis is exquisitely sensitive to insulin but, paradoxically, mitochondrial fat oxidation and ketogenesis are increased in NAFLD with hyperinsulinaemia [192,193]. Even obese leptin (ob/ob)-deficient mice with fatty liver also have increased fat oxidation [194], perhaps explained by increased concentrations of long-chain fatty acids activating PPAR-α, a nuclear receptor and a positive transcription factor for the CPT-1, MCAD (medium-chain acyl-CoA dehydrogenase) and HMG (3-hydroxy-3-methylglutaryl)-CoA synthetase genes. Some investigators have postulated that long-chain fatty-acid-mediated PPAR-α activation is specific to fatty acids from the de novo pathway [95], and there is increased expression of PPAR-α in ob/ob mice with steatosis [195]. Long-chain fatty acids increase the activity and expression of CPT-1 and, thereby, increase fat oxidation [196]. Additionally, there is a decreased affinity of CPT-1 activity for malonyl-CoA, a potent inhibitor of CPT-1 [197].

Activation of AMPK also enhances mitochondrial fat oxidation via phosphorylation (inactivation) of ACC. As a result, there is a reduced conversion of acetyl-CoA into malonyl-CoA and, therefore, decreased inhibition of CPT-1 by malonyl-CoA, resulting in increased β-oxidation [198].

In mitochondria oxidative phosphorylation, NADH and FADH2 are re-oxidized to NAD+ and FAD via the electron transport chain. As part of oxidative phosphorylation, this is a co-ordinated flow of electrons down an electrochemical gradient from complex I to IV with simultaneous extrusion of protons, creating a large electrochemical gradient. Consequently, a large gradient across the mitochondrial membrane drives ATPase synthase resulting in ATP production. Under physiological conditions, most of the electrons provided to the respiratory chain ‘transit’ through to the last complex, where they combine with oxygen and protons to form water. However, upstream of the production of water, electrons react with oxygen to form free radicals (superoxide anions). These radicals are dismutated by SOD (superoxide dismutase) to H2O2 and then to water by glutathione peroxidase. Glutathione peroxidase requires GSH to function, and decreased GSH may be accompanied by mitochondrial dysfunction and/or cell death [199]. In pathological conditions, it is plausible that increased free radical production results in further mitochondrial damage and dysfunction, with a consequent decrease in oxidative capacity. A decrease in mitochondrial oxidative capacity could potentially result in an imbalance between fat oxidation and lipogenesis, leading to hepatic fat accumulation and NAFLD.


Our group has published extensively in recent years showing that modification of intra-uterine nutrition during gestation affects epigenetic regulation of key metabolic genes and has the potential to modify the disease phenotype [200208]. Specifically, our group has shown for the first time that prenatal nutrition induces differential changes to the methylation of individual CpG dinucleotides in the hepatic PPAR-γ promoter, altering mRNA levels of the PPAR-α gene, downstream target genes and phenotype [202]. We have recently developed a mouse model of human NAFLD which shows that exposure to a high-fat diet in utero and during lactation exacerbates the NAFLD phenotype exhibited in offspring that were also fed a high-fat diet post-weaning (K. D. Bruce, F. Cagampang and C. D. Byrne, unpublished work).

Thus with our findings showing an influence of intrauterine nutritional exposure to affect the risk of the NAFLD phenotype we suggest that intra-uterine nutrition may have the potential to modify hepatic lipogenesis through an influence of PPAR-γ and modifying SREBP-1c function. We speculate that increased dietary exposure to a high-fat diet in the pregnant mother results in increased placental transfer of fatty acids to the fetus. It is possible that increased placental transfer of fatty acids to the fetus increases hepatic lipogenesis and oxidative stress in the vulnerable fetal liver, perhaps also increasing the demand for methyl donors to promote cell growth and the subsequent development of NAFLD in adulthood. It is conceivable that, if the increased demand for methyl donors is not met, epigenetic changes in gene expression would result that predispose to conditions of fat accumulation and inflammation (such as NAFLD) in the developing offspring.

The mitochondrial genome is particularly susceptible to oxidative damage due to the absence of protective histones and incomplete repair mechanisms in mitochondria [209]. Mitochondria and mtDNA are attractive vectors for the transmission of early environmental stress, as they can harbour environmental damage for many cell divisions before cell dysfunction becomes evident [210]. Mitochondria depend on both nuclear and mtDNA to function. mtDNA is sensitive to environmental changes during oocyte maturation and during perimplantation development [211]. Therefore mitochondrial dysfunction caused by changes in mtDNA or altered expression of nuclear genes could cause both immediate and delayed increases in oxidant stress which may alter nuclear gene expression and cell function.

As mentioned above, we have developed a mouse model showing that manipulating the maternal dietary exposure during pregnancy worsens development of the adult offspring NAFLD phenotype. We have obtained results from this model of the developmental origins of NAFLD of decreased mitochondrial complex activity. To investigate the effect on the offspring of manipulating the maternal diet during pregnancy, we have determined whether there were changes in electron transport chain enzyme complex activity in the liver from mouse offspring in adulthood which had been exposed to a high-fat or to a control diet both pre- and post-natally (K. D. Bruce, F. Cagampang and C. D. Byrne, unpublished work). Female C57 BL/6J black mice were randomly assigned to either a high-fat diet (HF, 45% kJ from fat, 20% kJ from protein and 35% kJ from carbohydrate; n=10) or to a standard laboratory chow diet (C, 21% kJ from fat, 18% kJ from protein and 63% kJ from carbohydrate; n=10). Dams were fed the diets for 4 weeks prior to conception, during gestation and lactation. At weaning, the offspring were assigned either the high-fat or control diets, generating four experimental groups: HF/HF (n=12), HF/C (n=12), C/HF (n=12) and C/C (n=12), which represents the pre-natal/post-natal diet respectively. In both male and female offspring, body weights, total fat mass and systolic BP measurements were significantly higher (P<0.01) in offspring from HF/HF compared with the C/C or the C/HF groups. Hepatic electron transport chain complex activity was decreased in offspring from HF/C and HF/HF groups compared with the C/C group, and the greatest decrease was observed in Complex I activity, where a 3.2- and 3.7-fold reduction in activity levels was seen in HF/HF offspring (P<0.05) and the HF/C offspring (P<0.05) respectively, compared with C/C offspring. These findings suggests that the pre-natal exposure to a high-fat diet impairs complex activity and mitochondrial function. Steatosis was visible in liver sections (most marked in the HF/HF animals (Figure 7), HF/HF Oil red staining), and there was a marked inflammatory infiltrate in sections from the HF/HF group (Figure 7, HF/HF H&E staining) that is characteristic of human NASH. Thus these preliminary results provide evidence to suggest that exposure to a high-fat diet in utero combined with post-natal high-fat exposure contributes to a more florid disease progression of fatty liver to NASH.

Figure 7 Liver histology of a mouse model of the ‘developmental origins’ of NAFLD

Sections were stained with either (upper panels) haematoxylin and eosin (H&E) or (lower panels) Oil Red. Magnification, ×40. C, control diet; HF, high-fat diet.

Our group has also shown in human studies that fatty liver is associated with increased NEFA supply to the liver [28,212,213], and we have shown in our mouse model that hepatic TAG is strongly correlated with CPT-1 expression (r=−0.55, P=0.033) [214], which mediates the transfer of long-chain fatty acyl-CoAs across the mitochondrial membrane. CPT-1 is regulated by PPAR-α activity and, thus, our finding that altered early nutrition influences PPAR-α methylation [200,202] suggests that altered early nutrition may influence the flux of fatty acids into the mitochondria for β-oxidation. We suggest that increased flux of fatty acids to the developing fetal liver increases susceptibility of the liver to damage (NAFLD) induced by a second post-natal exposure to a high-fat diet.


Precise estimates for the prevalence of NAFLD (and the different subgroups such as NASH) are difficult to obtain, because of an unmet need for simple diagnostic tests. Recent progress has been made in this field, and in the near future it may be possible to use simple algorithms to replace liver biopsy for the diagnosis of NAFLD. It has been shown recently that NASH is associated with CHD, and in Type 2 diabetes NAFLD may be associated with microvascular disease. To date there are no licensed treatments for NAFLD and the development of novel therapies has been hampered by a poor understanding of the molecular mechanisms contributing to NAFLD. This problem is now being addressed and there has been a marked improvement in an understanding of the pathogenesis of NAFLD over the last decade. It is now clear that factors operating outwith the liver (e.g. uncontrolled adipocyte lipolysis) have a powerful impact to increase liver fat accumulation. There has been a marked improvement in our understanding of the role of key transcription factors and the importance of three transcription factors SREBP-1c, LXR-α and ChREBP operating in the liver to co-ordinate the regulation of glycolysis and fatty acid synthesis. In the future, it may prove possible to modify the function of these transcription factors to ameliorate the development and progression of NAFLD. At the present time, there is a need to understand better the link between components of TAG synthesis and inflammation, as decreasing factors contributing to inflammation are likely to attenuate the development of fibrosis and progression of disease. Work from our laboratory suggests that factors affecting early fetal liver development, such as increased maternal dietary fat consumption, may act to increase the vulnerability of developing fetal liver to fat accumulation in adulthood. If the rapid progress in understanding of the pathogenesis of NAFLD is maintained, it is likely that we will have novel treatments for NAFLD and the tools for monitoring therapy without resorting to liver biopsy in the near future.


The unpublished work relating to Figure 7 was supported by the Biotechnology and Biological Sciences Research Council [grant number BB/D001633/1].


We thank Lucinda England for her help with the manuscript.

Abbreviations: ACC, acetyl-CoA carboxylase; ALT, alanine aminotransferase; AMPK, AMP-activated kinase; AP-1, activated protein-1; aRR, adjusted relative risk; ASO, antisense oligonucleotide; AST, aspartate aminotransferase; BMI, body mass index; BP, blood pressure; CCT, CTP:phosphocholine cytidylyltransferase; CHD, coronary heart disease; ChREBP, carbohydrate-responsive-element-binding protein; CI, confidence interval; CKD, chronic kidney disease; CPT, CDP-choline:1,2 diacylglycerol choline phosphotransferase; CPT-1, carnitine palmitoyltransferase-1; CVD, cardiovascular disease; DAG, diacylglycerol; DGAT-2, diacylglycerol acyltransferase-2; ELF score, European Liver Fibrosis score; ER, endoplasmic reticulum; FAS, fatty acid synthase; Fox, forkhead box; FRS, Framingham risk score; GFR, glomerular filtration rate; GGT, γ-glutamyltransferase; GLUT4, glucose transporter 4; HbA1c, glycated haemoglobin; HDL, high-density lipoprotein; HSL, hormone-sensitive lipase; IL, interleukin; IMT, intima-media thickness; INSIG-1, insulin induced gene-1; IRE-1, inositol-requiring enzyme-1; IRS, insulin receptor substrate; IU, international units; JNK, c-Jun N-terminal kinase; LDL, low-density lipoprotein; L-PK, liver-specific pyruvate kinase; LPS, lipopolysaccharide; LXR-α, liver X receptor-α; MAPK, mitogen-activated protein kinase; MCD, methionine- and choline-deficient; MRI, magnetic resonance imaging; MRS, magnetic resonance spectroscopy; mtDNA, mitochondrial DNA; mTOR, mammalian target of rapamycin; MUFA, mono-unsaturated fatty acid; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis; NEFA, non-esterified fatty acid (‘free fatty acid’); NF-κB, nuclear factor κB; IκB, inhibitor of NF-κB; IKK, IκB kinase; NEMO, NF-κB essential modulator; OR, odds ratio; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PEMT, phosphatidylethanolamine N-methyltransferase; PERK, PKR (double-stranded-RNA-dependent protein kinase)-like ER kinase; PI3K, phosphoinositide 3-kinase; PKC, protein kinase C; PP2A, protein phosphatase 2A; PPAR, peroxisome-proliferator-activated receptor; PTEN, phosphatase and tensin homologue deleted on chromosome 10; PUFA, polyunsaturated fatty acid; ROS, reactive oxygen species; SAM, S-adenosylmethionine; SCD-1, stearoyl-CoA desaturase-1; SHIP-2, SH2-containing inositol phosphatase-2; SREBP-1c, sterol-regulatory-element binding-protein-1c; TAG, triacylglycerol; TGF, transforming growth factor; TNF-α, tumour necrosis factor-α; UDCA, ursodeoxycholic acid; VLDL, very-LDL


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
  84. 84.
  85. 85.
  86. 86.
  87. 87.
  88. 88.
  89. 89.
  90. 90.
  91. 91.
  92. 92.
  93. 93.
  94. 94.
  95. 95.
  96. 96.
  97. 97.
  98. 98.
  99. 99.
  100. 100.
  101. 101.
  102. 102.
  103. 103.
  104. 104.
  105. 105.
  106. 106.
  107. 107.
  108. 108.
  109. 109.
  110. 110.
  111. 110a.
  112. 111.
  113. 112.
  114. 113.
  115. 114.
  116. 115.
  117. 116.
  118. 117.
  119. 118.
  120. 119.
  121. 120.
  122. 121.
  123. 122.
  124. 123.
  125. 124.
  126. 125.
  127. 126.
  128. 127.
  129. 128.
  130. 129.
  131. 130.
  132. 131.
  133. 132.
  134. 133.
  135. 134.
  136. 135.
  137. 136.
  138. 137.
  139. 138.
  140. 139.
  141. 140.
  142. 141.
  143. 142.
  144. 143.
  145. 144.
  146. 145.
  147. 146.
  148. 147.
  149. 148.
  150. 149.
  151. 150.
  152. 151.
  153. 152.
  154. 153.
  155. 154.
  156. 155.
  157. 156.
  158. 157.
  159. 158.
  160. 159.
  161. 160.
  162. 161.
  163. 162.
  164. 163.
  165. 164.
  166. 165.
  167. 166.
  168. 167.
  169. 168.
  170. 169.
  171. 170.
  172. 171.
  173. 172.
  174. 173.
  175. 174.
  176. 175.
  177. 176.
  178. 177.
  179. 178.
  180. 179.
  181. 180.
  182. 181.
  183. 182.
  184. 183.
  185. 184.
  186. 185.
  187. 186.
  188. 187.
  189. 188.
  190. 189.
  191. 190.
  192. 191.
  193. 192.
  194. 193.
  195. 194.
  196. 195.
  197. 196.
  198. 197.
  199. 198.
  200. 199.
  201. 200.
  202. 201.
  203. 202.
  204. 203.
  205. 204.
  206. 205.
  207. 206.
  208. 207.
  209. 208.
  210. 209.
  211. 210.
  212. 211.
  213. 212.
  214. 213.
  215. 214.
  216. 215.
  217. 216.
  218. 217.
  219. 218.
  220. 219.
  221. 220.
  222. 221.
  223. 222.
  224. 223.
  225. 224.
  226. 225.
  227. 226.
  228. 227.
View Abstract