Animal models in nonalcoholic steatohepatitis research: utility and clinical translation

Nonalcoholic fatty liver disease (NAFLD) represents a spectrum of metabolic syndrome-associated liver disease progressing from simple steatosis, through nonalcoholic steatohepatitis (NASH) and fibrosis to cirrhosis and hepatocellular carcinoma (1). Increasingly sedentary lifestyles and changing dietary patterns have allowed NAFLD/NASH to become the most common cause of liver dysfunction in many countries worldwide. Despite this, our understanding of disease pathogenesis remains incomplete and as yet specific therapies to delay disease progression or treat fibrosis are limited. Experience has shown that murine biological processes are similar to those in man and that rodent models are both genetically tractable and generally reliable surrogates for human biology. Although some research may be undertaken at the cellular level in vitro, in order to dissect multisystem pathophysiological processes such as those associated with NAFLD/NASH, it is necessary to study effects at the whole-organism level. Established inbred strains permit both genetic background (which may contain susceptibility alleles) and environmental factors to be controlled and therefore facilitate study of complex disease traits such as NAFLD/NASH. The contribution of murine models to this field already extends from basic research through to preclinical testing of novel therapeutics. Indeed, the ‘Two-Hit’ hypothesis was first proposed by Day and James in an editorial that accompanied a paper by Berson and colleagues describing a link between drug-induced oxidative stress and steatohepatitis in rats (2, 3). More recently, hepatocyte fatty acid (FFA) flux rather than histologically evident steatosis was shown to be the driver of progression to steatohepatitis in a mouse model of fibrosing NASH (4). In this important study siRNA techniques were used to block diacylglycerol acyltransferase 2 expression showing that, while triglyceride accumulation was reduced, hepatocellular oxidative stress, inflammation and necrotic/apoptotic cell death were increased (4). It is through these and other studies using rodent models that our understanding of disease pathogenesis has developed. Specifically, insulin resistance promotes greater hepatic FFA flux which leads to direct hepatocyte lipotoxicity, hepatocellular oxidative stress secondary to free radical production during β- and ω-FFA oxidation, endotoxin/TLR4-induced inflammation, cytokine release and endoplasmic reticulum stress (5). Consequent cellular damage triggers a mixture of immune-mediated hepatocellular injury and both necrotic and apoptotic cell death pathways which culminates in tissue fibrosis. This greater mechanistic understanding is now informing strategies for therapeutic intervention which can be tested preclinically in these same rodent models and may one day be applied in the clinic. Research models of NAFLD may be divided into two broad categories, those caused by either spontaneous or induced genetic mutation and those with an acquired NAFLD phenotype (1). The latter group may be produced by either dietary or pharmacological manipulation. Despite their proven utility, an apparent ‘phenotype-gap’ exists between some of the currently available models and the condition encountered at the bedside (1, 6). While the literature contains numerous different rodent models that exhibit histological evidence of hepatic steatosis and mild–moderate steatohepatitis, the features of more advanced steatohepatitis (including ballooning hepatocyte degeneration and an inflammatory infiltrate) combined with insulin resistance, are less frequently demonstrated (1). Depending on the background strain studied, a chronic (45 cal%) high-fat diet or sustained high fructose consumption provide acceptable models of the metabolic syndrome with histological features of a mild-moderate steatohepatitis with which to study effects of dietary/environmental manipulation, as well as hepatocellular lipid handling, fatty acid metabolism and associated cellular stresses (7, 8). However, the lack of progression to fibrosing steatohepatitis has limited such models' utility for therapeutics discovery in progressive disease. Attempts at using higher dietary fat levels (9) and/or hyperalimentation by gastrostomy feeding (10) have produced some increase in steatohepatitis severity but have not been widely adopted due to technical difficulties and the modest increase in disease severity generated. A more severe fibrosing steatohepatitis may be induced by feeding a diet deficient in methionine and choline (MCD). The metabolic consequences of MCD feeding are complex (11, 12) and an important caveat remains: steatohepatitis is induced through a nonphysiological process. Indeed, MCD dietary manipulation alone does not produce insulin resistance (12), but instead causes weight loss which can limit the extent to which disease may be allowed to progress. Several groups have attempted to bridge the phenotype gap by combining genetic and acquired (dietary) manipulation (or multiple separate dietary/pharmacological challenges) to produce a more severe, accelerated or fibrosing steatohepatitis. These include Abcb11 mutant mice fed MCD (13), apolipoprotein E mutant mice fed a high-fat diet (14) and PPARα null mice fed MCD (15). Currently, one of the most promising and more widely adopted models available for studying more advanced disease at the steatohepatitis–fibrosis interface is generated by feeding genetically hyperphagic, obese/diabetic db/db mice an MCD diet (4, 16, 17). The combined genetic and dietary manipulation provides a prosteatotic, metabolic syndrome background with histological features of severe fibrosing steatohepatitis evidenced at the transcriptome level by a 10-fold increase in hepatic procollagen type 1 expression in MCD fed db/db mice while their nonsteatotic db/m littermates have only a four-fold rise after MCD feeding (17). The publication of another rodent model of NAFLD/NASH prompts us to consider what are the specific characteristics required in a good fatty liver disease model (Table 1)? It is a given that a model must be easy to use, offer a genetically stable background on which to study disease modifiers and produce a reliable and predictable disease progression within a relatively short time frame so that studies deliver results in a timely fashion and do not consume excessive resources. The recognition that insulin resistance and FFA flux are at the heart of the disease process would suggest that a background insulin resistant, obese phenotype is required; however, what remains unresolved is whether a model must over time recapitulate every aspect of the clinical disease spectrum or whether it must simply be sufficient to the research question and stage of disease studied. What is certain is that the more faithful a preclinical model is to the clinical condition, the more likely that what we learn from it will be clinically relevant. In the current edition of the journal, Ogasawara et al. (18) describe a new ‘combined model’ generated by exploiting two acquired challenges in the search for a better NASH model. Rather than relying on genetic mutations such as leptin deficiency (ob/ob mice) or leptin receptor deletion (db/db mice and fa/fa rats) to induce a hyperphagic state, the authors have used the ability of gold-thioglucose (GTG) to induce hyperphagia, obesity and NASH through selective ablation of a ‘satiety centre’ within the venteromedial hypothalamic nucleus of C57BL/6 mice (19). This technique obviates concerns regarding absence of leptin signalling and the consequent downstream changes in innate and acquired immune response and fibrogenesis that may be levelled at leptin-deficient models such as the ob/ob mouse (1). Here, the central effects of GTG, first reported in Nature over 50 years ago (20), have been combined with a 12-week period of extreme high-fat diet (82 cal% fat) consumption to exacerbate the obese, metabolic syndrome phenotype and produce greater oxidative stress and a more severe histological steatohepatitis with mild zone 3 and pericellular hepatic fibrosis. Although the authors do not show progression to advanced fibrosis or cirrhosis, this model could facilitate the study of the transition from simple steatosis to steatohepatitis and the initiation of fibrogenesis while remaining faithful to many of the fundamental aetiological processes which underpin NAFLD/NASH and the metabolic syndrome. How widely this model will be adopted will be determined in part by how consistently reproducible the model is found to be. However, a model exhibiting a predictable, obesity-related, mild-moderate fibrosing steatohepatitis produced through voluntary hyperalimentation could serve as a useful platform for basic science and therapeutic investigation across an important spectrum of clinical disease. The evolution of rodent NASH models from those, such as the Ob mouse, that were limited to a single step in disease pathogenesis, to models such as the Db/MCD model (17) and the one discussed here (18), that blur the once ridged boundaries between steatosis, steatohepatitis and fibrosis are a welcome development in the translational investigation of NASH. Q. M. A. is the recipient of a Clinical Senior Lectureship Award from the Higher Education Funding Council for England (HEFCE). Q. M. A. is also the recipient of an MRC Programme Grant to identify and study novel murine models of liver disease.