Portal de Periódicos da CAPES

The epithelial-mesenchymal transition: Fact or fiction in cancer?

2009; Lippincott Williams & Wilkins; Volume: 50; Issue: 5 Linguagem: Inglês

10.1002/hep.23329

1527-3350

Gianluigi Giannelli,

Wnt/β-catenin signaling in development and cancer

The epithelial-mesenchymal transition (EMT) is a physiological embryogenetic process during which epithelial cells become mesenchymal cells. These cells are highly motile so as to be able to migrate to areas where organs, including epithelial organs, originate. This concept is widely reported in embryology textbooks. APC, adenomatous polyposis coli; CFI, cancer-free interval; EMT, epithelial-mesenchymal transition; HCC, hepatocellular carcinoma; OS, overall survival. The EMT was first recognized as an independent process in 1982, and since 1989 it has been linked to cancer, because cells from rat bladder carcinoma were seen to have a changed morphology, were missing desmosome junctions, and had a rearranged cytoskeleton.1, 2 Several other markers have been described as hallmarks of the EMT. Now we know that it is characterized by a weak expression of E-cadherin, inversely correlated to the zinc-finger proteins Snail, Slug, and Sip-1,3-6 or the transcriptional factor Twist.7 As a consequence, β-catenin interacts with the adenomatous polyposis coli (APC) protein and works as a negative regulator of the Wnt-1 pathway.8 Once inside the nuclei, β-catenin stimulates the transcription of a number of genes, including c-myc.9 One example of the clinical outcome of such a complicated cascade is the mutation of the APC tumor suppressor gene in a high percentage of patients with familial adenomatous polyposis.10 In a mouse experimental model of hepatocellular carcinoma (HCC), the up-regulation of c-Myc by β-catenin was responsible for inducing HCC, whereas its inactivation induced tumor dormancy, although the molecular signature was the same as in the active stage tumor.11 In humans, HCC is characterized by long periods of disease inactivity that can suddenly, and for no apparent reason, be replaced by a great aggressiveness. Unfortunately, the mechanisms responsible for the dormancy of HCC have not yet been demonstrated. In this issue of HEPATOLOGY, Yang and collaborators12 have investigated several EMT markers in a large cohort (123) of patients, stratifying the data according to the patients' clinical characteristics. The authors report a decreased expression of E-cadherin in 52.8% of patients, validating this finding by means of three different techniques (immunohistochemistry, enzyme-linked immunosorbent assay, and real-time reverse-transcription polymerase chain reaction). Interestingly, the down-regulation of E-cadherin was statistically significantly associated to a reduced cancer-free interval (CFI) in both univariate and multivariate analysis, and was also associated to a more aggressive phenotype. Moreover, the authors report an overexpression of two other important EMT markers, Snail and Twist, but not of Slug, associated with a down-regulation of E-cadherin and the nuclear translocation of β-catenin. Consistently, an increased expression of Snail or Twist is associated, in a statistically significant manner, with the tumor aggressiveness and with the CFI, as well as overall survival (OS). In addition, the patients with the worst prognosis and shortest CFI and OS were those with positivity to all the EMT markers. Finally, independent predictors of the CFI in multivariate analysis were a down-regulation of E-cadherin and an overexpression of Twist, whereas in multivariate analysis for OS, the coexpression of Snail and Twist was statistically significant. All these interesting data further support the hypothesis, advanced in previous studies, that EMT has a role in HCC,13 although the final statistical analysis is restrictive as compared to the plethora of data included by the authors. This apparent discrepancy can be explained by at least two different aspects: the heterogeneity that is one of the hallmarks of HCC and, on the other hand, the difficulty in getting the EMT at the right moment. This could explain why, in other studies, alterations of E-cadherin and the nuclear translocation of β-catenin in HCC human samples suggest different roles for these molecules from those reported in other tumors.14, 15 However, although a huge number of articles have focused on the role of the EMT in cancer development and progression, there is still no consensus in the scientific community on this issue. In particular, pathologists are skeptical about the real existence of the EMT, and a famous negative review appeared in 2005, written by Tarin, who concluded there was "no convincing evidence for conversion of epithelial cells into mesenchymal cell lineages in vivo".16 The analysis made in this review is correct, although the author's feelings about the EMT probably led to an extreme emphasis on the criticisms. Nevertheless, an important misleading factor is the lack of a unanimous definition of the EMT and, in particular, of how many markers are required to define EMT (Fig. 1). This has given rise to a number of articles that have probably contributed to the reader's confusion, whereas a more careful evaluation of the results could be important. Therefore, the best that can be done at the moment is to make a simultaneous analysis of several EMT markers, on serial sections by different techniques, as performed by the authors. But the EMT is a dynamic process that cannot be followed serially in humans, and this is another important limiting factor in several studies. In fact, it seems possible that cancer cells can reverse the EMT cascade by a mesenchymal-epithelial transition (MET), in a sort of "plastic", counterbalanced process. This can explain why some but not all the EMT markers are observed in tumors or, as in the study by Yang et al., are associated to the clinical outcome. In a more general view, this could also justify the difficulties found by pathologists in recognizing EMT-transformed cancer cells. However, if we can avoid sticking too closely to the strict meaning of the words, to avoid a possible semantic discrepancy, the term EMT could be changed to EMT-like, as also proposed by Thompson and Newgreen in response to Tarin's review. As these authors suggest "we don't consider this (EMT) in the embryologic sense". This is an important point that could finally bring about an agreement between pathologists and other researchers. All are agreed that cancer cells undergo a de-differentiation process; in the case of HCC, this is evident from the production of alpha-fetoprotein. In my opinion, what generates the ongoing debate about the EMT is mainly the word "mesenchymal" because, as Tarin argued in his review, we have never observed EMT-transformed cells "to make cartilage, bone tendons" as the real mesenchymal cells do.16 In this sense, Tarin and all the pathologists who are not convinced of the EMT are correct, but one way to overcome this problem might be to rename the word "mesenchymal" as "mesenchymal-like", or maybe "mesenchymal phenotype" would be better. In other words, the EMT would replace the term "de-differentiation", preserving the same concept. Schematic representation of different types of EMT. Type 1 EMT underlies mesoderm and endoderm formation during embryonic development. In particular, the primary mesenchyme originates from the epiblast through an EMT process. Type 2 EMT occurs in epithelial cells under prolonged inflammatory stresses. Under these conditions, epithelial cells can show intermediate aspects of epithelial and mesenchymal phenotype (partial EMT), such as in liver fibrosis. This has been proposed to occur during early stages of HCC tumorigenesis. Type 3 EMT is believed to play a role in cancer progression. Cancer cells could undergo the EMT process that favor invasion and metastasis. Moreover, it is believed that an opposite process termed "mesenchymal-epithelial transition" (MET) takes part in the metastatic foci at distant sites, thus explaining the apparent controversial data on EMT in cancer. So, is it just a matter of changing the name? I do not believe so, but because everybody agrees on the relevance of this process, nevermind the nomenclature, we can say that the EMT implicates a more dynamic way to recognize the tumoral de-differentiation process, thus dissecting the complicated process of carcinogenesis and tumoral progression. The main goal is to find a pathway or a molecule or a trigger of EMT, as in the case of transforming growth factor-β in HCC,17 which can be targeted by new biological therapies as recently proposed with a transforming growth factor-β receptor kinase inhibitor.18-20