DNA Damage Induces Alternative Lengthening of Telomeres (ALT)–Associated Promyelocytic Leukemia Bodies that Preferentially Associate with Linear Telomeric DNA

Introduction

Some tumors (10%) and in vitro immortalized cell lines maintain their telomeres using the recombination-mediated alternative lengthening of telomeres (ALT) mechanism ( 1). Human ALT cell lines have a number of characteristics in common. In addition to the lack of significant levels of telomerase activity, ALT cells have telomeres that are highly heterogeneous in length, ranging from very long to very short ( 1). This heterogeneity is generated by a combination of gradual telomere attrition and rapid lengthening or shortening events ( 2). These data suggest that ALT telomeres are maintained by a recombination-mediated lengthening mechanism, which was confirmed by the observation of movement of a telomere-targeted tag onto other telomeres in an ALT cell line ( 3). In contrast to telomerase-positive cells, which contain only linear extrachromosomal telomere repeats (ECTR), ALT cells contain both linear and circular ECTR ( 4, 5). ALT cells also contain unique nuclear structures called ALT-associated promyelocytic leukemia (PML) bodies (APB). These structures are a subset of PML bodies that contain telomeric DNA and telomere-associated proteins ( 6). Within an asynchronously dividing population of ALT cells, only 5% to 10% of the cells are APB-positive ( 6, 7). In addition to telomere associated components, APBs also contain proteins involved in recombination and repair (reviewed in ref. 8). The exact role of APBs in the ALT mechanism, however, is still unknown.

In the following study, we found that treatment with DNA-damaging agents caused an increase in the number of cells containing APBs. In addition to APBs, H₂O₂ and N-methyl-N′-nitro-N-nitrosoguanidine (MNNG) caused an ALT-specific induction of linear and circular ECTR. We analyzed an APB-enriched nuclear fraction from H₂O₂-treated cells and found that APBs preferentially associated with linear ECTR. We hypothesize that APBs may protect the cells from DNA damage–induced signaling by sequestering low molecular weight telomeric DNA with free ends.

Materials and Methods

Cell culture and treatments. The ALT cell lines GM847 (Coriell Cell Repositories, Camden, NJ) and JFCF/6-T.1J/1D (Dr. Paul Bonnefin, CMRI) and the telomerase-positive cell lines HT1080 and GM639 (American Type Culture Collection) were grown in DMEM (Invitrogen) supplemented with 10% fetal bovine serum (FBS) and in a 5% CO₂ incubator at 37°C. Treatment of cells with H₂O₂ (Merck) or MNNG (Sigma) was done at ∼70% confluence. Cells were exposed to 50 μmol/L H₂O₂ in serum-free DMEM for 45 min then cultured in DMEM containing 10% FBS for 48 h or were cultured in medium containing 50 μmol/L MNNG for 48 h.

APB staining and analysis. Cells grown on chambered slides were processed for detection of APBs as previously described ( 9). Slides were stained with 4′,6-diamidino-2-phenylindole (Sigma) mounted in antifade [2.33% (w/v) 1,4-diazabicyclo[2.2.2]octane (DABCO; Sigma) in 90% glycerol/20 mmol/L Tris (pH 8.0)] mounting media and stored at 4°C until analysis. Slides were analyzed on a Leica DMLB epifluorescence microscope (Leica) with appropriate filter sets, and images were taken with a cooled CCD camera (SPOT2, Diagnostic Instruments). For demonstrative purposes, images were further processed with Adobe Photoshop version 6.0 software. Cells were scored as APB positive if they contained one or more large PML bodies colocalized to telomeric DNA or at least three small PML/telomeric DNA colocalizations (i.e., wherein the telomeric DNA component was approximately the same size as the telomeres in that cell line). For confocal analysis, cells were processed as described ( 9), except that the secondary antibody was goat anti-rabbit IgG labeled with Alexa-633 which emits in the far red spectrum. Cells were imaged on a Leica TCS SP2 confocal microscope at 0.4-μm slices using the Leica Confocal Software. A single slice containing the brightest telomere signals was overlaid on the identical slice imaging the PML bodies using Photoshop 6.0.

Agarose plugs. Asynchronous and treated cells were collected and 3 × 10⁵ cells were resuspended in 2.0% low melt agarose. The plugs were digested in 2.0 μg/mL Proteinase K (Roche) overnight at 50°C, rinsed thoroughly and equilibrated in Tris-borate EDTA (TBE). The plugs were loaded into a 1.0% agarose gel overlaid with agarose and separated by electrophoresis at 4 V/cm for 5 h. The gel was dried and processed to visualize telomeric DNA as described previously ( 9).

APB enrichment. The isolation of an APB-enriched fraction of nuclei was obtained by modifying a protocol described by Lam et al. ( 10). Sonicated JFCF-6/T.1J/1D nuclei were resuspended in 1 mol/L sucrose and centrifuged at 3000×g for 15 min at 4°C. The supernatant was then resuspended to a final concentration of 1 mol/L sucrose, 15.4% percoll, 10 mmol/L Tris-HCl (pH 7.4), 0.5 mmol/L MgCl₂, 1% Triton-X and centrifuged at 235,000×g for 2 h at 4°C. The resulting pellet was resuspended in a final concentration of 0.5 mol/L sucrose, 15.4% percoll, 10 mmol/L PIPES, 0.5 mmol/L MgCl₂, 0.5 mg/mL heparin, 0.75% Triton-X. Marker beads of 1040 g/mL (blue) and 1052 g/mL (yellow) density (Amersham Biosciences) were added to the sample, which was centrifuged at 235,000×g for 1 h at 4°C. Fractions of 1 mL were collected, fraction 1 being above the blue beads and fraction 2 between the blue and yellow beads. The highest number fraction represents the pellet. Proteins were extracted by boiling the nuclear fractions in SDS, separated on a 10% acrylamide gel and detected with mouse anti-PML (Chemicon), 1:500, or rabbit anti-TRF2 (Dr. Lorel Colgin, CMRI) 1:1,000. Nuclear fractions were spotted onto slides, air-dried, and processed for APB detection as previously described ( 9).

Isolation and characterization of APB-associated telomeric DNA fractions. Nuclear fractions were phenol-chloroform extracted, ethanol precipitated, air dried, and resuspended in TE. DNA was separated by electrophoresis in a 1.0% agarose/1× TBE (0.045mol/L Tris-borate, 0.001mol/L EDTA) gel at 10 V/cm. In gel hybridization with a probe for telomere repeats was done as previously described ( 9).

Two-dimensional gel electrophoresis of low molecular weight DNA (Hirt lysate). Hirt lysate was isolated as previously described ( 11). Hirt lysate (30 μg, ALT positive or 60 μg, telomerase positive) was separated by two-dimensional gel electrophoresis as described by Cohen et al. ( 12). The telomere-specific probe consisted of 800 bp of telomere repeats excised and gel purified from the plasmid pTel ( 3). This probe was labeled by random priming according to the DECAprime II (Ambion) manufacturer's instructions.

Results and Discussion

DNA damage induces both linear and circular ECTR. The presence of circular and linear ECTR is characteristic of ALT cells. We hypothesized that a decrease in telomere length due to DNA damage ( 13, 14) would be reflected by an increase in ECTR. We embedded asynchronously growing and H₂O₂-treated or MNNG-treated GM847 (ALT positive) and GM639 (telomerase positive) cells in agarose, subjected them to electrophoresis in a 1.0% agarose gel, and probed for telomeric DNA ( Fig. 1A ). The GM847 cells showed a significant increase in the amount of ECTR after the treatments ( Fig. 1B), whereas GM639 did not ( Fig. 1A and B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

ECTR are increased in ALT cells after treatment with DNA damage agents. A, 3 × 10⁵ GM847 (ALT positive) or GM639 (telomerase positive) cells embedded in agarose plugs, subjected to electrophoresis, and hybridized with a ³²P-labeled telomere-specific oligonucleotide probe. Irrelevant lanes were removed from this gel using Photoshop 6.0. No other alterations were made. B, densitometric quantitation of low molecular weight genomic DNA (rectangle) confirms an increase in ECTR in the treated GM847 cells and no difference in the GM639 cells. Columns, mean from three independent experiments; bars, SD.

We determined the structure of this ECTR by two-dimensional gel electrophoresis. We speculated that the increase in ECTR might reflect an increase in circular ECTR, correlating with cleavage of the t-loop ( 5). Asynchronously growing GM847 and JFCF/6-T.1J/1D (also ALT positive) cells had a similar intensity of the arcs corresponding to linear and circular DNA ( Fig. 2A ). This observation was confirmed by densitometric analysis that indicated a ratio of linear to circular ECTR in asynchronous GM847 and JFCF/6-T.1J/1D of ∼1 ( Fig. 2B). Surprisingly, we found a similar pattern ( Fig. 2A) and ratio ( Fig. 2B) for H₂O₂-treated cells, suggesting that there was an increase in both linear and circular ECTR after telomeric DNA damage. Treatment of GM847 cells with MNNG resulted in a similar pattern and ratio as H₂O₂ treatment (data not shown). There was no increase in linear ECTR in H₂O₂-treated and MNNG-treated GM639 and no circular ECTR were seen (data not shown). It is likely that the origin of the linear ECTR is the telomere ends. The DNA in the arc corresponding to circular ECTR could arise either from double-strand breaks that cleave off t-loops with stems of various lengths or from recombination resolution events that result in the t-loops themselves being converted to free circles ( 4); the latter would suggest that telomere-specific DNA damage stimulates ALT activity.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

DNA damage does not affect the structure of ECTR. A, Hirt lysate from GM847 cells or JFCF-6/T.1J/1D cells was separated by two-dimensional gel electrophoresis. Asynchronous and H₂O₂-treated cultures contain similar amounts of linear and circular ECTR reflecting maintenance of the ratio between the linear and circular ECTR after DNA damage. B, densitometric analysis confirms the similar amounts of linear and circular ECTR in asynchronous as well as H₂O₂-treated cells. Columns, mean from two independent experiments; bars, SD.

DNA damage induces APBs. Previous work by our group and others ( 15) has shown that only ∼5% to 10% of exponentially growing ALT cells contain APBs, defined as PML bodies colocalizing with aggregates of telomeric DNA or telomere-binding proteins that are unambiguously larger than the telomeres within the same cell ( 6). These APBs can usually be detected with a 10× objective and often have the “donut” morphology characteristic of large PML bodies ( Fig. 3A ). Combined telomere fluorescence in situ hybridization (FISH) and indirect immunofluorescence also identifies smaller, telomere-sized foci that colocalize with smaller PML bodies and require a 40× or 100× objective for visualization ( Fig. 3A), and these are present in a larger proportion of cells within ALT populations ( Fig. 3B). These APBs were further analyzed by confocal microscopy to confirm colocalization of the telomere-sized foci with small PML bodies seen by conventional microscopy (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

DNA damage induces both large and small APBs in cells using the ALT telomere maintenance mechanism (TMM). A, in addition to the large APBs, ALT cells contain smaller telomere-sized APBs, both of which are induced after DNA damage with H₂O₂ and MNNG. At least 250 cells were counted in each category. B, large and small APBs can be distinguished in asynchronous and DNA- damaged cultures of both ALT (ALT+) cell lines GM847 and JFCF-6/T.1J/1D but not in the telomerase-positive (telomerase+) GM639 using combined telomere FISH (green) and indirect immunofluorescence against PML (red).

It has previously been shown that telomeres become shortened after exposure to DNA-damaging agents, such as H₂O₂ or MNNG ( 13, 14). We hypothesized that telomere shortening might be accompanied by an increase in ECTR DNA and that this might result in an increase in APB-positive cells. After treatment of GM847 and JFCF/6-T.1J/1D cells with either 50 μmol/L H₂O₂ or 50 μmol/L MNNG, there was a substantial increase in the number of cells containing APBs. In untreated populations of asynchronous GM847 cells, the proportion of cells containing large APBs was 11.7%, and 52.1% contained small APBs. This increased to 32.3% positive for large APBs and 83.5% for small APBs after H₂O₂ treatment, and JFCF/6-T.1J/1D increased from 21.5% (large)/64.7% (small) untreated to 33.9%/86.6% after H₂O₂ ( Fig. 3B). Similarly, MNNG treatment of GM847 and JFCF/6-T.1J/1D cells increased the proportion of APB-positive cells to 26.5%/75.0% and to 60.3%/84.6%, respectively ( Fig. 3B). Seven additional ALT cell lines have also shown induction of APBs by H₂O₂ treatment (data not shown), suggesting that this mechanism is common to APB formation and not specific for a particular cell line. No APBs of any size were detected in the telomerase-positive cell line, GM639, in asynchronous populations or after H₂O₂ or MNNG treatment ( Fig. 3A and B). In contrast, induction of DNA damage by UV light, hydroxyurea, or double thymidine block did not increase the number of APB-positive ALT cells (data not shown). These data show that ALT cells, when exposed to specific DNA damaging agents, have a striking propensity to increase the proportion of APB-positive cells. This ability to induce APBs has provided us with a method of further investigating the relationship of APBs to the ALT mechanism.

APB-associated telomeric DNA is predominantly linear. Because H₂O₂ induces both linear and circular ECTR, we wanted to determine the structure of APB-associated telomere repeats. We first adapted a method for purifying APBs from a protocol used to isolate Cajal bodies ( 10) using a series of sucrose gradients. The resulting fractions were analyzed for APBs and the APB-associated proteins TRF2 and PML. The majority of intact APBs could be visualized in fraction 2 by indirect immunofluorescence of the PML protein combined with FISH using a telomere-specific peptide nucleic acid probe ( Fig. 4A, top ). The enrichment of the APBs in fraction 2 corresponded with the enrichment of the telomere-binding protein, TRF2, and PML, as detected by Western blot analysis ( Fig. 4A, bottom). Although TRF2 was found in fraction 6, which represents the pellet after fractionation, PML was absent and no APBs were visualized (data not shown; Fig. 4A, bottom).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Fraction 2 contains APBs that predominantly associate with linear ECTR. A, colocalization of PML (green) and telomeric DNA (red) indicated the presence of APBs in one of the JFCF-6/T.1J/1D nuclear fractions. The enrichment for APBs in fraction 2 was corroborated by the presence of PML and TRF2, also found in fraction 6, by Western blot analysis. B, most of the genomic DNA was found in the first two fractions (1a and 1b) when stained with ethidium bromide. After hybridization with a (TTAGGG)₃ oligonucleotide probe, the majority of telomeric DNA was found in the APB-enriched fraction 2. C, two-dimensional gel electrophoresis indicated a significant amount of linear ECTR compared with circular ECTR in fraction 2. The predominance of linear ECTR DNA in fraction 2 was confirmed by densitometric analyses of the arc intensity ratios. D, to confirm the structure of the major arc in fraction 2, a linear marker, Spp-1, was separated with fraction 2 DNA by two-dimensional gel electrophoresis. Hybridization with the telomere-specific oligonucleotide revealed a major arc which, after stripping and reprobing with an Spp-1 probe, comigrated with the linear marker.

From these fractions, we were able to extract the APB-associated DNA and analyze it for telomeric content as well as structure. We first ran a single-dimension gel and stained it with ethidium bromide ( Fig. 4B, top). The majority of high molecular weight DNA was contained in fractions 1a, 1b, and 1c. The amount of DNA decreased steadily from fractions 2 through 6 ( Fig. 4B, top). Probing this gel with a telomere-specific oligo probe revealed an enrichment of telomeric DNA in fraction 2 ( Fig. 4B, bottom). Therefore, fraction 2 does not only contain most of the APBs (which contain the bulk of the telomeric DNA detectable by FISH staining) but also contains the majority of the telomeric DNA detectable by Southern blotting.

We analyzed fractions 2 and 6 by two-dimensional agarose gel electrophoresis. Fraction 2 contained a strong linear arc with a distinct, but less intense, circular arc ( Fig. 4C, left). The relative intensities of the APB-associated arcs in fraction 2 indicate a correlation between APBs and linear extrachromosomal telomeric DNA ( Fig. 4C, right) compared with the asynchronous and H₂O₂-treated parental JFCF/6-T.1J/1D. To positively identify the lower arc as linear DNA, we analyzed comigration of fraction 2 ECTR and 500 ng of linear marker DNA Spp-1 by two-dimensional agarose gel electrophoresis. We first identified the fraction 2 linear ECTR DNA using a telomere-specific probe ( Fig. 4D). After stripping and reprobing for the Spp-1 ladder ( Fig. 4D), we merged the fraction 2 image with the marker image and found that the arcs comigrated. Fraction 6 contained a single arc with the majority of DNA having a low molecular weight and high mobility structure (data not shown). As no APBs could be visualized by immunofluorescence microscopy of fraction 6, its associated telomeric DNA most likely represents a population of DNA sheared from the telomeres during the Hirt lysate extraction. Taken together, these data show that APBs associate primarily with linear telomere repeats in H₂O₂-treated cells. We therefore suggest that a function of APBs is to sequester linear telomeric DNA.

We have found that APBs and ECTR can be induced by DNA-damaging agents. The H₂O₂-induced APBs associate preferentially with linear ECTR. These data suggest a model in which linear ECTR may be recognized by the cell as a double-strand break or as an intermediate product in double-strand break repair. The small APBs may represent “new” sites of linear ECTR production resulting from rapid telomere deletion. Alternatively, they might be providing ECTR to telomeres for lengthening events. Circular ECTR DNA, which does not have free ends, presumably does not elicit the same response and mostly remains elsewhere in the nucleoplasm but could also participate in ALT activity, for example, by a rolling circle mechanism.

Several groups have reported the movement of PML bodies ( 16, 17). One group characterized the movements of telomeres in U-2 OS, an ALT cell line, and found transient associations of telomere-sized foci with larger telomeric aggregates ( 18). Although this group did not correlate PML with these larger telomere “clusters,” they most likely correspond to the large APBs. It is possible that the cell localizes the ECTR to a large APB either actively involved in homologous recombination or indirectly involved in removing the remaining components of the repair process. The budding yeast, Saccharomyces cerevisiae, forms a large repair focus, which is likened to a recombination center ( 19); multiple recombination/repair events occur within this one focus. It is possible that the large APBs represent a similar phenomenon.