Mitochondrial Dysfunction, Insulin Resistance, and Potential Genetic Implications

Abstract

Insulin resistance (IR) is fundamental to the development of type 2 diabetes (T2D) and is present in most prediabetic (preDM) individuals. Insulin resistance has both heritable and environmental determinants centered on energy storage and metabolism. Recent insights from human genetic studies, coupled with comprehensive in vivo and ex vivo metabolic studies in humans and rodents, have highlighted the critical role of reduced mitochondrial function as a predisposing condition for ectopic lipid deposition and IR. These studies support the hypothesis that reduced mitochondrial function, particularly in insulin-responsive tissues such as skeletal muscle, white adipose tissue, and the liver, is inextricably linked to tissue and whole body IR through the effects on cellular energy balance. Here we discuss these findings as well as address potential mechanisms that serve as the nexus between mitochondrial malfunction and IR.

More than 380 000 000 adults worldwide have type 2 diabetes (T2D) and this number will greatly increase in prevalence over the next 20 years, leading to marked increases in morbidity and mortality and contributing to massive increases in associated healthcare expenditures (1). Type 2 diabetes occurs when tissue insensitivity to insulin action (insulin resistance [IR]) is coupled with an inadequate secretion of insulin by the pancreas. The vast majority of those with prediabetes (preDM) have some degree of IR and the T2D pandemic is being largely driven by the deterioration of those with preDM to overt T2D.

Insulin resistance has strong environmental influences, including reduced daily physical activity and increased consumption of calorie-dense processed foods and beverages. Other factors implicated in IR also include physiological stressors, systemic inflammation, and oxidative stress. However, IR also has a significant heritable component, and in this review we will discuss the role that large human genetic studies have had in improving our understanding of the inherited basis of IR, in particular revealing a key role of reduced mitochondrial function. We will focus on an emerging explanation of the cellular basis of IR that ties together caloric demand/excess, intracellular nutrient oversupply, lipid-induced cellular alterations in insulin signaling, and the protective role that exercise may play at a cellular level.

Insulin action in nutrient oversupply and the relationship with ectopic fat-induced IR

Insulin action fosters the storage of nutrients in key metabolic organs, including skeletal muscle, the liver, and white adipose tissue (WAT). In the fed (non-fasting) state, a rise in blood glucose stimulates the secretion of insulin, and the binding of insulin to its receptor results in tyrosine auto phosphorylation (2), leading to phosphatidylinositol 3 (PI3)-kinase and Akt activation and translocation of glucose transporter type 4 (GLUT4) to the plasma membrane, glucose uptake, and glycogen synthesis in skeletal muscle. In WAT, insulin action increases glucose uptake and suppresses lipolysis. In the liver, insulin signaling through Akt promotes glycogen synthesis and inhibits gluconeogenesis. Taken together, IR is associated with decreases in insulin-mediated glucose uptake in skeletal muscle (with concomitant decreased glycogen synthesis) (3–5) and WAT (with concomitant increases in lipolysis and therefore increases in plasma fatty acid [FA] levels) (6), and an increase in hepatic glucose production. The deleterious results of IR in the separate tissues can negatively reinforce each other. For instance, excess FAs act to decrease insulin-stimulated muscle glucose uptake, glycogen synthesis, and glucose oxidation (7–11) as well as increased hepatic acetyl CoA content, which in turn leads to the activation of pyruvate carboxylase activity and increased rates of hepatic gluconeogenesis (12).

Insulin resistance is also associated with an increase in ectopic lipid deposition, and excess intracellular lipids serve as a good predictor of IR in humans (13, 14). Fatty acid metabolites such as sn-1,2 diacylglycerols (DAGs) and ceramides inhibit insulin signaling (15) and intracellular sn-1,2 DAGs are a trigger for lipid-induced skeletal muscle IR through the activation of novel protein kinase Cs (PKCθ, PKCε) in both human and rodent models (Fig. 1) (16–21). Protein kinase C activation and the phosphorylation of insulin receptor substrate (IRS) on serine residues leads to decreased insulin-stimulated glucose transport (16, 22). PKCθ-knockout mice are resistant to lipid-mediated IR (23). Increased serine phosphorylation of IRS-1 and decreased insulin-meditated Akt activation are observed in muscle biopsy samples obtained from the insulin resistant offspring of patients with T2D (24). Moreover, muscle IR caused by lipid-induced an accumulation of intramyocellular DAGs is associated with an increase in activity of PKCβII and PKCδ, and a reduction in IκB-α, an inhibitor of nuclear factor-kappaB (17). Besides DAGs, the buildup of ceramides in skeletal muscle may also contribute to IR by blocking Akt activation (25, 26). Excess intramyocellular triglycerides (TGs) are not always associated with insulin resistance as observed in elite, insulin sensitive endurance athletes (27). These results are consistent with prior studies that have dissociated TGs as the mediator of IR in muscle and the liver and reflects the importance of other lipid metabolites (eg, DAGs, ceramides) as the trigger of IR. Recent studies have also demonstrated that the compartmentation of sn-1,2 DAGs in the plasma membrane is also likely an important factor in mediating liver and muscle IR.

Figure 1.

Mechanism of sn-1,2 DAG-PKCθ/PKCε-mediated skeletal muscle IR. Increase in sn-1,2 DAGs caused by reduced mitochondrial function and fatty acid oversupply activates PKCθ and PKCε, leading to serine phosphorylation of IRS-1 and increased threonine phosphorylation of the insulin receptor on threonine 1160. Both of these in turn result in inhibition of insulin-induced PI3K/Akt2 activity. As a consequence, insulin-induced translocation of GLUT4 to the plasma membrane, glucose uptake, and glycogen synthesis are decreased.

Abbreviations: DAG, diacylglycerol; GLUT4, glucose transporter type 4; GS, glycogen synthase; IRS, insulin receptor substrate; PCK, protein kinase C; PI3K, phosphoinositide-3-kinase; NAT2, N-acetyltransferase 2.

Fat in the liver is also strongly associated with hepatic IR, and IR in the liver has many similarities to skeletal muscle. As with skeletal muscle, lipid-induced hepatic IR is associated with increased sn-1,2 DAGs in the plasma membrane. In the liver, accumulated sn-1,2 DAGs activate PKCε, which in turn stimulates threonine 1160 phosphorylation of the insulin receptor kinase (IRK) at its catalytic subunit (Fig. 2) (19, 28–31). As a consequence, increased IRK phosphorylation leads to decreases in tyrosine phosphorylation, IRK activity, and Akt phosphorylation, causing an inhibition of insulin signaling. In addition to the inhibition of insulin signaling at the level of the insulin receptor kinase, other targets of PKCε activation lead to the inhibition of insulin signaling downstream of the insulin receptor (32). Strong correlations between hepatic DAG and IR, and DAG content and PKCε activation are observed in liver biopsies from obese, nondiabetic individuals (33). Similar observations regarding a relationship between hepatic IR and DAG content are also reported by other groups (34–37). Thus, these findings confirm a detrimental effect of lipid excess (especially sn-1,2 DAGs in the plasma membrane) on insulin action, implying that ectopic lipid accumulation occurs early during the development of IR.

Figure 2.

Mechanism of sn-1,2 DAG-PKCε-mediated hepatic IR. Accumulation of sn-1,2 DAGs promote a translocation of PKCε to the plasma membrane, and PKCε phosphorylates insulin receptor at Threonine 1160, inhibiting insulin receptor kinase activity. Lowered insulin-stimulated PI3K/Akt activity leads to decreased insulin-stimulated glycogen synthesis through a reduction of GSK3/GS activity, and decreased insulin inhibition of gluconeogenesis through a reduction of FOXO1 phosphorylation. Increased FOXO1 nuclear translocation promotes transcription of gluconeogenic enzymes such as PEP-CK and G6P. Thus, DAG inhibits a direct effect of insulin on suppressing hepatic glucose production by inhibiting insulin-stimulated hepatic glycogen synthesis and increasing insulin-mediated transcription of gluconeogenic enzymes. Insulin action on WAT also indirectly regulates hepatic glucose production. In insulin resistant WAT, insulin inhibition of lipolysis is decreased, resulting in increased lipolysis and fatty acid and glycerol fluxes. Increased delivery of fatty acids to the liver leads to an increase in hepatic acetyl-CoA content and PC activity, promoting hepatic gluconeogenesis. Increased glycerol flux to the liver fosters a conversion of glycerol to glucose.

Abbreviations: DAG, diacylglycerol; FOXO1, transcription factor forkhead box O1; G6P, glucose-6-phosphatase; GS, glycogen synthase; GSK3, glycogen synthase kinase 3; IRS, insulin receptor substrate; NAT2, N-acetyltransferase 2; PC, pyruvate carboxylase; PCK, protein kinase C; PEP-CK, Phosphoenolpyruvate carboxykinase; PI3K, phosphoinositide-3-kinase; WAT, white adipose tissue.

Similar to accumulating an excess amount of fat in the body, having no dedicated WAT depots is problematic largely due to the resultant overflow of lipid, leading to ectopic lipid deposition in the liver and skeletal muscle. Severe IR and T2D are present in A-ZIP fatless mice, which is associated with defects in insulin activation of IRS-1/IRS-2-associated PI3-kinase activity, and a 2-fold increase in muscle and liver lipid content (38). Transplantation of fat tissue into these fatless mice normalizes IR and lipid content in muscle and the liver. Patients with lipodystrophy are markedly insulin resistant and leptin treatment both reduces abnormal levels of muscle and hepatic triglyceride content and improves insulin sensitivity (39). These results suggest that IR is not about how much fat we have, but about how fat distributes in the body, and that it is the intracellular lipid moieties (eg, sn-1,2 DAGs in the plasma membrane) that are responsible for triggering defects in insulin signaling and action.

IR and reduced mitochondrial function

Mitochondria use fat for energy production and decreased mitochondrial function is associated with increases in ectopic fat and IR. Resting adenosine triphosphate (ATP) synthesis in skeletal muscle in insulin resistant subjects is reduced when compared with insulin sensitive individuals, suggesting a contribution of mitochondrial dysfunction to IR (40). Reduced mitochondrial function in healthy, young, lean, insulin resistant offspring of patients with T2D is accompanied by an increase in the intramyocellular lipid concentration, a reduction of mitochondrial density/content, and a decrease in rates of oxidative phosphorylation in muscle mitochondria (14, 24, 41). Most (but not all) studies in humans (42) strongly support the association of reduced mitochondrial function with IR, as can be seen in the following:

• A decrement in mitochondrial number and electron transport chain activity in T2D and obese individuals compared with lean volunteers (43).

• A deterioration of mitochondrial function in skeletal muscle obtained from obese type 2 diabetic subjects (44).

• Nonresponsiveness of muscle mitochondrial ATP production to high-dose insulin infusion in type 2 diabetic subjects, suggesting impaired response to insulin and reduced mitochondrial function (45, 46).

• A modest decrease in mitochondrial ATP synthesis rates in nonobese patients with T2D and in nondiabetic older individuals when compared to nondiabetic young groups under fasting conditions and after insulin stimulation (47).

• Lower mitochondrial function caused by a reduction of basal ADP-stimulated and intrinsic mitochondrial respiratory capacity in type 2 diabetic subjects when compared with age- and BMI-matched control subjects (48, 49).

Mitochondrial malfunction and oxidative stress are inextricably linked and both are associated with IR and T2D (50). Mitochondria are the main source of reactive oxygen species (ROS), including superoxide anion and hydrogen peroxide (H₂O₂), whose overproduction underlines oxidative stress (51). During mitochondrial ATP production, electrons are transferred from carriers in the electron transport chain Complex I through Complex IV, where these electrons reduce oxygen to water (52). Single electrons that are not passed down on to Complex IV react with oxygen to form superoxide, which is the proximal mitochondrial ROS. Excessive ROS production reduces mitochondrial function and subsequently generates more ROS in a vicious cycle, as can be seen in the following:

• Excess mitochondrial ROS production and reduced mitochondrial function are enhanced by elevated FAs and/or hyperglycemia (53–55).

• Mitochondrial ROS attenuate insulin action in adipocytes, myotubes, and mice (56), and abolish insulin-stimulated GLUT4 translocation in 3T3L1 cells by interfering with the insulin-mediated redistribution of IRS-1 and PI3-kinase (57, 58).

• Obese, insulin-resistant humans and insulin-resistant rats fed a high fat diet have an enhancement of H₂O₂ produced by skeletal muscle mitochondria, suggesting a correlation of mitochondrial H₂O₂ emission to the pathogenesis of IR (59).

Similar to humans, aging in mice promotes mtDNA damage and reduced mitochondrial activity in skeletal muscle, which in turn is associated with increased intramyocellular lipid content and DAG-nPKC–mediated IR. All of these aging-associated effects are abrogated in mice with catalase targeted to the mitochondria, an enzyme that reduces H₂O₂ accumulation in mitochondria by catalyzing H₂O₂ to H₂O and oxygen (60). Moreover, branched-chain amino acid (BCAA) catabolism is linked to age-related reductions in mitochondrial biogenesis and ROS defense system (61), and decreased expression of branched-chain amino acid transaminase 1 (BCAT1), an enzyme involved in the initiation of BCAA catabolism, is observed during aging in C. elegans, zebrafish, and mice (62). Knockdown of BCAT1 in breast cancer cells lowers mitochondrial number, biogenesis, and function by repressing mTOR signaling, suggesting a role of mTOR signaling in BCAAs increased mitochondrial biogenesis and function (63).

Reduced mitochondrial function and IR: What is the chicken and what is the egg?

Given the links between caloric intake, physical activity, obesity, abnormal body fat distribution, and glucose metabolism at a systemic level, it is not surprising that a large body of literature supports a connection among ectopic lipid deposition, IR, and mitochondrial malfunction at a cellular level. Nevertheless, the studies discussed above cannot conclusively determine whether reduced mitochondrial function is a cause or consequence of IR. Studies in mice and in humans have shed light on this important issue.

Reduced mitochondrial function may serve as a primary causal etiology of IR in certain conditions.

It has been known for some time that diseases associated with severely reduced mitochondrial function, such as mitochondrial encephalomyopathy, are associated with IR (64–69). In mice, a specific deletion of skeletal muscle Tfam, a gene encoded for a key transcriptional regulator of mitochondrial proteins, have abnormal mitochondrial morphology, impaired respiratory chain, and increased mitochondrial mass in skeletal muscle (70, 71). However, improved glucose clearance, increased glucose uptake into the muscle, and no alteration in muscle ATP production are observed in the muscle-specific Tfam knockout mice. Similar observations are demonstrated in adipocyte-specific deletion of Tfam, whose mitochondrial DNA copy number and protein mitochondrial contents were decreased in adipose tissues (72).

However, it is important to note that not all studies have not observed an association between reduced mitochondrial function and IR and, in fact, have observed improved insulin sensitivity in certain animal models with reduced mitochondrial function. Oxidative damage and activation of the mitochondrial unfolded protein response in skeletal muscle of polymerase gamma mitochondrial DNA mutator (POLG) mice are associated with elevated expression of growth differentiation factor 15 (GDF15), a circulating factor that may protect against diet-induced obesity and IR (73). Improved whole-body insulin sensitivity is also observed in mice with a deficiency of mitochondrial fusion protein optic atrophy 1 (OPA1) in muscle that develop progressive nonlethal mitochondrial dysfunction (74). Moreover, a reduction in ATP/ADP exchange in mice with a deficiency of adenine nucleotide translocator isoform 1 (ANT1) results in insulin hypersensitivity and a resistance to high fat diet-induced toxicity, which are associated with mitochondrial hyperproliferation in skeletal muscle (75). Thus, these observations imply that “pure” mitochondrial dysfunction is not always an etiology of IR development. One possible explanation for the dissociation of IR from reduced mitochondrial function in these mouse models might have to do with the severity of reduction in mitochondrial function in these models leading to major reductions in the intracellular energy charge, which in turn would result in increased anaerobic glucose metabolism/adenosine monophosphate-activated protein kinase (AMPK) activation, which in turn would be predicted to lead to reduced plasma glucose and insulin concentrations. On the other hand, mice model mild (10–30%) reductions in mitochondrial function (as is typically observed in insulin resistant humans), predisposition to increased plasma membrane sn-1,2 DAG accumulation, and IR in the absence of major reductions in intracellular energy charge and anaerobic metabolism.

Decreased insulin action as an etiology of mitochondrial dysfunction.

In healthy subjects, hyperinsulinemia promotes increased muscle mitochondrial ATP production, protein synthesis, and enzyme activities of cytochrome c oxidase and citrate synthase. However, insulin resistant humans and rodents manifest blunted mitochondrial response to insulin (45, 76). Consistent with these findings, individuals with T2D also manifest reduced insulin-stimulated rates of muscle mitochondrial ATP (46). Moreover, double-deletion of both IRS-1 and IRS-2 in skeletal muscle in mice results in disrupted insulin sensitivity (77) accompanied by impaired oxidative phosphorylation and ATP production. Mice lacking both IRS-1 and IRS-2 in the liver become resistant to insulin and have larger and lucent mitochondria, and decreased mitochondrial number accompanied by impaired mitochondrial respiration, ATP production, and electron transport chain activity (78–80). When the transcription factor FOXO1 is abolished in the IRS-1/IRS-2 double knockout mice, abnormalities of mitochondrial morphology and function in the liver are restored, indicating that the signaling is mediated through IRS/PI3-kinase/FOXO pathway. Furthermore, increases in IRS-1 serine phosphorylation and a reduction in insulin-stimulated Akt activation are associated with reduced mitochondrial density and mitochondrial oxidative phosphorylation in skeletal muscle in young, lean, insulin-resistant, first-degree relatives of patients with T2D (24). Taken together, these data indicate a direct effect of disrupted insulin signaling and the subsequent IR on mitochondrial function, implying that defects of mitochondrial function may also, under certain circumstances, be a consequence of IR.

Mitochondrial function and insulin action in response to exercise.

Multiple lines of evidence from human studies suggest the beneficial effect of exercise in mitigating IR and decreasing the risk of T2D. Acute (81–86) and/or chronic (85, 87, 88) exercise interventions in healthy, insulin resistant and diabetic humans have consistently demonstrated improvements in insulin action, weight loss, ectopic fat distribution glucose tolerance, GLUT4 induction, phosphorylation of IRS-1 and Akt, inflammatory profiles (89), mitochondrial biogenesis, mitochondrial size, mitochondrial number, mitochondrial oxidative activity, mitochondrial enzyme activity (90), and ATP production (87). Decreases in hepatic de novo lipogenesis and hepatic TGs are also observed (86). Many of these findings are based on studies performed on skeletal muscle or liver biopsies of human pre- and postexercise interventions. Similar findings have been observed in rodent models (91–95).

Thus, a large body of data from human physiological and interventional studies suggest that IR is tightly linked to, and potentially causal for, reduced mitochondrial function, and that this may be at least partly due to adverse effects of ectopic lipid deposition, which can be mitigated by exercise. Nevertheless, these studies have not yet been able to explain the heritable nature of IR or to suggest why certain humans and human populations are more insulin resistant than others (96).

Human genetic studies implicate decreased mitochondrial function and ectopic lipid deposition with IR

Genetic analysis has been a powerful tool to identify metabolic risk loci, and alleles (sometimes) that contribute to IR and T2D risk. Other than a few very rare Mendelian forms of IR (lipodystrophy, IRS1 mutations [97–101]), genetic susceptibility to IR in human populations is complex genetically and moderately heritable. With the widespread application of genome-wide association studies (GWAS) over the last decade or so, we have been able to gain great insight into the genetic architecture of many complex conditions, including T2D.

We and others have shown that genetic susceptibility to T2D can be mediated by genetic variability at sites related to several processes, including β-cell function, proinsulin production, obesity, IR, and lipodystrophy/peripheral adipose storage capacity (102–105). Recognition of these “clusters” driving T2D risk not only provides biological insight into genetic susceptibility but also has clinical implications as certain clusters, especially the IR and lipodystrophy clusters, are more strongly associated with adverse cardiovascular outcomes (102).

While the majority of T2D susceptibility variants appear to be affecting pancreatic β-cell function and/or insulin production/secretion, several large studies have reported the identification of common genetic variation associated with surrogate measures of IR, including fasting plasma insulin concentrations or insulin and glucose responses during oral glucose tolerance tests (which are more highly correlated with reference measures of IR) (103, 104, 106–108). Some insulin resistant loci (eg, FTO, MC4R) clearly act through increasing the risk of obesity. Further refinement using cluster analyses incorporating additional phenotypic data has strongly implicated other loci that increase the risk of IR by decreasing peripheral adipose storage and increasing ectopic lipid deposition. These loci include (but are not limited to): IRS1, PPARG, GRB14, KLF14, ARL15, ADCY5, LYPLAL1, ADAMTS9, MACF1, and POU5F1 (102, 105, 109). The causal gene and molecular mechanism of action leading to IR is only understood for a small number of these loci (eg, IRS1, PPARG).

In addition, other studies have implicated a connection between IR and reduced mitochondrial function. Among these is our GWAS of whole-body insulin sensitivity assessed with gold standard methods in ~5000 Europeans, which identified human N-acetyltransferase 2 (NAT2) as a novel insulin sensitivity gene (110). Mice deficient of Nat1 (the mouse ortholog of human NAT2) have decreases in insulin sensitivity, basal metabolic rate, exercise capacity, and fat utilization, and marked increases in levels of blood glucose, insulin, TGs, and liver and muscle lipid content (30, 111). Studies in mouse adipocyte and skeletal muscle model systems demonstrated a decline in mitochondrial function characterized by increases in ROS and fragmentation, and decreases in ATP production, mitochondrial membrane potential, mitochondrial mass, and biogenesis as a result of Nat1 deficiency (111). Nat1-deficient mice are prone to fat buildup in the liver and muscle, and IR through the same sn-1,2 DAG–nPKC mechanism as previously discussed, where reduced mitochondrial fat oxidation promotes the accumulation of plasma membrane DAGs, resulting in liver and muscle IR.

Other recent works from Rusu et al (112) and Hoch et al (113) support the general idea that altered lipid metabolism in the liver affecting mitochondrial fatty acid oxidation and intracellular DAG levels may be a common pathway towards increased T2D susceptibility. Starting with GWASs, these investigators identified a risk haplotype responsible for 20% of the increased prevalence of T2D in Mexico. Subsequent fine mapping studies identified the SLC16A11 gene as the causal gene in the locus and functional studies demonstrated that this gene is an H⁺ coupled monocarboxylate transporter. Steady-state levels of intracellular acylcarnitines, DAGs, and TGs are significantly increased in human hepatocytes treated with siRNAs targeting SLC16A11, which is consistent with the observed effect of the risk allele, which decreases SLC16A11 expression and transport activity in the liver. While the full cellular mechanisms by which SLC16A11 deficiency leads to impaired mitochondrial function and metabolic processes remain unknown, the authors suggest that the effect on fatty acid β-oxidation by the mitochondria and increases in DAGs and TGs are consistent with those seen in the pathophysiology of IR and T2D (114–119). In contrast, studies by Zhao et al reported minimal metabolic effects in Slc16a11 knockout mice compared to wild-type control mice (120, 121). In the Slc16a11 knockout mice, re-expression of a mutant murine Slc16a11 (intended to model the human T2D-risk coding variants) leads to increases in liver TGs, glucose intolerance, and IR. Elevated TGs are also detected in hepatocytes overexpressing the mutant Slc16a11. Given that SLC16A11 has been linked to IR and T2D, further investigations are needed to address controversial data on how T2D variants affect SLC16A11 functional expression.

Taken together, these results suggest an association of reduced mitochondrial activity with decreased insulin sensitivity, and that mitochondrial malfunction (caused by perturbations in genes like NAT2 or SLC16A11) results in the ectopic accumulation of lipid in muscle and the liver even in the absence of obesity. Although additional supportive data are needed, these studies suggest that strategies to increase the activity of mitochondria and β-oxidation (including exercise) can prevent ectopic fat (and intracellular DAG) buildup in the liver and muscle, thereby improving insulin sensitivity.

Future Perspectives and Conclusions

Studies in the past years provide the evidence that intracellular lipid accumulation, abnormal lipid distribution, and altered mitochondrial function are closely associated with IR and T2D. Human genetic studies have lent unique mechanistic insights into the molecular and inherited basis of IR and have complemented clinical and physiological observations. Recent studies in both rodents and humans suggest that IR can result in reductions in mitochondrial function and conversely genetic alterations in certain genes (eg, NAT2, SLC16A11) can lead to reduced mitochondrial function and promote increased plasma membrane accumulation of DAGs, resulting in increased nPKC activation, and liver and muscle IR.

Future studies aimed at determining the exact molecular mechanisms for genes like NAT2 and SLC16A11 will be crucial to develop targeted therapies for IR and will help elucidate the potential role of genetic causes of mitochondrial dysfunction in the pathogenesis of metabolic syndrome, nonalcoholic fatty liver disease (NAFLD)/nonalcoholic steatohepatitis (NASH) and T2D.

Additional Information

Disclosure Summary: P.S. and J.W.K. declare no relevant financial relationships. G.I.S. is on the scientific advisory boards for Merck, NovoNordisk, Gilead Sciences, AstraZeneca, Aegerion, iMBP, and Janssen Research and Development and receives investigator-initiated support from Gilead Sciences, Merck, and AstraZeneca.

Data Availability. Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.