NEWS AND V IEWS

In an intriguing study reported on page 93 of this issue, Shima et al.1 describe a recessive mutation, Chaos3, identified in a genetic screen for spontaneous chromosomal instability in mice resulting from ENU mutagenesis. By con-ventional genetic mapping, the authors local-ized the mutant allele to an interval containing Mcm4, which encodes a protein required for initiation of DNA replication. Sequencing the Mcm4 gene in Chaos3 mice uncovered a muta-tion changing a conserved phenylalanine in the Mcm4 protein to isoleucine. After show-ing that the Chaos3 allele did not complement an Mcm4 null allele, Shima et al. established the functional significance of the mutation by showing that the same substitution renders yeast Mcm4 hypofunctional in DNA replication. Notably, mouse embryonic fibroblasts (MEFs) from Mcm4Chaos3/Chaos3 mice showed slightly slower cell cycle kinetics, with fewer cells in S phase in asynchronous populations, and also showed an increase in chromatid and chro-mosome breaks when subjected to replication stress. Remarkably, female Mcm4Chaos3/Chaos3 mice also showed a high incidence of spon-taneous mammary adenocarcinomas. Based on these results, the authors conclude that the Chaos3 mutation is a hypomorphic mutation of Mcm4 that promotes genomic instability and leads to cancer.

Defective licensingOrigins of replication in eukaryotic chromo-somes are licensed by the combined action of cellular replication initiation factors ORC, Cdc6 and Cdt1, resulting in the loading of the Mcm2-7 complex, the presumptive replicative helicase (Fig. 1a). Because of the essential nature of these genes in yeast, there has been a pronounced lack of interest in the field in making mice mutant for these genes. The few replication initiation factors that have been knocked out, Mcm4 (in this study) and Cdc45, have led to embry-onic lethality in mice1,2. However, the study by

Shima et al. shows the amazing power of hypo-morphic genetics to study a gene’s function in the intact animal. A hypomorphic mutation in a Drosophila melanogaster replication initiation factor, latheo (also known as Orc3), implicated this gene product in a neural function unrelated to DNA replication3. In the current study, the hypomorphic mutation implicates Mcm4 in genomic instability, a function more closely related to replication initiation. Although they allow for the possibility that the cancer predis-position could arise from a replication-indepen-dent function of Mcm4, such as transcriptional regulation or sensing of a checkpoint, the authors prefer the hypothesis that decreased licensing results from a decrease in MCM pro-teins, making the cells susceptible to chromo-some breaks. Decreasing the amount of MCM protein in a Xenopus laevis replication reaction or in C. elegans indeed allows DNA replication and normal cell cycling, but genomic stability is seriously compromised when the system is fur-ther stressed by replication stressors4.

Further work is needed to test whether the phenotypes in Mcm4Chaos3/Chaos3 mice are com-plemented by a molecular clone of wild-type

Mcm4. Without this, we cannot rule out the very rare possibility that the results are con-founded by a mutation in a gene tightly linked (even overlapping) with the Mcm4 locus. Finally, the Chaos3 allele seems to have a dom-inant phenotype, too, because Mcm4Chaos3/+ mice have a two- to fivefold-higher frequency of micronuclei in their erythrocytes compared with wild-type mice, which is not seen in Mcm4+/– mice. What is the mechanism of this dominant phenotype? Is licensing truly inhib-ited in the McmChaos3/+ mice? If so, this result might imply that Mcm4 acts in the Mcm2-7 complex in units of at least two (as in two hexamers) such that the protein encoded by mutant Mcm4 can poison the activity of a hex-amer containing the wild-type Mcm4 protein. Or does this result point to a function of Mcm4 (or a tightly linked gene) unrelated to its role in DNA replication? Clearly, many complemen-tation experiments are needed, particularly to ensure that there is no allele that separates the formation of micronuclei or predisposition to mammary carcinomas from the origin licens-ing defect, but at least these questions can now be asked.

Chaotic license for genetic instability and cancerAnindya Dutta

Replication initiation factors are essential for cell proliferation and thus are not expected to be disrupted in cancers. Challenging this notion, a new study in mice shows that a hypomorphic mutation in the gene encoding the replication initiation factor Mcm4 leads to genetic instability and predisposes to mammary adenocarcinomas.

Anindya Dutta is at the Department of Biochemistry and Molecular Genetics, University of Virginia, Charlottesville, Virginia 22908, USA.e-mail: ad8q@virginia.edu

Kat

ie R

is

a

ORC ORC

Cdc6

Cdt1

Geminin

Chk1Chk2

p53

p27 Cdk2

Cdc25/CDK

ORC

Cdc6

Cdt1

Mcm

7

3

2 5

4 6

Normal licensing

b Excessive licensing

Re-replication

c Inadequate licensing

Prereplicativecomplex

Figure 1 Regulation of DNA replication by licensing factors. (a) Licensing of an origin of replication by the six-subunit ORC, Cdc6 and Cdt1 (inhibited by geminin), leading to loading of the Mcm2-7 complex on DNA. (b,c) Excessive licensing (re-replication) (b) or inadequate licensing (c) activates checkpoint pathways to stop the cell cycle or promote apoptosis.

10 VOLUME 39 | NUMBER 1 | JANUARY 2007 | NATURE GENETICS

©20

07 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

ureg

enet

ics

NEWS AND V IEWS

Checkpoints and cancerJust as decreased licensing is shown here to promote genetic instability and cancer, exces-sive licensing in the same cell cycle seems to have similar effects. For example, repeated licensing in the same cell cycle (re-replication) is promoted by an excess of the licensing factor Cdt1 and its partner Cdc6 or by a decrease in an inhibitor of licensing, geminin5,6. Excess Cdt1 in mice promotes cancer7, whereas excess Cdt1 or Cdc6 in human cancer is correlated with excess genetic instability and a poorer prog-nosis8. In several studies, the genetic instability arising from deregulation of replication licens-ing activates checkpoint pathways that stop cell proliferation and eventually lead to cell death (Fig. 1b). For example, re-replication activates both the S-G2 checkpoint and the p53 tumor suppressor, and cancers arising in the context of Cdt1 overexpression often have mutations in the gene encoding p53 (reviewed in refs. 5,6). Conversely, inadequate licensing imposed experimentally by overexpression of geminin or knockdown of Orc2 leads to a G1 arrest that, at least in one case, is mediated by the stabilization of the cyclin-dependent kinase 2 (Cdk2) inhibi-

tor p27 (Fig. 1c)9,10. Thus, one point of interest will be whether Mcm4Chaos3/Chaos3 cells activate checkpoint pathways and whether mutations in p53, or activated degradation of p27, are important for the development of mammary carcinomas in Mcm4Chaos3/Chaos3 mice.

Although much attention has been paid to germline and somatic mutations in DNA dam-age repair genes that predispose to cancer or promote cancer progression, the current study highlights the importance of screening for mutations in genes involved in DNA replication. Any derangement in the proteins that maintain the high fidelity of replication and the close link between replication, cell cycle and cell divi-sion is expected to lead to genomic instability. Homozygous null mutations are unlikely to be seen in cancer cells or in a person predisposed to cancer, but hypomorphic mutations similar to Chaos3 are entirely possible. One hopes that the genes selected for screening in cancer genome projects include the complete assortment of genes involved in DNA replication. Decreases in the activity of a gene or protein can also arise from epigenetic and post-transcriptional mechanisms. For example, overexpression of

cyclins E, D and A has been often described in cancers, and these cyclins (paired with their relevant CDKs) are known to phosphorylate components of the prereplicative complex and prevent the loading of Mcm2-7 (ref. 11). Elevated levels of these cyclins (or decreased lev-els of CDK inhibitors like p27) could promote cancer development or progression through decreased origin licensing and genomic instabil-ity, similar to observations in Mcm4Chaos3/Chaos3 mice.

1. Shima, N. et al. Nat. Genet. 39, 93–98 (2007).2. Yoshida, K., Kuo, F., George, E.L., Sharpe, A.H. &

Dutta, A. Mol. Cell. Biol. 21, 4598–4603 (2001).3. Pinto, S. et al. Neuron 23, 45–54 (1999).4. Woodward, A.M. et al. J. Cell Biol. 173, 673–683

(2006).5. Blow, J.J. & Dutta, A. Nat. Rev. Mol. Cell Biol. 6, 476–

486 (2005).6. Machida, Y.J., Hamlin, J.L. & Dutta, A. Cell 123, 13–24

(2005).7. Arentson, E. et al. Oncogene 21, 1150–1158

(2002).8. Karakaidos, P. et al. Am. J. Pathol. 165, 1351–1365

(2004).9. Shreeram, S., Sparks, A., Lane, D.P. & Blow, J.J.

Oncogene 21, 6624–6632 (2002).10. Machida, Y.J., Teer, J.K. & Dutta, A. J. Biol. Chem. 280,

27624–27630 (2005).11. Ekholm-Reed, S. et al. J. Cell Biol. 165, 789–800

(2004).

Telomeres, p21 and the cancer-aging hypothesisJessica F Bell & Norman E Sharpless

Telomere dysfunction suppresses cancer through the p53 tumor suppressor pathway but also contributes to aging. A new study suggests that these effects of dysfunctional telomeres may be separable, such that aging—but not cancer suppression—depends on the p21 cell cycle inhibitor.

Jessica F. Bell and Norman E. Sharpless are in the Departments of Medicine and Genetics, The Lineberger Comprehensive Cancer Center, The University of North Carolina School of Medicine, Chapel Hill, North Carolina 27599, USA.e-mail: nes@med.unc.edu

The ‘cancer-aging hypothesis’ proposes that neoplastic disease and the age-induced degen-eration of many organs are closely linked by the failure or success of tumor suppressor mecha-nisms such as cellular senescence. For example, activation of the p16 tumor suppressor in self-renewing compartments such as hematopoi-etic stem cells (HSCs) seems to contribute to aging1, yet inactivation of p16 is an early step in the formation of many cancers2. An impor-tant prediction of the cancer-aging hypothesis is that certain well-intended efforts to prevent aging (growth hormone replacement to retard muscle loss, for example) may accelerate cancer

formation, and vice versa. Work from Aaheli Roy Choudhury and colleagues on page 99 of this issue3 is therefore surprising in that it shows that deficiency of the p53 effector p21 extends longevity and rescues stem cell function in mice with dysfunctional telomeres, without sacrific-ing tumor suppression.

Taking the bad with the goodStrong evidence from disparate systems sug-gests that telomere dysfunction is a critical tumor suppressor mechanism in humans4. In the absence of telomerase activity, persistent proliferation leads to telomere shortening and dysfunction, which in turn induces replicative senescence. Most human cancers have to solve the problem of telomere shortening, usually by reactivating telomerase activity. Evidence that telomere dysfunction also contributes to human aging is accumulating. For example, mice geneti-cally engineered to harbor short telomeres show premature aging5, and evidence of telomere

NATURE GENETICS | VOLUME 39 | NUMBER 1 | JANUARY 2007 11

dysfunction can be detected in primate skin dur-ing the physiological aging process6. Moreover, humans with mutations that compromise the telomerase RNA-protein holoenzyme complex develop premature bone marrow failure and a shortened lifespan7,8. This dual role in prevent-ing cancer and causing aging has made telo-mere-based replicative senescence a prominent component of the cancer-aging hypothesis.

The p53 tumor suppressor is a major deter-minant of the organismal response to telomere dysfunction (Fig. 1). Previous work has shown that telomere dysfunction is sensed in a man-ner similar to DNA double-strand breaks, with activation of p53 effecting either senescence or apoptosis4. In agreement with this model, p53 deficiency rescues many of the phenotypic consequences of telomere dysfunction such as impaired fertility and germ cell apoptosis in mice with pathologically shortened telomeres9. Despite these impressive phenotypic effects of telomere dysfunction, animals with short

©20

07 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

ureg

enet

ics