Epithelial–mesenchymal transition (EMT) is a complex process involved in embryonic development, wound healing and carcinogenesis. During this process, epithelial cells lose their defining characteristics and acquire mesenchymal properties: loss of cell–cell adhesion; increased motility and invasiveness; resistance to apoptosis and changes in cellular morphology. EMT has been implicated as a driver of metastasis and tumour invasion, as this process allows cells to detach from their niche and migrate through blood and lymphatic vessels to invade different organs. This transition involves a diverse range of transcription factors, including Twist, Snail and ZEB1, and downstream transcriptional targets, including E-cadherin, β-catenin, fibronectin and vimentin. Recent evidence indicates that cancer stem cells are required for metastatic tumours to become established at a distant site, and that cancer cells undergoing EMT may develop stem-cell characteristics as well as increased invasive potential. The role of EMT in cancer biology is newly emerging in the human field, and to date very little has been done in veterinary medicine. EMT may therefore be an important molecular determinant of tumour metastasis, and further understanding of this process may lead to novel drug targets to be exploited in both veterinary and human medicine.
Epithelial–mesenchymal transition describes a rapid and often reversible change of cell phenotype from epithelial to mesenchymal  and can facilitate cell migration. This program is essential throughout different embryonic stages such as organogenesis and gastrulation. Cells acquiring mesenchymal characteristics are able to move towards different places in the organism in order to generate organs and their anatomical structures. [2, 3] This process is also proposed to play a role in metastasis by increasing cell motility and invasiveness. Based upon the fundamental roles that EMT has, the process has been classified as follows:
Type I Embryogenesis and organ development
Type II Fibrosis – wound healing and tissue regeneration
Type III Metastasis – enabling neoplastic cells to invade different tissues and organs during cancer progression.
EMT essentially mediates reorganisation of the cytoskeleton by decreasing cell–cell contact and changing cell polarity. Epithelial cells adhere to each other by lateral cell-cell junctions. These cells are polarized such that the bottom is defined as basal, and the top as apical. The matrix binding areas are located at the basal aspect of the epithelium. Due to these properties, epithelial cells generally form groups with tight junctions. In contrast, mesenchymal cells have a more elongated shape and their polarity is reversed, therefore they do not form large groups, and easily migrate (Fig. 1).  These changes are induced by transcriptional repression of proteins, such as cadherins, occludin, claudin and desmoplakin. This repression can be accomplished by a range of transcription factors including ZEB1, Snail, Slug and Twist. Furthermore, epithelial cells undergo cytoskeletal changes, acquiring increased motility and invasiveness. [5, 6]
If a cell is to achieve colonization of a distant site, then EMT has to be reversed. This reversion is called mesenchymal to epithelial transition (MET), and without it, cells would not be able to form any tissue or organ, including any kind of metastatic colonization. After MET, cells gain back their lost polarity and adhere to each other by tight junctions. This is due to an upregulation of proteins that help create adherens junctions such as E-cadherin and β-catenin, and the downregulation of those which can enable cells to break their adhesions and change their shape and polarity, such as fibronectin and vimentin. All three sub-types of EMT (types I–III) require the reverse MET program to enable cells to attach to their new microenvironments and establish a niche. 
EMT in cancer and during metastasis
Recent data shows that EMT is commonly observed in primary tumours at their cancer invasion front.  It is also associated with resistance to anticancer agents such as EGFR inhibitors.  According to the previously explained classification, during EMT III cancer cells follow similar pathways as normal cells during embryonic stages and wound healing. The existence of a reversible program (MET) which would enable cells to attach to a new epithelial sheet has been proposed by Thiery et al.  and Hugo et al.  Some of these cancer cells may also pass through just a partial EMT, expressing both, mesenchymal and epithelial markers during the process. 
There are three important pathways in EMT; Wnt/β-catenin, FGF and TGF-β1/BMP. These are vital for embryogenesis, but they also play important roles in tumour formation and progression. The transforming growth factor β1 (TGF-β1) pathway is crucial in EMT induction due to its multiple downstream effectors capable of repressing E-cadherin and subsequently enabling upregulation of mesenchymal promoters as fibronectin and vimentin. [13, 14]
Signalling pathways and EMT inducers
To induce EMT during cancer, there are extrinsic and intrinsic stimuli. The microenvironment of a tumour plays a very important role as an extrinsic stimulus during cancer progression and metastasis. Some EMT inducing factors can be promoted by the interaction between the microenvironment and tumour cells, for example, members of the Snail family (E-cadherin repressors) can be activated by extracellular factors. Extrinsic stimuli include activators of Wnt, Hedgehog, Notch, nuclear factor-κB (NF-κB) and TGF-β signalling pathways. Subsequently, TGF-β is a potent EMT inducer during tumour progression. [9, 15, 16] Crosstalk between these signalling pathways has been mplicated in E-cadherin repression and EMT induction. For example, Wnt and Hedgehog signalling cascades induce Snail upregulation, which subsequently represses E-cadherin. Hedgehog signals interact with Notch and TGF-β signalling pathways to induce EMT.  Wnt signalling pathway has been shown to be disrupted during different cancer stages leading to EMT induction.
TGF-β is a member of a very complex multifunctional family of cytokines which act as mediators of development stages and tissue homeostasis. However, they have paradoxical functions depending on cell type. TGF-β can act as a physiological and pathophysiological regulator through different stages of life. [18, 19] For example; it can promote apoptosis and/or cell cycle arrest in normal epithelial, endothelial and haematopoietic cells, but when there is a deregulation or a mutation in the TGF-β pathway, apoptosis can be inhibited while cells acquire tumorigenic characteristics and also become motile and invasive to the surrounding tissues.  Mutations in the TGF-β type II receptor were demonstrated by Markowitz et al.  and Myeroff et al.  in different kinds of cancer showing loss of expression of this receptor. This resistance to its anti-proliferative characteristics can contribute to the formation of different kinds of cancer. The PI3K signalling pathway (induced by TGF-β), via its downstream effector Akt, can block apoptosis in mammary epithelial cells chronically treated with TGF-β by sequestering Smad3 into the cytoplasm after insulin treatment. [23, 24] Moreover, TGF-β acts as a tumour suppressor during early stages of tumourigenesis by its growth inhibitory effects and stimulation of apoptosis. Interestingly, during late stages of tumour progression, it can also act as a tumour promoter with prometastatic effects. This was confirmed by Shipitsin et al.  who observed a high level of expression of TGF-β in CD44+/CD24− metastatic breast cancer stem cells (CSCs). In addition, following TGF-β inhibition, these cells demonstrated a more epithelial behaviour. This cytokine can stimulate two of the most important features of cancer progression, which are invasion and metastasis.  Furthermore, when TGF-β binds to its receptors (TGF-βRI and TGF-βRII), a signalling cascade is initiated by their phosphorylation, which can activate different proteins, in which the Smad family of transcription factors is one of the most notable. Specifically, when Smad2 and Smad3 bind with co-smad (Smad4) after their activation via TGF-βRI, this complexes associate with different transcription factors like ZEB proteins to repress E-cadherin during EMT. [19, 27, 28]
TGF-β also activates a range of downstream effectors, which can induce EMT. One of the most important downstream effector is NF-κB. As yet, the exact way NF-κB works is not fully understood, but one of its major roles is to suppress apoptosis. It remains unclear how NF-κB regulates EMT and metastasis, however, Huber et al.  demonstrated that EMT can be induced in Ras-transformed mammary epithelial cells by gain of NF-κB. They also demonstrated that inactivation of NF-κB causes MET, suggesting that this signalling pathway not only induces EMT but also helps to maintain it. Inhibitory proteins known as Inhibitor κB (IκB) proteins are involved in NF-κB activity regulation.  And by using these, NF-κB function has been confirmed in different model systems. Huang et al. used a mutant form of IκBα to downregulate NF-κB activity in metastatic human prostate cancer and, as a result, they observed inhibition of tumour growth, invasion and metastasis in vitro and in vivo. [29, 30] Furthermore, they also confirmed this in human melanoma cells. [29, 31] NF-κB regulates Snail expression, leading to its increased transcription. Subsequently, Snail represses Raf kinase inhibitor (RKI), which can inhibit NF-κB activity. Therefore, Snail overexpression suppresses E-cadherin activity, inducing EMT while repressing RKI. 
Genetic alterations such as mutations/perturbations can act as intrinsic stimuli of EMT during cancer progression by potentiating EMT inducers or inhibiting EMT repressors. TGF-β can act as a tumour suppressor after acquiring a specific gene mutation, but its unmutated form can function as an EMT inducer to promote invasion and metastasis.  Lehmann et al.  showed that during cancer progression, the TGF-β growth inhibitory effect can have a negative feedback and cancer cells can proliferate due to an upregulation of Ras proteins and consequential activation of the MAP kinase pathway. The same authors demonstrated that TGF-β lost its pro-apoptotic but not its pro-invasive activities through this pathway. TGF-β activates a range of different signalling pathways leading to EMT during tumour progression (Fig. 2), such as Rho family of GTPases (Fig. 2A), PI3K/AKT (Fig. 2B), integrin-linked kinase (ILK) (Fig. 2C), NF-κB (Fig. 2D) and MAP kinases (Fig. 2E).  Each of these signalling pathways involves different features that are inter-related, for example, downregulation of EMT suppressors such as E-cadherin or β-catenin by a range of transcription factors including Snail, Slug, Twist, and ZEB1. 
Molecular features of EMT during tumour progression
There are four important stages in cancer development and cell migration which are (1) invasion, (2) intravasation, 3) extravasation and 4) metastatic colonization.  In order to complete each step during this process molecular changes are required to enable EMT. These molecular and genetic alterations are due to different transcription factors, which act as EMT inducers by stimulating and/or suppressing functions of different proteins involved in this complex process. Important transcription factors are Snail, Slug (Snail 2), Twist and ZEB1, they act in different ways, but all of them repress E-cadherin expression which is classified as a hallmark of EMT.
One of the principal characteristics of EMT is the loss of E-cadherin expression. E-cadherin participates in cell–cell adhesion and interacts with other molecules to form epithelial junctions. It connects epithelial cells by calcium-dependent homotypic interactions.  Its expression is inversely proportional to the tumour grade and stage, and patient prognosis. β-catenin is also essential for keeping cells attached to each other, it acts by binding E-cadherin and α-catenin to the actin cytoskeleton. [35-37]
Once epithelial cells detach from other cells and their epithelial sheet, they can acquire motility by changing their shape and polarity in order to migrate through different tissues and lymphatic and blood vessels. Vimentin is an intermediate filament protein with structural features which plays a main role in switching a cell's shape and making their cytoskeleton stronger so it can be more flexible and prevent damage. It binds with microtubules and actin microfilaments to make up the cytoskeleton. Without this protein, migrating cells would be very fragile. Not only is a robust cytoskeleton necessary for migrating cells, it is also necessary to guide these cells during migration. Fibronectin is a high molecular weight extracellular matrix glycoprotein which plays an important role in cell adhesion, growth, migration and differentiation.  It plays an important role by guiding cells and binding to collagen and fibrin to enable embryogenesis and wound healing. Genetic alterations in fibronectin have been associated with some pathologies like fibrosis and metastasis. [38-40]
The vast majority of molecular changes in these proteins are due to genetic alterations in different transcription factors which can be divided into different families. The Snail family is made up of Snail and Slug (Snail 2) and they are involved in different stages of development such as gastrulation, mesoderm formation, cell differentiation, cell motility and apoptosis. [41-44] They can be implicated in EMT by downregulating the expression of E-cadherin. [45-47] Olmeda et al.  showed that when Snail and Slug were inhibited in mammary tumour cell lines, tumour growth was impaired and their metastatic potential was lowered in mice. Twist is part of the basic helix-loop-helix protein family and one of its most important features is the ability to inhibit apoptosis. It is also known to trigger EMT by downregulating E-cadherin expression. [49, 50] Two members of the zinc finger E-box binding proteins family are ZEB1 and ZEB2, both of which can regulate E-cadherin and thus, are capable of stimulating EMT (Fig. 3). [51-53] Burk et al.  showed that ZEB1 was overexpressed in colorectal, pancreatic and breast cancer cell lines undergoing EMT induced by TGF-β1 and TNFα. Moreover, when they knocked down ZEB1, EMT was partially prevented. Each of these different epithelial/mesenchymal biomarkers and transcription factors has different characteristics and functions as shown in Table 1.
Table 1. EMT biomarkers (epithelial and mesenchymal) and transcription factors characteristics as well as their expression during EMT
Expression during EMT
Every component of this complex network has different characteristics; some of which would enable cells to acquire mesenchymal features whilst losing epithelial properties. Their expression can be downregulated or upregulated, depending on their role in EMT.
Recent studies have shown that a small population of cells within a tumour can have stem-cell characteristics, such as self-renewal and pluripotency. These cells have been named CSCs or tumour-initiating cells  and are required to produce all cancer cell types in a tumour.  The concept that tumours can grow from small populations of tumorigenic cells was proposed in the late nineteenth century, but was tested by Dick et al. [58-62] approximately 100 years later, in 1994 with haematopoietic stem cells and leukaemia stem cells. This seminal study, helped to define cancer stem cells, i.e., a single cell capable of generating a population of heterogeneous cells. [58-62] They confirmed that acute myelogenous leukaemia could be seen as a hierarchical model which emerged from a single haematopoietic stem cell by looking for expression of cell-surface markers. They found out that leukaemia-initiating cells that could form large number of colonies in transplanted severe combined immune deficient (SCID) mice were CD34+ CD38−. And cells that were CD34+ CD38+ and CD34− did not have the same colony-forming properties.  Different studies of neoplastic tissues are providing strong evidence of the existence of tumour-initiating cells.  These cells were first discovered in the haematopoietic system, but have subsequently been identified in several kinds of solid tumours including colon, breast and brain cancer. [64-67]
Linking EMT and cancer stem cells
Recent experiments have demonstrated that cancer cells induced to undergo EMT, produce cells with stem-cell characteristics.  In addition to morphological changes, epithelial cells induced to undergo EMT may also develop altered functional properties, such as tumour-seeding ability, tumoursphere formation and expression of transcription factors Twist and Snail. In addition, they also play an important role in invasiveness and migration in different types of cancer (Fig. 3).  Independent groups have demonstrated that cells that had undergone EMT by expressing different transcription factors or induced by TGF-β, showed stem-cell characteristics, such as the ability to form spheres and expressed stem cell associated cell-surface markers. [68, 70, 71]
Mani et al.  showed that by over expressing Twist or Snail in human mammary epithelial cells (HMECs), these cells became more mesenchymal, and their expression pattern demonstrated downregulation of epithelial markers, such as E-cadherin, and upregulation of mesenchymal markers such as vimentin and fibronectin. After confirming that EMT took place in these cells, they used flow cytometric analysis to see if they had stem-cell characteristics. They confirmed that these cells expressed CD44+CD24−, which are cell-surface markers for mammary epithelial stem cells and human breast cancer stem cells. They achieved similar results exposing epithelial cells to TGF-β.  Furthermore, they measured the sphere forming ability of normal mammary epithelial cells, compared with EMT induced cells with TGF-β and found that EMT induced cells were capable of forming at least 30-fold more mammospheres than the untreated cells. Moreover, they performed cell sorting in HMECs, mouse mammary stem cells and normal and neoplastic human breast stem cells and confirmed using RT-PCR the overexpression of mRNA's encoding mesenchymal markers and downregulation of epithelial markers in CD44+CD24− (human) and CD49fhigh CD24med (mouse). They also transfected immortalized HMECs, transformed by the HER2/neu oncogene, with a vector expressing the tamoxifen-activatable form of Snail or Twist transcription factors in order to confirm the possibility of EMT generating cells with stem-like properties. These treated cells were assayed for tumoursphere forming efficiency. They found that cells expressing Snail and/or Twist underwent EMT, and formed 10-fold tumourspheres than the untreated group (control).  Further, they showed that EMT-derived cells were capable of differentiating into other cell lineages. 
Additional studies have confirmed that tumour-initiating cells can originate from a more differentiated cell line. Morel et al.  confirmed this by utilizing FACS analysis to demonstrate that cells expressing CD44+CD24−/low can be derived from cells expressing CD44lowCD24+ through the activation of the Ras/MAPK signalling pathway, and most interestingly, that this process can be stimulated and accelerated by EMT. They compared the abilities of HMEC's and an oncogenic line (HMLER) to form mammospheres, a process associated with stem-cell properties. They confirmed that only HMLER cells were able to form mammospheres, and subsequently analyzed the cell phenotypes of these different cell lines by FACS confirming that HMEC's were CD44lowCD24+ and HMLER were CD44+CD24−/low. Furthermore, they performed cell sorting and single-cell cloning assays after H-RasV12 retroviral expression in HMECs. They seeded CD24+ cells into 96-well plates with limiting cloning conditions and saw that after three weeks, 19% of the population of cells were CD44+CD24−. This population of cells was able to grow tumours when injected into mammary pads of nude mice, compared with CD44lowCD24+, which were not able to establish tumour growth. CD44+CD24− cells in HMEC and MCF10 (immortal human mammary epithelial cell line) cell lines showed a spindle shape and expressed lower levels of E-cadherin and β-catenin (epithelial markers) and higher levels of fibronectin and vimentin (mesenchymal markers), suggesting that these stem-cell properties were related to EMT. Finally, they confirmed that EMT enabled cells to acquire stemness by treating the CD44−CD24+ cell lines with TGF-β1, which is one of the most potent EMT inducers, for eight days and showed upregulation of vimentin and downregulation of E-cadherin, as well as presence of CD44+CD24− cells.  Collectively, these results strongly suggest that EMT is potentially a precursor to generate cancer stem cells from more differentiated cell lines (Fig. 4). Further work is required to consolidate these theories.
Clinical and therapeutic implications of an EMT/CSC axis
Recent data demonstrates that epithelial cells induced by EMT may play an important role in invasiveness and migration in different types of cancer (Fig. 3).  The percentage of cancer stem cells in a tumour may vary between different tumours and patients (Fig. 5A).  The subpopulation of tumour-initiating cells identified in human breast cancer exhibit CD44+ and CD24− cell-surface markers, and were found to be more resistant to conventional therapies than the more differentiated cancer cells  suggesting that tumour relapse can be accomplished by these tumour-initiating cells after treatment (Fig. 5B). Therefore, CSC's must be eliminated to affect a cure on cancer (Fig. 5C).
As cancer stem cells are proposed to be more resistant to conventional therapies, it becomes necessary to find new ways to target them in order to treat cancer and prevent its recurrence. The central role that EMT has in tumour progression makes it an obvious target for therapeutic intervention. However, the inhibition of EMT may also have serious consequences for wound healing processes and also tissue remodelling during repair and regeneration.
Independent groups have demonstrated that cells that have undergone EMT exhibit stem-cell characteristics, including resistance to chemotherapy. [68, 70, 71, 75] Therefore targeting EMT inducers may help treating cancer. Olmeda et al. [76, 77] have shown that by silencing Snail with shRNA, they can block EMT by enabling E-cadherin expression and lead to a MET process. Knowing that EMT plays a main role in generating cancer stem cells is crucial in order to target its different pathways, and screening with different drugs to block this feature. The development of new therapies blocking EMT and its precursors (transcription factors) is of vital importance in this quest for curing cancer and avoid its progression. In different approaches regarding the blockage of EMT during cancer progression (tumour growth and proliferation), it has been shown that blocking oncomirs (microRNAs) can be a useful resource as described by Yan et al  after inhibiting proliferation of human breast cancer cells (MCF7) in vivo by miR-21 knockdown with peptide nucleic acids (PNAs).
If this crosslink between EMT and cancer stem cells is confirmed, treatment of cancer at early stages should include EMT blocking agents that could inhibit TGF-β downstream effectors like ZEB1. By these means, resistance to apoptosis and self-renewal characteristics can be nullified, preventing not only migration of cancer cells, but also the development of the primary tumour by stopping cancer stem-cells multiplication (Fig. 6).
Cellular reprogramming, EMT and E-cadherin
Cell reprogramming can be defined as the ability of specific transcription factors to change the identity of specialized differentiated cells into different cell lineages. [79, 80] As described by Yamanaka et al., it might also be an important step in the acquisition and maintenance of pluripotency of epithelial cells achieved by defined pluripotency transcription factors: OCT4, SOX2, KLF4 and c-MYC (OSKM), also known as the Yamanaka cocktail. [81, 82] This group showed that pluripotency can be induced in mouse embryonic fibroblasts (MEFs), and these can be reprogrammed by the above mentioned factors to become induced pluripotent stem cells (iPSCs), and this mechanism would be named cell reprogramming to pluripotency. 
Stadtfeld et al.  described that an MET is necessary for the reprogramming process, which was then confirmed by Redmer et al.  showing that E-cadherin might play an important role in maintaining pluripotency in mouse embryonic stem cells (mESCs), and it can also replace OCT4 during reprogramming after an MET. It is not well understood how E-cadherin maintains pluripotency. However, its role was confirmed by replacing OCT4 in the OSKM combination with an E-cadherin-expressing retrovirus (pMXs-Ecad) which demonstrated that this new combination of factors (ESKM) reprogrammed MEFs into iPSCs, the latter being able to form teratomas in non-obese diabetic/severe combined immunodeficiency (NOD/SCID) mice, while SKM clones just formed small lumps. 
The metastatic progression of different types of cancer is carried out by a complex network of pathways in which E-cadherin plays different roles regarding EMT, MET and as mentioned above, cell reprogramming (Fig. 7). The role of E-cadherin in cell reprogramming as a pluripotency inducer could be one of the essential features of metastasis due to its ability to change cellular identity once the cancer cells have reached their new niche after EMT and subsequent MET (in which E-cadherin is highly upregulated).
Furthermore, Liao et al.  showed that cell reprogramming through an MET could also be carried out by microRNA (miRNA) expression in mesenchymal cells. They demonstrated that overexpression of miR-302 complex in addition to miR-367 in fibroblasts was enough to downregulate a TGF-β receptor (TGF-βR2), thus, enabling cells to undergo a MET by upregulation of E-cadherin and cell reprogramming into iPSCs.
These findings would confirm a link between the epithelial phenotype and pluripotency, elucidating potential targets for prevention or treatment of metastatic consequences.
EMT in companion animals
Epithelial–mesenchymal transition in companion animals (dogs and cats) has not been widely studied. Chandler et al.  showed the importance of Slug as a transcription factor during cell migration in wound healing. They used 12 dogs to which they induced corneal wounds and measured Slug expression in wounded and unwounded corneas. They assessed the rate of wound healing with and without Slug transfection. Their results showed an upregulation of Slug in wounded corneas compared to the unwounded ones. They also observed downregulation of E-cadherin and β-catenin (epithelial markers) in healing tissues, meaning that transcription factor Slug downregulated these epithelial markers so the cells could lose their adherens junctions and start moving through the cornea. Finally, when transfecting wounded corneas ex vivo with Slug, the migration rate increased dramatically, suggesting that epithelial cells became mesenchymal through an EMT process.
In two different studies, Aresu et al. [87, 88] have been studying the role of EMT in canine renal fibrosis, in which EMT has been found to be essential to generate this abnormality that can cause several renal diseases. They showed that during renal inflammation-fibrosis, leading to diseases such as glomerulonephritis, epithelial markers E-cadherin and β-catenin  and cytokeratin  are downregulated whilst the mesenchymal marker vimentin is identified and upregulated at inflammation areas. As we can see, these research groups have found important results regarding type II EMT (wound healing and fibrosis), but these studies have yet to be extended to type III EMT (cancer progression and metastasis).
Several authors from different research groups have published interesting results regarding expression of cell adhesion molecules and their correlations with tumour growth and cell proliferation.
Han et al.  observed an increased expression of β-catenin accumulation in the cytoplasm and nucleus in canine cutaneous melanotic tumours. Normally, β-catenin is degraded in the cytoplasm; however, dysregulation of the Wnt/β-catenin signalling pathway leads to accumulation of this protein, causing uncontrolled cell proliferation.  These findings have been also documented in canine colorectal tumours  and canine osteosarcoma.  Moreover, Nowak et al. [93, 94] compared the reciprocal relations between extracellular matrix metalloproteinase (MMP-9), E-cadherin, the proliferation associated antigen Ki-67 and β-catenin in canine mammary adenocarcinoma, finding that the decreased expression of E-cadherin is inversely proportional to the expression of MMP-9, Ki-67 and nuclear-located β-catenin, whereas a direct correlation was observed between MMP-9 and Ki-67, and β-catenin and Ki-67. They also observed increased cell growth and proliferation associated with higher expression of nuclear accumulation of β-catenin.
In canine colorectal adenocarcinoma, Aresu et al.  observed lower expressions of E-cadherin and β-catenin in the cell membrane in the majority of analyzed tumours. They found a correlation between low expression of E-cadherin with higher grade (grade 4) tumours and higher mean age of patients. Reduced expression of cell membrane accumulation of β-catenin was also related to tumours with higher grade, but also with increased tumour size.
Furthermore, Ide et al.  found that increased expression of cell adhesion molecules N-cadherin, doublecortin and nuclear β-catenin was closely associated with progression of canine meningioma.
Most of these groups correlate the expression of these molecules associated with cancer cell growth and/or progression, but interestingly, they do not make any observation about EMT type III in small animals, nevertheless, all of their perspectives are of great importance to the development of deeper understanding of EMT in cancer development and progression.
In the authors' laboratory, we have studied EMT as a feature of human, canine and feline mammary carcinoma and feline squamous cell carcinoma in an attempt to elucidate the role of EMT in cancer progression and in the acquisition of stem-cell-like characteristics. We have demonstrated that dog and cat cells undergoing EMT by TGF-β stimulation show mesenchymal morphology (spindle shapes) and lose cell–cell contact, separating from the group of cells. These cells also overexpress mesenchymal markers (fibronectin and vimentin) while epithelial markers (E-cadherin and β-catenin) are downregulated. Interestingly, cells that underwent EMT (verified by protein expression and morphologic changes) are more capable of forming spheres, which are representative of cancer stem cells. 
These results offer the opportunity for exploring EMT as a potential therapeutic target in companion animal oncology.
Unanswered questions and future work
There is still much to be done regarding this difficult task of knowing and understanding the complex network of cancer and its progression mechanisms.
Some of the key questions include:
• What other mechanisms and factors are involved in EMT and MET?
• Can this process be targeted therapeutically and safely?
• What mechanisms link EMT, MET and pluripotency in cells?
• What is the role of microRNA's in these processes and can they be targeted
It is becoming clear that EMT and MET are key mechanisms in cancer progression. However, it is also clear that these processes may also be involved in the maintenance of pluripotency in normal cells. One would anticipate that there will be convergence of these respective research fields and it is likely that key questions will be answered by studying both normal and neoplastic processes simultaneously.
EMT and cancer progression is an emerging field of cancer research. EMT may be very relevant to cancer induction and progression research. Moreover, if we can confirm that EMT/MET process is an inducer of stem-cell characteristics, then we would be able to generate new therapies against cancer and its progression and metastasis.
This work was supported by the Mexican National Council on Science and Technology (CONACYT) and the Scottish Overseas Research Students Award Scheme (SORSAS).