Determinants of fibrotic injury in mouse models of liver disease

Chronic liver diseases (including liver cancer) are a common cause of morbidity and mortality worldwide, accounting for almost 2 million deaths per year, with non-alcoholic fatty liver disease (NAFLD) becoming a growing burden on the healthcare systems in westernised countries. Liver fibrosis, driven by persistent, low-grade inflammation, is a common feature of chronic liver diseases, characterised by the unregulated wound healing processes that result in excessive extracellular matrix formation. Irrespective of aetiology, liver fibrosis and its most severe form cirrhosis, contributes to declining liver function and predisposes to hepatocellular carcinoma. The mechanisms by which liver injury progresses to more advanced stages are still being elucidated, resulting in the slow development of effective therapeutics for fibrosis prevention or reduction, where no effective therapeutic modalities to effectively treat liver fibrosis currently exist. Therefore, there is an unmet need for identification of new therapeutic targets to prevent disease progression, requiring suitable modelling systems to understand the pathogenesis of fibrosis and for efficient drug development. Mice have been widely utilised as preclinical models to understand the molecular mechanisms driving liver disease progression in an in vivo setting. Significant variations in the methodologies have contributed to a failure to produce standardised models with known and predictable fibrotic outcomes, resulting in discrepancies in the literature. Differing techniques utilised to measure liver injury and fibrosis also influence the study outcomes. This thesis aims to compare and contrast commonly used fibrosis models (Carbon tetrachloride, thioacetamide and atherogenic diet) while controlling for mouse strain, sex and age to assess their congruence with expected fibrosis or cirrhotic outcomes observed in the human condition. Thioacetamide (TAA) administration in the drinking water method produced the most advanced cirrhotic phenotype though requires a long period for induction, whereas carbon tetrachloride (CCl4) model is more suitable as an early inflammatory model with moderate fibrosis to early-stage cirrhosis. Thioacetamide administration by intraperitoneal injection with increasing dosage over 6 weeks presented a rapid fibrosis model, intermediate between the oral TAA and CCl4 in terms of liver injury and fibrotic outcome. Atherogenic diet for 12 weeks resulted in simple steatosis without fibrotic scar. No single model was able to recapitulate human liver disease progression. The injury phenotype in these ‘standardised’ models is useful to study the development of progressive liver injury stage in a reproducible manner, enabling specific experimental questions to be addressed based on expected liver injury and fibrotic outcomes. In the search for new therapeutics for liver fibrosis, repositioning of approved drugs is advantageous in the drug discovery process by identification of targets common to liver injury pathways. Oxidative stress has been implicated as driver of hepatocyte injury and in the pathogenesis of liver fibrosis progression to cirrhosis. The effects of a new pro-drug NP202, with proven efficacy in prevention of myocardial fibrosis in other pre-clinical models of heart failure and injury via antioxidant pathways, has not been previously investigated in the liver disease context, with potential use as an anti-fibrotic therapeutic. Mouse models of fibrosis and NAFLD/NASH development with the addition of NP202 were used to determine the efficacy of this pro-drug in the prevention of TAA, western diet or atherogenic diet-induced liver injury during disease progression and resolution of fibrotic scar after injury cessation. NP202 treatment did not prevent fibrosis development in TAA-injured livers, and had little effect on liver injury, steatosis or fibrosis development in the dietary models of NAFLD/NASH. After injury cessation, continued administration of the pro-drug failed to improved resolution of liver injury, steatosis and fibrotic scar. Unexpectedly, some mice administered concurrent TAA and NP202 treatment developed primary liver tumours. As inflammation is a major driver of fibrotic injury, a tumour necrosis factor (TNF) knockout mouse was utilised to model inhibition of a major proinflammatory mediator, as a potential therapeutic strategy to prevent fibrogenesis in hepatotoxin or diet-induced liver injury. During early phase injury, both inflammation and fibrosis were impaired with hepatotoxin induction, resulting in significantly lower level of hepatic injury. While inhibition of TNF did not inhibit the fibrogenic processes, the rate of inflammatory injury and fibrotic scar deposition was slowed in oral TAA administered mice compared to control mice. Further, Loxl2-mediated collagen crosslinking was impaired in TNF-deficient mice undergoing TAA or CCl4 induced liver injury, with resolution of fibrotic scar also impaired in these mice. In contrast, atherogenic diet caused significant hepatocyte ballooning and steatotic injury in TNF-deficient animals, suggestive that the role of TNF is dynamic and heavily dependent on the type of hepatotoxicant. Taken together, anti-TNF therapies have the potential to reduce the inflammatory injury driving early fibrogenesis, though alone cannot prevent fibrotic scar formation. Furthermore, TNF is required of efficient fibrotic scar dissolution after cessation of injury. Overall, while there is no single liver injury mouse model that can effectively recapitulate the natural liver disease progression observed in humans, mouse models continue to serve as an essential and irreplaceable pre-clinical model for understanding of pathological processes. The utility of well characterised liver fibrosis models allows the study of inhibition of major drivers of liver injury such as TNF and oxidative stress in a controlled manner, which is particularly important when testing novel therapeutics for safety and efficacy purposes. The expected injury phenotypes and examples of their use in practice, as described in this thesis can assist in the identification of the most suitable liver injury model for pre-clinical assessment of future anti-fibrotic therapeutics