with regard to SNAIL (SNAI), known as SNAI1 and SNAI2 in
vertebrates. However, it is clear that the pathways involved are
quite complex and not well conserved in controlling cell shape
changes, the initiation of migration, and the engagement into
differentiation. For instance, in the sea urchin, none of the contributing pathways alone can complete gastrulation [4]. Thus, one
cannot infer that one EMT transcription factor is sufficient to
execute the full gastrulation process although, when overexpressed
in some cell lines in vitro, such a factor can lead to morphological
transitions. These findings point to the intricate nature of EMT
signal transduction that is established during evolution, a process
that appears to ensure the robustness of this fundamental event.
Our understanding of EMT has considerably improved over
the past 10 years [5], particularly following the discovery of the
contributions of epigenetics [6], splicing, [7], posttranslational
modifications [8], metabolism [9], and redox states [10]. Notably,
EMT is now shown to require ribosome biogenesis, which occurs at
the G1/S restriction point [11] in cell lines and in vivo during
neural crest cell dissociation from the neural fold, as well as in their
early migratory phase. Snail1 recruited to rDNA operons displaces
the repressive nucleolar chromatin complex and becomes associated
with the rRNA along with the core components of polymerase
1 and rictor, a component of the mammalian target of rapamycin.
This mechanism also contributes to the maintenance of the dedifferentiated state and the promotion of invasion and metastasis.
These events can be abrogated by CX-5461, a drug that interferes
with rRNA biogenesis. Some recent efforts have been made to
better understand the intricate feedback loop controlling EMT
and MET. A hysteretic mechanism has now been revealed to ensure
the robustness of the EMT process [12]. The mechanism—driven
by the mir-200/Zeb1 double-negative feedback loop—ensures
that the level of E-cadherin remains high in a concentration range
of the TGFβ inducer if the cells are in an epithelial state, and low in a
mesenchymal state. This so-called bi-stability of E-cadherin is lost if
the Zeb1 binding site on the E-cadherin promoter is deleted,
leading to a gradual progression of EMT. Such high sensitivity
and memory of hysteretic EMT allows cells to accomplish the full
metastatic cascade upon just a short exposure to the inducer.
Several EMT scoring methods have been published recently,
albeit with different degrees of robustness [13, 14]. The scoring
methods more precisely position cells in distinct intermediate
stages; however, these scores fail to tell us how these intermediate
stages are achieved. We now require a matrix of scores that provides
“n”-dimensional computations to position each of the mechanisms
that contribute to these intermediate stages. The debate surrounding whether EMT is dispensable for metastasis is directly linked to
36
Jean Paul Thiery
vertebrates. However, it is clear that the pathways involved are
quite complex and not well conserved in controlling cell shape
changes, the initiation of migration, and the engagement into
differentiation. For instance, in the sea urchin, none of the contributing pathways alone can complete gastrulation [4]. Thus, one
cannot infer that one EMT transcription factor is sufficient to
execute the full gastrulation process although, when overexpressed
in some cell lines in vitro, such a factor can lead to morphological
transitions. These findings point to the intricate nature of EMT
signal transduction that is established during evolution, a process
that appears to ensure the robustness of this fundamental event.
Our understanding of EMT has considerably improved over
the past 10 years [5], particularly following the discovery of the
contributions of epigenetics [6], splicing, [7], posttranslational
modifications [8], metabolism [9], and redox states [10]. Notably,
EMT is now shown to require ribosome biogenesis, which occurs at
the G1/S restriction point [11] in cell lines and in vivo during
neural crest cell dissociation from the neural fold, as well as in their
early migratory phase. Snail1 recruited to rDNA operons displaces
the repressive nucleolar chromatin complex and becomes associated
with the rRNA along with the core components of polymerase
1 and rictor, a component of the mammalian target of rapamycin.
This mechanism also contributes to the maintenance of the dedifferentiated state and the promotion of invasion and metastasis.
These events can be abrogated by CX-5461, a drug that interferes
with rRNA biogenesis. Some recent efforts have been made to
better understand the intricate feedback loop controlling EMT
and MET. A hysteretic mechanism has now been revealed to ensure
the robustness of the EMT process [12]. The mechanism—driven
by the mir-200/Zeb1 double-negative feedback loop—ensures
that the level of E-cadherin remains high in a concentration range
of the TGFβ inducer if the cells are in an epithelial state, and low in a
mesenchymal state. This so-called bi-stability of E-cadherin is lost if
the Zeb1 binding site on the E-cadherin promoter is deleted,
leading to a gradual progression of EMT. Such high sensitivity
and memory of hysteretic EMT allows cells to accomplish the full
metastatic cascade upon just a short exposure to the inducer.
Several EMT scoring methods have been published recently,
albeit with different degrees of robustness [13, 14]. The scoring
methods more precisely position cells in distinct intermediate
stages; however, these scores fail to tell us how these intermediate
stages are achieved. We now require a matrix of scores that provides
“n”-dimensional computations to position each of the mechanisms
that contribute to these intermediate stages. The debate surrounding whether EMT is dispensable for metastasis is directly linked to
36
Jean Paul Thiery