Ectopic mTERT expression in mouse embryonic stem cells does not affect differentiation but confers resistance to differentiation- and stress-induced p53-dependent apoptosis

Telomerase is a ribonucleoprotein complex comprising the catalytic subunit (TERT) and an RNA component (TR) that adds TTAGGG repeats to chromosome ends, thus maintaining the length of telomeres (Collins and Mitchell, 2002). Telomerase is present in human stem cells and progenitors but is hardly detectable in the vast majority of differentiated somatic tissues (Urquidi et al., 2000). However, telomerase is reactivated during tumorigenesis, preventing telomere erosion and immortalization (Kim et al., 1994). By contrast, telomerase activity is more readily detectable in mouse cells and tissues, which generally contain longer telomeres than humans (Kipling and Cooke, 1990; Martin-Rivera et al., 1998; Greenberg et al., 1998). Moreover, murine embryonic stem (ES) cells have been shown to exhibit high telomerase activity (Sharma et al., 1995).

The potential role of telomerase in regulating cell growth has been evaluated by overexpression and gene-ablation experiments. Overexpression of human TERT (hTERT) in various cell lines resulted in resistance to several forms of apoptosis and in extended cellular life-span (Fu et al., 2000; Akiyama et al., 2002; Gorbunova et al., 2002; Luiten et al., 2003). By contrast, ablation of telomerase activity by antisense approaches or by expression of the dominant-negative hTERT resulted in enhanced sensitivity to cell death and consequently, growth inhibition (Fu et al., 1999; Misawa et al., 2002; Yuan and Mei, 2002; Nakajima et al., 2003; Teng et al., 2003; Liu et al., 2004). However, limited experiments have been performed with mouse TERT (mTERT). Overexpression of mTERT in transgenic mice resulted in susceptibility to several forms of tumour (Oh et al., 2001; Artandi et al., 2002; Gonzalez-Suarez et al., 2002). Gene knockout experiments revealed that the Terc^–/– and mTERT^–/– mice had no obvious abnormalities in early generations, although significant telomere shortening was observed in successive generations, resulting in increased apoptosis and malignancies arising from genomic instability (Rudolph et al., 1999; Yuan et al., 1999). It was also demonstrated that telomerase can add telomeric repeats to chromosome breaks upon irradiation, suggesting that telomerase might have additional roles beyond telomere maintenance (Slijepcevic and Bryant, 1998). Together, these data suggest a crucial role for TERT in regulating cellular growth, although the mechanistic basis for some of these observations is not completely understood.

Telomerase activity is regulated during differentiation, being reduced after ES-cell differentiation and during placental differentiation (Sharma et al., 1995; Armstrong et al., 2000; Rama et al., 2001). Accordingly, tumours arising out of progenitor cells lose their tumorigenicity upon differentiation, during which the telomerase activity is also dramatically reduced (Xu et al., 1999; Koyanagi et al., 2000). Although ectopic overexpression of hTERT in human mesenchymal cells and leukaemic cells was shown not to affect differentiation (Kobune et al., 2003; Yamada et al., 2003), terminal differentiation of human HaCaT cells was blocked by constitutive hTERT expression (Cerezo et al., 2003). These data raise the possibility that telomerase activity must be tightly regulated during cellular differentiation, although it appears to be cell-type specific. We have therefore evaluated whether ectopic expression of mTERT in pluripotent mouse ES cells would affect their differentiation capacity, as well as their response to differentiation and stress-induced apoptosis.

CCE and p53^–/– (P1.1) murine ES cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% serum, 100 units/ml penicillin, 100 μg/ml streptomycin, 0.1 mM non-essential amino acids, 2 mM glutamine, 1 mM sodium pyruvate, 50 μM β-mercaptoethanol and 2000 international units (IU) leukaemia inhibitory factor (LIF), as described (Sabapathy et al., 1997). Differentiation of ES cells was induced by adding all-trans-retinoic acid to culture medium without LIF. A NotI/HindIII fragment containing the coding sequence of mTERT (Martin-Rivera et al., 1998) was subcloned into pcDNA vector (Invitrogen). Catalytically inactive mTERT (also cloned into pcDNA vector) is a mutant form of mTERT in which amino acids 1186 and 1187 (both aspartic acid) were changed to alanine (F. Ishikawa, Department of Gene Mechanisms, Graduate School of Biostudies, Kyoto University, Japan). For electroporation, plasmids were linearized with ScaI and purified by phenol-chloroform extraction, precipitated and dissolved in sterile water. 30 μg linearized plasmid was used to transfect 1×10⁷ ES cells. For CCE cells, stable ES clones were selected and maintained in 250 μg ml^–1 G418, and single colonies were isolated and analysed. For p53^–/– ES cells, linearized plasmid was co-transfected with pLoxPuro (Arakawa et al., 2001) at a ratio of 5:1 and selected with 1 μg ml^–1 puromycin.

For the semiquantitative reverse-transcription polymerase chain reaction (RT-PCR), total RNA was isolated by TRIZOL reagent (Invitrogen) from ES cells. First-strand cDNA was synthesized by SuperScript II (Invitrogen) and PCR was carried out with Taq DNA polymerase (Qiagen). Primer sequences and PCR conditions used were as described previously by others (Sabapathy et al., 1997; Nozawa et al., 2001; Mitsui et al., 2003).

The telomeric repeats amplification protocol (TRAP) assay was performed using TRAPeze telomerase detection kit (Intergen). Briefly, approximately 1×10⁶ cells were lysed in CHAPS lysis buffer and the protein concentration was determined. Enzymatic reaction and PCR were performed according to the protocols provided by manufacturer. The PCR product was resolved by 10% polyacrylamide-gel electrophoresis (PAGE) in TBE buffer and visualized by staining with ethidium bromide.

Anti-ECMA-7 antibody was obtained from culture supernatants of ECMA-7 hybridomas (Sabapathy et al., 1997). For detection of ECMA-7 antigen, ES cells were harvested and washed with PBS once and then incubated with anti-ECMA-7 antibody for 30 minutes at room temperature. Cells were washed once with 0.5% Tween 20, 1% bovine serum albumen (BSA) in PBS and incubated with fluorescein isothiocyanate (FITC) conjugated anti-mouse IgG antibody for 30 minutes at room temperature. Cells were washed once and resuspended in PBS for flow cytometry.

Two independent cell-death assays were used in this study: propidium iodide (PI) exclusion and annexin-V staining. For PI exclusion, cells were harvested and washed once in PBS. Cell suspension was briefly incubated with 50 μg ml^–1 PI and immediately analysed by flow cytometry to detected PI-positive cells, which represent the dead population. To detect annexin-V-positive cells, ES cells were incubated with FITC-conjugated annexin-V (BD Biosciences) and 50 μg ml^–1 PI in binding buffer (10 mM HEPES, pH 7.4, 0.14 M NaCl, 2.5 mM CaCl₂) for 15 minutes in the dark at room temperature. Cells were analysed by flow cytometry immediately after incubation.

ES cells were harvested and washed with PBS once and then lysed in lysis buffer [50 mM Tris-HCl, pH 7.6, 50 mM NaCl, 0.1% sodium-dodecyl-sulfate (SDS), 1% Triton X-100 and protease inhibitors]. To detect mTERT, nuclear extract was used as described previously (Shats et al., 2003). Cells were briefly lysed with hypertonic buffer and nuclei were precipitated. 50 μg protein from whole-cell lysate or 100 μg nuclear extract was used in 10% SDS-PAGE and 7% SDS-PAGE, respectively, and then transferred to PVDF membrane and incubated with anti-p53 antibody (CM5; Novocastra), anti-telomerase antibody, anti-topoisomerase-I antibody (Santa Cruz Biotechnology) or anti-actin antibody (Sigma). Blots were incubated with ECL western-blotting detection reagent (Amersham) and chemiluminescence was detected with Biomax MR X-ray film (Kodak).

Cleaved caspase-3 was measured using an FITC-conjugated antibody against active caspase-3 apoptosis kit (BD Biosciences). Briefly, cells were harvested and washed with PBS and resuspended in Cytofix/Cytoperm™ solution for 20 minutes on ice. Cells were then incubated with antibody at room temperature for 30 minutes. After incubation, cells were washed with Perm/wash™ buffer once and analysed by flow cytometry.

ES cells were incubated with 10 μM bromodeoxyuridine (BrdU) for 30 minutes, washed with PBS once and fixed in 70% ethanol overnight. 2 N HCl/0.5% Triton X-100 solution was used to denature DNA then neutralized with 0.1 M Ba₂B₄O₇·10 H₂O, pH 8.5 solution. Cells were then incubated with FITC-conjugated anti-BrdU antibody (BD Biosciences) diluted in 0.5% Tween 20, 1% BSA in PBS. Cells were washed once with 0.5% Tween 20, 1% BSA in PBS, pelleted and resuspended in 50 μg ml^–1 PI in PBS for flow cytometry.

Feeder-independent ES cells (CCE cells) (Robertson et al., 1986) were used to generate mTERT-overexpressing cells. CCE cells electroporated with a mTERT cDNA expression vector or an empty vector (pcDNA) were selected on G418 and single colonies were isolated. Expression of mTERT was assayed by RT-PCR and colonies expressing relatively high levels of mTERT compared with pcDNA-transfected cells were selected (Fig. 1A). Western-blot analysis also revealed that mTERT overexpression resulted in increased mTERT protein level compared with vector-expressing cells (Fig. 1B). The telomerase enzymatic activity of the cells was determined by TRAP assay, which confirmed the increase in telomerase activity in mTERT-transfected clones (Fig. 1C). It should be noted that vector-transfected ES cells also show telomerase activity correlating with high basal mTERT protein production, confirming previous results describing the presence of relatively high telomerase activity in undifferentiated ES cells (Sharma et al., 1995). However, the mTERT-overexpressing clones displayed significantly more telomerase activity than their pcDNA-transfected counterparts. All subsequent experiments were performed using two independent mTERT-overexpressing ES-cell clones, which gave similar and consistent results, excluding any effects caused by integration sites variations.

Because telomerase activity is downregulated upon differentiation of ES cells (Sharma et al., 1995), we next evaluated whether enforced overexpression of mTERT would affect the differentiation potential of these cells. To this end, vector-transfected and mTERT-overexpressing ES-cell clones were induced to differentiate with 0.1 μM (data not shown) and 0.3 μM (Fig. 2A) retinoic acid. The progress of differentiation was monitored both by morphological appearance and by determining the amount of ECMA-7 stem-cell-specific antigen expression, which is specifically expressed in undifferentiated ES cells (Kemler, 1980). Retinoic-acid-induced differentiation caused similar changes in the morphological characteristics of both vector-transfected and mTERT-overexpressing cells, resulting in a flattened appearance after 6 days (Fig. 2A). Similar results were obtained with other methods of differentiation [i.e. dimethyl-sulfoxide (DMSO)-induced differentiation] (data not shown). RT-PCR and immunoblot analysis indicated that expression of exogenous mTERT was maintained in differentiated mTERT-expressing cells, excluding the possibility that the expression of exogenous mTERT was not shut down during differentiation (Fig. 2B,C). Becase there were no obvious defects in the differentiation process, we determined whether the rate of differentiation was affected in mTERT-overexpressing ES cells by following the loss of ECMA-7 expression over 6 days. Flow-cytometric analysis revealed that the pcDNA-transfected ES cells lost a significant amount of ECMA-7 expression 2 days after differentiation induction, and the cells did not express any ECMA-7 by 4 days after differentiation induction (Fig. 2D, left). mTERT-overexpressing cells also lost ECMA-7 expression at a similar rate (Fig. 2D, right), suggesting that the rate of differentiation was unaffected by mTERT overexpression.

We also evaluated the differentiation capacity of mTERT-overexpressing ES cells by analysing several molecular changes associated with differentiation. Expression of rex 1, a stem cell marker that has been shown to be downregulated during differentiation (Rogers et al., 1991), was reduced in vector-transfected and mTERT-overexpressing ES cells upon differentiation (Fig. 2E). Moreover, expression of α foetal protein (afp), gata 4, goosecoid and pax 6, which are differentiation markers of primitive endoderm, mesoderm and neuroectoderm, respectively (Blum et al., 1992; Arceci et al., 1993; Abe et al., 1996; Renoncourt et al., 1998), were upregulated to equal extents in vector-transfected and mTERT-expressing ES cells upon differentiation (Fig. 2E). Taken together, these data suggest that mTERT overexpression does not affect ES-cell differentiation.

While monitoring alterations in differentiation rates, we noticed that colonies of mTERT-overexpressing cells were much larger than those of the vector-transfected cells (data not shown). We therefore determined whether mTERT overexpression would alter the growth properties of undifferentiated ES cells. Analysis of cumulative cell numbers over several days indicated that mTERT-overexpressing cells had a growth advantage, which was obvious after 3 days in culture (Fig. 3A). In order to elucidate the cause of the observed differences, we investigated whether there were differences in the proliferation or basal cell death rates of these cells. Asynchronous cultures of cells were labelled with BrdU to monitor proliferating cells and analysed by flow cytometry after staining with anti-BrdU antibody. The amounts of BrdU-labelled cells in both the vector-transfected and the mTERT-expressing populations were not significantly different (65.68±0.71% and 66.07±2.32%, respectively) (Fig. 3B).

We therefore determined whether there were differences in apoptosis. Analysis of basal cell death rates revealed that there was a small but statistically significant difference between the two populations. The basal level of cell death in vector-transfected cells was 3.19±0.2%, whereas the levels in the mTERT-overexpressing population was 2.11±0.24%, indicating that mTERT overexpression conferred resistance to apoptosis (P<0.05 by an unpaired Student's t test) (Fig. 3C). Because this effect was small, we further investigated whether the cell-death rates were altered upon differentiation, during which differentiation-associated apoptosis ensues (Castro-Obregon and Covarrubias, 1996). Both vector-transfected and mTERT-overexpressing ES cells were differentiated and the cumulative cell-death rate was monitored over the first 3 days of retinoic-acid-induced differentiation by the PI exclusion method. Differentiation-induced cell death was markedly reduced in the mTERT-overexpressing cells over the course of differentiation (P<0.05 by an unpaired Student's t-test) (Fig. 3D). These findings together suggest that, although ectopic mTERT expression does not alter the proliferation potential of ES cells, it does affect the basal cell-death rates both in the undifferentiated state and during differentiation, thereby conferring a growth advantage on mTERT-overexpressing ES cells.

Overexpression of hTERT protects several human cell lines against stress-induced cell death (Lu et al., 2001; Gorbunova et al., 2002). We found that mTERT overexpression also conferred a survival advantage upon retinoic-acid-induced differentiation, during which oxidative radicals are produced (Castro-Obregon and Covarrubias, 1996), and so we explored the possibility that these cells are also resistant to other forms of cellular insults, including oxidative stress. Hydrogen peroxide (H₂O₂) was used as a source of oxidative stress, and cell death was determined upon treatment with various doses of H₂O₂. mTERT-overexpressing cells were significantly resistant to H₂O₂-mediated cell death compared with vector-transfected cells over all doses of H₂O₂ tested (Fig. 4A). This suggests that mTERT can protect cells from oxidative stress and probably explains the reduced cell-death rates observed during retinoic-acid-induced differentiation (Fig. 3D).

In addition to oxidative stress, we investigated whether mTERT overexpression could also protect ES cells against other forms of genotoxic insult. To this end, cell death after treatment with various doses of chemotherapeutic agents such as cisplatin (Fig. 4B) and doxorubicin (Fig. 4C) was determined. Like H₂O₂, overexpression of mTERT conferred a significant survival advantage against both cisplatin and doxorubicin, compared with vector-transfected cells (Fig. 4B,C).

We extended the analysis on cell death by analysing the proportion of annexin-V staining by flow cytometry, which identify cells undergoing apoptosis (compared to necrosis), upon genotoxic insults. Doxorubicin (0.6 μg ml^–1) and H₂O₂ (1 mM) treatment resulted in accumulation of annexin-V-positive pcDNA-transfected cells (control vs doxorubicin and H₂O₂: 4% vs 44.5% and 16.4%, respectively) (Fig. 5, left). However, the mTERT-overexpressing cells were markedly resistant to apoptosis induced by both agents (control vs doxorubicin and H₂O₂: 3.4% vs 17.6% and 6.5%, respectively) (Fig. 5, right). A similar trend in resistance to apoptosis was observed on treatment of cells with various doses of stress agents (data not shown). Together, these findings suggest that mTERT overexpression in ES cells confers resistance to apoptosis.

It has been shown that hTERT protects breast-cancer cells from apoptosis independently of its catalytic activity (Cao et al., 2002). However, the enzymatic activity was required to protect HeLa cells from apoptosis (Zhang et al., 2003), suggesting that dependence on the telomerase enzymatic activity to suppress apoptosis might be cell-type specific. Therefore, we further investigated whether the catalytic activity of mTERT is required to protect ES cells against apoptosis. We generated stable ES cells overproducing the catalytically inactive mTERT, in which the aspartic residues at positions 1186 and 1187 were changed to alanines (F. Ishikawa, personal communication). The expression of the inactive mTERT was confirmed by RT-PCR analysis (Fig. 6A). Despite increased mTERT production, these ES cells did not show an increase in the telomerase enzyme activity compared with vector-transfected ES cells (Fig. 6B), confirming that the exogenous mTERT is indeed catalytically inactive. We then tested whether overproduction of catalytically inactive mTERT can protect ES cells from genotoxic insults. Cell death of inactive-mTERT-producing cells was comparable to that of vector-transfected cells (vector-transfected vs inactive-mTERT-producing ES cells: 37.1±0.57% vs 49.9±0.57% at 0.5 μg ml^–1 doxorubicin and 68.8±3.54% vs 69.65±0.78% at 1 μg ml^–1 doxorubicin) (Fig. 6C), whereas production of catalytically active mTERT drastically reduced cell death after doxorubicin treatment (Fig. 6C). Similar results were obtained with annexin-V staining (data not shown), confirming that production of inactive mTERT cannot protect ES cells from apoptosis and indicate that the enzyme activity of mTERT is required for its anti-apoptotic properties in ES cells.

ES cells express high levels of p53 that are reduced upon differentiation and undergo p53-dependent cell death during differentiation (Sabapathy et al., 1997). Moreover, all the genotoxic drugs used in the study induce apoptosis in a p53-dependent manner (de Vries et al., 2002; Siddik, 2003). Hence, we investigated whether mTERT-mediated resistance to cell death occurs by inhibiting p53-dependent cell death. We noticed that endogenous p53 was upregulated by all genotoxic insults to equal extents in vector and mTERT-overexpressing cells (Fig. 7A), suggesting that the induction of p53 is not affected by mTERT overexpression. Consistently, treatment with these genotoxic agents resulted in the upregulation of several p53-dependent apoptotic target genes, such as noxa and puma, in both vector and mTERT-overexpressing ES cells (Fig. 7B). There was no difference in the levels of bax expression during apoptosis in ES cells (Fig. 7B). The upregulation of p53 and its target genes suggested that p53 function is not compromised by mTERT overexpression. However, cleavage of caspase-3, the hallmark of apoptotic execution and a downstream event subsequent to p53 activation via the mitochondrial pathway (Schuler and Green, 2001; Wang, 2001), was compromised in mTERT-overexpressing cells compared with vector-transfected cells upon doxorubicin treatment (vector-transfected vs mTERT-overexpressing cells: 46.19% vs 16.28%) (Fig. 7C). Similar results were obtained with cisplatin treatment (data not shown). These findings suggested that telomerase might act downstream of or independent of p53 to inhibit apoptosis.

We therefore investigated whether mTERT overexpression could affect p53-mediated apoptosis. To this end, we assayed cell death in the absence or presence of pifithrin-α (PFTα), a chemical inhibitor of p53 (Komarov et al., 1999) (Fig. 8A). Cells were incubated with PFTα and subsequently treated with cisplatin to induce cell death. The proportion of cisplatin-induced cell death in vector-transfected ES cells incubated with PFTα was reduced and was strikingly similar to control (DMSO incubated) mTERT-overexpressing ES cells (PFTα-treated vector cells vs DMSO-treated mTERT-overexpressing cells: 13.14±0.67% vs 13.23± 0.12% at 5 μM cisplatin and 18.55±0.28% vs 18.17±0.19% at 10 μM cisplatin) (Fig. 8A), indicating that mTERT overexpression was as effective as inhibiting p53. In addition, incubation of mTERT-overexpressing ES cells with PFTα before cisplatin treatment only slightly reduced the apoptotic rates (DMSO- vs PFTα-treated: 13.23±0.23% vs 10.22±0.21% at 5 μM cisplatin and 18.17±0.19% vs 12.65±1.28% at 10 μM cisplatin) (Fig. 8A). These results suggest that the anti-apoptotic effect of mTERT overexpression probably occurs through the inhibition of the p53-dependent cell death pathway.

Recently, it was shown that PFTα also suppress cellular functions other than the inhibition of p53 (Komarova et al., 2003; Rocha et al., 2003). Therefore, we generated mTERT-overexpressing p53^–/– ES cells to confirm the role of p53 in mTERT-mediated anti-apoptotic effects. RT-PCR analysis confirmed that the p53^–/– (P1.1) ES cells do not express p53 (Fig. 8B) and that mTERT-overexpressing p53^–/– ES cells expressed significantly higher levels of mTERT mRNA (Fig. 8C). We then tested whether the exogenous mTERT expression conferred anti-apoptotic effects on p53^–/– ES cells. Similar to PFTα treatment, cell death of mTERT-overexpressing CCE ES cells (p53^+/+) was only slightly higher than that of vector-transfected p53^–/– ES cells, confirming that mTERT overexpression was almost as effective as knocking out p53 in terms of anti-apoptotic capacity, especially at lower doses of doxorubicin (mTERT-overexpressing CCE ES cells vs vector-transfected p53^–/– ES cells: 27.55±0.97% vs 18.76±1.35% at 1 μg ml^–1 doxorubicin and 39.75±1.35% vs 20.84±3.03% at 2 μg ml^–1 doxorubicin) (Fig. 8D). Moreover, production of exogenous mTERT or inactive mTERT in p53^–/– ES cells did not confer further resistance to doxorubicin treatment (vector-transfected vs mTERT-producing vs inactive-mTERT-producing p53^–/– ES cells: 18.76±1.35% vs 25.06±3.44% vs 24.72±3.12% at 1 μg ml^–1 doxorubicin and 20.84±3.03% vs 25.11±7.89% vs 20.22±8.36% at 2 μg ml^–1 doxorubicin, respectively) (Fig. 8D). Together, the data suggest that the survival advantage conferred by exogenous mTERT expression is through the suppression of p53-dependent apoptosis.

The results presented here indicate that ectopic overexpression of mTERT does not affect the differentiation capacity or the rate of differentiation of mouse ES cells. However, constitutive mTERT expression confers a survival advantage, especially during differentiation and stress-induced apoptosis. Furthermore, the mTERT-mediated anti-apoptotic effect appears to be dependent on its catalytic activity and is due to the inhibition of the p53-dependent apoptotic pathway.

Our data indicate that enforced mTERT expression does not affect the initial differentiation process, consistent with recent findings in human mesenchymal stem cells and leukaemic stem cells, in which differentiation was not affected by exogenous telomerase (Kobune et al., 2003; Yamada et al., 2003). These data suggest that the decrease in telomerase activity observed during differentiation in several cell systems is probably a consequence of the differentiation process, during which the pluripotence of the stem cells is lost. It is thus conceivable that differentiated cells do not require to maintain telomerase activity because their telomeres need not be protected any further upon loss of `stem-ness' and, hence, differentiation results in a reduction in telomerase activity. However, we cannot exclude the possibility that terminal differentiation, leading to quiescence of cells, could be affected by the overexpression of TERT.

Oxidative-stress-associated cell death occurs during ES-cell differentiation by retinoic acid (Castro-Obregon and Covarrubias, 1996). Consistently, we found that mTERT-expressing ES cells were more resistant to differentiation and oxidative-stress-induced cell death. Moreover, exogenous telomerase expression also conferred resistance to other genotoxic-stress-induced cell death in these cells. These findings are consistent with other studies showing that suppression of telomerase results in cell death in neurons (Zhu et al., 2000), PC12 cells (Fu et al., 1999), breast-cancer cells (Cao et al., 2002) and cervical cancer cells (Kushner et al., 2000), whereas ectopic telomerase expression protected PC12 cells (Lu et al., 2001) and fibroblasts (Gorbunova et al., 2002) from cell death. However, the mechanism by which telomerase protects cells against cell death remains controversial. Cao et al. (Cao et al., 2002) demonstrated that hTERT regulates cell survival independent of its enzymatic activity in breast cancer cells, whereas Zhang et al. (Zhang et al., 2003) showed that the catalytic activity and the binding to 14-3-3, a downstream target of p53, are essential for its anti-apoptotic function in HeLa cells. Our results also indicate that the catalytic activity of mTERT is required for its anti-apoptotic function, suggesting that the mechanisms of TERT-mediated anti-apoptosis could be different in different cell types.

ES cells are atypical in that they express high levels of p53, which is crucial for cell death during differentiation or caused by genotoxic insults (Sabapathy et al., 1997). Our data indicate that, although mTERT protects ES cells from apoptosis, its production did not affect p53 levels and the induction of p53-dependent apoptotic target genes. However, chemical inhibition of p53 function by PFTα did not significantly affect the apoptotic rates in mTERT-overexpressing cells. Moreover, mTERT overexpression did not confer a survival advantage on p53^–/– ES cells, suggesting that the mTERT-mediated protective effect might come about by antagonizing the p53 pathway, probably through targets downstream of p53. Consistently, Wiman and colleagues have recently found that hTERT antagonizes p53-mediated apoptosis (Rahman et al., 2004). Taken together, these data indicate that telomerase expression confers resistance to apoptosis through inhibition of the p53 pathway.

This study also highlights the differences between cell types with respect to the ability of telomerase to regulate cellular proliferation. Exogenous telomerase was shown to increase proliferation in somatic lens epithelial cells (Xiang et al., 2002), whereas mTERT overexpression did not affect ES-cell proliferation. This might be due to the differences in the regulation of cell-cycle progression between these cell types. hTERT overexpression was shown to increase cellular proliferation rates by regulating phosphorylation of the retinoblastoma protein (Rb) (Xiang et al., 2002). Unlike somatic cells, cell-cycle progression in ES cells is regulated by total Rb protein levels rather than the phosphorylation status of Rb (Savatier et al., 1994), which probably accounts for the lack of a proliferation advantage of mTERT-overexpressing ES cells.

Human cancers, many of which are derived from progenitor populations, exhibit increased telomerase activity and are often associated with poor prognosis (Kim et al., 1994; Ohali et al., 2003; Ulaner et al., 2003). Moreover, inhibition of telomerase activity often results in sensitivity to drug-induced apoptosis (Kondo et al., 1998; Misawa et al., 2002; Yuan and Mei, 2002). The data presented here demonstrate that resistance to drug-induced cell death is a direct consequence of increased telomerase activity, suggesting that reactivation of telomerase during tumorigenesis might contribute directly to survival advantage and chemotherapeutic drug resistance. Therefore, it is conceivable that therapeutic drugs that have the ability to reduce telomerase activity would be of clinical relevance in TERT-overproducing tumours. In this respect, histone-deacetylase inhibitors, which have been shown to decrease cellular telomerase activity (Suenaga et al., 2002), have also been demonstrated to have antitumour properties (Camphausen et al., 2004; Kuefer et al., 2004). Thus, future therapeutic strategies can be aimed at reducing the telomerase activity to combat cancers.

This study highlights the anti-apoptotic role of mTERT in ES cells through a mechanism involving the inhibition of p53-dependent apoptosis. In addition, telomerase downregulation appears to be a consequential event during cellular differentiation, because loss of telomerase activity in knockout mice and mTERT overexpression in ES cells appear not to affect the differentiation process. Thus, these findings can be extended to stem-cell therapies, because the overexpression of TERT has no negative effects on cellular differentiation and should be useful for increasing the viability and replicative potential of stem cells.

We thank L. Harrington (University of Toronto, Canada), F. Ishikawa (Kyoto University, Japan) and H. Awakawa (Heinrich-Pette-Institute, Germany) for their generous sharing of mTERT cDNA, catalytically inactive mTERT cDNA and pLoxPuro plasmid, respectively. This work was supported by grants from the National Medical Research Council and the Biomedical Research Council of Singapore to K.S.