Exploiting collateral damage

Some mutations in tumour cells play no part in causing cancer, but they generate cellular weak spots that may allow tumour cells to be selectively killed by drugs. See Article p.337

The central challenge in anticancer therapy is to kill tumour cells without harming other cells of the body. A defining property of cancer cells is that they carry genetic changes that distinguish them from normal cells and that should, at least in theory, allow them to be specifically killed by drugs that target them. But in practice this represents a formidable problem. On page 337 of this issue, Muller et al.1 outline a potential strategy that exploits 'passenger' mutations. These mutations can remove genetic redundancies in cancer cells, making them vulnerable and so facilitating their destruction.

Most cancer research focuses on the genetic mutations that cause cancer — 'driver' mutations that typically activate cancer-causing oncogenes or inactivate tumour-suppressor genes. Such mutations make logical drug targets, and, in recent decades, an explosion of knowledge about their mechanism of action has led to the development of several anticancer drugs. Examples include the successful inhibition of the proteins encoded by the oncogenes BCR-ABL, BRAF and ERBB2 in patients with chronic myelogenous leukaemia, melanoma and breast cancer, respectively2,3. But in many other cancers it has proved difficult to target the driver mutations effectively, even though they have been known for a long time.

An alternative approach would be to target the vulnerabilities created by passenger mutations. Although the development of a tumour may depend on only a small number of driver mutations, tumour cells usually also contain other genetic changes that accumulate as the tumours develop, often through exposure to environmental mutagens such as tobacco smoke and ultraviolet light4. Solid tumours can contain tens of thousands of such mutations4. The key point is that these passenger mutations also distinguish cancer cells from normal cells. So if the presence of a passenger mutation means that a cell becomes sensitized to the effects of a particular drug, then this should allow the tumour cells to be specifically killed.

Muller et al. outline one way in which this might be achieved. Many essential processes in biology are performed by two or more genes encoding proteins that have very similar biochemical functions. This 'partial redundancy' arises through the duplication of genes. It is a widespread phenomenon, and is maintained in genomes over vast evolutionary timescales5. One of the effects of partial redundancy is that the loss of one gene sometimes has little effect unless a second, related gene is also inhibited. Muller and colleagues' idea is that, if a passenger mutation removes or inactivates a duplicated gene, the tumour cells will become exquisitely sensitive to inhibition of the second, partially redundant gene (Fig. 1).

Figure 1: Therapy in the passenger seat.

The genomes of cancer cells often contain 'passenger' mutations, which are not directly related to the disease but which can create cellular vulnerabilities that allow the tumour cells to be killed by drugs targeting them. a, For example, Muller et al.1 show that in the case of the brain cancer glioblastoma, a cancer-causing deletion of a tumour-suppressor gene also leads to loss of the ENO1 gene, which encodes the enzyme enolase. This makes the tumour cells dependent on ENO2, which also encodes enolase but is expressed at lower levels than ENO1 in brain cells. b, Normal cells will survive treatment with a drug that inhibits enolase activity (indicated by the green circle), because of their high expression of ENO1. The low levels of enolase in the tumour cells, however, mean that they will be killed by this treatment.

To test this idea, the authors hunted for passenger mutations that resulted in the deletion of a duplicated gene in a type of brain cancer called glioblastoma. They found that, in their tumour samples, the gene ENO1 was deleted in several cases — most probably because it is located in a region of the genome that also contains one or more tumour-suppressor genes that, when deleted, promote uncontrolled proliferation of the tumour cells. ENO1 is one of three human genes that encode enolase, an enzyme required for the metabolic pathway that converts sugar to energy. These genes originated from duplications of an ancestral enolase gene more than 500 million years ago, and each gene is expressed to varying extents in different tissues. In the brain, for example, most enolase activity is provided by the ENO1 protein, with a smaller contribution from ENO2.

Consistent with their hypothesis, the authors found that glioblastoma cells in which ENO1 is deleted are much more sensitive to inhibition of ENO2 than other cells. They tested this sensitivity in cultured cells using two methods of inhibition — a known small-molecule inhibitor of enolase and a short hairpin RNA molecule (shRNA) that binds specifically to enolase-encoding messenger RNAs and interferes with the production of enolase proteins. The authors found that in normal cells, because of the high levels of ENO1 protein present, ENO2-specific shRNA molecules or the enolase inhibitor had little effect. But in glioblastoma cells in which ENO1 was deleted, both treatments killed the cells (Fig. 1). Indeed, when the authors inhibited ENO2 activity in the ENO1-deficient glioblastoma cells, the cells could no longer form tumours when transferred into mice.

The study therefore provides a powerful proof of concept that the loss of redundancy in cancer genomes can be exploited to selectively kill tumour cells. In glioblastoma, the strategy works because the deleted duplicate gene (ENO1) normally provides most of the enolase activity, and there seems to be no compensatory increase in protein production from the second gene (ENO2). But compensatory changes in gene expression from duplicate genes may actually be common6,7, and in other types of cancer the overall difference in protein activity between the tumour and normal cells may be much harder to target therapeutically. Moreover, the strategy must still be tested in animal models and in patients, particularly as a combination therapy with other drugs, because compensatory increases in the expression of the remaining gene duplicate may lead to the rapid development of drug resistance.

Finally, although not mentioned by Muller and colleagues, the vulnerabilities created by common passenger mutations are likely to extend far beyond the loss of protein redundancy8. Systematic screens to identify 'synthetic lethal' interactions9,10, in which a drug or a gene inhibition kills only cells carrying a common passenger mutation, may be one way to identify these additional vulnerabilities. For instance, another recent study11 presents an alternative strategy for targeting passenger mutations. It takes advantage of situations in which the activity of a single gene is reduced in cancer cells, for example through deletion of one of the two copies of the gene, rendering the cells highly sensitive to further inhibition of the same gene.

Muller and colleagues' results provide an impetus for two major conceptual advances: first, that the loss of protein redundancy provides a therapeutic opportunity to kill specific cells, and second, that passenger mutations may be the Achilles heel of cancer genomes. In our opinion, this second idea is the more important, and the plethora of passenger mutations present in cancer genomes should create many opportunities for personalized therapies. This underlines how crucial it is that, rather than simply focusing on cancer-causing genes and pathways, researchers also consider the therapeutic opportunities created by common passenger mutations.


  1. 1

    Muller, F. L. et al. Nature 488, 337–342 (2012).

    CAS  ADS  Article  Google Scholar 

  2. 2

    Sellers, W. R. Cell 147, 26–31 (2011).

    CAS  Article  Google Scholar 

  3. 3

    Workman, P. & de Bono, J. Curr. Opin. Pharmacol. 8, 359–362 (2008).

    CAS  Article  Google Scholar 

  4. 4

    Stratton, M. R., Campbell, P. J. & Futreal, P. A. Nature 458, 719–724 (2009).

    CAS  ADS  Article  Google Scholar 

  5. 5

    Vavouri, T., Semple, J. I. & Lehner, B. Trends Genet. 24, 485–488 (2008).

    CAS  Article  Google Scholar 

  6. 6

    Kafri, R., Bar-Even, A. & Pilpel, Y. Nature Genet. 37, 295–299 (2005).

    CAS  Article  Google Scholar 

  7. 7

    Burga, A., Casanueva, M. O. & Lehner, B. Nature 480, 250–253 (2011).

    CAS  ADS  Article  Google Scholar 

  8. 8

    Lehner, B. Trends Genet. 27, 323–331 (2011).

    CAS  Article  Google Scholar 

  9. 9

    Hartwell, L. H., Szankasi, P., Roberts, C. J., Murray, A. W. & Friend, S. H. Science 278, 1064–1068 (1997).

    CAS  ADS  Article  Google Scholar 

  10. 10

    Chan, D. A. & Giaccia, A. J. Nature Rev. Drug Discov. 10, 351–364 (2011).

    CAS  Article  Google Scholar 

  11. 11

    Nijhawan, D. et al. Cell (2012).

Download references

Author information



Corresponding author

Correspondence to Ben Lehner.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Lehner, B., Park, S. Exploiting collateral damage. Nature 488, 284–285 (2012).

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.