|We recently established novel genetically engineered mouse models of hepatocellular carcinoma (HCC) and cholangiocarcinoma (CC), which have afforded new insights into molecular mechanisms and cellular origins of HCC and CC. First, we generated a nonalcoholic steatohepatitis-driven HCC mouse model by feeding high-fat diet (HFD) to MUP-uPA mice. In this model, the vicious cycle of endoplasmic reticulum stress and hypernutrition synergistically aggravated lipid accumulation and subsequently induced continuous cell death and compensatory proliferation, and eventually lead to HCC. Using this model, we analyzed the roles of Sox9+ periportal hepatocytes in liver regeneration and HCC development by genetic lineage tracing. Although Sox9+ periportal hepatocytes did not participate in maintaining adult liver homeostasis, those cells underwent extensive proliferation during chronic hepatocyte injury in MUP-uPA mice. Interestingly, despite their high regenerative potential, periportal hepatocytes did not give rise to HCC in HFD-fed MUP-uPA mice. On the other hand, Axin2+ pericentral hepatocytes also contributed to liver regeneration in MUP-uPA mice, and their tumorigenic capacity is now under investigation. Second, we established a mouse model of biliary injury-related extrahepatic CC (ECC) by tamoxifen-inducible duct cell-specific activation of Kras and deletion of TGFbR2 and E-cadherin. In this model, E-cadherin-deleted biliary epithelial cells (BECs) lining the bile duct lumen died and released alarmin IL-33 which stimulated bile duct regeneration by peribiliary gland (PBG), a BEC stem cell niche, and eventually lead to ECC. Cell lineage tracing suggested PBGs as the cellular origin of ECC. Thus, these mouse models shed some light on the nature of hepatobiliary carcinogenic process.