Alternative Lengthening of Telomeres in Human Glioma Stem Cells

Abstract

Cancer stem cells are increasingly recognized as major therapeutic targets. We report here the isolation of glioma stem cells (GSCs) maintaining telomere length through a telomerase-independent mechanism known as alternative lengthening of telomeres (ALTs). TG20 cells were isolated from a glioblastoma multiforme, which had the ALT phenotype. They have no detectable telomerase activity and extremely long and heterogeneous telomeres colocalizing with promyelocytic leukemia bodies. The cancer stem cell potential of TG20 cells was confirmed based on their expression of neural stem cell markers, their capacity of in vitro long-term proliferation and to form intracranial tumors in immune-deficient mice. Interestingly, we found that both in vitro and in vivo TG20 cells were significantly more resistant to ionizing radiation than GSCs with telomerase activity. Analysis of DNA damage foci, DNA double-strand breaks repair, and chromosome instability suggest that radiation resistance was related to interference of ALT pathway with DNA damage response. Therefore, our data show for the first time that the ALT pathway can confer to cancer stem cells the capacity to sustain long-term proliferation as telomerase activity and importantly may also affect treatment efficiency. TG20 cells are thus the first cellular model of GSCs displaying ALT and should prove to be useful for the development of specific treatment strategies.

Tumor stem-like cells, Alternative lengthening of telomeres, Human glioma, Radioresistance

Introduction

Telomeres are specialized DNA-protein complexes found at the ends of chromosomes, which play a critical role in protecting chromosomes from end degradation and end-to-end fusions [1]. In humans, they consist of linear tandem arrays of TTAGGG repeats ending in a single-stranded G-rich 3′ overhang that participates in a higher-order terminal loop structure, the t loop. Telomeric DNA is lost with each cell division in most somatic cells, mostly due to the end replication problem. Short telomeres drive eukaryotic cells into replicative senescence, so the maintenance of functional telomeres is crucial for continued proliferation. Germ cells and somatic stem and progenitor cells overcome this problem by expressing telomerase activity, which is lost during differentiation [2, 3]. The reactivation of telomerase has been shown to be the most prominent feature of cancer cells. However, some cancer cells display no detectable telomerase activity, using a different mechanism for maintaining telomeres, referred to as alternative lengthening of telomeres (ALTs), which is based on homologous recombination at telomeres [4].

Glioblastoma multiforme (GBM) is the most aggressive form of brain tumors, with an extremely poor prognosis [5]. Radiotherapy is currently the primary treatment for this malignant tumor. However, even when combined with surgery, the median survival of patients with GBM is less than 1 year [6], with no significant improvement in prognosis over the last 20 years [7]. ALT activity is detected in about 30% of human glioma biopsy specimens and has been shown to be associated with a better prognostic [8–11], but no ALT cell lines or primary cultures from ALT gliomas have been described to date.

Cells resembling stem cells have recently been identified in gliomas (glioma stem-like cells, GSCs) [12–14]. These cells have certain characteristics in common with normal neural stem cells, such as the capacity for self-renewal and long-term proliferation, the formation of neurospheres, and the ability to differentiate into multiple nervous system lineages. However, they differ from normal neural stem cells in displaying the aberrant expression of differentiation markers, abnormal karyotypes and in generating tumors following their injection into immune-deficient mouse brain [13, 14]. GSCs are currently thought to be highly resistant to radiotherapy, surviving this treatment and subsequently repopulating the tumor [15].

Strikingly, only telomerase+ GSCs, and more generally telomerase+ cancer stem cells, have been described to date [16, 17], raising the possibility that GSCs may be absent from gliomas displaying ALT and/or that the stem cell potential of GSCs is necessarily associated with the reactivation of telomerase, as in normal neural stem cells. We report here for the first time the isolation and characterization of adult human GSCs displaying telomere maintenance by the ALT pathway. Our study shows that this alternative mechanism is compatible with the cancer stem cell status, which should be useful for understanding the physiopathology of ALT cancers. Interestingly, ALT GSCs appeared to be less responsive than telomerase+ GSCs to γ-irradiation. Analysis of DNA damage foci, DNA double-strand breaks repair, and chromosome instability suggest that this resistance results directly from interference of ALT pathway with DNA damage response (DDR).

Materials and Methods

Cell Cultures and Treatment

Tumor samples were obtained from surgical resections carried out on patients at Sainte Anne Hospital (Paris, France). This study was approved by the Institutional Review Board and informed consent was obtained from all patients. All the tumors used in this study were high-grade gliomas (glioblastoma or oligoastrocytoma III), according to the WHO classification, or malignant glio-neuronal tumors [18] according to the Sainte-Anne Hospital classification (Table 1). Glioma stem-like cells were obtained as previously described [18]. Briefly, tumor samples were dissociated to form a single-cell suspension, which was plated on serum-free Dulbecco's Modified Eagle Medium (DMEM)/F12 supplemented with B27 (Invitrogen, Carlsbad, CA, http://www.invitrogen.com), heparin (Stem Cell Technologies, Vancouver, Canada, http://www.stemcell.com), and human recombinant epidermal growth factor (EGF) and basic fibroblast growth factor (FGF-2) (Sigma, St. Louis, MO, http://www. sigmaaldrich.com), both at a final concentration of 20 ng/ml. Neurosphere cultures were then passaged every 11 days, by mechanical dissociation, to give a concentration of 50,000 cells per milliliter in fresh medium, in noncoated T25 or T75 flasks. Viable cells were counted by trypan blue exclusion.

Table 1

TP53 status of GSC cultures and characteristics of original tumors

For the monitoring of intracranial tumor growth, TG20 cells producing green fluorescent protein were obtained by lentiviral infection.

In some experiments, as specified in the figure legends, cells cultured in the same medium as used in neurosphere cultures were allowed to adhere to laminin (Sigma)-coated flasks, as previously described [21]. For differentiation experiments, cells were cultured in DMEM/F12 supplemented with 10% fetal bovine serum (FBS; Gibco-BRL, Gaithersburg, MD, http://www.gibcobrl.com) in poly L-ornithine (Sigma)-coated flasks.

Irradiation was carried out with a ¹³⁷Cs source (IBL637, CIS BIO International, Saclay, France, http://www.cisbiointernational.fr), at a dose rate of 0.94 Gy/minute, in laminin-coated plates.

Telomerase Activity Assay

We determined the telomerase activity of exponentially growing neurosphere cultures (7 days after seeding), using the TRAPeze enzyme-linked immunosorbent assay (ELISA) Telomerase Detection Kit (Chemicon, Temecula, CA, http://www.chemicon.com) according to the manufacturer's instructions. Protein concentrations were determined by the Bradford Method (BioRad, Hercules, CA, http://www.bio-rad.com).

Southern Blot Analysis of Telomere Restriction Fragments

DNA was purified with the DNA Isolation Kit for Cells and Tissues (Roche, Basel, Switzerland, http://www.roche-applied-science.com). Telomere restriction fragments (TRFs) were generated from 2 μg of total DNA, by restriction with HinfI and RsaI (Roche). Fragments were separated by electrophoresis in 0.8% agarose gels and visualized with the TeloTAGGG Telomere Length Assay Kit (Roche), according to the manufacturer's instructions, after hybridization with a telomere-specific, digoxigenin-labeled probe.

Telomere Fluorescence In Situ Hybridization and Metaphase Telomere-Induced Foci Assay

Telomere-fluorescent in situ hybridization (Telo-FISH) was carried out on metaphase spreads as previously described [22] using a telomeric cyanin-3-conjugated (C3AT2)3 peptide nucleic acid (PNA) probe (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com) complementary to the G-rich telomeric strand. In the quantification of chromosome damage after irradiation, a FITC-AAACACTCTTTTTGTAGA probe for centromeres (Eurogentec, Liège, Belgium, http://www.eurogentec.be) was used. Chromosome preparations were counterstained with 4′,6′-diamidino-2-phenylindole (DAPI) and observed under a fluorescence microscope (Olympus AX70). Digital images were recorded and analyzed with Cytovision system. Metaphase telomere-induced foci (Meta-TIF) assays were performed on adherent cultures treated for 2 hours with colchicin (0.1 mg/ml) and collected by cytospin onto polylysin-treated glass slides, as previously described [23].

Colocalization of Promyelocytic Leukemia/Telomeres

Dissociated neurospheres were collected by cytospin onto polylysin-treated glass slides. Cells were fixed by incubation for 10 minutes in 4% paraformaldehyde and permeabilized by incubation with 0.1% Triton X-100 and 0.15% bovine serum albumin (BSA) in phosphate-buffered saline (PBS) for 5 minutes. They were blocked by incubation for 1 hour in 7.5% BSA or 7.5% FBS or 0.1 Triton X-100 in PBS. Cells were incubated overnight at 4°C with an antibody against promyelocytic leukemia protein (PML; 1:100, Santa Cruz, Santa Cruz, CA, http://www.scbt.com) and then with fluorescein isothiocyanate (FITC)-conjugated secondary antibody, at room temperature, for 1 hour. Hybridization with the telomeric cyanin-3-conjugated (C3AT2)3 PNA probe was done as described above. Preparations were counterstained with DAPI, and images were acquired by confocal microscopy (DM 2500, Leica, Heerbrugg, Switzerland, http://www.leica.com).

Frozen tumor biopsy specimens were cut into 10-μm sections, on which telomere or PML colocalization analysis was performed with a slightly modified version of a protocol described elsewhere [9].

FACS Analysis

Neurosphere cultures were mechanically disaggregated and stained with PE-conjugated anti-CD133 (Miltenyi Biotec, Bergisch Gladbach, Germany, http://www.miltenyibiotec.com) and FITC-conjugated anti-CD15 (BD Biosciences, Franklin Lakes, NJ, http://www.bdbiosciences.com) antibodies, at a ratio of 5 μl of each antibody per 10⁶ cells, in a total volume of 100 μl, for 30 minutes at 4°C. Isotype controls coupled to the same fluorophores were also used. Data were acquired with a FACScalibur (BD Biosciences) and analyzed with FlowJo software (Tree Star Inc., Ashland, OR, http://www.treestar.com). For clonogenic analysis, cells were sorted into 96-well plates (1 cell per well) and incubated for 1 month before analysis.

Immunofluorescence Analysis

Cells were grown in Lab-Tek II Chambered Coverglass systems (Nunc, Rochester, NY, http://www.nuncbrand.com). They were fixed by incubation with 4% PFA for 10 minutes at room temperature and permeabilized by incubation in 0.1% Triton X-100.

They were then incubated overnight at 4°C with primary antibodies at the following dilutions: Sox2 (1:500; Chemicon), glial fibrillary acidic protein (GFAP) (1:400; Millipore, Lake Placid, NY, http://www.millipore.com), O4 (1:200; Millipore), MAP2 (1:100; Chemicon), CD15 (1:50; BD Biosciences), and γ-H2AX (1:400; Millipore). Secondary antibodies conjugated to FITC or rhodamine were then applied for 1 hour at room temperature (1:400; Molecular Probes, Eugene, OR, http://probes.invitrogen.com). Preparations were counterstained with DAPI and images were acquired by confocal microscopy (DM 2500, Leica).

Xenotransplantation

GSCs were injected stereotaxically into the striatum of anesthetized 3- to 4-month-old NOD-SCID-IL2Rγ mice (NSG; Jackson Laboratory, Bar Harbor, ME, http://www.jax.org). Animals were sacrificed 10 weeks later and immunofluorescence analysis was performed on brain slices with an anti-human Nestin antibody (1:400; R&D Systems, Minneapolis, MN, http://www.rndsystems.com) or an anti-GFP antibody (1:200; Abcam, Cambridge, U.K., http://www.abcam.com), as previously described [21]. All animal-related procedures were performed in accordance with the National Institutes of Health Guide for the care and use of laboratory animals.

TP53 Mutation Analysis

RNA was extracted from cells with RNeasy Mini Kit (Qiagen, Hilden, Germany, http://www1.qiagen.com) and retrotranscribed. PCR was performed by using 1 μl of the reverse transcriptase reaction, using Platinum Taq DNA Polymerase High Fidelity (Invitrogen). The primers were designed to overlap TP53 exons 4/5 (5′CTGTGACTTGCACGTACTCC3′) and exons 8/9 (5′TTGGGC AGTGCTCGCTTAGT3′). The thermocycling conditions were 5 minutes at 94°C for enzyme activation, followed by 35 cycles of 94°C for 30 seconds, 60°C for 1 minute, and 68°C for 1 minute. The PCR products (568-bp length) were purified and sequenced (Biofidal, Vaulx en Velin, France, http://biofidal.com).

Matrigel Invasion Assays

Matrigel invasion assays were performed in growth factor reduced Matrigel invasion chambers (BD Biosciences). Chambers were thawed for 1 hour and then hydrated in DMEM/F12 for 2 hours at 37°C and 5% CO₂. A total of 100,000 cells were loaded in a total volume of 500 μl of media containing 0.2 ng/ml EGF and FGF-2 (Sigma) to the top of the transwell chamber. The well in which each chamber rested was filled with 500 μl of growing media. Chambers were left for 26 hours at 37°C and 5% CO₂. Cells present on the upper surface of the filters were removed with cotton swabs. The chambers were then fixed in 4% PFA for 10 minutes followed by three washes of PBS and staining in crystal violet. Images of each chamber were taken using an inverted light microscope at ×40 magnification.

Neutral Comet Assay

For the detection of DNA double-strand breaks, a neutral Comet assay (single-cell gel electrophoresis assay) was performed according to the manufacturer's instructions (Trevigen, Gaithersburg, MD, http://www.trevigen.com) on adherent cultures. The quantification of tail moments was determined by Komet software. For each experimental point, 60–80 cells were evaluated.

Results

TG20 Cells Derived from a High-Grade Glioma Have an ALT Phenotype

Although 30% of GBMs have been reported to have an ALT phenotype, there is not any glioma cell line with an activated ALT pathway reported to date. We, therefore, determined telomerase activity with the TRAPeze ELISA Telomerase Detection Kit in glioma cells grown under neural stem cell culture conditions after isolation from six adult human glioma biopsy specimens of unknown telomerase status (Table 1 and [18]). Five of these glioma stem-like cell cultures (telomerase+ GSCs) displayed significant levels of telomerase activity, whereas no telomerase activity was detectable in the remaining culture, TG20, isolated from a second-relapse GBM in a 38-year-old patient (Fig. 1A).

1

Demonstration that cultured TG20 cells are of the alternative lengthening of telomeres (ALT) type. (A): Telomerase activity, as determined by PCR/ELISA (see Materials and Methods), in neurosphere cultures 7 days after passage. (B): Detection, by chemiluminescence, of the telomere restriction fragments of the designated cell lines by Southern blotting with a dioxigenin-labeled telomeric probe. (C): Representative metaphase spreads from telomerase+ glioma stem cells (GSCs; TG1N, left panel) and ALT GSCs (TG20, right panel) hybridized with a telomeric peptide nucleic acid probe (red) in Telo-fluorescence in situ hybridization experiments. DNA was counterstained with DAPI (blue). (D): Immunofluorescence staining of PML (first row, green) and telomeric DNA (second row, red) in neurosphere-derived cells of the designated cell lines. The third row shows the merging of these two images. Nuclear shape was visualized by DAPI staining (fourth row). Scale bar = 10 μm. (E): Colocalization analysis of PML bodies (left panel, green), telomeric DNA (middle panel, red), and the merged for both signals (right panel) on cryosections from the original TG20 tumor. Inset, the merged image for the two signals shows representative colocalization spots. Nuclear shape was visualized by DAPI staining (right panel, blue). Scale bar = 20 μm. Abbreviations: DAPI, 4′,6′-diamidino-2-phenylindole; ELISA, enzyme-linked immunosorbent assay; MWM, molecular weight marker; PCR, polymerase chain reaction; PML, promyelocytic leukemia. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

The molecular basis of ALT remains unclear, but this process involves telomere-telomere recombination [24, 25], generating a unique pattern of telomere length heterogeneity. We visualized the telomeric profile of TG20 cells by TRF Southern blot analysis (Fig. 1B). The telomere length distribution of TG20 cells resembled that of SAOS-2 cells, a well-known ALT osteosarcoma cell line [26] in which telomeres are long and heterogeneous in length (2–50 kb), whereas telomerase+ GSCs had the short, homogeneous telomeres typical of tumoral cells with telomerase activity, like the T98G glioma cell line used as a control. The heterogeneity of telomere length in TG20 cells was further confirmed by Telo-FISH on metaphase chromosomes, which revealed a much greater heterogeneity of telomeric signals at chromatid ends than for telomerase+ GSCs (Fig. 1C). Moreover, extrachromosomal telomeric DNA, another characteristic of cells displaying ALT, was abundant in TG20 cells (Fig. 1C) [27, 28].

Finally, cells displaying ALT have ALT-associated PML nuclear bodies (APBs) containing (TTAGGG)n DNA and telomere-specific binding proteins. APBs are a subset of PML bodies, present only in ALT cells and absent from telomerase-positive cells [29]. The colocalization of telomeric DNA and PML bodies, as shown by confocal microscopy, confirmed the ALT phenotype of TG20 cells, whereas no colocalization was found in any of the telomerase+ GSC cultures (Fig. 1D).

We next performed an in situ analysis of the tumor sample from which TG20 cells were isolated to determine whether the original tumor presented also the ALT phenotype. Results showed the colocalization between telomeric DNA and PML bodies in tumor cells, indicating the ALT status of this GBM (Fig. 1E), and therefore, that the ALT phenotype of TG20 cells is not a culture artifact but mirrors the ALT status of the original tumor. Taken together, these results demonstrate that TG20 cells are the first culture model described to date of human ALT glioma.

TG20 Cells Have the Properties of Cancer Stem Cells

We then evaluated the potential of TG20 cells to act as cancer stem cells. The widely accepted criteria defining GSCs are the ability to self-renew in defined stem cell culture conditions, the expression of normal neural stem cell markers, the ability to generate differentiated progeny, and the capacity to initiate brain tumors in immune-deficient mice [12, 13, 30, 31]. TG20 cells were grown as neurospheres for approximately 30 passages, that is, in medium containing EGF and FGF2 as a source of growth factors, allowing the proliferation of stem and progenitor cells [32]. They proliferated at a constant rate, with a population doubling time (PDT) of 9.07 ± 0.83 days (calculated from 16 passages), which is within the range of PDT of the other telomerase+ GSC cultured in parallel (from 3.78 to 11.96 days). Moreover, clonogenic assays, performed by seeding 1 cell per well in 96-well plates, evidenced that TG20 neurospheres contained 9.67% ± 0.73% long-term colony-initiating cells. These data demonstrate that ALT, like telomerase activity, is able to confer long-term self-renewing capacity to glioma cells.

Immunofluorescence staining of TG20 cells showed strong nuclear expression of Sox2, a transcription factor present in neural stem and progenitor cells during development and throughout adulthood (Fig. 2A). This transcription factor has also been detected in GSCs [33]. By contrast, CD133 expression was not detected by flow cytometry in TG20 cells or any of the other GSCs cultures used in this study (Fig. 2C and supporting information Fig. 1). CD133 has been identified as a putative marker of GSCs [13, 34], but several examples of GSCs lacking CD133 expression have already been reported [35, 36]. About 40%–60% of TG20 cells expressed CD15, another cell surface marker of neural stem cells and GSCs [37], as assessed by flow cytometry and immunofluorescence assays (Fig. 2A, 2C). None of the other GSCs used in these study were found to express CD15 (supporting information Fig. 1). Altogether, our data show that TG20 cells expressed neural stem cells markers such as Sox2 and CD15.

2

In vitro expression of stem/differentiated cell markers in human glioma-derived ALT cultures. (A): Immunofluorescence analysis of stem (SOX2, CD15) and differentiated cell markers after 14 days in culture, in the conditions specified in adherent cultures. Nuclear shape was visualized by DAPI staining. Scale bar = 50 μm. (B): Quantification of cells expressing stem cell/differentiation markers in the indicated conditions, as explained in (A). Error bars indicate the SEM (performed in triplicate, ***, p < .0001, as determined by Student's t test). (C): Flow cytometry analysis of surface neural stem cell marker expression (CD15 and CD133) in TG20 neurospheres cultured in stem cell conditions. Abbreviations: DAPI, 4′,6′-diamidino-2-phenylindole; EGF, epidermal growth factor; FBS, fetal bovine serum; FGF, fibroblast growth factor; FITC, fluorescein isothiocyanate; GFAP, glial fibrillary acidic protein. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Elimination of FGF2 and EGF and/or the addition of 10% FBS to the culture medium induce the differentiation of GSCs [30, 38, 39]. In this latter condition, TG20 cells lost Sox2 and CD15 expression (Fig. 2A, 2B). Conversely, most cells acquired expression of both GFAP (87.58% ± 4.1%) and MAP2 (92.45% ± 1.19%), markers of astrocytes and neurons, respectively, which were not detected in the parental TG20 cells cultured in stem cell medium. O4, a marker of oligodendrocytes, was not detected in either set of conditions. Thus, TG20 cells have the ability to differentiate into cells expressing both astrocyte and neuron markers, losing at the same time expression of stem cell markers, an important feature of GSCs [19, 38, 40].

Assessment of the ability of TG20 cells to form tumors in vivo was carried out using intracranial transplantation of 100,000 cells per severely immune-deficient NSG mouse. TG18 (Telomerase+) and TG20 (ALT) cell cultures expanded in stem cell conditions were compared. In vitro, the PDTs of these two cultures were similar (11.75 ± 0.54 days for TG18 vs. 9.07 ± 0.83 days for TG20). Mice were sacrificed 10 weeks after injection, revealing large numbers of engrafted human Nestin-positive cells that had infiltrated the host brain, in the case of TG18 cells (Fig. 3A and supporting information Fig. 2). By contrast, most of TG20 cells remained close to the injection site and were much less invasive (Fig. 3B and supporting information Fig. 2). Similar results were obtained after xenotransplantation of TG20 cells transduced with a construct encoding GFP that allowed identification of the grafted cells with an antibody against GFP (Fig. 3C). We then performed Matrigel invasion assays on TG18 and TG20 cells. In this in vitro assay, TG18 and TG20 cells exhibited similar invasion efficiencies (supporting information Fig. 2). This result suggests that the differing in vivo behaviors of both cell cultures were not directly related to decreased intrinsic invasion competence of ALT cells (TG20) as compared with telomerase-positive GSCs (TG18).

3

In vivo cancer stem cell properties of human glioma-derived ALT cultures. Cells (100,000 per mouse) from the specified adherent cultures were stereotaxically injected into the striatum of NSG mice. The animals were euthanized 10 weeks later, and engrafted cells were visualized by staining with the specified antibodies. (A): TG18 cells cultured in stem cell conditions, human Nestin staining (n = 3). (B): TG20 cells cultured in stem cell conditions, human Nestin staining (n = 4). (C): TG20 cells expressing GFP, cultured in stem cell conditions, GFP staining (n = 3). (D): TG20 cells, cultured in medium supplemented with 10% FBS for 7 days before injection, human Nestin staining (n = 3). Abbreviations: DAPI, 4′,6′-diamidino-2-phenylindole; EGF, epidermal growth factor; FBS, fetal bovine serum; FGF, fibroblast growth factor; GFP, green fluorescent protein. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Finally, we investigated whether FBS-induced TG20 differentiation was accompanied with a loss of its ability to form tumors in xenografts, as reported for other GSCs [30]. No human cells were found 10 weeks after the intracerebral injection of 100,000 TG20 cells previously cultured in the presence of FBS for 7 days (Fig. 3D and supporting information Fig. 3), clearly confirming the loss of tumorigenic potential of these cells in this situation.

Altogether, these data show that TG20 cells display the main characteristics of GSCs according to the current model.

TG20 Cells Are More Radiation Resistant Than Their Telomerase+ Counterparts

It has been suggested that GSCs are responsible for the repopulation of gliomas after chemotherapy or radiotherapy [34, 41, 42]. We, therefore, compared the effect of γ-irradiation on TG20 cells and on five telomerase+ GSCs cultures to evaluate the extent to which ALT interferes with the response to radiation.

Cells were cultured in adherent conditions (4,000 cells per well), as previously described [19], and were irradiated with doses of γ-irradiation ranging from 2 to 10 Gy, 24 hours after seeding in laminin-coated wells. Cell viability was assessed 5 and 10 days after irradiation, by the water-soluble tetrazolium salt-1 assay. TG20 cells were significantly more resistant to irradiation than the five telomerase+ GSC cultures. This was evident 5 and 10 days after irradiation for all tested doses (from 2 to 10 Gy; Fig. 4A).

4

ALT glioma stem cell cells are less sensitive to irradiation than their telomerase+ counterparts. (A): Cells of the indicated cultures (4,000 cells per well) were seeded in laminin-coated 96-well plates. Five (left panel) and ten (right panel) days after irradiation, cell viability was determined by the water soluble tetrazolium salt-1 assay. Quadruplicate determinations were carried out for each set of experimental conditions. The percentages given are the means ± SEM of two independent experiments. *, p < .05; **, p < .005; as determined with Student's t test. (B): Cells from the specified adherent cultures where irradiated as indicated and then stereotaxically injected into the left (nonirradiated) or right (irradiated at 5 Gy) striatum of NSG mice (100,000 cells per condition, n = 3 animals for each cell culture). The animals were euthanized 10 weeks later and engrafted cells visualized by staining with and anti-human Nestin antibody. Below each hemisphere, a magnification of the boxed area is shown. Red, human Nestin staining; blue, DAPI staining. Abbreviation: DAPI, 4′,6′-diamidino-2-phenylindole. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

To validate these results in vivo, we compared the tumorigenicities of irradiated (5 Gy) TG20 cells and TG1N (telomerase+) (Fig. 4B). Immediately after irradiation, 100,000 cells were injected stereotaxically into the ipsilateral striatum of NSG mice, whereas 100,000 nonirradiated control cells were injected into the contralateral striatum. Ten weeks after injection, animals were sacrificed and xenografted cells detected using an anti-human Nestin antibody. As shown in Figure 4B, only a few irradiated TG1N cells were found at the site of injection, contrary to nonirradiated cells, which proliferated and largely invaded the contralateral hemisphere. Confirming the radiation resistance of the ALT GSCs, irradiated TG20 cells were able to form tumors of a similar size as those generated by untreated TG20 cells (Fig. 4B). As a whole, we show that both in vitro and in vivo TG20 cells are less affected by ionizing irradiation than telomerase+ GSCs.

To gain insight into the mechanistic basis of this behavior, we evaluated TP53 status on these GSCs. In human gliomas, mutant TP53 correlates strongly with the ALT mechanism and with good prognosis, whereas for patients with telomerase-positive tumors, mutant TP53 confers a worse prognosis [10]. Furthermore, to date, U-2OS human osteosarcoma cells are the only ALT cell line known to have wild-type TP53. Mutations on this gene were thus analyzed by sequencing of exons 4, 5, 8, and 9 (Table 1). TG20, as well as TG10 and TG16 cells, presented mutated TP53. Furthermore, in the case of TG20, the same mutation was found on the original tumor. On the other hand, OB1, TG1N, and TG18 did not bare any mutations on TP53. These data are thus consistent with the relation between ALT gliomas and TP53 mutations [10] and indicate that TP53 status is not the single cause of the better resistance of TG20 to radiation.

Higher Activation of DDR in TG20 Cells

It has been previously demonstrated that glioma stem cells (GSCs) are more resistant to radiation compared with matched nonstem glioma cells due to preferential activation of the DDR pathway [34]. We, thus, tested if that could explain as well the relative radioresistance of TG20 cells by examining irradiated cells for the DDR marker phosphorylated histone H2AX (γ-H2AX). Four hours after 5 Gy of irradiation, TG20 cells presented a significant higher number of radiation-induced γ-H2AX nuclear foci compared with three different telomerase+ GSCs (Fig. 5A, 5B and supporting information Fig. 4). Interestingly, we also found that at the basal level TG20 cells presented a higher number of γ-H2AX foci than the telomerase+ GSCs (Fig. 5A, 5B and supporting information Fig. 4). It has been already shown that ALT immortalized cultures present a higher rate of spontaneous occurrence of telomeric DDR compared with telomerase+ counterparts [23]. Therefore, to determine whether spontaneous DDR took place at telomeres in unirradiated TG20 cells as other ALT cell lines, we measured telomere dysfunction-induced γ-H2AX foci at metaphase chromosomes (Meta-TIFs) [43]. As it can be seen in both Figure 5C and 5D, TG20 cells bear a higher number of spontaneous Meta-TIFs compared with three telomerase+ GSCs consistently with the higher number of γ-H2AX foci in interphasic TG20 nuclei.

5

ALT glioma stem cells activate higher DNA damage response compared with telomerase+ cells. (A): Quantification of γ-H2AX foci, determined by immunofluorescence, per nucleus on nonirradiated cells (left panel) or 4 hours after irradiating at 5 Gy (right panel). Cells were cultured in adherent conditions. Values are the means ± SEM (**, p < .005; ***, p < .0001, as determined by Student's t test). In parentheses is indicated the number of nuclei analyzed per condition. For irradiated cells, values are the difference between foci number at 5 Gy and 0 Gy. (B): Representative images depicting γ-H2AX foci (red) on the conditions stated on (A). Nuclear shape was visualized by DAPI staining (blue). Scale bar = 10 μm. (C): Spontaneous γ-H2AX immunofluorescence signal at metaphase chromosome ends (metaphase telomere-induced foci [Meta-TIF]) occurring on the stated cell lines grown in adherent conditions, measured as the percentage of telomeres colocalizing with γ-H2AX on one or both chromatids. Percentages are the means ± SEM (***, p < .0001; as determined by Student's t test). In parentheses is indicated the number of metaphases analyzed per condition. (D): Representative metaphase spreads from OB1 (left panel) and TG20 (right panel) cultures stained with γ-H2AX antibody (green), telomere fluorescence in situ hybridization (FISH) (red), and DAPI (blue). Arrows point to spontaneous Meta-TIFs. (E): Representative metaphase spreads 4 hours after 5 Gy of γ-irradiation stained with γ-H2AX antibody (green), telomere FISH (red) and DAPI (blue), from the stated cell cultures. Abbreviation: DAPI, 4′,6′-diamidino-2-phenylindole. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

We then analyzed H2AX phosphorylation pattern on metaphase chromosomes prepared 4 hours after 5 Gy of irradiation. As shown in Figure 5E, irradiation induced a high level of H2AX phosphorylation spanning large areas of metaphase chromosomes with no particular telomeric localization in both TG20 and telomerase+ GSCs. Therefore, our data show that TG20 cells are characterized by a higher level of (a) spontaneous telomeric DDR and (b) nontelomeric DDR after DNA damage, than telomerase+ GSCs.

TG20 Cells Exhibit Similar Efficiency of Double-Strand Breaks Repair but Lower Formation of Dicentric Chromosomes After Irradiation Compared with Telomerase + GSCs

Differences in survival after irradiation could be due to a better capacity to repair radiation-induced double-strand breaks (DSBs). To address this issue, we determined DSBs repair kinetics after irradiation by using the neutral Comet assay. As it can be seen on Figure 6A and 6B, both the telomerase+ OB1 and TG20 cells were found to repair DSBs with similar efficiencies at 1 and 20 hours postirradiation despite significant differences in radiation sensitivity (Fig. 4A).

6

Lower formation of dicentric chromosomes in ALT glioma stem cells. (A): OB1 and TG20 cells were irradiated with 5 Gy of IR. The presence of DNA double-strand breaks 1 and 20 hours after damage was assessed by single-cell gel electrophoresis assay under neutral conditions and comet tail moment measured. Cell viability was >90%, as determined by Trypan blue exclusion. Data are the means ± SEM (***, p < .0001; as determined by Student's t test). One representative experiment of two performed is shown. Cells were cultured in adherent conditions. (B): Representative micrographs from the conditions displayed in (A). DNA was visualized by SYBR green staining. (C, D): Percentage of chromatid breaks (C) and dicentric chromosomes (D) 24 hours after 2 and 5 Gy of γ irradiation on adherent cultures analyzed by Telo-fluorescence in situ hybridization. At least 50 metaphases were photographed and scored for each condition. Percentages are the means ± SEM (*, p < .05; **, p < .005; ***, p < .0001, as determined by Student's t test). p values above bars were calculated versus 0 Gy. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

As generation of lethal chromosome aberrations is an important consequence of irradiation, we next analyzed metaphase spreads prepared 24 hours after 2 or 5 Gy of irradiation (Fig. 6C, 6D). Whereas no differences in induction of chromatid breaks were found (Fig. 6C), TG20 showed significantly lower induction of dicentric chromosomes compared with telomerase+ GSCs (Fig. 6D).

Altogether, these data show that TG20 cells could not be distinguished from telomerase+ GSCs by the efficiency of DSBs repair but by the induction of lethal chromosome aberrations after irradiation.

Discussion

We report here the first identification and characterization of cancer stem cells maintaining their telomeres by an alternative mechanism not involving telomerase. TG20 cells fulfilled the widely accepted criteria for GSCs: (a) capacity to proliferate in vitro in stem cell culture conditions, (b) expression of normal neural stem cell markers, (c) differentiation capacity, and (d) initiation of intracranial tumors after their injection into immune-deficient mice. Thus, ability to sustain the long-term proliferation of cancer stem cells does not necessarily require telomerase activity, by contrast to what has been reported for normal stem cells. Instead, it may involve the ALT pathway despite this mechanism presumably maintains telomeres less efficiently than telomerase.

The original tumor from which TG20 cells were obtained had an ALT phenotype, like almost 30% of glioblastomas [8, 9, 11]. This was the third resection of an original tumor diagnosed 2 years previously (Table 1). The tumors induced by TG20 cells in animals progressed more slowly than those induced by their telomerase+ counterparts. These data are thus consistent with the ALT mechanism being an indicator of better prognosis for gliomas [9, 11].

Our results demonstrate that ALT gliomas may also contain a subpopulation of GSCs, as already shown for telomerase+ gliomas [16, 17]. This is important to understand the development of this type of tumor. Furthermore, TG20 cells provide the first cell culture model for ALT glioma, as no ALT glioma-derived cell line has ever been described before. The capacity of these cells to initiate tumors in mice, establishing the first animal model of this type of tumor, may be useful for the development of specific therapeutic approaches.

There is growing evidence that suggests that cancer stem cells are major potential targets for treatment [41]. It has been proposed that GSCs are resistant to radiation due to preferential activation of DNA repair pathways [34]. Possible changes in the response of cells to irradiation due to activation of the ALT pathway have been little investigated. We have shown here by using both in vitro and in vivo assays that ALT GSCs may be distinguished from telomerase+ GSCs on the basis of their greater resistance to radiation. TG20 cells were TP53-mutated, consistently with the relation between ALT gliomas and TP53 mutations [10]. But, as TG20 was found more radiation resistant than the other TP53 mutant/telomerase+ GSCs tested in this study, TP53 status could not be considered as the only cause of the radiation resistance of TG20. This suggests, therefore, that the ALT pathway may interfere with the cellular response to DNA damage and that radiation-induced DSBs are processed differently in ALT and telomerase+ cells. This is supported by our data showing higher induction after irradiation of DDR markers in TG20 cells than in telomerase+ GSCs. This increase was not associated with specific chromosome regions, indicating that radiation resistance of TG20 is not directly linked to telomere maintenance.

Moreover, as previously reported for other ALT cell lines [23], TG20 cells were characterized by a high level of spontaneous telomeric DDR, known to induce cell cycle arrest and senescence in other cell types [44]. This tolerance to telomeric DDR appears as an important feature of ALT cells, playing a crucial role in the mechanism of telomere maintenance [23]. Our results suggest that TG20 cells could also tolerate a higher threshold of nontelomeric DDR.

The ALT mechanism is dependent on homologous recombination at telomeres (for review see [25]). It remains unclear whether higher levels of recombination are restricted to telomeres, as some minisatellite instability occurs in ALT cells [45]. However, no increase in sister chromatid exchange was observed elsewhere in the genome of ALT cells [46, 47], and no increase in activity has been detected with a recombination reporter assay [46]. Accordingly, neutral Comet assays showed that TG20 cells were indistinguishable from telomerase+ GSCs concerning DSB repair efficiency, suggesting that the better radiation resistance of TG20 was not linked to a higher activation of the DNA repair machinery.

We observed a significantly lower induction of dicentric chromosomes in irradiated TG20 cells compared with all telomerase+ GSCs tested. Radiation-induced dicentric chromosomes are well-known lethal chromosome aberrations, which are the consequence of DSBs leading to defective chromosome end capping and further end-to-end chromosome fusions. As they are a major cause of breakage-fusion-bridge cycles, their lower induction is a possible cause for the relative radiation-resistance of TG20 cells.

It has been recently shown that ALT cells present a spontaneous occurrence of telomeric DDR in the absence of chromosome fusions, differing in that from non-ALT cells [23]. Cesare and Reddel [25] have proposed that ALT cells repress covalent ligation of telomeres eliciting DDR. Further experiments are needed to demonstrate whether similar repression of end-to-end fusions occurs at nontelomeric sites in TG20 cells, as suggested by our results, what may cause their better tolerance to radiation-induced double strand-breaks. This would be consistent with the higher activation of nontelomeric DDR markers observed in these cells, as compared with their telomerase+ counterparts.

Conclusion

In summary, we have shown for the first time that the ALT mechanism is compatible with the cancer stem cell status. TG20 cells, as the first model of ALT cancer stem cells, represent an interesting alternative to currently available immortalized ALT cell lines to study both in vitro and in vivo this mechanism of telomere maintenance. Our demonstration that the pathogenesis of ALT glioma could involve cancer stem cells and that ALT mechanisms could interfere with treatment efficacy should prove useful for the understanding of this particular type of glioma and for the development of specific therapeutic strategies.

Acknowledgements

We thank Dr. Cécile Thirant for GSCs isolation, Dr. Pierre Fouchet for technical assistance with the generation of TG20-eGFP-expressing cells, Julien Tilliet for assistance in the animal facilities, Jan Baijer and Nathalie Déchamps for assistance with clonogenic analysis, and Morgane Thion and Cristovao Sousa for technical help with in vitro invasion assays. This work was supported by funds from CEA (Programme Plasticité et Instabilité des Génomes) and DIM STEM-Pôle.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

Author notes

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

Author contributions: D.C.S.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript; J.R.P. and F.H.: collection and assembly of data, data analysis and interpretation; J.-M.S. and M.-A.M.: collection and/or assembly of data; F.P., M.-P.J., and H.C.: provision of study material or patients and data interpretation; F.D.B.: conception and design, financial support, data analysis and interpretation, manuscript writing, final approval of manuscript.

First published online in STEM CELLS  EXPRESS January 14, 2011.

Telephone: [phone], Fax: [phone]