regulated by a set of evolutionarily conserved molecules [8]. An
epithelium is organized by cells that have uniform cellular
polarity and form relatively stable intercellular junctions and
cell–matrix interactions, whereas a mesenchyme is a collection of
non-uniformly polarized cells with more dynamic intercellular and
cell–matrix interactions. The value of any molecular or cellular
descriptor of EMT/MET, therefore, would depend on how it can
effectively and faithfully distinguish the mesenchymal and epithelial
populations. For example, the presence or absence of E-Cadherin
(E-cad) staining is not a good indicator of an epithelium, but its
sub-cellular localization to the apical adherens junctions and the
lateral membrane is. Such descriptive value is further enhanced by
differentiating E-cad from P-Cadherin (P-cad), which is a closely
related cadherin with distinct expression patterns but is recognized
by most E-cad antibodies [9, 10]. Similarly, predictive value of most
EMT/MET markers can be substantially increased if attention is
paid to these (and other similar) cellular and molecular details. It is
reasonable to expect that, for any given full or partial EMT/MET, a
set of descriptors could be extracted in the near future by means of
bioinformatics and computational biology.
With increased precision in EMT/MET description, one can
start to address the fundamental question of whether a limited set
of molecular descriptors would be sufficient to characterize a myriad of partial EMT/METs and whether all partial EMT/METs
could fit in a 1- or 2-dimensional continuum between the two
end states (fully epithelial and fully mesenchymal) (Fig. 2). With
extreme diversity in physiological circumstances underlying various
normal and disease EMT/METs, one would be tempted to believe
otherwise. However, a more optimist view would emerge if we view
Fig. 2 Schematic representation of EMT/MET and partial EMT/MET features.
Tumor metastasis preferentially occurs in partial epithelial status
Partial EMT/MET: An Army of One
31
epithelium is organized by cells that have uniform cellular
polarity and form relatively stable intercellular junctions and
cell–matrix interactions, whereas a mesenchyme is a collection of
non-uniformly polarized cells with more dynamic intercellular and
cell–matrix interactions. The value of any molecular or cellular
descriptor of EMT/MET, therefore, would depend on how it can
effectively and faithfully distinguish the mesenchymal and epithelial
populations. For example, the presence or absence of E-Cadherin
(E-cad) staining is not a good indicator of an epithelium, but its
sub-cellular localization to the apical adherens junctions and the
lateral membrane is. Such descriptive value is further enhanced by
differentiating E-cad from P-Cadherin (P-cad), which is a closely
related cadherin with distinct expression patterns but is recognized
by most E-cad antibodies [9, 10]. Similarly, predictive value of most
EMT/MET markers can be substantially increased if attention is
paid to these (and other similar) cellular and molecular details. It is
reasonable to expect that, for any given full or partial EMT/MET, a
set of descriptors could be extracted in the near future by means of
bioinformatics and computational biology.
With increased precision in EMT/MET description, one can
start to address the fundamental question of whether a limited set
of molecular descriptors would be sufficient to characterize a myriad of partial EMT/METs and whether all partial EMT/METs
could fit in a 1- or 2-dimensional continuum between the two
end states (fully epithelial and fully mesenchymal) (Fig. 2). With
extreme diversity in physiological circumstances underlying various
normal and disease EMT/METs, one would be tempted to believe
otherwise. However, a more optimist view would emerge if we view
Fig. 2 Schematic representation of EMT/MET and partial EMT/MET features.
Tumor metastasis preferentially occurs in partial epithelial status
Partial EMT/MET: An Army of One
31