Statistics from Altmetric.com
See article on page 206
As obesity creeps in epidemic proportions through increasingly inert and affluent “developed” societies, metabolic imbalance has become the commonest cause of liver disease. Fatty liver, although itself benign, predisposes the liver to NASH or non-alcoholic steatohepatitis, a chronic disorder characterised by steatosis, mixed cell type inflammation, focal hepatocyte degeneration, and perivenular or pericellular fibrosis.1 2 NASH is slowly progressive, occasionally resulting in cirrhosis with the potentially fatal complications of portal hypertension, liver failure, and hepatocellular carcinoma.2 3
Most cases of NASH appear to have a multifactorial aetiopathogenesis. The predisposing factors include obesity, type II diabetes, insulin resistance, hypertriglyceridaemia, and rapid weight loss, each of which can cause hepatic steatosis. The trigger that sets off injury and inflammation against this background of oxidisable fatty acid excess, and the mechanisms that perpetuate steatohepatitis and fibrogenesis are less clear.4 A focus of recent experimental studies has been on biochemical processes that reduce oxygen to reactive oxygen species (ROS); these can potentially damage tissues by the process of oxidative stress. Oxidative stress also increases expression of cellular adhesion molecules, and secretion of chemokines and cytokines, thereby initiating the recruitment of an hepatic inflammatory response.5 6
Weltman and colleagues demonstrated that cytochrome P450 (CYP)2E1 is over expressed in the livers of patients with NASH7 as well as in a model of steatohepatitis induced by feeding rats a high fat, methionine and choline deficient (MCD) diet.8Leclercq and colleagues then showed that CYP2E1 catalyses lipid peroxidation in the murine MCD dietary model,9contributing to profound oxidative stress. Alternatively, in Cyp2e1 nullizygous mice, CYP4A enzymes are upregulated and function as alternative catalysts of microsomal lipoperoxidation in experimental NASH.9 Other studies have implicated activation of the peroxisome proliferator activated receptor α (PPARα) in a pathway leading to NASH.10 Activated PPARα is a transcription factor that governs expression of the peroxisomal and microsomal (via CYP4A) pathways of lipid oxidation, with consequent over production of ROS.
As opposed to the possible primary role of “biochemical oxidative stress”, it is possible that the inflammatory response can be the primary mediator of liver cell injury in steatohepatitis. This is a favoured mechanism in alcoholic hepatitis in which gut derived endotoxin appears to incite necroinflammatory change,6even in the absence of CYP2E1.11 Endotoxin is a potent releaser of proinflammatory and cytotoxic cytokines such as tumour necrosis factor α (TNF-α). The cell types involved include activated macrophages, including Kupffer cells, lymphocytes, and neutrophils. In alcoholic hepatitis,6 and potentially in NASH, such cell types could be recruited to the liver where they not only contribute further to oxidative stress by release of ROS and nitroradicals12 but also release cytokines that contribute to liver cell injury and hepatic fibrogenesis.
The distinction between NASH and steatosis without inflammation, injury, or fibrosis is never clearer than in theob/ob leptin deficient mouse. This animal exhibits exogenous obesity, diabetes, and a steatotic liver that resembles Swiss cheese without the rind (there is no fibrosis!). Inob/ob mice, as well as in other rodent models of uncomplicated hepatic steatosis, administration of endotoxin provokes liver inflammation with focal hepatocyte injury, probably by releasing a shower of TNF-α from Kupffer cells, macrophages, and other cell types.13 14
Until now it has been unclear if these experimental findings have any relevance to the pathogenesis of NASH in the human liver. An earlier clue came from the severe steatohepatitis, occasionally fatal, that occurred after jejunoileal bypass (JIB) surgery for obesity.15 NASH has also been described in adults during total parenteral nutrition,16 and in a case of multiple jejunal diverticulae with bacterial overgrowth of the small intestine.17 In these conditions, metronidazole therapy reduced liver injury in some cases, although use of tetracycline was less effective.17 Furthermore, in a rat model of JIB liver disease, resection of the redundant intestinal loop with its bacterial flora, or treatment with metronidazole or tetracycline, improved liver disease.18 The mechanism appeared to be improved absorption of micronutrients, including vitamin E, a critical membrane antioxidant, and possibly essential amino acids required to synthesise the principal cellular antioxidant, glutathione.18
Despite the dire consequences of severe NASH after earlier JIB, there has seemed little reason to suspect that the intestinal flora has much to do with the usually insidious process of NASH. The report from Adelaide by Wigg and colleagues published in this issue ofGut challenges this complacency (see page206).19 The investigators studied small intestinal function and serum TNF-α levels in 22 patients with NASH and 23 controls. Using a combined 14C-d-xylose and lactulose breath test to define small intestinal bacterial overgrowth, they found that half of the patients with NASH had evidence of bacterial overgrowth compared with 22% of controls. Furthermore, serum TNF-α levels were significantly increased in the NASH group compared with controls. These findings entice us to consider the plausible proposal that gut derived bacterial toxins could be involved in triggering liver injury in the context of hepatic steatosis—or to paraphrase the title of this commentary: Do gut bugs trash the stash of liver hash into NASH?
Before accepting this contribution as a real advance towards understanding the pathogenesis of NASH, the limitations of the study should be considered. The patient groups were small and imperfectly matched in some critical variables, such as obesity, diabetes, and hypertriglyceridaemia. A positive 14C-d-xylose and lactulose breath test was strongly associated with diabetes, as might be expected from the decreased gut motility that commonly results from diabetic neuropathy (albeit that an impressive slowing of gut transit time could not be demonstrated). Could this have biased the findings so that the apparent association with NASH is spurious? To answer this, further studies comparing diabetics with and without NASH are required.
In the Adelaide study, the definition of small intestinal bacterial overgrowth rested on a single parameter, and the apparent high frequency of abnormalities in relatively healthy controls begs the issue of test reliability to define an apparently subtle form of bacterial overgrowth. The definition NASH may also have been inadequate1: it requires histological assessment because liver test abnormalities and hepatic imaging do not reliably discriminate between uncomplicated steatosis and NASH. In this study, three of the NASH patients (and all of the controls) were not subjected to liver biopsy, leaving open the opportunity for misclassification of cases.
Despite the finding that TNF-α levels were increased in patients with NASH, Wigg et al were unable to demonstrate either a “leaky” infected small intestine (as measured by the lactulose-rhamnose sugar test) or endotoxaemia.19 At first glance, this seems counterintuitive for the proposal that bacterial overgrowth of the small intestine plays a pathogenic role in NASH. Some plausible explanations for the paradox were suggested: limitations of the limulus assay, binding of endotoxin to plasma proteins, and systemic levels may not reflect portal endotoxaemia. Furthermore, other bacterial products such as peptidoglycan-polysaccharide polymers rather than endotoxin could stimulate release of TNF-α. The latter concept is particularly cogent because Bacteroides species rather than aerobic Gram negative bacteria such asEscherichia coli, the source of endotoxin, appear to be implicated in the pathogenesis of small intestinal bacterial overgrowth.20 Measurement of peptidoglycan-polysaccharide polymers in patients with NASH would be an interesting direction of investigation in the future.
In the Adelaide study, there was no relationship between body mass index and serum TNF-α levels, but a link between obesity and raised serum TNF-α has been described by others.21 Yang and colleagues noted that after endotoxin administration in leptin deficient ob/ob mice, hepatic induction of IFN-γ is increased whereas IL-10 induction is inhibited.13 IFN-γ increases hepatocyte sensitivity to TNF-α while IL-10 appears to inhibit the tissue response to TNF-α. These findings were interpreted as indicating possible macrophage dysfunction in obesity in a way that could promote steatohepatitis by sensitising hepatocytes to endotoxin.13 Guebe-Xabieret al have also shown thatob/ob mice have a selective reduction of hepatic CD4+ NK T cells, and this is associated with and possibly mediated by upregulation of IL-18 and IL-12.14Whether these abnormalities of lymphocyte populations and cytokine responses are due to obesity per se or to leptin deficiency (which is not a feature of human obesity) remains to be determined.
NASH can be regarded as the hepatic consequence of the metabolic syndrome (central obesity, insulin resistance, type II diabetes, arterial hypertension, hyperlipidaemia).22 23 Attention has shifted from the reasons for steatosis, much of which is benign or resolves in the advanced stages of cirrhosis, to the mechanisms for hepatocellular injury, inflammation, and fibrosis.4 9 13The findings reported by Wigg et al, while not definitive, may provide a new clue to the importance of cytokines in mediating liver cell injury in NASH. Whether the release of TNF-α is a consequence of small intestinal bacterial overgrowth, obesity, or oxidative stress will require further study.
See article on page 206