Article DirectoryArticle Directory – Free Online Article Directory

The expression of several EMT inducers, such as Snaill, Snail2, Twist, and SIP1, has been demonstrated in diverse human tumors using different methodologies. Generally, the expression of these transcription factors is correlated with E-cadherin down-regulation As discussed earlier, based on experimental and developmental studies, we proposed a hierarchical model for the action of different repressors during tumor progression. However, the exact contribution of these transcription factors in tumorogenesis in vivo is not fully understood.

Expression of Snaill mRNA has been detected in biopsies from patients with breast ovarian, gastric, colon, and hepatocellular carcinomas. Of these reports, only one analyzed Snaill expression directly in epithelial neoplastic cells. Thus, the study by Blanco et al. (2000) using in situ hibridization reported that mRNA expression was detected at the invasive areas and more frequent among poorly differentiated breast carcinomas with lymph node metastasis. As mentioned before, in breast cancer a role for Snaill in tumor recurrence has also been proposed. Comparative transcriptome analysis of human primary breast cancers suggests that elevated Snaill expression is correlated with decreased relapse-free survival. In addition, Toyama et al. observed that node-negative infiltrating ductal carcinomas showing low Snaill mRNA expression do not developed distant metastasis, whereas 12%, showing high expression of Snaill mRNA, did. Collectively, these data suggest that Snaill expression is a poor prognostic factor in breast carcinomas and is probably involved in local recurrence and in lymph node and distant metastasis.

Additionally, it has been reported that Snaill expression in colon cancer was associated with down-regulation of E-cadherin and VDR (vitamin D receptor) gene products. These findings suggest that Snaill may be associated with loss of responsiveness to vitamin D analogs and may thus be used as an indicator of patients who are unlikely to respond to this therapy. Another study reported Snaill mRNA overexpression in 16% of hepatocellular carcinomas compared with adjacent noncancerous liver tissue. E-Cadherin protein expression was found significantly down-regulated in cases with Snaill mRNA overexpression. In addi­tion, the tumor and nontumor ratio of Snaill mRNA correlated independently with tumor invasiveness. This indicates that Snaill both down-regulates E-cadherin expression and promotes invasion in human hepatocelular carcinoma. Finally, up-regulation of Snaill associated with reduced or negative E-cadherin expression has also been reported in diffuse gastric cancer.

Despite these data, the exact role of Snaill in clinical tumor invasion and pro­gression remains to be established. In most of the aforementioned studies, the precise contribution of stromal and carcinoma cells to the expression of Snaill was not assessed. In fact, it has been demonstrated that Snaill is expressed by tumor stromal fibroblasts and endothelial cells. A recent study using immunohis-tochemical analysis of Snaill detected this protein only in occasional neoplastic epithelial cells situated at the edge of the tumor mass. Some of them had an undif-ferentiated phenotype while others retained an epithelial phenotype. Analysis of E-cadherin in these cells demonstrated that although most of the Snail-positive cells were negative for E-cadherin, others coexpressed both proteins and may rep­resent cells at the initial stages of the EMT, where E-cadherin gene transcription has been switched off but the protein remains. These results fit nicely in our hierarchical model of Snail 1 action described previously.

Regarding Snail2, initial in vitro studies suggested that Snail2 is a likely in vivo repressor of E-cadherin in breast cancer and participates in the metastatic poten­tial of melanoma. In human clinical samples, although the expression of Snail2 has been explored only occasionally, some data suggest that Snail2 expression might be an indicator of poor prognosis. Thus, a high expression of Snail2 mRNA in lung cancer tissue was significantly associated with postoperative relapse and shorter patient survival. In addition, positive expression of Snail2 was found in 58% of colorectal carcinomas. The positive expression of Snail2 was signifi­cantly associated with higher Dukes stage and distant metastasis and had a negative impact on overall patient survival. In breast cancer, increased Snail2 expression was associated with metastatic disease or tumor recurrence.

With regard to the expression of other EMT-mediating transcription factors in human tumors (e.g., SIP1, Twist), little information is currently available. SIP1 expression has been analyzed in gastric and gynecological cancer and in squamous cell carcinomas. In 20 intestinal-type gastric cancer samples, reduced E-cadherin expression was found in 60% of the cases, which correlated with up-regulation of SIP1 at the mRNA level. However, in this study, SIP1 overexpression could not be linked to down-regulated E-cadherin in diffuse-type gastric cancer. In ovarian cancer, expression of SIP1 correlated with worse outcome. Twist was studied in primary human gastric, breast, and colon cancers. In diffuse gastric cancer, overexpression of Twist correlated with N-cadherin expression, but this association was not found in intestinal gastric cancer or in colon cancer. In breast carcinomas, Twist overexpression was associated with poor outcome. Additionally, Twist expression was up-regulated in lobular breast cancer, a tumor characterized by the lack of E-cadherin expression. Since lobular carcinomas invade in a distinctive diffuse pattern (with cells infiltrating Indian-file or as single cells), the authors suggest that in this tumor type, Twist may be involved in silencing E-cadherin and promoting this EMT-like invasion pattern. However, there is evidence contrary to this assumption. First, lobular tumors consistently express cytokeratins (even in metastsic lesions) but rarely mesenchymal markers such as vimentin. Second, E-cadherin gene is irreversibly inactivated by genetic (mutations, allelic loss) and epigenetic (promoter hypermethylation) alterations in lobular breast cancer. Third, a recent report with a larger series of tumors (n = 144) failed to find such an association between Twist expression and the lobular histotype. In contrast, studies in prostate cancer appear to support a role for Twist in this tumor type. Thus, Twist was found highly expressed in the majority (90%) of prostate cancer tissues but only in a small percentage (6.7%) of benign prostate hyperplasia. In addition, the Twist expression levels were positively correlated with Gleason grading and metastasis, indicating its role in the development and progression of prostate cancer.

One important drawback on the studies discussed above is that most of them rely on detection of the factors by RT-PCR (or qRT-PCR) on the whole-tumor samples or on antibodies, the specificity of which has not been clearly established. Provided that EMT can be transient and/or spatiotemporally restricted in most carcinomas, additional studies in a larger series of tumors using highly specific reagents are clearly needed to establish the role of the various EMT factors definitively. It is also possible that overt EMT might not always be a general mechanism for a carcinoma to invade and metastasize. In fact, many carcinomas invade as multi-cellular aggregates and sometimes form glandular structures resembling normal epithelium, in which the carcinoma cells retain epithelial characteristics, includ­ing adherens junctions and apical-basal polarity. Therefore, these cells can invade and metastasize in a process known as collective migration} Thus, it is reasonable to assume that EMT may preferentially occur in specific biological contexts.