Cancer and the Knockout Mouse Sary for Tumorigenesis

news and views

SMADs are feedback regulators of TGF-ȋ 2. Imamura, T. et al. Nature 389, 622–626 (1997). capacity in culture, or any reduction in the 3. Tsuneizumi, K. et al. Nature 389, 627–631 (1997). signalling opens up the question of how the 4. Nakao, A. et al. Nature 389, 631–635 (1997). ability to generate tumours when oncogene- expression and regulation of these molecules 5. Hayashi, H. et al. Cell 89, 1165–1174 (1997). transformed cells are injected into nude participates in the patterning of TGF-ȋ 6. Topper, J. N. et al. Proc. Natl Acad. Sci. USA 94, 9314–9319 (immunodeficient) mice. responses in developing tissues. (1997). It has been suggested that progressive 7. Alevizopoulos, A. & Mermod, N. Bioessays 19, 581–591 (1997). Malcolm Whitman is in the Department of Cell 8. Neumann, C. & Cohen, S. Bioessays 19, 721–729 (1997). erosion of telomeres in somatic cells (which Biology, Harvard Medical School, 240 Longwood 9. Chen, X. et al. Nature 389, 85–89 (1997). have physiologically repressed telomerase) Avenue, Boston, Massachusetts 02115, USA. 10.Kim, J., Johnson, K., Chen, H. J., Carroll, S. & Laughon, A. represents a barrier to indefinite cell prolifer- Nature 388, 304–308 (1997). e-mail: [email protected] 11.Raftery, L. A., Twombly, V., Wharton, K. & Gelbart, W. Genetics ation. Tumour cells are thought to overcome 1. Massagué, J., Hata, A. & Liu, F. Trends Cell Biol. 7, 187–192 (1997). 139, 241–254 (1995). this block by reactivating telomerase. But the apparently undiminished tumorigenic abili- Telomerase ty of oncogene-transformed mTR –/– cells suggests that telomerase may not be neces- Cancer and the knockout mouse sary for tumorigenesis. Data showing that, in mice, telomerase can be up-regulated at David Wynford-Thomas and David Kipling an early stage in tumour development5 — before any significant telomere erosion has elomeres are specialized structures at any acute toxicity. But telomere shortening is occurred — leads the authors to hint that the the ends of linear chromosomes. They occurring: by the latest generation studied current dogma may be wrong. In other Tusually consist of tandem arrays of (G6), there is clear evidence of telomere words, telomerase may be nothing more a short DNA sequence (TTAGGG in ver- erosion, with around five per cent of chro- than a passive bystander, rather than facili- tebrates) and associated proteins, and they mosomes in embryonic fibroblasts lacking tating tumour growth4. If so, this would have are thought to have at least two essential detectable TTAGGG, accompanied by considerable implications for telomerase as a functions1. First, they stop natural chromo- increasingly frequent chromosome fusions. target for cancer therapy (although it does some ends behaving as random breaks, This ‘delay’ in manifestation of the phe- not necessarily detract from its value as a which might otherwise generate inter- notype almost certainly reflects the unusual- diagnostic marker). chromosomal fusions and activate DNA- ly long telomere-repeat arrays in the germ There are, however, good reasons for damage-induced cell-cycle arrest. Second, line of this mouse species. These repeats are exercising caution when extrapolating from they provide the structural basis for solving not exhausted until many generations have the mouse model. The main limitation is the end-replication problem2 — the inability elapsed — a human ‘knockout’ would be that the phenotype is probably not yet com- of DNA polymerases to completely replicate predicted to show a much more rapid onset. pletely developed. Although telomere fail- the end of a DNA duplex (see box). Yet, despite reaching what seems to be a criti- ure is occurring at G6, most of the chromo- In the germ line, telomeres are main- cal state of erosion, cells from generation G6 somes retain detectable TTAGGG. The very tained by the compensatory addition of do not show an obvious reduction in growth fact that viable G6 animals exist suggests TTAGGG repeats to chromosome ends. These repeats are synthesized by an enzyme The end-replication problem called telomerase (Fig. 1, overleaf), and a lack of telomerase activity — such as occurs physiologically in most adult human somat- For lagging-strand DNA a b 3’ 3’ ic cells — leads to progressive telomere replication, short RNA 5’ 5’ shortening with every cell cycle3. If such cells primers (blue) are made are forced to grow for long enough, they by RNA primase. These eventually lose telomere function leading to are then extended by 5’ 5’ 3’ 3’ end-to-end chromosome fusions and cell DNA polymerase to form death. What, then, might be the effect of Okazaki fragments. a constitutional (germline) deficiency of When these RNA 5’ 5’ telomerase? An intriguing (although still primers are removed, 3’ 3’ partial) answer is provided in the latest issue there is no way to of Cell, where Blasco et al.4 report their synthesize lagging- Primer gap Primer gap studies of a telomerase-deficient mouse. strand sequence that is 3’ extension (b). mechanism to ensure Telomerase contains an essential RNA complementary to the Why are such natural that natural chromosome component that provides the template for small region at the end blunt ends not ends are even more the specific synthesis of TTAGGG repeats. of the chromosome recognized as DNA different. This involves a Blasco et al. used transgenic technology to (which is at least as damage? One possibility degradative pathway create a germline deletion of the mTR gene, large as an RNA is that this is because which, even in which encodes the RNA component of primer). So, with unnatural ends have a telomerase-deficient mouse telomerase. The authors bred continuing cell division, slightly different human cells, results in 3’ homozygous, null mTR–/– mice which sequence is lost from chemical structure; for extensions for most turned out to be both viable and fertile, the ends of linear example, radiation- chromosome ends9. despite having no detectable telomerase chromosomes. Some induced blunt ends can This may be analogous activity. Moreover, the mice have now been blunt-ended daughter have terminal 3’ to a degradative maintained for six generations (G1–6). The molecules are produced phosphoglycolate pathway that has absence of any initial phenotype is striking. by this scheme, residues whereas been described in So is the fact that the next few generations irrespective of whether physiological ends do budding yeast, involving have also shown no symptoms, suggesting the starting terminus is not. Human cells may Cdc13p and other that the telomerase inhibitors that have been blunt-ended (a) or has a have an additional proteins10. D.W.-T. & D.K. envisaged for cancer therapy will not have

Figure 1 Synthesis of telomeric DNA Daedalus a by telomerase. a, A short region of * * 3’ the upper strand of the telomere Smooth ascent (composed of TTAGGG repeats) 5’ cannot be copied by normal DNA A solid-fuel rocket is simply an enormous Telomerase replication (see box) and is indicated firework. Once lit, it cannot be stopped or by asterisks. b, Telomerase controlled. Furthermore, it burns in b circumvents this by synthesizing new massively turbulent flow, generating 3’ TTAGGG repeats (blue), allowing the extreme vibration. Everything that has to ends to be copied by conventional fly in a rocket must pass a searching 5’ DNA replication (c). Mouse vibration test. Daedalus is now devising a DNA polymerase telomerase carries an RNA molecule less vibrant way to climb into space. that acts as a template for synthesis of A typical solid fuel is a mixture of 4 c the TTAGGG sequence. Blasco et al. aluminium powder and ammonium * * 3’ have deleted the mTR gene (which perchlorate, formed into a resilient solid encodes this RNA molecule) in the by a polymeric binder. DREADCO 5’ mouse germ line.The resultant chemists are replacing the aluminium knockout mice have no telomerase powder by fine aluminium wire, and the Primer gap activity, and their telomeres become binder by a water-based gel. The product is shorter with continuing cell division. just damp enough so that ignition cannot spread through it spontaneously; but each that interference with chromosome func- cell death), and natural chromosome ends element can be ignited by passing a current tion may need to be more widespread to (which do not). It has been assumed that in through the local aluminium wire. The compromise cell proliferation severely. It mammalian cells, specific terminal DNA charge will be fired controllably, element seems improbable that the observed telo- sequences and associated proteins somehow by element, by passing a programmed mere erosion could continue unabated ‘shield’ the natural end of the chromosome sequence of current pulses through the without eventually stopping growth. For and prevent it from activating genome- wires. example, viral-oncogene-transformed hu- damage-monitoring systems. The later- In this simple form, the idea is clearly man cells eventually undergo ‘crisis’ (cell generation mTR –/– cells show an apparent impractical. A big rocket motor could have death) during continuous culture, unless loss of telomere-protective function (as millions of ignitable elements, far too telomere erosion is halted by activation of judged by end-to-end fusions in metaphase many for each to have its own electrical either telomerase3 or of an alternative path- chromosome preparations). But the fact leads. But Daedalus recalls the old flashbar way6. A remarkable counterpart in vivo may that such cells have reached metaphase units for small cameras, which carried ten be the spontaneous tumour regression that suggests that they have passed through at separate flash-bulbs. As each bulb fired, it occurs in a subset of childhood neuroblast- least one damage-sensing checkpoint (at carbonized and caused a lead to the next omas (so-called stage IV-S), which appar- G2–M), despite having critically eroded bulb to become conducting. Each time the ently fail to up-regulate telomerase7. So, telomeres. shutter was depressed, the trigger current if telomere erosion continues, oncogene- If the absence of telomere-specific DNA set off the next live bulb. This trick, transformed mTR –/– cells might be expect- sequences and proteins does not activate suitably developed, would allow thousands ed, eventually, to lose their ability to gener- cell-cycle arrest, then what does? It is possible of elements in the motor to be set off ate tumours, once they have undergone that the physical structure of the DNA-end controllably by a single set of leads. A few enough cell divisions. itself is differentially recognized — a break thousand sets of such leads could It is also risky to extrapolate from the caused by ionizing radiation is not only command the complete motor. mouse with regard to the timing of telo- physiologically abnormal, but it is chemical- The resulting rocket will be a merase up-regulation in tumour develop- ly very different from a natural chromosome masterpiece of smooth power. Its thrust ment — not least because telomerase seems terminus formed by DNA replication. will be controlled from instant to instant to be more tightly regulated in man. Many We now await with interest the pheno- by the rate at which its elements are fired. (but not all) human cancers possess very type of later generations of these mice (if Its vibration will be damped by fast-acting short telomeres, consistent with the classic they remain fertile), and later passages of acceleration sensors feeding its control model of telomere erosion followed by selec- cell cultures derived from them. Whatever computer, which will stagger the firing tion for telomerase reactivation during an in the outcome, it is certain that interest in sequence to the next elements so as to vivo equivalent of crisis. Of course, some telomeres has not come to an end. cancel or minimize the instantaneous tumours may acquire telomerase ahead of David Wynford-Thomas and David Kipling are in fluctuations of thrust. If required, the critical telomere erosion; for example, if the Department of Pathology, University of Wales whole motor can be shut off instantly, and telomerase is reactivated as part of a ‘pack- College of Medicine, Heath Park, Cardiff started again later; once in space, it can be age’ of changes in gene expression that CF4 4XN, UK. operated on command at very low thrust occurs after some other genetic event (the 1. Kipling, D. The Telomere (Oxford Univ. Press, 1995). for orientation or course correction. ‘co-selection’ hypothesis4,8). But such antici- 2. Olovnikov, A. M. Dokl. Akad. Nauk 201, 1496–1499 (1971). The technology could even be adapted 3. Counter, C. M. et al. EMBO J. 11, 1921–1929 (1992). patory up-regulation does not mean that 4. Blasco, M. A. et al. Cell 91, 25–34 (1997). for other purposes, such as blasting, telomerase is not required at a later stage, 5. Broccoli, D., Godley, L. A., Donehower, L. A., Varmus, H. E. demolition and explosive forming. Instead so this should not dampen enthusiasm for & de Lange, T. Mol. Cell. Biol. 16, 3765–3772 (1996). of many explosives for different jobs, one anti-telomerase therapies. 6. Bryan, T. M. & Reddel, R. R. Eur. J. Cancer 33, 767–773 all-purpose DREADCO product will do. It (1997). We may also need to revise our ideas 7. Hiyama, E. et al. Nature Med. 1, 249–255 (1995). can be programmed to act as a propellant, about how cells distinguish between unnat- 8. Kipling, D. Eur. J. Cancer 33, 792–800 (1997). a low, medium or high explosive, or even a ural genomic breaks such as those caused by 9. Makarov, V. L., Hirose, Y. & Langmore, J. P. Cell 88, 657–666 subtle combination of each. (1997). ionizing radiation (which usually lead to 10. Lin, J.-J. & Zakian, V. A. Proc. Natl Acad. Sci. USA 93, David Jones immediate cell-cycle arrest or programmed 13760–13765 (1996).