Principles of stem cell biology and cancer: future applications and therapeutics. Edited by T. Regad, T. J. Sayers and R. C. Rees. John Wiley & Sons (2015)
Part II. Cancer stem cells
The majority of the carcinoma cells present within a growing tumour exhibit primarily epithelial characteristics and properties, including robust expression of E-cadherin, high levels of proliferation and limited motility (Scheel and Weinberg, 2012; Tam and Weinberg, 2013). However, some cancer cells develop the ability to metastasize, resulting in the growth of tumours at locations distant from the primary tumour. For these cells to invade local tissue, disseminate to distant organs and tissues through the systemic circulation and eventually give rise to metastatic colonies, they must undergo a phenotypic and functional shift towards a more motile, invasive mesenchymal cell (Chaffer and Weinberg, 2011; Scheel and Weinberg, 2012). Activation of EMT is the mechanism by which cancer cells acquire these attributes. Strong concordance exists between the mesenchymal phenotype of cancer cells and the phenotypic characteristics required by cancer cells for metastasis. During EMT, epithelial cancer cells lose their differentiated epithelial characteristics, including intercellular adhesion, polarity and nonmotility, and acquire mesenchymal traits, including motility, invasiveness, resistance to apoptosis and resistance to treatment (Nieto, 2013; Tam and Weinberg, 2013). Prototypic features of EMT in the context of cancer are similar to those critical for EMT in development: loss of epithelial cell polarity and cell – cell contacts; loss of the epithelial marker E-cadherin; acquisition of mesenchymal markers such as vimentin, N-cadherin and SMA; upregulation of cell migration and invasion; and resistance to treatment (Tam and Weinberg, 2013; Tsai and Yang, 2013). These changes result in a more aggressive, metastatic phenotype. Thus, EMT can provide epithelial cells with the critical traits required for metastasis (Nieto, 2013). This section discusses the role of EMT and MET in the various steps of metastasis.
6.3.1. Mesenchymal conversion and local invasion
Local invasion of the surrounding ECM is one of the earliest steps in the process of metastasis and is an important determinant of the metastatic potential of cancer cells. In order to invade local tissue, epithelial cancer cells must undergo adaptive structural changes that allow them to simultaneously loosen cell – cell adhesions, degrade and remodel the ECM and become motile (Nieto, 2013; Tsai and Yang, 2013) (Figures 6.2 and 6.3). EMT core regulators such as Snail1 repress the expression of E-cadherin and other junctional proteins and simultaneously induce the expression of soluble and membrane-tethered (MT) matrix metalloproteases (e.g. MT1-MMP, MT2-MMP, MMP2, MMP9) to digest or degrade the basement membrane (Tsai and Yang, 2013). In addition to their degradative properties, these proteases may also be involved in regulating the disruption of cell – cell junctional complexes, suggesting that there may be coordinated induction of proteases and loss of cellular junctions in the promotion of the invasion process (Tsai and Yang, 2013).
In addition to MMP activation, emerging evidence suggests that the activation of EMT TFs by inducers of EMT can result in the formation of invadopodia, specialized actin-based cellular protrusions capable of recruiting proteases to sites of cell – ECM contact in order to focus matrix breakdown at specific locations (Murphy and Courtneidge, 2011; Tsai and Yang, 2013) (Figure 6.2). For example, growth factor signalling coupled with the activation of Twist is reported to promote the formation of invadopodia in cultured cancer cells, as is Zeppo1, a novel promoter of metastasis that represses E-cadherin expression (Tsai and Yang, 2013). Collectively, the repression of epithelial traits and upregulation of mesenchymal properties results in cells being permissive for invasion.
Figure 6.2. Role of the EMT programme in local invasion. Stimuli present in the tumour microenvironment, including TGF-β produced by CAFs and hypoxia, activate EMT in epithelial cancer cells. Activation of EMT results in the repression of E-cadherin (a hallmark of the cellular programme), acquisition of mesenchymal phenotypic traits and activation of ECM degradative enzymes. EMT activation also induces the formation of invadopodia, to allow for focused matrix breakdown. These events culminate in increased migration and invasion of tumour cells, allowing them to invade the surrounding tissue and intravasate into the vasculature.
Figure 6.3. Role of EMT in metastasis cascade, epithelial – mesenchymal plasticity and CSCs. Activation of EMT by paracrine signals present in the tumour microenvironment induces a phenotypic and behavioural shift in epithelial cancer cells towards a more mesenchymal cell, including increased migratory, invasive and survival properties that allow for local invasion. Individual cells advance to different points along the EMT pathway, with some cells undergoing ‘partial EMT’ and some undergoing ‘complete EMT’, yielding cells with different metastatic propensities. Cancer cells that have undergone EMT exhibit hallmarks of CSCs and maintain their mesenchymal phenotype through both paracrine and autocrine signalling. Mesenchymal CSCs invade local tissue and intravasate into vessels to disseminate into the systemic circulation. Survival of these CTCs is critical for metastasis and is mediated, in part, through maintenance of the EMT programme by platelet-derived paracrine signalling and adhesion. Successful CTCs may then extravasate at distant sites and establish in the tissue stroma, a process that is in?uenced by cellular plasticity and requires the ability of mesenchymal cells (CSCs) to revert to an epithelial state through MET. Proliferation of the established tumour cells results in the formation of clinically relevant macrometastases.
6.3.2. Tumour cell intravasation and transport within the systemic circulation
Subsequent to local invasion, tumour cells must intravasate into the vasculature to allow for systemic dissemination by the vascular system (Chaffer and Weinberg, 2011) (Figures 6.2 and 6.3). The modulation of properties acquired by cancer cells during the activation of EMT, especially those involved in migration and invasion, likely plays a role in the transit of invasive tumour cells across the endothelial cell barrier. However, the detailed mechanisms involved in the process of intravasation by tumour cells that have undergone EMT are still being delineated (Tsai and Yang, 2013). Recent studies have shown that the expression of Zeb1 in PC-3 prostate cancer cells is required for increased migration through the endothelial cell barrier, while the overexpression of Snail1 has been reported to promote intravasation through activation of membrane-bound but not secreted MMPs, providing initial evidence of a role for EMT in this process (Tsai and Yang, 2013). Indeed, EMT may be an enabling force, providing cells with the tools required to intravasate.
Invading tumour cells that successfully intravasate and enter the systemic circulation are termed ‘CTCs’ (Ksiazkiewicz et al., 2012; Kang and Pantel, 2013; Krebs et al., 2014). Their presence in the circulation represents the mechanism for the transport of cancer cells with metastatic potential to distant sites, where they may disseminate and eventually form metastases. A critical requirement of CTCs is their ability to resist anoikis and survive in the circulation (Krebs et al., 2014). Maintenance of the EMT programme by CTCs is thought to help provide this capacity.
Several recent studies have demonstrated the presence of mesenchymal markers on CTCs, both in preclinical models of cancer and in human cancer patients (Tsai and Yang, 2013), strongly suggesting that CTCs have undergone EMT. Clinical studies have shown that the presence of increased numbers of CTCs is an indicator of poor prognosis and of distant relapse (Tsai and Yang, 2013; Krebs et al., 2014). The presence of mesenchymal CTCs has also been correlated with resistance to treatment. Importantly, many of the technologies presently available for the detection and isolation of CTCs rely on epithelial markers, thus missing the subset of mesenchymal cells – a fact that has resulted in the development of assays capable of detecting mesenchymal CTCs (Krebs et al., 2014). Recent interrogation of CTCs in patients, particularly when through non-antigen-dependent techniques, has demonstrated clear heterogeneity with respect to EMT markers, further suggesting the existence of a continuum of EMT phenotypes (Krebs et al., 2014).
The maintenance of a mesenchymal state by CTCs in the systemic circulation may promote survival, an enabling characteristic of tumour cell dissemination. In particular, EMT may be important in preventing single tumour cells from suffering from detachment-induced anoikis by promoting CTC aggregation to leukocytes and platelets in the circulation (Tsai and Yang, 2013; Krebs et al., 2014). Indeed, platelets may be important for the maintenance of the mesenchymal phenotype of CTCs (Figure 6.3). In addition to providing CTCs with physical protection in the circulation, platelets may promote EMT in these cells through TGF-β and NF-kB signalling and thus prime them for metastasis (Tsai and Yang, 2013; Krebs et al., 2014). Breast cancer CTCs enriched in TGF-β-pathway components have been found clustered to platelets, a major source of the EMT inducer TGF-β, and depletion of platlet-derived TGF-β reduces distant metastasis, suggesting that maintenance of a mesenchymal state by CTCs may be important for metastasis.
6.3.3. Tumour cell extravasation
The exit of tumour cells from the systemic circulation through the endothelial barrier, a process called extravasation, is a critical step in the metastatic cascade (Figure 6.3). While one can imagine that this step, like that of intravasation, would require substantial modulation of migration and invasion processes, the precise role of EMT in extravasation remains to be extensively investigated. A major challenge to the study of EMT in this context, especially in vivo, is the paucity of experimental models that faithfully recapitulate physiologically relevant levels of tumour cell dissemination and extravasation, which are highly inefficient (Tsai and Yang, 2013). Most studies have relied on experimental metastasis assays, such as tail vein injection or intracardiac injection, to interrogate the mechanisms of extravasation, but these methods involve injecting very large numbers of tumour cells directly into the circulation, resulting in intravessel growth in the lung microvasculature (Tsai and Yang, 2013). Recent studies have demonstrated that tumour cells arriving at distant metastatic sites develop integrin-containing filopodium-like protrusions (FLPs) in order to interact with the surrounding ECM. FLPs can be induced by Twist1 and Snail1 expression. Their generation is correlated with the mesenchymal state of the tumour cells and their formation has been shown to be essential for successful metastasis, suggesting a role for EMT in the process of extravasation (Scheel and Weinberg, 2012; Tam and Weinberg, 2013).
6.3.4. MET and metastatic colonization
Subsequent to extravasation, disseminated tumour cells must survive in their new environment and proliferate to form first the micrometastases and then the macrometastases that are ultimately identified as clinically relevant metastatic disease (Chaffer and Weinberg, 2011; Scheel and Weinberg, 2012). The development of micro and macrometastases at distant organ sites highlights the dynamic plasticity of the EMT programme (Figure 6.3). Specifically, while locally invading cells and CTCs demonstrate many of the cellular and molecular markers of EMT, cells comprising metastatic foci are largely of epithelial phenotype. These observations have suggested that, once metastatic cells have ‘set up shop’ at a secondary site, they undergo MET and revert to epithelial cells to expand the metastasis (Scheel and Weinberg, 2012; Nieto, 2013). Recent experiments suggest the presence of epithelial – mesenchymal plasticity during metastasis. In an inducible Twist1 mouse skin tumour model, activation of EMT promoted invasion, intravasation and extravasation, while subsequent loss of the EMT-inducing signal was required for cell proliferation and formation of metastases (Scheel and Weinberg, 2012). In addition, lung metastasis formation following tail vein injection of BT-549 breast cancer cells required the downregulation of a novel EMT inducer, Prrx-1, and reversion of EMT following initial Prrx1 – Twist1 cooperation in EMT promotion (Scheel and Weinberg, 2012). The reversion to the epithelial state by tumour cells may be required for the re-engagement of the proliferative programmes necessary for the development of metastases. EMT regulators such as Snail1, Zeb2 and Twist1 suppress cell division, although the signalling pathways that couple EMT to proliferation remain to be elucidated.
The signals and switches required for the reversion of cancer cells from a mesenchymal phenotype to an epithelial state are not yet understood (Scheel and Weinberg, 2012; Tsai and Yang, 2013). For example, it is possible that the absence of an EMT-inducing signal may be sufficient for the reversion process. It is equally possible that specific MET-inducing signals may be necessary to actively engage MET. Alternatively, the balance between the removal of EMT-inducing signals coupled with the onset of MET-inducing signals may dictate the phenotypic direction of the cell. It is likely that signal inputs from the metastatic microenvironment may also help regulate the EMT – MET conversion, although how this occurs remains to be investigated.