To the Editor:

After reading the article by Taura et al. with great interest, we really appreciate the ingenious work they have done.1 In this study, they found type I collagen-producing cells do not originate from hepatocytes in a triple transgenic mouse model. Hepatocytes in vivo neither express mesenchymal markers nor exhibit a morphological change. All these results strongly challenge the earlier concept that hepatocytes in vivo acquire a mesenchymal phenotype through epithelial-mesenchymal transition (EMT) to produce the extracellular matrix (ECM), leading to liver fibrosis. Although they have provided solid evidence to show that EMT does not contribute to ECM in the process of liver fibrosis, several issues need to be further discussed.

The EMT phenomenon was first clearly demonstrated by Kaimori2 and Zeisberg3 simultaneously in mouse liver fibrosis. Although Taura et al. challenge the concept in vivo, the three groups independently showed that isolated primary hepatocytes exhibited fibroblast-like morphological change. Therefore, the different biological behaviors of hepatocytes in liver fibrosis between in vivo and in vitro models need to be reassessed. Studies on gene expression profiles of activated hepatic stellate cells revealed the almost identical gene expression patterns after bile duct ligation or carbon tetrachloride (CCl4) treatment, whereas culture activation only partially reproduced the gene expression changes,4 suggesting in vivo activation should be considered as the gold standard for the study of liver fibrosis. However, the etiology of liver fibrosis includes viruses (hepatitis B and C viruses), alcohol intoxication, obesity, diabetes, and hereditary metabolism disorders, and so could the currently prevalent liver fibrosis models truly reflect the changes of the hepatocytes in liver injury? The existence of EMT of hepatocytes in liver fibrosis still seems to be an open question.

With triple transgenic mice ROSA26 stop β-gal, AlbCre, and collagenase green fluorescent protein (GFP), the double-positive cells for GFP and X-gal were not observed in situ at different stages of liver injury, including the chronic phase (after 16 injections with CCl4), indicating collagen-producing cells do not originate from hepatocytes. The authors further demonstrated the isolated hepatocytes from CCl4-induced transgenic mice do not express mesenchymal markers including α-smooth muscle actin. However, in our CCl4-induced mouse liver fibrosis sections, α-smooth muscle actin was detected in the cytoplasm of hyperplastic hepatocytes by immunohistochemistry even though it is expressed prominently in the perisinusoidal space (Fig. 1). Therefore, we suggest the authors should evaluate again the double staining for myofibroblastic phenotypes and X-gal in situ. Recently, Zulehner et al. reported EMT is involved in hepatocarcinogenesis in a mouse model and loss of plasma membrane E-cadherin expression in poorly differentiated human hepatocellular carcinoma, suggesting EMT of hepatocytes in this stage.5 Cirrhotic liver–derived hepatocytes from a mouse cirrhosis model with characteristics of EMT exhibit decreased apoptosis via a mitogen-activated protein kinase–dependent cell survival pathway, implying EMT as an outcome of antiapoptosis in carcinogenesis.6, 7 Because of the insufficient evidence from the literature and limitations of the study as mentioned by the authors, more detailed studies with a translational medicine methodology are needed to verify the existence of EMT of hepatocytes followed by investigation of its related role in liver diseases, including liver fibrosis and hepatocellular carcinoma development.

thumbnail image

Figure 1. Expression of α-SMA was detected in (A) Kunming mouse liver and (B) CCl4-induced liver by immunohistochemistry. Sections were incubated with antibody against α-SMA (Abcam, Cambridge, MA; diluted 1:50) followed by visualization with the Histostain-Plus Kit (Invitrogen, Carlsbad, CA) after treatment with 3,3-diaminobenzidine and counterstain with hematoxylin (magnification = ×1000).

Download figure to PowerPoint


  1. Top of page
  • 1
    Taura K, Miura K, Iwaisako K, Österreicher CH, Kodama Y, Penz-Österreicher M, et al. Hepatocytes do not undergo epithelial-mesenchymal transition in liver fibrosis in mice. HEPATOLOGY 2009; doi:10.1002/hep.23368.
  • 2
    Kaimori A, Potter J, Kaimori JY, Wang C, Mezey E, Koteish A. Transforming growth factor-beta1 induces an epithelial-to-mesenchymal transition state in mouse hepatocytes in vitro. J Biol Chem 2007; 282: 22089-22101.
  • 3
    Zeisberg M, Yang C, Martino M, Duncan MB, Rieder F, Tanjore H, et al. Fibroblasts derive from hepatocytes in liver fibrosis via epithelial to mesenchymal transition. J Biol Chem 2007; 282: 23337-23347.
  • 4
    De Minicis S, Seki E, Uchinami H, Kluwe J, Zhang Y, Brenner DA, et al. Gene expression profiles during hepatic stellate cell activation in culture and in vivo. Gastroenterology 2007; 132: 1937-1946.
  • 5
    Zulehner G, Mikula M, Schneller D, van Zijl F, Huber H, Sieghart W, et al. Nuclear beta-catenin induces an early liver progenitor phenotype in hepatocellular carcinoma and promotes tumor recurrence. Am J Pathol 2010; 176: 472-481.
  • 6
    Nitta T, Kim JS, Mohuczy D, Behrns KE. Murine cirrhosis induces hepatocyte epithelial mesenchymal transition and alterations in survival signaling pathways. HEPATOLOGY 2008; 48: 909-919.
  • 7
    Valdés F, Alvarez AM, Locascio A, Vega S, Herrera B, Fernández M, et al. The epithelial mesenchymal transition confers resistance to the apoptotic effects of transforming growth factor Beta in fetal rat hepatocytes. Mol Cancer Res 2002; 1: 68-78.

Da-Wei Zhang*, Huijie Bian*, * Cell Engineering Research Center and Department of Cell Biology, State Key Laboratory of Cancer Biology, Fourth Military Medical University, Xi'an, China.