Perspectives

PERSPECTIVES

In support of a dominant phenotype for OPINION chromosome instability, sequencing colorectal cancer cell genomes has revealed cancer-specific dominant mutations in cer- Telomere dysfunction and the tain kinetochore checkpoint genes, such as BUB1 and BUBR1 (REF. 6). Expression of initiation of genome instability mutant BUB1 in stable diploid cancer cells induces whole chromosome gains and losses. The idea that dominant genetic David M. Feldser, Jennifer A. Hackett and Carol W. Greider changes can result in a chromosome-instability phenotype is provocative; however, few Tumour growth is an evolutionary process Telomere dysfunction examples of genetic mutations that cause that is characterized by the selection of Rather than mutations in specific genes, chromosome instability have been docu- clonal populations of cells that acquire telomere defects might be a significant cause mented7. This means that another mecha- distinct genetic changes. Many cancer of chromosome instability. Telomerase activ- nism could be responsible for generating the therapies aim to exploit the specific ity is absent from most somatic cells. Cell complex chromosome rearrangements that changes that occur in cancer cells, but division in the absence of telomerase causes are observed in most tumour cells. understanding the underlying mechanisms progressive telomere shortening, which could Telomere dysfunction is an alternative of genomic instability that cause these ultimately lead to telomere dysfunction. The mechanism to specific gene mutations in mutations could lead to more effective ongoing chromosomal instability that occurs initiating chromosome instability. As dis- therapies. If common mechanisms exist when telomere function is lost could drive the cussed in detail below, telomere dysfunction for initiating genomic instability in tumours, tumorigenic process by increasing the rate can induce a dominant chromosome-insta- selection could explain the differences in of mutation of oncogenes and tumour- bility phenotype. Telomere dysfunction can specific gene mutations that accumulate suppressor genes8. Telomerase activation late be induced by mutations in specific genes in different tumour types. The cause of in tumorigenesis is ultimately required to that cause cancer-predisposition syn- genomic instability in human tumours is limit chromosome instability and to allow dromes, including dyskeratosis congenita, unclear, although there is evidence to progression towards malignancy (FIG. 1). Bloom’s syndrome, Werner’s syndrome and, indicate that telomere dysfunction could Such a role for telomere dysfunction in perhaps, ataxia telangiectasia. However, make an important contribution. tumour initiation accounts for many aspects many tumour cells undergo telomere short- of chromosome instability in human cancers. ening without mutations in genes that Most human malignancies are characterized These include the timing of chromosome encode telomerase components, which also cytogenetically — by marked aneuploidy and instability, the rarity of specific chromosome- leads to chromosome instability. complex chromosomal rearrangements that instability-inducing genetic mutations, the include non-reciprocal translocations, DNA karyotypic profile found in cancers, the dom- Contribution to chromosome instability. fragmentation and chromosome fusions1.This inant nature of chromosome instability and Telomere function has long been known to be chromosome instability might be a ‘mutator the requirement for telomerase activation late essential for the preservation of chromosomal phenotype’2 that allows an increased rate of in tumorigenesis. integrity. Hermann J. Muller’s study of irradi- chromosomal aberrations per cell division (BOX ation-induced breaks in chromosomes 1).Chromosome instability is present very early Chromosome instability is dominant. The showed that the telomere functions to distin- in colorectal tumorigenesis3 and an increased mechanism by which complex karyotypes are guish a chromosome’s natural DNA end from rate of chromosomal aberrations has been doc- generated in tumours is unclear. When a DNA break9. When the function of the umented in colorectal cancer cell lines4,5. In tumour cells without chromosome instability telomere is compromised — either by loss of support of this, rare genetic mutations have are fused with tumour cells that show chro- various telomere-capping proteins or by criti- been found that induce chromosome instabil- mosome instability, the fused cells are chro- cal shortening of telomeric repeat sequences ity in karyotypically stable cell lines6, but mosomally unstable4. This result indicates — the telomere becomes a substrate for despite extensive sequencing of candidate that chromosome instability is a dominant DNA-repair processes, such as end-to-end genes, mutations in other genes that contribute phenotype, so loss of a tumour-suppressor chromosome fusion and exonucleolytic to this phenotype have not yet been found7. gene could not cause chromosome instability. sequence degradation. As shown by Barbara

mechanism (FIG. 2c), as breakage of fused Box 1 | Genomic instability chromosomes results in a random distribu- Instability refers specifically to an increased rate of mutation (that is, an increased frequency of tion of breakpoints along the chromosome. mutation per cell division); the presence of a mutation is not evidence for instability. A further The initial chromosome instability induced distinction has been made regarding the types of mutations involved in cancer cells54. Instability by end resection at dysfunctional telomeres involving whole or partial regions of chromosomes is referred to as chromosomal instability. could initiate BFB cycles. In support of this Instability at the nucleotide level is frequently referred to as microsatellite instability. Genomic idea, the chromosome fusion events observed instability is a general term that refers to all types of instability. in mouse cells that have telomere dysfunction have extensive sequence loss at the fusion junctions, indicating that end resection McClintock, dicentric chromosomes can ini- chromosome fusions11. The chromosomes might occur before fusion12. The types of tiate ongoing chromosomal instability10 via that participate in these fusions are those with chromosome rearrangements that are breakage–fusion–bridge (BFB) cycles. If the the shortest telomeres in the population12. observed as a result of telomere dysfunction, two centromeres of a dicentric chromosome These findings are supported by studies in including non-reciprocal translocations, are pulled to opposite poles during mitosis, telomerase-deficient yeast, in which telomere chromosome losses, and regional amplifica- the chromosome can break. These breaks can shortening directly contributes to an increased tions and deletions13,19,20, are similar to the then contribute to translocations or form new rate of chromosomal rearrangements13. rearrangements that are observed in human dicentric chromosomes that continue the tumours. Moreover, BFB cycles can be self- process of instability. Causing cancer-like karyotypes. The types of perpetuating21, and so might explain the Experiments that disrupted the telom- chromosome rearrangements that are dominant nature of the chromosome-insta- erase RNA gene (mTR) in mice offered the observed in human cancer cells change dur- bility phenotype of human tumours. This first indication that telomere shortening can ing the course of tumour development. A model also provides an explanation for the initiate chromosomal instability. These mice cytogenetic analysis of human osteosarcomas way in which chromosome instability occurs develop shortened telomeres, and their and pancreatic tumours indicated that early in the absence of a specific mutator gene. cells have a high frequency of end-to-end tumours with few cytogenetic abnormalities had rearrangement breakpoints that were Initiation of tumorigenesis. Chromosome predominantly located near the ends of chro- instability occurs at early stages of tumorige- Telomere shortening mosomes. However, in tumours with highly nesis in humans3, although there are experi- (telomerase inactive) complex abnormalities, chromosomal break- mental limitations to how early tumours can

Normal somatic cell points were mostly interstitial and cen- be detected. Recently, mathematical model- tromeric14. Importantly, the complexity of ling of the initiation of colon cancer showed chromosomal aberrations was inversely cor- that chromosome instability can occur early Telomere dysfunction related with telomere length. The localization enough to precede inactivating mutations in of chromosome breakpoints near chromo- the tumour-suppressor gene APC 22. By con- some ends, coupled with short telomeres trast, given the low rate of new mutations in observed in early tumour samples, indicates normal cells, it is difficult to imagine how a Chromosomal that telomere dysfunction could contribute mutation that induces chromosome instabil- instability begins Initial mutation to chromosome instability14,15. ity could occur before loss of function of a The clustering of chromosome break- tumour suppressor such as APC 23. However, Multiple points near telomeres could result from if telomere dysfunction is the initiating Telomerase genetic exonucleolytic end resection of dysfunctional event, the chance of chromosomal instability activation changes telomeres (FIG. 2a). Rapid and extensive occurring is related to the telomere lengths exonucleolytic degradation of sequence from on each chromosome end in every somatic chromosome ends occurs in yeast that have cell. So, telomere length is not directly related lost telomere end-binding proteins16–18 or to the basal mutation rate, but rather to the that lack telomerase (J.A.H. and C.W.G., proliferative history of the cell. unpublished observations). In fact, this In telomerase-deficient mice, short dys- Carcinoma process causes a marked increase in chromo- functional telomeres can increase the fre- Figure 1 | Telomere shortening in a simplified some rearrangements that are clustered pre- quency of tumorigenesis8,20,24–26. Significantly, cancer-progression model. Normal somatic dominantly near chromosome ends. The telomere dysfunction has also been directly cells lose telomeric repeats as they divide in the extent to which sequence loss at dysfunc- linked to an increased frequency of tumour absence of telomerase. Eventually, telomeres Min become dysfunctional and can cause tional telomeres in human cells can initiate initiation in the Apc mouse model of colon 26 +/– chromosomal instability. Chromosomal instability non-reciprocal translocations is unknown; cancer . These animals are Apc and invari- occurs very early in tumorigenesis. It begins either however, end resection at dysfunctional ably develop multiple intestinal neoplasia before or soon after the initial mutation in a tumour telomeres can explain the presence of mostly (Min) following loss of the remaining wild- suppressor or oncogene. Chromosomal instability telomeric breakpoints in human cancers that type Apc allele. Telomerase-deficient Apc Min then drives the multiple genetic changes that are have limited chromosome instability (FIG. 2b). mice develop more early-stage microadeno- required for the formation of a carcinoma. The chromosome rearrangements that mas than Apc Min mice. However, the inability Telomerase activation occurs late in tumorigenesis as it can not be detected until the carcinoma occur in more advanced tumours can also be to activate telomerase in this genetic back- stage49. Telomerase increases the replicative a result of telomere dysfunction. Such ground suppresses the progression of these potential of a tumour, facilitating tumour growth. rearrangements might occur through a BFB adenomas (see below).

Normal 3ʹ chromosome end 5ʹ

Telomere repeat

Dysfunctional telomere

BFB Resected end

Limited chromosomal Exonuclease Extensive chromosomal instability instability

b One resected end c Two resected ends or Resected chromosome Intact chromosome additional DNA damage

Chromosome with ds break

Chromosome end acquisition

Chromosome fusion +

Terminal breakpoints Random breakpoints

Figure 2 | Model for the mechanisms by which dysfunctional telomeres contribute to genomic instability. a | In the absence of telomerase, telomeres progressively shorten and can become dysfunctional. Dysfunctional telomeres are subject to end resection by exonucleases18,53 (J.A.H. and C.W.G., manuscript in preparation). A resected chromosome end can initiate different types of chromosome rearrangements in cells with different degrees of chromosome instability. b | If there is limited chromosome instability, a resected chromosome end is likely to use an intact chromosome as a substrate for repair (blue chromosome end). The resected end could copy the end of the intact chromosome, using break-induced replication or another mechanism, leading to a non-reciprocal translocation involving chromosome-end acquisition on the broken chromosome13 (J.A.H. and C.W.G., manuscript in preparation). Break-induced replication could copy the end from the homologous chromosome, or from an unrelated chromosome, using shared regions of homology such as repetitive elements. So, telomere dysfunction in a setting of limited chromosomal instability will produce primarily terminal chromosome breakpoints (bottom panel). This has been observed in tumours with limited chromosome instability14. c | If there is extensive chromosome instability in a cell, there are several possible substrates for repair of a resected chromosome end. These substrates include an intact chromosome (as in b), a second resected chromosome end or a chromosome with a double-strand (ds) break (red chromosomes). Such repair can generate end-to-end chromosome fusions, which can accentuate chromosome instability by generating new DNA breaks through breakage–fusion–bridge (BFB) cycles. BFB cycles can produce random chromosome breakpoints (bottom panel), as have been observed in tumours with extensive chromosome instability14.

Although there is much evidence to indi- This progressive telomere shortening is consis- To determine whether dysfunctional cate that short telomeres contribute to tumour tent with a progressive increase in tumorigen- telomeres actually initiate tumorigenesis, it initiation in mice27, the evidence that a similar esis with age. Cancer cells generally have is important to examine telomeres during effect promotes human tumour formation is significantly shorter telomeres than their nor- the earliest stages of tumour development. only correlative. The absence of telomerase mal counterparts29, but these short telomeres Recently, a fluorescence in situ hybridiza- activity in most human somatic cells leads to could simply be a byproduct of the extensive tion protocol was developed that allows the telomere shortening with each cell division28. cellular divisions that occur in the cancer cell30. measurement of telomere length directly in

progression. Alternatively, telomere short- occasions,a process called alternative lengthen- ening could have a more direct effect, by ing of telomeres (ALT) can elongate telomeres inducing the chromosomal instability that — probably through a DNA-recombination leads to tumorigenesis. mechanism51.In support of the idea that telomere maintenance is required,when telomerase- Cellular senescence deficient mice were crossed to Apc Min mice, an Telomere shortening has been proposed to be increased incidence of tumour initiation was a tumour-suppressor mechanism30,39,40, so it is observed; however, these tumours did not paradoxical to think of it as a mechanism for progress past the early stages26. Therefore, tumour initiation. The idea that telomere restoration of telomere function seems to be an shortening suppresses tumour formation is important step during the later stages of based on the behaviour of primary human tumorigenesis. Reactivation of telomerase fibroblasts in culture. These cells undergo a might provide a simple mechanism to increase limited number of population doublings the proliferative capacity of cells in a tumour by Figure 3 | Telomere fluorescence in situ before they undergo cellular senescence, ceas- stabilizing telomere structure. hybridization in a tissue section of colonic ing to divide41. Cellular senescence is corre- mucosa. A fluorescent probe indicates the lated with telomere shortening28, as well as a Future directions presence of telomeres (red) in the DNA (blue) of marked increase in the frequency of end-to- To establish a firm link between telomere dys- crypt cells31. The fluorescent signal is very low or end chromosome fusions42–44. Reintroduction function, telomerase activation and human absent in cells of the adenomatous crypt (left), whereas normal crypt cells (right) have long of telomerase elongates telomeres and allows cancer, at least four important questions must telomeres, and therefore emit a very bright signal. cells to avoid the induction of cellular senes- be addressed. Does telomere dysfunction cause Image provided with permission from Alan cence. Therefore, the absence of telomerase instability in human cancer and, if so, when Meeker, Johns Hopkins University, USA. activity in most human somatic cells could does instability begin? Does telomerase activa- serve as a tumour-suppressor mechanism, by tion suppress genomic instability? Finally, will inducing senescence. However, it is not inhibiting telomerase specifically cause cancer- archival tissue sections31 (FIG. 3). Using this known whether cellular senescence occurs cell death? To address the question of whether technique, telomere shortening was in vivo in the many different tissue types that telomere dysfunction causes instability in reported to be the earliest detectable genetic give rise to tumours. If this mechanism does human cancer it will be necessary to study change in pancreatic cancer cells32. Loss of exist, the pathways that recognize short tumours at their earliest stages of development. telomere signal in these tissues is nearly telomeres and arrest cell division might need It is clear that short telomeres are present in universal during all stages of pancreatic to be disrupted for telomere shortening to precursor lesions to many tumour types32,33, intraepithelial neoplasia. In addition, in accelerate tumorigenesis. Initially, if dysfunc- and in late-stage tumours29. However,it is tech- high-grade prostatic intraepithelial neopla- tional telomeres arise in a normal cell and at nically difficult to establish that these telomeres sia — a putative precursor lesion to some low frequency, normal checkpoint con- are dysfunctional.It will therefore be necessary prostate cancer — telomeres were signifi- trol could be evaded. In this scenario, short to develop a molecular marker for telomere cantly shorter than adjacent non-neoplastic telomeres could initiate chromosomal dysfunction, the presence of which can be lesions in more than 96% of the samples33. instability, leading to tumorigenesis as other detected in very early tumours. If telomere shortening does contribute mutations arise. In addition to understanding whether to tumorigenesis, there should be higher telomere dysfunction can initiate tumorige- frequencies of tumour formation in tissues Limiting chromosome instability. Chromo- nesis, understanding the role of telomerase that undergo high rates of cell turnover. somal instability in cancer cells provides a activation could be more important for pos- Several chronic human diseases, including tumour with genetic diversity. Through this sible cancer therapy. The question ‘does Barrett’s oesophagus, ulcerative colitis34,35, diversity, some cells acquire the ability to telomerase activation suppress genome liver cirrhosis36 and pancreatitis32, are asso- subvert the initial barriers to continued instability?’ has been largely unexplored in ciated with high cell turnover, as well as an growth, and therefore undergo positive selec- mammalian cancer cells, although evidence increased risk of cancer in the affected tion7. Although genomic instability might from yeast systems indicates that this might organ. Telomere shortening has been be necessary for tumour development, too be the case13,52. Finally, it is essential to know detected in tissue samples taken from much instability can be detrimental to whether inhibiting telomerase specifically patients with ulcerative colitis35,37,38. Most tumour growth45. If telomere dysfunction, kills human tumour cells in vivo. If telom- strikingly, patients whose biopsies show the rather than a specific gene mutation, is a erase activation provides stability to the can- highest degree of telomere shortening not cause of chromosome instability, then telom- cer-cell genome, telomerase inhibition only have the highest degree of chromo- erase reactivation would provide a built-in might be an effective therapeutic strategy. some instability, but also are most likely to mechanism to limit instability. In yeast, the Targeting telomerase might limit tumour progress to cancer37. The cause of tumori- reactivation of telomerase in telomerase- progression while increasing the degree of genesis in this setting is still unclear. The null cells has been shown to limit ongoing chromosomal instability specifically in high rate of cell turnover in these diseases chromosomal instability 46. tumour cells. Normal cells that express could promote tumorigenesis simply by All tumours acquire mechanisms to telomerase might be more resistant to the providing more cell divisions, allowing elongate their telomeres by late stages of tumor- effects of telomerase inhibition than cancer more cancer-causing mutations to accumu- igenesis, which could serve to limit instabil- cells. These cells would be expected to have late. If this is the case, telomere shortening ity 47–50. Most of the time, this is accomplished longer telomeres and divide less frequently might simply be a side effect of cancer through telomerase reactivation47, but on rare than cancer cells.

David M. Feldser and Jennifer A. Hackett 18. Garvik, B., Carson, M. & Hartwell, L. Single-stranded 41. Hayflick, L. & Moorhead, P. S. Exp. Cell Res. 25, DNA arising at telomeres in cdc13 mutants may 585–621 (1961). are part of the Predoctoral Training Program in constitute a specific signal for the RAD9 checkpoint. Mol. 42. Counter, C. M. et al. Telomere shortening associated with Human Genetics and Molecular Biology, Cell. Biol. 15, 6128–6138 (1995). chromosome instability is arrested in immortal cells which Johns Hopkins University School of Medicine, 19. O’Hagan, R. C. et al. Telomere dysfunction provokes express telomerase activity. EMBO J. 11, 1921–1929 617 Hunterian Building, 725 N. Wolfe Street, regional amplification and deletion in cancer genomes. (1992). Baltimore, Maryland 21205, USA. Cancer Cell 2, 149–155 (2002). 43. Sherwood, S. W., Rush, D., Ellsworth, J. L. & Schimke, R. T. 20. Artandi, S. E. et al. Telomere dysfunction promotes non- Defining cellular senescence in IMR-90 cells: a flow reciprocal translocations and epithelial cancers in mice. cytometric analysis. Proc. Natl Acad. Sci. USA 85, Carol W. Greider is at the Nature 406, 641–645 (2000). 9086–9090 (1988). Department of Molecular Biology and Genetics, 21. McClintock, B. The behavior in successive nuclear 44. Benn, P. A. Specific chromosome aberrations in Johns Hopkins University School of Medicine, 617 divisions of a chromosome broken at meiosis. Proc. Natl senescent fibroblast cell lines derived from human Hunterian Building, Acad. Sci. USA 25, 405–416 (1939). embryos. Am. J. Hum. Genet. 28, 465–473 22. Nowak, M. A. et al. The role of chromosomal instability in (1976). 725 N. Wolfe Street, Baltimore, tumor initiation. Proc. Natl Acad. Sci. USA 99, 45. Cahill, D. P., Kinzler, K. W., Vogelstein, B. & Lengauer, C. Maryland 21205, USA. 16226–16231 (2002). Genetic instability and darwinian selection in tumours. 23. Tomlinson, I. P., Novelli, M. R. & Bodmer, W. F. The Trends Cell Biol. 9, M57–M60 (1999). Correspondence to C.W.G. mutation rate and cancer. Proc. Natl Acad. Sci. USA 93, 46. Hackett, J. A. & Greider, C. W. Balancing instability: e-mail: [email protected] 14800–14803 (1996). dual roles for telomerase and telomere dysfunction in 24. Rudolph, K. L. et al. Longevity, stress response, and tumorigenesis. Oncogene 21, 619–626 (2002). doi:10.1038/nrc1142 cancer in aging telomerase-deficient mice. Cell 96, 47. Kim, N. W. et al. Specific association of human 701–712 (1999). telomerase activity with immortal cells and cancer. 1. Mitelman, F. (ed. Bertil Johansson, F. M.) Catalog of 25. Chin, L. et al. p53 deficiency rescues the adverse effects Science 266, 2011–2015 (1994). Chromosome Aberrations in Cancer (Wiley–Liss, New of telomere loss and cooperates with telomere 48. Shay, J. W. & Bacchetti, S. A survey of telomerase York, 1994). dysfunction to accelerate carcinogenesis. Cell 97, activity in human cancer. Eur. J. Cancer 33, 787–791 2. Loeb, L. A., Springgate, C. F. & Battula, N. Errors in DNA 527–538 (1999). (1997). replication as a basis of malignant changes. Cancer Res. 26. Rudolph, K. L., Millard, M., Bosenberg, M. W. & DePinho, R. A. 49. Chadeneau, C., Hay, K., Hirte, H. W., Gallinger, S. & 34, 2311–2321 (1974). Telomere dysfunction and evolution of intestinal Bacchetti, S. Telomerase activity associated with 3. Shih, I. M. et al. Evidence that genetic instability occurs at carcinoma in mice and humans. Nature Genet. 28, acquisition of malignancy in human colorectal cancer. an early stage of colorectal tumorigenesis. Cancer Res. 155–159 (2001). Cancer Res. 55, 2533–2536 (1995). 61, 818–822 (2001). 27. Chang, S., Khoo, C. & DePinho, R. A. Modeling 50. Tang, R., Cheng, A. J., Wang, J. Y. & Wang, T. C. 4. Lengauer, C., Kinzler, K. W. & Vogelstein, B. Genetic chromosomal instability and epithelial carcinogenesis in Close correlation between telomerase expression and instability in colorectal cancers. Nature 386, 623–627 the telomerase-deficient mouse. Semin. Cancer Biol. 11, adenomatous polyp progression in multistep (1997). 227–239 (2001). colorectal carcinogenesis. Cancer Res. 58, 5. Phear, G., Bhattacharyya, N. P. & Meuth, M. Loss of 28. Harley, C. B., Futcher, A. B. & Greider, C. W. Telomeres 4052–4054 (1998). heterozygosity and base substitution at the APRT locus shorten during ageing of human fibroblasts. Nature 345, 51. Bryan, T. M., Englezou, A., Dunham, M. A. & Reddel, R. R. in mismatch-repair-proficient and-deficient colorectal 458–460 (1990). Telomere length dynamics in telomerase-positive carcinoma cell lines. Mol. Cell. Biol. 16, 6516–6523 29. de Lange, T. et al. Structure and variability of human immortal human cell populations. Exp. Cell Res. 239, (1996). chromosome ends. Mol. Cell. Biol. 10, 518–527 (1990). 370–378 (1998). 6. Cahill, D. P. et al. Mutations of mitotic checkpoint genes 30. de Lange, T. in Telomeres (eds. Blackburn, E. H. & 52. Myung, K., Chen, C. & Kolodner, R. D. Multiple pathways in human cancers. Nature 392, 300–303 (1998). Greider, C. W.) 265–293 (Cold Spring Harbor Laboratory cooperate in the suppression of genome instability in 7. Da Costa, L. T. & Lengauer, C. Exploring and exploiting Press, New York, 1995). Saccharomyces cerevisiae. Nature 411, 1073–1076 instability. Cancer Biol. Ther. 1, 212–225 (2002). 31. Meeker, A. K. et al. Telomere length assessment in (2001). 8. Maser, R. S. & DePinho, R. A. Connecting human archival tissues: combined telomere fluorescence 53. Maringele, L. & Lydall, D. EXO1-dependent single- chromosomes, crisis, and cancer. Science 297, 565–569 in situ hybridization and immunostaining. Am. J. Pathol. stranded DNA at telomeres activates subsets of DNA (2002). 160, 1259–1268 (2002). damage and spindle checkpoint pathways in budding 9. Muller, H. J. The remaking of chromosomes. Collecting 32. van Heek, N. T. et al. Telomere shortening is nearly yeast yku70Delta mutants. Genes Dev. 16, 1919–1933 Net. 13, 181–198 (1938). universal in pancreatic intraepithelial neoplasia. Am. (2002). 10. McClintock, B. The stability of broken ends of J. Pathol. 161, 1541–1547 (2002). 54. Lengauer, C., Kinzler, K. W. & Vogelstein, B. Genetic chromosomes in Zea mays. Genetics 26, 234–282 33. Meeker, A. K. et al. Telomere shortening is an early instabilities in human cancers. Nature 396, 643–649 (1941). somatic DNA alteration in human prostate tumorigenesis. (1998). 11. Blasco, M. A. et al. Telomere shortening and tumor Cancer Res. 62, 6405–6409 (2002). formation by mouse cells lacking telomerase RNA. Cell 34. Ekbom, A., Helmick, C., Zack, M. & Adami, H. O. Acknowledgements 91, 25–34 (1997). Ulcerative colitis and colorectal cancer. A population- We thank members of the Greider lab for reading the manuscript 12. Hemann, M. T., Strong, M. A., Hao, L. Y. & Greider, C. W. based study. N. Engl. J. Med. 323, 1228–1233 and A. I. Meeker for providing the image in figure 3. This work is The shortest telomere, not average telomere length, is (1990). supported by grants from the National Institutes of Health. critical for cell viability and chromosome stability. Cell 35. Kinouchi, Y. et al. Telomere shortening in the colonic 107, 67–77 (2001). mucosa of patients with ulcerative colitis. 13. Hackett, J. A., Feldser, D. M. & Greider, C. W. Telomere J. Gastroenterol. 33, 343–348 (1998). Online links dysfunction increases mutation rate and genomic 36. Kitada, T., Seki, S., Kawakita, N., Kuroki, T. & Monna, T. instability. Cell 106, 275–286 (2001). Telomere shortening in chronic liver diseases. Biochem. DATABASES 14. Gisselsson, D. et al. Telomere dysfunction triggers Biophys. Res. Commun. 211, 33–39 (1995). The following terms in this article are linked online to: extensive DNA fragmentation and evolution of complex 37. O’Sullivan, J. N. et al. Chromosomal instability in Cancer.gov: http://www.cancer.gov/cancer_information/ chromosome abnormalities in human malignant ulcerative colitis is related to telomere shortening. Nature colorectal cancer | osteosarcomas | pancreatic cancer | tumors. Proc. Natl Acad. Sci. USA 98, 12683–12688 Genet. 32, 280–284 (2002). prostate cancer (2001). 38. Engelhardt, M., Drullinsky, P., Guillem, J. & Moore, M. A. LocusLink: http://www.ncbi.nlm.nih.gov/LocusLink/ 15. Lengauer, C. How do tumors make ends meet? Proc. Telomerase and telomere length in the development and Apc | APC | BUB1 | BUBR1 | mTR Natl Acad. Sci. USA 98, 12331–12333 (2001). progression of premalignant lesions to colorectal cancer. OMIM: http://www.ncbi.nlm.nih.gov/Omim/ 16. Baumann, P. & Cech, T. R. Pot1, the putative telomere Clin. Cancer Res. 3, 1931–1941 (1997). ataxia telangiectasia | Bloom’s syndrome | dyskeratosis end-binding protein in fission yeast and humans. Science 39. Campisi, J. Replicative senescence: an old lives’ tale? congenita | Werner’s syndrome 292, 1171–1175 (2001). Cell 84, 497–500 (1996). 17. Carson, M. & Hartwell, L. CDC 17: an essential gene that 40. Greider, C. W. Cellular responses to telomere shortening: FURTHER INFORMATION prevents telomere elongation in yeast. Cell 42, 249–257 cellular senescence as a tumor suppressor mechanism. Carol Greider’s lab: http://telomerase.bs.jhmi.edu/greiderlab (1985). Harvey Lect. 96, 33–50 (2000). Access to this interactive links box is free online.