Concise Review: Telomere Biology in Normal and Leukemic Hematopoietic Stem Cells

Authors


Abstract

The measurement of telomere length can give an insight into the replicative history of the cells in question. Much of the observed telomere loss occurs at the stem and progenitor cell level, even though these populations express the enzyme telomerase. Telomerase-transfected hematopoietic stem cells (HSC), although able to maintain telomere length, are still limited in terms of ability to undergo sequential transplantation, and other factors require to be addressed to achieve optimal levels of stem cell expansion. Unchecked telomere loss by HSC, meanwhile, would appear to play a significant role in the pathogenesis of bone marrow failure, as observed in the condition dyskeratosis congenita. This heterogeneous inherited condition appears to exhibit telomerase dysfunction as a common final pathogenic mechanism. Although less well-established for acquired marrow failure syndromes, mutations in key telomerase components have been described. The identification of the leukemic stem cell (LSC), along with the desire to target this population with anti-leukemia therapy, demands that telomerase biology be fully understood in this cell compartment. Future studies using primary selected LSC-rich samples are required. A better understanding of telomerase regulation in this population may allow effective targeting of the telomerase enzyme complex using small molecule inhibitors or additional novel approaches.

Disclosure of potential conflicts of interest is found at the end of this article.

Introduction

The notion that cancer may be underpinned by the malignant transformation of a critical stem cell population, although relatively recent in terms of solid tumor biology, is one that has been accepted for many years by scientists studying leukemia. The ability to isolate and characterize the leukemic stem cell (LSC) population has facilitated steady progress in understanding how such cells may attain an immortal phenotype. One prerequisite is the ability to maintain telomere length via expression of the widely conserved enzyme telomerase. This review will summarize the current understanding of telomere biology as it relates to hematopoietic stem cells, from steady state normality through to stem cell transplantation and manipulation and finally in hematological malignancy.

The bone marrow of an adult human being is estimated to produce in excess of 1011 blood cells per day, all of which are ultimately derived from pluripotent hematopoietic stem cells (HSC) (reviewed in [1]). These unique cells, with their paradoxical leukemogenic and therapeutic potential (for example, in HSC transplantation), have stimulated enormous research effort. All mature circulating blood cells are derived from a relatively small number of HSC via a successive series of intermediate progenitors displaying steadily increasing lineage commitment (reviewed in [2]). Such a hierarchical model of hemopoiesis has provided a paradigm for the development of other tissues and, more recently, leukemogenesis and tumorigenesis [3]. Candidate HSC can be isolated and characterized both functionally (e.g., by the ability of single cells to reconstitute multilineage hemopoiesis in a mouse model or by the ability to establish long-term cell culture in vitro) and by immunophenotyping. The latter is commonly performed in the clinical and laboratory setting using the CD34 antigen (a surface glycoprotein) as a surrogate HSC marker [4]. The properties of CD34+ cells are well-characterized from both research and clinical perspectives, although it is apparent that even the CD34+ progenitor cell population is in itself extremely heterogeneous [5]. True HSC most likely represent a tiny proportion of CD34+ cells: CD34+ cells negative (or low) for CD38, thy-1, and CD71 are enriched for HSC activity [6]. Such cells are predominantly quiescent (i.e., in G0) [7], and it has been estimated by mathematical modeling that they divide only once every 1–2 years [8]. Unfortunately, such cells are difficult to study in vitro; in addition to their rarity, the mechanism of self-renewal versus differentiation is not clear, and candidate HSC will expand and differentiate in culture conditions, with acquisition of lineage-specific markers and concomitant loss of stem cell phenotype.

Although HSC have an enormous capacity to self-renew and/or differentiate, this capacity appears to be finite; for example, murine HSC can only be serially transplanted 5–7 times in mice before hemopoiesis is exhausted [9]. Although it is possible that extrinsic stress (i.e., ex vivo handling of cells, engraftment in a hostile bone marrow microenvironment, etc.) may contribute to this observation, the concept that an intrinsic mechanism is operating to limit cell expansion, by registering accumulated cell divisions, has always been attractive. Therefore, the discovery that telomeres (DNA-protein “caps” that exist at all eukaryotic chromosomal termini) shorten with each cell division in vitro [10] and with age in vivo [11, [12]–13] subsequently implicated this as a likely mechanism.

The “End-Replication Problemand Telomerase

Semiconservative replication of DNA presents a unique problem: the process only works in the 5′ to 3′ direction, and DNA polymerase requires binding of an RNA primer. Olovnikov [14] and Watson [15] predicted the consequences of this long before the telomere was characterized and termed it the end-replication problem. This anticipated the loss of a small 5′ nucleotide segment as DNA synthesis took place, with progressive replication-induced telomere shortening (Fig. 1). The cellular consequence of unchecked telomere loss is replicative senescence, a viable physiological state but one from which no further cell division can occur. It has been hypothesized that this serves as an evolutionary conserved tumor-suppressor mechanism, analogous to checkpoints during cell division [16].

Figure Figure 1..

The end-replication problem. The 3′ end of the molecule is replicated continuously to the very end, having started at the opposite, 5′, terminus. However, synthesis of the lagging strand (also in the 5′ to 3′ direction) is discontinuous because of the requirement for RNA primer binding and unidirectional growth of the new strand. Upon removal of the RNA primers, the gaps in the discontinuous lagging strand are filled in and ligated. However, there is no provision for such a process at the immediate 5′ end, thus leaving a gap. Since the DNA duplex is antiparallel, each daughter molecule will be shortened on its 5′ end after replication, with successively shorter daughter chromosomes resulting from additional cycles of cell division.

The widely conserved enzyme telomerase (via its ability to reinstate lost telomeric DNA-repeat sequence) enables cells to bypass replicative senescence and can confer cellular immortality [17]. The two crucial constituents of telomerase (which together can reconstitute telomerase activity in vitro in a rabbit reticulocyte-lysate model [18]) consist of a catalytic protein known as human telomerase reverse transcriptase (hTERT) [19, 20] and an RNA template molecule that contains a sequence complimentary to the human telomeric DNA (hTR) [21, [22]–23]. Whereas hTR can be detected in most tissues, hTERT is restricted in its expression [20, 24, 25]. Introduction of the latter is sufficient to reconstitute telomerase activity, and this has led to the concept that hTERT is the primary rate-limiting component of telomerase. This is, however, a simplistic view, as hTR can be limiting for telomere maintenance in some circumstances (described below).

Telomerase-deficient murine HSC are considerably less able to undergo serial transplantation than wild-type [26]. Human HSC from umbilical cord blood (UCB) (with long telomere sequence) have greater proliferative potential than bone marrow (BM) HSC from older donors [27]. Thus, for successful ex vivo expansion of HSC (e.g., for tissue engineering or gene therapy, both long-held ambitions of HSC biologists), the ability to manipulate telomerase (and consequently the telomere) is likely to be critical.

Telomerase activity is highly expressed germline cells, thereby protecting them against telomere shortening. Human germline telomeres are typically maintained at approximately 15 kilobase pairs (kbp) [28]. In contrast, telomerase is not expressed in most normal somatic cells, which is in line with the observation that these cells have a finite life span in vitro and undergo cellular replicative senescence after 40–80 population doublings [29], after considerable telomere shortening. High levels of telomerase activity have been detected in transformed cell lines and malignant tumors by the sensitive telomere repeat amplification protocol (TRAP) assay [30]. It has therefore been postulated that the process of immortalization typically requires the reactivation of telomerase for telomere stabilization and maintenance in tumor cells [31, 32]. That many leukemias (and indeed other cancers) are thought to arise in a stem cell population has given further import to understanding telomere biology and the role of telomerase in this rare but critically important cell type.

Telomere and Telomerase Biology in the Normal and Transplanted HSC Population

Steady-State Hemopoiesis

The primary determinant of telomere length in individuals has been shown to be genetic, as confirmed by twin studies [12, 33, 34], strain-specific variation in telomere length in mice [35], and the fact that telomere length on specific chromosome arms from different tissues of the same individual are similar [36]. Presumably as a result of subsequent cell division, cross-sectional studies have demonstrated age-related telomere loss in a range of tissues. For example, HSC and progenitor cell telomere loss is detectable in their peripheral blood leukocyte (PBL) progeny. Rufer et al., in a study of over 300 normal individuals of all ages, detected telomere shortening in peripheral blood (PB) granulocytes and lymphocytes of 39 and 59 base pairs (bp) per year, respectively [12], an observation confirmed by others [37]. Interestingly, more rapid loss (1,000–3,000 bp/year) was observed in the first 2 years of life. If one assumes that the number of divisions between HSC and mature progeny are identical throughout life and that replication-induced telomere loss is constant with each division, then PBL telomere length reflects the replicative history of HSC [16]. These data predict that the average HSC divides every 1–2 years (based on measured loss of 50–100 bp per division) after the age of 2 years or so, which is consistent with data derived from mathematical modeling of the mutation rate in red cells [8]. Interestingly, recent data suggest a further acceleration of telomere loss after the age of 60 years [16]. The reasons for accelerating telomere shortening in the elderly are not clear, but it could be due to replicative stress within a diminishing HSC pool or to the consequences of ineffectual DNA repair and/or antioxidant defense mechanisms.

From these data, it is apparent that telomerase is unable to prevent telomere shortening in the HSC compartment. Despite this observation, telomerase activity is readily detectable in HSC: it is expressed at a low level in steady-state CD34+ cells, is highest in the CD34+/38+ subpopulation (which contains proliferating progenitor cells), upregulates in response to cytokine stimulation in vitro within 48 hours (peaking at day 7 and declining to baseline after 3–4 weeks), and correlates with cell cycle status [38]. However, despite telomerase activity, telomeres shortened by 1–2 kbp over the 4-week culture period (albeit at a slower rate during the period of telomerase expression, with an overall loss of 73 bp per division). Telomerase activity was low in the quiescent CD34+/38 population and could also be downmodulated with transforming growth factor β. Therefore, despite these cells being telomerase-competent, telomere loss still occurs. Thus, in vitro telomerase activity per se is not adequate to maintain telomere length and is therefore not synonymous with telomere maintenance.

Telomere Loss After HSC Transplantation

Notaro et al. first demonstrated significant telomere shortening, of the order of 79–1,446 bp, in the peripheral blood granulocytes of allogeneic stem cell transplant (SCT) recipients as compared with their donors [39]. The inverse correlation between the number of mononuclear cells received and the degree of telomere shortening was interpreted as reflecting the additional cell divisions required to achieve hemopoietic reconstitution. Telomere loss post-transplant (an average of approximately 1 kilobase [kbp]) has been shown to be restricted to the first year of regeneration after SCT, followed by a more gradual loss (comparable to the normal rate) thereafter [40, [41]–42]. The results of studies in children post-SCT are similar, demonstrating telomere shortening equivalent to approximately 15 years of aging [43]. As to the pathological significance of these observations, little is known. It must be remembered that late failure of a successful graft is rare, and some individuals in these studies had no demonstrable telomere loss post-transplant. Early onset of clonal hemopoietic disorders might be predicted, in a fashion analogous to that of aplastic anemia, where telomere shortening correlated with chromosomal abnormalities and where clonal disorders have been described [44]. Further study, for example on the role of telomere shortening in dysplastic states observed post-autologous SCT, is required, although clearly there will be many confounding variables, such as the mutagenic effects of treatment. Its significance may only eventually be realized with the ex vivo manipulation of stem cells, such as that used in gene therapy or UCB transplant protocols, where a small number of stem cells are required to fully reconstitute an adult marrow. In this regard, a recent report described stem cell expansion with preserved telomerase activity for 18 weeks in culture [45]. This was achieved purely by optimization of culture conditions, including cytokine and stromal support. In an analogous study, in vivo administration of granulocyte colony-stimulating factor has been postulated to attenuate telomere attrition induced by repeated cycles of chemotherapy [46].

In wild-type mice serially transplanted with transgenic mTERT-expressing BM or HSC (c-Kithi, Sca-1hi, Thy1.1lo, Linneg), no telomere shortening was observed after four serial transplants [47]. This was in comparison to the wild-type transplanted animals, in which 40% of the serially transplanted donor telomeric DNA was lost. However, it was not possible to transplant beyond a fourth attempt; therefore, telomere maintenance with mTERT alone did not extend replicative capacity. The authors of that study suggested that this may reflect the influence of telomere-independent senescence mechanisms, due to repeated HSC handling and cell stress, pointing to a telomere-independent “culture-shock” that has been described by others [47, 48]. Potentially, this has major implications for tissue engineering, ex vivo cell expansion, and so on. Overexpression of hTERT per se is unlikely to be the sole requirement for unlimited cell expansion, and it is likely that meticulous attention to cell culture conditions and similar factors will be required to avoid inducing cell senescence via other routes.

The Telomerase Knockout Mouse

Mice lacking mTR exhibit excessive telomere shortening (4–5 kbp loss per mouse generation [49]) and lack telomerase activity. Late-generation mTR−/− mice demonstrate infertility and reduced viability and often exhibit defects in highly proliferative tissues [50], such as the hematopoietic system [51]. Late-generation mTR−/− B6 mice show splenic atrophy, abnormal hematopoietic function, and impaired B- and T-cell reaction to mitogen stimulation. Defective germinal center reactions following immunization have also been noted [52]. A significant decrease in the total number of committed hematopoietic progenitor cells in late-generation as opposed to early-generation mTR−/− mice was described [50]. Serial transplantation experiments in mice have compared wild-type (mTR+/+) with telomerase knockout (KO) HSC (mTR−/−), with the former successfully transplantable over four rounds, compared with only two for the telomerase-deficient population [53]. Furthermore, the latter cells exhibited a degree of telomere loss nearly twice that of cells from the wild-type population. These results support telomere maintenance as a prerequisite for stem cell replicative capacity.

Telomere Dynamics in the Lymphoid Population

Mature lymphocytes must undergo massive clonal expansion in response to antigenic stimulation. Consequently, memory CD4+ and CD8+ cells have significantly shorter telomere sequences than their naïve counterparts [54, 55]. These additional replications likely explain the more rapid telomere loss in lymphocytes (as compared with granulocytes) noted in cross-sectional population studies. Again, this underscores the principle that telomere loss remains significant, even in telomerase-competent cells such as lymphocytes. Potential sequelae of such telomere loss, primarily immunosenescence, have been discussed extensively and are beyond the scope of this review (this subject is reviewed in [56, 57]). Such a mechanism could, however, explain the inefficiency of immune surveillance post-transplant of BM or HSC (as suggested by the progressive increase in risk of solid tumors), as well as loss of immune function in the elderly. Again, these in vivo effects may be exaggerated by future ex vivo manipulation of such cells.

Telomere Length in Blood Donors

The total number of blood cells donated over the lifetime of a volunteer donor can be equivalent to several years of steady-state blood cell production [58]. Telomere length in PB cells of whole blood and apheresis platelet donors, however, failed to demonstrate a difference in telomere length between donor and nondonor. Furthermore, there was no correlation between telomere length and the total number or the total years of donation. At least as indicated by telomere length analysis, the impact of frequent blood donations on the replicative capacity of the human HSC pool would appear to be insignificant [58].

Telomere Dysfunction in Hematopoietic Disease

Dyskeratosis Congenita: An Inherited Disorder of Telomere Maintenance

Reduced telomere length has been documented in patients with the progressive BM failure syndrome dyskeratosis congenita (DKC) [59]. Abnormalities in these patients include skin pigmentation, nail dystrophy, and leukoplakia, and similarities have been noted with the telomerase KO mouse. At first sight, it is a genetically heterogeneous condition, with X-linked, autosomal recessive, and autosomal dominant inheritance all having been described in different pedigrees. The clinical spectrum of the disease ranges from severely affected individuals in the first decade of life (as in Hoyeraal-Hreidarsson syndrome) to asymptomatic patients with only minor changes in blood counts (reviewed in [60]). It is becoming apparent, however, that disruption of telomere maintenance may be a final common pathway in the pathogenesis of this disorder. The X-linked form of the disease is due to missense mutations in the gene DKC1 on the long arm of the X chromosome [61]. The affected protein, dyskerin, is a nucleolar protein that is associated with hTR [62]. Furthermore, in patients with autosomal dominant DKC, both deletions, as well as mutations involving hTR, have recently been described [63]. Mutations in the DKC1 or hTR gene can be identified in less than 50% of the patients [64]; however, these findings lend considerable weight to a link between BM failure syndromes and failure of telomere maintenance. More recently, a mutation in a functional SP1 binding site of telomerase has been identified as an additional mechanism [65] in an acquired bone marrow failure syndrome, in this case paroxysmal nocturnal hemoglobinuria (PNH) [66, 67]. The possibility of “telomerase therapy” has been proposed for these individuals and is currently being studied preclinically.

Acquired Bone Marrow Failure Syndromes

Acquired aplastic anemia (AA) is thought to result from damage to the stem and progenitor compartment caused by autoreactive T cells. Although many patients recover with immunosuppressive therapy, the majority exhibit a deficit in stem cell numbers despite apparent hematological recovery [68]. Severe AA also confers an increased risk for development of late clonal marrow disorders, such as PNH (25% at 15 years), myelodysplastic syndromes (MDS) (probability ranging from 10% to 47%), and acute leukemia. Although this might be due to impaired immune surveillance, a corresponding rise in solid cancers has not been described as might be expected if this were the primary mechanism. A more attractive hypothesis might involve disruption of intrinsic cell processes resulting in premature aging or genetic instability in the marrow stem cell compartment, ultimately leading to clonal hemopoiesis.

In granulocytes from patients with AA, not only was telomere length found to be significantly shorter compared with age-adjusted controls [44], but the degree of shortening correlated significantly with the severity of cytopenia [69]. Intriguingly, telomere length in granulocytes from AA patients successfully treated with immunosuppressive therapy did not differ significantly from controls. Untreated and nonresponding patients with persistent severe pancytopenia continued to exhibit significant telomere shortening [69]. Although these results imply extensive proliferation of HSC in subgroups of AA patients (reviewed in [70]), the question remains as to whether this is a secondary phenomenon (reflecting increased HSC turnover as a consequence of damage to the stem cell compartment) or indicates telomere-mediated replicative exhaustion of the HSC pool. Support for the latter has come from recently described mutations of the hTR gene in hematopoietic cells from patients with acquired BM failure [71]. However, these mutations occur at such a low frequency (2 of 150 patients with AA; 0 of 13 patients with PNH) that they would appear not to be the major cause of otherwise clinically typical BM failure syndromes [72]. More recently, heterozygous mutations in the TERT gene have been detected in patients with AA and shown to impair telomerase activity via haploinsufficiency [73].

Telomere Biology and Hematological Malignancy

Progressive telomere shortening is well described in hematological malignancies [37, 74, [75], [76]–77] and is thought to result from marked clonal expansion, although oxidative damage [78] or telomerase dysregulation may be contributory at least in the early stages of some leukemias [79]. Furthermore, there is evidence that critical telomere shortening with resulting genetic instability may promote tumor evolution and telomerase activation or upregulation, during which critically short telomeres are stabilized and ongoing tumor growth is facilitated. A biphasic pattern of telomere and telomerase kinetics has been proposed, at least as regards progression of hematological malignancy such as chronic myeloid leukemia (CML) [37, 75]. Although these are difficult concepts to prove experimentally, much circumstantial evidence would appear to fit this model, as detailed below. Furthermore, given its widely reported association with advanced disease, critically short telomeres, and genetic instability (reviewed in [77]), telomerase may be relevant both as a prognostic or “staging” marker and as a therapeutic target, and many investigators continue to address these problems.

Telomerase Expression and Cell Cycle Status

Expression of significant levels of telomerase can dramatically increase proliferative life span and promote cellular immortality, thereby contributing to the malignant phenotype [80]. It is therefore important that its contribution to the burden of human cancer be understood. Initial reports of telomerase expression seemed to support a specific association with a malignant, immortal, or germline phenotype [31]. However, by using a more sensitive assay, telomerase has been detected in proliferating normal somatic tissues. For example, high levels of telomerase activity are detected in activated lymphoid cells [81] and proliferating endometrium [82], with low levels present in other somatic tissues. It would appear that such low-level expression represents rare telomerase-expressing stem or progenitor cells rather than uniform expression in all cells in the specimen [83, [84]–85], although there may be more widespread expression in proliferating endometrium [86]. Telomerase activity appears to be tightly associated with proliferation status, although how this is regulated remains unclear. One practical consequence is that many observations of telomerase modulators in vitro or of telomerase activity between different primary tissues (including tumors) may be confounded by altered or inherently different cell cycle activation status rather than by direct modulation of telomerase activity per se. In this regard, studies of telomerase regulation during cell cycle progression have produced conflicting results. Zhu et al. demonstrated upregulation of telomerase activity in human tumor cells during S-phase, using phase-selective cell cycle inhibitors [87]. However, a subsequent study, using selected cycling cell populations based on DNA content, demonstrated no change in telomerase activity during cycle progression [88]. When cells entered G0, however, telomerase was repressed and telomerase activity generally correlated with growth rate. By analogy, cells that are postmitotic should not express telomerase; rapid downregulation of activity has been described in several cell lines during terminal differentiation [89, 90], apparently mediated by a rapid reduction in hTERT mRNA levels occurring independently of simple cell cycle arrest and requiring de novo protein synthesis [91].

The implications of these observations are frequently overlooked. They would imply that a (if not the) primary determinant of TRAP activity in a tumor or tissue sample is the proportion of cycling cells (i.e., stem or progenitor cells, whether they are normal or malignant). Studies on hematological malignancy support this notion; for example, in early chronic lymphocytic leukemia (a disorder characterized by slow accumulation of mature B lymphocytes displaying resistance to apoptosis rather than increased turnover), TRAP activity is not raised until the disease transforms or accelerates (when cell turnover is increased) [92]. In keeping with this, our recent study (using purified primary CD34+BCR-ABL+ cells from chronic-phase CML at diagnosis) demonstrated increased cell cycle activation compared with controls, with a significant inverse correlation between the proportion of G0 cells and TRAP activity [79]. These data also explain a well-documented paradox, namely reduced telomere length in the face of increased telomerase activity; overall, an increased proportion of cycling cells would appear to elevate “whole tissue” TRAP levels; however, telomere maintenance in any individual cycling cell must remain suboptimal as shortening continues.

Telomerase and the Leukemic Stem Cell

Whether the mechanism of telomerase positivity in neoplasia is expansion of a pre-existing telomerase-competent clone (i.e., the LSC population) or upregulation within clonal, previously telomerase-negative cells (secondary to acquired genetic events) remains unclear. Given the increasing evidence that many hematopoietic malignancies arise in a stem or progenitor cell compartment, the former hypothesis appears the more attractive. There are few studies that have directly addressed telomerase biology at the LSC level; the majority have used unselected leukemia (or tumor) cell populations and therefore contain a heterogeneous population of malignant and benign cells at various stages of maturation. A case in point is CML, in which the progeny, as detected in unselected PBL and BM, are largely postmitotic differentiated cells in which telomerase is not expressed to any significant degree. In these circumstances, studies performed on sorted CD34+BCR-ABL+ populations are the more relevant [79]. We have demonstrated subtly, but significantly, elevated TRAP levels in this population; this elevation (as discussed above) is largely explicable by differing cycling populations. Intriguingly, however, expression of the major telomerase components was found to be dysregulated, with a fivefold reduction in levels of hTR in the clonal CD34+ cells [79]. Other groups have also found dysregulated expression of telomerase components [93] in this population. One may therefore surmise, at least in early-phase tumorigenesis, that a degree of telomerase dysfunction may accelerate telomere loss. In later tumor stages, when telomere shortening begins to exert selection pressure, telomerase-high (but telomere-short) progenitor populations may predominate. This has been documented during the progression of CML from chronic phase to blastic phase (M.W. Drummond, unpublished observations; [37] and reviewed in [77]). Similarly, in acute myeloid leukemia (AML), a hierarchy of LSC organization and differentiation has been described [94, 95], although studies on telomerase expression in this disorder have largely used downstream blast populations and reported widely varying degrees of telomerase expression. Such heterogeneity at the stem and progenitor level may go some considerable way in explaining the wide variation in telomerase expression, telomere length, and changes during disease progression that have been reported for AML [96, [97], [98], [99]–100], as discussed in more detail below.

Acute Myeloid Leukemia and Myelodysplasia

The most important prognostic factor in adult AML, as confirmed by the U.K. Medical Research Council (MRC) studies, are cytogenetic abnormalities at diagnosis [101]. However, considerable work has been undertaken to determine whether telomerase activity can further refine these prognostic data. Whether telomere shortening (with resulting genetic instability) might be involved in the evolution of these recurring genetic events is unknown but remains under consideration. The overall importance of telomerase in the pathogenesis of AML has recently been confirmed by the demonstration that hTERT is necessary for growth of primary AML cells in a mouse model [102]. A number of studies have investigated telomerase activity and telomere length in mononuclear cells (MNC) from patients with MDS and AML [76, 103, [104]–105, [106], [107]–108]. Telomere shortening was significantly more pronounced in patients with cytogenetic alterations as compared with patients with normal karyotypes [100]. In this study, the shortest median telomere length was found in the group with complex cytogenetic abnormalities. hTERT was overexpressed in patients with complex karyotypes, followed by patients with noncomplex karyotypes and patients without karyotypic changes [100]. This might suggest that with increasing telomere attrition, by either replication-dependent or replication-independent mechanisms, karyotypic abnormalities becomes more pronounced and, as a consequence, telomerase upregulation becomes essential to prevent replicative senescence of the malignant clone. However, it has recently been suggested that telomerase expression in the context of short telomeres does not necessarily prevent cells from reaching replicative exhaustion [109]. Whatever the role of hTERT in determining the prognosis in AML, it is unlikely to surpass that of cytogenetics in discriminating risk groups; recent work by our group on nearly 170 patients from the U.K. MRC AML12 trial has shown prognostic relevance of minor significance at best using quantitative polymerase chain reaction for hTERT (N.E. Jordanides, W.N. Keith, R.K. Hills, K. Wheatley, Q.T. Luong, A.K. Burnett, T.L. Holyoake, M.W. Drummond, unpublished observations). In multivariate analysis, only cytogenetics and white count at diagnosis remained significant.

Because of the uneven distribution of telomere length on individual chromosome arms [36], critical shortening of telomeres on particular chromosomes could promote the formation of chromosomal aberrations and contribute to clonal evolution. This hypothesis remains relevant even if the average telomere length remains well above the critical level of shortening [110]. Distinct groups of AML that are characterized either by aberrations that could result from telomere dysfunction (terminal deletions, gains/losses of chromosome parts, or nonreciprocal translocations) or by aberrations that are unlikely to result from telomere dysfunction (e.g., reciprocal translocations or inversions) could serve as an ideal model to study the effect of telomere shortening and telomerase activity during tumorigenesis [111].

Chronic Myeloid Leukemia

In terms of telomere and telomerase biology, CML is arguably the best characterized of human malignancies. A number of factors single it out as an ideal model for such studies. First, the malignant HSC (as defined by the presence of the BCR-ABL translocation) are characterized by increased cellular turnover (comprehensively reviewed in [112]). Second, the disease is characterized clinically by a relatively stable chronic phase (CP) that can last for many years. Third, progression to the accelerated phase or to blastic phase (BP) of the disease is associated with increased genetic instability and with the acquisition of additional cytogenetic abnormalities and mutations that are responsible for the altered and more aggressive growth pattern of the malignant clone (Fig. 2). To investigate the impact of telomere biology on disease progression in malignant HSC in this disease, we comparatively analyzed telomere length in BCR-ABL+ PBL and BCR-ABL (i.e., polyclonal) T lymphocytes, respectively. We found that telomeres in malignant cells were indeed significantly shorter than BCR-ABL T lymphocytes [75]. Successful therapy with the tyrosine kinase inhibitor imatinib mesylate was associated with an increase in mean telomere length as polyclonal hematopoiesis was restored [113, 114]. Furthermore, age-adjusted telomere shortening was found to be correlated with disease stage, remaining duration of CP before onset of BP [75], and Hasford risk score [37]. A number of studies have confirmed ongoing telomere shortening in the context of CML progression (Table 1); however, the role of telomerase is less certain. Our studies on CD34+-selected, BCR-ABL+ (>90%) primary LSC as described above suggest that any elevation in whole population telomerase activity (as measured by the TRAP assay) is due to an increased proportion of cycling progenitors. Indeed, at an individual LSC level, it is possible that the major telomerase components hTR and hTERT are dysregulated [79, 93]. It is therefore not surprising that telomere loss occurs; indeed, it may be accelerated to 10–20 times the normal rate, as measured in leukemic PBL [37]. Contributory factors, over and above replication and dysregulation of telomerase-induced attrition, include oxidative damage to telomere sequence with eventual loss. Interestingly, the BCR-ABL fusion protein significantly increased generation of reactive oxygen species in transfected cells [115] and is one potential mechanism linking ongoing genetic damage, clonal evolution, and telomere loss during CML progression from CP. Furthermore, oxidative stress would appear to result in translocation of endogenous hTERT out of the nucleus and into the cytosol, in a process that left overall TRAP activity unchanged [116]. It is therefore important to stress once again that TRAP activity is not synonymous with telomere maintenance.

Figure Figure 2..

Time course of chronic myeloid leukemia (CML) progression and relationship to clinical, morphologic, and cytogenetic features. Ionizing radiation remains the only clear etiological factor identified for CML. Based on look-back studies by the Hiroshima Foundation, the mean time to clinical development of CML after the atomic explosion (and therefore after acquisition of BCR-ABL in an irradiated stem cell) was 7 years [118]. A typical triphasic disease pattern is illustrated, with the median duration of each phase shown. A common feature of disease progression is the development of additional nonrandom cytogenetic abnormalities in approximately 70%–80% of patients, as illustrated.

Table Table 1.. Telomere shortening during progression of CML
original image

Conclusions and Future Directions

Simultaneous advances in stem cell and telomerase biology have given major insights into mechanisms that are central to both leukemogenesis and stem cell-based therapies and biology. A paradox is apparent in that both promotion and inhibition of telomerase activity seem attractive therapeutic techniques, albeit in very different circumstances. For example, in vitro optimization of telomere maintenance would be desirable during stem cell expansion protocols, whereas an in vivo approach might be a rational therapeutic intervention in DKC. More exciting is the prospect of telomerase-directed therapeutics in cancer. A variety of strategies are being pursued (with some in early phase clinical trials), including immunization with hTERT peptides and telomerase inhibition with modified oligonucleotides and hTERT-activated oncolytic viruses (recently reviewed in [117]). Such approaches may combine well with conventional leukemia chemotherapy, the latter to debulk the tumor and prevent early death and the former introduced to inhibit telomere maintenance (and tumor regrowth) in the chemoresistant LSC population. Much remains to be done, but hopefully a decade of impressive basic research will soon translate into therapeutic reality.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

Ancillary