Study of the roles of dishevelled-3 in stemness and cell migration in hepatocellular carcinoma

Abstract

Hepatocellular carcinoma (HCC) is a common malignancy worldwide and particularly common in China and Southeast Asia. It ranks the 2nd and 4th most common fatal cancer in males and females, respectively, in Hong Kong. Current treatments are not always effective, as recurrence and metastasis in HCC are difficult to tackle and the underlying mechanisms not fully understood. Aberration of Wnt signaling has been implicated in HCC; in this study, we investigated the underlying mechanisms of how aberrant Wnt signaling promoted HCC development. With Taqman Low Density Array (LDA) analysis on 38 pairs of HCC and the corresponding non-tumorous livers for 59 Wnt signaling related-genes, we found significant overexpression of the Wnt signaling intermediate, Dishevelled (Dvl)-3, in HCC (p = 0.014). This observation in LDA was confirmed in 36 additional HCC cases. Among a total of 74 cases studied, 28.38% showed more than 3-fold overexpression in the tumors as compared with the corresponding non-tumorous livers. Dvl3 overexpression positively correlated with the presence of venous invasion. We also observed significant correlation of Dvl3 expression with accumulation of β-catenin, a downstream effecter of Wnt/β-catenin signaling (p=0.028).

We further characterized the functional roles of Dvl3 in contributing to the stem cell-like and metastatic properties of HCC. We found that Dvl3 knockdown in HCC cells suppressed cell proliferation, sphere formation, tumorigenicity in immunodeficient mice, chemo-resistance, and expression of stemness genes. We then examined whether Wnt/β-catenin was effectively modulated by Dvl3 and found that Dvl3 overexpression and knockdown, respectively, promoted and reduced the TOP/FOP luciferase reporter activity in HCC cells. This was accompanied by the expression of β-catenin target genes, EpCAM and LGR5, both of which are associated with HCC stemness. Furthermore, rescue with wild-type or constitutively active β-catenin partially restored the in vivo tumorigenicity suppressed by Dvl3 knockdown, indicating a partial role of β-catenin in mediating the effects of Dvl3 on HCC stemness.

In addition, since cell migration is a critical determinant in metastasis, we assessed the HCC cell migratory ability in vitro using transwell migration assays and observed suppression of the cell migration ability upon Dvl3 knockdown. Also, the in vivo orthotopic model confirmed a role of Dvl3 in promoting metastasis, as stable Dvl3 knockdown in HCC cells resulted in a reduction in lung metastasis. Interestingly, the effect of Dvl3 on cell migration was independent of β-catenin, as knockdown of β-catenin had no effect on HCC cell migration in vitro. It was also not related to the phosphorylation of MYPT in Rho-ROCK signaling, which itself was previously implicated in HCC cells metastasis and reported as a downstream signaling of Dvl in development.

In summary, our study has identified roles of Dvl3 in HCC stemness properties and cell migration and this may provide functional implication of Dvl3 overexpression, which significantly correlated with venous invasion in human HCCs. Also, β-catenin is partly responsible for the role of Dvl3 in HCC stemness but independent of that in cell migration.

Functional characterization of Dvl3 in HCC may help future development of therapy targeting Dvl3 of Wnt signaling pathways.