Whether non-alcoholic fatty liver disease (NAFLD) precedes insulin resistance (IR) or IR preludes/causes NAFLD has been long debated. Recent studies have shown that there are two phenotypes of NAFLD, ‘genetic’ vs ‘metabolic’ NAFLD. The former patients are more at risk of hepatocellular carcinoma and chronic liver disease the latter are more IR and at increased risk of type 2 diabetes (T2D). Even if they are not yet diabetics, from a metabolic point of view having NAFLD is equivalent to T2D with reduced peripheral glucose disposal and impaired suppression of hepatic glucose production, but without fasting hyperglycaemia. T2D develops only when hepatic autoregulation is lost and glucose production exceeds the capacity of muscle glucose disposal.
In NAFLD adipocytes are resistant to the effect of insulin, lipolysis is increased and excess plasma free fatty acids (FFA) are taken up by other organs (mainly liver) where they are stored as lipid droplets or oxidized. Increased adiposity is associated with worsen severity of both ‘genetic’ and ‘metabolic’ NAFLD. FFA oxidative metabolism is increased in NAFLD and not shifted towards glucose during insulin infusion. Although this reduced metabolic flexibility is an early predictor of T2D, it can be seen also as a protective mechanism against excess FFA.
In conclusion, IR precedes and causes ‘metabolic’ NAFLD, but not ‘genetic’ NAFLD. Reduced metabolic flexibility in NAFLD might be seen as a protective mechanism against FFA overflow, but together with IR remains a strong risk factor for T2D that develops with the worsening of hepatic regulation of glucose production.
Non-alcoholic fatty liver disease (NAFLD) is finally recognized as a serious disease that is associated with increased risk of comorbidities like type 2 diabetes (T2D) and cardiovascular diseases (CVD) and can progress to non-alcoholic steatohepatitis (NASH) and hepatocellular carcinoma (HCC) [1,2]. The global prevalence of NAFLD (approximately 25%) is rising together with the prevalence of obesity, and currently the highest rate (approximately 30%) is in the Middle East and South America and the lowest in Africa (13%) . The early stages of NAFLD (simple steatosis) are a sign of metabolic dysfunction; however, steatosis can progress to NASH and chronic liver disease (CLD) and both the American  and the  European Guidelines for the Management of NAFLD established that it is important to diagnose and treat this disease in the earliest stages.
The majority of patients with NAFLD are obese so it is not surprising to find that most of them have insulin resistance (IR) and/or T2D [4,5]. Although muscle is the tissue that is mainly affected by IR, in subjects with NAFLD impairment in insulin action is present also in liver and adipose tissue, even before T2D becomes manifest [6,7]. In the study published in this issue of the journal , Brouwers et al. added further knowledge studying glucose and lipid metabolism of non-diabetic (ND) and diabetic (T2D) subjects with/without NAFLD, well matched for gender, age (59 years) and BMI (30 kg/m2). As expected NAFLD-ND were more IR at the level of the muscle than controls, i.e., NDs without NAFLD (CT); peripheral glucose disposal (Rd), that reflects muscle insulin sensitivity, was markedly reduced (45% less than in CT), but unexpectedly insulin sensitivity in NAFLD-ND was no better than in T2D with or without NAFLD. Another finding of this paper was that NAFLD-ND have IR also at the level of the liver since insulin is not able to properly suppress endogenous glucose production (EGP) neither during fasting nor during insulin infusion, confirming previous results [6,7]. Hepatic IR in NAFLD-ND was comparable to T2Ds (with/without NAFLD) during insulin infusion (−60% in CT vs −30% in the other three groups), while during fasting the defect was evident only in NAFLD (both ND and T2D) . From the analysis of these results we can conclude that having NAFLD is equivalent to T2D regarding the defects in muscle and liver glucose metabolism, but without fasting hyperglycaemia. The main determinant of fasting hyperglycemia and T2D is increased EGP, but this occurs only when hepatic ‘autoregulation’ (that limits the contribution of gluconeogenesis (GNG) and glycogenolysis to EGP) is lost and the increased EGP flux, together with reduced peripheral glucose disposal, results in high blood glucose concentrations [9,10]. In NAFLD GNG does not correlate with the amount of total fat . It is likely that hyperglycemia does not develop because excess circulating substrates that reach the liver are diverted towards metabolic pathways alternative to GNG, such as hepatic glyceroneogenesis and de novo lipogenesis (DNL), that in NAFLD have been shown to be enhanced . It can be speculated that in NAFLD fasting glucose concentrations are maintained within normal ranges till hepatic autoregulation is preserved.
NAFLD has a dual origin ‘genetic’ or ‘metabolic’. ‘Genetic NAFLD’, either from PNPLA3, hypo-betalipoproteinemia, DGAT or TM6SF2 predicts NASH and severe liver disease but not T2D [4,12]. The genetic forms of NAFLD are not associated to increased IR in muscle or liver when patients are compared with control subjects with similar BMI [9,13–15]. The exact mechanisms why ‘genetic NAFLD’ accumulates fat in the liver are only in part clear: it has been shown that it tends to accumulate mainly polyunsaturated triacylglycerols  but the risk of progression to more severe forms of this disease is increased . In ‘genetic NAFLD’ an increase in total fat accumulation determines an increase in IR and together they synergistically increase the risk associated with PNPLA3, TM6SF2 and GCKR genotype [12,17].
In subjects without genetic risk, NAFLD occurs because of high caloric intake; IR and excess energy intake favour hepatic fat accumulation and this is generally referred as ‘metabolic’ NAFLD (Figure 1). In ‘metabolic’ NAFLD, hepatic triglyceride (TG) accumulation is not due to impairment in either VLDL secretion or FFA oxidation that is even higher than in non-NAFLD , but is the consequence of the hepatic lipid overflow. Peripheral IR requires higher insulin secretion; lipolysis is not reduced despite high insulin levels rather the combination of high insulin and high lipids promotes hepatic FFA uptake and TG synthesis. Thus, in subjects with IR that consume excess calories, NAFLDs are more likely to develop and then progress in the disease. The impairment in insulin action at the level of the liver and adipose tissue is proportional to the amount of both hepatic and visceral fat [6–9].
Proposed mechanisms for the development of NAFLD: ‘genetic’ vs ‘metabolic’
The adipose tissue metabolism is crucial for the development of NAFLD [5,18]. It is now recognized that if the capacity of subcutaneous adipose tissue (the biggest fat depot) to expand is compromised, excess lipids are drifted towards visceral fat and organs such as the liver and pancreas where they are stored as ectopic fat [5,18] (Figure 1). The overfeeding studies showed that increased fat (mainly subcutaneous) and muscle IR are among the early effects of positive energy balance, while ectopic fat accumulation, as hepatic or visceral fat, occurs only later . For example Knudesen et al.  showed that 14 days of inactivity and overfeeding (+50%) decreased peripheral insulin sensitivity already at day 3 while body fat and visceral fat were increased significantly only after 14 days. Peterson et al.  have shown that a 40% overfeeding for 8 weeks decreased peripheral glucose disposal at low (10 mU/min m2) insulin infusion rates, in particular non-oxidative disposal rate, and although it increased body weight by 7.6 kg (of which +4.2 kg of body fat) there was no relevant change in hepatic fat (from 1.5 to 2.2% at the end of study) or visceral fat (from 0.58 to 0.94 kg). However, some patients may have both conditions (genetic + metabolic); in this case the genetic risk is synergistically increased by total adiposity and IR [12,17].
FFA oxidation is also up-regulated in NAFLD, not only during fasting, but even during insulin infusion [7,8] when glucose is the predominant energy substrate . This reduced capacity to switch from fat to carbohydrate oxidation during insulin infusion is defined as ‘metabolic inflexibility’ and together with IR is a risk factor for the development of T2D . However, the reduced metabolic flexibility observed by Brouwers et al. in NAFLD-ND can be seen also as a protective mechanism against hepatic fatty acids overflow .
Although large prospective studies on this topic are still lacking, it seems that the majority of subjects with the ‘metabolic NAFLD’ tend to accumulate ‘toxic lipids’, are more insulin resistant and at increased risk of T2D than those with ‘genetic NAFLD’ [4,5] that on the other hand have a higher risk of severe liver disease .
In conclusion, the ‘metabolic’ and the ‘genetic’ NAFLD identify two different phenotypes: the former is associated with IR and reduced metabolic flexibility and many evidences indicate that IR may precede ‘metabolic’ NAFLD; ‘genetic’ NAFLD is not necessarily associated to IR or lipotoxicity, but the excess caloric intake and the increase in total fat synergistically increase the risk of progression of liver disease in ‘genetic’ NAFLD. We should probably distinguish a type 1 (genetic) vs type 2 (metabolic) NAFLD, the former related to genetic and the latter to metabolic alterations (Figure 1). At the moment there is no approved drug therapy for NAFLD, but the evaluation of the clinical phenotype should become an important step to develop the most effective way to treat these patients.
Amalia Gastaldelli is a member of the EPoS (Elucidating Pathways of Steatohepatitis) consortium, which is funded by the Horizon 2020 Framework Program of the European Union under Grant Agreement 634413.
The author declares that there are no competing interests associated with the manuscript.
chronic liver disease
de novo lipogenesis
endogenous glucose production
free fatty acids
non-alcoholic fatty liver disease
type 2 diabetes
very low density lipoproteins