Liver injury in the setting of steatosis: Crosstalk between adipokine and cytokine


  • Mark J. Czaja

    Corresponding author
    1. Department of Medicine and Marion Bessin Liver Research Center, Albert Einstein College of Medicine, Bronx, NY
    • Department of Medicine and Marion Bessin Liver Research Center Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461
    Search for more papers by this author
    • fax: (718) 430-8975

  • See Articles on Pages 46 and 177.

Central to the mechanisms underlying a variety of forms of liver injury are factors produced from the accompanying inflammatory response. Liver injury triggers the recruitment and activation of macrophages and neutrophils, a process mediated in part by lipopolysaccharide.1 These activated cells produce a number of potentially injurious factors that may promote hepatocyte injury. The most critical of these is the cytokine tumor necrosis factor (TNF)-α. Through either direct toxic or proinflammatory effects, TNF may promote injury from toxins,2 ischemia/reperfusion,3 and hepatitis viruses.4 Strategies to block the production, activity, or death pathway signaling of TNF therefore represent potential therapies for a variety of liver diseases. However, while progress has been made in understanding the mechanisms of hepatocyte sensitization to TNF cytotoxicity, effective means of suppressing this effect in vivo are still lacking.


BMI, body mass index; GalN, galactosamine; LPS, lipopolysaccharide; PPAR, peroxisome proliferator-activated receptor; TNF, tumor necrosis factor; TNFR, TNF receptor.

Crucial to understanding liver injury in nonalcoholic fatty liver disease (NAFLD) is the delineation of mechanisms of progression from benign fatty liver to hepatocyte injury and steatohepatitis. Several facts suggest that the inflammatory response and TNF in particular may promote this liver injury. First, the disease lacks an apparent direct death stimulus such as a hepatotoxin or hypoxia, suggesting that indirect factors must cause cell injury. Second, steatohepatitis frequently occurs in the setting of obesity and insulin resistance or diabetes which are increasingly recognized as proinflammatory states with increased oxidative stress, cytokine production, and cellular stress pathway signaling.5, 6 Obesity in particular is associated with increased TNF production from fat stores. The ability of TNF to act as a hepatotoxin, and the presence of increased levels of this cytokine in conditions associated with NAFLD, make TNF a prime candidate to promote progression to steatohepatitis. This possibility is supported by studies in leptin deficient ob/ob mice with obesity and fatty liver in which TNF inhibition reduced steatosis and liver injury.7, 8 However, studies of ob/ob mice lacking type I and II TNF receptors (TNFR1 and TNFR2) have suggested that TNF is not involved in their liver disease.9 How TNF could cause steatotic cell injury is unclear because TNF is normally a mitogen unless hepatocellular resistance mechanisms to the toxic effects of TNF are somehow circumvented. One possibility is that hepatocyte sensitization to TNF injury in the setting of NAFLD results from the cellular effects of overexpression of the pro-oxidant enzyme cytochrome P450 2E1 in this disease.10

An additional factor in the regulation of liver injury occurring in the setting of obesity and insulin resistance is the influence of adipocyte produced proteins.11 Adipocytes not only store excess energy, but also respond to metabolic signals by secreting proteins that exert local, central, and peripheral effects. Principal among these adipocyte specific or enriched factors, or adipokines, are leptin, resistin, and adiponectin. Leptin has multiple actions that include decreasing food intake and increasing energy expenditure. Serum leptin levels increase in proportion with body mass index (BMI). Whether leptin levels are altered in NAFLD is controversial, with some studies demonstrating increased levels in this disease,12 but others finding no correlation between serum leptin and the development of steatohepatitis.13 Resistin is an adipocyte produced protein whose main effect is to increase hepatic glucose production.14 Serum resistin levels are increased in human obesity,15 but levels in NAFLD have not yet been examined. Thus, no clear evidence yet implicates either leptin or resistin in the development of NASH.

Adiponectin is present in significant concentrations in human serum (5–30 nM), and circulates in several forms including dimers, trimers, and a high-molecular weight complex consisting of up to six trimers.16 Two adiponectin receptors, AdipoR1 and AdipoR2, have been cloned.17 Liver expresses both receptor genes and has the highest expression of AdipoR2 among organs. Adiponection is secreted by adipocytes in inverse proportion to BMI.18 Serum adiponectin levels are also reduced with insulin resistance and diabetes.19 Metabolically adiponectin acts to reduce body fat,20 improve hepatic and peripheral insulin sensitivity,21 and decrease serum fatty acid levels in association with increased fatty acid oxidation in muscle.22 However, adiponectin has significant anti-inflammatory as well as metabolic effects. Adiponectin blocks macrophage phagocytosis and lipopolysaccharide-induced TNF release in vitro, possibly through inhibition of NF-κB activation.23–25 These dual metabolic and anti-inflammatory beneficial effects of adiponectin have been utilized to effectively treat atherosclerosis.26

Recent studies have now suggested that adiponectin can also prevent liver disease. Initially adiponectin null mice were reported to develop more extensive carbon tetrachloride-induced hepatic fibrosis than wild-type mice.27 A direct antifibrotic effect of adiponectin was suggested by findings of adiponectin receptor gene expression in hepatic stellate cells, and the inhibition of stellate cell proliferation, migration, and transforming growth factorβ1 expression by adiponectin treatment.27 A second study examined the effects of adiponectin administration on both ob/ob mice and mice fed a high fat, ethanol-containing diet.28 In both models, adiponectin significantly decreased levels of steatosis, liver injury and serum TNF. Ethanol-induced steatohepatitis was associated with a reduction in serum adiponectin levels, suggesting that a relative deficiency of this adipokine may have promoted liver disease.

The article by Masaki et al. in this issue of HEPATOLOGY further expands the spectrum of adiponectin's protective effects during liver injury.29 The authors examined the effect of adiponectin administration on liver injury in KK-Ay obese mice sensitized to acute hepatotoxicity from LPS or TNF by cotreatment with the toxin galactosamine (GalN). Obese mice were markedly more sensitive to both forms of injury than lean controls, consistent with previous findings of LPS sensitivity in obese mice and rats.30 As expected, the obese mice had a 30% decrease in serum and adipose levels of adiponectin. Pretreatment with adiponectin reduced mortality, transaminase elevations and the amount of apoptosis induced by GalN/LPS by approximately 50%. This reduction in liver injury was associated with marked decreases in serum and hepatic TNF levels that rose with GalN/LPS treatment, leading the authors to conclude that the protective effect of adiponectin was mediated by its inhibition of TNF synthesis or release. However, adiponectin also significantly decreased liver injury from GalN/TNF administration in this study, suggesting that the major mechanism of adiponectin's action was downstream of TNF activation. The reduction in TNF could have been a secondary manifestation of the decrease in liver injury and therefore the stimulus for inflammation. It would be interesting to determine in this model whether adiponectin inhibits NF-κB activation, and therefore the production of other cytokines that may modulate liver injury. An alternative mechanism of adiponectin's effect may have been the ability of this adipokine to prevent the GalN/LPS-induced reduction in peroxisome proliferator-activated receptor-α (PPAR-α) expression, because PPAR-α-mediated signaling has been previously demonstrated to block liver injury in experimental NAFLD.31 While the mechanism of adiponectin's protective effect requires further study, these investigations provide the first demonstration that adiponectin can prevent acute hepatic injury from LPS/TNF in a steatotic liver.

The exciting evidence of protective effects of adiponectin in liver injury in obese mice raises the possibility that adiponectin may be a treatment for human NAFLD. Hui and colleagues provide additional support for this concept in a second article in this month's Journal.32 In a study of over 100 patients with NAFLD, multivariate analysis revealed that decreased serum adiponectin levels and increased TNF and soluble TNFR2 levels correlated with the presence of NASH independent of the presence of insulin resistance. Levels of adiponectin were lower in NASH than in simple steatosis, and correlated with the degree of hepatic necroinflammation. This important study provides both concrete evidence for the involvement of TNF and adiponectin in human NAFLD, and suggests that adiponectin may be a critical factor in the progression of this disease.

These two studies along with previous investigations suggest an emerging concept of the interrelationship of adiponectin with the liver (Fig. 1). In addition, they provide three forms of cogent evidence that support further study of adiponectin as a therapy for NAFLD: (1) this molecule has metabolic and anti-inflammatory properties that could alleviate both the fat accumulation and liver injury that mark this disease; (2) adiponectin has been effective in inhibiting liver injury in two animal models of hepatic steatosis; and (3) NASH may occur in the setting of relative adiponectin “deficiency.” A possible mechanistic scenario for NAFLD is that conditions such as insulin resistance and obesity lead to increased levels of fatty acids and the development of hepatic steatosis. However, these states also suppress adiponectin levels leading to a proinflammatory condition and the generation of injurious factors such as TNF. Individual variation in the degree of adiponectin suppression, the amount of inflammatory cell activation, and/or the hepatic susceptibility to injury from inflammatory mediators may determine which individuals progress to NASH. However, the actions of adiponectin are likely to be more complex than simple TNF suppression, and may include direct protective effects on hepatocytes and anti-fibrotic effects on hepatic stellate cells. Thus, adiponectin therapy is potentially applicable to liver diseases other than NAFLD. The role of adiponectin in other steatotic diseases such as alcoholic steatohepatitis, and any liver disease dependent on inflammatory mediators, requires further study as well.

Figure 1.

Effects of adiponectin on hepatic steatosis and injury. Adiponectin produced by adipose tissue has both physiological effects on lipid homeostasis (blue arrows) and anti-inflammatory effects (red lines) that may modulate hepatic steatosis and injury. Local effects of adiponectin on adipose tissue include a reduction in body fat and tumor necrosis factor (TNF) production. Adiponectin's peripheral effects on muscle lead to increased fatty acid oxidation and insulin sensitivity. The net result of these effects on adipose and muscle tissue is a decrease in serum fatty acid levels that may prevent hepatic fat accumulation. In the setting of hepatic injury, adiponectin may additionally act to inhibit both liver injury and fibrosis. This adipokine may inhibit Kupffer cell (KC) activation and release of injurious substances such as the cytokine TNF. This effect together with reduced adipose production of TNF may prevent hepatocyte injury. Adiponectin may also have direct effects on the hepatic stellate cell (HSC), blocking its proliferation and secretion of the profibrogenic cytokine transforming growth factor-β1 (TGF-β1). Finally, adiponectin also has as yet poorly described central effects on the brain that may impact on liver steatosis, injury or fibrosis.