EMT or EMT-Promoting Transcription Factors, Where to Focus the Light?

The epithelial-to-mesenchymal transition (EMT) is commonly considered as a main driving force of the metastatic cascade. An impressive series of experiments and observations worldwide supports its pivotal role in promoting cancer cell dissemination at the invasive fronts of tumors, intravasation, cell survival in fluids, and extravasation, with secondary site colonization being the only step requiring a return to an epithelial phenotype (1). Beyond this, commitment of epithelial cells into an EMT program has been associated with resistance to chemo, radio-, and hormono-therapies, definitively extending the interest of studying EMT from the embryologist community to the oncology and medical fields. Obviously, EMT turns out to constitute a transdifferentiation program, allowing epithelial cells to escape from numerous stresses, including mechanical constrains, hypoxia, nutrient depletion, and unfortunately therapeutic treatments (1). EMT is orchestrated by a restricted number of transcription factors mainly the three Snail, Twist, and Zeb families (EMT-TFs). With inclusion of miRNAs, these factors constitute a complex interactome, able to sensor multiple signals received from the proximal microenvironment and relay them onto gene expression (2, 3). Generally undetectable in adult epithelial cells in homeostatic conditions, EMTTFs were found to be recurrently expressed in various types of cancers including multiple carcinomas, an expression often associated with a high metastatic risk. An outstanding amount of work has been performed over the last years evaluating their relative contribution to the initiation and maintenance of the EMT process, with a significant impact on the prognostic value occurring following to their detection in primary tumors or in disseminated cancer cells (4, 5). Does this mean that the oncogenic potential of these factors in epithelial cells only rely on their ability to promote EMT or does their oncogenic activity extend further than EMT induction? By performing in vitro functional assays, we and others firstly demonstrated that several of these embryonic transcription factors, including the TWIST proteins, alleviated fail-safe program induction. Through this, they were shown to cooperate with mitogenic oncoproteins such as RAS and N-MYC in promoting cell transformation in vitro, as well as in lung and breast carcinoma development in transgenic mouse models (4, 6–9). The underlying mechanisms have been largely explored. The TWIST1 protein was found to directly interact with p53 and to destabilize the oncosuppressive protein by altering specific post-translational modifications (10). Furthermore, TWIST1 alleviates induction of cyclin-dependent kinase inhibitors (CDKN1A,CDKN2A, and CDKN2B), thereby sustaining cell proliferation (8, 11). These pleiotropic properties probably provide TWIST proteins unique properties. As TWIST proteins experimentally are inefficient in triggering EMT as compared with SNAIL and ZEB proteins, their main functions may consist in protecting cells from fail-safe programs during the EMT-associated genetic reprograming. In this respect, they might be considered as survival factors rather than EMT inducers. Although SNAIL and ZEB proteins, unlike the TWIST transcription factors, fail to prevent HRASG12Vinduced senescence in murine embryonic fibroblasts (B. Gras, unpublished data), we obviously cannot exclude that they facilitate the escape from oncogene-induced fail-safe programs in other cellular settings and/or experimental conditions. In support of this assumption, ZEB proteins were indeed reported to protect lung cancer cells from EGFR-induced senescence through their ability to down-modulate CDKN1A and CDKN2B (12). In line with this observation, ZEB1 was demonstrated to be positioned downstream of RB and to contribute to fibroblast immortalization induced by RB and RB-like protein depletion (13). Enforced expression of SNAIL or ZEB proteins in non-transformed mammary epithelial cells, and the consequent activation of RAS-downstream pathways, predominantly triggers EMT. Nonetheless, it also accidentally promotes the commitment of cells into a senescence program, unveiled by their SA-β-galactosidase activity (B Gras and SA, unpublished data). This observation is consistent with the reported anti-proliferative properties of SNAIL and ZEB proteins in epithelial cells (14) and with the recognized antagonism between cell proliferation and dissemination. It additionally gives a rationale to the restricted staining of these transcription factors to the tumor-stromal interface, stabilized by microenvironmental EMTpermissive conditions (15–17). The need to maintain ZEB and SNAIL proteins at