Investigating the tumour suppressive role of autophagy-related genes in liver cancer formation

Liver cancer is predicted to be the third leading cause of cancer-related deaths in the UK by 2030. Hepatocellular carcinoma, accounting for 80% of liver cancers, is driven by chronic rounds of inflammation resulting in fibrosis, cirrhosis and, ultimately, cancer. Autophagy, an intracellular recycling mechanism, has a dual role in cancer by preventing cancer initiation, but also supporting cancer growth and progression in established tumours. In the liver, autophagy prevents inflammation, fibrosis and tumorigenesis and suppresses the formation of atypical ductular cells in the liver parenchyma, a process named ductular reaction. The ductular reaction is a repair mechanism that develops following chronic liver injuries when hepatocyte function is severely impaired to regenerate new hepatocyte and restore liver function. This details a specific role for autophagy in the liver, however, the origin and the role of the ductular reaction remain unclear in autophagy-deficient livers. In addition, it is unclear how autophagy is modulated during liver tumorigenesis. Here, we report a new genetically-engineered mouse model for ductular reaction and hepatocellular carcinoma driven by the loss of macroautophagy and phosphatase and tensin homolog (PTEN) in the liver. In this model, Pten deletion exacerbated the macroautophagy-deficient phenotype (inflammation, fibrosis, ductular reaction and tumorigenesis) in the liver. In this model, we identified that ductular cells are biliary-like liver progenitor cells expressing stem cell markers. Using a lineage-tracing reporter, we established that liver progenitor cells originate from recombined hepatocytes in macroautophagy-deficient livers. Next, we determined that liver progenitor cells give rise to liver tumours in our model using lineage tracing and genetic deletion. Mechanistically, we show that hepatic autophagy loss activates the co-transcriptional activators YAP and TAZ in dedifferentiated liver progenitor cells and the combined, but not individual, deletion of these factors, completely reverses the dedifferentiation capacity and tumorigenesis in our model. Subsequently, I identified that hemizygosity of Atg7, a gene involved in autophagy initiation, is lost during liver tumorigenesis in a murine model of non- alcoholic-steatohepatitis-induced hepatocellular carcinoma, resulting in autophagy-deficiency. I report that restoring ATG7 expression using gene therapy in preneoplastic hepatocytes prevents liver tumorigenesis in this model. To validate these findings in human hepatocellular carcinoma, we quantified ATG7 expression in human tissue microarrays of liver cancer and discovered that ATG7 expression is transiently downregulated in low-grade hepatocellular carcinoma, and later re-expressed in high-grade hepatocellular carcinoma. Interestingly, we identified that hemizygosity of Atg5, another gene involved in autophagy initiation, is not lost in the same murine model of non-alcoholic-steatohepatitis- induced hepatocellular carcinoma is maintained. This dictates a specific role for ATG7 modulation during liver tumorigenesis. These findings indicate that autophagy suppresses the formation of hepatocyte- derived ductular progenitor cells giving rise to cancer in the liver. Moreover, the transient loss of ATG7 drives autophagy inhibition during tumorigenesis of alcoholic-steatohepatitis livers. These results therefore shed more light on the cell of origin of hepatocellular carcinoma development, the role and regulation of autophagy in cancer and highlight new potential points for therapeutic intervention