Breast cancer is the most commonly diagnosed cancer (1,2). It is also the second leading cause of cancer-related mortalities in females. Among these patients, more than 90% of breast cancer-related mortalities are caused not by the primary tumor, but by its metastases at distant sites (2). In 6–10% of breast cancer diagnoses, cancer has already metastasized to other parts of the body, and ~30% of patients with early-stage breast cancer have a metastatic or recurrent disease (3). Understanding the breast cancer metastasis mechanism is important for early diagnosis and treatment of breast cancer. The precise mechanisms that are involved in the transition of the subset of non-invasive tumor cells into cells with metastatic potential are still not well understood. However, accumulating evidence suggests that an epithelial-mesenchymal transition (EMT)-like process is one of the main mechanisms involved in breast cancer metastasis (4).

EMT describes a process in which cells lose epithelial traits and gain mesenchymal characteristics. EMT is characterized by loss of cell adhesion and phenotypic change from typical cuboidal to an elongated spindle shape, leading to enhanced migratory capacity (5). In the early stage of tumor metastasis (Fig. 1), cancer cells from the primary tumor could acquire invasive properties and gain access to the blood or lymphatic vascular systems as circulating tumor cells (CTCs) (6,7). This procedure is aided by neo-angiogenesis and remodeling of the basement membrane (8–10). In the bloodstream and lymphatic vessels, CTCs are capable of surviving and eventually reach distant secondary sites, including the bone, lungs, liver and brain. This process is accomplished mainly by the mesenchymal-epithelial transition (MET), which is a process opposite to the initial EMT at the primary tumor site, and is considered to contribute substantially to the colonization of CTCs into metastatic tumors at the secondary site (11–14). Such dynamic EMT/MET state transitions may play a critical role during tumor metastasis.

The purpose of this review is to present the growing evidence (4) that EMT plays a significant role in the invasion and metastasis of breast cancer. Numerous EMT-related genes, proteins, microRNAs (miRNAs or miRs) and kinases contribute to EMT, and they are associated with breast cancer invasion and metastasis. However, there is no systematic summary for the association between EMT markers and breast cancer. In the present study, the association between EMT markers and breast cancer invasion and metastasis will be summarized.

A variety of biomarkers have been used to demonstrate EMT in breast cancer (Fig. 2). Here, we examine various of the most common markers, some of which are acquired, while some of which are attenuated, during the transition.

E-cadherin is a calcium-dependent cell-surface protein that facilitates adhesion between epithelial cells (15). E-cadherin is characterized by long cytoplasmic and extracellular domains, which create homophilic interactions between adjacent cells to facilitate adhesion (16). A change in the expression of E-cadherin is the typical epithelial cell marker of EMT (17). Suppression of E-cadherin function or expression leads to mesenchymal morphology and increased cell migration and invasion (18,19) as well as metastasis (20). In breast cancer, partial or total loss of E-cadherin expression correlates with loss of differentiation characteristics, acquisition of invasiveness, increased tumor grade, metastatic behavior and poor prognosis (21–24). Furthermore, its reduced expression can also be associated with some non-lobular breast carcinomas of triple-negative phenotype such as metaplastic carcinomas (25,26). Choi et al analyses revealed that the loss of E-cadherin in invasive carcinoma is greater than in pure ductal carcinoma in situ (DCIS) in basal-like subtype cancer (27). Recently, changes in the level of expression of different cadherins have been increasingly used to monitor EMT. Indeed, the cadherin switch from E-cadherin to N-cadherin has often been used to monitor the progress of EMT during cancer progression. This switch increases cell motility and the abilities of invasion and metastasis (28). In addition, E-cadherin also plays a major role in the process of MET. In a previous study, Chao et al reported the re-expression of E-cadherin at distant metastatic tumors arising from E-cadherin-low or E-cadherin-negative primary tumors (11). The authors reported strong E-cadherin expression in >50% of liver, brain and lung metastases originating from infiltrating ductal carcinoma of the breast (29).

Cluster of differentiation (CD) 44 is a cell-surface protein that modulates cellular signaling by forming co-receptor complexes with various receptor tyrosine kinases. It plays an important role in the metastasis of breast cancer (30). Some studies have shown an upregulation of CD44 in a metastatic cell line as compared with a non-metastatic cell line (31). There is a shift in CD44 expression from the variant isoform (CD44v) to the standard isoform (CD44s) during EMT. The splicing factor epithelial splicing regulatory protein 1 could control the CD44 isoform switch, which is critical for regulating the EMT phenotype. CD44s expression is upregulated in high-grade human breast neoplasms, and is correlated with the level of the mesenchymal marker N-cadherin in these tumors (30).

Discoidin domain receptor 2 (DDR2) is an atypical receptor tyrosine kinase. It is the collagen-specific receptor that reflects adaptation to the altered ECM microenvironment associated with the EMT (32). In adult tissues, the expression of DDR2 is confined to subsets of fibroblasts or vascular smooth muscle cells (33,34). In cancer, DDR2 facilitates prostate cancer cells to adhere to type I collagen. It plays an important role in prostate cancer bone metastasis (35). In breast cancer, DDR2 expression correlates with increased invasiveness, thus demonstrating its utility in identifying EMT (36). It has been reported that activation of the collagen I receptor DDR2 regulates Snail1 protein stability by stimulating extracellular signal-regulated kinase 2 (ERK2) activity. Activated ERK2 directly phosphorylates Snail1 and reduces its ubiquitination; as a result, the half-life of Snail1 increases. DDR2-mediated stabilization of Snail1 promotes breast cancer cell invasion and metastasis in vivo (37). Recent research suggests that DDR2 facilitates breast cancer cell metastasis in vivo as well as hypoxia-induced cell migration, invasion and EMT in vitro. DDR2 could be a potential target to treat breast cancer metastasis (38).

β-catenin is a cytoplasmic plaque protein that plays an important role in EMT (39). β-catenin is localized in the cell membrane of normal epithelial cells and noninvasive tumor cells. In cells undergoing EMT, β-catenin is located either in the cytoplasm or in the nucleus of the cells. β-catenin localization in the cytoplasm is reflective of its dissociation from E-cadherin (40). Subsequently, β-catenin translocates to the nucleus to promote the transcription of genes that induce EMT (41). Prasad et al provided clinical evidence to support the upregulation of Wnt/β-catenin signaling in invasive ductal carcinoma of breast (42). Choi et al reported a higher expression of β-catenin in invasive carcinomas than in pure DCIS, especially in basal-like subtype breast cancers (27). Moreover, the notable reduction of β-catenin expression and β-catenin/transcription factor 4 (TCF4) transcriptional activity by β-catenin short hairpin RNAs (shRNAs) in HMLE cells expressing cyclin-dependent kinase-like 2 (CDKL2) resulted in decreased zinc finger E-box binding homeobox 1 expression level and promoter activity, as well as in increased E-cadherin expression and redistribution from the perinuclear region to the membrane (43). This suggests that the silencing of β-catenin can reverse CDKL2-induced EMT in breast cancer.

Vimentin is an intermediate filament that is used as a marker of mesenchymal cells to distinguish them from epithelial cells (41). Vimentin is expressed at sites of cellular elongation, and is associated with a migratory phenotype. Increased vimentin expression is frequently used as an EMT marker in cancer (44,45). There is a positive correlation of vimentin expression with augmented invasiveness and metastasis. In breast cancer, it was observed that Smad-interacting protein 1 (SIP1) could regulate vimentin expression in epithelial breast tumor cells, and that vimentin was distinctly related to SIP1 expression in invasive cell lines (46). These data suggest that the regulation of vimentin by SIP1 can be independent of E-cadherin expression and does not necessarily rely on the modulation of the β-catenin/TCF pathway. Nevertheless, many other indirect mechanisms could be involved (46). In addition, vimentin expression was found to be significantly higher in patients with stage IV disease, Bloom Richardson score 4 and progesterone receptor negativity. That study revealed that vimentin expression was a significant biomarker for predicting reduced disease-free survival and overall survival in breast cancer (47).

α-smooth muscle actin (α-SMA) is one of the six actin family members (48). Cells expressing α-SMA contribute to EMT in embryogenesis and to wound healing in normal epithelial cells (41). In adults, significant α-SMA expression can be found in vascular smooth muscle cells and myoepithelial cells (46). In cancer, evidence that the EMT is associated with α-SMA is mostly confined to breast cancer (49), where α-SMA is largely detected in breast tumors of the ‘basal phenotype’ (50). A previous study used 60 patients with a known prognosis of invasive breast cancer to identify that α-SMA and lymph node metastasis were independent predictive factors of metastasis. The result indicated that the metastasis group showed significantly higher α-SMA expression compared with the non-metastasis group (51).

The expression of E-cadherin plays a very important role in the process of EMT. EMT transcription factors can be classified based on their ability to repress E-cadherin directly or indirectly.

miRNAs, functioning as co-activators or co-repressors, are key players in cell plasticity, being specifically involved in cell regulation with EMT-related transcription factors. miRNAs that contribute to EMT in breast cancer are categorized as either EMT inducers or EMT repressors (Table I).

Several kinases regulate EMT, stemness or metastasis, including FYN proto-oncogene, Src family tyrosine kinase, platelet-derived growth factor receptor α, BRAF, Fms-related tyrosine kinase 1, LYN proto-oncogene, Src family tyrosine kinase and YES proto-oncogene 1, Src family tyrosine kinase (92–94). In the present report, we review two recent studies on breast cancer occurrence and EMT-associated protein kinases.

AXL receptor tyrosine kinase (AXL) is a member of the Tyro3-AXL-Mer family of receptor tyrosine kinases. AXL is overexpressed in a wide variety of human cancers, with significant correlation with tumor stage in breast cancer patients. It plays a important role in cancer progression and metastasis (95–97). Asiedu et al (98) reported that AXL overexpression in HMLE cells downregulated E-cadherin, while the expression of mesenchymal markers such as N-cadherin, Snail and Slug was upregulated. A similar result was observed by forced expression of AXL in the normal human mammary epithelial cell line MCF10A. On the contrary, the silencing of AXL by viral-mediated shRNA led to the upregulation of epithelial markers, while mesenchymal markers were downregulated. Inactivation of AXL also led to downregulation of the NF-κB pathway and reduced tumor formation in vivo (98). In addition, AXL is overexpressed in highly invasive breast cancer cell lines. By contrast, weakly invasive breast cancer cell lines do not or merely express limited quantities of AXL. AXL may be an important therapeutic target in inflammatory breast cancer (99).

CDKL2 is one of the most distant members of the cell division cycle protein 2-related serine/threonine protein kinase and mitogen-activated protein kinase family, which is also known as p56 or KKIAMRE (100,101). A recent study (100) demonstrated that CDKL2 activated a positive feedback loop, consisting of ZEB1/E-cadherin/β-catenin, to induce EMT in breast cancer. As a result, E-cadherin expression was reduced, and the epithelial barrier was broken down, which led to nuclear translocation of β-catenin as well as elevated β-catenin/TCF4 transcriptional activity. Activated β-catenin increased ZEB1 promoter activity and transcription, which in turn resulted in further suppression of E-cadherin expression and continuous activation of the positive feedback loop (100) This result suggested that CDKL2 can be a potential prognostic factor for poor outcome and a therapeutic target for human invasive breast cancer.

Currently, the role of EMT and MET in breast tumors is under investigation, and the molecular mechanism of EMT and MET is being revealed. With the development of individualized breast cancer therapies, new prognostic and predictive biomarkers are required to facilitate clinical decision-making processes. Further studies on the link between EMT markers and breast cancer, such as the link between EMT and biomarkers, the regulatory association between transcription factors and miRNAs, and the link between EMT and protein kinases, will contribute to the identification of biomarkers for predicting early breast cancer metastasis and to identify intervention therapeutic targets in breast cancer, as well as to provide new ideas and methods for the treatment of breast cancer.