• Potential conflict of interest: Nothing to report.


We appreciate the thoughtful comments of Li and colleagues with regard to our recent studies on microRNA-181 (miR-181) and its functional connection to epithelial cell adhesion molecule (EpCAM)-expressing liver cancer stem cells.1 In their letter, Li et al. made four valid points. They have correctly pointed out that there is a heterogeneous expression of EpCAM in both HuH1 and HuH7 hepatocellular carcinoma cell lines. These cells behave as cancer stem cells and possess both self-renewal and differentiation capacities in a hierarchy manner. In contrast, most of the cells in several other hepatoma cell lines such as Hep3B and HepG2 express EpCAM.2, 3 It is noticeable that both HuH1 and HuH7 cells contain a functional Wnt/β-catenin signaling pathway whereas HepG2 cells contain a mutant β-catenin and Hep3B cells contain a persistent phosphorylation of glycogen synthase kinase-3β, resulting in a constitutive activation and nuclear localization of β-catenin in these cells.4 The parallel expression of EpCAM and β-catenin in these cell lines is consistent with our finding that EpCAM is a transcriptional target of β-catenin.2 Such a correlation can also be observed with EpCAM-positive and EpCAM-negative cells isolated from hepatocellular carcinoma cell lines carrying wild-type β-catenin. We hypothesize that there is a major distinction between liver cancer stem cells with an activated wild-type β-catenin and liver cancer cells carrying a constitutively activated β-catenin.

We agree with the authors' comment in regard to a potential direct transcriptional regulation of miR-181 expression by β-catenin. Our ongoing studies are designed to determine whether miR-181s are direct transcriptional targets of Wnt/β-catenin signaling. These studies utilize both genetic and biochemical approaches to test physical and functional interactions of the miR-181 loci and Wnt/β-signaling pathway in liver cells.

It remains unclear whether and how miR-181s are regulated by microRNA biogenesis. It is interesting to note that there is a difference in expression levels between miR-181c and miR-181d. However, such a difference is only evident in mouse fetal livers but not in human fetal livers and human liver stem cells. The nature of this discrepancy is unclear. It is interesting to point out that if microRNA biogenesis plays a role in the regulation of miR-181s, such a mechanism apparently is not evolutionarily preserved.

We agree that a positive feedback loop model we proposed in our recent article1 has a missing link, which, we believe, has been found in our recent unpublished data. We found that Wnt signaling can directly induce miR-181 expression (Ji et al., manuscript in preparation). Thus, it appears that Wnt/β-catenin signaling is critical in inducing miR-181s, which then inactivate NLK (nemo-like kinase) to further boost Wnt signaling activity. We believe such a positive feedback mechanism is designed to maintain stemness of liver cancer stem cells. Thus, the Wnt/β-catenin signaling pathway may be a sensible therapeutic target for eradication of liver cancer stem cells, the entities that have been linked to poor prognosis, drug resistance, and metastasis in liver cancer.

Junfang Ji*, Xin Wei Wang*, * Liver Carcinogenesis Section, Laboratory of Human Carcinogenesis, Center for Cancer Research, National Cancer Institute, Bethesda, MD.