vimentin) (reviewed in [1, 5]), thus completing its mesenchymalization. Recently, however, our perception of what constitutes
“complete” EMT has expanded from this former all-or-nothing
view to one that instead exists as a spectrum [3, 6].
During developmental EMT, mesenchymalization requires the
loss of stable basement membrane adhesions and, subsequently, a
loss of apico-basal polarity. Beyond this delamination event, there
exists a wide array of “mesenchymal” phenotypes (based on stereotypical marker expression and morphology) across organisms and
migratory cell types (reviewed in [1, 7]). Neural crest cells are a
classic example of the ranges of developmental EMT; in the chick,
these cells lose epithelial characteristics to delaminate from the
neuroepithelium and acquire mesenchymal characteristics to
migrate largely as individual cells. In Xenopus, neural crest cells
undergo collective migration, wherein cells retain some epithelial
characteristics and cell–cell connections as they migrate (reviewed
in [8, 9]). Interestingly, the collective migration of lateral line cells
in zebrafish and border cells in Drosophila exhibit even greater
retention of epithelial characteristics (reviewed in [6]). This span
of migration and EMT phenotypes observed across development
supports the notion that there exists a sliding scale of EMT, where
partial mesenchymalization may be a normal end-point for EMT in
certain contexts.
The notion that this “partial” EMT is observed in developmental contexts is interesting in light of recent discoveries describing a
similar partial EMT phenomenon in cancer cells [10–12]. Singlecell transcriptomics have uncovered tremendous heterogeneity in
tumor cells, and importantly, revealed partial EMT programs
through which these cells transit to become invasive [13]. Cancer
cells that undergo partial EMT often exhibit collective migration
[14], which imbues them with increased therapy resistance and
generates tumor cells that, based on molecular signatures, retain
greater plasticity and ability to transdifferentiate (reviewed in [2]).
Thus, from a biomedical perspective, complete understanding of
the spectrum of tumor EMT states, and the molecular factors
driving these transitions, will likely provide novel and improved
targets for cancer therapies.
The mechanisms underlying EMT have previously been largely
attributed to “classical” transcription-dependent cellular changes
(reviewed in [1, 15]); though undoubtedly important, there are
additional modes of regulation beyond transcription—from posttranscriptional to post-translational—that are emerging as critical
regulators of EMT (reviewed in [2, 3, 12]). Direct modification to
or localization of EMT-related proteins, from transcription factors
to adhesion molecules, has critical roles in regulating the progression of EMT programs [11, 16, 17]. In addition, posttranscriptional regulatory mechanisms mediated by miRNAs,
lncRNAs, and RNA-binding proteins are increasingly implicated
4
Erica J. Hutchins and Marianne E. Bronner
“complete” EMT has expanded from this former all-or-nothing
view to one that instead exists as a spectrum [3, 6].
During developmental EMT, mesenchymalization requires the
loss of stable basement membrane adhesions and, subsequently, a
loss of apico-basal polarity. Beyond this delamination event, there
exists a wide array of “mesenchymal” phenotypes (based on stereotypical marker expression and morphology) across organisms and
migratory cell types (reviewed in [1, 7]). Neural crest cells are a
classic example of the ranges of developmental EMT; in the chick,
these cells lose epithelial characteristics to delaminate from the
neuroepithelium and acquire mesenchymal characteristics to
migrate largely as individual cells. In Xenopus, neural crest cells
undergo collective migration, wherein cells retain some epithelial
characteristics and cell–cell connections as they migrate (reviewed
in [8, 9]). Interestingly, the collective migration of lateral line cells
in zebrafish and border cells in Drosophila exhibit even greater
retention of epithelial characteristics (reviewed in [6]). This span
of migration and EMT phenotypes observed across development
supports the notion that there exists a sliding scale of EMT, where
partial mesenchymalization may be a normal end-point for EMT in
certain contexts.
The notion that this “partial” EMT is observed in developmental contexts is interesting in light of recent discoveries describing a
similar partial EMT phenomenon in cancer cells [10–12]. Singlecell transcriptomics have uncovered tremendous heterogeneity in
tumor cells, and importantly, revealed partial EMT programs
through which these cells transit to become invasive [13]. Cancer
cells that undergo partial EMT often exhibit collective migration
[14], which imbues them with increased therapy resistance and
generates tumor cells that, based on molecular signatures, retain
greater plasticity and ability to transdifferentiate (reviewed in [2]).
Thus, from a biomedical perspective, complete understanding of
the spectrum of tumor EMT states, and the molecular factors
driving these transitions, will likely provide novel and improved
targets for cancer therapies.
The mechanisms underlying EMT have previously been largely
attributed to “classical” transcription-dependent cellular changes
(reviewed in [1, 15]); though undoubtedly important, there are
additional modes of regulation beyond transcription—from posttranscriptional to post-translational—that are emerging as critical
regulators of EMT (reviewed in [2, 3, 12]). Direct modification to
or localization of EMT-related proteins, from transcription factors
to adhesion molecules, has critical roles in regulating the progression of EMT programs [11, 16, 17]. In addition, posttranscriptional regulatory mechanisms mediated by miRNAs,
lncRNAs, and RNA-binding proteins are increasingly implicated
4
Erica J. Hutchins and Marianne E. Bronner