Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Circadian rhythms

The cancer connection

The Per2 gene is a core component of the circadian clock in mammals. It now seems that the mouse Per2 gene is also involved in suppressing tumours, through other genes that affect cell proliferation and death.

If the mantra in real estate is 'location, location, location', in genetics it would be 'phenotype, phenotype, phenotype'. There is simply no substitute for a detailed phenotypic analysis of a mutant strain (study of the overt manifestation of a mutated gene in the organism). This has the potential to reveal unanticipated — and sometimes truly surprising — relationships between genotype and phenotype, or between a primary phenotype and a secondary one. Such was the case for an analysis published recently in Cell by Lee and colleagues1.

The organisms under study here were mice in which both copies of the mPer2 gene were mutated — a genotype shown previously2 to cause a strong defect in circadian rhythms. Most organisms have endogenous 'clocks' that control rhythms of physiology and behaviour with roughly 24-hour (circadian) periodicity. But the period is shortened, and rhythmicity is lost, in the mPer2 mutant mice. This is reminiscent of the effects of the original perS or per0 mutations in fruitflies, described in the 1971 landmark paper from Konopka and Benzer3.

By observing their mutant mice for a couple of years, however, Lee and colleagues1 made an unexpected discovery: the animals were unusually cancer prone. At six months of age they began to show excessive cell proliferation in the salivary glands, as well as teratomas — tumours that originate from germ cells and comprise a mix of cell types. Thirty per cent of the mutant mice died before the age of 16 months, half of these from spontaneous lymphomas. In contrast, such lymphomas were first found in normal mice at the age of 20 months, a highly significant difference. The mutant animals were also more sensitive to γ-radiation, as indicated by premature hair greying and hair loss, and an increased rate of tumour formation — this last effect stemming, at least in part, from a decreased likelihood of cell death (apoptosis) in response to radiation.

So, what could be the story here? The current picture of the central circadian clock in animals is of a self-sustaining transcription–translation feedback loop, involving the transcription of key clock genes, their translation into protein, and the proteins' repression of transcription of the same key genes4 (as well as of downstream, clock-controlled genes). In fact, given the importance of post-translational mechanisms — such as protein phosphorylation and turnover — and the lack of translational control in the current picture, it might be more accurate to describe it as a macromolecular feedback loop. In any case, the mPer2 protein is a key clock component: it contributes to the circadian regulation of transcription of both mPer2 and downstream genes5.

All of which begs the question: is there a tight relationship between γ-radiation and clock genes? And could disruption of circadian transcriptional regulation cause the defects in cell proliferation and death (together termed cell growth) seen even in the absence of γ-radiation in mPer2 mutant mice? In other words, is there a transcriptional cascade from clock genes, to downstream growth-control genes, to growth-effector genes? The answers all appear to be yes.

Lee and colleagues' results show that, in normal mice, the expression of several core clock genes was rapidly and potently upregulated in the liver in response to γ-radiation. But in the mutants this response was absent or severely attenuated. Even more surprising was the authors' analysis of a few key genes concerned with cell growth. They found that expression of the Myc gene, as judged by levels of its messenger RNA, was circadian in wild-type liver; but in the mutants the expression pattern was modestly shifted and levels of Myc mRNA were dramatically increased. Moreover, experiments in cultured cells suggested that Myc transcription is directly regulated by the circadian clock. The authors also looked at the expression of cyclin D1 and Gadd45α, two Myc-regulated mRNAs, and found that the levels of both fluctuated in a circadian pattern in wild-type livers; in the mutants, both patterns were altered.

So Lee and colleagues propose that the key effect of inactivating mPer2 is to derepress Myc expression, leading to excessive cell growth and tumour formation. The effect is exacerbated by γ-radiation, which normally upregulates clock genes and thereby presumably leads to Myc repression. This fails in the mPer2 mutants. If these proposals are true, there are some testable predictions. First, overexpresssing Myc in an otherwise normal genetic background should have the same growth-promoting effects as mutating mPer2. Second, and more important, inhibiting Myc expression should suppress tumour formation in mPer2 mutants.

One caveat is that all of these experiments were performed under conditions of 12 hours' light, 12 hours' darkness, so it could be that the mRNA cycles were merely light-driven, not clock-driven. In this context, it is notable that Myc and Gadd45α were not identified as cycling genes (although cyclin D1 was) in three out of four microarray studies of liver mRNAs6,7,8,9. But the marked effects of the mPer2 mutation suggest that, at the very least, there is a strong connection between cell growth and the circadian clock. The discrepancy also reinforces the importance of taking microarray data — especially negative data — with a pinch of salt. In our opinion, careful biochemical analyses are more credible.

More generally, the new results1 have brought into proximity two previously disparate fields of study: circadian rhythms and cell-growth control. One reason why they have hitherto been infrequent bedfellows is that the mammalian circadian rhythm field has historically focused on the brain and, more narrowly, on the suprachiasmatic nucleus — the region of the hypothalmus that is essential for directing cycles of locomotor activity. Most adult neurons do not divide yet manifest potent molecular rhythms, suggesting that these rhythms are controlled separately from cell division. A similar conclusion stems from work on rat fibroblast cells, where molecular rhythms persist even if cell division is blocked10. And classical work in microbes indicates that rapid cell division can be completely uncoupled from the 24-hour timing of the circadian system in the same cells11.

On the other hand, cell division in microbes can be driven by the circadian cycle when the periodicity of division is not far from 24 hours. This suggests that the circadian system is important for proper growth control, and is consistent with the apparent circadian regulation of cell proliferation and apoptosis1. Moreover, the striking time-dependent response of wild-type mice to γ-radiation1 reinforces the potential importance of circadian principles in cancer and cancer therapies. It seems that cancer can be a direct consequence of the absence of circadian regulation. Perhaps this applies to other diseases, too. For instance, shift workers tend to have increased health problems. These could result indirectly from disruption of physiological systems that are under circadian control — but there might also be a direct connection to the clock.


  1. 1

    Fu, L., Pelicano, H., Liu, J., Huang, P. & Lee, C. C. Cell 111, 41–50 (2002).

    CAS  Article  Google Scholar 

  2. 2

    Zheng, B. H. et al. Nature 400, 169–173 (1999).

    ADS  CAS  Article  Google Scholar 

  3. 3

    Konopka, R. J. & Benzer, S. Proc. Natl Acad. Sci. USA 68, 2112–2116 (1971).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Allada, R., Emery, P., Takahashi, J. S. & Rosbash, M. Annu. Rev. Neurosci. 24, 1091–1119 (2001).

    CAS  Article  Google Scholar 

  5. 5

    Zheng, B. H. et al. Cell 105, 683–694 (2001).

    CAS  Article  Google Scholar 

  6. 6

    Akhtar, R. A. et al. Curr. Biol. 12, 540–550 (2002).

    CAS  Article  Google Scholar 

  7. 7

    Panda, S. et al. Cell 109, 307–320 (2002).

    CAS  Article  Google Scholar 

  8. 8

    Storch, K. F. et al. Nature 417, 78–83 (2002).

    ADS  CAS  Article  Google Scholar 

  9. 9

    Ueda, H. R. et al. Nature 418, 534–539 (2002).

    ADS  CAS  Article  Google Scholar 

  10. 10

    Balsalobre, A., Damiola, F. & Schibler, U. Cell 93, 929–937 (1998).

    CAS  Article  Google Scholar 

  11. 11

    Johnson, C. H. & Golden, S. S. Annu. Rev. Microbiol. 53, 389–409 (1999).

    CAS  Article  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Rosbash, M., Takahashi, J. The cancer connection. Nature 420, 373–374 (2002).

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.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

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