Telomeres are nucleoprotein structures that protect the ends of human chromosomes through the formation of a ‘cap’, thus preventing exonucleolytic degradation, inter- and intra-chromosomal fusion, and subsequent chromosomal instability. During aging, telomere shortening correlates with tissue dysfunction and loss of renewal capacity. In human cancer, telomere dysfunction is involved in early chromosome instability, long-term cellular proliferation, and possibly other processes related to cell survival and microenvironment. Telomeres constitute an attractive target for the development of novel small-molecule anti-cancer drugs. In particular, individual protein components of the core telomere higher-order chromatin structure (known as the telosome or ‘shelterin’ complex) are promising candidate targets for cancer therapy.
The 2009 Nobel Prize in Physiology or Medicine was awarded to Elizabeth Blackburn, Carol Greider and Jack Szostak, who discovered the telomere and telomerase, which are linked to crucial biological problems with important medical implications.
The importance of telomeres and telomerase in biomedicine and the presumed therapeutic utility of telomere-related molecules is easily grasped from the following numbers: 92 telomeres contribute to the protection of the 23 pairs of chromosomes in the nucleus of every normal human somatic cell; 10365 papers about telomeres are included in the PubMed database; and in 2008 alone, telomeres were the subject of 3685 articles, of which at least 800 involved clinical studies and another 400 were clinical trials involving preclinical molecules targeting telomeres. Currently, telomeres and telomerase are considered the prime suspects in aging-related pathologies, tumorigenesis, congenital diseases and other pre-aging syndromes . No other DNA structure is as closely involved in issues pertaining to aging as telomeres, and telomerase activity is recognized as one of the broadest-spectrum biomarkers for detecting advanced cancers. Telomere function and telomerase activity contribute to the mechanisms underlying human aging and cancer formation. Telomeres are not irrelevant pieces of chromatin inside our cells. They are the subject of intense study by doctors searching for effective medications against cancer and aging-related diseases. For doctors, telomeres go beyond being a target for disease treatment; instead, they have emerged as an early diagnostic or prophylactic tool. This explains the continued focus of scientists on the telomere story from the perspective of both basic research and clinical application.
Telomere and chromosome end stability
Telomeres are structures at the ends of linear chromosomes (Figure 1). They are composed of telomeric DNA, which consists of short, tandemly repeated sequences and various telomere-associated proteins . Most telomeric DNA is in duplex form; however, the 3′-end of telomeric DNA consists of a G-rich overhang (the G tail). Telomeric DNA and its associated proteins form a specialized structure that controls the stability and function of DNA ends . Notably, telomeres prevent chromosome termini from being recognized as DSBs (double-stranded breaks) and from triggering DNA damage checkpoint-induced responses  (Figure 1). Telomeres are essential for various chromosome functions and the control of cellular proliferation; in contrast, DSBs represent a threat to the integrity of the genome due to their ability to generate chromosome rearrangements. Despite these functional and structural differences, DSBs and telomeres share a common set of proteins. At present, the integrated functioning of the various telomere components is poorly understood.
Telomere length as a determinant of aging
During each round of replication, human telomeres lose 20–100 bp of repeated sequence. After 40–80 divisions, human telomeres become so short that they are interpreted by the cell as dangerous and trigger either replicative senescence or apoptosis [5–8] (Figure 1). Therefore the progressive shortening of telomeres determines the lifespan of somatic cells . Evidence from cultured human fibroblasts shows that cells with shorter telomeres undergo significantly fewer doublings compared with those with longer telomeres . Studies in mice indicate that short telomeres reduce body mass and the potential for tissue regeneration, angiogenesis and wound healing .
Numerous studies have shown that telomere length is negatively correlated with age (Figure 2). Telomere arrays are much shorter in adult colonic mucosa and blood than in fetal tissue and sperm [12,13]. Interestingly, telomere length may predict pathological aging. For instance, elderly individuals with short average telomere lengths have increased rates of organ failure and mortality in chronological aging diseases of various origins, including heart failure, hypertension, myocardial infarction, immunodeficiency, digestive tract atrophy, atherosclerosis, liver cirrhosis and aplastic anaemia [14,15].
The use of telomere length as a predictive biomarker of pathological aging is attractive; however, telomere length is highly variable among individuals as a result of genetic influences and lifestyle, which includes social pressure, diet, smoking, inflammation and oxidative stress [16–19]. Nevertheless, an analysis of 175 same-sex Swedish twins revealed that telomere length predicts survival beyond the impact of one’s early living environment and genetic influence, and represents a major risk factor for mortality .
On the other hand, the use of telomere functional markers might be more appropriate than simply the length of telomeric DNA. This might be achieved by assaying the recruitment to telomeres of checkpoint and repair proteins [21,22]. Interestingly, dysfunctional critically short telomeres not only trigger intrinsic cell pathways, they also trigger extrinsic mechanisms leading to impaired stem cell function and possibly contributing to aging . Such extracellular factors secreted by cells suffering from telomere dysfunction might also be useful markers of pathological aging and cancer risk .
Telomerase and tissue renewal
The progressive loss of sequence from telomeres at each round of cell division is like a counting mechanism that limits cell-renewal capacity. In certain cell types, however, telomere loss can be overcome by a reverse transcriptase named telomerase . This enzyme is composed of a catalytic subunit [TERT (telomerase reverse transcriptase)] and an RNA template [TERC (telomerase RNA component) or TR]. Forced expression of telomerase prevents cultured fibroblasts from entering senescence [26,27]; therefore cells with telomerase activity may become immortal .
The catalytic subunit of telomerase is expressed in the stem cell compartment of several adult tissues, although its level does not appear to be sufficient to fully compensate for replicative erosion. Human telomerase expression is reduced dramatically in many somatic cells during embryonic development [29,30]. Germline mutations in the human TERT and TERC genes are found in families with premature aging syndrome, indicating that telomeres and telomerase are involved in the human aging process. Early mortality through bone marrow failure in X-linked and TERC-related dyskeratosis congenita is due to a defect in telomerase activity . In idiopathic pulmonary fibrosis, telomerase impairment limits tissue renewal capacity in the lungs and germline [31,32]. These data suggest that increased telomerase expression extends the renewal capacities of tissues. Indeed, excess TERT expression can rescue the renewal capacity of epidermal stem cells ; the reintroduction of telomerase TERC in telomerase-deficient mice with inherited short telomeres prevents further telomere shortening, chromosomal instability and loss of organismal viability ; and TERC delivery to mice with end-stage liver cirrhosis can rescue telomere function and improve liver function .
The dual role of telomerase in oncogenesis
A major problem with using telomerase therapy to counter the adverse effects of aging is that telomerase is also overexpressed in approx. 85% of human cancers . Moreover, telomerase acts as an oncogene, contributing to the transformation of cultured primary human cells into neoplastic cells [37,38]. Therefore telomerase activity is a link between cellular immortalization and malignant transformation. How then can one manipulate the telomerase to cure cancer without accelerating aging? How can one manipulate telomerase to prevent aging without facilitating cancer progression? Can it be used to safely prolong the renewal potential of cells? What are the telomere-related processes that determine the balance between aging and cancer?
If telomerase is required for cellular immortalization and oncogenesis, one would expect that patients with reduced telomerase capacity should be more resistant to cancer. However, patients with congenital dyskeratosis also have an increased predisposition to cancer. In addition, numerous clinical studies of tumour samples have shown that carcinogenesis, including hepatocarcinogenesis, is associated with a reduction in telomere length and an increase in chromosome instability [39,40]. Moreover, short telomeres in non-transformed leucocytes may be a risk factor for lymphomagenesis . Telomerase-null mice mimic the phenotypes associated with telomere shortening, including a shortened lifespan and increased incidence of spontaneous malignancies .
Finally, it should be kept in mind that telomerase activity is undetectable in 5–10% of human cancers; in such cases, an alternative method of telomere elongation [ALT (alternative lengthening of telomeres)] is activated . Telomerase-negative cells that escape senescence/crisis may maintain the length of their telomeres via recombination.
Interestingly, if telomere dysfunction in telomerase-deficient mice increases the incidence of cancer, the tumours usually remain in situ and are resistant to progression [44,45]. This suggests that an absence of telomerase in the somatic cells of the elderly contributes to cancer initiation, whereas subsequent telomerase overexpression is required for cancer progression. Indeed, the absence of telomerase in p53 heterozygotes increases the incidence of carcinoma in elderly mice . This result also suggests that p53 is a major barrier against cancer initiation in response to telomere dysfunction.
Telomerase as a universal biomarker of cancer
Telomerase activity is thought to be one of the most widely expressed cancer biomarkers, and it is even more sensitive and specific than p53 mutations . Owing to its high levels of expression and sensitivity, telomerase activity is also detectable in tiny samples, including those from tissues that are difficult to access or are present in limited amounts. Clinical studies of telomerase activation from gastric lavage fluid in patients with gastric cancer, in fine-needle-aspirated samples and in ascitic fluid from metastatic carcinoma patients, indicate that telomerase activity could be a convenient, less-invasive diagnostic marker or prognostic factor in the surveillance of cancer therapy. These findings suggest the use of telomerase activity as a clinical biomarker in a broad spectrum of cancers [47–51].
Telomere nucleoprotein structure
A universal characteristic of telomeric DNA is the presence of a 3′-overhang corresponding to the G-rich strand. This G-tail can form two different structures. First, the G-tail can invade the duplex portion of telomeric DNA, forming a t-loop structure, which has been proposed to hide the chromosome ends from DNA damage response signalling and to prevent inappropriate DNA repair at these sites . Secondly, Hoogsteen G–G base-pairing may lead to the formation of a four-stranded structure known as a G-quadruplex (or G4). This complex structure may delay the progression of the replication fork, preventing efficient telomere replication. G4 structures are also involved in telomerase regulation .
Recently, it was found that telomeric DNA is transcribed into a non-coding RNA called TERRA (telomeric repeat-containing RNA). This RNA remains attached to the telomere, suggesting that it plays an important role in the folding of telomeric chromatin. Although the function of TERRA at telomeres is unknown, its presence suggests that telomeric structure may affect the epigenetic regulation of gene expression through chromosome end regression [54–56].
Several telomere-associated proteins have been shown to play roles in the protection and stabilization of telomere structure and function .
TRF1 (telomeric-repeat-binding factor 1) and TRF2 (telomeric-repeat-binding factor 2) are paralogue telomere-binding proteins in vertebrates [58–61]. They share a common structural domain consisting of a TRFH (TRF homology) domain and a C-terminal SANT/Myb DNA-binding motif (also called a telobox), which binds specifically to duplex TTAGGG repeats . TRF1 and TRF2 bind DNA as homodimers or oligomers through homotypic interactions at the TRFH domain. The TRFH domains of TRF1 and TRF2 function as telomeric protein docking sites that recruit other telomeric DNA-binding proteins, including TIN2, POT1 (protection of telomeres protein 1), TPP1 (three-prime phosphatase 1) and hRap1 (human repressor activator protein), which have distinct functions at chromosome ends [2,57,62–64]. These six telomere-related subunits are essential for telomere function through the formation of a large multimeric complex known as a shelterin or telosome [2,3,57,65,66].
In the shelterin multimeric complex, TRF1 and TRF2 bind to a double-stranded telomeric sequence, whereas POT1 binds to the 3′-overhang. hRap1 is recruited to telomeres by TRF2. TPP1 interacts with telomeres through interactions with TIN2 and POT1. TIN2 binds to TRF1, TRF2 and TPP1, thereby bridging the shelterin components bound to double- and single-stranded telomeric DNA [67,68].
A second telomeric complex, t-RPA, was discovered in yeast through the capping properties of the single-stranded binding protein Cdc13, which turned out to be a heterotrimer named CST (for Cdc13-Stn1-Ten1) with structural properties similar to RPA. A wealth of recent data show that the CST complex is present and required for telomere capping in a wide range of organisms, including plants and mammals .
Telomere capping by the shelterin complex
Dysfunctional telomeres elicit the DNA-damage response [21,22]. In general, TRF2 protects telomeres against activation of the ATM (ataxia telangiectasia mutated)-dependent DNA-damage response and POT1 against the ATR (ataxia telangiectasia mutated- and Rad3-related)-dependent DNA-damage response  (Figure 1). Importantly, telomere capping depends both on the length of telomeric DNA (i.e. the possibility of forming a telomere-specific nucleoprotein structure) and on the abundance of telomere-binding factors. For instance, a long telomere in cells with low levels of bound TRF2 becomes uncapped and triggers the DNA-damage response.
TRF2 protects against DNA-damage checkpoints and repair in several ways. First, it inhibits the kinase activity of ATM  and CHK2 (checkpoint kinase 2) . In addition, TRF2 remodels and stabilizes linear DNA into t-loops, a structure believed to protect the 3′-overhang from the DNA-damage response [52,73–76]; it also prevents illegitimate fusion between telomeres, independent of its role in t-loop formation , and it serves as a chaperone for G-quadruplex formation .
In checkpoint-compromised cells, TRF2-deficient telomeres are processed by the DNA ligase IV-dependent NHEJ (non-homologous end-joining) reaction, which mediates DSB repair . This leads to an end-to-end chromosome fusion phenotype, which manifests itself as a disappearance of the 3′-overhang .
Unlike TRF2, POT1 represses the ATR pathway at chromosome ends (Figure 1). POT1 binds ssDNA (single-stranded DNA) in response to TPP1 recruitment, and inhibits the localization of RPA, another DNA-damage sensor, to telomeres . In human cells, POT1 also determines the structure of the G-tail. POT1 binds the 3′-overhang and may also coat the G-rich strand displaced at the t-loop. POT1 maintains the G-tail in an unfolded state by preventing G4 formation [3,82,83]. During telomere replication, POT1 may bind the single G-rich strands that accumulate when mechanisms that replicate that strand are defective. This binding most probably prevents RPA binding, thus allowing the replication on the other (C-rich) strand to continue .
Telomere proteins and telomere length regulation
The general picture of how shelterin and CST are involved in telomere length regulation is currently unclear; however, several of these proteins are part of subcomplexes with opposite roles in telomere length dynamics. The emerging view is that complexes that protect against the DNA-damage response at telomeres inhibit telomerase in cis, whereas telomerase access to the telomere requires transient uncapping at the time of telomere replication. This is expected to account for the negative-feedback mechanism that determines the mean telomere length in dividing cells [85–90]. For instance, POT1 exerts its negative role in telomere elongation by interacting with TRF1 , whereas POT1 and TPP1 are required for efficient telomerase recruitment and processivity [91,92]. On the other hand, the CST complex stimulates the polymerase α-primase, thereby preventing telomerase action.
From these observations, it has been suggested that TRF1 acts in cis to control the access of telomerase to chromosomal termini via a ‘protein counting’ mechanism analogous to that described in budding yeast [86,93]. In cases where TRF1, but not TRF2, is removed, the telomeres first lengthen then stabilize , suggesting the involvement of another feedback mechanism. This alternative pathway could involve TRF2, possibly by promoting t-loop recombination [95,96].
Shelterin expression in human cancers
Clinical studies indicate that telomeric protein expression is often higher in cancer tissues than in normal tissues. Interestingly, this usually occurs at the early stages of malignant transformation, as in atypical adenomatous hyperplasia lung tissue  or during hepatocarcinogenesis . In non-small-cell lung cancer, the expression of TRF1 and TRF2 is elevated, similar to hTERT (human TERT), and is closely related to the T-status and TNM (tumour, node, metastasis) stage of the tumour . TRF1 and TRF2 mRNA are overexpressed in gastric carcinomas and adult T-cell leukemias [99,100]. TRF2 is overexpressed in gastric, breast and skin cancer tissues [101,102]. Moreover, TRF2, but not TRF1, is significantly up-regulated in drug-resistant gastric cancer cells . Finally, the increased TRF2 expression detected in AML (acute myelotic leukaemia) patients with a poor prognosis suggests that TRF2 expression is related to the prognosis of leukaemia .
However, some clinical investigations indicate that the expression of telomeric proteins is reduced or unbalanced compared with normal cells in such conditions as B-CLL (B-cell chronic lymphocytic leukaemia), suggesting that changes in telomeric protein composition are involved in the pathogenesis of this lymphoid disorder . In another clinical study of non-small-cell lung cancer, TRF1 mRNA expression was significantly lower in tumour tissues than in adjacent normal tissue, and the level of RAP1 was highly predictive of overall survival, whereas the level of TRF2 was correlated with tumour grade .
Overall, various levels of telomere functional organization are likely to be important parameters in clinical oncology. These parameters might include the dosage and activity of factors that regulate the structure and function of telomeres (e.g. shelterin, CST and TERRA), the co-localization of PML (promyelocytic leukaemia) bodies with the components of telomeric chromatin (a signature of ALT cells ), and the quantification of telomeric dysfunction, including anaphase bridges, chromosome breaks visualized either in interphasic cells or in metaphase spreads, telomeric fusion in metaphase spreads, and DNA-repair foci (Figure 2).
Of note, no mutation in the genes encoding telomere-binding proteins is known to be involved in human oncogenesis (except indirectly, since they may provoke dyskeratosis congenita). This raises the interesting possibility that these genes act as oncogenes, and that their overexpression increases the proliferative capacity of tumour cells.
The role of TRF2 in oncogenesis
To mimic the TRF2 overexpression observed during carcinogenesis (see above), a mouse model overexpressing TRF2 in stratified epithelia has been studied (k5-Terf2). These mice have severe epidermal stem cell dysfunction [108,109]. The deletion of p53 restores epidermal stem cell behaviour, although it results in further telomere shortening. In addition, p53−/− K5trf2 mice have an increased incidence and accelerated onset of TRF2-induced skin cancer . These carcinogenic effects of TRF2 correlate with important telomeric DNA loss; thus it is unclear whether the effects of TRF2 overexpression in this system result from telomere shortening or a gain-of-function effect. Such telomere shortening might explain why k5-Terf2 mice show premature aging, in a similar manner to telomerase-deficient mice .
Several pieces of evidence suggest that TRF2 overexpression contributes to malignant transformation of human cells: (i) TRF2 inhibition reduces tumorigenicity in human melanoma cells ; (ii) the depletion of TRF2 by RNAi (RNA interference) in drug-resistant gastric cancer cells partially reverses the resistant phenotype, whereas an increased dosage of TRF2 exacerbates the resistant phenotype to several different chemotherapeutic agents ; and (iii) down-regulation of TRF2 by RNAi reduces the proliferative capacity of human colorectal carcinoma cells .
The role of TRF2 in carcinogenesis is elusive. It might help prevent senescence, apoptosis and growth arrest, as has been shown in several cellular models [113–116,117]. It could also be involved in general cell reprogramming processes favouring malignant transformation owing to its emerging role in chromatin structure and transcriptional regulation [117,118].
Toward combined telomerase–telomere anti-cancer therapies
We have seen that telomerase is reactivated in the vast majority of tumours, and that the disruption of telomere maintenance limits the cellular lifespan in human cancer cells, validating human telomerase as a valuable target in chemotherapy . Consequently, various anti-telomerase strategies are currently being developed to limit tumour proliferation in cancer patients. However, increasing experimental evidence suggests that unusual activities of this enzyme beyond telomere maintenance could play a role in cell transformation [119,120]. Changes in other specific components of telomeric chromatin (e.g. shelterin subunits) contribute to oncogenesis, and telomerase-independent pathways may maintain telomere length in tumour cells . Moreover, telomerase is expressed continuously throughout the cell cycle in dividing immortal cells, but it is repressed in quiescent cells [121,122]. This suggests that quiescent stem cells with the potential to express telomerase may be invulnerable to anti-telomerase cancer therapies.
Therefore it is necessary to target not only telomerase, but other telomeric functions. For instance, a wealth of data indicate that uncapped telomeres can also be obtained by pharmacological G-quadruplex stabilization . RHPS4 (3,11-difluoro-6,8,13-trimethyl-8H-quino[4,3,2-kl]acridinium methosulfate) is one of the most effective G4 ligands, showing high selectivity for the quadruplex DNA structure . By stabilizing G4 DNA at telomeres, this agent may impair both fork progression and telomere processing, resulting in telomere dysfunction [125,126]. Interestingly, in terms of clinical application, RHPS4 is active in vivo as a single agent with a good toxicological profile and it potentiates the anti-tumour efficacy of TOPO I (topoisomerase I) inhibitors in preclinical models of solid tumours .
Interestingly enough, telomere dysfunction enhances the sensitivity of cells to agents that induce DSB [128,129], suggesting that the combined use of DSB-inducing agents (either ionizing radiation or clastogenic molecules) and telomere inhibition is an effective anti-cancer therapeutic approach. Obviously, certain questions need to be addressed. What is the role of the non-canonical activities of telomerase in cancer and aging? When are these activities required? In view of these additional roles, should we envisage additional telomerase therapeutic targets? What is the role of telomeric proteins and TERRA in cell transformation and aging? For example, we can imagine that different telomere statuses exist in the same tumour, reducing the efficacy, and therefore the success, of therapies targeting only reverse transcriptase activity. Also, certain therapeutic protocols could select for ALT cells, favouring the proliferation of such cells.
In summary, we believe that the analysis of telomere changes during aging and cancer goes beyond the classical parameters of telomere length maintenance and telomerase activity. We anticipate that the functional exploration of telomere function during normal and pathological aging, as well as in tumours, will be extremely useful not only in the diagnosis and prognosis of human diseases, but also in predicting the patient response to therapy.
• Telomeres are structures at the ends of linear chromosomes, which consist of short, tandemly repeated DNA sequences, non-coding RNA and various telomere-associated proteins.
• The progressive loss of telomeric DNA sequence at each round of cell division limits cell renewal capacity.
• Telomere length is negatively correlated with age.
• In germinal and cancer cells, telomere loss is overcome by a reverse transcriptase named telomerase.
• Telomere shortening contributes to cancer initiation, whereas subsequent telomere maintenance is required for cancer progression.
• Telomerase activity is one of the most widely expressed cancer biomarkers.
• In addition to telomerase, telomere-associated proteins play roles in the protection and stabilization of telomere structure and function.
• Critically short or unfolded telomeres elicit the DNA-damage response
• Telomeric protein expression is often higher in cancer tissue than in normal tissue.
• Telomerase and telomere-binding proteins are considered as valuable targets in chemotherapy.
Work in the authors’ laboratory was supported by a grant from the ARC (Association pour la recherche sur le cancer).
- © The Authors Journal compilation © 2010 Biochemical Society