Nonalcoholic fatty liver disease (NAFLD) is now the most common liver disease in the United States, and it has quickly become a global problem, paralleling the rise in the prevalence of obesity. In the hunt for a dietary trigger for NAFLD, fructose has been identified as a tempting candidate, particularly because of its tendency to promote visceral adiposity and hypertriglyceridemia.1 Fructose is a six-carbon ketone sugar that is common in the diet and makes up approximately 10% of calories for a typical American.2 The primary sources for fructose are added sugars, such as high fructose corn syrup and sucrose, that are found in beverages and processed foods.2 Growing clinical evidence links fructose consumption and NAFLD.3, 4 The article by Spruss et al.5 examines one possible mechanism for fructose-induced hepatic steatosis in mice, namely, the interaction(s) between Toll-like receptor 4 (TLR4) and gut-derived endotoxin [lipopolysaccharide (LPS)].
Fructose metabolism is distinct from glucose metabolism. Dietary fructose is absorbed into the intestine via a saturable, facilitative glucose transporter (GLUT5), and healthy persons have the ability to absorb up to 25 g; malabsorption occurs at higher doses and can lead to increased fructose fermentation by gut bacteria.6 After transport across the basolateral membrane, fructose is taken up by the liver with a high rate of extraction in comparison with glucose. In the hepatocyte, fructose is rapidly phosphorylated to form fructose-1-phosphate in a reaction catalyzed by fructokinase. The next steps in fructose metabolism result in the production of glyceraldehyde, dihydroxyacetone phosphate, and glyceraldehyde-3-phosphate. This is the point at which glucose metabolism and fructose metabolism merge; however, fructose metabolites reach this stage without passing through the rate-limiting step of phosphofructokinase, effectively avoiding the regulating action of insulin. It is this lack of regulation that has been often hypothesized to contribute to the differential effects of fructose feeding versus glucose feeding.1
Research has clearly shown that fructose feeding of animals (reviewed by Havel1) and humans results in increased de novo lipogenesis, elevated plasma triglyceride levels, increased hepatic lipids, and visceral adiposity.6–8 Oxidative stress has been shown in animals, and more recently, several articles have demonstrated links between oxidative stress and fructose in humans as well.7, 9 However, the mechanisms for these effects are not known. In their article, Spruss et al.5 use a fructose-fed animal model to examine possible mechanisms for the development of steatosis, specifically looking at the role of innate immunity, bacterial overgrowth, and endotoxin using a TLR4 mutant mouse model. The authors seek to demonstrate an independent effect of intestinal translocation of bacterial endotoxin as a partial or additional source of the increased hepatic steatosis seen in the model, rather than simply the effects assumed from rapid and relatively unregulated hepatic fructose metabolism. They demonstrate that fructose-fed mice had endotoxemia, hepatic steatosis with increased alanine aminotransferase, indicators of oxidative stress, and increased myeloid differentiation protein 88 (MyD88) and tumor necrosis factor (TNF) messenger RNA production, all of which, except endotoxemia, were attenuated/blocked in TLR4 mutant mice.
The approach taken by Spruss et al.5 is similar to that used in animal models of alcoholic liver disease. It is well documented that alcohol increases gut permeability through multiple mechanisms, including alterations in signaling molecules such as nitric oxide, effects of alcohol/acetaldehyde on gut barrier proteins, and zinc deficiency induced by chronic alcohol intake.10 Moreover, alcohol-fed animals and humans have alterations in gut flora.10, 11 These gut alterations are associated with endotoxemia, activation of TLR4 signaling, increased TNF production, and liver injury. Moreover, when C3H/HeJ TLR4 mutant mice are fed alcohol, fatty liver is attenuated, and liver injury is blocked.12 Multiple research groups have used similar strategies in diet-induced NAFLD produced with a high-fat diet. Mice consuming a high-fat diet exhibit significantly increased plasma LPS, which has been postulated to play a major role in complications of metabolic syndrome, and this has been termed metabolic endotoxemia (reviewed by Cani and Delzenne13). High-fat-fed rodents also have an altered gut microbiota profile and increased gut permeability. Mice fed a high-fat diet develop not only NAFLD but also other complications of metabolic syndrome, such as vascular inflammation and insulin resistance, which are markedly attenuated in animals lacking the TLR4 receptor (reviewed by Cani and Delzenne13). Elegant work from Cani and coworkers14 demonstrated that mice chronically infused with a very low dose of LPS (to simulate blood levels induced by a high-fat diet) mimic the high-fat phenotype of obesity, steatosis, hepatic insulin resistance, and so forth. Lastly, translational research has shown that high-fat feeding in humans causes low-grade postprandial metabolic endotoxemia (reviewed by Cani and Delzenne13). The endotoxin absorption following high-fat feeding has recently been postulated to occur not only through increased gut permeability but also through binding to chylomicrons formed after high fat consumption.15 Thus, both chronic alcohol intake and high fat intake mimic the gut permeability, endotoxemia, increased hepatic TNF production, and hepatic steatosis observed by Spruss et al. with high fructose feeding, and the liver/metabolic abnormalities in all three models are blocked/attenuated with a loss of TLR4 function (Fig. 1).
Toll-like receptors are highly conserved pattern recognition receptors that function as pathogen sensors and play a critical role in the innate immune system, and endotoxin is the best studied TLR4 ligand. In liver disease, the role of gut-derived endotoxin has been recognized for over a half-century. Pioneering studies by Broitman et al.16 showed that neomycin attenuated liver injury and blocked cirrhosis in choline-deficient rats. When endotoxin was added to the diet in neomycin-treated animals, liver injury and cirrhosis again developed. Very importantly, it is increasingly clear that there are other agents that activate TLR4 in addition to endotoxin, with certain saturated fatty acids, such as palmitate, representing one example of potential relevance to NAFLD.17 LPS stimulates an inflammatory response, causes an elevation in levels of free fatty acids, and impairs insulin sensitivity. Endotoxin increases lipolysis in adipose tissue, elevates circulating free fatty acids levels, and induces insulin resistance in rodents. The lipolytic action of endotoxin is mediated by its lipid A moiety and is blocked by anti-endotoxin peptides.18 Fatty acids such as palmitate have been shown to activate macrophages to produce TNF and other cytokines such as interleukin 6.17 In in vivo animal models of obesity/NAFLD, it is unclear if it is a direct effect of endotoxin that is inducing TLR4 activation or whether other mediators such as fatty acids are mediating this activation. In this study by Spruss et al.,5 fatty acid levels were not determined in the liver or serum of these animals. To complicate the TLR4 story further, there are other nonendotoxin ligands for TLR4 that could play a role in NAFLD, with the alarmin, or danger signal high mobility group box 1, being an excellent example. This nuclear protein can be released by cell necrosis or by activation of monocytes/Kupffer cells, and it is increasingly being appreciated as a mediator of several forms of liver injury.19
Certain saturated fatty acids also have been shown to cause activation of TLR4 signaling and inflammation in the hypothalamus, which has major implications for obesity.20 Indeed, both the TLR4 mutation leading to loss of function and the immunological inhibition of TLR4 protected mice from diet-induced obesity. It is important to note that in the study by Spruss et al.5 the TLR4 mutant C3H/HeJ mice weighed less than the wild-type mice fed fructose. Thus, the lower weight gain may also have attenuated the fatty liver observed in the fructose-fed TLR4 mutant mice.
Although there are many similarities between diet/obesity-induced and alcohol-induced hepatic steatosis and liver injury, there also are important differences. For example, TLR4 signaling through the MyD88 pathway appears to be important for diet/obesity-induced hepatic steatosis and liver injury, and that appears to be the case in the fructose model. On the other hand, the MyD88-independent pathway appears critical for alcohol-induced liver injury.21 The reasons for this and which cell types are most critical (e.g., Kupffer cells versus hepatocytes) are unclear at this time. Immunohistochemical studies by Spruss et al.5 indicated that MyD88 was up-regulated in Kupffer cells in their fructose-fed mice. Further defining the interactions of the microbiome, gut permeability, the importance of various TLR4 ligands, and the differences in TLR4 signaling pathways will be critical for understanding alcoholic, high-fat–induced, and fructose-induced fatty liver disease and for defining targets for therapeutic interventions.