• Potential conflict of interest: Nothing to report.


We thank Evert et al. for their interesting comments. We do acknowledge their elegant study concluding that intrahepatic cholangiocarcinoma (ICC) can arise from adult hepatocytes after delivering Notch1/Akt plasmids to adult mouse livers by means of hydrodynamic tail vein injection.[1]

We support this overall conclusion. However, especially when new concepts are coined, potential limitations of the methods used must be discussed carefully. Evert et al. argue that transfection of periportally located liver progenitors after hydrodynamic tail vein injection of plasmids is highly unlikely since this would “contradict the physiological principles of hydrodynamic gene delivery.” Surprisingly, in a recent article this is exactly what the authors propose to be a possible mechanism for hepatocellular carcinoma (HCC)/ICC development after hydrodynamic Akt/N-Ras delivery.[2] Nevertheless, the lineage tracing data that Evert et al. now strengthen by additional morphological data clearly support the concept of hepatocytes as the cells of origin of ICC in the Notch1/Akt mouse model. Although hepatic carcinogenesis was beyond the focus of our study, our notion that Notch2-IC rapidly converts adult hepatocytes to the biliary lineage further corroborates this concept.[3] Importantly, cholangiocarcinoma development also appears to rely on biliary transdifferentiation of pericentral hepatocytes in chronic thioacetamide-induced liver injury in mice without genetic modulation of oncogenic pathways.[4]

We proposed Notch2 to be the primary Notch receptor to mediate biliary conversion of adult hepatocytes similar to intrahepatic biliary development.[5] Fan et al.[1] reported that delivery of Notch1-IC alone did not result in any alterations of liver morphology within 10 weeks, while we observed biliary transdifferentiation of hepatocytes within a few days after overexpression of Notch2-IC in adult mouse livers.[3] Nonetheless, as discussed in our study, we are well aware that, i.e., transgene levels may also contribute to these differences and studies are lacking that directly compare the effects of Notch1-IC and Notch2-IC in the adult liver. Future Notch1 and Notch2 loss-of-function studies in liver regeneration or tumor models where biliary transdifferentiation of hepatocytes occurs may better characterize the importance of individual receptors.

Owing to the histopathological heterogeneity of ICC several cells of origin including hepatic progenitors have been proposed.[6, 7] Currently, however, taking into account how easily hepatocytes are turned to a biliary fate and how easily these cells express progenitor markers under experimental conditions, it remains to be determined whether biliary conversion of mature hepatocytes represents just an alternative or rather the predominant route to at least certain subtypes of ICC (i.e., cholangiolocarcinoma or small duct ICC). The pancreas and the liver share many similarities in development and disease. Notably, in the controversy of the cellular paternity of pancreatic ductal adenocarcinoma, the pendulum seems to swing to the acinar rather than the ductal compartment.[8] Nevertheless, due to the tremendous plasticity of adult liver cells we at least must accept that the morphological appearance and molecular profiles of cells at a given time do not allow us to deduce its cell of origin and only further well-controlled compartment-specific lineage tracing studies can reveal the individual susceptibility of different liver cell compartments to malignant transformation.

  • Petia Jeliazkova, M.D.Simone Jörs, Ph.D.Jens T. Siveke, M.D.Fabian Geisler, M.D.

  • 2nd Department of Internal Medicine Klinikum rechts der Isar Technical University of MunichMunich, Germany