National Cancer Institute NCI Cancer Bulletin: A Trusted Source for Cancer Research News
March 8, 2011 • Volume 8 / Number 5

A Closer Look

The Epithelial-Mesenchymal Transition in Cancer: A Potentially Fatal Transformation?

The question remains one of cancer biology’s most perplexing: How do cancer cells from a primary tumor navigate to other parts of the body to form metastatic tumors?

Most solid tumors arise from epithelial cells, which normally stick together in tightly bound sheets to provide the structural foundations of many organs. In principle, epithelial cells lack the ability to escape those bonds and wiggle and jostle their way into nearby tissues, other organs, or the circulatory system. Yet, somehow cancerous epithelial cells—also known as carcinoma cells—do end up in other areas of the body, and the outcome is often dire: metastatic tumors are responsible for the vast majority of cancer-associated deaths.

Tumor cells from a highly aggressive ovarian cancer cell line before and after researchers introduced microRNA-429 (Images courtesy of Dr. John McDonald, Georgia Institute of Technology) Tumor cells from a highly aggressive ovarian cancer cell line are shown on the left. They have a mesenchymal shape (spindle-like) and express typical mesenchymal markers. On the right, cells from the same cell line reverted to an epithelial shape (rounded) and expressed epithelial cell markers after researchers introduced microRNA-429. (Images courtesy of Dr. John McDonald, Georgia Institute of Technology)

Over the past decade, accumulating evidence has shown that epithelial tumor cells can undergo an important change that transforms them into migratory mesenchymal cells. Some researchers now believe this conversion, known as the epithelial-to-mesenchymal transition, or EMT, can explain how epithelial tumor cells escape their primary residence in the breast, for instance, and travel to the liver, lungs, or bone marrow.

Recently, the role of EMT in cancer has come under even greater scrutiny because of evidence linking the process to the formation of so-called cancer stem cells, explained Dr. Suresh Mohla, associate director of NCI’s Division of Cancer Biology and program director of DCB’s Tumor Microenvironment Network (TMEN) program. Cancer stem cells, he noted, have been strongly associated with resistance to treatment and disease recurrence.

Although the evidence to support a role for EMT in cancer comes mostly from cell-line and animal-model studies, the evidence so far has been consistent with what’s been observed in human cancers, argued Dr. Jean-Paul Thiery, of the A*STAR Institute of Cellular and Molecular Biology in Singapore. A better understanding of EMT, he believes, could lead to an “improved rationale for effectively treating cancer.”

Gathering the Evidence

EMT is best known for its role in embryonic development, a time when epithelial cells readily bounce back and forth between epithelial and mesenchymal states as part of organ development. After the completion of embryogenesis, it was widely assumed that there was little biological use for EMT; that it was switched off and put to rest, perhaps permanently. At most, it appeared, EMT might be reactivated briefly during wound healing.

Over the last decade, however, researchers have found evidence that cancer cells reactivate EMT in an effort to escape their normal confines. For a carcinoma cell to escape and eventually form a micrometastasis in a distant tissue EMT may even be essential, explained Dr. Robert Weinberg, of the Whitehead Institute for Biomedical Research and the Massachusetts Institute of Technology, who is a TMEN investigator.

“If a carcinoma cell were still fully epithelial, to my mind it would be incapable of executing a number of the steps required for its physical dissemination from the primary tumor to a distant tissue site,” Dr. Weinberg said.

Cells that have undergone some degree of EMT are typically identified by the excess presence of certain molecules, such as vimentin and fibronectin, that promote a flexible structure and the ability to move, as well as by their shape, which is more spindle- and amoeboid-like than the typically rounded epithelial cell. Studies have also linked the activation of EMT to an overabundance of certain proteins and microRNAs. In some human tumors, an excess of these same proteins and microRNAs has been associated with more aggressive, metastatic disease.

Additional evidence of EMT-induced tumor cell migration in cancer has come from NCI-supported studies of mice in which an “imaging window” has been implanted in the skin above mammary tumors. In these studies, individual cells can be tracked over time with the use of light-sensitive proteins.

Digging Deeper into the Tumor Microenvironment

NCI is funding many research projects focused on learning more about the role of the epithelial-to-mesenchymal transition in cancer and, more generally, how the tumor microenvironment contributes to cancer initiation, progression, and metastasis.

A key component of this effort is NCI’s Tumor Microenvironment Network (TMEN) program. Launched in 2006, TMEN is composed of 10 research programs that are collectively working to better delineate the biological mechanisms that govern the interactions between a tumor and its microenvironment. Much of this work is focused on specific tumor sites, using human cancer tissues and models.

Fostering collaboration is a critical aspect of the TMEN program, explained Dr. Suresh Mohla, program director of TMEN. Research teams supported by the program are identifying and developing tools and resources, such as critical reagents and new cancer models, for use by the research community. Seven working groups help to facilitate the efforts of the 10 TMEN-funded programs.

One NCI-supported study found that cancer cells only traveled in certain areas of the tumor, depending on the makeup of the surrounding microenvironment, including the presence of blood vessels and tumor-infiltrating cells such as macrophages. The migrating tumor cells lacked the traditional symmetrical structure and shape of epithelial cells and thus appeared to have undergone EMT, said Dr. John Condeelis, also a TMEN-supported investigator whose lab at the Albert Einstein College of Medicine developed the mouse model.

The microenvironment surrounding cancer cells within a tissue is a critical component in the EMT, Dr. Condeelis explained. Signals from the microenvironment, in some cases caused by inflammation or perhaps in response to oxygen depletion in the tumor itself, may trigger the process of EMT and tumor cell migration. High-resolution imaging of this type, he continued, makes it easy “to believe in microenvironment-dependent tumor cell dissemination, because that’s what we see.”

Evidence of EMT and cell migration in human tumors has, however, been tough to come by. Studies suggest that mesenchymal tumor cells are not common and that they probably form only “at the [tumor’s] leading edge,” where the tumor interacts with its microenvironment, said Dr. Mohla. “Unless you are specifically looking for them, you’re probably not going to find them.”

Dr. Thiery pointed out that tumor cells that have undergone a complete EMT are hard to differentiate from other mesenchymal cells, such as fibroblasts, which are critical components of normal connective tissue. And Drs. Weinberg and Thiery both stressed that cancer cells may be in a mesenchymal state for only a short time before they revert to an epithelial state.

But some human evidence is beginning to emerge. In colorectal tumor samples, histopathology studies have identified “tumor budding,” small bunches of loosely grouped cancer cells at a tumor’s invading front. And recently published data from The Cancer Genome Atlas (TCGA) identified four distinct subtypes of glioblastoma, including a “mesenchymal type” that is highly aggressive and expresses markers indicative of EMT. In unpublished work, Dr. Thiery’s lab, using ovarian cancer samples acquired via TCGA, found that 45 percent of the tumors they analyzed had cells that expressed classic EMT markers.

A Transition to Treatment

Although there remain many unknowns about EMT’s significance in cancer, the phenomenon is already being used to guide new treatment approaches. OSI Pharmaceuticals, for example, is pursuing treatment approaches that are driven by EMT markers, explained Dr. David Epstein, the company’s chief scientific officer.

Retrospective analyses from two clinical trials involving the targeted therapy erlotinib (Tarceva) in patients with advanced non-small cell lung cancer (NSCLC) showed that responders had increased expression of the classic epithelial cell marker, E-cadherin. However, patients with low E-cadherin expression appeared to be less responsive to the drug. Further studies in cell lines, Dr. Epstein said, have indicated that mesenchymal-like NSCLC cells lose their reliance on erlotinib’s molecular target, the epidermal growth factor receptor.

“When we…think about EMT, we think about the rewiring that occurs as a function of that transition,” Dr. Epstein said. “That’s what’s kept us looking at the pharmacology, how we can use this to better design combination therapies and better delineate response markers in the clinical setting.”

Dr. Thiery believes EMT can also be exploited by forcing tumor cells that have become mesenchymal to revert to a more epithelial-like state. The approach ties back to the findings of NCI-supported work by Dr. Weinberg and others, primarily in breast cancer. Cells they have identified as cancer stem cells—cancer cells that on their own can produce other tumor cells that then further differentiate and form tumors—often express mesenchymal-like markers.

In the tumor types in which they’ve been identified, cancer stem cells have demonstrated resistance to current therapies and are associated with disease recurrence. Thus, forcing these cells to return to an epithelial-like state could potentially make them more responsive to existing therapies.

In January, researchers from the Georgia Institute of Technology reported some success along these lines. They found that introducing a specific microRNA into a highly metastatic ovarian cancer cell line changed the appearance of these mesenchymal-like cancer cells, making them more rounded and epithelial-like, a transformation that was accompanied by decreased levels of mesenchymal markers and an increase in E-cadherin levels.

There is still a tremendous amount of research to do around EMT, Dr. Condeelis acknowledged. “The next step is to better understand the context of EMT in normal embryonic development and why these processes are repeated in tumors,” he said. “We’re still missing much of the story.”

Carmen Phillips