Apoptosis is normally a rapid process, occurring within a period between one hour and one day. It is elicited likewise by rapid signals, such as the binding of cytokines to death receptors at the cell membrane or the activation of TP53 by ATM following a DNA double-strand break. Moreover, cells subjected to apoptosis vanish quite rapidly by phagocytosis. In all these respects, replicative senescence differs. It sets in slowly, it is usually elicited by signals that accumulate gradually, and cells persist, at least in the short run.

Replicative senescence can be evoked by two different signals which use overlapping pathways for execution. One type of signal emanates from short telomeres, and the second type from CDK inhibitors.

Telomeres in human cells are 5-30 kb long and made up of 1000-5000 repeats of TTAGGG hexamers. The bulk of each telomere consists of double-stranded DNA, but 75-150 nt at the ends are single-stranded. Normally, these single strands are folded back into the double strand, forming a T-loop (Figure 7.6). This is a structure similar to the D-loops occuring during DNA repair by homologous recombination (^■3.3). In humans, telomeric DNA is wrapped around nucleosomes. Therefore, core histones are present, but in addition an unusual assembly of further proteins. The TRF2 (telomeric repeat binding factor 2) protein induces and seals T-loops. It also serves as an anchor for a number of further proteins that are located to the telomere under normal circumstances, in particular the RAD50/NBS/MRE11 complex. This complex processes double-strand breaks during DNA repair (^3.3). The KU70 and KU80 proteins which mark and protect DNA double-strand breaks during repair are also present at telomeres. Thus, telomeres appear to serve as reservoirs for these proteins on one hand, but on the other the repair proteins are strategically placed for dealing with damage to the telomeres themselves. A further protein, TRF1, which is homologous to TRF2, limits telomere length, being regulated itself by tankyrase, a poly-adenosine diphosphate ribosylase and TRF1-interacting nuclear protein 2. TRF1 also helps to maintain the RAD50/NBS/MRE11 complex at the telomere.

Figure 7.6 Structure of human telomeres The T-loop structure of human telomeres with some proteins located there. Their arrangement is largely hypothetical.

With each DNA replication in somatic cells, telomeres shorten. This is caused in principle by the end-replication problem. The top strand (with a 5'-end at the telomere) is replicated by elongation of an RNA primer at or near its end. When it is removed by RNase H after DNA synthesis has proceeded, the resulting gap cannot be filled, since DNA polymerases work invariably in the 5'^3' direction. This end-replication dilemma predicts a theoretical minimum loss of telomere sequences during each replication. In reality, its extent can be larger and is regulated by TRF1.

In germ-line cells, the decrease in telomere length is prevented by a specialized enzyme, telomerase (Figure 7.7). Accordingly, telomeres in germ line cells are approximately twice as long as in somatic cells. Telomerase is a specialized reverse transcriptase that uses an RNA template (AAUCCC) provided by its hTERC subunit to elongate telomeres. While the hTERC subunit is expressed in almost all human cells, the catalytic subunit hTERT is restricted to a small set of cells with high replicative potential, like germ-line cells, tissue stem cells and memory immune cells. Expression of the TERT gene is induced by a number of proliferation-



Figure 7.6 Structure of human telomeres The T-loop structure of human telomeres with some proteins located there. Their arrangement is largely hypothetical.

Figure 7.7 Structure of the human telomerase catalytic subunit The 127 kDa human hTERT enzyme contains a motif (T) shared by all telomerases and several motifs characteristic of reverse transcriptases (RT), including those of endogenous human LINE-1 retrotransposons, HIV and HBV.

stimulating and stem-cell maintaining factors. In particular, its promoter is a target of MYC proteins.

Shortening of telomeres to below a certain length causes replicative senescence. In cells cultured over longer periods, actually two successive steps can be distinguished, which are called M1 and M2. They are operationally defined: M1 can be bypassed by obliteration of RB1 and TP53 function. In the laboratory, this can be achieved by introduction of viral proteins such as SV40 large T antigen (^5.3). After 40 - 50 further doublings, senescence sets in irreversibly at the M2 point. Circumventing the M2 point requires activation of telomerase. It is not precisely known what happens at M1 and M2. Human telomeres are very variable. So, one idea is that M1 is triggered by the first telomere reaching a critical length. This would then activate a checkpoint response through the RB1 and TP53 pathways. At M2, further telomere shrinking has taken place. Some telomeres may have become so short that they can no longer form a T-loop. In addition, they may not be capable of storing DNA repair proteins any more. So, some sort of DNA double-strand repair response may be initiated, likely through ATM, which induces replicative senescence once and for all. While some of these ideas are not fully proven, DNA damage signaling is certainly involved in replicative senescence.

Telomere shortening leads to chromosomal instability. Of course, shortened unsealed telomeres are expected to become substrates for exonucleases which would gradually degrade a chromosome. In fact, a greater danger to genomic integrity may be recombination between different telomeres that are not protected by proteins. Recombination between the telomeres of two chromosomes can generate a dicentric chromosome. During mitosis, this may become missegregated or be pulled to opposite sides of the spindle and disrupted. Disruption would cause two open chromosome ends which could again fuse to other chromosomes and form further dicentrics to continue the cycle. Of note, in this classical breakage-fusion-bridge sequence (^Figure 2.7), the breakpoints tend to move from the telomeres towards the centromere.

In human cancers, establishment of replicative senescence as a consequence of telomere shortening is impeded, often at Ml as well as M2. Many human cancers contain defects in the RB1 and TP53 pathway. While these have many other consequences as well (^6.4, ^6.6), loss of RB1 and TP53 function would be expected to permit the bypassing of the Ml limit. This would establish a population of cells continuing to proliferate with at least some critically shortened telomeres and therefore an enhanced potential for genomic instability. Dicentric chromosomes and a movement of chromosome breaks towards the centromere are quite common observations in carcinoma cells. Moreover, telomere instability due to telomerase dysfuntion is the cause of a human disease, dyskeratosis congenita. Patients with this rare inherited affliction do not only present with defects in skin, hair, and the hematopoetic system, but are also prone to cancer.

In addition to the defects in the RB1 and TP53 pathways, many human cancers express hTERT which can be shown to be enzymatically active in tissue extracts. In some tissues, hTERT expression or activity could therefore serve as a cancer biomarker. In cancers with hTERT expression, telomere lengths are at least stabilized at a low level, albeit they do not always rebound.

There is evidence for a different, alternative mechanism of telomere stabilization, named ALT, in some cancers and even in normal tissues, where telomeres are stabilized or even expanded in the absence of detectable telomerase activity. The unspecific designation ALT reveals that the mechanism is presently mostly based on conjecture, with hints from alternative mechanisms employed in organisms that lack telomerase. There, telomere expansion can be achieved by a kind of homologous recombination double-strand repair (^3.3). Indeed, there is some evidence for such a mechanism in humans and, specifically, that the WRN helicase might be involved.

Telomere erosion is certainly to a large degree responsible for the limited lifespan of cultured human cells. It can be regarded as a mechanism counting the number of cycles a cell has undergone. A second mechanism appears to rely on CDK inhibitor proteins, in particular p16INK4A, p21CIP1, and p57KIP2 (Figure 7.8).

Among the CDK inhibitors, p21CIP1 is strongly induced by TP53 and may be largely reponsible for the arrest of the cell cycle after telomere shortening. However, it is thought that p21CIP1 also accumulates in cells that proliferate continuously, independently of TP53, since it is induced by many proliferative stimuli. This is certainly so for p16INK4A which is not regulated by TP53. In somatic human cells p 16INK4A is induced by E2F and other transcription factors activated during cell cycle progression. Because the protein has a relatively long half-life, it accumulates when successive cell cycles follow rapidly upon each other. In some cell types that express p57KIP2, this inhibitor behaves in a similar fashion. So, the level of certain CDK inhibitors - like telomere length - depends on the number of successive cell cycles. This may provide a second counting mechanism.

However, in this mechanism counting not only depends on the actual number of cell cycles, but more critically on how quickly they follow each other and on which

Figure 7.8 CDK inhibitors as regulators of replicative senescence The width of the arrows indicates the presumed strength of the influences.

signals elicit proliferation. An extreme case is hyperproliferation induced by oncogenes such as RAS and MYC. In human cells, such hyperproliferation induces not only p14ARF1 to sensitize TP53, but also p16INK4A. Together, these proteins lead to a rather quick arrest of the cell cycle, certainly more rapidly than the telomere shortening mechanism would. This mechanism could account for the different lifespans of different human cell types in culture, because it may be more sensitive in epithelial cells that become relatively soon senescent in culture. More generally, the involvement of both p14ARF1 and p16INK4A in the response to hyperproliferation in human cells may explain why the CDKN2A locus is such a frequent target for inactivation in such a wide variety ofhuman cancers (^5.3). Specifically, it may solve the enigma why p16INK4A of all INK4 proteins is the most important tumor suppressor.

The mechanisms involved in the regulation of replicative senescence constitute one of the more important differences between humans and rodents with regard to cancer. Since these mechanisms may be related to organism aging (Box 7.1), this is plausible. A two year old mouse is approaching old age, whereas a two year old human is a toddler and a long way from maturity. Moreover, 70 kg humans living for 70 years or so may require additional mechanisms for protection against cancer than 50 g mice living for 30 months. On a less intuitive argument, it has been observed for a long time that human cells are much more difficult to transform in vitro than rodent cells. It had been a long-standing speculation that there might be (at least) one additional mechanism that protects them from becoming cancerous. It is now established that somatic cells in rodents more generally express telomerase and telomeres in rodents are longer than in humans. Moreover, the regulation of

CDK inhibitors is different, particularly that of p16INK4A. There is good reason to believe that the long-sought difference may reside here.

Further reading

Johnson FB, Sinclair DA, Guarante L (1999) Molecular biology of aging. Cell 96, 291-302

Los M et al (2001) Caspases: more than just killers? Trends Immunol. 22, 31-34

Chen G, Goeddel DV (2002) TNF-R1 signaling: a beautiful pathway. Science 296, 1634-1635

Wajant H (2002) The Fas signaling pathway: more than a paradigm. Science 296, 1635-1636

Brenner C, Le Bras M, Kroemer G (2003) Insights into the mitochondrial signaling pathway: what lessons for chemotherapy? J. Clin. Immunol. 23, 73-80 Cory S, Huang DC, Adams JM (2003) The Bcl-2 family: roles in cell survival and oncogenesis. Oncogene 22, 8590-8607

Feldser DM, Hackett JA, Greider CW (2003) Telomere dysfunction and the initiation of genomic instability. Nat. Rev. Cancer 3, 623-627 Franke TF et al (2003) PI3K/Akt and apoptosis: size matters. Oncogene 22, 8983-8998 Hahn WC (2003) Role of telomeres and telomerase in the pathogenesis of human cancer. J. Clin. Oncol. 21, 2034-2043

Kucharczak J et al (2003) To be, or not to be: NFkB is the answer - role of Rel/ NFkB in the regulation of apoptosis. Oncogene 22, 8961-8982 Schwerk C, Schulze-Osthoff K (2003) Non-apoptotic functions of caspases in cellular proliferation and differentiation. Biochem Pharmacol. 66, 1453-1458 Ben-Porath I, Weinberg RA (2004) When cells get stressed: an integrative view of cellular senescence. J. Clin. Invest. 113, 8-13

Castedo M et al (2004) Cell death by mitotic catastrophe: a molecular definition. Oncogene 23, 28252837

Debatin KM, Krammer PH (2004) Death receptors in chemotherapy and cancer. Oncogene 23, 2950-2966 Sharpless NE, DePinho RA (2004) Telomeres, stem cells, senescence, and cancer. J. Clin. Invest. 113, 160-168

Box 7.1: Human aging and cancer

Theories on the causes of human aging basically fall into two groups. One group assumes that the phenotypic changes associated with aging are caused by the accumulation of unrepaired damage to tissues, cells and macromolecules. One variant emphasizes oxidative damage by reactive oxygen species. Indeed, changes progressing with age can be found in extracellular tissue and in cells, including base mutations and epigenetic changes in the nuclear and mitochondrial genomes.

A second group of theories stresses that the regularity of the changes occurring with age reminds one of a genetic program - just like that controlling development and maturation. A minimum version of this sort of theory suggests that humans are 'build' to last for a certain period sufficient for reproduction, protective mechanisms holding out only so long. This version is easily reconciled with damage theories.

Replicative senescence is defined at the cell level. Although 'senescent' cells can be observed in aging humans, it is not clear to what extent this phenomenon contributes to human aging at the tissue level and the entire organism. Replicative senescence can be straightforwardly integrated into theories of programmed aging, but is neither incompatible with damage accumulation theories.

The fact that the majority of human cancers arise in older people and the incidence, prevalence and mortality of many cancers increase with age (cf Fig. 2.9) is compatible with both theories, perhaps better with the damage theory. In fact, there are indications that at very old age (>85 years), the incidence and aggressiveness of cancers also diminish. Again, both theories hold explanations for this (uncertain) effect, but the explanation by program theories is more elegant, i.e. cancer cells, too, are affected by the programmed loss of 'vigour'.

One might have thought that the elucidation of the genetic basis of human premature aging syndromes would have decided the debate. Their very existence has traditionally been used as an argument in favor of program theories. While premature aging is observed in several syndromes, including some resulting from defects in DNA repair and cell protection (^3.4), the prototypic diseases are the Hutchison-Gilford and Werner syndromes, which differ in the age of onset and the range of symptoms. Hutchison-Gilford syndrome is caused by mutations in the lamin A gene. This is puzzling, since it is everything but clear why defects in the nuclear membrane should be associated with prepubertal aging. The Werner syndrome is caused by muations inactivating the WRN helicase-exonuclease. The protein is certainly involved in DNA repair, making a good case for damage accumulation theories. Still, it may be particularly important for the maintenance of telomeres (^7.4), as might be expected for a protein involved in programmed aging. Moreover, the syndrome sets in at puberty, apparently dependent on hormonal changes, thereby fulfilling another postulate of program theories.

Hisama FM, Weissman SM, Martin GM (eds) Chromosomal instability and aging, Marcel Dekker, 2003. Hayflick L (1994) How and why we age. Ballantine.

0 0

Post a comment