Subtelomeric DNA hypomethylation is not required for telomeric sister chromatid exchanges in ALT cells

Article metrics

Abstract

Most human tumor cells acquire immortality by activating the expression of telomerase, a ribonucleoprotein that maintains stable telomere lengths at chromosome ends throughout cell divisions. Other tumors use an alternative mechanism of telomere lengthening (ALT), characterized by high frequencies of telomeric sister chromatid exchanges (T-SCEs). Mechanisms of ALT activation are still poorly understood, but recent studies suggest that DNA hypomethylation of chromosome ends might contribute to the process by facilitating T-SCEs. Here, we show that ALT/T-SCEhigh tumor cells display low DNA-methylation levels at the D4Z4 and DNF92 subtelomeric sequences. Surprisingly, however, the same sequences retained high methylation levels in ALT/T-SCEhigh SV40-immortalized fibroblasts. Moreover, T-SCE rates were efficiently reduced by ectopic expression of active telomerase in ALT tumor cells, even though subtelomeric sequences remained hypomethylated. We also show that hypomethylation of subtelomeric sequences in ALT tumor cells is correlated with genome-wide hypomethylation of Alu repeats and pericentromeric Sat2 DNA sequences. Overall, this study suggests that, although subtelomeric DNA hypomethylation is often coincident with the ALT process in human tumor cells, it is not required for T-SCE.

Introduction

Telomeres are specialized structures that cap the ends of eukaryotic chromosomes. In human somatic cells, telomeres consist of TTAGGG repeat sequences extending over 5–15 kb (Blackburn, 2001). A progressive shortening of telomeres at each normal somatic cell division is known as the ‘end replication problem’ (Watson, 1972) and results in an irreversible arrest termed replicative senescence (Hayflick and Moorhead, 1961). This progressive shortening is compensated by the expression of telomerase, a ribonucleoprotein complex absent in most normal human somatic cells but activated in the majority of tumor-derived cell lines (Kim et al., 1994; Shay and Bacchetti, 1997). A subset of tumors and about 50% of SV40-immortalized cell lines maintain their telomeres by one or more telomerase-independent mechanisms referred to as alternative lengthening of telomeres (ALT) (Bryan et al., 1997). ALT pathways involve homologous recombination to replicate telomeric DNA (Dunham et al., 2000; Bailey et al., 2004). Analysis of post-replicative telomere exchanges using the chromosome orientation fluorescence in situ hybridization (CO-FISH) method showed a tight correlation between telomeric sister chromatid exchanges (T-SCEs) and ALT pathways (Bechter et al., 2004; Londono-Vallejo et al., 2004).

Adjacent to telomeres, subtelomeric regions consist of duplicated DNA tracts extending up to 500 kb from the chromosome end (Riethman et al., 2005). Recent studies have shown that mammalian telomeres and subtelomeres display epigenetic features characteristic of heterochromatin, and that alterations of these epigenetic marks are associated with telomere deregulation and changes in telomere dynamics (Blasco, 2007; Michishita et al., 2008). Notably, a lack of DNA methyltransferases (DNMTs) in mouse embryonic stem cells alters the DNA methylation status of subtelomeric regions, leading to telomere elongation and increased T-SCE frequency (Gonzalo et al., 2006). Similar observations were made for chromosome ends of mouse cells lacking histone methyltransferases (Benetti et al., 2007b). In turn, progressive telomere shortening in Terc-/- mice has been shown to affect the epigenetic status of chromosome ends, resulting in decreased subtelomeric DNA methylation accompanied by an increased frequency of telomeric recombination, providing a possible mechanism for the activation of ALT in telomerase-deficient cells (Benetti et al., 2007a). On the other hand, hypomethylation of subtelomeric DNA in cells isolated from patients with mutations in the DNMT3b gene was associated with abnormally short telomeres and an increased level of transcripts emanating from telomeric regions, but no increase of T-SCE (Yehezkel et al., 2008).

As heterochromatin features may potentially impact on the rate of telomeric recombination, we aimed to investigate the relationship between subtelomeric DNA methylation level and T-SCE. To this end, we compared the methylation levels of D4Z4 and DNF92 subtelomeric sequences, located on the distal portions of 4qter, 10qter, 1pter, 5qter, 6qter, 8pter and 17qter genomic loci, with the T-SCE frequencies in human cancer cell lines and in vitro immortalized fibroblasts relying on either ALT- or telomerase-dependent pathway. Next, we analysed the DNA methylation status of D4Z4 subtelomeric regions in ALT cancer cells, in which T-SCE frequencies were reduced following a forced expression of telomerase. Finally, we analysed the expression level of DNMTs and the extent of genome-wide DNA methylation in ALT and in telomerase-positive cell lines.

Results

Hypomethylation of subtelomeric DNA in ALT cancer cell lines

Given recent reports indicating that a defect in DNA methylation of mouse chromosome ends may facilitate T-SCE (Gonzalo et al., 2006), a hallmark of ALT cells, we analysed the methylation status of subtelomeric DNA isolated from 11 human cancer cell lines relying on either telomerase or ALT pathways for telomere maintenance.

The telomere maintenance mechanism of these tumor cell lines was first characterized by the measurement of hTERT/hTR telomerase subunit expression and telomere length (Figures 1a and b). The ALT phenotype of LB188 cell line, established from a rhabdomyosarcoma tumor (F Brasseur, unpublished data), was further confirmed by the absence of telomerase activity assessed by the telomeric repeat amplification protocol (Supplementary Figure 1). For the evaluation of subtelomeric DNA methylation levels, we chose to analyse the methylation status of the non-satellite D4Z4 subtelomeric repeats located at 4q35.2 and 10q26.3, as (1) the nucleotide sequence of the corresponding 3.3-kb D4Z4 repeats has been determined (Cacurri et al., 1998, 2) this region is known to be subjected to methylation on cytosine residues in normal tissues (van Overveld et al., 2003) and (3) the last D4Z4 repeat is mapped 23–24 kb from the 4q or 10q telomere in the current reference genome build (36.3) and further analyses indicated that D4Z4 is the last repetitive element in 4q and 10q (Riethman, 2008). Primers were designed to specifically amplify a 380-bp region of the D4Z4 repeat unit derived from 4q35.2 and 10q26.3 and containing 31 CpG dinucleotides. As expected from earlier studies (de Lange et al., 1990; Brock et al., 1999), chromosome ends were heavily methylated in normal human dermal fibroblasts (HFF2), with D4Z4 methylation levels amounting to 77% (Figure 2a), a value that agrees with the earlier D4Z4 CpG DNA methylation of 72–79% measured in normal brain samples (Cadieux et al., 2006). In comparison, the mean methylation level of D4Z4 repeats in telomerase-positive cell lines was 56% (Figures 2a and c). In agreement with the proposed role of subtelomeric DNA hypomethylation in the ALT process, the subtelomeric D4Z4 DNA in the three ALT cancer cell lines was largely demethylated, as only 23, 37 and 28% of CpG were methylated in U2OS and SaOS-2 osteosarcoma and LB188 rhabdomyosarcoma cell lines, respectively. These values contrast with the 77% CpG methylation detected in LB23 telomerase-positive rhabdomyosarcoma cells (Figure 2c). To check whether the decreased D4Z4 methylation levels that we observed in ALT cell lines reflect mislocalization of these motifs within chromosomes, we hybridized metaphase spreads from U2OS and LB188 cell lines with a fluorescent probe corresponding to 4q35, in which most D4Z4 motifs are located. As expected, FISH signals were exclusively detected at chromosome ends in both HFF2 normal fibroblasts and ALT cell lines (Figure 2b). For a better overview of subtelomeric loci, we analysed the methylation level of a 553-bp fragment of DNF92 minisatellite. The DNF92 sequence (accession number Y13543) not only represents the most distal segment of the proximal subtelomeric region found mainly at chromosomes 1 (1pter), 5 (5qter), 6 (6qter), 8 (8pter) and 17 (17qter) but also is detected at a lower frequency at eight other chromosome ends (Monfouilloux et al., 1998; Der-Sarkissian et al., 2002). This segment is separated from the telomeric tract by the distal subtelomeric region, which spans around 10 kb (Flint et al., 1997; Monfouilloux et al., 1998). As expected, normal fibroblasts displayed a high level of DNF92 methylation (80%) (Figures 2a and c). In tumor cell lines, data were consistent with the analysis of D4Z4 methylation except for two telomerase-positive cell lines, MG63 and LB37, which showed a reduced level of D4Z4 methylation but DNF92 methylation levels amounting to 81 and 66%, respectively (Figures 2a and c). Overall, the mean methylation level of DNF92 sequences in telomerase-positive cell lines was 62%. In agreement with D4Z4 analysis, ALT tumor cell lines showed low levels of DNF92 methylation, ranging from 27 to 45%.

Figure 1
figure1

Characterization of telomeres and telomerase expression in tumor-derived and in vitro immortalized cell lines. (a) Terminal restriction fragment (TRF) analysis of telomeres from cancer cell lines (MZ2, LB45, LB373 and LB929 melanoma; LB23 and LB188 rhabdomyosarcoma; HeLa cervix cancer; LB37 NSCLC and U2OS, SaOS-2 and MG63 osteosarcoma-derived cell lines), SV40-immortalized fibroblasts (VA13, SW39, IMRB and IMRG) and HFF2 normal human foreskin fibroblasts. (b) Expression of telomerase subunit genes. hTERT, hTR and ACTB cDNA levels were measured by qRT–PCR and expressed as (hTERT or hTR cDNA/ACTB cDNA) × 104. NSCLC, non-small-cell lung cancer; qRT–PCR, real-time quantitative PCR.

Figure 2
figure2

Subtelomeric DNA methylation in tumor-derived and in vitro immortalized cell lines. (a) Sequencing of PCR-amplified D4Z4 and DNF92 fragments from bisulfite-treated genomic DNA. Each line corresponds to the CpG residues of an individual clone of the D4Z4 or DNF92 PCR. Upto 10 representative sequences are shown for each cell line. White and gray boxes represent unmethylated and methylated cytosines, respectively. Black boxes indicate a TpA dinucleotide in the bisulfite-treated DNA, presumably due to the spontaneous deamination of methylated CpGs on the opposite DNA strand (Cadieux et al., 2006). The total number of CpG analysed for either D4Z4 or DNF92 is given for each cell line under ‘n’. (b) Fluorescent in situ hybridization (FISH) with subtelomeric 4q35 probe on metaphase spreads of normal human foreskin fibroblasts as control (HFF2) and ALT cell lines. Arrowheads indicate the FISH signals on chromosome ends. (c) Quantification of D4Z4 (gray bars) and DNF92 (black bars) methylation levels calculated from (a). A χ2-test of association was applied to compare D4Z4 or DNF92 DNA methylation in ALT and TEL+ tumor-derived cell lines. ALT, alternative lengthening of telomeres.

The above data indicate that subtelomeric DNA methylation is reduced in ALT, but not in most telomerase-positive cancer cells.

No subtelomeric hypomethylation in ALT SV40-immortalized fibroblasts

To further address the relationship between subtelomeric DNA methylation and ALT pathways, we analysed the methylation of D4Z4 and DNF92 DNA in ALT (VA13, IMRB and IMRG) and telomerase-positive (SW39) SV40-immortalized fibroblasts (Figures 2a and c). Surprisingly, subtelomeric DNA methylation levels in the ALT cell lines were high and similar to those in telomerase-positive SW39 fibroblasts. FISH experiments with the 4q35 fluorescent probe confirmed subtelomeric location of the locus in ALT (VA13 and IMRB)-immortalized fibroblasts (Figure 2b). These data suggest that ALT mechanisms can be activated in the absence of subtelomeric DNA hypomethylation.

ALT cancer cell lines and ALT SV40-immortalized fibroblasts display similar T-SCE frequencies

As D4Z4 and DNF92 methylation analyses revealed considerable discrepancies between ALT cells derived from tumors and in vitro immortalized cells, we quantified T-SCE events in these cell lines to assess whether the distinct subtelomeric DNA methylation profiles may be correlated with differences in T-SCE frequencies (Figure 3). To this end, we used CO-FISH to detect SCE events at telomeres (Bechter et al., 2004; Londono-Vallejo et al., 2004). We observed similar levels of T-SCE in VA13 ALT fibroblasts (an average of 3.5 exchanges/100 chromosome extremities) and U2OS or LB188 ALT cancer cell lines (3.3 and 3.7%, respectively) (Figures 3a and b). T-SCE events were more frequent in IMRB ALT fibroblasts (6.4%) and SaOS-2 ALT osteosarcoma cell line (5.6%). As expected, T-SCE frequencies measured in LB23 rhabdomyosarcoma, MG63 osteosarcoma and MZ2 melanoma telomerase-positive cancer cells were much lower (0.3, 0.4 and 0.07%, respectively) (Figure 3b). The above data show that a reduced level of subtelomeric DNA methylation is not required for telomeric DNA exchanges in ALT cells.

Figure 3
figure3

Detection of telomeric sister chromatid exchanges in tumor-derived and SV40-immortalized cell lines. (a) T-SCEs detected by CO-FISH. Representative images are shown for U2OS and LB188 ALT cancer cell lines, LB23 and MZ2 telomerase-positive cancer cell lines and IMRB and VA13 ALT SV40-immortalized fibroblasts. T-SCEs, detected by the presence of Telo-FISH signals (red) on both sister chromatids of ALT chromosomes, are surrounded by blue circles. Only very rare metaphases display a double signal in telomerase-positive cells (not shown). (b) Quantification of T-SCE frequencies in cell lines from (a), SaOS-2 and MG63 tumor cell lines. Results are given as T-SCE events for 100 chromosome extremities (n=total number of chromosome extremities analysed for each cell line). Error bars represent standard deviations from three (IMRB) and two (U2OS and MZ2) independent CO-FISH experiments. Note that the T-SCE frequencies obtained here for U2OS cells are about fivefold higher than those reported earlier (Londono-Vallejo et al., 2004), thanks to the improvement of CO-FISH experimental conditions. ALT, alternative lengthening of telomeres; CO-FISH, chromosome orientation fluorescence in situ hybridization; T-SCE, telomeric sister chromatid exchange.

Ectopic expression of telomerase in ALT cancer cells reduces T-SCE but does not restore subtelomeric DNA methylation

We further assessed the relationship between subtelomeric DNA methylation and T-SCE in ALT cancer cells by manipulating T-SCE rates in ALT tumor cells and testing potential concurrent changes in subtelomeric DNA methylation levels. To this end, we ectopically expressed both subunits of telomerase in U2OS osteosarcoma ALT cells. Earlier studies reported distinct outcomes after telomerase introduction into ALT SV40T-immortalized fibroblasts: either ALT was preserved and T-SCEs were unaffected (Cerone et al., 2001) or ALT was abolished and T-SCE frequency was reduced (Ford et al., 2001; Bechter et al., 2004). These apparent discrepancies may possibly be due to distinct telomerase levels in the cells (Henson et al., 2002).

After transfection with the hTR subunit and selection for stable integrants, U2OS/hTR cells were transfected with an hTERT construct. The expression level of both telomerase subunits was very high in the resulting stable integrants (Figure 4a). U2OS/hTR/hTERT cells, referred to as U2OS Telo, were cultured for more than 265 population doublings. Ectopically expressed telomerase was detected in U2OS Telo cell extracts (Figure 4b), and was also shown to act at the endogenous telomeres of these cells, as revealed by the increased frequency of chromosome ends with telomeric DNA in later passages of U2OS Telo cells (Figure 4c). Notably, although 19.2% of chromosome extremities did not show any FISH signal in control U2OS cells, this number was reduced to 6.9% in U2OS Telo cells after 218 population doublings, a value close to the 4.1% of undetectable telomeres in telomerase-positive LB23 cells using these experimental settings (Figure 4c). In agreement with an earlier study (Bechter et al., 2004), the active telomerase reduced T-SCE frequency by more than fourfold in U2OS Telo cells (Figure 4d). Importantly, D4Z4 sequences remained mostly hypomethylated in U2OS Telo cells (Figure 4e), although we cannot exclude the possibility that a small fraction of subtelomeres were remethylated in these cells. These data indicate that T-SCE rates are not directly linked to overall subtelomeric DNA methylation levels.

Figure 4
figure4

Ectopic expression of telomerase in U2OS ALT cells. (a) Expression levels of hTERT catalytic subunit and hTR RNA subunit of telomerase measured by qRT–PCR in U2OS cell line and U2OS cells transfected with telomerase (U2OS Telo). Values are given as (hTERT cDNA/ACTB cDNA) × 10 000 and (hTR cDNA/ACTB cDNA) × 1000. U2OS cells were first transfected with pBabe::U1-hTR construct and selected for stable integration. Population doubling (PD) counting of U2OS Telo cells started after transfection with pBMN::hTERT construct. Measurements were performed in triplicate. Error bars represent standard deviations. (b) Telomerase activity in U2OS Telo cells at PD141 detected by the telomerase repeat amplification protocol (TRAP) assay. NTC is a negative control with lysis buffer alone and IC is an internal control of PCR. (c) Telomeric repeats detected by Telo-FISH with a Cy3-labeled LNA telomeric probe on metaphase spreads from U2OS and U2OS Telo cells at PD94 and PD218. The graph gives the percentage of chromosome ends with undetectable FISH signal. Measurements in ALT IMRB SV40-immortalized fibroblasts and telomerase-positive LB23 and SW39 cells are given for comparison. χ2-tests of association were applied to compare telomerization levels. (d) Quantification of T-SCE on metaphase spreads from U2OS and U2OS Telo cells at PD265 by the CO-FISH technique. Results are given as T-SCE events for 100 chromosome extremities. Error bars give standard deviations for two independent experiments. A χ2-test of association was applied to compare T-SCE frequencies. (e) Analysis of D4Z4 DNA methylation in U2OS and U2OS Telo cell lines by bisulfite sequencing. See legend of Figure 2a for details. ALT, alternative lengthening of telomeres; CO-FISH, chromosome orientation fluorescence in situ hybridization; qRT–PCR, real-time quantitative PCR; T-SCE, telomeric sister chromatid exchange.

Hypomethylation of subtelomeric DNA is correlated with a reduced genome-wide methylation in ALT cancer cell lines

We next searched to determine the mechanisms underlying subtelomeric DNA demethylation in ALT tumor cells. Tumor development is often accompanied by a process of genome-wide DNA demethylation (Ehrlich, 2002). As a result, tumors show varying extents of DNA hypomethylation within repeat elements, pericentromeric sequences and single-copy genes (Ehrlich, 2006). Here, we evaluated if this genome-wide DNA demethylation process also underlies hypomethylation of D4Z4 and DNF92 subtelomeric DNA in ALT tumor cells. To this end, we measured the methylation level of Alu-interspersed repetitive elements and satellite 2 pericentromeric DNA at chromosomes 1 and 16 in different tumor cell lines by the quantitative MethylLight assay (Weisenberger et al., 2005). Combined MethylLight-based measurement of Sat2 and Alu methylation levels was reported earlier to be highly correlative with global DNA methylation measurements performed by high-performance liquid chromatography (Weisenberger et al., 2005).

As expected, our data revealed a decrease of pericentromeric Sat2 and Alu DNA methylation in cancer cell lines compared with normal dermal fibroblasts (Figure 5b). Interestingly, significantly lower values of Sat2/Alu methylation were measured in ALT when compared with telomerase-positive cancer cell lines (P=0.004, Student's t test) (Figure 5b). A very good correlation was found between the level of Sat2/Alu methylation and the average methylation level of D4Z4 and DNF92 subtelomeric sequences (Figure 5d). Profound genomic DNA demethylation in ALT cancer cells was further confirmed by the expression of the cancer germline gene MAGE-A1 (Figure 5c), the expression of which was correlated earlier with global genomic DNA demethylation in tumors (De Smet et al., 1996; Cadieux et al., 2006). Altogether, these data suggest that hypomethylation of subtelomeric sequences in ALT tumor cells results from the genome-wide DNA demethylation process often associated with tumorigenesis.

Figure 5
figure5

Analysis of genome-wide methylation in cancer cell lines using either ALT mechanisms or telomerase activity. (a) Quantification of subtelomeric DNA methylation ((%CpG D4Z4 +%CpG DNF92)/2) in ALT cancer cell lines (U2OS, LB188 and SaOS-2) and telomerase-positive cancer cell lines (MG63, LB37, MZ2, LB45, LB373, LB929, LB23 and HeLa) calculated from Figure 2a data. The HFF2 normal human foreskin fibroblasts are shown as reference. Values are given as the percentage of methylated CpG dinucleotides. A χ2-test of association was applied to compare subtelomeric DNA methylation in ALT and TEL+ tumor cell lines. (b) Quantification of Alu/Sat2 methylation levels by MethylLight analysis on bisulfite-converted DNA. Measurement of methylation-independent Alu-C4 DNA was used as a control of input DNA (Weisenberger et al., 2005). ((Sat2-M1+Alu-M2)/2 × Alu-C4) ratios are calculated for each sample and compared with the ratio obtained for in vitro M.SssI-methylated BJhTERT genomic DNA. (c) MAGE-A1 gene expression measured by qRT–PCR. Values were normalized with β-actin expression and are given as % of MAGE-A1 expression in LB188 cells. All measurements were performed in triplicate and error bars represent standard deviations. (d) Correlation between global genomic and subtelomeric DNA methylation in tumor-derived cell lines. Linear regression was applied to test the linear association between Alu/Sat2 methylation levels from (b) and subtelomeric DNA methylation from (a). Normal HFF2 fibroblasts are also shown (dotted circle). ALT, alternative lengthening of telomeres; qRT–PCR, real-time quantitative PCR.

Reduced levels of DNMTs could possibly account for the steady-state DNA hypomethylation of ALT tumor cell lines, both at the global and at the subtelomeric level. To test this hypothesis, we used real-time quantitative PCR to quantify the expression of DNMT1, -3a and -3b genes encoding the enzymes that catalyse the transfer of methyl groups to cytosine in mammals (Figures 6a-c). Overall, the expression of the DNMTs was highly variable in all cell lines tested and DNMT1/3b expression levels were increased in cell lines compared with normal dermal fibroblasts, in agreement with earlier observations (Siedlecki and Zielenkiewicz, 2006). However, a reduced expression of DNMT genes in ALT tumor cell lines was not generally observed and cannot therefore account for DNA hypomethylation.

Figure 6
figure6

Quantification of DNMT1, -3a and -3b expression in tumor-derived and in vitro immortalized cell lines. cDNA levels of DNMT1 (a), DNMT3a (b) and DNMT3b (c) were measured by qRT–PCR. Values were normalized by the number of ACTB cDNA molecules. The graphs give the relative expression of DNMTs compared with normal foreskin fibroblasts HFF2. Measurements were performed in triplicate and error bars give the standard deviations. ALT, alternative lengthening of telomeres; DNMTs, DNA methyltransferases; qRT–PCR, real-time quantitative PCR.

Discussion

Recent studies reported evidences indicating that chromatin modifications of chromosome ends, through the action of histone and DNMTs, regulate mammalian telomeres (Blasco, 2007). In particular, an association between hypomethylation of subtelomeric DNA and increased T-SCE frequency (Gonzalo et al., 2006; Benetti et al., 2007a), a hallmark of ALT cells (Londono-Vallejo et al., 2004), was observed in murine cells. The role of DNA methylation in telomere exchanges is, however, still controversial, as mutations in human DNMT3b affect telomere length but not T-SCE (Yehezkel et al., 2008). Here, we analysed the methylation status of two types of subtelomeric DNA sequences, D4Z4 and DNF92, in telomerase-positive and ALT tumor cell lines. Altogether, these two sequences cover at least seven human subtelomeres. Our results indicate that the methylation of these subtelomeric regions is significantly reduced in ALT tumor cell lines compared with telomerase-positive cell lines that show varying subtelomeric methylation levels, irrespective of their origin (melanoma, sarcoma, non-small-cell lung cancer and cervix cancer) and their telomere length. This observation is in agreement with the hypothesis that lower subtelomeric DNA methylation levels may favor ALT activation (Blasco, 2007). Surprisingly, however, we found no evidence of such subtelomeric DNA hypomethylation in ALT SV40-immortalized fibroblasts, even though these cells exhibited high T-SCE frequencies. This indicates that subtelomeric DNA hypomethylation is not necessary for ALT-associated T-SCE. Moreover, when ALT tumor cells were forced to express active telomerase, we observed a fourfold reduction in T-SCE frequency, despite persistent subtelomeric DNA hypomethylation. This indicates that the level of T-SCE is not strictly dependent on the overall methylation level of subtelomeric sequences.

Considering the established association between the presence of telomerase at chromosome ends and DNA methylation at subtelomeric repeats, it was surprising to find no increase in D4Z4 methylation in the telomerase-expressing ALT tumor cells. One possible explanation could be that telomerase did not completely revert the ALT pathway and/or the associated telomeric heterochromatinization, as suggested earlier (Perrem et al., 2001; Cerone et al., 2005). However, the telomeric repeat addition at short chromosome ends that we observed upon telomerase expression suggests that telomerase acted at the telomeres of these ALT cells. Another explanation could be that, whereas telomerase efficiently contributes to the maintenance of pre-established methylation marks within subtelomeric regions (Benetti et al., 2007a), it is not able to induce de novo methylation of previously unmethylated sequences.

What is the origin of subtelomeric DNA hypomethylation in ALT tumor cells? Our results indicate that this may be a consequence of the extensive genomic DNA demethylation process that usually occurs during tumorigenesis (Ehrlich, 2002). This is evidenced by the low methylation level of pericentromeric Sat2 DNA and Alu repeats, two markers of genome-wide demethylation in tumors (Weisenberger et al., 2005; Cadieux et al., 2006; Ehrlich, 2006), in ALT tumor cell lines. It is worth noting that when this article was in preparation, a study on epigenetic regulation of human telomeres reported no correlation between the methylation level of subtelomeric DNA and the pericentromeric repeat NBL2 DNA (Vera et al., 2008). We believe that the discrepancy between this study and the present one may be related to the fact that NBL2-repeated sequences are known to undergo opposite cancer-linked epigenetic changes, displaying either hypomethylation or, more frequently, predominant hypermethylation in tumors (Ehrlich, 2006). This is not the case for cancer-linked hypomethylation of pericentromeric satellite 2 DNA, which is significantly associated with global cancer-linked DNA hypomethylation (Weisenberger et al., 2005; Ehrlich, 2006). Deep genomic DNA demethylation in ALT tumor cells was further shown in our study by the expression of the cancer germline gene MAGE-A1, which was correlated earlier with genome-wide hypomethylation in tumors (De Smet et al., 1996; Cadieux et al., 2006).

The reason why genome-wide DNA hypomethylation is prevalent in ALT tumor cells is an open question. We showed that this cannot be explained by lower levels of DNMT1/3a/3b expression in ALT tumor cells. Alternatively, DNA hypomethylation in ALT cells could be related to their origin as most ALT cell lines derive from sarcomas. However, the LB23 and MG63 telomerase-positive sarcoma cell lines included in this study do not show any sign of deep genome-wide demethylation. On the basis of the assumption that genome-wide DNA demethylation occurs gradually during tumorigenesis and terminates when cells acquire immortalization (Wilson and Jones, 1983; Young et al., 2003), we propose instead the following hypothesis: extensive genome-wide DNA hypomethylation in ALT tumor cells may possibly reflect the difficulty to set up alternative mechanisms of telomere maintenance, which would generally occur later than telomerase gene activation, when DNA demethylation has spread to most parts of the genome, including subtelomeric regions (Figure 7). This model is compatible with the established impact of telomerase on subtelomeric DNA methylation and on overall chromatin maintenance (Masutomi et al., 2005; Blasco, 2007), as it is conceivable that the lack of telomerase in developing ALT tumors facilitates the spreading of the DNA demethylation process to most parts of the genome.

Figure 7
figure7

Model to account for the hypomethylated subtelomeric DNA in ALT cancer cells. During tumorigenesis, cells are subjected to genome-wide demethylation (Ehrlich, 2002). On the basis of the assumption that DNA demethylation occurs gradually during tumorigenesis and terminates when cells acquire immortalization (Wilson and Jones, 1983; Young et al., 2003), we propose that extensive genome-wide DNA hypomethylation in ALT tumor cells may reflect the difficulty to set up ALT, which would occur later than telomerase gene activation, when DNA methylation has spread to most parts of the genome, including subtelomeric regions. Profound demethylation of subtelomeric DNA is, however, not restricted to ALT tumors and was also detected in a minority of telomerase-positive cell lines. ALT, alternative lengthening of telomeres.

The lack of subtelomeric DNA hypomethylation in ALT SV40-immortalized fibroblasts may be due to the fact that SV40-immortalized cells do not undergo the DNA demethylation process that characterizes most tumor cells. However, the low level of Sat2 DNA methylation measured in both ALT and TEL+ SV40-immortalized fibroblasts (Supplementary Figure 2) does not support this hypothesis. Another hypothesis would be that the ALT process is acquired earlier during in vitro immortalization, thereby preventing DNA hypomethylation of subtelomeric regions. Accordingly, Sat2 DNA methylation was higher in ALT-immortalized fibroblasts compared with ALT tumor-derived cell lines (Supplementary Figure 2).

If subtelomeric DNA hypomethylation does not seem to be required for telomere exchanges in ALT cells, it remains to be seen whether other epigenetic alterations, such as histone modifications, may affect T-SCE in these cells. Understanding how epigenetic alterations occurring during tumorigenesis affect telomere maintenance of cancer cells is undoubtedly a major future research goal.

Materials and methods

Cell lines and cell culture

HFF2 normal human foreskin fibroblasts were purchased from ATCC (Rockville, MD, USA). WI38 VA13/2RA (here referred to as VA13) and IMR-90-SV40 (Coriell, Camden, NJ, USA) are ALT SV40-immortalized fetal lung fibroblasts; SW39 is a telomerase-positive SV40-transformed fetal lung fibroblast cell line kindly provided by W Wright (UT Southwestern Medical Center, Dallas). IMRB and IMRG clones have been isolated from the IMR-90-SV40 population. U2OS, SaOS-2 and MG63 osteosarcoma cell lines were kindly provided by F Fuks, F Journe and H Id Boufker (ULB, Belgium). HeLa cells are derived from cervix cancer. Other cancer cell lines were derived from melanoma (LB929, MZ2, LB45 and LB373), non-small-cell lung cancer (LB37) and rhabdomyosarcoma (LB188 and LB23) tumors obtained from patients undergoing surgery and were kindly provided by M Swinarska and F Brasseur (Ludwig Institute for Cancer Research, Brussels). Normal and immortalized fibroblasts were cultured in minimum essential medium (Gibco, Invitrogen, Carlsbad, CA, USA) supplemented with 1% non-essential amino acids (Gibco, Invitrogen) in humidified incubators (37 °C) with 5% CO2. Cancer cell lines were grown at 37 °C in humidified incubators with 8% CO2 in either Iscove's medium or Dulbecco's modified Eagle's medium. All media were supplemented with 10% fetal bovine serum (HyClone, Perbio Science, Logan, UT, USA).

Plasmids and transfection

The pBMN::hTERT-puro plasmid, encoding the catalytic subunit of telomerase, was kindly provided by C Heirman (VUB, Belgium). The pBABE::U1-hTR-hygro plasmid, with the hTR RNA subunit of telomerase, was constructed by inserting the 1.2-kb EcoRI/SalI fragment isolated from pBS::U1-hTR plasmid (kindly provided by J Lingner, ISREC, Switzerland) into EcoRI/SalI-cleaved pBABE-hygro vector. A total of 5 × 105 U2OS cells were transfected with 10 μg pBMN::hTERT-puro or 10 μg pBABE::U1-hTR-hygro using calcium phosphate precipitation method as described earlier (Traversari et al., 1992). Stable transfectants were selected with 200 μg/ml hygromycin (Sigma, St Louis, MO, USA) or 0.8 μg/ml puromycin (InvivoGen, San Diego, CA, USA).

Terminal restriction fragment analysis

Telomere length was measured using the Telo TAGGG Telomere Length Assay kit from Roche Applied Science (Mannheim, Germany).

Telo-FISH and CO-FISH

Preparation of metaphase spreads and hybridization with 100 nM of Cy3-labeled (GGGTTA)6 LNA probe (Exiqon, Vedbaek, Denmark) were performed as described earlier (Londono-Vallejo et al., 2004). For CO-FISH, cells were first grown in the presence of 30 μM bromodeoxyuridine (Sigma) for 24 h prior colcemid addition. Degradation of newly synthesized strands was realized as described earlier (Cornforth and Eberle, 2001) with the following adaptations: ultraviolet irradiation was performed at 365 nm and chromosomes were incubated with Exonuclease III (New England Biolabs, Ipswich, MA, USA) for 10 min.

4q35 subtelomere FISH

TelVysion 4q DNA probe (Vysis; Abbott Molecular, Louvain-la-Neuve, Belgium) was designed to hybridize to a locus located within 300 kb of human 4q chromosome end. Hybridization to metaphase spreads and the rapid wash procedure were performed according to the manufacturer's instructions.

Telomerase repeat amplification protocol assay

Telomerase repeat amplification protocol assay was performed as described earlier (Kim et al., 1994; Kim and Wu, 1997). Briefly, telomerase-synthesized telomeric repeats onto TS oligonucleotide (5′-IndexTermAATCCGTCGAGCAGAGTT) and elongation products were amplified with primers ACX (5′-IndexTermGCGCGG[CTTACC]3CTAACC) and TS. An internal control TSNT (5′-IndexTermAATCCGTCGAGCAGAGTTAAAAGGCCGAGAAGCGAT) was amplified by primers TS and NT (5′-IndexTermATCGCTTCTCGGCCTTTT). Conditions for telomere elongation, PCR amplification and detection of telomeric products were based on instructions from TRAPeze Telomerase Detection Kit (Chemicon, Temecula, CA, USA).

cDNA preparation and real-time quantitative PCR

Total RNA was extracted from cell lines using TriPure Isolation Reagent (Roche Applied Science). Reverse transcription was performed on 2 μg of total RNA using random hexamers in the presence of 200 U M-MLV reverse transcriptase (Invitrogen) for 1 h at 42 °C in a final volume of 20 μl. Reaction volume was then adjusted to 100 μl with water. Gene expression levels were measured by real-time quantitative PCR TaqMan technology as described earlier (Tilman et al., 2007). Primers and probes for quantification of β-actin (ACTB), hTERT, hTR, DNMT1, DNMT3a, DNMT3b and MAGE-A1 were purchased from Eurogentec (Seraing, Belgium) and were described earlier (Nakamura et al., 1998; Shimojima et al., 2004; Loriot et al., 2006; Tilman et al., 2007; Kholmanskikh et al., 2008).

Bisulfite genomic DNA sequencing

Genomic DNA was extracted with the Invisorb Spin Tissue Mini kit (Invitek, Westburg, Leusden, The Netherlands). Sodium bisulfite treatment of 20–30 μg of genomic DNA was performed as described earlier (De Smet et al., 2004). After purification, bisulfite-modified DNA was resuspended into 100 μl of water. For D4Z4 and DNF92 methylation analysis, semi-nested PCR amplifications were realized using 2 μl of bisulfite-modified DNA for the first amplification (final volume of 20 μl, primers 5′-IndexTermGTATATTTTTAGGTTGAGTTTTGTAA (FE-D4Z4) and 5′-IndexTermAATATACCAAACCCTCTCTC (RE-D4Z4) or 5′-IndexTermGTGGAGGTAGTATTAGGTAGA (FE-DNF92) and 5′-IndexTermCCCTACTTTCTCAAAATACTCCTA (RI-DNF92)) and 5 μl of a 1/300 dilution of the first PCR product for the second amplification (final volume of 50 μl, primers 5′-IndexTermGTGTGGTTTTGTTTTTTGTAAGG (FI-D4Z4) and RE-D4Z4 or 5′-IndexTermGGGGTTGAGGTATATTAGTATTAA (FI-DNF92) and RI-DNF92). PCR conditions were as follows: 2 min at 94 °C followed by 50 s at 94 °C, 50 s at 55 °C and 2 min at 72 °C for 35 cycles and 3 min at 72 °C. Two microliters of the final PCR product were cloned into pJET 1.2/blunt vector using CloneJET PCR Cloning kit (Fermentas, St Leon-Rot, Germany) and the ligation mixture was directly used for bacterial electroporation. Plasmid inserts were sequenced using the BigDye Terminator kit (PE Applied Biosystems, Foster City, CA, USA).

MethylLight reactions

MethylLight reactions for the quantification of methylated Alu and Sat2 sequences were performed as described earlier (Weisenberger et al., 2005) using the Alu-M2 and Sat2-M1 primer sets, respectively. Values were corrected for the amount of input sodium bisulfite-treated DNA, which was determined by quantitative PCR amplification of CpG-depleted Alu sequences using the Alu-C4 primer set (Weisenberger et al., 2005). Alu-M2/Alu-C4 and Sat2-M1/Alu-C4 ratios were expressed as the percentage of the ratio obtained for genomic DNA extracted from BJhTERT human dermal fibroblasts (kindly provided by F d'Adda di Fagagna) and methylated in vitro with bacterial SssI methyltransferase (NEB, Ipswich, MA, USA) before bisulfite treatment as described earlier (Loriot et al., 2006).

Validation of the MethylLight assay was performed by two ways. First, we measured Sat2 and Alu methylation levels in in vitro methylated genomic DNA and DNA isolated from cells treated with the DNA-demethylating agent 5-aza-2′-deoxycytidine (Supplementary Figure 3a and b). Second, we correlated MethylLight-based analysis with M.SssI/dam 3H-methyl incorporation ratio measurement for quantification of overall methyl CpG content in MZ2, LB188, LB23 and LB45 tumor cell lines and normal fibroblasts (De Smet et al., 1996) (Supplementary Figure 3c).

Statistical analysis

χ2-tests were applied to evaluate the statistical significance of the observed differences in chromosome end telomerization (Figure 4c); T-SCE frequency (Figure 4d) and D4Z4/DNF92 DNA methylation (Figures 2c and 5a). Linear regression analysis was applied to test the linear correlation between (1) Sat2/Alu methylation levels and subtelomeric DNA methylation (D4Z4/DNF92) (Figure 5) and (2) SssI/dam 3H-methyl incorporation ratio and Sat2/Alu methylation levels (Supplementary Figure 3c). Statistical analyses were performed on http://faculty.vassar.edu/lowry/VassarStats.html.

Accession codes

Accessions

GenBank/EMBL/DDBJ

References

  1. Bailey SM, Brenneman MA, Goodwin EH . (2004). Frequent recombination in telomeric DNA may extend the proliferative life of telomerase-negative cells. Nucleic Acids Res 32: 3743–3751.

  2. Bechter OE, Shay JW, Wright WE . (2004). The frequency of homologous recombination in human ALT cells. Cell Cycle 3: 547–549.

  3. Benetti R, Garcia-Cao M, Blasco MA . (2007a). Telomere length regulates the epigenetic status of mammalian telomeres and subtelomeres. Nat Genet 39: 243–250.

  4. Benetti R, Gonzalo S, Jaco I, Schotta G, Klatt P, Jenuwein T et al. (2007b). Suv4-20h deficiency results in telomere elongation and derepression of telomere recombination. J Cell Biol 178: 925–936.

  5. Blackburn EH . (2001). Switching and signaling at the telomere. Cell 106: 661–673.

  6. Blasco MA . (2007). The epigenetic regulation of mammalian telomeres. Nat Rev Genet 8: 299–309.

  7. Brock GJ, Charlton J, Bird A . (1999). Densely methylated sequences that are preferentially localized at telomere-proximal regions of human chromosomes. Gene 240: 269–277.

  8. Bryan TM, Englezou A, Dalla-Pozza L, Dunham MA, Reddel RR . (1997). Evidence for an alternative mechanism for maintaining telomere length in human tumors and tumor-derived cell lines. Nat Med 3: 1271–1274.

  9. Cacurri S, Piazzo N, Deidda G, Vigneti E, Galluzzi G, Colantoni L et al. (1998). Sequence homology between 4qter and 10qter loci facilitates the instability of subtelomeric KpnI repeat units implicated in facioscapulohumeral muscular dystrophy. Am J Hum Genet 63: 181–190.

  10. Cadieux B, Ching TT, VandenBerg SR, Costello JF . (2006). Genome-wide hypomethylation in human glioblastomas associated with specific copy number alteration, methylenetetrahydrofolate reductase allele status, and increased proliferation. Cancer Res 66: 8469–8476.

  11. Cerone MA, Autexier C, Londono-Vallejo JA, Bacchetti S . (2005). A human cell line that maintains telomeres in the absence of telomerase and of key markers of ALT. Oncogene 24: 7893–7901.

  12. Cerone MA, Londono-Vallejo JA, Bacchetti S . (2001). Telomere maintenance by telomerase and by recombination can coexist in human cells. Hum Mol Genet 10: 1945–1952.

  13. Cornforth MN, Eberle RL . (2001). Termini of human chromosomes display elevated rates of mitotic recombination. Mutagenesis 16: 85–89.

  14. de Lange T, Shiue L, Myers RM, Cox DR, Naylor SL, Killery AM et al. (1990). Structure and variability of human chromosome ends. Mol Cell Biol 10: 518–527.

  15. De Smet C, De Backer O, Faraoni I, Lurquin C, Brasseur F, Boon T . (1996). The activation of human gene MAGE-1 in tumor cells is correlated with genome-wide demethylation. Proc Natl Acad Sci USA 93: 7149–7153.

  16. De Smet C, Loriot A, Boon T . (2004). Promoter-dependent mechanism leading to selective hypomethylation within the 5' region of gene MAGE-A1 in tumor cells. Mol Cell Biol 24: 4781–4790.

  17. Der-Sarkissian H, Vergnaud G, Borde YM, Thomas G, Londono-Vallejo JA . (2002). Segmental polymorphisms in the proterminal regions of a subset of human chromosomes. Genome Res 12: 1673–1678.

  18. Dunham MA, Neumann AA, Fasching CL, Reddel RR . (2000). Telomere maintenance by recombination in human cells. Nat Genet 26: 447–450.

  19. Ehrlich M . (2002). DNA methylation in cancer: too much, but also too little. Oncogene 21: 5400–5413.

  20. Ehrlich M . (2006). Cancer-linked DNA hypomethylation and its relationship to hypermethylation. Curr Top Microbiol Immunol 310: 251–274.

  21. Flint J, Bates GP, Clark K, Dorman A, Willingham D, Roe BA et al. (1997). Sequence comparison of human and yeast telomeres identifies structurally distinct subtelomeric domains. Hum Mol Genet 6: 1305–1313.

  22. Ford LP, Zou Y, Pongracz K, Gryaznov SM, Shay JW, Wright WE . (2001). Telomerase can inhibit the recombination-based pathway of telomere maintenance in human cells. J Biol Chem 276: 32198–32203.

  23. Gonzalo S, Jaco I, Fraga MF, Chen T, Li E, Esteller M et al. (2006). DNA methyltransferases control telomere length and telomere recombination in mammalian cells. Nat Cell Biol 8: 416–424.

  24. Hayflick L, Moorhead PS . (1961). The serial cultivation of human diploid cell strains. Exp Cell Res 25: 585–621.

  25. Henson JD, Neumann AA, Yeager TR, Reddel RR . (2002). Alternative lengthening of telomeres in mammalian cells. Oncogene 21: 598–610.

  26. Kholmanskikh O, Loriot A, Brasseur F, De Plaen E, De Smet C . (2008). Expression of BORIS in melanoma: lack of association with MAGE-A1 activation. Int J Cancer 122: 777–784.

  27. Kim NW, Piatyszek MA, Prowse KR, Harley CB, West MD, Ho PL et al. (1994). Specific association of human telomerase activity with immortal cells and cancer. Science 266: 2011–2015.

  28. Kim NW, Wu F . (1997). Advances in quantification and characterization of telomerase activity by the telomeric repeat amplification protocol (TRAP). Nucleic Acids Res 25: 2595–2597.

  29. Londono-Vallejo JA, Der-Sarkissian H, Cazes L, Bacchetti S, Reddel RR . (2004). Alternative lengthening of telomeres is characterized by high rates of telomeric exchange. Cancer Res 64: 2324–2327.

  30. Loriot A, De Plaen E, Boon T, De Smet C . (2006). Transient down-regulation of DNMT1 methyltransferase leads to activation and stable hypomethylation of MAGE-A1 in melanoma cells. J Biol Chem 281: 10118–10126.

  31. Masutomi K, Possemato R, Wong JM, Currier JL, Tothova Z, Manola JB et al. (2005). The telomerase reverse transcriptase regulates chromatin state and DNA damage responses. Proc Natl Acad Sci USA 102: 8222–8227.

  32. Michishita E, McCord RA, Berber E, Kioi M, Padilla-Nash H, Damian M et al. (2008). SIRT6 is a histone H3 lysine 9 deacetylase that modulates telomeric chromatin. Nature 452: 492–496.

  33. Monfouilloux S, Avet-Loiseau H, Amarger V, Balazs I, Pourcel C, Vergnaud G . (1998). Recent human-specific spreading of a subtelomeric domain. Genomics 51: 165–176.

  34. Nakamura TM, Cooper JP, Cech TR . (1998). Two modes of survival of fission yeast without telomerase. Science 282: 493–496.

  35. Perrem K, Colgin LM, Neumann AA, Yeager TR, Reddel RR . (2001). Coexistence of alternative lengthening of telomeres and telomerase in hTERT-transfected GM847 cells. Mol Cell Biol 21: 3862–3875.

  36. Riethman H . (2008). Human telomere structure and biology. Annu Rev Genomics Hum Genet 9: 1–19.

  37. Riethman H, Ambrosini A, Paul S . (2005). Human subtelomere structure and variation. Chromosome Res 13: 505–515.

  38. Shay JW, Bacchetti S . (1997). A survey of telomerase activity in human cancer. Eur J Cancer 33: 787–791.

  39. Shimojima M, Komine F, Hisatomi H, Shimizu T, Moriyama M, Arakawa Y . (2004). Detection of telomerase activity, telomerase RNA component, and telomerase reverse transcriptase in human hepatocellular carcinoma. Hepatol Res 29: 31–38.

  40. Siedlecki P, Zielenkiewicz P . (2006). Mammalian DNA methyltransferases. Acta Biochim Pol 53: 245–256.

  41. Tilman G, Mattiussi M, Brasseur F, van Baren N, Decottignies A . (2007). Human periostin gene expression in normal tissues, tumors and melanoma: evidences for periostin production by both stromal and melanoma cells. Mol Cancer 6: 80.

  42. Traversari C, van der Bruggen P, Van den Eynde B, Hainaut P, Lemoine C, Ohta N et al. (1992). Transfection and expression of a gene coding for a human melanoma antigen recognized by autologous cytolytic T lymphocytes. Immunogenetics 35: 145–152.

  43. van Overveld PG, Lemmers RJ, Sandkuijl LA, Enthoven L, Winokur ST, Bakels F et al. (2003). Hypomethylation of D4Z4 in 4q-linked and non-4q-linked facioscapulohumeral muscular dystrophy. Nat Genet 35: 315–317.

  44. Vera E, Canela A, Fraga MF, Esteller M, Blasco MA . (2008). Epigenetic regulation of telomeres in human cancer. Oncogene 27: 6817–6833.

  45. Watson JD . (1972). Origin of concatemeric T7 DNA. Nat New Biol 239: 197–201.

  46. Weisenberger DJ, Campan M, Long TI, Kim M, Woods C, Fiala E et al. (2005). Analysis of repetitive element DNA methylation by MethyLight. Nucleic Acids Res 33: 6823–6836.

  47. Wilson VL, Jones PA . (1983). DNA methylation decreases in aging but not in immortal cells. Science 220: 1055–1057.

  48. Yehezkel S, Segev Y, Viegas-Pequignot E, Skorecki K, Selig S . (2008). Hypomethylation of subtelomeric regions in ICF syndrome is associated with abnormally short telomeres and enhanced transcription from telomeric regions. Hum Mol Genet 17: 2776–2789.

  49. Young JI, Sedivy JM, Smith JR . (2003). Telomerase expression in normal human fibroblasts stabilizes DNA 5-methylcytosine transferase I. J Biol Chem 278: 19904–19908.

Download references

Acknowledgements

We are grateful to M Swinarska, F Brasseur, W Wright, F Fuks, J Lingner, F d'Adda di Fagagna, F Journe, H Id Boufker and C Heirman for the generous gifts of cell lines and plasmids. We thank all the members of the GENEPI group for their constant support and help. This study was supported by the Fonds National de la Recherche Scientifique (FNRS), Belgium. GT is supported by a PhD fellowship grant from Télévie/FNRS. AL is supported by a post-doctoral grant from the FNRS. Work in the ‘Telomere&Cancer’ laboratory is supported by grants from ANR, ARC, INCa and La Ligue.

Author information

Correspondence to A Decottignies.

Additional information

Conflict of interest

The authors declare no conflict of interest.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Tilman, G., Loriot, A., Van Beneden, A. et al. Subtelomeric DNA hypomethylation is not required for telomeric sister chromatid exchanges in ALT cells. Oncogene 28, 1682–1693 (2009) doi:10.1038/onc.2009.23

Download citation

Keywords

  • telomeres
  • DNA methylation
  • T-SCE
  • ALT
  • subtelomeric D4Z4 and DNF92

Further reading