Studies Reveal a New Twist in EMT's Role in Metastasis
By Anette Breindl
Arguments have gone back and forth for some time now on whether cancer cells use a developmental program called the epithelial to mesenchymal transition, or EMT, to metastasize.
Two new studies suggested they do – but only for the first half of the metastatic process, pulling up stakes and leaving the primary tumor. In order to actually establish metastases at distant sites, though, it is critical that those wandering cells turn the EMT program back off.
In retrospective studies, an EMT gene signature in the primary tumor correlates with a poor prognosis. And it certainly makes sense that transitioning to a mesenchymal-like state would help tumor cells metastasize. Mesenchymal cells can move more easily and are less connected to their neighbors than epithelial ones, and are more stem cell-like in certain respects.
The thing is, though, that those EMT cells are nowhere to be found in the metastases themselves.
"Especially in humans, it is very clear that the distant metastases have very similar morphology to the primary tumor," Jing Yang told BioWorld Today. Yang is at the University of California at San Diego and the senior author of one of the papers, which appeared together in the Nov. 29, 2012, online and December 2012 print editions of Cancer Cell.
If metastases are derived from cells that have undergone the epithelial to mesenchymal transition, she elaborated, one might expect that "they should all look like sarcomas," which are derived from mesenchymal cells. Instead, metastases look like the epithelial-derived carcinomas that they are – to the point that a lung metastasis of a breast tumor is clearly more like that breast tumor than like a lung tumor.
For their studies, Yang and her team developed mice in which they could turn on the Twist1 gene, which starts up the gene expression program that supports the epithelial to mesenchymal transition. EMT has been challenging to study, because most cancer cell lines do not undergo it – they are in a permanently mesenchymal-like state. Yang noted that her team's model "might be useful for studying a variety of metastases" because the development of those metastases can be precisely controlled, a feat that is true in "very few mouse tumor models."
They then induced skin cancer in such mice by exposing their skin to chemicals, and turned Twist1 on, either temporarily or permanently. Mice whose Twist1 was turned on for a period of time developed lung metastases, while those whose Twist1 remained on permanently developed next to no such metastases.
The team also looked at matched samples from the primary tumors and lymph node metastases of a group of breast cancer patients, and found that while primary tumor cells had high levels of Twist1 expression, Twist1 expression was once again low in most of the metastases. And the authors of the second paper identified a new transcription factor, Prrx1 , whose suppression let cells revert to an epithelial-like state and was necessary for metastases to form in mice with tumors.
Yang's interpretation of the data is that EMT is important for metastasis to begin, but that in order to grow into detectable metastases, the cells that seed those metastases have to revert to a more epithelial-like state once they have colonized the lung or other organs.
If that interpretation is correct, it suggests that suppressing EMT – which is one therapeutic approach that is being tested to prevent metastases – might actually backfire. If patients already have micrometastases, or even tumor cells that have pulled up stakes and are now circulating in the bloodstream, suppressing EMT could be exactly the wrong thing, pushing such cells into becoming larger metastases rather than preventing such metastases.
Yang said she thinks the late reversal of EMT might also explain tumor dormancy, where metastases suddenly develop many years after a tumor has been treated with seeming success.
For now, it is not clear whether EMT reversal happens by chance in some cells or whether the tumor microenvironment, or active processes in either the tumor cells themselves or the microenvironment are to blame. Yang favors the notion that active processes underlie EMT reversal, although for now, there's no evidence one way or another.
"Is it by chance? I doubt it," she said. "But I don't have any data to show you."