News & Views | Published:


The to and fro of tumour spread

Nature volume 493, pages 487488 (24 January 2013) | Download Citation

Two studies shed light on the role of cellular transitions between the epithelial and mesenchymal states during cancer metastasis, and provide food for thought as to which cellular processes should be targeted in cancer treatment.

The spread of cells from the primary site of a solid tumour to distant sites remains the major cause of disease and death associated with these cancers. For tumour cells to spread, or metastasize, they must modify their 'anchored' state and detach from their neighbouring cells; migrate through tissues into the blood and lymph systems; survive in these circulation systems; and then leave the vessels at an appropriate site to form another tumour1. Many of these events are favoured by conversions between two cellular states — the epithelial and mesenchymal phenotypes. But the role of these transitions in cancer metastasis is controversial. Writing in Cancer Cell, Tsai et al.2 and Ocaña et al.3 help to clarify this issue.

Most cancers in adults are carcinomas, which arise from abnormal growth of epithelial cells that line the surfaces and cavities of organs such as the breasts, colon, lungs and liver. Epithelial cells are fully differentiated and non-mobile, whereas mesenchymal cells migrate easily. In the developing embryo, a process called the epithelial–mesenchymal transition (EMT) allows epithelial cells to take on mesenchymal-cell characteristics and move sites, after which they revert back to an epithelial phenotype through the mesenchymal–epithelial transition (MET), and continue to form body structures. EMT is also involved in metastasis, but whether it is an absolute requirement is controversial, in part because metastatic tumours display an epithelial phenotype and lack mesenchymal markers4.

We have coined the term epithelial–mesenchymal plasticity for the dynamic interchange between epithelial and mesenchymal phenotypes5. Ten years ago, it was hypothesized that tumour cells that had undergone EMT and disseminated would then go through MET to successfully colonize a secondary site and form a metastatic tumour6; and preliminary evidence for MET has been reported in cancers of the colon, breast, bladder and prostate7. Although the abundance of studies on EMT still dwarfs those on MET, it is emerging as a crucial and possibly rate-limiting step in metastasis, and this is supported by the two new papers.

Previous work has shown that the EMT-inducing transcription-factor protein Twist1 is essential for spontaneous metastasis in a mouse model of mammary cancer8. Tsai and co-workers have built on these findings using a mouse model of squamous-cell carcinoma in which Twist1 expression can be turned on and off. The authors found that Twist1 induces EMT and tumour-cell transit into the bloodstream (Fig. 1), but that metastases did not form if the cancer cells could not switch Twist1 off after disseminating. This suggests that halting EMT, and thereby allowing MET, is required for completion of the metastatic process.

Figure 1: Cellular transitions in cancer metastasis.
Figure 1

Dissemination of cells from a primary solid tumour is facilitated by the epithelial–mesenchymal transition (EMT). This process allows epithelial tumour cells (carcinoma cells), which are fully differentiated and non-mobile, to acquire the more invasive characteristics of mesenchymal cells. Tsai et al.2 show that EMT is induced by a transcription factor called Twist1, and that Twist1 expression also increases the number of metastable cells, which express both epithelial and mesenchymal characteristics, in the bloodstream of mice. However, the authors found that sustained Twist1 expression prevents the formation of secondary tumours, possibly by preventing the cells from undergoing the mesenchymal–epithelial transistion (MET) that is needed for them to revert to the epithelial state and form metastases. Ocaña et al.3 characterize another EMT-inducing transcription factor, Prrx1, and show that MET can be achieved with Prrx1 suppression, even during continued Twist1 expression.

Ocaña et al. have characterized a new EMT-inducing transcription factor, Prrx1, that they show drives EMT during embryonic development in chicks. The authors also show that, similarly to other embryonic EMT-driving factors, Prrx1 seems to play a part in the invasive, migratory phenotype of breast carcinoma cells undergoing metastasis. Moreover, the authors found that forced, continuous expression of Prrx1 blocked the capacity of otherwise metastasis-competent cells to produce metastatic tumours, and that Prrx1 suppression is needed for MET to proceed — consistent with Tsai and colleagues' observations for Twist1 (Fig. 1).

A notable advance in our appreciation of the role that EMT may have in solid-tumour biology came from observations that 'breast cancer stem cells' (cells isolated from clinical samples that have particularly strong cancerous properties and the ability to initiate new tumours from small numbers of cells) exhibit a mesenchymal profile9. Conversely, both normal and cancerous mammary epithelial cells that are induced to undergo EMT become more stem-cell-like8. This relationship was further refined when cells with progenitor-cell properties — those committed more towards differentiation than the mesenchymal-like stem cells — were seen to have higher cancerous potential9,10.

Adding to this picture, Ocaña et al. show that the stem-cell-like features of carcinoma cells can be segregated from EMT by manipulating the expression of Prrx1. Remarkably, they show that abrogation of Prrx1 in carcinoma-derived cells (called BT-549 cells) that would not normally form tumours when injected into a mouse makes them both tumorigenic and metastatic. Moreover, they show that reducing Prrx1 levels in cancer cells causes a reduction in EMT, but a gain of stem-cell activities. These findings are consistent with recent reports that pluripotency (the ability to differentiate into various cell types), which is associated with the epithelial phenotype more than the mesenchymal phenotype, is a major driver of both cancerous and metastatic potential10,11. It is thus implicit that exhibiting the enhanced malignant properties associated with the mesenchymal stem-cell state requires an ability to spontaneously progress towards an epithelial state with pluripotent characteristics.

An additional twist to the tale comes from the fact that some tumour cells simultaneously show mesenchymal and epithelial characteristics12,13. Also called the intermediate, or metastable, phenotype, it has long been recognized that co-expression of both phenotypes may offer cells a form of pluripotency, allowing them to dynamically adjust to the circumstances they encounter. Such plasticity seems to be key for the full spectrum of metastatic competence, because cells that are rigidly locked into one state or another are less capable of metastasis or even of primary-tumour growth, as seems to be the case with BT-549 cells. Furthermore, other studies10,14 have shown that cooperation between mesenchymal and epithelial variants of some cancer cells can allow the epithelial variants to escape their tumour site and metastasize. Thus, the combined evidence from these studies and the present papers suggests that both mesenchymal and epithelial phenotypes are required for metastatic competence, and that cancer cells must have sufficient plasticity between these phenotypes to spread.

The reports by Tsai et al. and Ocaña et al. leave little doubt as to the importance of MET and/or the epithelial phenotype for carcinoma metastasis. As stated by both sets of authors, this introduces the question of whether the therapies that many laboratories are chasing — treatments designed to inhibit cancer cells in the mesenchymal state — might in fact have a stimulatory effect on established metastases, or even activate dormant cancer cells. Models and experimental designs, such as those used for these studies, will be crucial to resolving this question, and they herald an exciting chapter in our understanding of metastasis.


  1. 1.

    Semin. Cancer Biol. 21, 71 (2011).

  2. 2.

    , , , & Cancer Cell 22, 725–736 (2012).

  3. 3.

    et al. Cancer Cell 22, 709–724 (2012).

  4. 4.

    , & Cancer Res. 65, 5996–6000 (2005).

  5. 5.

    & Nature Med. 17, 1048–1049 (2011).

  6. 6.

    Nature Rev. Cancer 2, 442–454 (2002).

  7. 7.

    , , & Cancer Metastasis Rev. 31, 469–478 (2012).

  8. 8.

    et al. Cell 117, 927–939 (2004).

  9. 9.

    , & J. Mammary Gland Biol. Neoplasia 14, 29–43 (2009).

  10. 10.

    et al. J. Clin. Invest. 122, 1849–1868 (2012).

  11. 11.

    , , , & Stem Cells 30, 292–303 (2012).

  12. 12.

    , , & J. Cell Biol. 172, 973–981 (2006).

  13. 13.

    & Am. J. Pathol. 174, 1588–1593 (2009).

  14. 14.

    et al. Cancer Res. 68, 10377–10386 (2008).

Download references

Author information


  1. Bryce J. W. van Denderen and and Erik W. Thompson are at St Vincent's Institute, Fitzroy, Victoria 3065, Australia.

    • Bryce J. W. van Denderen
    •  & Erik W. Thompson
  2. B.J.W.v. D. is also in the Department of Medicine, University of Melbourne

    • Bryce J. W. van Denderen
  3. E.W.T. is in the Department of Surgery, University of Melbourne, St Vincent's Hospital, Fitzroy.

    • Erik W. Thompson


  1. Search for Bryce J. W. van Denderen in:

  2. Search for Erik W. Thompson in:

Corresponding authors

Correspondence to Bryce J. W. van Denderen or Erik W. Thompson.

About this article

Publication history



Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing