The metastatic spread is a multistep process that requires basal membrane invasion by the cells, their intravasation in blood or lymph vessels and their extravasation to reach and grow in distant tissues. At each step, cells undergo morphological changes driven by the cortical actin cytoskeleton. Proteins of the ERM (Ezrin, Radixin, Moesin) family provide a regulated linkage between the plasma membrane and the actin cytoskeleton. They have been implicated in the determination of cell shape, membrane organization, cell polarization, migration, division and they participate in various signaling pathways. Their activity is negatively regulated by intramolecular association between their N-terminal FERM and C-terminal domains thus preventing their binding to F-actin and membrane proteins. Conformational changes triggered by the binding of the FERM domain to PIP2 followed by the phosphorylation of a conserved C-terminus threonine (558 for moesin, 567 for ezrin and 564 for radixin) regulate the association of ERM proteins with both the membrane and the actin cytoskeleton.

In the recent years, there has been mounting evidence that ezrin is actively involved in the metastatic spread of various neoplasms. Following studies correlating the expression levels of ezrin to the metastatic potential of different types of tumors, experimental models have demonstrated the implication of ezrin in the metastatic spread of osteosarcoma, rhabdomyosarcoma and mammary tumor cells (Elliott et al., 2005; Khanna et al., 2004; Yu et al., 2004). Less documented is the role of moesin in tumor invasion whereas, so far, radixin has not been implicated in this process. If the role of ezrin and to a lesser extent of moesin in promoting tumor progression is well-established the molecular mechanisms by which these proteins trigger cell invasion and metastasis are far from being elucidated.

In this context, the paper by Estecha et al. provides interesting observations on the role of moesin and ezrin in cortical melanoma cell organization and invasion. Using highly invasive melanoma cell lines, the authors show that depletion of moesin strongly decreases the colonization of the lung several hours after injection of the cells in the tail vein of mice, but not at early time points suggesting that moesin is not involved in the attachment of the cells to the blood vessels but more likely in the extravasation step. Interestingly, ezrin-depleted cells colonize the lung to the same extent as control cells.

Then the authors explored the role of ezrin and moesin in melanoma cell invasion of 3D collagen matrices and in transmigration through endothelial cells. They showed that following early attachment of the cells to the collagen matrix or to the endothelium, moesin distribution was clearly polarized, accumulating to the dorsal surface of the cell, and being excluded from the attachment point. In contrast, ezrin distribution was found in blebs at the sites of cell invasion. Moreover, whereas moesin-depleted cells were less invasive in 3D collagen matrix, ezrin-depleted cells acquired an elongated shape and were more prone to 3D matrix invasion.

This polarized distribution of moesin was also observed at the initial step of melanoma cell spreading on 2D collagen-coated surface. Moesin was concentrated at the dorsal surface of the cells as was PIP2 which trigger the activation of ERM proteins. Consistent with this observation, moesin phosphorylated at threonine 558 was also enriched at the dorsal surface of the cells. The phosphorylation of moesin was sensitive both to PKC and ROCK inhibitors suggesting that RhoA is an important regulator of moesin activity during cell invasion. In addition, phosphorylation of myosin light chain (MLC), a marker of acto-myosin contractility, was also dorsally localized during early and intermediate spreading. These observations led to the proposal that moesin plays a role in the establishment of a stiff rounded dorsal cortex. This was confirmed by the observation that moesin depletion in melanoma cells caused a flattened phenotype and a faster spreading on a collagen matrix.

Overall, the study clearly demonstrates that moesin is required for melanoma cell invasion both in vivo and in vitro. These observations are compatible with a model where by early attachment to the matrix would lead to the activation of moesin with its concomitant relocalization to the dorsal surface of the cells where PIP2 is enriched. Activated moesin would promote RhoA activation and MLC-dependent contractility and would represent a link between these two processes. These events would in turn orientate the contractile forces toward the leading edge during invasion. In another study, it has been shown that ezrin participates to rhabdomyosarcoma metastasis and that its action is, at least partially, mediated by RhoA (Yu et al., 2004). Thus, RhoA appears as a common factor in ezrin- and moesin-mediated metastasis. ERM proteins have been shown to function upstream and downstream of the Rho pathways. In the experiments reported by Estecha et al. such regulatory loop would exist where by phosphorylated (active) moesin would activate RhoA which in turn would sustain moesin activation. In this context, ezrin would be inactive suggesting that a complex interplay may exist between the ERM proteins during the invasion process.

The importance of moesin in the regulation of the cortical cytoskeleton contractility and cell shape has already been demonstrated in mitotic Drosophila cells (Rosenblatt, 2008). In these cells, moesin depletion results in the decrease of cortical stiffness during mitosis and the loss of a round shape. This rounding was not dependent on myosin since cells lacking MLC but expressing an active form of moesin were still able to adopt a round shape. It is however worth mentioning that whereas moesin is the only ERM protein to be expressed in Drosophila, melanoma cells express the three ERM proteins.

It is becoming evident that although ERM proteins have overlapping functions they also display specific ones. In many reports, cell metastasis has been correlated with the over-expression of one ERM protein, however, the contribution of the two other, in the same context, has not been investigated. A key issue is therefore to identify the respective role of ezrin and moesin in tumor progression and to determine which functions exerted by these proteins are subverted by tumor cells. Thus, it has been shown that in a mouse model of osteosarcoma metastasis, ezrin provides a survival advantage mediated largely by the activation of the MAPK pathway (Khanna et al., 2004). The study by Estecha et al. demonstrates that ezrin and moesin have different roles in cell invasion and metastasis. Their results suggest that cellular compartimentation may provide a mean for these proteins to exert different functions. While moesin would regulate the stability of the dorsal cortex, ezrin present at the ventral surface of the cells would control cell adhesion and migration. Further studies will need to determine whether these proteins have synergetic or antagonist role in the metastatic process.

Finally, although the present study clearly provides substantial progress in the understanding of moesin function during cell invasion in vitro and in vivo, the relevance for human melanoma metastasis remains to be demonstrated. Indeed, in a study by Ichikawa et al., moesin level was found to be elevated in benign melanocytic naevi and malignant melanomas but was significantly reduced in invasive melanomas and metastasis (Ichikawa et al., 1998) whereas, in another study, elevated level of ezrin was associated with increased melanoma tumor growth and invasion (Ilmonen et al., 2005). Therefore, more studies will be necessary to elucidate the role of moesin, as well as the other ERM members in melanoma progression and metastasis.


  1. Top of page
  2. References
  • Elliott, B.E., Meens, J.A., Sengupta, S.K., Louvard, D., and Arpin, M. (2005). The membrane-cytoskeletal crosslinker ezrin is required for metastasis of breast carcinoma cells. Breast Cancer Res. 7, 365373.
  • Ichikawa, T., Masumoto, J., Kaneko, M., Saida, T., Sagara, J., and Taniguchi, S. (1998). Moesin and CD44 expression in cutaneous melanocytic tumours. Br. J. Dermatol. 138, 763768.
    Direct Link:
  • Ilmonen, S., Vaheri, A., Asko-Seljavaara, S., and Carpen, O. (2005). Ezrin in primary cutaneous melanoma. Mod. Pathol. 18, 503510.
  • Khanna, C., Wan, X., Bose, S., Cassaday, R., Olomu, O., Mendoza, A., Yeung, C., Gorlick, R., Hewitt, S.M., and Helman, L.J. (2004). The membrane-cytoskeleton linker ezrin is necessary for osteosarcoma metastasis. Nat. Med. 10, 182186.
  • Rosenblatt, J. (2008). Mitosis: moesin and the importance of being round. Curr. Biol. 18, R292293.
  • Yu, Y., Khan, J., Khanna, C., Helman, L., Meltzer, P.S., and Merlino, G. (2004). Expression profiling identifies the cytoskeletal organizer ezrin and the developmental homeoprotein Six-1 as key metastatic regulators. Nat. Med. 10, 175181.