Emt: 2016

Leading Edge Review EMT: 2016 M. Angela Nieto,1,6,* Ruby Yun-Ju Huang,2,3,6 Rebecca A. Jackson,4,6 and Jean Paul Thiery3,4,5,6,* 1Instituto de Neurociencias CSIC-UMH, Avda. Ramón y Cajal s/n, 03550 San Juan de Alicante, Spain of Obstetrics & Gynaecology, National University Hospital, Singapore 119228, Singapore 3Cancer Science Institute of Singapore, National University of Singapore, Singapore 117599, Singapore 4Department of Biochemistry, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 117596, Singapore 5Institute of Molecular and Cell Biology, A-STAR, Singapore 138673, Singapore 6Co-first authors *Correspondence: [email protected] (M.A.N.), [email protected] (J.P.T.) http://dx.doi.org/10.1016/j.cell.2016.06.028 2Department The significant parallels between cell plasticity during embryonic development and carcinoma progression have helped us understand the importance of the epithelial-mesenchymal transition (EMT) in human disease. Our expanding knowledge of EMT has led to a clarification of the EMT program as a set of multiple and dynamic transitional states between the epithelial and mesenchymal phenotypes, as opposed to a process involving a single binary decision. EMT and its intermediate states have recently been identified as crucial drivers of organ fibrosis and tumor progression, although there is some need for caution when interpreting its contribution to metastatic colonization. Here, we discuss the current state-of-the-art and latest findings regarding the concept of cellular plasticity and heterogeneity in EMT. We raise some of the questions pending and identify the challenges faced in this fast-moving field. Introduction During embryonic development, cells can transition between epithelial and mesenchymal states in a highly plastic and dynamic manner. A shift toward the mesenchymal state, in a process referred to as epithelial-mesenchymal transition (EMT), modifies the adhesion molecules expressed by the cell, allowing it to adopt a migratory and invasive behavior. The reverse of this process—mesenchymal-epithelial transition (MET)—is associated with a loss of this migratory freedom, with cells adopting an apico-basal polarization and expressing the junctional complexes that are hallmarks of epithelial tissues (Thiery et al., 2009). The EMT is executed in response to pleiotropic signaling factors that induce the expression of specific transcription factors (TFs) called EMT-TFs (e.g., Snail, Zeb, Twist, and others) and miRNAs together with epigenetic and post-translational regulators, many of which are involved in embryonic development, wound healing, fibrosis, and cancer metastasis. Indeed, the significant parallels between cell plasticity in embryonic development and carcinoma progression led to the proposition that EMT is a central driver of epithelial-derived tumor malignancies (Cano et al., 2000; Thiery, 2002). EMT has since been shown to trigger the dissociation of carcinoma cells from primary carcinomas, which subsequently migrate and disseminate to distant sites. Significantly, it is the MET that is then believed to trigger the cessation of migration, inducing the same cells to proliferate and seed the new tumor. The classic description of EMT as the transformation of epithelial cells into mesenchymal cells (Hay, 1995) may have invoked the perception of this process as a shift between two alternative states, mesenchymal or epithelial. However, the change from ‘‘transformation’’ to ‘‘transition’’ more than a decade ago (first meeting of The Epithelial-Mesenchymal Transition International Association [TEMTIA]) already conveyed the concept of plasticity and transitional states. It is therefore not surprising that more recent work points to a greater flexibility in this transitional process, and cells are no longer thought to oscillate between the full epithelial and full mesenchymal states, but rather, they move through a spectrum of intermediary phases. This purported plasticity means that cells could linger in intermediary stages and that they may frequently undergo a partial EMT program (Figure 1). Much of the work into the mechanisms regulating EMT has been carried out on cultured cells. However, the identification of hybrid intermediate EMT states during organ fibrosis and in circulating tumor cells provides evidence that the spectrum of intermediary phases previously observed in cultured cells reflects what happens in vivo. In this review, we will discuss these intermediate states of EMT, assessing whether the acquisition of stemness depends on EMT and exploring some of the latest findings regarding EMT in development and disease. We will also deliberate on the impetus to exploit EMT for therapeutic gain and pose some pending questions and challenges in this expanding field. Broadening the Reductionist’s View: Intermediate EMT Phenotypes From a reductionist’s viewpoint, the operational definition of EMT describes the loss and gain of certain characteristics during a single transition between two states. As such, most experimental models tend to consider a dramatic change in the expression of selected epithelial and mesenchymal markers to confirm EMT (Thiery and Sleeman, 2006): E-cadherin, occludins, and Cell 166, June 30, 2016 ª 2016 Elsevier Inc. 21 Figure 1. Transitions through the Different States along the EMT Spectrum EMT can be considered as a continuum, whereby cells exhibit epithelial (E), intermediate (EM), and mesenchymal (M) phenotypes. As cells make the transition from left to right, they sequentially lose apico-basal polarity and cell-cell adhesions and they gain front-back polarity and enhanced cellmatrix interactions. The colored spectrum denotes hypothetical transitions (x axis). Different models predict that both stability and metastability could occur during the phase transitions where there is a point of low energy or a net thermodynamic equilibrium (y axis), in function of the intrinsic, multi-parametric EMT regulators (z axis). These regulators include transcription factors (SNAI1/2, ZEB, TWIST1, GRHL2, OVOL1/2, and PRRX1), post-transcriptional regulators (miRNAs), and the proposed epigenetic controls at the promoters of epithelial genes (Tam and Weinberg, 2013). When the system is perturbed in response to EMT signals (e.g., hypoxia, growth factors, etc.), cells could pass through a thermodynamic hump that leads to a distinct phenotype (e.g., EM1; EM3). The backward arrows represent reversal to more epithelial states. It is worth noting that the existence of discrete metastable states (e.g., EM1, EM3) has not been proven and that the existence of quasi-stable transitional states is also predicted and found in different contexts, including fibrosis (e.g., EM2). Importantly, although full reversibility is evident during embryonic development, it is not clear whether there is a ‘‘point of no return’’ in the EMT undertaken by adult cells. Finally, it seems that, although the mesenchymal state prevents metastatic outgrowth, a full reversion to the epithelial phenotype is not required for metastatic colonization. TJ, tight junction; AJ, adherens junction; DS, desmosome. cytokeratins are the most commonly used markers for the epithelial trait and N-cadherin and vimentin for the mesenchymal (Thiery et al., 2009). This oversimplification ultimately led to some confusion and considerable debate as to whether EMT actually occurs in some circumstances, such as during carcinoma progression (Tarin et al., 2005). However, the definition of EMT has now been broadened, based on many observations and particularly those regarding the so-called partial EMT. A partial EMT state has been noted in association with many developmental, wound healing, fibrosis, and cancer processes (Arnoux et al., 2008; Blanco et al., 2007; Futterman et al., 2011; Grande et al., 2015; Leroy and Mostov, 2007; Lovisa et al., 2015; Grigore et al., 2016), evident through the existence of intermediate hybrid epithelial and mesenchymal phenotypes (Abell et al., 2011; Huang et al., 2013; Jordan et al., 2011; Yu et al., 2013). Cells bearing this hybrid phenotype have been referred to as ‘‘metastable’’ (Lee et al., 2006; Tam and Weinberg, 2013), reflecting the flexibility of these cells to induce or reverse the EMT process (Rosanò et al., 2006). In physics and chemistry, the metastable state denotes an isolated system that is maintained in a delicate equilibrium and not in a state of least energy (Cheng and Keller, 1998). Thus, metastability suggests a dynamic phase transition along the energy gradient. Adapting this concept of metastability to EMT probably fails to accurately reflect thermodynamic laws. Direct evidence of the energy transition that occurs during the EMT process is limited, although a 22 Cell 166, June 30, 2016 linear energy gradient based on epigenetic changes between different EMT states has been hypothesized (Figure 1) (Tam and Weinberg, 2013). More recently, the free-energy landscape of EMT was modeled, proposing three thermodynamically and kinetically stable states for EMT: the epithelial, maturation, and mesenchymal states (Zadran et al., 2014). Energy is released after exiting the epithelial state as the cell moves to a maturation state, which corresponds to an intermediate EMT state (Zadran et al., 2014), and this energy is then consumed for the cell to reach the mesenchymal state. Thus, the intermediate EMT state in this model is not equivalent to the metastable state described with biophysical models. Computational modeling, including those that consider the mutual inhibitory loops between several microRNAs (miRNAs) and EMT transcriptional drivers like Snail1 and Zeb1, also accepts an intermediate hybrid EMT state that could favor the progress of developmental programs and metastatic potential (Jolly et al., 2015; Lu et al., 2013; Tian et al., 2013; Zhang et al., 2014). The inclusion of additional reciprocal inhibitory loops that involve other transcription factors (e.g., Zeb1 with Ovol2 and Grhl2) and the description of these as phenotypic stability factors indicates that the network is capable of generating additional intermediate stabilized states that, therefore, are not necessarily metastable (Hong et al., 2015; Jolly et al., 2016). Figure 1 depicts putative intermediate states in a hypothetical diagram to highlight the transitional nature of the process, which can include stable and metastable states. Figure 2. The Core Regulatory Machinery of EMT Although multiple non-coding RNAs control EMT, two regulatory networks have been described that can be considered as the core regulatory machinery: the miR34-SNAI1 and miR200-ZEB1 axes. These two miR-transcription factor (TF) axes employ a double-negative feedback mechanism in which miR34-SNAI1 and miR200-ZEB1 repress each other. SNAI1 and ZEB1 further repress miR200 and miR34, respectively. At this level, other TFs, such OVOL1/2 and GRHL2, further feed into this core regulatory machinery to protect the epithelial phenotype. The miR34-SNAI1 and miR200-ZEB1 core not only contributes to the epigenetic control of EMT, but are also targets for epigenetic modifications. Downstream of the miR34-SNAI1 and miR200-ZEB1 axes, regulation of transcript processing would shape the landscape of epithelial and mesenchymal effectors, for example, through alternative splicing. These EMT effectors may also be subject to epigenetic modifications. How other TFs, like TWIST1 and PRRX1, affect the network and its epigenetic control requires further study, although it is clear that, unlike Snail and Zeb, they are much more potent mesenchymal promoters than epithelial repressors. Evidence of transitional EMT states comes from multiple contexts. Whereas many studies use the co-expression of epithelial and mesenchymal markers to define the hybrid state (Huang et al., 2013; Yu et al., 2013), cells do not need to gain mesenchymal traits in a partial EMT. Indeed, early EMT might only involve a dampening of epithelial properties like apico-basal polarity and a remodeling of junctional complexes in favor of cell-substrate adhesions (Huang et al., 2012). Even when a mesenchymal trait is gained, there is still much heterogeneity in terms of the mesenchymal characteristic induced. In an analysis of 43 ovarian cancer cell lines, only 50% of cells with an intermediate phenotype (co-expressing cytokeratin and vimentin) induced N-cadherin (Huang et al., 2013). This heterogeneity may reflect different biological consequences of the intermediate states, such that cells losing E-cadherin that do not gain N-cadherin will most likely behave distinctly during migration and invasion to those that do gain N-cadherin, given their different adhesive behavior (Chu et al., 2006; Halbleib and Nelson, 2006). It is also worth noting that the partial EMT seems to be a final state in some cases, as described during organ fibrosis. Dedifferentiated renal epithelial cells do not seem to undergo a full EMT nor do they undergo reversal during the course of the disease. Importantly, they do not engage in the invasion program but rather they remain integrated in the tubules (Grande et al., 2015). Hence, it is important to consider not only epithelial or mesenchymal traits during epithelial transitions but also other programs associated with EMT, such as invasion, increased survival or decreased proliferation. In sum, intermediate states probably reflect the delicate balance of transcriptional drivers and suppressors of EMT. This equilibrium is inevitably affected by epigenetic changes (Tam and Weinberg, 2013) and by the main effectors in the cyto- skeleton, the cellular machinery driving migration and invasion. This wider view of EMT offers a more dynamic interpretation of the fluidity and plasticity of this phenomenon (Figure 1). Complex Regulatory Networks of EMT: Beyond Transcriptional Control EMT is modulated at different levels by integrating epigenetic modifications, transcriptional control, alternative splicing, protein stability, and subcellular localization. Thus, the mechanisms that govern the intermediary phases of EMT are non-linear. Despite the efforts to identify common regulatory networks in normal development and disease, it is clear that some pathways may be unique to a given EMT event in a particular tissue or tumor subset. The regulation of classical EMT centers on the transcriptional suppression of the prototypic adhesion molecule, E-cadherin (Lamouille et al., 2014; Thiery et al., 2009). The activities of major EMT-TFs, such as SNAI1, SNAI2, ZEB1, ZEB1, and TWIST1, have been described and reviewed extensively (Lamouille et al., 2014; Peinado et al., 2007). With these EMT-TFs taking center stage over the past decade, our understanding of EMT has evolved to explain how their transcripts are expressed and processed by miRNAs (Lamouille et al., 2013) or alternative splicing (Warzecha and Carstens, 2012). Multiple miRNAs are thought to govern EMT, as reviewed elsewhere (Dı́az-López et al., 2014). As a prototypic regulatory model, we focus here on the more elaborate networks involving miR-200 and miR-34, together with ZEB1 and SNAI1. This regulatory network (Figure 2) establishes a negative feedback that is thought to maintain epithelial homeostasis under normal conditions (Puisieux et al., 2014). Indeed, various epithelial markers are downstream targets of this regulatory loop, including E-cadherin, claudins, and occludins (Lamouille et al., 2014). Alternative Cell 166, June 30, 2016 23 splicing mediated by epithelial-specific regulatory protein 1 and 2 (ESRP1 and ESRP2) also affects the maintenance of epithelial features (Warzecha and Carstens, 2012). A panel of transcripts involved in cell-cell adhesion, cell motility, and cell-matrix adhesion are regulated by these ESRPs, including fibroblast growth factor receptor 2 (FGFR2), p120-catenin, CD44, and Mena (Warzecha and Carstens, 2012). Conversely, splicing programs promoted by Quaking (QKI), SRSF2, and RBFOX2 are upregulated during EMT (Bonomi et al., 2013; Braeutigam et al., 2014; Conn et al., 2015). QKI promotes the generation of circular RNAs (circRNAs) by non-sequential back-splicing of premRNAs, and RBFOX2 has multiple targets that include the receptor tyrosine kinase Ron, cortactin, and dynamin (Figure 2). Therefore, the maintenance or loss of epithelial features is governed by how transcripts are processed to their mature forms prior to translation. This systems view provides an alternative way to describe the EMT spectrum in terms of regulatory networks, both quantitatively and qualitatively. Importantly, the miRNA-TF regulatory circuits are not linear in EMT, as there is substantial diversity in the regulation exerted by different miRNAs in EMT. Within the miR-200 family, miR-200a/ b/c, miR-141, and miR-429 all target ZEB1, yet they probably inhibit ZEB1 translation to a different extent. It is unclear how other novel TFs recently implicated in EMT or MET affect the interaction between the miR34/SNAIL and miR200/ZEB regulatory circuits, such as PRRX1 (Ocaña et al., 2012), grainyheadlike 2 (GRHL2) (Cieply et al., 2012), or connective tissue growth factor (CTGF) (Chang et al., 2013). GRHL2 and OVOL1/2 form a double-negative feedback loop with ZEB1 (Figure 2) (Hong et al., 2015), and GRHL2 directly regulates the transcription of miR-200 and OVOL1/2 upstream of ZEB1 (Chung et al., 2016; Cieply et al., 2012). PRRX1 induces EMT in a SNAI1-independent manner, yet cooperates with TWIST1 to drive invasiveness (Ocaña et al., 2012). Both PRRX1 and TWIST1 are more potent mesenchymal inducers than epithelial repressors, and SNAI1 and ZEB1 are strong epithelial repressors and weaker mesenchymal promoters. In contrast to SNAIL and ZEB family members, there is still little information regarding the regulation of TWIST and PRRX by miRNAs or by other non-transcriptional mechanisms (Figure 2). Upstream and downstream epigenetic controls largely affect the EMT miRNA-SNAIL-ZEB networks, as these EMT transcriptional drivers are involved in a complex epigenetic regulatory loop (Bedi et al., 2014). Interestingly, their regulation might involve transitions through intermediate states (Jordan et al., 2011). SNAI1 recruits chromatin modifiers, including SIN3A, histone deacetylase 1 (HDAC1), HDAC2 (Peinado et al., 2004), lysine-specific demethylase1 (LSD1) (Lin et al., 2010), components of the polycomb repressor-2 (PRC2) complex (Herranz et al., 2008), and the G9a and Suv39H1 histone methyltransferases (Dong et al., 2013; Lin et al., 2014). SNAI1 binds transiently to its target promoters, triggering both transient and long-lasting chromatin changes (Javaid et al., 2013). In addition, SNAI1 can interact with a lysyl oxidase, LOXL2 (Peinado et al., 2005), and regulate heterochromatin transcription (Millanes-Romero et al., 2013). On the other hand, ZEB1 represses its target genes by recruiting the LSD1-containing co-repressor complex (Wang et al., 2007), as well as HDAC1 and HDAC2 (Aghdassi 24 Cell 166, June 30, 2016 et al., 2012). ZEB proteins are required for the co-repression in Jurkat cells mediated by a histone acetyltransferase (HAT) domain-containing protein, Tip60 (Hlubek et al., 2001), and ZEB1 itself is affected by the recruitment of PRC2 by the histone methyltransferase, PRDM14 (Chan et al., 2013). Both inactive (H3K27Me3) and active (H3K4Me3) histone marks can be found at the ZEB1 promoter region, presumably to facilitate a switch during EMT (Chaffer et al., 2013; Spaderna et al., 2008). Therefore, the consequences of the epigenetic changes initiated by these EMT-TFs during EMT are expected to confer dynamic control over chromatin and nuclear organization during the transcription of downstream EMT effector genes, thereby contributing to the plasticity evident during transitions. The miR-200 family is also subjected to epigenetic modifications. The expression of miR-200 family members is regulated by a histone demethylase, KDM5B (Enkhbaatar et al., 2013). The miR-200b-200a-429 cluster is regulated by histone modifications mediated by the polycomb group and the EMT suppressor GRHL2, and the miR-200c-141 cluster is regulated by DNA methylation (Lim et al., 2013; Chung et al., 2016). This epigenetic regulation of these EMT miRNA-TF networks further complicates the entire regulatory landscape of EMT, especially given that many EMT effector genes can also be directly regulated by histone modifications and methylation, including E-cadherin (Figure 2) (Wu et al., 2012). Histone acetylation is altered in epithelial, intermediate, and mesenchymal trophoblast stem cells, and these changes have been correlated with the expression of different TF combinations (Abell et al., 2011; Jordan et al., 2011). Accordingly, an ‘‘EMT acetylome’’ could be derived by crossing these changes in acetylation with known EMT signatures (Jordan et al., 2011). At the protein level, EMT is also regulated by post-translational modifications including the ubiquitin-mediated degradation of important effectors, such as that following the phosphorylation of Snail1 and Twist1 by GSK3b or MAP kinases, respectively (Hong et al., 2011; Zhou et al., 2004). F-box proteins also target these and other EMT-TFs to the proteasome (Diaz and de Herreros, 2016). Phosphorylation also impinges into Snail1 subcellular localization (Du et al., 2010; Zhang et al., 2012a), in which nuclear import and export pathways have also been characterized (Domı́nguez et al., 2003; Mingot et al., 2009, 2013). As such, it seems to be a daunting task to construct a systems landscape that encompasses the regulatory networks involving miRNA-TF loops, transcript processing, and the epigenetic control of protein stability to determine what controls the continuous transitions through the complete EMT spectrum. However, the recent demonstration that a generic EMT signature is common to various cancers (Tan et al., 2014) and other advances in our understanding of EMT in physiological and pathological conditions makes it reasonable to believe that such broad systems analyses are not so far away. New Insights for EMT in Development EMT drives important aspects of embryonic development. This phenomenon became apparent to developmental biologists in the late nineteenth century, presenting clear histological evidence that, during gastrulation, mesenchymal cells originated from the adjacent epiblast. EMT was established experimentally by showing that embryonic and adult epithelia embedded in 3D collagen gels converted into migratory, invasive fibroblast-like cells (Greenburg and Hay, 1982). This seminal study provided evidence that, contrary to popular belief, embryonic and also adult epithelial tissues were less stable than expected. Different model systems were explored to define the molecular mechanisms driving this remarkable conversion. One important finding was that rapid changes in epithelial cell shape were driven by the complementary modulation of cell-cell adhesion and cell-substrate interactions, mediated by N-CAM and E-cadherin, as well as extracellular matrix (ECM) proteins such as fibronectin and integrin receptors, respectively (Boucaut et al., 1984; Edelman et al., 1983; Takeichi, 1988). Another landmark was the discovery that an ortholog of the Drosophila snail gene, expressed during the formation of the ventral furrow (Leptin and Grunewald, 1990), was critical to the formation of the primary mesenchyme during gastrulation in the chicken embryo (Nieto et al., 1994). The first signal-transduction pathways leading to EMT were then defined from genetic studies in Drosophila and in vertebrate embryos. Here, we integrate some recent findings on EMT in development with the knowledge accumulated over the past two decades. Gastrulation Genetic studies in Drosophila revealed how dorsoventral polarity, mediated by Toll receptor activation, engages a program orchestrated by Twist and Snail. This program involves cell-cycle repression by Tribbles, and the activation of actomyosin contractility by T48 and Fog through RhoGEF2. EMT of the invaginated epithelium in the ventral furrow is then promoted by the activation of heartless, an FGFR tyrosine kinase (Lim and Thiery, 2012). Snail promotes an E- to N-cadherin switch (Oda et al., 1998) that, in conjunction with FGF signaling mediated by heartless activation, initiates EMT in the nascent fly mesoderm. Recent findings question whether the repression of E-cadherin during the invagination process and/or the gain of N-cadherin are in fact key events during fly gastrulation (Schäfer et al., 2014), with the suggestion that other mechanisms control the changes in morphology and the individualization of the cells in the mesoderm. One possibility to reconcile earlier data with these findings could be that E-cadherin turnover is accentuated upon forced overexpression in order to reduce the strength of intercellular adhesion, giving rise to a much weaker phenotype than that anticipated. Clearly, more work is required to clarify how this intricate network is autoregulated when challenged, leading to unpredictable and weak phenotypes and emphasizing the robustness of gastrulation in Drosophila. The analysis of the gene regulatory networks in sea urchin gastrulation also indicated that Snail and Twist execute the EMT process in primary mesenchyme formation (Saunders and McClay, 2014). However, a detailed analysis of blastula epithelial cell behavior defined five distinct sub-circuits involved in EMT: basement membrane remodeling, motility, apical constriction, loss of apico-basal polarity, and changes in intercellular adhesion. Thirteen transcription factors contribute to the execution of EMT, including Snail and Twist, but no single transcription factor is expressed in all five sub-circuits, highlighting the complexity and robustness of the system even in sea urchin (Saunders and McClay, 2014). These findings are consistent with the phenotype observed in mouse embryos, where Snail1 mutant embryos do not survive gastrulation (Carver et al., 2001). However, the specific conditional deletion of Snail1 in the epiblast affects mesoderm migration, but not its initial formation (Lomelı́ et al., 2009; Murray and Gridley, 2006). Strong gastrulation defects are observed in FGFR1 knockout mice (Ciruna and Rossant, 2001). FGF signaling, together with Nodal and Wnt3 signaling, is required for the generation of migratory mesenchymal mesodermal cells (Lim and Thiery, 2012). FGF signaling maintains Snail1 expression, and Snail1 and Sox2/3 mutually repress each other, with Sox3 preventing epithelial cell ingression and Snail1 promoting it by inhibiting Sox3 in the primitive streak in chick and mouse embryos (Acloque et al., 2011). The winged helix-turn-helix Ets2 transcription factor is expressed in extraembryonic ectoderm trophoblasts, and it plays a role in the formation and elongation of the primitive streak in mouse embryos (Polydorou and Georgiades, 2013). Indeed, Ets2 promotes the expression of Elf5 and Cdx2 in order to maintain trophoblast stemness, and it also induces bone morphogenetic protein 4 (BMP4) expression, which acts on the epiblast in a paracrine manner to induce Wnt3. Wnt3 elevates Nodal and induces Snail1, promoting primitive streak elongation and EMT, respectively. These findings add another layer of complexity to EMT signaling in mouse gastrulation. Neural Crest and Mesectoderm At the neurula stage, a group of stem/progenitor cells appears at the dorsal border of the neural plate of vertebrates, the neural crest (NC) (Baggiolini et al., 2015; Buitrago-Delgado et al., 2015), which generates glial cells, most neurons in the peripheral nervous system, some endocrine and paraendocrine cells, and all melanocytes. NC cells share intercellular adhesion systems with adjacent neural epithelial cells that are destined to give rise to the CNS, yet they engage in EMT to form a population of migratory cells that populate distant territories. The mechanisms driving EMT induction in NC cells are only partially shared across species, although a generic gene regulatory network has been established (Simões-Costa and Bronner, 2015), whereby the EMT signaling module is believed to be embedded in a rather complex network. Multiple feed-back loops regulate NC development and guarantee the robustness of the system, anticipating compensatory regulatory mechanisms. Epigenetic regulation also occurs during NC specification and migration, controlling the expression of important EMT regulators (Hu et al., 2014). In the chick embryo, the NC territory at the trunk level is induced by reciprocal rostro-caudal gradients of retinoic acid (RA) and FGF. The territory’s borders are refined by downstream signaling that involves opposing gradients of BMP4/noggin and Wnt1, as well as epigenetic controls that establish poised chromatin domains of genes like Snail, which are activated before cell emigration. In the NC territory, BMP4 induces N-cadherin cleavage by ADAM10 (Shoval et al., 2007), and it creates two opposing gradients to control the downregulation of N-cadherin and the expression of cadherin-6 (Park and Gumbiner, 2010), events that transiently mark the territory from which the NC dissociates. Cadherin-6 is also cleaved by ADAM10 and g-secretase (Schiffmacher et al., 2014), and thus, it is downregulated Cell 166, June 30, 2016 25 following crest cell emigration when these cells acquire cadherin-7. In zebrafish, cadherin-6 acts autonomously in NC progenitors to localize activated RhoA in the apical domain. This restricted distribution promotes actomyosin contractility and constriction of the apical domain, releasing the apical junctional complexes of NC cells to allow them to emigrate from the neural fold (Clay and Halloran, 2014). In summary, the loss of apicobasal polarity, the modulation of cadherins and other adhesion molecules, and the profound remodeling of the actin cytoskeleton allow EMT to occur and cells to acquire a mesenchymal migratory morphology. At cranial levels, mesectodermal cells can also give rise to mesodermal derivatives that include bone, cartilage, and meninges. Believed to be part of the NC, recent findings—still subjected to debate—propose that this may be a distinct yet adjacent ectodermal cell population that also undergoes EMT. These cells would downregulate E-cadherin rather than N-cadherin as they dissociate from the non-neural epithelium of the neural fold (Lee et al., 2013; Weston and Thiery, 2015). However, in Xenopus embryos, migratory cranial cells still express N-cadherin during migration as they pursue collective cell migration, having undergone only a partial EMT (Scarpa et al., 2015). Heart The heart develops from a group of mesodermal mesenchymal cells during gastrulation that assembles as two cardiogenic mesodermal layers (the primary and secondary heart fields), completing a first round of EMT and MET. The second round is activated in the splanchnopleuric mesoderm to form the inner endothelial layer, and it is followed by a third round to form the cardiac valves that arise from the dissociation of endothelial cells in the atrio-ventricular region and outflow track (OFT). This third round should perhaps more precisely be termed an endothelialmesenchymal transition (EndMT). Quite distinct mechanisms drive the successive rounds of EMT/EndMT and MET during heart development (Lim and Thiery, 2012; von Gise and Pu, 2012). Whereas the first round is basically a subprogram of gastrulation that depends on Wnt, FGF, and TGFb signaling, the third is initially driven by the BMP2 produced by myocardial cells. Nevertheless, BMP nullhomozygotes do not exhibit valve defects, indicating that compensatory mechanisms exist similar to those seen in gastrulation and NC development. However, altering downstream canonical signaling, such as interfering with Smad4, leads to a loss of the cardiac cushion. BMP2 can drive local EMT in the atrioventricular region through the activity of its downstream target Tbx2. The atrioventricular endocardium responds to the myocardium through secreted BMP2 and TGFb1/2, triggering canonical Smad pathways. All four ErbB receptors are important for cardiac valve formation, and while ErbB2-ErbB3 heterodimers are activated by neuregulin in the endocardium in an autocrine manner (Lim and Thiery, 2012), ErbB3 signaling is also evident in the myocardium and is driven by the hyaluronic acid produced by hyaluronan synthase (HAS)-2, a Tbx2 target. Notch signaling is another critical pathway activated in the endocardium (Luxán et al., 2016), and all three pathways, Notch, BMP/ TGFb, and ErbB, contribute to endocardial EndMT to establish the valve interstitial cells (Moskowitz et al., 2011; von Gise and Pu, 2012). 26 Cell 166, June 30, 2016 The OFT obeys the same controls in the formation of the conotruncal endocardial cushion, albeit with the additional complexity that the cells arise from the second heart field and the NC. EndMT operates under the Tbx2-BMP2 axis, and when deregulated, retinoic acid (RA) can directly target Tbx2 transcription and reduce BMP2 levels. Indeed, ectopic RA causes defects in the OFT cushion (Sakabe et al., 2012). Mesothelial cells of the sinus venosus located at the posterior end of the embryonic heart form the proepicardium anlage. Bmp2 expression in this anlage is weak, inducing the expression of Wilms tumor gene 1 (Wt1) and Tbx18 to sustain epicardial identity while the cells migrate to the heart primordium to form the primitive epicardium. Pre-epicardial cells migrate and assemble as a single-layered epithelial sheet around the myocardium through a MET-like process. Subsequently, the epicardial layer undergoes EMT and gives rise to resident interstitial fibroblasts within the myocardium, as well as smooth muscle cells in the coronary arteries and at least 20% of the endothelium (Cano et al., 2016). This process has recently been considered as an extended process of the EMT of mesothelial cells, as the primitive epicardium never undergoes a full MET (Ariza et al., 2016). Epicardial EMT is controlled by the neurofibromatosis type 1 (NF1) gene, as its deletion induces premature and more extensive EMT in the epicardium (Baek and Tallquist, 2012). Wt1 is also essential for this process, inducing the expression of either Snail1 or b-catenin to promote cellular dissociation and migration in the subepicardial layer. The basic-loop-helix TCF21 also influences the EMT and lineage specification of epicardial cells (Acharya et al., 2012). On the other hand, the Notch pathway promotes EMT by activating platelet-derived growth factor receptor (PDGFR) and producing TGFb in cooperation with Wt1, which can induce RA production by the Raldh2 enzyme (Lim and Thiery, 2012; Pérez-Pomares and de la Pompa, 2011). Together, these findings show that proper heart development requires a plethora of growth factors and pathways to drive epicardium-myocardium interactions, which are still to be comprehensively integrated (as required in other developmental processes). It is widely accepted that pathways converge on the activation of different combinations of EMT-TFs and that an ‘‘EMT code’’ determines cell behavior, producing heterogeneity and tissue specificity. Establishing such gene regulatory networks and feedback loops ensures the robustness of pivotal developmental processes, as observed during gastrulation and NC development. Although EMT in development has inspired cancer researchers to identify similar mechanisms in the progression of carcinomas, it is now clear that the regulation observed in normal development would not only be used, but also expanded on in cancer, and that additional effort is needed to describe the signaling networks that drive invasion and metastasis. Protection of the Epithelial Phenotype Despite the high degree of epithelial plasticity observed during embryonic development, it is appropriate to stress that, in normal adult epithelial tissues, various mechanisms guarantee epithelial homeostasis and protect the epithelial phenotype in order to maintain tissue integrity and organ function. In addition to the aforementioned global program driven by epithelial-specific, ESRP-mediated splicing (Warzecha and Carstens, 2012), the epithelial-specific H2BK5Ac epigenetic mark also exists (Abell et al., 2011). Moreover, Ovol zinc-finger TFs appear to act as protectors of the epithelial phenotype, with Ovol1/Ovol2 inhibiting the cell’s capacity to engage into EMT by activating epithelial genes and repressing multiple EMT drivers and mesenchymal genes (Lee et al., 2014; Kitazawa et al., 2016). Ovol2 deletion dramatically augments the EMT phenotype in epithelial cells of the mammary gland, provoking considerable structural perturbation during extension of the terminal end buds (Watanabe et al., 2014). Elf5 shares similar properties in protecting the epithelial state by suppressing Snail2 during mammary gland development and metastasis (Chakrabarti et al., 2012). New functions for p53 extend its role of ‘‘guardian of the genome’’ (Lane, 1992) to include ‘‘guardian of the epithelial phenotype,’’ as p53 also safeguards epithelial homeostasis by directly activating the miR-200 family. In mammary epithelial cells, the loss of p53 causes a decrease in miR-200c thereby increasing EMT, with a concomitant increase in the stem cell population (Chang et al., 2011). The latter is in part mediated by the upregulation of Bmi1 (Shimono et al., 2009), which is expressed by cisplatin-resistant bladder cancer cells that behave as cancer stem cell (CSC)-like cells, undergoing EMT with the potential to form spheres (Zhang et al., 2012b). EMT is also prevented by miR-200b, which inhibits mammosphere formation through targeting Suz12, a subunit of the polycomb PRC2 complex (Iliopoulos et al., 2010). Another PRC2 component, EZH2, supresses EMT, invasion, and metastasis by repressing Snail1 in colorectal cancer cells (Wang et al., 2015). Other protectors of the epithelial phenotype include proteins associated with DNA damage repair, such as the ubiquitin ligases FBXW7 and Siah, which prevent EMT by targeting Zeb1 for degradation (Chen et al., 2015; Yang et al., 2015), as well as ataxia-telangiectasia-mutated (ATM) kinase, which represses Sox9 expression (Russell et al., 2015). In some cell contexts, Sox9 behaves as an important transcription factor associated with EMT and stemness. The programs implemented to protect the epithelial phenotype are counterbalanced by other programs that promote EMT. As such, epigenetically driven global reprogramming of specific chromatin domains and specific splicing programs occur during the induction of EMT (Bonomi et al., 2013; Braeutigam et al., 2014; McDonald et al., 2011). While the final response would depend on the relative contribution of these positive and negative EMT programs, in the healthy adult, the protection of epithelial homeostasis will generally prevail. Nevertheless, the remarkable epithelial plasticity observed under physiological conditions in certain tissues is worth noting, such as in the mammary gland during puberty and pregnancy. Ductal elongation of the mammary gland during puberty is driven by terminal end buds. These multi-layered structures exhibit significant epithelial cell plasticity to allow fat pad invasion through the collective migration of cells with an intermediate EMT phenotype (Kurley et al., 2012; Shamir and Ewald, 2015). As the reactivation of EMT programs in the adult go far beyond this process, the following sections discuss the implication of EMT in wound healing, fibrosis, and cancer, with a focus on partial phenotypes in terms of both physiology and pathology. The Partial EMT Associated with Wound Healing During epithelial wound healing, cells at the edge of the wound move into damaged area to re-establish normal tissue architecture in a process referred to as re-epithelialization. Since epithelial cells are fairly immotile, they must undergo behavioral changes to move as a coordinated group of cells and heal the wound, ultimately reverting to their epithelial phenotype to reconstitute epithelial sheet integrity. Both basal and suprabasal cells are involved in the repair process that involves a partial EMT (Shaw and Martin, 2016). The partial EMT is thought to rely on Snail2 under the control of the epidermal growth factor receptor (EGFR)-extracellular signal-regulated kinase 5 pathway in mouse skin explants (Arnoux et al., 2008). As such, wound healing is defective in adult mice deficient for Snail2 (Hudson et al., 2009). EGFR signaling also promotes re-epithelialization in N-acetylglucosaminyltransferase V (GnT-V) transgenic mice through the upregulation of Snail and Twist (Terao et al., 2011). In addition, axon guidance molecules play an important role in this partial transition during wound healing in the mouse skin. As such, upregulation of EphrinB1 in the basal and suprabasal cells downregulate components of the cell-cell adhesion complexes, loosening the junctions to release tension and enable cell movement (Nunan et al., 2015). In fly larvae, Netrin A controls EMT in the squamous epithelium by downregulating Frazzled, a member of the deleted in colorectal carcinoma (DCC) family of surface receptors, which are known to stabilize adherens junctions through the phosphorylation of moesin (Manhire-Heath et al., 2013). These studies offer insight into epithelial breakdown during EMT, which is important given that Netrin is overexpressed in various cancers, and it induces cancer cell proliferation and migration through the activation of YAP signaling (Qi et al., 2015). In the lung, wound healing is also associated with an EMT phenotype, typified by a reduction in cell-cell junctions, cell flattening, the acquisition of migratory properties, and the expression of vimentin. Yet, this transient phenotypic change is limited to the basal cells of the airway and there is no evidence that type II alveolar cells, which are responsible for the production and secretion of surfactant to reduce surface tension, undergo EMT or adopt mesenchymal features. By contrast, the club (or clara) cells in the bronchiole airways appear to undergo a transient EMT program during the regeneration of the bronchiolar epithelium. Club cells act as stem cells, and they can terminally differentiate into ciliated cells during lung tissue repair (Vaughan and Chapman, 2013). Wound healing is often investigated using the dorsal closure model in Drosophila, as EMT- and MET-like events are coordinated during this process. During dorsal closure, epidermal flanks—amnioserosa—migrate to extend over the gap in the epidermis, and as the epidermis progressively expands on the outer surface of the dorsal territory, the transient amnioserosa epithelium becomes internalized. Epidermal pioneer cells at the leading edge undergo a partial EMT, forming filopodia and lamellipodia that actively promote cell migration through Cdc42/Rac-mediated activation of actomyosin contractility. The Scribble apico-basal polarity complex recruits p21-activated kinase (PAK) to the leading edge for polarized cell migration, and when closure is complete, PAK recruits the Scrib complex to Cell 166, June 30, 2016 27 restore septate junctions and apico-basal polarity (Bahri et al., 2010). GRH, the homolog of mammalian GRHL-1, -2 and -3, regulates wound healing in Drosophila amnioserosa through the establishment of septate junctions (Narasimha et al., 2008). GRHL3 is also involved in epithelial homeostasis in the mouse (Ting et al., 2005), where it contributes to wound healing by activating transglutaminase 1 and RhoGEF19 (Darido and Jane, 2010). However, GRHL2 can compensate for the loss of GRHL3 to heal the wounds of hind limb amputated mouse embryos (Boglev et al., 2011), suggesting redundancy is at play. It is important to note that, while EMT in cancer is deleterious, wound-healing-driven EMT induced in response to injury is beneficial. Therefore, analyzing the similarities and differences between these two types of EMT will be essential to better design specific therapies (Leopold et al., 2012). Nonetheless, some EMT-healing responses may also be deleterious, such as opacification of the posterior capsular (PCO), a major complication after surgery for cataract treatment (Xiao et al., 2015), or the exaggerated healing responses that lead to fibrosis and scarring. Interestingly, defective healing and increased scarring was recently associated with prolonged inflammation (Qian et al., 2016), highlighting the tight connection between a sustained inflammatory response and the progression of fibrosis (as discussed below). Reconciling the Role of EMT in Fibrosis Fibrosis, or tissue scarring, is a hallmark of many chronic degenerative disorders and is associated with reduced organ function and eventual organ failure. Fibrotic disease is on the increase; for example, idiopathic pulmonary fibrosis (IPF) is estimated to match the incidence of stomach, liver, testicular, and cervical malignancies (Hutchinson et al., 2015). Fibrosis typically ensues through the progressive accumulation of ECM deposited by myofibroblasts. The origin of these myofibroblasts has been debated for many years, with EMT being considered by some authors to drive the transformation of epithelial and endothelial cells into myofibroblasts (Iwano et al., 2002; Zeisberg et al., 2007b), whereas lineage tracing in transgenic mice indicates that the contribution of those cells to the population of myofibroblasts is negligible (Humphreys et al., 2010; Li et al., 2010a; LeBleu et al., 2013). Recent advances in fibrosis-EMT research has helped to shed light on some of these issues. The EMT Debate in Renal Fibrosis May Be Explained by Partial Activation of the Program Early insight into the potential benefit of abrogating EMT in fibrotic conditions came from the finding that mice lacking Smad3 were protected against tubular interstitial fibrosis (TIF) (Sato et al., 2003). Smad3 acts downstream of TGFb, one of the main EMT inducers, and indeed, Smad2/3 activation mediates renal fibrosis in response to TGFb and angiotensin II (Yang et al., 2010a). Furthermore, reactivation of Snail and EMT in renal epithelial cells can induce renal fibrosis and renal failure (Boutet et al., 2006). However, there is no evidence of the migratory potential of renal epithelial cells in vivo despite their accumulation of mesenchymal markers when observed in culture upon treatment with TGFb (Humphreys et al., 2010; LeBleu et al., 2013). New evidence may help to understand the mechanisms behind renal fibrosis and resolve this controversy (Grande 28 Cell 166, June 30, 2016 et al., 2015; Lovisa et al., 2015). It seems that, although renal epithelial cells do not directly generate myofibroblasts, deletion of Twist or Snail in renal epithelial cells significantly attenuates interstitial fibrosis in mouse models of TIF: unilateral ureteral obstruction (UUO), folic acid administration (FA), or nephrotoxic-serum-induced nephritis (NTN) (Grande et al., 2015; Lovisa et al., 2015). Yet, how can EMT-TFs be so relevant for the development of fibrosis if kidney epithelial cells are not the source of myofibroblasts? These two recent studies both suggest that renal epithelial cells undergo a partial EMT: one pointing to perturbations in transporter proteins and cell-cycle regulation (Lovisa et al., 2015) and the other showing that epithelial cells lose epithelial markers and apico-basal polarity, but they remain integrated in the tubules (Grande et al., 2015). Importantly, the injured epithelial cells relay signals to the interstitium to promote the differentiation of fibroblasts into myofibroblasts and the subsequent fibrogenesis. The cells that underwent a partial EMT also secrete exosomes and cytokines that help recruit (1) bone-marrow-derived mesenchymal cells that also differentiate into myofibroblasts to promote fibrogenesis and (2) macrophages that sustain inflammation (Borges et al., 2013; Grande et al., 2015; Lovisa et al., 2015). Interestingly, although Snail factors were considered purely as transcriptional repressors, Snail is converted into an activator upon its binding to Twist (Rembold et al., 2014) and also upon binding to CREB-binding protein (CBP). The latter allows Snail1 to directly activate the transcription of inflammatory cytokines genes in cancer cells and during fibrosis (Grande et al., 2015; Hsu et al., 2014; Lovisa et al., 2015). Thus, even though kidney tubular cells do not become collagen-producing fibroblasts, they undergo a partial EMT, and in a paracrine manner, they contribute significantly to fibrogenesis and inflammation, two hallmarks of fibrosis (Figure 3). Is There a Role for EMT in Liver and Lung Fibrosis? Evidence to clarify the role of EMT in liver fibrosis has been less forthcoming. Hepatocytes do not seem to undergo EMT in vivo, and they are not the source of collagen-producing cells in liver fibrosis (Taura et al., 2010). In fact, the transdifferentiation of hepatic stellate cells are reportedly responsible for 90% of the myofibroblast population in transgenic mice (Mederacke et al., 2013), with neither hepatocytes nor cholangiocytes (the epithelial cells of the bile duct) seeming to undergo EMT (Chu et al., 2011; Scholten et al., 2010). These findings contrast with earlier lineage-tracing studies showing that fibroblasts were derived from hepatocytes that underwent EMT (Zeisberg et al., 2007b). Others have reported that hepatic stellate cells secrete collagen I to trigger EMT in epithelial hepatoma cells (Yang et al., 2014), that fibrosis can be repressed by inhibiting TGFb signaling (Park et al., 2015), and that miRNAs can modulate EMT in cultured hepatocytes through TGFb inhibition (Brockhausen et al., 2015; Zhao et al., 2014; Zhao et al., 2015). Importantly, Snail deletion in hepatocytes attenuates liver fibrosis in adult transgenic mice (Rowe et al., 2011), also suggesting that hepatocyte-EMT may play a role in liver fibrosis. In the light of these studies and the new data garnered from renal fibrosis (Grande et al., 2015; Lovisa et al., 2015), it is tempting to speculate that EMT contributes to the development of liver fibrosis. In parallel with that observed in the kidney, Figure 3. EMT and Fibrosis Following one of unilateral ureteral obstruction (UUO), folic acid treatment (FA), or nephrotoxic serum-induced nephritis (NTN), renal epithelial cells undergo a partial EMT characterized by the loss of epithelial and specific differentiation markers and accompanied by the attenuation of their proliferative capacity. These dedifferentiated damaged renal epithelial cells persist in the tubules, relaying instructive signals to the interstitium to promote myofibroblast differentiation and the recruitment of immune cells, thereby promoting fibrogenesis and sustaining inflammation. All these effects are mainly due to the activation of EMT-TFs, including Snail and Twist, as their specific deletion in renal epithelial cells strongly attenuates the development of the hallmarks of fibrosis (adapted from Grande et al., 2015). damaged hepatocytes could secrete signals after undergoing a partial EMT, which promotes stellate cell transdifferentiation and enhances inflammation. This would be consistent with data obtained in mouse models of fibrosis in Snail-deficient hepatocytes (Rowe et al., 2011). This partial EMT hypothesis is also supported by findings in lung fibrosis, in which a subset of epithelial cells in patients with IPF expresses both epithelial and mesenchymal markers (Marmai et al., 2011). Therapeutic Strategies in Fibrosis As well as being a very potent EMT inducer, aberrant TGFb signaling is the main cause of fibrosis, and it provokes a massive accumulation of ECM. How or why TGFb signaling becomes constitutive is unknown (Marmai et al., 2011), but nevertheless, attenuating TGFb signaling has been the main strategy to fight fibrosis (as in colorectal cancer, in which TGFb inhibitors block the crosstalk between the stroma and cancer cells to halt disease progression) (Calon et al., 2015). We describe below different strategies proposed to inhibit the TGFb pathway as a proof of concept of the benefits of targeting EMT, albeit some targets may not fulfill the criteria for the optimal targeting in patients. STAT3 is a downstream mediator of TGFb signaling and induces EMT in cooperation with active K-Ras by increasing Snail expression in cancer cells (Saitoh et al., 2016). STAT3 is also reported to be elevated in lung biopsies from patients with IPF and in mice with fibrotic lungs. Inhibiting STAT3 with a synthetic smallmolecule inhibitor (C-188-9) attenuates fibrosis in bleomycin-treated mice (Pedroza et al., 2016). Agonists of another endogenous inhibitor of TGFb signaling, the orphan nuclear receptor 4A1 (NR4A1), can also inhibit fibrosis in skin, liver, lung, and kidney (Palumbo-Zerr et al., 2015), and this is consistent with the inhibition of TGFb-induced EMT and metastasis by the expression of NR4A1 itself (Zhou et al., 2014). Others have targeted TGFb signaling in mouse models by modulating the activity of BMP components, including administration of BMP7 (Zeisberg et al., 2003) and the use of small peptides to specifically inhibit the BMP-Alk3 receptor interaction (Sugimoto et al., 2012). The prevention or reversal of fibrosis requires long-lasting suppression and presumably, tissue remodeling. Epigenetic modifications can perpetuate fibroblast activation in renal fibrosis (Bechtel et al., 2010) and, given that some epigenetic modifications can be reversible, epigenetic modulators such as miRNAs are increasingly being targeted for the prevention or reversal of fibrosis, although there are still concerns about the off-target or toxic side effects associated with the delivery of miRNA inhibitors (Pottier et al., 2014). Examples of putative candidates are miR-21 and miR-181, linked with EMT-driven fibrosis in epicardial mesothelial cells (Brønnum et al., 2013), hepatic stellate cells, and hepatocytes (Zhao et al., 2014; Brockhausen et al., 2015). Other miRNAs have been proposed as beneficial against fibrosis, after showing that they inhibit EMT: miR-200b and miR34c inhibit EMT in TIF (Morizane et al., 2014; Oba et al., 2010; Tang et al., 2013); the miR-106b-25 cluster (miR-106b, miR-93, and miR-25) is markedly downregulated in TGFb1-induced Cell 166, June 30, 2016 29 EMT in renal HK2 cells (Tang et al., 2014); the loss of miR-101 promotes EMT and fibrosis in hepatocytes through TGFb activation (Zhao et al., 2015); and overexpression of let-7 miRNA in primary fibroblasts causes changes consistent with the inhibition of EMT (Huleihel et al., 2014). These miRNAs could be considered in anti-fibrotic therapeutic strategies that seek to restore their function. As such, the intravenous injection of synthetic RNA duplexes that mimic miR-29 can recover its function and even block pulmonary fibrosis in mice (Montgomery et al., 2014). Heat shock proteins (HSPs), molecular chaperones that assist in protein folding, are overexpressed in numerous cancers (Lianos et al., 2015), and they have been linked with fibrosis (Bellaye et al., 2014). HSP27 inhibition (OGX-427) destabilizes Snail and inhibits TGFb-induced EMT and fibrosis (Wettstein et al., 2013). In fact, blocking HSP27 with a synthetic tetrapeptide (Ac-SDKP) or with the OGX-427 antisense oligonucleotide (an anticancer agent in phase II clinical trials) can impair EMT (Deng et al., 2014, Wettstein et al., 2013). Fibrosis is associated with most cardiac diseases, leading to stiffness of the cardiac tissue and dysfunctional cardiac contraction (Krenning et al., 2010). Patient survival rates are reported to deteriorate as the proportion of fibroblasts rises (Gulati et al., 2013). In the heart, a subset of endothelial cells undergo TGFb-mediated EndMT, and these fibroblasts are associated with heart fibrosis (Zeisberg et al., 2007a). In addition, recent work shows that approximately 35% of cardiac fibroblasts undergo transdifferentiation into endothelial cells in a p53-dependent manner through a mesenchymal to endothelial transition (MEndoT), contributing to neovascularization after ischemia (Ubil et al., 2014). Again, p53 could be a beneficial agent, and reversion to the epithelial or endothelial phenotype should be the preferred strategy in anti-fibrotic therapies. In summary, the extensive exploration of EMT in fibrotic diseases has helped identify therapeutic targets that predominantly hinder TGFb signaling (Chen et al., 2013; Wang et al., 2013). However, inhibiting TGFb will also impede its beneficial effects (Mehal et al., 2011) and it may therefore be more desirable to target its principal downstream pro-fibrotic effectors (e.g., STAT3 or NR4A1), or alternatively, restore the expression of epithelial miRNAs. These potential therapeutic targets can ameliorate fibrosis by attenuating EMT and EndMT, offering an opportunity to reverse fibrosis via MET. As such, inactivating Snail1 in mice with established renal fibrosis improves organ morphology and function, significantly attenuating the disease (Grande et al., 2015). EMT in Cancer EMT is thought to be activated in cancer cells, linked to their dissociation from the primary tumor and their intravasation into blood vessels (Thiery et al., 2009). However, the impact of the EMT in cancer progression and patient survival is still far from fully understood. From skepticism (Tarin et al., 2005) to fundamental support (Lamouille et al., 2014; Thiery et al., 2009; Ye and Weinberg, 2015), more recent data suggest notes of caution on the true role of EMT in cancer progression (Fischer et al., 2015; Zheng et al., 2015). Nevertheless, studies in the last decade have fueled the interest in EMT in the cancer field, making it necessary to discuss recent advances regarding this complex biological process. 30 Cell 166, June 30, 2016 Despite the vast body of literature regarding the role of EMT in cancer, its applicability in cancer diagnosis and treatment has remained limited, in part due to the intrinsic heterogeneity of the actual tumor cells and tumor microenvironment when compared with the relative homogeneity of cell culture models. Thus, there are frequent debates as to whether EMT is truly relevant to cancer in vivo. The execution of EMT in cancer is not homogeneous, lending further support to the need to view EMT as a spectrum of intermediate states. Indeed, the description of the tumor invasive front as being functionally distinct from the main tumor bulk (Brabletz et al., 2001) points to the heterogeneous nature of the EMT program executed within the tumor microenvironment. The tumor invasive front, in which the EMT program is executed, has the hallmarks of a mesenchymal phenotype, with a weakened cell adhesion system. By contrast, the main tumor bulk remains largely epithelial (Figure 4). Thus, there is probably an EMT gradient from full, (some focal events at the tumor invasive front) to partial, to no (main tumor bulk) EMT (Huang et al., 2013). Furthermore, this gradient may be more or less steep in different tumors, reflecting the intrinsic molecular heterogeneity arising from diverse driver mutation profiles or altered signaling networks. For example, the circulating tumor cells (CTCs) in breast cancers with a triple-negative molecular subtype (ER–/PR–/HER2–) tend to have a more mesenchymal phenotype. Therefore, the overall biology of the tumor will likely dictate the frequency of identifying CTC subpopulations with a partial or full EMT. The tissue of origin may also influence the heterogeneity in the execution of EMT. A quantitative EMT scoring system was recently established based on gene expression profiles, ranking the EMT state from 1.0 to +1.0 (Tan et al., 2014). This system was used to highlight that the developmental lineage of each cancer type is defined by both a micro- and macro-EMT gradient, and each cancer type has a distinct propensity to exhibit diverse EMT states. Expectedly, cancers that arise from the mesoderm or NC lineages consistently show higher EMT scores in both in vivo tumors and cultured cell lines. In these cancers (e.g., osteosarcoma and malignant melanoma), the range of intrinsic EMT gradients is more restricted than in solid tumors of epithelial origin, such as ovarian, breast, and lung carcinomas. These different EMT gradients might help explain the inconclusive clinical significance of EMT found in the literature. Given the wide range of EMT states in ovarian, breast, and lung carcinomas, their clinical significance should be considered in conjunction with their intrinsic molecular profiles, such as the gene expression-based molecular subtype (GEMS) for ovarian and breast cancers and the EGFR mutational status or ALK fusion transcript status for lung cancer. In ovarian cancer, the EMT status is closely linked to the reported GEMS, with worse GEMS prognosis correlated with higher EMT scores (Tan et al., 2013). Furthermore, the EMT status consistently predicts overall- (OS) and disease-free survival (DFS) (Tan et al., 2014). Collectively, these findings offer a new perspective on the translational applicability of EMT in prognosis. In the absence of cell lineage analyses, the expression of epithelial and mesenchymal markers in primary tumors is a good indicator of a potential EMT. The existence of intermediate phenotypes should also allow cells in the stromal compartment Figure 4. The Metastatic Cascade This scheme aims to illustrate the complexity of the metastatic cascade from the point of view of cell plasticity, from carcinoma cell heterogeneity in the primary tumor to the different strategies that cancer cells adopt to survive in the bloodstream and colonize distant sites. EMT is a limited focal event in the primary tumor that could occur upon the interaction of carcinoma cells with the microenvironment, in which cancer-associated fibroblasts (CAFs) and immune cells can be found, including different types of natural killer T and B lymphocytes and inflammatory cells (not shown). The stroma also includes tumor-associated macrophages (TAMs) and other bone-marrow-derived cells. Some cells are possibly recruited prior to the establishment of primary and metastatic malignancies (Barcellos-Hoff et al., 2013). CAFs are like myofibroblasts (Lau et al., 2016) and TAMs are M2-type macrophages (Hsu et al., 2014), indicating that the microenvironment in the primary tumor is similar to that found in fibrosis, including an important inflammatory component. Other stromal cells also play a major role in tumor progression (Gentles et al., 2015). As expected from a focal event, the proportion of carcinoma cells with EMT features in primary breast tumors does not exceed 3% in estrogen receptor (ER)-positive tumors and 10% to 15% in ER-negative tumors. However, the majority of circulating tumor cells (CTCs) exhibits an intermediate EMT phenotype. Cells that undergo EMT in the primary tumor can also help epithelial cancer cells to invade. CTCs may be derived from carcinoma cells that undergo EMT in situ in the primary tumor or they may acquire intermediate EMT phenotypes in circulation, particularly when in clusters exposed to high concentrations of TGFb originating from associated platelets. Recent data suggest that microemboli may be able to extravasate and that they exhibit higher clonogenic potential (Aceto et al., 2014; Au et al., 2016). Cancer cells in metastatic outgrowths are clearly epithelial-like, and they can be identified morphologically and molecularly as having been derived from the primary tumor, again highlighting the plasticity of tumor cells. This is also consistent with data showing that mesenchymal cells must revert to the epithelial phenotype to colonize their new destination. The plasticity of carcinoma cells is therefore of upmost importance to escape death during the different stages of tumor progression. Mesenchymal-like cells are better adapted to cell deformation, resisting shear stress and being more resistant to drugs. Tumor dissemination in the lymph nodes is much less documented, albeit it is a major aspect of tumor staging (not shown). Histopathological examination of proximal (sentinel) lymph nodes show either dispersed carcinoma single cells, micrometastases (<2 mm), or massive invasion. The latter is compatible with findings indicating that lymph nodes can be colonized after collective cell migration (Giampieri et al., 2009). (which maintain the epithelial markers of their tissue of origin) to be identified. The distribution of a number of epithelial and mesenchymal markers was first established in a breast cancer tissue microarray (Sarrió et al., 2008). More recently, mesenchymal markers were found in 12% of the basal molecular subtypes of breast carcinoma. This percentage is likely to be higher in the claudin-low molecular subtype. Furthermore, cells expressing mesenchymal markers are also likely to exist among epithelial carcinoma cells in ErbB2 and luminal molecular subtypes; albeit, to a lesser extent (Tan et al., 2014; Yu et al., 2013). It is also interesting to note that, in cancer, as in embryonic development, EMT may be a focal rather than a global event, probably a response of cancer cells to their local microenvironment (Figure 4). Macrophages participate in EMT induction within primary tumors in transgenic mouse models, orthotopic xenografts (Wyckoff et al., 2004), and primary human breast carcinomas, interacting with carcinoma cells adjacent to the endothelial cells (Robinson et al., 2009). The latter suggests that macrophages may be also critical for local intravasation. M2a macrophages migrate toward carcinoma aggregates to induce the dissociation and migration of single carcinoma cells through the local production of TGFb and direct ICAM-1-b2 integrinmediated cell-cell contact. Endothelial cells also contribute to their dissociation through diffusible scatter factors, such as HGF (Bai et al., 2015a; Bai et al., 2015b). Nevertheless, it is crucial to consider the nature of CTCs as indicators of the next step in the metastatic cascade. EMT and Circulating Tumor Cells Most CTCs express both epithelial and mesenchymal markers (Khoo et al., 2015; Yu et al., 2013), suggesting that an active EMT process is likely to operate during the dissemination of carcinoma cells (Figure 4) (Thiery and Lim, 2013). In addition, after transient amplification of breast cancer CTCs, the cells exhibit a full range of EMT phenotypes (Khoo et al., 2015), with evidence indicating that CTCs express ECM components potentially involved in the successful colonization of distant organs (Ting et al., 2014). A landmark study revealed that CTCs in blood from metastatic breast cancer patients can serve to evaluate prognosis and response to treatment (Cristofanilli et al., 2004), and this was recently confirmed when CTCs quantified in metastatic patients were shown to predict OS and progression-free survival (Bidard et al., 2014). However, a recent clinical trial could not identify an effective second-line chemotherapy strategy for patients whose CTC numbers were not diminished by first-line chemotherapy (Smerage et al., 2014). Clearly, there is an urgent need to understand more about the molecular attributes of CTCs, particularly how their presence and EMT status reflect treatment response Cell 166, June 30, 2016 31 and malignant potential. Several studies have addressed the value of CTC monitoring and phenotyping during neoadjuvant and adjuvant therapies (Alix-Panabières and Pantel, 2014; Khoo et al., 2015) and their use in drug sensitivity testing (Yu et al., 2013). Importantly, cancer therapy affects CTC number and phenotype, with refractory patients having more mesenchymal-like CTCs and patients who are responding to treatment demonstrating significantly fewer CTCs with a more epitheliallike phenotype (Yu et al., 2013). These findings are in agreement with recent studies highlighting the importance of EMT in conferring chemoresistance in breast and pancreatic cancer models (Fischer et al., 2015; Zheng et al., 2015). Intriguingly, the EMT program may be engaged at premalignant stages, as shown in an experimental model of pancreatic cancer, especially at sites of inflammation (Rhim et al., 2012). This suggests that EMT may be a very early event in tumorigenesis and is consistent with the presence of disseminated tumor cells (DTCs) at very early tumor stages in breast cancer patients (Bidard et al., 2008; Hüsemann et al., 2008). CTCs are mainly identified with the EpCAM antibody, which may exclude cells that have undergone a more complete EMT and have lost this epithelial marker. Thus, lineage tracing experiments in animal models are required to identify CTCs in different states, and this may also offer a way to explore whether CTCs only undergo transitions after reaching the metastatic site or if MET can also occur in the bloodstream. Interestingly, circulating tumor clusters are more effective in colonizing secondary sites than single mesenchymal CTCs (Aceto et al., 2014). These clusters comprise distinct clonal carcinoma cells which are held together through plakoglobin-mediated intercellular adhesion. Again, this highlights the requirement for mesenchymal cancer cells to at least partially reverse to the epithelial state for metastatic growth (Nieto, 2013). CTC clusters formed through platelet interactions are also correlated with a worse prognosis, and mesenchymal CTCs may interact with epithelial-like carcinoma cells and drag them along to secondary sites (Figure 4). It was unclear how such aggregates can reach secondary sites through the 5–10-mm lumens of blood capillaries in the lung, but recent data show that clusters of %20 cells can traverse these constrictions (Au et al., 2016). Extravasation is a very efficient process and even untransformed epithelial mammary cells can reach the lung after being injected into the bloodstream (Podsypanina et al., 2008). If epithelial cells, and cell clusters, can efficiently disseminate and extravasate, the EMT/MET pathway may not be necessary in all cells for secondary tumor formation. This directly impinges on the requirement for EMT in tumor progression. Is EMT an Essential Element of the Metastatic Cascade? Most cancer EMT studies, including mechanistic analyses, are conducted on cultured cells, and the vast majority of in vivo studies are carried out after injection or xenografting of parental or manipulated cancer cells, which are likely to show distinct EMT features. Some mouse breast cancer reporter models have been used to follow the endogenous activation of Snail TFs (Tran et al., 2014; Ye et al., 2015), showing that Snail1 activation and endogenous EMT occurs in the primary tumor, and that, although important for invasion and the formation of CTCs, EMT is not required for metastatic colonization. These 32 Cell 166, June 30, 2016 data are consistent with the lack of metastatic outgrowth from cells with constitutive mesenchymal traits, and it highlights the transient nature of EMT and the cell’s need to revert to a more epithelial phenotype for colonization (Ocaña et al., 2012; Tsai et al., 2012). To better understand the contribution of EMT to metastatic colonization, Fsp-1 or vimentin GFP-reporter lines were used to monitor the expression of mesenchymal markers in a MMTV-PyMT breast cancer model (Fischer et al., 2015). While this provided evidence of EMT in a small proportion of cells in the primary tumor, these GFP-positive cells—even though proportionally enriched in the CTC population—did not seem to contribute significantly to the formation of metastatic outgrowths. Others show that pancreas-specific loss of either Snail or Twist does not block systemic dissemination or the formation of metastasis in pancreatic ductal adenoma carcinoma (PDAC) (Zheng et al., 2015). Together, these findings prompted the authors to suggest that EMT may be dispensable for carcinoma progression to a metastatic state. Although these data are provocative, EMT cannot yet be ruled out as a driver of cancer progression (Maheswaran and Haber, 2015). First, the cooperation between the multiple EMT-TFs suggests that loss of individual factors may be insufficient to prevent cell invasion and dissemination, particularly in PDAC, in which even normal cells do not exhibit a strong epithelial phenotype. Furthermore, Fsp-1 is not a universal marker of EMT and it would not be activated in the primary tumors of Neu and PyMT breast cancer models (Trimboli et al., 2008), albeit EMT has been described in both (Ye et al., 2015). Second, the few cells that underwent EMT in the primary tumor (Fischer et al., 2015) are consistent with the numbers observed in other studies (Tan et al., 2014). Third, the relative enrichment of mesenchymal markers in CTCs confirms their enhanced ability to invade and intravasate. Fourth, the comparatively low contribution of the vimentin-expressing cells to metastases calls for an analysis of other markers that may better assess intermediate EMT phenotypes and for the evaluation of a possible cooperation between mesenchymal and epithelial cells (Tsuji et al., 2009). Indeed, EMT may facilitate the invasion and intravasation of other cells that retain their epithelial character (Figure 4). As such, even if EMT were more limited than anticipated, it would still be required for tumor progression to the metastatic state (Li and Kang, 2016). A recent study using intravital microscopy in mice has detected a pool of breast tumor cells that spontaneously undergo EMT, become motile, and disseminate and then reverse to the epithelial state upon metastatic outgrowth (Beerling et al., 2016). More studies like these are needed in other types of cancer in order to assess the contribution of EMT as a crucial requirement for the progression of primary tumors toward the metastatic state, as consistently suggested in different models and patient datasets. EMT and Cancer Stem Cells The induction of both EMT and stemness at the invasive fronts of tumors was first suggested to explain how CSCs disseminate and seed fully heterogeneous tumors at secondary sites (Brabletz et al., 2005). Indeed, the EMT-derived stem cell phenotype in the mammary gland—capable of forming mammospheres following EMT induction and expressing a CD44highCD24low profile (Mani et al., 2008; Morel et al., 2008)—was reminiscent of that of CSCs. This profile has since been identified in numerous CSC subpopulations (Gupta et al., 2009a), with other groups reporting high intracellular aldehyde dehydrogenase 1 (ALDH1) activity as a marker of stemness (Ginestier et al., 2007). Recent indications submit that CD44high and ALDH1 activity do not generally coexist in CSCs (Liu et al., 2013a), suggesting the existence of different CSC types. The gene signature identified for the induction of pluripotency is also relevant to the persistence of CSCs, with Oct4 and Nanog ectopically inducing CSC-like properties and enhanced features of EMT (Kong et al., 2010). This is similar to other genes involved in maintaining self-renewal capacity, including b-catenin, Bmi1, Gli1, and Pou5f1. Within the CD44highCD24low population of triple-negative breast cancers, there are discrete epithelial and mesenchymal cell types that are differentiated by their expression of a6-integrin splice variants—a6A (epithelial) and a6B (mesenchymal)—with the a6Bb1-integrin noted as essential for CSC function, mammosphere formation, and anchorage-independent growth (Goel et al., 2014). Both a6-integrin and CD44 are alternatively spliced by ESRPs, which as already discussed, protect the epithelial phenotype and, hence, promote the presence of epithelial a6 and CD44 isoforms in these cells (Goel et al., 2014; Warzecha et al., 2009). Accordingly, it would be interesting to correlate the combinations of integrins and CD44 isoforms with ALDH1 activity in primary tumors and their metastatic outgrowths. EMT involves the expression of one or several classical EMTTFs, and since the expression of these TFs often accompanies features of stemness, the retention of this latter state has been linked to EMT. ZEB1 has been specifically detected in poorly differentiated pancreatic carcinomas, and it is thought to repress the expression of stemness-inhibiting miRNAs to maintain a stem-like phenotype (Wellner et al., 2009). CSC expansion is also thought to occur after Snail1 mediates a shift from asymmetric (one stem cell, one differentiating cell) to symmetric (two stem cells) cell division, implying a role for EMT in increasing the stem cell pool within the tumor (Hwang et al., 2014). Similarly, other inducers of EMT, such as the homeobox protein Six1 or ladybird homeobox-1 protein (LBX1), also induce stemness and mammary spheroidogenesis when ectopically expressed in the mouse mammary gland, presumably through their effects on one of their transcriptional targets (ZEB1, ZEB2, SNAIL1, and TGFb2) (McCoy et al., 2009; Yu et al., 2009). Nonetheless, it is noteworthy that the EMT program implicated in normal stem cells and CSCs may differ, as recently shown in the murine mammary gland, in which they involve Snail2 and Snail1, respectively (Ye et al., 2015). Much of the plasticity associated with CSCs arises from the local niche, both at the primary and secondary tumor sites (Nieto, 2013), and the presumed CSC reservoir is thought to be responsible for tumor relapse and poor clinical outcome. TGFb induces a program in stromal cells that correlates with poor prognosis in colorectal cancer patients, in which cancer-associated fibroblasts (CAFs) have been linked with an increase in tumor-initiating cells (Calon et al., 2015). Similarly, in hepatocellular carcinoma that usually develops in the context of fibrosis, CAFs with characteristics of myofibroblasts promote the appearance of liver-tumor-initiating cells (Lau et al., 2016). Ongoing molecular rearrangements were recently proposed to drive adaptive plasticity in highly metastatic neuroblastoma cells, cells that strongly express EMT genes and that display a stem-like phenotype, with high SOX2 and NANOG expression (Pandian et al., 2015). Collectively, these studies highlight the regulatory effects of the microenvironment on cell fate. Uncoupling EMT and Stemness Despite the evidence linking stemness and the EMT, recent studies consider that these two aspects of CSC development may occur in parallel rather than through the activation of the same pathway. It is now widely accepted that, although EMT is instigated in CSCs to drive their dissemination, they undergo MET on reaching a desirable metastatic niche to seed a new tumor, presumably by suppressing driver EMT-TFs (Nieto, 2013). The MET at the new metastatic site is surely influenced by CAFs and a specific type of metastasis-associated macrophages (MAMs) (Qian et al., 2015), as well as mesenchymal stem cells, immunological mediators, and other factors involved in hypoxia and angiogenesis (Plaks et al., 2015). As such, recent data indicate that MAMs promote the conversion of hepatic stellate cells into myofibroblasts, resulting in a fibrotic microenvironment that sustains metastatic tumor growth in the liver (Nielsen et al., 2016). Mesenchymal DTCs activate CAFs, which in turn, promote the transition of DTCs to a more epithelial and proliferative phenotype that favors metastatic colonization (Del Pozo Martin et al., 2015). Parallels can again be seen between the induction of tumorinitiating capacities in cancer cells and the acquisition of pluripotency during fibroblast reprogramming to induced pluripotent stem cells. The latter also requires a MET (Li et al., 2010b; Samavarchi-Tehrani et al., 2010), compatible with the fact that embryonic stem cells are epithelial-like. Recent data indicate that the serial delivery of four transcription factors in a specific order— Oct-4 with Klf4, followed by c-Myc, followed by Sox2 (Liu et al., 2013b)—induces a higher rate of pluripotency. Oct4 first promotes fibroblasts with an even more mesenchymal-like phenotype through the upregulation of Snail (Oct4), such that these cells are then more responsive to Sox2, which is needed to repress Snail and induce MET in embryos and mouse fibroblasts (Acloque et al., 2011; Li et al., 2010b). This delay in MET is thought to provide time for epigenetic reorganization before pluripotency is induced (Gaeta et al., 2013). Interestingly, cells engaged in intermediate reprogramming closely resemble those found during mesendoderm formation, which is driven by a MET-like process in mesenchymal mesodermal cells in the primitive steak (Takahashi et al., 2014). Again this highlights the similarities with embryonic development and the importance of intermediate phenotypes in the induction of EMT. A prototypical example of this is the loss of an epigenetic mark in trophoblast epithelial cells that allows them to undergo a partial EMT while maintaining their self-renewal capacity and multipotency (Abell et al., 2011). Further evidence for uncoupling EMT and stemness derives from studies of the homeobox transcription factor, PRRX1, a potent EMT inducer in both embryos and cancer cells. In human breast cancer BT-549 cells, downregulation of PRRX1 is linked with MET and increased proliferation, but also with a gain in mammosphere formation and the emergence of a predominant CD44high cell population (Ocaña et al., 2012), all favoring Cell 166, June 30, 2016 33 metastatic colonization. This again suggests that EMT is not necessarily associated with stemness. Interestingly, CTGF and the inhibitor of differentiation (Id) both induce MET and the expression of pluripotency genes (Chang et al., 2013; Stankic et al., 2013), which suggest that the morphological phenotype and stemness can be regulated independently. Indeed, TGFb is essential for the initial EMT, after which the EGFR-Ras pathway independently drives stem cell properties (Voon et al., 2013). This is also consistent with Snail- or Twist-induced expression of miR-424, which promotes mesenchymal traits while inhibiting tumor initiation (Drasin et al., 2015). In a more extreme case, the mutual exclusivity of EMT and stem cell-like traits is evident in prostate and bladder cancer cell lines, in which constitutive Snail1 activation suppresses stemness (Celià-Terrassa et al., 2012). Furthermore, in the skin, low Twist levels induce tumor initiation and confer stem cell properties while higher levels are required for the induction of EMT (Beck et al., 2015). Consistently, Twist-induced tumor-initiating capacity is only manifested after Twist downregulation at the metastatic site (Schmidt et al., 2015), and decreasing Snail1, Twist1, or Prrx1 expression is a requisite for EMT attenuation to allow CSCs to populate a new metastatic site (Ocaña et al., 2012; Tran et al., 2014; Tsai et al., 2012). Can Partial EMT States Explain the Stemness-EMT-MET Link? The aforementioned data show that EMT and stemness are not necessarily linked and that the acquisition of both traits can be uncoupled. However, they also suggest that an intermediate EMT state may make cells more prone to exhibiting CSC-like properties. Transient Twist1 activation in mammary epithelial cells induces their long-term invasive potential, as well as their capacity to colonize distant organs and tendency to express traits typical of both epithelial and mesenchymal cells. By contrast, constitutive Twist1 activation induces a migratory but not a metastatic phenotype, as cells do not demonstrate tumor-initiating capacities (Schmidt et al., 2015). This is consistent with the finding that a complete EMT is not required for Twist1induced invasion (Shamir et al., 2014) and that plasticity at the metastatic site does not imply reversion to the fully differentiated epithelial phenotype (Nieto, 2013). In summary, stemness seems to be better manifested at intermediate epitheloid states, consistent with those observed in cultured mammospheres (Mani et al., 2008; Ocaña et al., 2012) and in embryonic stem cells. The intermediate EMT state associated with the stem cell properties of trophoblast cells (Abell et al., 2011) is similar to that observed in claudin-low breast cancer cells (Jordan et al., 2011; Prat et al., 2010). This is also the basis of the proposed existence of ‘‘migratory CSCs’’ at the invasive front of tumors (Brabletz et al., 2005). Indeed, transitions between EMT and MET are likely to be controlled by many factors that sway the balance from one extreme to the other. Thus, a tissue could display more mesenchymal or epithelial characteristics at any given time, yet still with evidence of the other. Indeed, pancreatic cancer cells are shown to have stem-like features and an intermediate EMT phenotype in vivo, with low levels of E-cadherin and concomitant mesenchymal features (Rhim et al., 2012). Along similar lines, the upregulation of miR-424 during Twist1- or Snai1-induced EMT in breast cancer cells can induce mesen34 Cell 166, June 30, 2016 chymal genes without affecting the expression of the epithelial genes (Drasin et al., 2015). Hybrid, partial transitions in cells lead to an expression of both epithelial and mesenchymal traits, which can potentially endow cells with migratory potential while retaining some degree of cell-cell adhesion, promoting a more efficient coordinated migration, as observed during embryonic development (Theveneau and Mayor, 2011). Stemness is regulated by a LIN28/ let-7 inhibitory loop (Yang et al., 2010b), with a balanced, intermediate level of the two helping to maintain an intermediate level of Oct4 for stemness (Niwa et al., 2000). Building on this, a ‘‘stemness window’’ that regulates the interplay between LIN28 and Let-7 has been proposed to control stemness and EMT (Jolly et al., 2015). Through the use of theoretical framework modeling, it was shown that strong inhibition of LIN28 by miR200 pushes the stemness window toward the mesenchymal end of the EMT axis, whereas strong inhibition of ZEB by Let-7 forces it in the other direction. From this, it was suggested that stemness can be gained by cells with different phenotypes based on the actions of Let-7 and Lin28 but also in response to changes in other signals like Snail1 and, presumably, Twist or other prominent inducers of EMT. A sub-population of ‘‘metastasis stem cells’’ was recently described among the CTC population (Baccelli et al., 2013), although it remains unclear whether CTCs harbor a small fraction of CSCs or if CTCs derive from cells with the potential to become CSCs, themselves a separate entity. Resistance to treatment may be acquired at metastatic sites in micrometastatic stages. Therefore, it will be necessary to determine whether stemness is a property of subsets of CTCs or whether EMT is solely a mechanism critical for dissemination, with stemness re-acquired at the site of arrest in the parenchyma of the target organ through MET (Beck et al., 2015; Brabletz, 2012; Jung and Yang, 2015; Polyak and Weinberg, 2009). Stemness, or the attainment of stem cell properties, is no longer considered a ‘‘fixed’’ trait, with such cells displaying a high degree of plasticity (Jolly et al., 2015). Stemness is well defined in embryonic development and it is a hallmark of cancer within a particular window during disease progression. Like other hallmarks of cancer, stemness can be lost or acquired, and it may be affected by the expression of certain phenotypic traits as cancer cells and tumors progress toward the metastatic disease (Antoniou et al., 2013). The Spectrum of EMT Plasticity in Cancer: Therapeutic Challenges Re-differentiation represents an attractive therapeutic option to reverse EMT (Beug, 2009; Huang et al., 2012; Nieto, 2013; Thiery and Sleeman, 2006), a feature that provides a unique way to tap into current mainstream therapies (Chua et al., 2011). However, there are several concerns that must first be considered. Given the existence of EMT gradients, we have yet to determine how far the cells should be reversed. Furthermore, it remains unclear what specific EMT score would be effective to abolish metastasis and/or enhance drug sensitivity. Evidence indicates that a full EMT reversal might not be optimal for metastatic colonization, and thus, the effective shift is likely to be context dependent and based on the different patterns of EMT in distinct cancer types. Indeed, the re-expression of E-cadherin in fibroblasts does not reverse them to the epithelial state (Navarro et al., 1993), and while restoring E-cadherin to ovarian cancer cells in culture abolishes their invasive properties, their mesenchymal traits remain unaffected (Huang et al., 2013). The drug discovery platform for EMT reversal is limited, although it is feasible to conceive drug pipelines that preferentially induce cytotoxicity in cells that underwent EMT. Lead compounds to target CSCs in breast cancers have been identified in screening programs based on EMT models, such as human mammary epithelial cells bearing shRNA against E-cadherin (Carmody et al., 2012; Gupta et al., 2009b). There are also pipelines for anti-cancer treatments that target specific components of signaling pathways, including HGFR, insulin-like growth factor-1 receptor (IGF-1R), EGFR, PDGFR, TGFbR, phosphatidylinositol 3-kinase (PI3K)/Akt, and ERK/MAPK (Lamouille et al., 2014). While these treatments could be considered as antiEMT drugs, these pipelines were not designed with a specific anti-EMT concept in mind, and therefore, the anti-proliferative and EMT-specific effects of these inhibitors might be difficult to assess independently. Few platforms have been developed that screen for non-cytotoxic agents that reverse EMT. A screening model based on a growth-factor-induced scatter assay (including HGF, IGF-1, and EGF) in NBT-2 rat bladder carcinoma cells has been adapted for this purpose (Chua et al., 2012). Its use highlighted compounds that target activin A receptor type II-like kinase (ALK5)/TGFbR1, MAPK, Src, and PI3K, none of which exhibited significant effects against proliferation. However, this model requires additional validation in a secondary screen on human carcinomas. Recent data indicate that activation of the PKA pathway induces epigenetic changes that can modulate EMT together with the loss of tumor-initiating capacities in a xenograft breast cancer model (Pattabiraman et al., 2016). TGFb inhibitors are the most intensively investigated antiEMT compounds, and recent Phase I studies tested the use of LY2157299 as a TGFb inhibitor in glioblastoma and hepatocellular carcinoma (Giannelli et al., 2014; Rodon et al., 2015). However, the effect of LY2157299 on EMT reversal warrants further study. It is worth noting here that TGFb can also act as a tumor suppressor through the activation of a particular type of lethal EMT in pancreatic cancer cells (David et al., 2016). Src kinase inhibitors also clearly reverse EMT (Huang et al., 2013; Vultur et al., 2008), yet their clinical effects as a monotherapy or in combination with other agents have been disappointing (Kim et al., 2009; Puls et al., 2011). An inhibitor of the closely related focal adhesion kinase (FAK), PF-00562271, is also being tested in a Phase I dose-escalation trial against solid tumors with promising results (Infante et al., 2012). These Src and FAK inhibitors must undergo lethality screens to determine the appropriate combinations of the compounds to induce EMT reversal. These drug screening platforms suffer from the same lack of microenvironmental modeling as cell culture models. However, microfluidic co-culture systems have recently been developed to mimic the tumor microenvironment and induce the formation of 3-dimensional tumor spheroids in order to screen for EMTblocking agents (Aref et al., 2013). This type of model more accurately assesses the efficacy of EMT reversal agents to combat the appearance of sprouting endothelial cells and/or immune cells. Finding a way to eradicate these epithelial cells would identify a mechanism through which the lethality of the cells can be enhanced following EMT reversal. One example would be the use of epigenetic modifiers, such as histone deacetylase (HDAC) inhibitors, to reverse ZEB1-regulated EMT and sensitize pancreatic cancers to gemcitabine (Meidhof et al., 2015). This strategy could be applied in the front-line or in chemoresistance settings. Finally, anti-EMT therapeutics should be carefully defined and positioned. Should we reverse or kill the cells that underwent EMT? Are we comfortable with reversing EMT and maintaining a stable disease status with these cells regaining an epitheliallike phenotype since they are less migratory and invasive? An important note of caution here comes from the evidence that reversal is also required for metastatic colonization in vivo (Alderton, 2013; Brabletz, 2012; Ocaña et al., 2012; Tsai et al., 2012; Beerling et al., 2016)). Therefore, promoting EMT reversal includes the potential risk of enhancing the establishment of secondary metastasis by DTCs (Nieto, 2013). Thus, it will be imperative to either define a proper therapeutic window in which to administer an EMT-reversing agent, or alternatively, the goal should be to eradicate the cells that underwent EMT. If effective, specifically eliminating those cells appears to be the optimal strategy. Even in the absence of metastatic disease, most locally advanced patients still relapse, and it is known that recurrence is accompanied by EMT, with Snail1 proving sufficient to promote mammary tumor recurrence in vivo (Moody et al., 2005). Interestingly, specifically targeting cells that underwent EMT might also target dormant cells in distant organs, as they are likely to be in a mesenchymal state. As such, NR2F1 has recently been considered a critical node in the induction of dormancy in DTCs (Sosa et al., 2015). Interestingly, NR2F1 is known to lie at the core of the EMT regulatory chromatin landscape in the neural crest (RadaIglesias et al., 2012). EMT Confers Resistance to Chemotherapy and Immunotherapy EMT confers resistance to cell death induced through various means both in embryos and in cancer cells (Vega et al., 2004; Thiery et al., 2009), including chemotherapy (Singh and Settleman, 2010). Even in studies that find a limited contribution of cells that have undergone EMT to established metastases, the role of EMT in conferring chemoresistance is clear (Fischer et al., 2015; Zheng et al., 2015). Therapeutic strategies should therefore encompass an arm that targets network rewiring when resistance emerges. For instance, a rewiring event occurs in EGFR-mutated non-small-cell lung carcinoma (NSCLC) that switches EGFR to Axl receptor tyrosine kinase in association with a mesenchymal phenotype (Byers et al., 2013). One can therefore propose targeting these erlotinib-resistant mesenchymal NSCLC cells with potent inhibitors of Axl (Sheridan, 2013), although it might also be envisaged that resistance to Axl will be acquired when network rewiring repeats. Nevertheless, there is increased interest in targeting Axl signaling (Gjerdrum et al., 2010) as a first-in-line class of therapeutic agent for EMT (clinicaltrials.gov; NCT02488408). Immunotherapy for cancer treatment has generated much interest in recent years, with notable successes against several Cell 166, June 30, 2016 35 tumors, including melanoma, NSCLCs, and renal carcinoma. These approaches are based on the blockade of immune inhibitory checkpoints, such as the use of monoclonal antibodies to programmed death-1 (PD1) or to cytotoxic T lymphocyte-associated antigen 4 (CTL4), two inhibitory T cell receptors (Okazaki et al., 2013). Further improvement in these therapeutic interventions would permit longer survival than that achieved using targeted therapeutics. Numerous somatic mutations—some of which are considered to be trunk and branch drivers—have been described in many tumors. However, new, dominating mutations are often induced by conventional and targeted therapeutics, and they have the potential to make cells refractory to these treatments (McGranahan and Swanton, 2015). In principle, although such rapid drift in mutation profiles may not affect CTL cancer cell lysis, EMT of the target cancer cell may potentially abrogate CTL lysis. Melanoma cells that undergo Snail-mediated EMT are much more metastatic than their parental cells, which was attributed to the emergence of T regulatory (Treg) CD4 cells expressing Foxp3, a known inducer of immunosuppression in Treg cells (Kudo-Saito et al., 2009). The emergence of these Treg cells was in part driven by the TGFb and thrombospondin secreted by these Snail1-expressing melanoma cells. Furthermore, a Snail small interfering RNA (siRNA) injected in vivo reduced the immunosuppressive and metastatic potential of melanoma cells (Kudo-Saito et al., 2009). Plasticity of carcinoma cell phenotypes can be achieved by conventional and targeted therapies and through the tumor microenvironment (Hölzel et al., 2013). Thus, as well as abrogating immune suppression in CTLs, one might consider modifying the EMT phenotype of carcinoma cells. Indeed, CTL lysis of carcinoma cells was considerably reduced when Snail-mediated EMT of MCF7 cells occurred in response to chronic exposure to TNFa due to the induction of autophagy (Akalay et al., 2013). EMT of the target cells affects the maturation of the immune synapse, although defining the mechanism that drives this plasticity in the cell cortex requires further study. It must also be ascertained what causes the defects in the organization and positioning of T cells, and of the adhesive receptors arranged as concentric rings in the immune synapse (Chouaib et al., 2014). Intriguingly, tumors that respond best to CTL-A4, and more recently to PD1 and PD1L antibodies, exhibit a higher EMT score (Tan et al., 2014). These tumors include melanomas and renal, bladder, and NSCL carcinomas (Topalian et al., 2012; Motzer et al., 2015; Rizvi et al., 2015). Thus, these antibodies could specifically restore a functional immune synapse in carcinomas that exhibit a mesenchymal- rather than an epithelial-like phenotype (Engl et al., 2015). This proposal is supported by recent data indicating that an increase in target cell tension potentiates killing by CTLs (Basu et al., 2016). In addition to immunotherapy, a vaccine was recently developed against a mesenchymal-associated T-box TF, brachyury (Hamilton et al., 2013), which is currently undergoing clinical trials for the treatment of solid tumors. In general, these fast-developing fields in immunotherapy and cancer vaccine development will most likely provide additional therapeutic options and insights into how to target and eradicate the population of cells that have undergone EMT in cancer. 36 Cell 166, June 30, 2016 Challenges, Future Directions, and Perspectives Originally described as a major event in embryonic morphogenesis, EMT is successfully recapitulated in various normal and abnormal processes, a phenomenon that has become well accepted since it was first proposed in the early 2000s. Yet, despite extensive forays into the role of EMT in disease progression, we are still far from reaching a consensus as to the true role EMT plays in these processes. While is seems clear that reversing EMT is beneficial for the treatment of fibrosis, the inherent complexity of cancer suggests that a better strategy may be to eradicate the cells that underwent EMT in the carcinoma. However, this still requires the development of better animal models together with technological advances to better visualize the EMT in vivo at the organ, tissue, cellular, and subcellular levels including advanced genomics in FACS-enriched cell populations. These approaches will directly test some of the concepts discussed here and help comprehend the dynamics of EMT to better design anticancer therapies. An emerging challenge will be to decipher inter- and intratumoral heterogeneity, including the full spectrum of intermediate EMT states and how cells make the transitions between these states. Targeted therapeutics face the formidable challenge of rapid tumor evolution, as revealed by the genomic landscape of distinct biopsies in primary and metastatic tumors (McGranahan and Swanton, 2015). It is also important to evaluate to what extent stochastic events could generate cells with a stem-cell-like phenotype and how the cell of origin influences the susceptibility to undergo EMT. Another important issue is to better understand the mechanisms for lymphatic dissemination in different cancer types. Although the EMT does not seem to be required for lymph node colonization (Giampieri et al., 2009), its relative contribution versus collective cell migration deserves further investigation. Bioinformatics algorithms must be developed to go far beyond gene ontology, principal component analyses, and the currently available networks. Most importantly, the predictions from mathematical models and hypothetical gene regulatory networks need to be tested in appropriate animal models. Furthermore, powerful bioinformatics analysis can also help in the new avenues that link EMT with metabolic changes (Choudhary et al., 2016; Jiang et al., 2015; Mathow et al., 2015; Shaul et al., 2014). Intravasation has been the least studied process in the metastatic cascade, and in light of recent data that suggests that EMT may be more limited than previously anticipated (Fischer et al., 2015; Zheng et al., 2015), attention should be paid to the possibility of cancer epithelial cells being able to efficiently intravasate. Do just a few mesenchymal cells help epithelial cancer cells intravasate? Is the leakiness of the tumor vessels sufficient to allow intravasation of epithelial cell clusters? As intravasation is clearly a crucial step in dissemination, further attention should be now paid to this process, particularly probing the recent proposal that TAMs and vascular mimicry acquired by cancinoma cells help them to intravasate (Harney et al., 2015; Wagenblast et al., 2015). It will also be relevant to define at what point in the lifespan of a CTC the cell undergoes a MET-like process: before it reaches the metastatic site (i.e., in the bloodstream) or after the metastasis has started to grow at a new site. Answering this question will surely be aided by identifying better markers for CTCs, particularly as EpCAM fails to detect all CTC states. It is also important to ask whether CSCs travel with CTCs in collective migration, the extent to which CSCs convert into non-CSCs (and vice versa) (Chaffer et al., 2011; Schwitalla et al., 2013), and how is such plasticity controlled. Cancer stroma crosstalk appears to be context specific, and thus, we must better understand the nature of the stroma that respectively promotes or represses EMT in the primary tumor or in the metastatic niche. In relation to this, the description of a pre-metastatic niche has been an important development (Kaplan et al., 2006), as has the role of membrane vesicles derived from tumor cells, called exosomes, which can induce vascular leakiness and bone marrow progenitor mobilization to ‘‘educate’’ the niche (Peinado et al., 2012). Exosomes derived from pancreatic ductal carcinoma can initiate the pre-metastatic niche in the liver (Costa-Silva et al., 2015). The target cells are the Kupffer cells, which instruct stellate cells to transdifferentiate into myofibroblasts and generate a fibrotic microenvironment that favors the formation of the metastatic niche. This is reminiscent of the situation in renal fibrosis, where damaged epithelial cells secrete exosomes that initiate the differentiation of fibroblasts into myofibroblasts (Borges et al., 2013). Importantly, the homing specificity of exosomes directed by the expression of different integrins defines the target tissue where cancer cells will preferentially form metastatic outgrowths (Hoshino et al., 2015). Thus, a deeper analysis of the metastatic niche and its interaction with exosomes, DTCs, and dormant cancer cells is now warranted. Most intriguing is how we can target the EMT program for therapeutic gain. Specific deletion of the cells that have undergone EMT would be optimal to prevent the colonization of distant sites by newly generated DTCs or long-standing dormant cells. This implies specifically rendering those cells sensitive to cytotoxic drugs, which is a challenge if we consider that they are resistant to multiple death-inducing signals and also to chemo- and immunotherapy (Fischer et al., 2015; Kudo-Saito et al., 2009; Voon et al., 2013; Zheng et al., 2015). Many of these challenges will require innovative approaches. We are at the dawn of a new era of personalized medicine and the recent advances in pharmacogenomics will provide clear-cut opportunities to tailor individualized cancer treatment strategies. With better treatments for cancer should come additional insights into how cancer progresses, and knowledge as to which are the most relevant drivers, pathways, and modes of dissemination. These advances may help clarify the extent to which EMT is involved in the various disease states and point to avenues through which our current understanding of the EMT pathway and transitional events can be exploited to target tumors and/or make them more susceptible to treatment regimes. ACKNOWLEDGMENTS We sincerely apologize to colleagues whose work has been omitted from this review owing to space limitations. We thank Stuart Ingham from the Instituto de Neurociencias for his help with the figures. This work has been supported by grants from the Spanish Ministry of Economy and Competitiveness (MINECO-BFU2014-53128-R), the Generalitat Valenciana (Prometeo II/2013/ 002 and ISIC 2012/010), and the European Research Council (ERC AdG322694) to M.A.N. and from Department of Biochemistry and Cancer Sci- ence Institute National University of Singapore and Institute of Molecular Cell Biology A*STAR Singapore to J.P.T. The Instituto de Neurociencias is a Centre of Excellence Severo Ochoa. This review is dedicated to the late Professor Gerald Maurice Edelman who was very influential on the research carried out by J.P.T. and to Mr. Angel Nieto who passed away recently and always was a role model for M.A.N. REFERENCES Abell, A.N., Jordan, N.V., Huang, W., Prat, A., Midland, A.A., Johnson, N.L., Granger, D.A., Mieczkowski, P.A., Perou, C.M., Gomez, S.M., et al. (2011). MAP3K4/CBP-regulated H2B acetylation controls epithelial-mesenchymal transition in trophoblast stem cells. Cell Stem Cell 8, 525–537. Aceto, N., Bardia, A., Miyamoto, D.T., Donaldson, M.C., Wittner, B.S., Spencer, J.A., Yu, M., Pely, A., Engstrom, A., Zhu, H., et al. (2014). Circulating tumor cell clusters are oligoclonal precursors of breast cancer metastasis. Cell 158, 1110–1122. Acharya, A., Baek, S.T., Huang, G., Eskiocak, B., Goetsch, S., Sung, C.Y., Banfi, S., Sauer, M.F., Olsen, G.S., Duffield, J.S., et al. (2012). The bHLH transcription factor Tcf21 is required for lineage-specific EMT of cardiac fibroblast progenitors. Development 139, 2139–2149. Acloque, H., Ocaña, O.H., Matheu, A., Rizzoti, K., Wise, C., Lovell-Badge, R., and Nieto, M.A. (2011). Reciprocal repression between Sox3 and snail transcription factors defines embryonic territories at gastrulation. Dev. Cell 21, 546–558. Aghdassi, A., Sendler, M., Guenther, A., Mayerle, J., Behn, C.O., Heidecke, C.D., Friess, H., Büchler, M., Evert, M., Lerch, M.M., and Weiss, F.U. (2012). Recruitment of histone deacetylases HDAC1 and HDAC2 by the transcriptional repressor ZEB1 downregulates E-cadherin expression in pancreatic cancer. Gut 61, 439–448. Akalay, I., Janji, B., Hasmim, M., Noman, M.Z., André, F., De Cremoux, P., Bertheau, P., Badoual, C., Vielh, P., Larsen, A.K., et al. (2013). Epithelial-tomesenchymal transition and autophagy induction in breast carcinoma promote escape from T-cell-mediated lysis. Cancer Res. 73, 2418–2427. Alderton, G.K. (2013). Metastasis: Epithelial to mesenchymal and back again. Nat. Rev. Cancer 13, 3. Alix-Panabières, C., and Pantel, K. (2014). Technologies for detection of circulating tumor cells: facts and vision. Lab Chip 14, 57–62. Antoniou, A., Hébrant, A., Dom, G., Dumont, J.E., and Maenhaut, C. (2013). Cancer stem cells, a fuzzy evolving concept: a cell population or a cell property? Cell Cycle 12, 3743–3748. Aref, A.R., Huang, R.Y., Yu, W., Chua, K.N., Sun, W., Tu, T.Y., Bai, J., Sim, W.J., Zervantonakis, I.K., Thiery, J.P., and Kamm, R.D. (2013). Screening therapeutic EMT blocking agents in a three-dimensional microenvironment. Integr Biol (Camb) 5, 381–389. Ariza, L., Carmona, R., Canete, A., Cano, E., and Munoz-Chapuli, R. (2016). Coelomic epithelium-derived cells in visceral morphogenesis. Dev. Dyn. 245, 307–322. Arnoux, V., Nassour, M., L’Helgoualc’h, A., Hipskind, R.A., and Savagner, P. (2008). Erk5 controls Slug expression and keratinocyte activation during wound healing. Mol. Biol. Cell 19, 4738–4749. Au, S.H., Storey, B.D., Moore, J.C., Tang, Q., Chen, Y.-L., Javaid, S., Sarioglu, A.F., Sullivan, R., Madden, M.W., O’Keefe, R., et al. (2016). Clusters of circulating tumor cells traverse capillary-sized vessels. Proc. Natl. Acad. Sci. USA 113, 4947–4952. Baccelli, I., Schneeweiss, A., Riethdorf, S., Stenzinger, A., Schillert, A., Vogel, V., Klein, C., Saini, M., Bäuerle, T., Wallwiener, M., et al. (2013). Identification of a population of blood circulating tumor cells from breast cancer patients that initiates metastasis in a xenograft assay. Nat. Biotechnol. 31, 539–544. Baek, S.T., and Tallquist, M.D. (2012). Nf1 limits epicardial derivative expansion by regulating epithelial to mesenchymal transition and proliferation. Development 139, 2040–2049. Baggiolini, A., Varum, S., Mateos, J.M., Bettosini, D., John, N., Bonalli, M., Ziegler, U., Dimou, L., Clevers, H., Furrer, R., and Sommer, L. (2015). Premigratory Cell 166, June 30, 2016 37 and migratory neural crest cells are multipotent in vivo. Cell Stem Cell 16, 314–322. exosomes from injured epithelial cells activate fibroblasts to initiate tissue regenerative responses and fibrosis. J. Am. Soc. Nephrol. 24, 385–392. Bahri, S., Wang, S., Conder, R., Choy, J., Vlachos, S., Dong, K., Merino, C., Sigrist, S., Molnar, C., Yang, X., et al. (2010). The leading edge during dorsal closure as a model for epithelial plasticity: Pak is required for recruitment of the Scribble complex and septate junction formation. Development 137, 2023–2032. Boucaut, J.C., Darribère, T., Boulekbache, H., and Thiery, J.P. (1984). Prevention of gastrulation but not neurulation by antibodies to fibronectin in amphibian embryos. Nature 307, 364–367. Bai, J., Adriani, G., Dang, T.M., Tu, T.Y., Penny, H.X., Wong, S.C., Kamm, R.D., and Thiery, J.P. (2015a). Contact-dependent carcinoma aggregate dispersion by M2a macrophages via ICAM-1 and b2 integrin interactions. Oncotarget 6, 25295–25307. Bai, J., Tu, T.Y., Kim, C., Thiery, J.P., and Kamm, R.D. (2015b). Identification of drugs as single agents or in combination to prevent carcinoma dissemination in a microfluidic 3D environment. Oncotarget 6, 36603–36614. Barcellos-Hoff, M.H., Lyden, D., and Wang, T.C. (2013). The evolution of the cancer niche during multistage carcinogenesis. Nat. Rev. Cancer 13, 511–518. Basu, R., Whitlock, B.M., Husson, J., Le Floc’h, A., Jin, W., Oyler-Yaniv, A., Dotiwala, F., Giannone, G., Hivroz, C., Biais, N., et al. (2016). Cytotoxic T Cells Use Mechanical Force to Potentiate Target Cell Killing. Cell 165, 100–110. Bechtel, W., McGoohan, S., Zeisberg, E.M., Müller, G.A., Kalbacher, H., Salant, D.J., Müller, C.A., Kalluri, R., and Zeisberg, M. (2010). Methylation determines fibroblast activation and fibrogenesis in the kidney. Nat. Med. 16, 544–550. Beck, B., Lapouge, G., Rorive, S., Drogat, B., Desaedelaere, K., Delafaille, S., Dubois, C., Salmon, I., Willekens, K., Marine, J.C., and Blanpain, C. (2015). Different levels of Twist1 regulate skin tumor initiation, stemness, and progression. Cell Stem Cell 16, 67–79. Bedi, U., Mishra, V.K., Wasilewski, D., Scheel, C., and Johnsen, S.A. (2014). Epigenetic plasticity: a central regulator of epithelial-to-mesenchymal transition in cancer. Oncotarget 5, 2016–2029. Beerling, E., Seinstra, D., de Wit, E., Kester, L., van der Velden, D., Maynard, C., Schäfer, R., van Diest, P., Voest, E., van Oudenaarden, A., et al. (2016). Plasticity between Epithelial and Mesenchymal States Unlinks EMT from Metastasis-Enhancing Stem Cell Capacity. Cell Rep. 14, 2281–2288. Bellaye, P.S., Burgy, O., Causse, S., Garrido, C., and Bonniaud, P. (2014). Heat shock proteins in fibrosis and wound healing: good or evil? Pharmacol. Ther. 143, 119–132. Beug, H. (2009). Breast cancer stem cells: eradication by differentiation therapy? Cell 138, 623–625. Bidard, F.C., Vincent-Salomon, A., Gomme, S., Nos, C., de Rycke, Y., Thiery, J.P., Sigal-Zafrani, B., Mignot, L., Sastre-Garau, X., and Pierga, J.Y.; Institut Curie Breast Cancer Study Group (2008). Disseminated tumor cells of breast cancer patients: a strong prognostic factor for distant and local relapse. Clin. Cancer Res. 14, 3306–3311. Bidard, F.C., Peeters, D.J., Fehm, T., Nolé, F., Gisbert-Criado, R., Mavroudis, D., Grisanti, S., Generali, D., Garcia-Saenz, J.A., Stebbing, J., et al. (2014). Clinical validity of circulating tumour cells in patients with metastatic breast cancer: a pooled analysis of individual patient data. Lancet Oncol. 15, 406–414. Blanco, M.J., Barrallo-Gimeno, A., Acloque, H., Reyes, A.E., Tada, M., Allende, M.L., Mayor, R., and Nieto, M.A. (2007). Snail1a and Snail1b cooperate in the anterior migration of the axial mesendoderm in the zebrafish embryo. Development 134, 4073–4081. Boutet, A., De Frutos, C.A., Maxwell, P.H., Mayol, M.J., Romero, J., and Nieto, M.A. (2006). Snail activation disrupts tissue homeostasis and induces fibrosis in the adult kidney. EMBO J. 25, 5603–5613. Brabletz, T. (2012). EMT and MET in metastasis: where are the cancer stem cells? Cancer Cell 22, 699–701. Brabletz, T., Jung, A., Reu, S., Porzner, M., Hlubek, F., Kunz-Schughart, L.A., Knuechel, R., and Kirchner, T. (2001). Variable beta-catenin expression in colorectal cancers indicates tumor progression driven by the tumor environment. Proc. Natl. Acad. Sci. USA 98, 10356–10361. Brabletz, T., Jung, A., Spaderna, S., Hlubek, F., and Kirchner, T. (2005). Opinion: migrating cancer stem cells - an integrated concept of malignant tumour progression. Nat. Rev. Cancer 5, 744–749. Braeutigam, C., Rago, L., Rolke, A., Waldmeier, L., Christofori, G., and Winter, J. (2014). The RNA-binding protein Rbfox2: an essential regulator of EMTdriven alternative splicing and a mediator of cellular invasion. Oncogene 33, 1082–1092. Brockhausen, J., Tay, S.S., Grzelak, C.A., Bertolino, P., Bowen, D.G., d’Avigdor, W.M., Teoh, N., Pok, S., Shackel, N., Gamble, J.R., et al. (2015). miR-181a mediates TGF-b-induced hepatocyte EMT and is dysregulated in cirrhosis and hepatocellular cancer. Liver Int. 35, 240–253. Brønnum, H., Andersen, D.C., Schneider, M., Sandberg, M.B., Eskildsen, T., Nielsen, S.B., Kalluri, R., and Sheikh, S.P. (2013). miR-21 promotes fibrogenic epithelial-to-mesenchymal transition of epicardial mesothelial cells involving Programmed Cell Death 4 and Sprouty-1. PLoS ONE 8, e56280. Buitrago-Delgado, E., Nordin, K., Rao, A., Geary, L., and LaBonne, C. (2015). NEURODEVELOPMENT. Shared regulatory programs suggest retention of blastula-stage potential in neural crest cells. Science 348, 1332–1335. Byers, L.A., Diao, L., Wang, J., Saintigny, P., Girard, L., Peyton, M., Shen, L., Fan, Y., Giri, U., Tumula, P.K., et al. (2013). An epithelial-mesenchymal transition gene signature predicts resistance to EGFR and PI3K inhibitors and identifies Axl as a therapeutic target for overcoming EGFR inhibitor resistance. Clin. Cancer Res. 19, 279–290. Calon, A., Lonardo, E., Berenguer-Llergo, A., Espinet, E., Hernando-Momblona, X., Iglesias, M., Sevillano, M., Palomo-Ponce, S., Tauriello, D.V., Byrom, D., et al. (2015). Stromal gene expression defines poor-prognosis subtypes in colorectal cancer. Nat. Genet. 47, 320–329. Cano, A., Pérez-Moreno, M.A., Rodrigo, I., Locascio, A., Blanco, M.J., del Barrio, M.G., Portillo, F., and Nieto, M.A. (2000). The transcription factor snail controls epithelial-mesenchymal transitions by repressing E-cadherin expression. Nat. Cell Biol. 2, 76–83. Cano, E., Carmona, R., Ruiz-Villalba, A., Rojas, A., Chau, Y.-Y., Wagner, N., Hastie, N.D., Muñoz-Chápuli, R., and Pérez-Pomares, J.M. (2016). Extracardiac septum transversum/proepicardial endothelial cells pattern embryonic coronary arterio-venous connections. Proc. Natl. Acad. Sci. USA 113, 656–661. Carmody, L.C., Germain, A.R., VerPlank, L., Nag, P.P., Muñoz, B., Perez, J.R., and Palmer, M.A. (2012). Phenotypic high-throughput screening elucidates target pathway in breast cancer stem cell-like cells. J. Biomol. Screen. 17, 1204–1210. Boglev, Y., Wilanowski, T., Caddy, J., Parekh, V., Auden, A., Darido, C., Hislop, N.R., Cangkrama, M., Ting, S.B., and Jane, S.M. (2011). The unique and cooperative roles of the Grainy head-like transcription factors in epidermal development reflect unexpected target gene specificity. Dev. Biol. 349, 512–522. Carver, E.A., Jiang, R., Lan, Y., Oram, K.F., and Gridley, T. (2001). The mouse snail gene encodes a key regulator of the epithelial-mesenchymal transition. Mol. Cell. Biol. 21, 8184–8188. Bonomi, S., di Matteo, A., Buratti, E., Cabianca, D.S., Baralle, F.E., Ghigna, C., and Biamonti, G. (2013). HnRNP A1 controls a splicing regulatory circuit promoting mesenchymal-to-epithelial transition. Nucleic Acids Res. 41, 8665– 8679. Celià-Terrassa, T., Meca-Cortés, O., Mateo, F., Martı́nez de Paz, A., Rubio, N., Arnal-Estapé, A., Ell, B.J., Bermudo, R., Dı́az, A., Guerra-Rebollo, M., et al. (2012). Epithelial-mesenchymal transition can suppress major attributes of human epithelial tumor-initiating cells. J. Clin. Invest. 122, 1849–1868. Borges, F.T., Melo, S.A., Özdemir, B.C., Kato, N., Revuelta, I., Miller, C.A., Gattone, V.H., 2nd, LeBleu, V.S., and Kalluri, R. (2013). TGF-b1-containing Chaffer, C.L., Brueckmann, I., Scheel, C., Kaestli, A.J., Wiggins, P.A., Rodrigues, L.O., Brooks, M., Reinhardt, F., Su, Y., Polyak, K., et al. (2011). Normal 38 Cell 166, June 30, 2016 and neoplastic nonstem cells can spontaneously convert to a stem-like state. Proc. Natl. Acad. Sci. USA 108, 7950–7955. Chaffer, C.L., Marjanovic, N.D., Lee, T., Bell, G., Kleer, C.G., Reinhardt, F., D’Alessio, A.C., Young, R.A., and Weinberg, R.A. (2013). Poised chromatin at the ZEB1 promoter enables breast cancer cell plasticity and enhances tumorigenicity. Cell 154, 61–74. isin, M., Chakrabarti, R., Hwang, J., Andres Blanco, M., Wei, Y., Lukac Romano, R.A., Smalley, K., Liu, S., Yang, Q., Ibrahim, T., et al. (2012). Elf5 inhibits the epithelial-mesenchymal transition in mammary gland development and breast cancer metastasis by transcriptionally repressing Snail2. Nat. Cell Biol. 14, 1212–1222. Chan, Y.S., Göke, J., Lu, X., Venkatesan, N., Feng, B., Su, I.H., and Ng, H.H. (2013). A PRC2-dependent repressive role of PRDM14 in human embryonic stem cells and induced pluripotent stem cell reprogramming. Stem Cells 31, 682–692. Chang, C.J., Chao, C.H., Xia, W., Yang, J.Y., Xiong, Y., Li, C.W., Yu, W.H., Rehman, S.K., Hsu, J.L., Lee, H.H., et al. (2011). p53 regulates epithelial-mesenchymal transition and stem cell properties through modulating miRNAs. Nat. Cell Biol. 13, 317–323. Chang, C.C., Hsu, W.H., Wang, C.C., Chou, C.H., Kuo, M.Y., Lin, B.R., Chen, S.T., Tai, S.K., Kuo, M.L., and Yang, M.H. (2013). Connective tissue growth factor activates pluripotency genes and mesenchymal-epithelial transition in head and neck cancer cells. Cancer Res. 73, 4147–4157. Chen, Y.L., Zhang, X., Bai, J., Gai, L., Ye, X.L., Zhang, L., Xu, Q., Zhang, Y.X., Xu, L., Li, H.P., and Ding, X. (2013). Sorafenib ameliorates bleomycin-induced pulmonary fibrosis: potential roles in the inhibition of epithelial-mesenchymal transition and fibroblast activation. Cell Death Dis. 4, e665. Chen, A., Wong, C.S., Liu, M.C., House, C.M., Sceneay, J., Bowtell, D.D., Thompson, E.W., and Möller, A. (2015). The ubiquitin ligase Siah is a novel regulator of Zeb1 in breast cancer. Oncotarget 6, 862–873. Clay, M.R., and Halloran, M.C. (2014). Cadherin 6 promotes neural crest cell detachment via F-actin regulation and influences active Rho distribution during epithelial-to-mesenchymal transition. Development 141, 2506–2515. Conn, S.J., Pillman, K.A., Toubia, J., Conn, V.M., Salmanidis, M., Phillips, C.A., Roslan, S., Schreiber, A.W., Gregory, P.A., and Goodall, G.J. (2015). The RNA binding protein quaking regulates formation of circRNAs. Cell 160, 1125–1134. Costa-Silva, B., Aiello, N.M., Ocean, A.J., Singh, S., Zhang, H., Thakur, B.K., Becker, A., Hoshino, A., Mark, M.T., Molina, H., et al. (2015). Pancreatic cancer exosomes initiate pre-metastatic niche formation in the liver. Nat. Cell Biol. 17, 816–826. Cristofanilli, M., Budd, G.T., Ellis, M.J., Stopeck, A., Matera, J., Miller, M.C., Reuben, J.M., Doyle, G.V., Allard, W.J., Terstappen, L.W., and Hayes, D.F. (2004). Circulating tumor cells, disease progression, and survival in metastatic breast cancer. N. Engl. J. Med. 351, 781–791. Darido, C., and Jane, S.M. (2010). Grhl3 and GEF19 in the front rho. Small GTPases 1, 104–107. David, C.J., Huang, Y.H., Chen, M., Su, J., Zou, Y., Bardeesy, N., IacobuzioDonahue, C.A., and Massagué, J. (2016). TGF-b Tumor Suppression through a Lethal EMT. Cell 164, 1015–1030. Del Pozo Martin, Y., Park, D., Ramachandran, A., Ombrato, L., Calvo, F., Chakravarty, P., Spencer-Dene, B., Derzsi, S., Hill, C.S., Sahai, E., and Malanchi, I. (2015). Mesenchymal Cancer Cell-Stroma Crosstalk Promotes Niche Activation, Epithelial Reversion, and Metastatic Colonization. Cell Rep. 13, 2456– 2469. Deng, H., Yang, F., Xu, H., Sun, Y., Xue, X., Du, S., Wang, X., Li, S., Liu, Y., and Wang, R. (2014). Ac-SDKP suppresses epithelial-mesenchymal transition in A549 cells via HSP27 signaling. Exp. Mol. Pathol. 97, 176–183. Diaz, V.M., and de Herreros, A.G. (2016). F-box proteins: Keeping the epithelial-to-mesenchymal transition (EMT) in check. Semin. Cancer Biol. 36, 71–79. Cheng, S.Z.D., and Keller, A. (1998). The role of metastable states in polymer phase transitions: Concepts, principles, and experimental observations. Annu. Rev. Mater. Sci. 28, 533–562. Dı́az-López, A., Moreno-Bueno, G., and Cano, A. (2014). Role of microRNA in epithelial to mesenchymal transition and metastasis and clinical perspectives. Cancer Manag. Res. 6, 205–216. Chouaib, S., Janji, B., Tittarelli, A., Eggermont, A., and Thiery, J.P. (2014). Tumor plasticity interferes with anti-tumor immunity. Crit. Rev. Immunol. 34, 91–102. Domı́nguez, D., Montserrat-Sentı́s, B., Virgós-Soler, A., Guaita, S., Grueso, J., Porta, M., Puig, I., Baulida, J., Francı́, C., and Garcı́a de Herreros, A. (2003). Phosphorylation regulates the subcellular location and activity of the snail transcriptional repressor. Mol. Cell. Biol. 23, 5078–5089. Choudhary, K.S., Rohatgi, N., Halldorsson, S., Briem, E., Gudjonsson, T., Gudmundsson, S., and Rolfsson, O. (2016). EGFR Signal-Network Reconstruction Demonstrates Metabolic Crosstalk in EMT. PLoS Comput. Biol. 12, e1004924. Chu, Y.S., Eder, O., Thomas, W.A., Simcha, I., Pincet, F., Ben-Ze’ev, A., Perez, E., Thiery, J.P., and Dufour, S. (2006). Prototypical type I E-cadherin and type II cadherin-7 mediate very distinct adhesiveness through their extracellular domains. J. Biol. Chem. 281, 2901–2910. Chu, A.S., Diaz, R., Hui, J.J., Yanger, K., Zong, Y., Alpini, G., Stanger, B.Z., and Wells, R.G. (2011). Lineage tracing demonstrates no evidence of cholangiocyte epithelial-to-mesenchymal transition in murine models of hepatic fibrosis. Hepatology 53, 1685–1695. Chua, K.N., Poon, K.L., Lim, J., Sim, W.J., Huang, R.Y., and Thiery, J.P. (2011). Target cell movement in tumor and cardiovascular diseases based on the epithelial-mesenchymal transition concept. Adv. Drug Deliv. Rev. 63, 558–567. Chua, K.N., Sim, W.J., Racine, V., Lee, S.Y., Goh, B.C., and Thiery, J.P. (2012). A cell-based small molecule screening method for identifying inhibitors of epithelial-mesenchymal transition in carcinoma. PLoS ONE 7, e33183. Chung, V.Y., Tan, T.Z., Tan, M., Wong, M.K., Kuay, K.T., Yang, Z., Ye, J., Muller, J., Koh, C.M.-Y., Guccione, E., et al. (2016). GRHL2-miR-200-ZEB1 maintains the epithelial status of ovarian cancer through transcriptional regulation and histone modification. Sci. Rep. 6, 19943. http://dx.doi.org/10.1038/ srep19943. Cieply, B., Riley, P., 4th, Pifer, P.M., Widmeyer, J., Addison, J.B., Ivanov, A.V., Denvir, J., and Frisch, S.M. (2012). Suppression of the epithelial-mesenchymal transition by Grainyhead-like-2. Cancer Res. 72, 2440–2453. Ciruna, B., and Rossant, J. (2001). FGF signaling regulates mesoderm cell fate specification and morphogenetic movement at the primitive streak. Dev. Cell 1, 37–49. Dong, C., Wu, Y., Wang, Y., Wang, C., Kang, T., Rychahou, P.G., Chi, Y.I., Evers, B.M., and Zhou, B.P. (2013). Interaction with Suv39H1 is critical for Snail-mediated E-cadherin repression in breast cancer. Oncogene 32, 1351– 1362. Drasin, D.J., Guarnieri, A.L., Neelakantan, D., Kim, J., Cabrera, J.H., Wang, C.A., Zaberezhnyy, V., Gasparini, P., Cascione, L., Huebner, K., et al. (2015). TWIST1-Induced miR-424 Reversibly Drives Mesenchymal Programming while Inhibiting Tumor Initiation. Cancer Res. 75, 1908–1921. Du, C., Zhang, C., Hassan, S., Biswas, M.H., and Balaji, K.C. (2010). Protein kinase D1 suppresses epithelial-to-mesenchymal transition through phosphorylation of snail. Cancer Res. 70, 7810–7819. Edelman, G.M., Gallin, W.J., Delouvée, A., Cunningham, B.A., and Thiery, J.P. (1983). Early epochal maps of two different cell adhesion molecules. Proc. Natl. Acad. Sci. USA 80, 4384–4388. Engl, W., Viasnoff, V., and Thiery, J.P. (2015). Epithelial Mesenchymal Transition Influence on CTL Activity. In Resistance of Cancer Cells to CTL-Mediated Immunotherapy, B. Bonavida and S. Chouaib, eds. (Springer International Publishing), pp. 267–284. Enkhbaatar, Z., Terashima, M., Oktyabri, D., Tange, S., Ishimura, A., Yano, S., and Suzuki, T. (2013). KDM5B histone demethylase controls epithelial-mesenchymal transition of cancer cells by regulating the expression of the microRNA-200 family. Cell Cycle 12, 2100–2112. Fischer, K.R., Durrans, A., Lee, S., Sheng, J., Li, F., Wong, S.T., Choi, H., El Rayes, T., Ryu, S., Troeger, J., et al. (2015). Epithelial-to-mesenchymal transition is not required for lung metastasis but contributes to chemoresistance. Nature 527, 472–476. Cell 166, June 30, 2016 39 Futterman, M.A., Garcı́a, A.J., and Zamir, E.A. (2011). Evidence for partial epithelial-to-mesenchymal transition (pEMT) and recruitment of motile blastoderm edge cells during avian epiboly. Dev. Dyn. 240, 1502–1511. Hlubek, F., Löhberg, C., Meiler, J., Jung, A., Kirchner, T., and Brabletz, T. (2001). Tip60 is a cell-type-specific transcriptional regulator. J. Biochem. 129, 635–641. Gaeta, X., Xie, Y., and Lowry, W.E. (2013). Sequential addition of reprogramming factors improves efficiency. Nat. Cell Biol. 15, 725–727. Hölzel, M., Bovier, A., and Tüting, T. (2013). Plasticity of tumour and immune cells: a source of heterogeneity and a cause for therapy resistance? Nat. Rev. Cancer 13, 365–376. Gentles, A.J., Newman, A.M., Liu, C.L., Bratman, S.V., Feng, W., Kim, D., Nair, V.S., Xu, Y., Khuong, A., Hoang, C.D., et al. (2015). The prognostic landscape of genes and infiltrating immune cells across human cancers. Nat. Med. 21, 938–945. Giampieri, S., Manning, C., Hooper, S., Jones, L., Hill, C.S., and Sahai, E. (2009). Localized and reversible TGFbeta signalling switches breast cancer cells from cohesive to single cell motility. Nat. Cell Biol. 11, 1287–1296. Giannelli, G., Villa, E., and Lahn, M. (2014). Transforming growth factor-b as a therapeutic target in hepatocellular carcinoma. Cancer Res. 74, 1890–1894. Ginestier, C., Hur, M.H., Charafe-Jauffret, E., Monville, F., Dutcher, J., Brown, M., Jacquemier, J., Viens, P., Kleer, C.G., Liu, S., et al. (2007). ALDH1 is a marker of normal and malignant human mammary stem cells and a predictor of poor clinical outcome. Cell Stem Cell 1, 555–567. Gjerdrum, C., Tiron, C., Høiby, T., Stefansson, I., Haugen, H., Sandal, T., Collett, K., Li, S., McCormack, E., Gjertsen, B.T., et al. (2010). Axl is an essential epithelial-to-mesenchymal transition-induced regulator of breast cancer metastasis and patient survival. Proc. Natl. Acad. Sci. USA 107, 1124–1129. Hong, J., Zhou, J., Fu, J., He, T., Qin, J., Wang, L., Liao, L., and Xu, J. (2011). Phosphorylation of serine 68 of Twist1 by MAPKs stabilizes Twist1 protein and promotes breast cancer cell invasiveness. Cancer Res. 71, 3980–3990. Hong, T., Watanabe, K., Ta, C.H., Villarreal-Ponce, A., Nie, Q., and Dai, X. (2015). An Ovol2-Zeb1 Mutual Inhibitory Circuit Governs Bidirectional and Multi-step Transition between Epithelial and Mesenchymal States. PLoS Comput. Biol. 11, e1004569. Hoshino, A., Costa-Silva, B., Shen, T.L., Rodrigues, G., Hashimoto, A., Tesic Mark, M., Molina, H., Kohsaka, S., Di Giannatale, A., Ceder, S., et al. (2015). Tumour exosome integrins determine organotropic metastasis. Nature 527, 329–335. Hsu, D.S., Wang, H.J., Tai, S.K., Chou, C.H., Hsieh, C.H., Chiu, P.H., Chen, N.J., and Yang, M.H. (2014). Acetylation of snail modulates the cytokinome of cancer cells to enhance the recruitment of macrophages. Cancer Cell 26, 534–548. Hu, N., Strobl-Mazzulla, P.H., and Bronner, M.E. (2014). Epigenetic regulation in neural crest development. Dev. Biol. 396, 159–168. Goel, H.L., Gritsko, T., Pursell, B., Chang, C., Shultz, L.D., Greiner, D.L., Norum, J.H., Toftgard, R., Shaw, L.M., and Mercurio, A.M. (2014). Regulated splicing of the a6 integrin cytoplasmic domain determines the fate of breast cancer stem cells. Cell Rep. 7, 747–761. Huang, R.Y., Guilford, P., and Thiery, J.P. (2012). Early events in cell adhesion and polarity during epithelial-mesenchymal transition. J. Cell Sci. 125, 4417– 4422. Grande, M.T., Sánchez-Laorden, B., López-Blau, C., De Frutos, C.A., Boutet, A., Arévalo, M., Rowe, R.G., Weiss, S.J., López-Novoa, J.M., and Nieto, M.A. (2015). Snail1-induced partial epithelial-to-mesenchymal transition drives renal fibrosis in mice and can be targeted to reverse established disease. Nat. Med. 21, 989–997. Huang, R.Y., Wong, M.K., Tan, T.Z., Kuay, K.T., Ng, A.H., Chung, V.Y., Chu, Y.S., Matsumura, N., Lai, H.C., Lee, Y.F., et al. (2013). An EMT spectrum defines an anoikis-resistant and spheroidogenic intermediate mesenchymal state that is sensitive to e-cadherin restoration by a src-kinase inhibitor, saracatinib (AZD0530). Cell Death Dis. 4, e915. Greenburg, G., and Hay, E.D. (1982). Epithelia suspended in collagen gels can lose polarity and express characteristics of migrating mesenchymal cells. J. Cell Biol. 95, 333–339. Hudson, L.G., Newkirk, K.M., Chandler, H.L., Choi, C., Fossey, S.L., Parent, A.E., and Kusewitt, D.F. (2009). Cutaneous wound reepithelialization is compromised in mice lacking functional Slug (Snai2). J. Dermatol. Sci. 56, 19–26. Grigore, A.D., Jolly, M.K., Jia, D., Farach-Carson, M.C., and Levine, H. (2016). Tumor Budding: The Name is EMT. Partial EMT. J. Clin. Med. 5, E51. Gulati, A., Jabbour, A., Ismail, T.F., Guha, K., Khwaja, J., Raza, S., Morarji, K., Brown, T.D., Ismail, N.A., Dweck, M.R., et al. (2013). Association of fibrosis with mortality and sudden cardiac death in patients with nonischemic dilated cardiomyopathy. JAMA 309, 896–908. Gupta, P.B., Chaffer, C.L., and Weinberg, R.A. (2009a). Cancer stem cells: mirage or reality? Nat. Med. 15, 1010–1012. Gupta, P.B., Onder, T.T., Jiang, G., Tao, K., Kuperwasser, C., Weinberg, R.A., and Lander, E.S. (2009b). Identification of selective inhibitors of cancer stem cells by high-throughput screening. Cell 138, 645–659. Halbleib, J.M., and Nelson, W.J. (2006). Cadherins in development: cell adhesion, sorting, and tissue morphogenesis. Genes Dev. 20, 3199–3214. Hamilton, D.H., Litzinger, M.T., Jales, A., Huang, B., Fernando, R.I., Hodge, J.W., Ardiani, A., Apelian, D., Schlom, J., and Palena, C. (2013). Immunological targeting of tumor cells undergoing an epithelial-mesenchymal transition via a recombinant brachyury-yeast vaccine. Oncotarget 4, 1777–1790. Harney, A.S., Arwert, E.N., Entenberg, D., Wang, Y., Guo, P., Qian, B.Z., Oktay, M.H., Pollard, J.W., Jones, J.G., and Condeelis, J.S. (2015). Real-Time Imaging Reveals Local, Transient Vascular Permeability, and Tumor Cell Intravasation Stimulated by TIE2hi Macrophage-Derived VEGFA. Cancer Discov. 5, 932–943. Hay, E.D. (1995). An overview of epithelio-mesenchymal transformation. Acta Anat. (Basel) 154, 8–20. Herranz, N., Pasini, D., Dı́az, V.M., Francı́, C., Gutierrez, A., Dave, N., Escrivà, M., Hernandez-Muñoz, I., Di Croce, L., Helin, K., et al. (2008). Polycomb complex 2 is required for E-cadherin repression by the Snail1 transcription factor. Mol. Cell. Biol. 28, 4772–4781. 40 Cell 166, June 30, 2016 Huleihel, L., Ben-Yehudah, A., Milosevic, J., Yu, G., Pandit, K., Sakamoto, K., Yousef, H., LeJeune, M., Coon, T.A., Redinger, C.J., et al. (2014). Let-7d microRNA affects mesenchymal phenotypic properties of lung fibroblasts. Am. J. Physiol. Lung Cell. Mol. Physiol. 306, L534–L542. Humphreys, B.D., Lin, S.L., Kobayashi, A., Hudson, T.E., Nowlin, B.T., Bonventre, J.V., Valerius, M.T., McMahon, A.P., and Duffield, J.S. (2010). Fate tracing reveals the pericyte and not epithelial origin of myofibroblasts in kidney fibrosis. Am. J. Pathol. 176, 85–97. Hüsemann, Y., Geigl, J.B., Schubert, F., Musiani, P., Meyer, M., Burghart, E., Forni, G., Eils, R., Fehm, T., Riethmüller, G., and Klein, C.A. (2008). Systemic spread is an early step in breast cancer. Cancer Cell 13, 58–68. Hutchinson, J., Fogarty, A., Hubbard, R., and McKeever, T. (2015). Global incidence and mortality of idiopathic pulmonary fibrosis: a systematic review. Eur. Respir. J. 46, 795–806. Hwang, W.L., Jiang, J.K., Yang, S.H., Huang, T.S., Lan, H.Y., Teng, H.W., Yang, C.Y., Tsai, Y.P., Lin, C.H., Wang, H.W., and Yang, M.H. (2014). MicroRNA-146a directs the symmetric division of Snail-dominant colorectal cancer stem cells. Nat. Cell Biol. 16, 268–280. Iliopoulos, D., Lindahl-Allen, M., Polytarchou, C., Hirsch, H.A., Tsichlis, P.N., and Struhl, K. (2010). Loss of miR-200 inhibition of Suz12 leads to polycomb-mediated repression required for the formation and maintenance of cancer stem cells. Mol. Cell 39, 761–772. Infante, J.R., Camidge, D.R., Mileshkin, L.R., Chen, E.X., Hicks, R.J., Rischin, D., Fingert, H., Pierce, K.J., Xu, H., Roberts, W.G., et al. (2012). Safety, pharmacokinetic, and pharmacodynamic phase I dose-escalation trial of PF-00562271, an inhibitor of focal adhesion kinase, in advanced solid tumors. J. Clin. Oncol. 30, 1527–1533. Iwano, M., Plieth, D., Danoff, T.M., Xue, C., Okada, H., and Neilson, E.G. (2002). Evidence that fibroblasts derive from epithelium during tissue fibrosis. J. Clin. Invest. 110, 341–350. Javaid, S., Zhang, J., Anderssen, E., Black, J.C., Wittner, B.S., Tajima, K., Ting, D.T., Smolen, G.A., Zubrowski, M., Desai, R., et al. (2013). Dynamic chromatin modification sustains epithelial-mesenchymal transition following inducible expression of Snail-1. Cell Rep. 5, 1679–1689. Jiang, L., Xiao, L., Sugiura, H., Huang, X., Ali, A., Kuro-o, M., Deberardinis, R.J., and Boothman, D.A. (2015). Metabolic reprogramming during TGFb1induced epithelial-to-mesenchymal transition. Oncogene 34, 3908–3916. Jolly, M.K., Boareto, M., Huang, B., Jia, D., Lu, M., Ben-Jacob, E., Onuchic, J.N., and Levine, H. (2015). Implications of the Hybrid Epithelial/Mesenchymal Phenotype in Metastasis. Front. Oncol. 5, 155. Jolly, M.K., Tripathi, S.C., Jia, D., Mooney, S.M., Celiktas, M., Hanash, S.M., Mani, S.A., Pienta, K.J., Ben-Jacob, E., and Levine, H. (2016). Stability of the hybrid epithelial/mesenchymal phenotype. Oncotarget. http://dx.doi.org/10. 18632/oncotarget.8166. Jordan, N.V., Johnson, G.L., and Abell, A.N. (2011). Tracking the intermediate stages of epithelial-mesenchymal transition in epithelial stem cells and cancer. Cell Cycle 10, 2865–2873. Jung, H.Y., and Yang, J. (2015). Unraveling the TWIST between EMT and cancer stemness. Cell Stem Cell 16, 1–2. Kaplan, R.N., Rafii, S., and Lyden, D. (2006). Preparing the ‘‘soil’’: the premetastatic niche. Cancer Res. 66, 11089–11093. Khoo, B.L., Lee, S.C., Kumar, P., Tan, T.Z., Warkiani, M.E., Ow, S.G., Nandi, S., Lim, C.T., and Thiery, J.P. (2015). Short-term expansion of breast circulating cancer cells predicts response to anti-cancer therapy. Oncotarget 6, 15578–15593. Lee, R.T., Nagai, H., Nakaya, Y., Sheng, G., Trainor, P.A., Weston, J.A., and Thiery, J.P. (2013). Cell delamination in the mesencephalic neural fold and its implication for the origin of ectomesenchyme. Development 140, 4890– 4902. Lee, B., Villarreal-Ponce, A., Fallahi, M., Ovadia, J., Sun, P., Yu, Q.C., Ito, S., Sinha, S., Nie, Q., and Dai, X. (2014). Transcriptional mechanisms link epithelial plasticity to adhesion and differentiation of epidermal progenitor cells. Dev. Cell 29, 47–58. Leopold, P.L., Vincent, J., and Wang, H. (2012). A comparison of epithelialto-mesenchymal transition and re-epithelialization. Semin. Cancer Biol. 22, 471–483. Leptin, M., and Grunewald, B. (1990). Cell shape changes during gastrulation in Drosophila. Development 110, 73–84. Leroy, P., and Mostov, K.E. (2007). Slug is required for cell survival during partial epithelial-mesenchymal transition of HGF-induced tubulogenesis. Mol. Biol. Cell 18, 1943–1952. Li, W., and Kang, Y. (2016). Probing the fifty shades of EMT in metastasis. Trends Cancer 2, 65–67. Li, L., Zepeda-Orozco, D., Black, R., and Lin, F. (2010a). Autophagy is a component of epithelial cell fate in obstructive uropathy. Am. J. Pathol. 176, 1767–1778. Li, R., Liang, J., Ni, S., Zhou, T., Qing, X., Li, H., He, W., Chen, J., Li, F., Zhuang, Q., et al. (2010b). A mesenchymal-to-epithelial transition initiates and is required for the nuclear reprogramming of mouse fibroblasts. Cell Stem Cell 7, 51–63. Lianos, G.D., Alexiou, G.A., Mangano, A., Mangano, A., Rausei, S., Boni, L., Dionigi, G., and Roukos, D.H. (2015). The role of heat shock proteins in cancer. Cancer Lett. 360, 114–118. Kim, L.C., Song, L., and Haura, E.B. (2009). Src kinases as therapeutic targets for cancer. Nat. Rev. Clin. Oncol. 6, 587–595. Lim, J., and Thiery, J.P. (2012). Epithelial-mesenchymal transitions: insights from development. Development 139, 3471–3486. Kitazawa, K., Hikichi, T., Nakamura, T., Mitsunaga, K., Tanaka, A., Nakamura, M., Yamakawa, T., Furukawa, S., Takasaka, M., Goshima, N., et al. (2016). OVOL2 Maintains the Transcriptional Program of Human Corneal Epithelium by Suppressing Epithelial-to-Mesenchymal Transition. Cell Rep. 15, 1359–1368. Lim, Y.Y., Wright, J.A., Attema, J.L., Gregory, P.A., Bert, A.G., Smith, E., Thomas, D., Lopez, A.F., Drew, P.A., Khew-Goodall, Y., and Goodall, G.J. (2013). Epigenetic modulation of the miR-200 family is associated with transition to a breast cancer stem-cell-like state. J. Cell Sci. 126, 2256–2266. Kong, D., Banerjee, S., Ahmad, A., Li, Y., Wang, Z., Sethi, S., and Sarkar, F.H. (2010). Epithelial to mesenchymal transition is mechanistically linked with stem cell signatures in prostate cancer cells. PLoS ONE 5, e12445. Lin, Y., Wu, Y., Li, J., Dong, C., Ye, X., Chi, Y.I., Evers, B.M., and Zhou, B.P. (2010). The SNAG domain of Snail1 functions as a molecular hook for recruiting lysine-specific demethylase 1. EMBO J. 29, 1803–1816. Krenning, G., Zeisberg, E.M., and Kalluri, R. (2010). The origin of fibroblasts and mechanism of cardiac fibrosis. J. Cell. Physiol. 225, 631–637. Lin, Y., Dong, C., and Zhou, B.P. (2014). Epigenetic regulation of EMT: the Snail story. Curr. Pharm. Des. 20, 1698–1705. Kudo-Saito, C., Shirako, H., Takeuchi, T., and Kawakami, Y. (2009). Cancer metastasis is accelerated through immunosuppression during Snail-induced EMT of cancer cells. Cancer Cell 15, 195–206. Liu, S., Cong, Y., Wang, D., Sun, Y., Deng, L., Liu, Y., Martin-Trevino, R., Shang, L., McDermott, S.P., Landis, M.D., et al. (2013a). Breast cancer stem cells transition between epithelial and mesenchymal states reflective of their normal counterparts. Stem Cell Reports 2, 78–91. Kurley, S.J., Bierie, B., Carnahan, R.H., Lobdell, N.A., Davis, M.A., Hofmann, I., Moses, H.L., Muller, W.J., and Reynolds, A.B. (2012). p120-catenin is essential for terminal end bud function and mammary morphogenesis. Development 139, 1754–1764. Lamouille, S., Subramanyam, D., Blelloch, R., and Derynck, R. (2013). Regulation of epithelial-mesenchymal and mesenchymal-epithelial transitions by microRNAs. Curr. Opin. Cell Biol. 25, 200–207. Lamouille, S., Xu, J., and Derynck, R. (2014). Molecular mechanisms of epithelial-mesenchymal transition. Nat. Rev. Mol. Cell Biol. 15, 178–196. Lane, D.P. (1992). Cancer. p53, guardian of the genome. Nature 358, 15–16. Lau, E.Y., Lo, J., Cheng, B.Y., Ma, M.K., Lee, J.M., Ng, J.K., Chai, S., Lin, C.H., Tsang, S.Y., Ma, S., et al. (2016). Cancer-Associated Fibroblasts Regulate Tumor-Initiating Cell Plasticity in Hepatocellular Carcinoma through c-Met/ FRA1/HEY1 Signaling. Cell Rep. 15, 1175–1189. LeBleu, V.S., Taduri, G., O’Connell, J., Teng, Y., Cooke, V.G., Woda, C., Sugimoto, H., and Kalluri, R. (2013). Origin and function of myofibroblasts in kidney fibrosis. Nat. Med. 19, 1047–1053. Lee, J.M., Dedhar, S., Kalluri, R., and Thompson, E.W. (2006). The epithelialmesenchymal transition: new insights in signaling, development, and disease. J. Cell Biol. 172, 973–981. Liu, X., Sun, H., Qi, J., Wang, L., He, S., Liu, J., Feng, C., Chen, C., Li, W., Guo, Y., et al. (2013b). Sequential introduction of reprogramming factors reveals a time-sensitive requirement for individual factors and a sequential EMT-MET mechanism for optimal reprogramming. Nat. Cell Biol. 15, 829–838. Lomelı́, H., Starling, C., and Gridley, T. (2009). Epiblast-specific Snai1 deletion results in embryonic lethality due to multiple vascular defects. BMC Res. Notes 2, 22. Lovisa, S., LeBleu, V.S., Tampe, B., Sugimoto, H., Vadnagara, K., Carstens, J.L., Wu, C.C., Hagos, Y., Burckhardt, B.C., Pentcheva-Hoang, T., et al. (2015). Epithelial-to-mesenchymal transition induces cell cycle arrest and parenchymal damage in renal fibrosis. Nat. Med. 21, 998–1009. Lu, M., Jolly, M.K., Levine, H., Onuchic, J.N., and Ben-Jacob, E. (2013). MicroRNA-based regulation of epithelial-hybrid-mesenchymal fate determination. Proc. Natl. Acad. Sci. USA 110, 18144–18149. Luxán, G., D’Amato, G., MacGrogan, D., and de la Pompa, J.L. (2016). Endocardial Notch Signaling in Cardiac Development and Disease. Circ. Res. 118, e1–e18. Maheswaran, S., and Haber, D.A. (2015). Cell fate: Transition loses its invasive edge. Nature 527, 452–453. Cell 166, June 30, 2016 41 Manhire-Heath, R., Golenkina, S., Saint, R., and Murray, M.J. (2013). Netrindependent downregulation of Frazzled/DCC is required for the dissociation of the peripodial epithelium in Drosophila. Nat. Commun. 4, 2790. Murray, S.A., and Gridley, T. (2006). Snail family genes are required for leftright asymmetry determination, but not neural crest formation, in mice. Proc. Natl. Acad. Sci. USA 103, 10300–10304. Mani, S.A., Guo, W., Liao, M.J., Eaton, E.N., Ayyanan, A., Zhou, A.Y., Brooks, M., Reinhard, F., Zhang, C.C., Shipitsin, M., et al. (2008). The epithelial-mesenchymal transition generates cells with properties of stem cells. Cell 133, 704–715. Narasimha, M., Uv, A., Krejci, A., Brown, N.H., and Bray, S.J. (2008). Grainy head promotes expression of septate junction proteins and influences epithelial morphogenesis. J. Cell Sci. 121, 747–752. Marmai, C., Sutherland, R.E., Kim, K.K., Dolganov, G.M., Fang, X., Kim, S.S., Jiang, S., Golden, J.A., Hoopes, C.W., Matthay, M.A., et al. (2011). Alveolar epithelial cells express mesenchymal proteins in patients with idiopathic pulmonary fibrosis. Am. J. Physiol. Lung Cell. Mol. Physiol. 301, L71–L78. Navarro, P., Lozano, E., and Cano, A. (1993). Expression of E- or P-cadherin is not sufficient to modify the morphology and the tumorigenic behavior of murine spindle carcinoma cells. Possible involvement of plakoglobin. J. Cell Sci. 105, 923–934. Mathow, D., Chessa, F., Rabionet, M., Kaden, S., Jennemann, R., Sandhoff, R., Gröne, H.J., and Feuerborn, A. (2015). Zeb1 affects epithelial cell adhesion by diverting glycosphingolipid metabolism. EMBO Rep. 16, 321–331. McCoy, E.L., Iwanaga, R., Jedlicka, P., Abbey, N.S., Chodosh, L.A., Heichman, K.A., Welm, A.L., and Ford, H.L. (2009). Six1 expands the mouse mammary epithelial stem/progenitor cell pool and induces mammary tumors that undergo epithelial-mesenchymal transition. J. Clin. Invest. 119, 2663–2677. Nielsen, S.R., Quaranta, V., Linford, A., Emeagi, P., Rainer, C., Santos, A., Ireland, L., Sakai, T., Sakai, K., Kim, Y.S., et al. (2016). Macrophage-secreted granulin supports pancreatic cancer metastasis by inducing liver fibrosis. Nat. Cell Biol. 18, 549–560. Nieto, M.A. (2013). Epithelial plasticity: a common theme in embryonic and cancer cells. Science 342, 1234850. McDonald, O.G., Wu, H., Timp, W., Doi, A., and Feinberg, A.P. (2011). Genome-scale epigenetic reprogramming during epithelial-to-mesenchymal transition. Nat. Struct. Mol. Biol. 18, 867–874. Nieto, M.A., Sargent, M.G., Wilkinson, D.G., and Cooke, J. (1994). Control of cell behavior during vertebrate development by Slug, a zinc finger gene. Science 264, 835–839. McGranahan, N., and Swanton, C. (2015). Biological and therapeutic impact of intratumor heterogeneity in cancer evolution. Cancer Cell 27, 15–26. Niwa, H., Miyazaki, J., and Smith, A.G. (2000). Quantitative expression of Oct-3/4 defines differentiation, dedifferentiation or self-renewal of ES cells. Nat. Genet. 24, 372–376. Mederacke, I., Hsu, C.C., Troeger, J.S., Huebener, P., Mu, X., Dapito, D.H., Pradere, J.P., and Schwabe, R.F. (2013). Fate tracing reveals hepatic stellate cells as dominant contributors to liver fibrosis independent of its aetiology. Nat. Commun. 4, 2823. Mehal, W.Z., Iredale, J., and Friedman, S.L. (2011). Scraping fibrosis: expressway to the core of fibrosis. Nat. Med. 17, 552–553. Meidhof, S., Brabletz, S., Lehmann, W., Preca, B.T., Mock, K., Ruh, M., Schuler, J., Berthold, M., Weber, A., Burk, U., et al. (2015). ZEB1-associated drug resistance in cancer cells is reversed by the class I HDAC inhibitor mocetinostat. EMBO Mol. Med. 7, 831–847. Millanes-Romero, A., Herranz, N., Perrera, V., Iturbide, A., Loubat-Casanovas, J., Gil, J., Jenuwein, T., Garcı́a de Herreros, A., and Peiró, S. (2013). Regulation of heterochromatin transcription by Snail1/LOXL2 during epithelial-to-mesenchymal transition. Mol. Cell 52, 746–757. Mingot, J.M., Vega, S., Maestro, B., Sanz, J.M., and Nieto, M.A. (2009). Characterization of Snail nuclear import pathways as representatives of C2H2 zinc finger transcription factors. J. Cell Sci. 122, 1452–1460. Mingot, J.M., Vega, S., Cano, A., Portillo, F., and Nieto, M.A. (2013). eEF1A mediates the nuclear export of SNAG-containing proteins via the Exportin5aminoacyl-tRNA complex. Cell Rep. 5, 727–737. Montgomery, R.L., Yu, G., Latimer, P.A., Stack, C., Robinson, K., Dalby, C.M., Kaminski, N., and van Rooij, E. (2014). MicroRNA mimicry blocks pulmonary fibrosis. EMBO Mol. Med. 6, 1347–1356. Moody, S.E., Perez, D., Pan, T.C., Sarkisian, C.J., Portocarrero, C.P., Sterner, C.J., Notorfrancesco, K.L., Cardiff, R.D., and Chodosh, L.A. (2005). The transcriptional repressor Snail promotes mammary tumor recurrence. Cancer Cell 8, 197–209. Morel, A.P., Lièvre, M., Thomas, C., Hinkal, G., Ansieau, S., and Puisieux, A. (2008). Generation of breast cancer stem cells through epithelial-mesenchymal transition. PLoS ONE 3, e2888. Nunan, R., Campbell, J., Mori, R., Pitulescu, M.E., Jiang, W.G., Harding, K.G., Adams, R.H., Nobes, C.D., and Martin, P. (2015). Ephrin-Bs Drive Junctional Downregulation and Actin Stress Fiber Disassembly to Enable Wound Reepithelialization. Cell Rep. 13, 1380–1395. Oba, S., Kumano, S., Suzuki, E., Nishimatsu, H., Takahashi, M., Takamori, H., Kasuya, M., Ogawa, Y., Sato, K., Kimura, K., et al. (2010). miR-200b precursor can ameliorate renal tubulointerstitial fibrosis. PLoS ONE 5, e13614. Ocaña, O.H., Córcoles, R., Fabra, A., Moreno-Bueno, G., Acloque, H., Vega, S., Barrallo-Gimeno, A., Cano, A., and Nieto, M.A. (2012). Metastatic colonization requires the repression of the epithelial-mesenchymal transition inducer Prrx1. Cancer Cell 22, 709–724. Oda, H., Tsukita, S., and Takeichi, M. (1998). Dynamic behavior of the cadherin-based cell-cell adhesion system during Drosophila gastrulation. Dev. Biol. 203, 435–450. Okazaki, T., Chikuma, S., Iwai, Y., Fagarasan, S., and Honjo, T. (2013). A rheostat for immune responses: the unique properties of PD-1 and their advantages for clinical application. Nat. Immunol. 14, 1212–1218. Palumbo-Zerr, K., Zerr, P., Distler, A., Fliehr, J., Mancuso, R., Huang, J., Mielenz, D., Tomcik, M., Fürnrohr, B.G., Scholtysek, C., et al. (2015). Orphan nuclear receptor NR4A1 regulates transforming growth factor-b signaling and fibrosis. Nat. Med. 21, 150–158. Pandian, V., Ramraj, S., Khan, F.H., Azim, T., and Aravindan, N. (2015). Metastatic neuroblastoma cancer stem cells exhibit flexible plasticity and adaptive stemness signaling. Stem Cell Res. Ther. 6, 2. Park, K.S., and Gumbiner, B.M. (2010). Cadherin 6B induces BMP signaling and de-epithelialization during the epithelial mesenchymal transition of the neural crest. Development 137, 2691–2701. Morizane, R., Fujii, S., Monkawa, T., Hiratsuka, K., Yamaguchi, S., Homma, K., and Itoh, H. (2014). miR-34c attenuates epithelial-mesenchymal transition and kidney fibrosis with ureteral obstruction. Sci. Rep. 4, 4578. Park, J.H., Yoon, J., Lee, K.Y., and Park, B. (2015). Effects of geniposide on hepatocytes undergoing epithelial-mesenchymal transition in hepatic fibrosis by targeting TGFb/Smad and ERK-MAPK signaling pathways. Biochimie 113, 26–34. Moskowitz, I.P., Wang, J., Peterson, M.A., Pu, W.T., Mackinnon, A.C., Oxburgh, L., Chu, G.C., Sarkar, M., Berul, C., Smoot, L., et al. (2011). Transcription factor genes Smad4 and Gata4 cooperatively regulate cardiac valve development. [corrected]. Proc. Natl. Acad. Sci. USA 108, 4006–4011. Pattabiraman, D.R., Bierie, B., Kober, K.I., Thiru, P., Krall, J.A., Zill, C., Reinhardt, F., Tam, W.L., and Weinberg, R.A. (2016). Activation of PKA leads to mesenchymal-to-epithelial transition and loss of tumor-initiating ability. Science 351, aad3680. Motzer, R.J., Rini, B.I., McDermott, D.F., Redman, B.G., Kuzel, T.M., Harrison, M.R., Vaishampayan, U.N., Drabkin, H.A., George, S., Logan, T.F., et al. (2015). Nivolumab for Metastatic Renal Cell Carcinoma: Results of a Randomized Phase II Trial. J. Clin. Oncol. 33, 1430–1437. Pedroza, M., Le, T.T., Lewis, K., Karmouty-Quintana, H., To, S., George, A.T., Blackburn, M.R., Tweardy, D.J., and Agarwal, S.K. (2016). STAT-3 contributes to pulmonary fibrosis through epithelial injury and fibroblast-myofibroblast differentiation. FASEB J. 30, 129–140. 42 Cell 166, June 30, 2016 Peinado, H., Ballestar, E., Esteller, M., and Cano, A. (2004). Snail mediates E-cadherin repression by the recruitment of the Sin3A/histone deacetylase 1 (HDAC1)/HDAC2 complex. Mol. Cell. Biol. 24, 306–319. Peinado, H., Portillo, F., and Cano, A. (2005). Switching on-off Snail: LOXL2 versus GSK3beta. Cell Cycle 4, 1749–1752. Peinado, H., Olmeda, D., and Cano, A. (2007). Snail, Zeb and bHLH factors in tumour progression: an alliance against the epithelial phenotype? Nat. Rev. Cancer 7, 415–428. kovic , M., Lavotshkin, S., Matei, I., Costa-Silva, B., MorenoPeinado, H., Alec Bueno, G., Hergueta-Redondo, M., Williams, C., Garcı́a-Santos, G., Ghajar, C., et al. (2012). Melanoma exosomes educate bone marrow progenitor cells toward a pro-metastatic phenotype through MET. Nat. Med. 18, 883–891. Pérez-Pomares, J.M., and de la Pompa, J.L. (2011). Signaling during epicardium and coronary vessel development. Circ. Res. 109, 1429–1442. Plaks, V., Kong, N., and Werb, Z. (2015). The cancer stem cell niche: how essential is the niche in regulating stemness of tumor cells? Cell Stem Cell 16, 225–238. Podsypanina, K., Du, Y.C., Jechlinger, M., Beverly, L.J., Hambardzumyan, D., and Varmus, H. (2008). Seeding and propagation of untransformed mouse mammary cells in the lung. Science 321, 1841–1844. Polyak, K., and Weinberg, R.A. (2009). Transitions between epithelial and mesenchymal states: acquisition of malignant and stem cell traits. Nat. Rev. Cancer 9, 265–273. Polydorou, C., and Georgiades, P. (2013). Ets2-dependent trophoblast signalling is required for gastrulation progression after primitive streak initiation. Nat. Commun. 4, 1658. Robinson, B.D., Sica, G.L., Liu, Y.F., Rohan, T.E., Gertler, F.B., Condeelis, J.S., and Jones, J.G. (2009). Tumor microenvironment of metastasis in human breast carcinoma: a potential prognostic marker linked to hematogenous dissemination. Clin. Cancer Res. 15, 2433–2441. Rodon, J., Carducci, M.A., Sepulveda-Sánchez, J.M., Azaro, A., Calvo, E., Seoane, J., Braña, I., Sicart, E., Gueorguieva, I., Cleverly, A.L., et al. (2015). First-in-human dose study of the novel transforming growth factor-b receptor I kinase inhibitor LY2157299 monohydrate in patients with advanced cancer and glioma. Clin. Cancer Res. 21, 553–560. Rosanò, L., Spinella, F., Di Castro, V., Decandia, S., Nicotra, M.R., Natali, P.G., and Bagnato, A. (2006). Endothelin-1 is required during epithelial to mesenchymal transition in ovarian cancer progression. Exp. Biol. Med. (Maywood) 231, 1128–1131. Rowe, R.G., Lin, Y., Shimizu-Hirota, R., Hanada, S., Neilson, E.G., Greenson, J.K., and Weiss, S.J. (2011). Hepatocyte-derived Snail1 propagates liver fibrosis progression. Mol. Cell. Biol. 31, 2392–2403. Russell, R., Perkhofer, L., Liebau, S., Lin, Q., Lechel, A., Feld, F.M., Hessmann, E., Gaedcke, J., Güthle, M., Zenke, M., et al. (2015). Loss of ATM accelerates pancreatic cancer formation and epithelial-mesenchymal transition. Nat. Commun. 6, 7677. Saitoh, M., Endo, K., Furuya, S., Minami, M., Fukasawa, A., Imamura, T., and Miyazawa, K. (2016). STAT3 integrates cooperative Ras and TGF-beta signals that induce Snail expression. Oncogene. 35, 1049–1057. Sakabe, M., Kokubo, H., Nakajima, Y., and Saga, Y. (2012). Ectopic retinoic acid signaling affects outflow tract cushion development through suppression of the myocardial Tbx2-Tgfb2 pathway. Development 139, 385–395. Pottier, N., Cauffiez, C., Perrais, M., Barbry, P., and Mari, B. (2014). FibromiRs: translating molecular discoveries into new anti-fibrotic drugs. Trends Pharmacol. Sci. 35, 119–126. Samavarchi-Tehrani, P., Golipour, A., David, L., Sung, H.K., Beyer, T.A., Datti, A., Woltjen, K., Nagy, A., and Wrana, J.L. (2010). Functional genomics reveals a BMP-driven mesenchymal-to-epithelial transition in the initiation of somatic cell reprogramming. Cell Stem Cell 7, 64–77. Prat, A., Parker, J.S., Karginova, O., Fan, C., Livasy, C., Herschkowitz, J.I., He, X., and Perou, C.M. (2010). Phenotypic and molecular characterization of the claudin-low intrinsic subtype of breast cancer. Breast Cancer Res. 12, R68. Sarrió, D., Rodriguez-Pinilla, S.M., Hardisson, D., Cano, A., Moreno-Bueno, G., and Palacios, J. (2008). Epithelial-mesenchymal transition in breast cancer relates to the basal-like phenotype. Cancer Res. 68, 989–997. Puisieux, A., Brabletz, T., and Caramel, J. (2014). Oncogenic roles of EMTinducing transcription factors. Nat. Cell Biol. 16, 488–494. Puls, L.N., Eadens, M., and Messersmith, W. (2011). Current status of SRC inhibitors in solid tumor malignancies. Oncologist 16, 566–578. Sato, M., Muragaki, Y., Saika, S., Roberts, A.B., and Ooshima, A. (2003). Targeted disruption of TGF-beta1/Smad3 signaling protects against renal tubulointerstitial fibrosis induced by unilateral ureteral obstruction. J. Clin. Invest. 112, 1486–1494. Qi, Q., Li, D.Y., Luo, H.R., Guan, K.L., and Ye, K. (2015). Netrin-1 exerts oncogenic activities through enhancing Yes-associated protein stability. Proc. Natl. Acad. Sci. USA 112, 7255–7260. Saunders, L.R., and McClay, D.R. (2014). Sub-circuits of a gene regulatory network control a developmental epithelial-mesenchymal transition. Development 141, 1503–1513. Qian, B.Z., Zhang, H., Li, J., He, T., Yeo, E.J., Soong, D.Y., Carragher, N.O., Munro, A., Chang, A., Bresnick, A.R., et al. (2015). FLT1 signaling in metastasis-associated macrophages activates an inflammatory signature that promotes breast cancer metastasis. J. Exp. Med. 212, 1433–1448. Scarpa, E., Szabó, A., Bibonne, A., Theveneau, E., Parsons, M., and Mayor, R. (2015). Cadherin Switch during EMT in Neural Crest Cells Leads to Contact Inhibition of Locomotion via Repolarization of Forces. Dev. Cell 34, 421–434. Qian, L.W., Fourcaudot, A.B., Yamane, K., You, T., Chan, R.K., and Leung, K.P. (2016). Exacerbated and prolonged inflammation impairs wound healing and increases scarring. Wound Repair Regen. 24, 26–34. Schäfer, G., Narasimha, M., Vogelsang, E., and Leptin, M. (2014). Cadherin switching during the formation and differentiation of the Drosophila mesoderm - implications for epithelial-to-mesenchymal transitions. J. Cell Sci. 127, 1511–1522. Rada-Iglesias, A., Bajpai, R., Prescott, S., Brugmann, S.A., Swigut, T., and Wysocka, J. (2012). Epigenomic annotation of enhancers predicts transcriptional regulators of human neural crest. Cell Stem Cell 11, 633–648. Schiffmacher, A.T., Padmanabhan, R., Jhingory, S., and Taneyhill, L.A. (2014). Cadherin-6B is proteolytically processed during epithelial-to-mesenchymal transitions of the cranial neural crest. Mol. Biol. Cell 25, 41–54. Rembold, M., Ciglar, L., Yáñez-Cuna, J.O., Zinzen, R.P., Girardot, C., Jain, A., Welte, M.A., Stark, A., Leptin, M., and Furlong, E.E. (2014). A conserved role for Snail as a potentiator of active transcription. Genes Dev. 28, 167–181. Schmidt, J.M., Panzilius, E., Bartsch, H.S., Irmler, M., Beckers, J., Kari, V., Linnemann, J.R., Dragoi, D., Hirschi, B., Kloos, U.J., et al. (2015). Stem-celllike properties and epithelial plasticity arise as stable traits after transient Twist1 activation. Cell Rep. 10, 131–139. Rhim, A.D., Mirek, E.T., Aiello, N.M., Maitra, A., Bailey, J.M., McAllister, F., Reichert, M., Beatty, G.L., Rustgi, A.K., Vonderheide, R.H., et al. (2012). EMT and dissemination precede pancreatic tumor formation. Cell 148, 349–361. Rizvi, N.A., Mazières, J., Planchard, D., Stinchcombe, T.E., Dy, G.K., Antonia, S.J., Horn, L., Lena, H., Minenza, E., Mennecier, B., et al. (2015). Activity and safety of nivolumab, an anti-PD-1 immune checkpoint inhibitor, for patients with advanced, refractory squamous non-small-cell lung cancer (CheckMate 063): a phase 2, single-arm trial. Lancet Oncol. 16, 257–265. Scholten, D., Osterreicher, C.H., Scholten, A., Iwaisako, K., Gu, G., Brenner, D.A., and Kisseleva, T. (2010). Genetic labeling does not detect epithelial-tomesenchymal transition of cholangiocytes in liver fibrosis in mice. Gastroenterology 139, 987–998. Schwitalla, S., Fingerle, A.A., Cammareri, P., Nebelsiek, T., Göktuna, S.I., Ziegler, P.K., Canli, O., Heijmans, J., Huels, D.J., Moreaux, G., et al. (2013). Intestinal tumorigenesis initiated by dedifferentiation and acquisition of stem-cell-like properties. Cell 152, 25–38. Cell 166, June 30, 2016 43 Shamir, E.R., and Ewald, A.J. (2015). Adhesion in mammary development: novel roles for E-cadherin in individual and collective cell migration. Curr. Top. Dev. Biol. 112, 353–382. Shamir, E.R., Pappalardo, E., Jorgens, D.M., Coutinho, K., Tsai, W.T., Aziz, K., Auer, M., Tran, P.T., Bader, J.S., and Ewald, A.J. (2014). Twist1-induced dissemination preserves epithelial identity and requires E-cadherin. J. Cell Biol. 204, 839–856. Shaul, Y.D., Freinkman, E., Comb, W.C., Cantor, J.R., Tam, W.L., Thiru, P., Kim, D., Kanarek, N., Pacold, M.E., Chen, W.W., et al. (2014). Dihydropyrimidine accumulation is required for the epithelial-mesenchymal transition. Cell 158, 1094–1109. Shaw, T.J., and Martin, P. (2016). Wound repair: a showcase for cell plasticity and migration. Curr. Opin. Cell Biol. 42, 29–37. Sheridan, C. (2013). First Axl inhibitor enters clinical trials. Nat. Biotechnol. 31, 775–776. Shimono, Y., Zabala, M., Cho, R.W., Lobo, N., Dalerba, P., Qian, D., Diehn, M., Liu, H., Panula, S.P., Chiao, E., et al. (2009). Downregulation of miRNA-200c links breast cancer stem cells with normal stem cells. Cell 138, 592–603. Tang, O., Chen, X.M., Shen, S., Hahn, M., and Pollock, C.A. (2013). MiRNA200b represses transforming growth factor-b1-induced EMT and fibronectin expression in kidney proximal tubular cells. Am. J. Physiol. Renal Physiol. 304, F1266–F1273. Tang, Q., Zhong, H., Xie, F., Xie, J., Chen, H., and Yao, G. (2014). Expression of miR-106b-25 induced by salvianolic acid B inhibits epithelial-to-mesenchymal transition in HK-2 cells. Eur. J. Pharmacol. 741, 97–103. Tarin, D., Thompson, E.W., and Newgreen, D.F. (2005). The fallacy of epithelial mesenchymal transition in neoplasia. Cancer Res. 65, 5996–6001. Taura, K., Miura, K., Iwaisako, K., Osterreicher, C.H., Kodama, Y., Penz-Osterreicher, M., and Brenner, D.A. (2010). Hepatocytes do not undergo epithelialmesenchymal transition in liver fibrosis in mice. Hepatology 51, 1027–1036. Terao, M., Ishikawa, A., Nakahara, S., Kimura, A., Kato, A., Moriwaki, K., Kamada, Y., Murota, H., Taniguchi, N., Katayama, I., and Miyoshi, E. (2011). Enhanced epithelial-mesenchymal transition-like phenotype in N-acetylglucosaminyltransferase V transgenic mouse skin promotes wound healing. J. Biol. Chem. 286, 28303–28311. Theveneau, E., and Mayor, R. (2011). Can mesenchymal cells undergo collective cell migration? The case of the neural crest. Cell Adhes. Migr. 5, 490–498. Shoval, I., Ludwig, A., and Kalcheim, C. (2007). Antagonistic roles of full-length N-cadherin and its soluble BMP cleavage product in neural crest delamination. Development 134, 491–501. Thiery, J.P. (2002). Epithelial-mesenchymal transitions in tumour progression. Nat. Rev. Cancer 2, 442–454. Simões-Costa, M., and Bronner, M.E. (2015). Establishing neural crest identity: a gene regulatory recipe. Development 142, 242–257. Thiery, J.P., and Lim, C.T. (2013). Tumor dissemination: an EMT affair. Cancer Cell 23, 272–273. Singh, A., and Settleman, J. (2010). EMT, cancer stem cells and drug resistance: an emerging axis of evil in the war on cancer. Oncogene 29, 4741–4751. Thiery, J.P., and Sleeman, J.P. (2006). Complex networks orchestrate epithelial-mesenchymal transitions. Nat. Rev. Mol. Cell Biol. 7, 131–142. Smerage, J.B., Barlow, W.E., Hortobagyi, G.N., Winer, E.P., Leyland-Jones, B., Srkalovic, G., Tejwani, S., Schott, A.F., O’Rourke, M.A., Lew, D.L., et al. (2014). Circulating tumor cells and response to chemotherapy in metastatic breast cancer: SWOG S0500. J. Clin. Oncol. 32, 3483–3489. Thiery, J.P., Acloque, H., Huang, R.Y., and Nieto, M.A. (2009). Epithelialmesenchymal transitions in development and disease. Cell 139, 871–890. Sosa, M.S., Parikh, F., Maia, A.G., Estrada, Y., Bosch, A., Bragado, P., Ekpin, E., George, A., Zheng, Y., Lam, H.M., et al. (2015). NR2F1 controls tumour cell dormancy via SOX9- and RARb-driven quiescence programmes. Nat. Commun. 6, 6170. Spaderna, S., Schmalhofer, O., Wahlbuhl, M., Dimmler, A., Bauer, K., Sultan, A., Hlubek, F., Jung, A., Strand, D., Eger, A., et al. (2008). The transcriptional repressor ZEB1 promotes metastasis and loss of cell polarity in cancer. Cancer Res. 68, 537–544. Stankic, M., Pavlovic, S., Chin, Y., Brogi, E., Padua, D., Norton, L., Massagué, J., and Benezra, R. (2013). TGF-b-Id1 signaling opposes Twist1 and promotes metastatic colonization via a mesenchymal-to-epithelial transition. Cell Rep. 5, 1228–1242. Sugimoto, H., LeBleu, V.S., Bosukonda, D., Keck, P., Taduri, G., Bechtel, W., Okada, H., Carlson, W., Jr., Bey, P., Rusckowski, M., et al. (2012). Activin-like kinase 3 is important for kidney regeneration and reversal of fibrosis. Nat. Med. 18, 396–404. Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Sasaki, A., Yamamoto, M., Nakamura, M., Sutou, K., Osafune, K., and Yamanaka, S. (2014). Induction of pluripotency in human somatic cells via a transient state resembling primitive streak-like mesendoderm. Nat. Commun. 5, 3678. Takeichi, M. (1988). The cadherins: cell-cell adhesion molecules controlling animal morphogenesis. Development 102, 639–655. Tam, W.L., and Weinberg, R.A. (2013). The epigenetics of epithelial-mesenchymal plasticity in cancer. Nat. Med. 19, 1438–1449. Tan, T.Z., Miow, Q.H., Huang, R.Y., Wong, M.K., Ye, J., Lau, J.A., Wu, M.C., Bin Abdul Hadi, L.H., Soong, R., Choolani, M., et al. (2013). Functional genomics identifies five distinct molecular subtypes with clinical relevance and pathways for growth control in epithelial ovarian cancer. EMBO Mol. Med. 5, 983–998. Tan, T.Z., Miow, Q.H., Miki, Y., Noda, T., Mori, S., Huang, R.Y., and Thiery, J.P. (2014). Epithelial-mesenchymal transition spectrum quantification and its efficacy in deciphering survival and drug responses of cancer patients. EMBO Mol. Med. 6, 1279–1293. 44 Cell 166, June 30, 2016 Tian, X.J., Zhang, H., and Xing, J. (2013). Coupled reversible and irreversible bistable switches underlying TGFb-induced epithelial to mesenchymal transition. Biophys. J. 105, 1079–1089. Ting, S.B., Caddy, J., Hislop, N., Wilanowski, T., Auden, A., Zhao, L.L., Ellis, S., Kaur, P., Uchida, Y., Holleran, W.M., et al. (2005). A homolog of Drosophila grainy head is essential for epidermal integrity in mice. Science 308, 411–413. Ting, D.T., Wittner, B.S., Ligorio, M., Vincent Jordan, N., Shah, A.M., Miyamoto, D.T., Aceto, N., Bersani, F., Brannigan, B.W., Xega, K., et al. (2014). Single-cell RNA sequencing identifies extracellular matrix gene expression by pancreatic circulating tumor cells. Cell Rep. 8, 1905–1918. Topalian, S.L., Hodi, F.S., Brahmer, J.R., Gettinger, S.N., Smith, D.C., McDermott, D.F., Powderly, J.D., Carvajal, R.D., Sosman, J.A., Atkins, M.B., et al. (2012). Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N. Engl. J. Med. 366, 2443–2454. Tran, H.D., Luitel, K., Kim, M., Zhang, K., Longmore, G.D., and Tran, D.D. (2014). Transient SNAIL1 expression is necessary for metastatic competence in breast cancer. Cancer Res. 74, 6330–6340. Trimboli, A.J., Fukino, K., de Bruin, A., Wei, G., Shen, L., Tanner, S.M., Creasap, N., Rosol, T.J., Robinson, M.L., Eng, C., et al. (2008). Direct evidence for epithelial-mesenchymal transitions in breast cancer. Cancer Res. 68, 937–945. Tsai, J.H., Donaher, J.L., Murphy, D.A., Chau, S., and Yang, J. (2012). Spatiotemporal regulation of epithelial-mesenchymal transition is essential for squamous cell carcinoma metastasis. Cancer Cell 22, 725–736. Tsuji, T., Ibaragi, S., and Hu, G.F. (2009). Epithelial-mesenchymal transition and cell cooperativity in metastasis. Cancer Res. 69, 7135–7139. Ubil, E., Duan, J., Pillai, I.C., Rosa-Garrido, M., Wu, Y., Bargiacchi, F., Lu, Y., Stanbouly, S., Huang, J., Rojas, M., et al. (2014). Mesenchymal-endothelial transition contributes to cardiac neovascularization. Nature 514, 585–590. Vaughan, A.E., and Chapman, H.A. (2013). Regenerative activity of the lung after epithelial injury. Biochim. Biophys. Acta 1832, 922–930. Vega, S., Morales, A.V., Ocaña, O.H., Valdés, F., Fabregat, I., and Nieto, M.A. (2004). Snail blocks the cell cycle and confers resistance to cell death. Genes Dev. 18, 1131–1143. von Gise, A., and Pu, W.T. (2012). Endocardial and epicardial epithelial to mesenchymal transitions in heart development and disease. Circ. Res. 110, 1628–1645. Yang, M.C., Wang, C.J., Liao, P.C., Yen, C.J., and Shan, Y.S. (2014). Hepatic stellate cells secretes type I collagen to trigger epithelial mesenchymal transition of hepatoma cells. Am. J. Cancer Res. 4, 751–763. Voon, D.C., Wang, H., Koo, J.K., Chai, J.H., Hor, Y.T., Tan, T.Z., Chu, Y.S., Mori, S., and Ito, Y. (2013). EMT-induced stemness and tumorigenicity are fueled by the EGFR/Ras pathway. PLoS ONE 8, e70427. Yang, H., Lu, X., Liu, Z., Chen, L., Xu, Y., Wang, Y., Wei, G., and Chen, Y. (2015). FBXW7 suppresses epithelial-mesenchymal transition, stemness and metastatic potential of cholangiocarcinoma cells. Oncotarget 6, 6310–6325. Vultur, A., Buettner, R., Kowolik, C., Liang, W., Smith, D., Boschelli, F., and Jove, R. (2008). SKI-606 (bosutinib), a novel Src kinase inhibitor, suppresses migration and invasion of human breast cancer cells. Mol. Cancer Ther. 7, 1185–1194. Ye, X., and Weinberg, R.A. (2015). Epithelial-Mesenchymal Plasticity: A Central Regulator of Cancer Progression. Trends Cell Biol. 25, 675–686. Wagenblast, E., Soto, M., Gutiérrez-Ángel, S., Hartl, C.A., Gable, A.L., Maceli, A.R., Erard, N., Williams, A.M., Kim, S.Y., Dickopf, S., et al. (2015). A model of breast cancer heterogeneity reveals vascular mimicry as a driver of metastasis. Nature 520, 358–362. Wang, J., Scully, K., Zhu, X., Cai, L., Zhang, J., Prefontaine, G.G., Krones, A., Ohgi, K.A., Zhu, P., Garcia-Bassets, I., et al. (2007). Opposing LSD1 complexes function in developmental gene activation and repression programmes. Nature 446, 882–887. Wang, C., Song, X., Li, Y., Han, F., Gao, S., Wang, X., Xie, S., and Lv, C. (2013). Low-dose paclitaxel ameliorates pulmonary fibrosis by suppressing TGF-b1/ Smad3 pathway via miR-140 upregulation. PLoS ONE 8, e70725. Wang, J., Zhu, X., Hu, J., He, G., Li, X., Wu, P., Ren, X., Wang, F., Liao, W., Liang, L., and Ding, Y. (2015). The positive feedback between Snail and DAB2IP regulates EMT, invasion and metastasis in colorectal cancer. Oncotarget 6, 27427–27439. Warzecha, C.C., and Carstens, R.P. (2012). Complex changes in alternative pre-mRNA splicing play a central role in the epithelial-to-mesenchymal transition (EMT). Semin. Cancer Biol. 22, 417–427. Warzecha, C.C., Sato, T.K., Nabet, B., Hogenesch, J.B., and Carstens, R.P. (2009). ESRP1 and ESRP2 are epithelial cell-type-specific regulators of FGFR2 splicing. Mol. Cell 33, 591–601. Watanabe, K., Villarreal-Ponce, A., Sun, P., Salmans, M.L., Fallahi, M., Andersen, B., and Dai, X. (2014). Mammary morphogenesis and regeneration require the inhibition of EMT at terminal end buds by Ovol2 transcriptional repressor. Dev. Cell 29, 59–74. Ye, X., Tam, W.L., Shibue, T., Kaygusuz, Y., Reinhardt, F., Ng Eaton, E., and Weinberg, R.A. (2015). Distinct EMT programs control normal mammary stem cells and tumour-initiating cells. Nature 525, 256–260. Yu, M., Smolen, G.A., Zhang, J., Wittner, B., Schott, B.J., Brachtel, E., Ramaswamy, S., Maheswaran, S., and Haber, D.A. (2009). A developmentally regulated inducer of EMT, LBX1, contributes to breast cancer progression. Genes Dev. 23, 1737–1742. Yu, M., Bardia, A., Wittner, B.S., Stott, S.L., Smas, M.E., Ting, D.T., Isakoff, S.J., Ciciliano, J.C., Wells, M.N., Shah, A.M., et al. (2013). Circulating breast tumor cells exhibit dynamic changes in epithelial and mesenchymal composition. Science 339, 580–584. Zadran, S., Arumugam, R., Herschman, H., Phelps, M.E., and Levine, R.D. (2014). Surprisal analysis characterizes the free energy time course of cancer cells undergoing epithelial-to-mesenchymal transition. Proc. Natl. Acad. Sci. USA 111, 13235–13240. Zeisberg, M., Hanai, J., Sugimoto, H., Mammoto, T., Charytan, D., Strutz, F., and Kalluri, R. (2003). BMP-7 counteracts TGF-beta1-induced epithelialto-mesenchymal transition and reverses chronic renal injury. Nat. Med. 9, 964–968. Zeisberg, E.M., Tarnavski, O., Zeisberg, M., Dorfman, A.L., McMullen, J.R., Gustafsson, E., Chandraker, A., Yuan, X., Pu, W.T., Roberts, A.B., et al. (2007a). Endothelial-to-mesenchymal transition contributes to cardiac fibrosis. Nat. Med. 13, 952–961. Zeisberg, M., Yang, C., Martino, M., Duncan, M.B., Rieder, F., Tanjore, H., and Kalluri, R. (2007b). Fibroblasts derive from hepatocytes in liver fibrosis via epithelial to mesenchymal transition. J. Biol. Chem. 282, 23337–23347. Wellner, U., Schubert, J., Burk, U.C., Schmalhofer, O., Zhu, F., Sonntag, A., Waldvogel, B., Vannier, C., Darling, D., zur Hausen, A., et al. (2009). The EMT-activator ZEB1 promotes tumorigenicity by repressing stemness-inhibiting microRNAs. Nat. Cell Biol. 11, 1487–1495. Zhang, K., Rodriguez-Aznar, E., Yabuta, N., Owen, R.J., Mingot, J.M., Nojima, H., Nieto, M.A., and Longmore, G.D. (2012a). Lats2 kinase potentiates Snail1 activity by promoting nuclear retention upon phosphorylation. EMBO J. 31, 29–43. Weston, J.A., and Thiery, J.P. (2015). Pentimento: Neural Crest and the origin of mesectoderm. Dev. Biol. 401, 37–61. Zhang, Y., Wang, Z., Yu, J., Shi, Jz., Wang, C., Fu, Wh., Chen, Zw., and Yang, J. (2012b). Cancer stem-like cells contribute to cisplatin resistance and progression in bladder cancer. Cancer Lett. 322, 70–77. Wettstein, G., Bellaye, P.S., Kolb, M., Hammann, A., Crestani, B., Soler, P., Marchal-Somme, J., Hazoume, A., Gauldie, J., Gunther, A., et al. (2013). Inhibition of HSP27 blocks fibrosis development and EMT features by promoting Snail degradation. FASEB J. 27, 1549–1560. Wu, C.Y., Tsai, Y.P., Wu, M.Z., Teng, S.C., and Wu, K.J. (2012). Epigenetic reprogramming and post-transcriptional regulation during the epithelialmesenchymal transition. Trends Genet. 28, 454–463. Wyckoff, J., Wang, W., Lin, E.Y., Wang, Y., Pixley, F., Stanley, E.R., Graf, T., Pollard, J.W., Segall, J., and Condeelis, J. (2004). A paracrine loop between tumor cells and macrophages is required for tumor cell migration in mammary tumors. Cancer Res. 64, 7022–7029. Zhang, J., Tian, X.J., Zhang, H., Teng, Y., Li, R., Bai, F., Elankumaran, S., and Xing, J. (2014). TGF-b-induced epithelial-to-mesenchymal transition proceeds through stepwise activation of multiple feedback loops. Sci. Signal. 7, ra91. Zhao, J., Tang, N., Wu, K., Dai, W., Ye, C., Shi, J., Zhang, J., Ning, B., Zeng, X., and Lin, Y. (2014). MiR-21 simultaneously regulates ERK1 signaling in HSC activation and hepatocyte EMT in hepatic fibrosis. PLoS ONE 9, e108005. Zhao, S., Zhang, Y., Zheng, X., Tu, X., Li, H., Chen, J., Zang, Y., and Zhang, J. (2015). Loss of MicroRNA-101 Promotes Epithelial to Mesenchymal Transition in Hepatocytes. J. Cell. Physiol. 230, 2706–2717. Xiao, W., Chen, X., Li, W., Ye, S., Wang, W., Luo, L., and Liu, Y. (2015). Quantitative analysis of injury-induced anterior subcapsular cataract in the mouse: a model of lens epithelial cells proliferation and epithelial-mesenchymal transition. Sci. Rep. 5, 8362. Zheng, X., Carstens, J.L., Kim, J., Scheible, M., Kaye, J., Sugimoto, H., Wu, C.C., LeBleu, V.S., and Kalluri, R. (2015). Epithelial-to-mesenchymal transition is dispensable for metastasis but induces chemoresistance in pancreatic cancer. Nature 527, 525–530. Yang, F., Huang, X.R., Chung, A.C., Hou, C.C., Lai, K.N., and Lan, H.Y. (2010a). Essential role for Smad3 in angiotensin II-induced tubular epithelial-mesenchymal transition. J. Pathol. 221, 390–401. Zhou, B.P., Deng, J., Xia, W., Xu, J., Li, Y.M., Gunduz, M., and Hung, M.C. (2004). Dual regulation of Snail by GSK-3beta-mediated phosphorylation in control of epithelial-mesenchymal transition. Nat. Cell Biol. 6, 931–940. Yang, X., Lin, X., Zhong, X., Kaur, S., Li, N., Liang, S., Lassus, H., Wang, L., Katsaros, D., Montone, K., et al. (2010b). Double-negative feedback loop between reprogramming factor LIN28 and microRNA let-7 regulates aldehyde dehydrogenase 1-positive cancer stem cells. Cancer Res. 70, 9463–9472. Zhou, F., Drabsch, Y., Dekker, T.J., de Vinuesa, A.G., Li, Y., Hawinkels, L.J., Sheppard, K.A., Goumans, M.J., Luwor, R.B., de Vries, C.J., et al. (2014). Nuclear receptor NR4A1 promotes breast cancer invasion and metastasis by activating TGF-b signalling. Nat. Commun. 5, 3388. Cell 166, June 30, 2016 45