Cell biology

Healthy skin rejects cancer

Article metrics

Live imaging shows that healthy skin cells surround and expel neighbours that have cancer-promoting mutations, revealing that tissues can recognize and eliminate mutant cells to prevent tumour initiation. See Letter p.334

Skin must constantly renew itself to replace damaged cells in its role as a barrier against the outside world. Such turnover is largely beneficial, but perpetual proliferation also increases the risk that cancer-causing mutations will arise and lead to uncontrolled growth. On page 334, Brown et al.1 use state-of-the-art imaging techniques to demonstrate that healthy skin cells routinely recognize and eliminate neighbouring mutant cells, pointing to an innate cellular ability to guard against over-proliferation and tumour initiation.

The authors focused on stem cells in hair follicles in the skin, which proliferate to mediate hair growth and repair of the outer skin layer. They used a genetic mouse model in which a few hair-follicle stem cells (HFSCs) were engineered to express a mutant form of the gene that encodes the protein β-catenin, which increases cell proliferation in many tumour types2. They followed a protocol3 developed by their group in 2012 to track individual hair follicles over the course of weeks using fluorescence microscopy, which enabled them to visualize the behaviour of mutant HFSCs in live mice.

Brown et al. observed that individual mutant HFSCs created abnormal outgrowths in the hair follicle, as expected. But rather than causing a tumour, the outgrowths eventually regressed and disappeared (Fig. 1). The authors repeated variations of this basic experiment: increasing the number and types of cell carrying mutant β-catenin; inducing expression of a different mutant gene, Hras, to increase proliferation another way; and burning the hair follicle with a laser to test the cells' response to physical damage. In each variation, the abnormal cells — whether stem cells or differentiated cells — were almost always eventually surrounded by non-mutant cells and expelled from the follicle.

Figure 1: Dynamics of mutant-cell growth in the hair follicle.

a, Brown et al.1 imaged growing hair follicles in the skin of mice in which a few stem cells had been induced to express mutant genes that promote hyperproliferation. b, The authors found that the mutant population grew and spread. c, Healthy neighbouring cells then increased their own proliferation rates to completely encompass the mutant population in the middle of the follicle. d, The healthy cells eventually expelled all mutant cells (black arrow), thereby protecting healthy tissue growth.

The authors' results suggest an active role for healthy cells in policing their neighbours. To test this idea, the researchers engineered the mutant cells such that they could not secrete Wnt proteins, with which the cells signal to neighbours to promote proliferation. And, in a separate experiment, the authors engineered the healthy neighbours to prevent them from proliferating. In both cases, expulsion of the mutants was impaired.

Brown and colleagues' findings point to a previously unknown antitumour mechanism in the skin. However, the study also raises many questions. For instance, how do healthy cells sense neighbouring mutants? The authors' experiments involving Wnt proteins point to a mechanism by which such sensing might occur. It remains unclear, however, whether the protective response occurs in other tissues that have different self-renewal strategies, and, if so, whether they use similar signalling mechanisms.

Do healthy neighbours also eliminate cells that harbour proliferation-driving mutations other than those tested in the current study? In human skin, mutations in β-catenin tend to cause benign tumours rather than malignant cancers, and HRAS mutations are found in only 6% of skin cancers4 — much less frequently than mutations in the gene Patched, for instance, which occur in about 30% of cases of the skin cancer basal cell carcinoma5. Perhaps each tissue has variable mutation-quashing capacity: skin might handle mutations in β-catenin better than those in Patched, and better than the large intestine, where β-catenin-pathway mutations are commonly found in colorectal cancer6.

How do mutant cells escape this 'neighbourhood watch' mechanism, in the cases in which they do go on to form tumours? One explanation is cell plasticity. Cell identity is proving to be fluid — cells that differentiate and stop dividing do not necessarily remain in this state; they can often become progenitors once again if needed. This phenomenon has been demonstrated in the intestine7 and stomach8, which, like the skin epidermis, renew rapidly. In organs that lack active stem-cell populations, such as the pancreas, reversion of differentiated cells is the only way in which large-scale tissue repair can happen. Such plasticity can make mutations silent in certain contexts and dangerous in others.

As an example, consider mutant KRAS (a relative of HRAS). This mutant is found in nearly all pancreatic ductal adenocarcinomas, which are the most common form of pancreatic cancer. However, KRAS mutations have little effect on differentiated, non-proliferating pancreatic cells until the cells re-enter the cell cycle following injury9. Once they are proliferating, mutant KRAS prevents them from redifferentiating, causing growth that leads to tumour formation7. Thus, mutations can hide and accumulate in the differentiated cells of a tissue, escaping detection until they are unmasked by damage that triggers a cell-identity switch and makes the mutation more deleterious. In the skin, cells that accumulate mutations but do not show increased levels of proliferation or Wnt signalling might evade the neighbourhood watch until their mutation load is sufficient to overcome the protective mechanism uncovered by Brown and co-workers.

The authors' analysis of how cells deal with neighbours that have acquired single mutations is unusual — other papers in this area have typically focused on how neighbouring cells react to established tumours. Future studies should continue to tread this path, to improve our understanding of the phenomenon the researchers have uncovered. For instance, what allows mutations in genes of the RAS family to initiate tumours in the pancreas, when they are unable to overcome this newly described protective mechanism in the skin? Among other possibilities, the crucial variable could be that normal neighbouring cells in the skin and pancreas respond differently to various types of mutant; that the mutations arise in different cellular contexts (for example, in cells undergoing dedifferentiation in the pancreas, but not in the skin); or that there are typically higher levels of damage or inflammation in the pancreas during mutant-cell expansion.

Finally, although Brown and colleagues' demonstration of cell competition is one of the first in vivo in mammals10, there are robust data on interactions between mutant and healthy cells in fruit flies. In flies, mutations that disrupt cell polarity and increase proliferation can trigger normal cells to outcompete their highly proliferating neighbours — a phenomenon dubbed neoplastic tumour suppression11. Researchers can now use the model described in the current work to investigate whether the extensively studied signalling pathways and mechanisms in flies might be conserved across species and tissue contexts. In short, the innate cellular defence mechanism discovered by Brown and colleagues might have important implications for our understanding of cell plasticity, cell competition and tumour initiation. Footnote 1


  1. 1.

    See all news & views


  1. 1

    Brown, S. et al. Nature 548, 334–337 (2017).

  2. 2

    Chan, E., Gat, U., McNiff, J. M. & Fuchs, E. Nature Genet. 21, 410–413 (1999).

  3. 3

    Rompolas, P. et al. Nature 487, 496–499 (2012).

  4. 4

    Prior, I. A., Lewis, P. D. & Mattos, C. Cancer Res. 72, 2457–2467 (2012).

  5. 5

    Gailani, M. R. et al. Nature Genet. 14, 78–81 (1996).

  6. 6

    Morin, P. J. et al. Science 275, 1787–1790 (1997).

  7. 7

    Mills, J. C. & Sansom, O. J. Sci. Signal. 8, re8 (2015).

  8. 8

    Burclaff, J., Osaki, L. H., Liu, D., Goldenring, J. R. & Mills, J. C. Gastroenterology 152, 762–766 (2017).

  9. 9

    Guerra, C. et al. Cancer Cell 11, 291–302 (2007).

  10. 10

    Kon, S. et al. Nature Cell Biol. 19, 530–541 (2017).

  11. 11

    Yamamoto, M., Ohsawa, S., Kunimasa, K. & Igaki, T. Nature 542, 246–250 (2017).

Download references

Author information

Correspondence to Joseph Burclaff or Jason C. Mills.

Related links

Related links

Related links in Nature Research

Regenerative biology: Skin, heal thyself

Biological techniques: An embryonic view of tumour development

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Burclaff, J., Mills, J. Healthy skin rejects cancer. Nature 548, 289–290 (2017) doi:10.1038/nature23534

Download citation

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.