alternative lengthening of telomeres: recurrent ... · cancer genome evolution that shapesthe...

Alternative Lengtheningof Telomeres: RecurrentCytogenetic Aberrations andChromosome Stability underExtreme Telomere Dysfunction1,2

Despoina Sakellariou, Maria Chiourea,Christina Raftopoulou and Sarantis Gagos

Laboratory of Genetics and Gene Therapy, Centerof Basic Research II, Biomedical Research Foundationof the Academy of Athens, Athens, Greece

AbstractHuman tumors using the alternative lengthening of telomeres (ALT) exert high rates of telomere dysfunction. Numer-ical chromosomal aberrations are very frequent, and structural rearrangements are widely scattered among thegenome. This challenging context allows the study of telomere dysfunction–driven chromosomal instability in neo-plasia (CIN) in a massive scale. We used molecular cytogenetics to achieve detailed karyotyping in 10 human ALTneoplastic cell lines. We identified 518 clonal recombinant chromosomes affected by 649 structural rearrangements.While all human chromosomes were involved in random or clonal, terminal, or pericentromeric rearrangementsand were capable to undergo telomere healing at broken ends, a differential recombinatorial propensity of specificgenomic regions was noted.We show that ALT cells undergo epigenetic modifications rendering polycentric chromo-somes functionally monocentric, and because of increased terminal recombinogenicity, they generate clonal recom-binant chromosomes with interstitial telomeric repeats. Losses of chromosomes 13, X, and 22, gains of 2, 3, 5, and20, and translocation/deletion events involving several common chromosomal fragile sites (CFSs) were recurrent.Long-term reconstitution of telomerase activity in ALT cells reduced significantly the rates of random ongoing telo-meric and pericentromeric CIN. However, the contribution of CFS in overall CIN remained unaffected, suggestingthat in ALT cells whole-genome replication stress is not suppressed by telomerase activation. Our results providenovel insights into ALT-driven CIN, unveiling in parallel specific genomic sites that may harbor genes critical forALT cancerous cell growth.

Neoplasia (2013) 15, 1301–1313

IntroductionMitotic chromosome integrity in humans relies on efficient DNAdamage responses (DDR), unfailing cell cycle checkpoints, as well asfunctional telomeres and centromeres [1–4]. Centrosomes, kineto-chores, chromatid cohesion, and nuclear and microtubule architecturealso play important roles in preserving faithful mitotic chromosomesegregation [5,6]. Chromosomal instability in neoplasia (CIN) is anextremely aggravated form of ongoing mitotic infidelity that is observedin most cancer cell populations [4]. Randomly dispersed CIN generatesclonal tumorigenic chromosome aberrations, contributes dramaticallyto intratumor genomic heterogeneity, and is mainly responsible forcancer genome evolution that shapes the multistep process of neoplasia[3,7]. Even more, CIN is related to advanced, incurable malignancyand is thought to complicate all current and future oncotherapeuticstrategies [4]. Understanding the patterns and driving mechanisms of

CIN may provide new tools toward personalized therapeutical schemesthat will be capable to defeat advanced cancers [8].In every neoplastic cell division, stability of chromosome content

is challenged by inherent impaired DDR, oncogene-induced DNA

Address all correspondence to: Sarantis Gagos, PhD, Soranou Efessiou 4, Athens 11527,Greece. E-mail: [email protected] work was supported by the Biomedical Research Foundation of the Academy ofAthens (Athens, Greece) in an intramural funding for S.G., by grant 05NON-EU-449of the Greek Secretariat of Research of the Greek Ministry of Development, and bythe EU COST Action BM0703 “Cangenin.”2This article refers to supplementary materials, which are designated by Tables W1 toW3 and Figures W1 to W5 and are available online at www.neoplasia.com.Received 3 September 2013; Revised 17 October 2013; Accepted 21 October 2013

Copyright © 2013 Neoplasia Press, Inc. All rights reserved 1522-8002/13/$25.00DOI 10.1593/neo.131574

www.neoplasia.com

Volume 15 Number 11 November 2013 pp. 1301–1313 1301

replication stress, and telomere deprotection [9,10]. Replication stressdue to chemical agents, activated oncogenes, or genetic interventionshas been shown to cause random illegitimate recombinogenicity ofcancer chromosomes that occurs frequently at common chromosomalfragile sites (CFSs) and can create novel clonal rearrangements [9–12].CFSs are AT-rich chromosomal regions that preferentially form

cytologically visible gaps or breaks on metaphase chromosomes underreplication stress [13]. The DNA polymerase inhibitor aphidicolinintroduces replication stress and induces 77 of 88 known human CFSs[13]. Fragile sites are conserved among mammals and are also foundin lower eukaryotes including yeast and flies [9]. CFSs are hotspotsfor gene amplification and viral integration, and they have been alsoimplicated in sister chromatid exchanges and in the generation ofconstitutional or acquired deletions and translocations [9].Telomeres protect the ends of eukaryotic chromosomes [14]. In most

human somatic tissues, these specialized nucleoprotein complexesare challenged after each round of DNA replication. From yeast tohumans, replicative loss of telomeric DNA is replenished by the actionof the RNP enzyme telomerase or by the telomerase-independentalternative lengthening of telomeres (ALT) [15]. Most normal humantissues do not possess a constitutive means to fully maintain theirtelomeres; thus, actively dividing cells demonstrate progressive telomericloss and deprotection [14]. Critical impairment of telomere protectionactivates DDR, and the cell cycle becomes arrested [16]. In normal cells,senescence and apoptosis are biologic barriers that prevent neoplastictransformation [1]. To overcome these barriers, human malignanciessustain continuous cellular growth by activating telomerase [14,17] orby using the alternative pathway of telomere lengthening (ALT) [15].The ALT pathway for telomere elongation was originally described

in yeast and in mammalian immortalized and cancer cells lacking telo-merase [15,18]. Although relatively rare in human neoplasia, the ALTpathway has been frequently observed in various types of aggressivehuman tumors such as osteosarcomas, undifferentiated pleomorphicsarcomas, leiomyosarcomas, astrocytic tumors (grades 2 and 3), andpancreatic neuroendocrine tumors [19]. In addition, the engagement ofthe ALT pathway may confer acquired resistance to cancer therapy intelomerase-positive cancer cells treated with telomerase inhibitors andhas been considered a major burden for current and future telomere-based antitumor therapeutics [20].Although not well understood, the mechanisms of ALT are thought

to engage non-homologous end joining (NHEJ) to seed “neo-telomeres”at broken chromosome ends [21]. However, the extent of this pro-cess still remains unknown [22]. ALT also implies the assembly andactivation of the complex homologous recombination-mediated DNAreplication machinery that elongates telomeres in the absence of telo-merase [23]. Cells using the ALT pathway display extensive telomericlength heterogeneity, ALT-associated promyelocytic leukemia (PML)nuclear bodies (APBs), extrachromosomal telomeric C-circles, massivetelomere dysfunction, large-scale epigenetic modifications at chromo-some termini, as well as high rates of structural and numerical chromo-some instability [23].Telomere dysfunction–driven genome damage is operated by fre-

quent chromosomal break-fusion-bridge (B/F/B) cycles [24,25]. TheB/F/B cycles can generate various types of oncogenic structural chromo-some rearrangements and may lead to extensive loss of genomic materialthrough anaphase lagging of whole chromosomes or chromosomal seg-ments [3,26]. In highly proliferating ALT cells, a continuous processof frequent B/F/B cycles generates very complex karyotypes that up tonow have not been analyzed in detail [27]. Numerical chromosomal

aberrations are very frequent, whereas structural rearrangements affectalmost every single chromosome [24]. This extremely challenging con-text provides excellent grounds for comparative analysis of clonal andrandom chromosome anomalies and allows the study of processesrelated to recombinant chromosome functionality in a large scale.In several histopathologic types of malignancies, specific chromosomal

aberrations have been associated with biologic mechanisms, along withthe diagnosis or prognosis of the disease [28]. Little is known aboutrecurrent chromosome rearrangements that might characterize the ALTpathway and may reflect known or unravel unknown genes or biologicpathways operating in this type of malignancy [27,29]. Furthermore,our knowledge on the distribution of CIN between different chromo-somes of the human genome and the processes related to clonal per-petuation of recombinant chromosomes in neoplasia still remains poor.We combined multicolor fluorescence in situ hybridization (M-FISH),

inverted 4,6-diamidino-2-phenylindole (DAPI) banding, telomere- orcentromere-specific FISH, strand-specific chromatid orientation FISH(CO-FISH), and immunocytochemistry to achieve detailed molecularkaryotyping in 10 human ALT cell lines. In a cohort of 649 uniqueclonal structural aberrations, we present the frequencies and chromo-somal locations of different types of structural rearrangements anddescribe cytogenetic markers of ALT continuous growth. Our datademonstrate that ALT-rearranged chromosomes are products of fre-quent ongoing telomeric, pericentromeric, and CFS recombinogenicity,and their clonality is preserved through extensive telomere healing andby widespread centromere epigenetic modifications that allow polycen-tric chromosomes to become functionally monocentric and perpetuatein culture. However, beyond extreme complexity compelled by frequentB/F/B cycles and despite the highly increased rates of pericentromericand CFS instability, human ALT-neoplastic karyotypes revealed recur-rent numerical and structural chromosome anomalies that unravelbiologic pathways related to cancerous cellular growth.

Materials and Methods

Cell Lines and Culture ConditionsThe osteosarcoma cell line Saos-2 was obtained from the American

Type Culture Collection (Wesel, Germany), and the osteosarcomacell line U2-OS was a gift from E. Gonos (National Hellenic Re-search Foundation, Athens, Greece). The SV-40 large T-antigen trans-formed ALT cell lines GM-847, VA-13, and IMR-90 were donated byA. Londoño-Vallejo (Institute Curie, Paris, France) [30]. In addition,we used a VA-13 derivative cell line that stably expresses human telo-merase RNA component (hTERC) and human telomerase reverse tran-scriptase (hTERT) and has reconstituted telomerase activity (VA-13TA)as described by Ford et al. [31]. The VA-13+hTERC+hTERT (VA-13TA) was a gift from J. W. Shay. The ALT liposarcoma Lisa-2 andLs-2 and the SV-40–transformed human ovarian surface epithelium celllines HIO107 and HIO118 were kindly provided by D. Broccoli (FoxChase Cancer Center, Philadelphia, PA) [32]. The T-24 bladder cancercells were provided by T. Vlahou (Biomedical Research Foundation ofthe Academy of Athens, Athens, Greece). HeLa and breast cancerMCF-7 cells were gifts from I. Irminger-Finger (GenevaMedical School,Geneva, Switzerland). The colon cancer SW-480 cell line was obtainedfrom the American Type Culture Collection. The osteosarcoma cellline KH-OS was provided by E. Gonos (Greek National Institute forResearch, Athens, Greece). The HT-29 colon cancer and NCI-H-460lung cancer cell lines, as well as the glioblastoma cell line SF-268,were provided by C. Dimas (Biomedical Research Foundation of the

1302 ALT Molecular Cytogenetics Sakellariou et al. Neoplasia Vol. 15, No. 11, 2013

Academy of Athens). All cell cultures were grown at 37°C and 5%CO2 in Dulbecco’s modified Eagle’s medium (Gibco, Grand Island,NY) supplemented with 10% FBS (Gibco), 25 units/ml penicillin(Sigma, St Louis, MO), and 25 pg/ml streptomycin (Invitrogen,Grand Island, NY).

Construction of Representative KaryotypesLogarithmically growing cell cultures were exposed to colcemid

(0.1μg/ml) (Gibco) for 1 to 3 hours, at 37°C in 5% CO2. Cells wereharvested by trypsinization (Gibco), suspended in medium, and spundown (10minutes at 1000 rpm). Supernatant was removed completely,and 5 ml of 0.075 M KCl (Sigma) at room temperature was addeddrop by drop. The cells were incubated for 20 minutes at roomtemperature, and then 1 ml of fixative [3× methanol (ApplichemGmbH, Darmstadt, Germany)/1× CH3COOH (Merck, Darmstadt,Germany)] was added. Cells were spun down (10 minutes at 1000 rpm),supernatant was removed, fixative was added, and the cells were re-centrifuged for 10 minutes at 1000 rpm. Finally, cells were droppedonto wet microscope slides and left to air-dry. For karyotypic analysis,we combined inverted DAPI staining, G-banding, and molecularkaryotyping byM-FISH (MetaSystems GmbH, Altlussheim, Germany).G-banding was performed after treatment with 0.25% trypsin (Gibco)and Giemsa (Carl Roth GmbH, Karlsruhe, Germany) staining. M-FISHwas performed according to themanufacturer’s protocols (MetaSystems).For inverted DAPI banding, slides were counterstained and mountedwith 0.1μg/ml DAPI in VECTASHIELD antifade medium (VectorLaboratories, Burlingame, CA). Cytogenetic analyses were performedusing a ×63 magnification lens on a fluorescent Axio-Imager Z1, Zeissmicroscope, equipped with a MetaSystems charge-coupled devicecamera and the MetaSystems Ikaros or Isis software. As representatives,we considered the karyotypes of the major clone per cell harvest (atleast 25 metaphases stained with M-FISH and inverted DAPI band-ing from one harvest were completely analyzed). Because of extremecomplexity of chromosomal rearrangements, karyotypes were writtenin the extended form according to International System for HumanCytogenetic Nomenclature (ISCN) 2009 [33,34].

Fluorescence In Situ HybridizationFor FISH, we used centromere-specific satellite DNA probes tar-

geting either all human centromeres or specific for chromosomes 1[yellow, red + green signals; green, fluorescein isothiocyanate (FITC);red, rhodamine], 9 (green, fluorescein), and 18 (green, FITC or purple/blue, spectrum aqua). Probes were purchased from Vysis (Abbott Park,IL) and Cytocell (Cambridge, United Kingdom). In brief, our protocolwas based on pepsin (Invitrogen) pretreatment, formamide (Appli-chem) or NaOH (Sigma) target denaturation, overnight hybridization,and high stringency post-hybridization washes. Telomeric peptidenucleic acid (PNA) FISH was performed using a Cy3-(CCCTAA)3PNA Probe according to the manufacturer’s instructions (DakoCytomation, Glostrup, Denmark). Briefly slides were incubated at3.7% formaldehyde (Carlo Erba Reagenti SpA, Milano, Italy), washedwith 1× TBS (Dako Cytomation), immersed in pretreatment solution(Dako Cytomation), and dehydrated with cold ethanol series (VWR,Radnor, PA). Probe and target DNAwere denatured at 80°C for 5 min-utes, and then slides were incubated for 1 hour at room temperature inthe dark. For M-FISH, we used the 24XCyte Kit from MetaSystemsGmbH. Staining was performed according to the manufacturer’sinstructions. All FISH preparations were mounted and counterstainedwith VECTASHIELD antifade medium containing 0.1μg/ml DAPI

(Vector Laboratories). Digital images were captured and enhanced ina MetaSystems workstation as described above.

ImmunocytochemistryMetaphase preparations were obtained through a modification of the

standard cytogenetic harvest technique that excluded acetic acid fixation.Freshly fed and subconfluent cell cultures were incubated with colcemid(0.01-0.04 mg/ml) for 0.5 to 16 hours at 37°C. Cells were dislodgedwith trypsin-EDTA solution and centrifuged at 1000 rpm for 10 min-utes, and the supernatant was discarded. Hypotonic solution (0.075 MKCl, prewarmed to 37°C) was added to the cell pellet, gently mixed bypipetting, and incubated for 20 to 30 minutes at 37°C. The cells werewashed with methanol. Preparations were cytospined for 10 minutes at2000 rpm, and slides were left to air-dry and immediately placed to4°C. Fluorescence immunocytochemistry was performed after phos-phate-buffered saline (PBS) washes, permeabilization with Triton X-100 (Applichem), and normal serum blocking. The CENP-A (StressgenBiotechnologies Corporation San Diego, CA) and CENP-C (Santa CruzBiotechnology, Santa Cruz, CA) antibodies were applied at dilutionsof 1:50 to 1:500 for 2 hours at room temperature. For antibody de-tection, we used a suitable Alexa Fluor–conjugated secondary anti-body (1:500-1:1000 in PBS; Molecular Probes, Grand Island, NY).Slides were incubated for 1 hour and washed with PBS, and then FISHwas applied as described above. For antibody detection, we usedAlexa Fluor 568 (Molecular Probes)– and IgG-FITC (Santa CruzBiotechnology)–conjugated antibodies, respectively (1:500 in 1% BSA).

Chromatid Orientation FISHStrand-specific telomeric CO-FISH was performed according to

Bailey et al. [35], with minor modifications. In brief, subconfluent cellmonolayers were cultured for 24 hours into medium containing 3 ×10−3 mg/ml 5′-bromo-2’-deoxyuridine (Sigma). Colcemid (0.1μg/ml;Gibco) was added for an hour before cell harvest. Metaphase spreadswere prepared by conventional cytogenetic methods. Chromosomepreparations were treated with 0.5 mg/ml RNase-A (Roche, Athens,Greece) for 10 minutes at 37°C, stained with Hoechst 33258 (0.5μg/ml; Sigma), incubated in 2× SSC (Invitrogen) for 15 minutes at roomtemperature, and exposed to 365-nm UV light (Stratalinker 1800 UVirradiator) for 30 minutes. The 5′-bromo-2’-deoxyuridine–substitutedDNA was digested with Exonuclease III (Promega, Madison, WI) in abuffer supplied by the manufacturer (5 mM DTT, 5 mMMgCl2, and50 mM Tris-HCl, pH 8.0) for 10 minutes at room temperature.Digested strand denaturation was performed at 70°C, with 70% form-amide (Applichem) in 2× SSC for 1 minute. The slides were then dehy-drated through a cold ethanol series (70%, 85%, and 100%) and air-dried. PNA-FISH was performed as above.

Statistical AnalysisOne-way analysis of variance (ANOVA) or paired t tests were

performed using the MINITAB software.

Results

ALT Karyotypes Derive from Polyploidization Followedby Frequent Chromosome Gains and LossesThe representative chromosome counts in metaphase spreads of the

10 cell lines of our panel varied from hypo-triploidy to near-heptaploidy(according to ISCN 2009) [33]. High chromosome numbers, andfrequent karyotypic presence of two or multiple copies of specific clonal

Neoplasia Vol. 15, No. 11, 2013 ALT Molecular Cytogenetics Sakellariou et al. 1303

recombinant chromosomes, indicated that one, two, or more roundsof genome reduplication have taken place and were selected to prevailin the cell lines of this panel (Figures W1 and W2 and SupplementaryData Set 1). In fact, all 10 karyotypes exhibited deviant modal chromo-some numbers from those expected after complete dosage duplication ormultiplication of the whole normal human genome content. Consistentwith the hypothesis of Duesberg et al. [36,37], this discrepancy can beexplained by the increased susceptibility of polyploid cancer cells toundergo chromosome losses. However, in the highly aberrant ALTbackground, clonal chromosomal gains into the representative ploidyindex were equally frequent to clonal losses (79 vs 83, respectively). Inaccordance to our recent report [38], all 10 cell lines of our panel exerteda continuous propensity to generate subclones with reduplicated ormultiplied genomic material compared to their major clones. Typically,the majority of observed polyploid metaphases were derivatives ofwhole-genome duplication of a preexisting subclone that had the sizeof the representative chromosome number (FigureW3). Hyperpolyploidcells (i.e., mitotic nuclei that had undergone three or more rounds ofgenome reduplication) were present but rare (<1%). These results in-dicate that the ALT karyotypes examined in this study are clonal evolu-tion products of an ongoing pressure for polyploidization that isaccompanied by equally frequent whole chromosome gains and losses.

Large Deletions, Terminal Structural Rearrangements,and Whole-Genome Pericentromeric RecombinogenicityCompared to telomerase-positive examples [39–42], the incidence

of clonal structural chromosome anomalies in the karyotypes of the

ALT cell lines of this panel was elevated by 3.7-fold (Figure 1A).Although monoclonal in origin, all ALT cell lines displayed variablepercentages of co-existing cytogenetically distinct subpopulations,characterized by unique clonal rearrangements (Figure W3). Accordingto ISCN 2009 [33], clonality of structural chromosome aberrationsin a cancer cell population is defined by the presence of an identicalmarker chromosome, in at least two co-dividing cells. As representative,we considered a particular karyotype of a series of at least 25 fullyanalyzed mitotic cells from the same harvest, containing most, if notall, of the identical chromosome markers shared by the majority ofthe cells. In the representative karyotypes of the 9 unique ALT cell linesof our panel, consisting of a total of 867 clonal autonomous chromo-somes, 518 were identified as structurally aberrant (59.75%). Severalrearranged chromosomes were found in multiple copies or exertedmore than one structural anomaly (Figures W1–W3 and Supplemen-tary Data Set 1). In a total of 649 identified unique clonal structuralaberrations, deletions of large genomic segments mostly spanning upto chromosomal termini were the most frequent (41.4%), indicatinga robust pressure toward ALT chromosome trimming. More than halfof total deletions (65.7%) encompassed pericentromeric breakpoints.Translocations between chromosomal segments composed the 30.1%of total aberrations. They were mostly unbalanced and often complex(involving more than two chromosomes). Consistent to our previouswork [24], a high proportion of the translocation breakpoints waspericentromeric (52.8%). Clonal pseudo-dicentric or pseudo-polycentricchromosomes were frequent (13.3%). Inverted duplications wereencountered in 4.8% of the total clonal rearrangements, whereas

Figure 1. Clonal structural chromosome rearrangements in human neoplastic cell lines: (A) Significant differences in the frequencies ofclonal structural aberrations in the karyotypes of eight ALT and eight telomerase-positive human cancer or transformed continuous celllines based on the results presented in this study and in [39–42] (ANOVA). (B) Frequencies of different types of aberrations in 649 uniqueclonal chromosome rearrangements, identified by M-FISH and inverted DAPI banding in nine human ALT cell lines.


isochromosome formations, inversions, or duplications were more rarelydetected (0.15%-1%; Figures 1B, W1, and W2). Hence, the increasedrates of telomere dysfunction operating in the ALT pathway lead pre-dominantly to large-scale deletions and unbalanced translocations thatfrequently occur at the pericentromeric regions.

Distribution of Telomere Dysfunction–Driven ClonalRearrangements along Chromosomal Termini andTelomere Healing of Recombinant ChromosomesFrom 1036 chromosome arms involved in ALT clonal structural

aberrations, 235 (22.7%) took part in terminal fusions. This percentageis raised up to 39%, if we take into account the incidence of clonalpseudo-polycentric chromosomes and inverted terminal duplications.Interestingly, with the exception of chromosome 9, which showedremarkably increased rates of telomeric recombinogenicity in both arms,smaller chromosomes (13-22) showed a significantly higher propensity toundergo clonal terminal p- or q-fusions (P = .031 by one-way ANOVA;Figure 2A). In a highly proliferating cell population, clonal maintenanceof structurally rearranged chromosomes requires the presence and func-tionality of the two main chromosomal organelles: the centromere andthe telomere [3,43]. ALT recombinant chromosomes may obtain telo-meres by using a preexisting telomere of another chromosome (canonicalcapping) or by introducing a tract of telomeric repeats through homolo-gous recombination or NHEJ at a “blunt” chromosome end to heal it(telomere capture or healing; Figure 2B) [21,44]. From 1036 arms ofthe rearranged chromosomes described in this study, 54.9% were foundcanonically capped. A substantial proportion (31.7%) of the recombinantarms that were terminally protected by telomere healing acquired telo-meres at pericentromeric regions (Figure 2C ). Expanding our earlier

observations [24], every human ALT centromere involved in peri-centromeric whole-arm deletions was found capable to capture telomeresand to become stabilized as a neo-acrocentric chromosome. Centro-meres of chromosomes 18 to 20 were proved highly recombinogenicin both telomeric and centromeric regions (Figure 2A). From 219 non-centromeric breakpoints healed by telomere capture, 89 (40.6%) co-incided with common CFSs, whereas 21.9% were found at genomicregions bearing conserved interstitial telomeric repeats (ITRs; Fig-ure 2D) [45]. These results indicate that in the ALT context, everychromosome of the human karyotype can be involved in telomeric orpericentromeric recombinogenicity and can acquire telomeric repeats atbroken ends.However, specific chromosomes display elevated propensityfor telomeric or centromeric instability. In addition, in a significantproportion of structural rearrangements, clonality is entailed throughthe acquisition of “neo-telomeres” near CFSs or conserved ITRs.

Clonal Pseudo-Polycentric Chromosomes and RecombinantChromosomes with Large ITRsCompared to telomerase-positive examples, the ALT cell lines of this

panel demonstrated significantly higher frequencies of pseudo-dicentricor pseudo-polycentric recombinant chromosomes as well as rearrangedchromosomes with cytologically visible blocks of telomeric repeats atthe recombination breakpoints (Figure 3, A and C). In humans, clon-ality of polycentric recombinant chromosomes implies activity of onecentromere and inactivation of all other extra centromeres [46]. Aproportion of the rearranged chromosomes of our cell line panel wasinterpreted as pseudo-dicentric by M-FISH and inverted DAPI band-ing (Figures 3A, W1, and W2 and Supplementary Data Set 1). Wefurther investigated these peculiar entities of ALT genomic instability

Figure 2. Terminal fusions, pericentromeric rearrangements, and telomere healing: Virtually all chromosome arms of the human karyotypeare involved in ALT clonal end-to-end fusions, and all centromeres are capable of taking part in clonal translocations and of undergoingtelomere capture (p-arms of acrocentric chromosomes were not examined). (A) Chromosome 9 and the small chromosomes (18-22)seemed more prone to take part in terminal or pericentric rearrangements (P < .01 by ANOVA). (B) A karyogram of the ALT U2-OS cellline, stained by inverted DAPI banding and a telomere-specific PNA FISH probe, labeled with Cy3 (red), indicates chromosomal sites oftelomere healing (arrows). (C) Telomere healing in the ALT pathway occurs in a massive scale because 45.1% of the recombinant chromo-some arms maintain clonality through capture of telomeric repeats. (D) Sites of frequent telomere capture.


by centromere- and telomere-specific FISH and immunocytochemistryusing antibodies specific for CENP-A and CENP-C proteins. Fluores-cent microscopy revealed that the excessive centromeres of clonal ALTdicentric or polycentric chromosomes are indeed inactivated [47](Figure 3B). In addition, telomere-specific FISH showed that severalALT clonal aberrant chromosomes deriving from end-to-end fusionsmaintained cytologically visible blocks of telomeric repeats at recom-bination sites (Figure 3D). Examination of the orientation of recom-binant interstitial telomeric sequences, by strand-specific CO-FISH,indicated that they represented NHEJ processes between telomeres atan antiparallel orientation or between telomeres and non-telomericgenomic sites. The above results demonstrate that in the ALT pathwaythe inert tendency for aberrant large-scale epigenetic modifications isnot confined only at the telomeres [23,48] but may affect severalcentromeric regions, facilitating inactivation of centromeres in excess,in polycentric chromosomes produced by B/F/B cycles. In addi-tion, increased terminal recombinogenicity generates recombinantchromosomes with large ITRs.

Recurrent Chromosomal Rearrangements in the ALT PathwayWe then attempted to identify recurrent chromosome aberrations

that might be related to the genetic pathways driving ALT continuous

growth or may reflect specific patterns of ALT chromosome evolutionunder increased telomeric instability. As recurrent, we considered allunique chromosome anomalies, affecting the same chromosome, inat least three of nine ALT cell lines. Losses of genomic material fromwhole chromosome 13 were recurrently observed in seven of ninekaryotypes (77.7%). Other chromosomes found to be frequently lostwere 22 (77.7%), X (55.5%), 8, 14, and 18 (44.4%), as well as 1, 7, 9,10, 11, 15, 16, and 21 (33.3%; Table W1). The same chromosomessuffered also from large terminal or whole-arm deletions. Recurrentchromosome gains beyond ploidy index involving chromosomes 2, 3,5, and 20 were observed in 50% of the cases (Table W1). Recurrentpericentromeric rearrangements are described in Table W2. Deletionsor unbalanced translocations of both arms of chromosome X wereobserved in all the cell lines of this panel. The breakpoint that wasmore frequently affected by terminal deletions was between Xq21.1and Xq22.3 (in six of nine lines). A recombination/deletion hotspotwas localized at the short arm of chromosome 1 (between 1p32 and1p36; observed in six of nine cell lines); in addition, chromosome 1was also found to be affected by terminal deletions originating at 1q21(four cases) with the other cases displaying either larger 1q deletions orlosses of the whole chromosome 1. Chromosome 2 exerted a recom-bination hotspot at 2q31-2q36 (five of nine lines), whereas seven of

Figure 3. Distinctive cytogenetic findings of the ALT pathway: (A) Incidence of clonal pseudo-polycentric chromosomes in eighttelomerase-positive and eight ALT cell lines (P = .019 by ANOVA). Examples of two “cryptic,” clonal pseudo-dicentric recombinantchromosomes of the ALT U2-OS karyotype: At the upper row, a neo-acrocentric derivative of chromosome 1, lacking the whole1p arm, displays alphoid centromeric repeats specific for human centromeres 1 and 18, located in close proximity to the telomere(M-FISH: yellow, specific for chromosome 1; purple, specific for chromosome 18; inverted DAPI: gray/black, ×630). At the lowerrow, a complex rearrangement between chromosomes 1 and 9 maintains alphoid repeats from both centromeres 1 and 9. (B) Lack ofchromatid cohesion and negative staining for antibodies specific for CENP-A or CENP-C indicate that the terminally positioned centro-mere 1 is inactivated, whereas centromere 9 remains active. (C) Frequencies of clonal rearranged chromosomes bearing cytologicallyvisible ITRs in a panel of eight telomerase-positive and eight ALT cell lines (P = .014 by ANOVA). (D) Strand-specific CO-FISH revealstwo types of clonal ITR: those representing NHEJ-mediated telomeric fusions between telomeres at antiparallel orientation (yellowarrows) and fusions of telomeric repeats with non-telomeric genomic regions (white arrows; ×630).


nine ALT lines exerted increased rates of recombinogenicity for theregions between 3q21 and 3q26. Chromosome 4 was found highlyrecombinogenic at bands 4q21 to 4q23 (six of nine lines). Deletionsof the short arm of chromosome 5 were observed in six lines (66.6%).A recombination hot region on this chromosome was identified between5q31 and 5q35 (in six lines). Six of nine lines displayed rearrangementsinto the regions 6q22.1 to 6q24.3, whereas four of them and two othercell lines displayed a hot region between 6p21.3 and 6p23. Deletionsof the short arm of chromosome 8 were present in six of nine lines,whereas five lines showed breakage at 8q21-8q24. The terminal regionsof chromosomes 7, 9, 11, and 18 to 21 showed a highly increasedtendency to undergo end-to-end fusions or terminal translocations(88.8%). Pericentromeric rearrangements of chromosome 10 were

found in 88.8% of the ALT lines. A clustering of breakpoints was alsonoted between bands 10q21.2 and 10q22.1. Six cell lines displayedterminal deletions of 14q21-22. The pericentromeric region of chro-mosome 16 exerted high rates of breakage and recombination, mainlyaffecting p11.2 (in seven cell lines). Chromosome 18 showed alsoremarkable pericentromeric recombinogenicity. Seven of nine cell linesdisplayed large deletions of the short arm of chromosome 18, whereasthe other two lines showed full losses of the same chromosome. Similarwas the case with 18q. In addition to increased fragility that gave rise toterminal deletions of either the short or the long arms, pericentromericregions of chromosome 18 were often taking part in translocationswith pericentric sites of other chromosomes such as 1, 7, 9, 10, 15,16, 17, and 20. Losses of 19p were observed in six lines of our panel.

Figure 4. Structural chromosome instability before and after long-term reconstitution of telomerase activity in ALT cells: Reconstitutionof telomerase activity in ALT VA-13 cells significantly reduced overall genomic structural CIN in two independent harvests (a and b) ofthe double-transfected isogenic VA-13TA cell line, grown in the presence of telomerase activity for more than 250 PDs. We analyzed1095 VA-13 and 3416 VA-13TA chromosomes (1657 from harvest a and 1759 from b) using M-FISH/inverted DAPI. (A and B) Distributionof random (non-clonal) structural rearrangements per chromosome, in the VA-13 cells, reveals that, in the ALT pathway, any chromosomecan be affected by increased recombinogenicity that predominantly affects terminal and pericentromeric regions and can be substantiallysuppressed on long-term activation of telomerase. Random chromosome rearrangements were classified as telomeric, pericentromeric,genomic (non-telomeric or centromeric), and unidentified. (C) Long-term expression of telomerase activity led to significant suppression oftelomeric and centromeric vulnerability (P < .0001 for VA-13TAa and for b, by paired t test). (D) The contribution of CFS in the proportion ofidentified random chromosome breakpoints was not decreased, indicating that telomerase activation cannot repress ALT whole-genomereplication stress.


Chromosome 20 showed increased propensity for pericentromericrearrangements and terminal fusions (88.8%). The long arm ofchromosome 21 showed recurrent propensity to undergo deletions ortranslocations. A recombinatorial hotspot was found at 21q22.1. Asignificant proportion (37.3%) of the above recurrent rearrangementswas localizing at common CFS (Tables W2 and W3), suggesting thatvirtually all ALT chromosomes suffer from generalized replicationstress. Therefore, beyond extreme complexity, ALT karyotypes arecharacterized by numerous recurrent chromosomal imbalances thatmay uncover genes and biologic pathways related to ALT.

Reconstitution of Telomerase Activity in ALT CellsSuppresses Terminal and Pericentromeric CINReconstitution of telomerase activity in human ALT cells has been

related to decreased frequencies of telomere dysfunction and reducedCIN [31]. The VA-13+hTERC+hTERT (VA-13TA) cells are stablytransfected to express telomerase activity for more than 250 populationdoublings (PDs) and display high prevalence of near-pentaploidy[31,38]. To address, in a cell-by-cell basis and in a chromosome band-by-band basis, the extent of suppression of ALT-driven CIN betweenco-dividing cells, we examined, before and after long-term restorationof telomerase activity, the frequencies of non-clonal (unsystematic) struc-tural rearrangements in 15 randomly selected metaphase spreads ob-tained from the same harvest of the ALT VA-13 and 30 mitoses fromtwo independent harvests of the telomerase-positive VA-13TA stainedby M-FISH and inverted DAPI (Figures W3–W5). The VA-13TAkaryograms maintained, in duplicate, most of signature anomaliesobserved in parental VA-13 cells (Figures W4 and W5). As expectedby the high rates of identified ALT clonal rearrangements, 238 of1095 VA-13 chromosomes (21.7%) demonstrated random telomeric,pericentromeric, or genomic recombinogenicity (Figure 4A). Reconstitu-tion of telomerase activity resulted in a 2.3- to 3.6-fold decrease of thisphenotype (Figure 4B). Similar to the distribution of clonal anomalies,random terminal or pericentromeric CIN affected all of the chromo-somes of VA-13 and VA-13TA karyotypes (Figure 4A). Telomeraseactivity significantly suppressed telomere fusions (4.6- up to 7.4-folddecrease). Although frequent in both cell lines, unsystematic peri-centromeric instability was also significantly repressed after restorationof telomere dysfunction by ectopic telomerase expression (1.9- to3.6-fold; Figure 4C). However, in both VA-13 and VA-13TA cell lines,a significant proportion of identified random genomic rearrangementscoincided with common CFSs (16.2% in VA-13 and 20.5% to28.6% for VA-13TAa and b, respectively; Figure 4D). These resultssuggest that the breakpoint patterns of random ongoing chromosomalinstability in ALT cells are analogous to those of clonal rearrangements.In addition, a proportion of the extremely high pericentromeric, but notof CFS, recombinogenicity in the ALT karyotypes may be attributed tomechanical forces generated during telomere dysfunction–driven B/F/Bcycles, because they are repressed by the introduction of telomerase [38].

DiscussionTumor genome evolution in humans is considered a rather slow pro-cess that may take several years to produce a highly malignant genome[7,8,49,50]. Cellular immortalization and continuous culture growththrough the ALT pathway are accompanied by extremely high ratesof CIN that generate a plethora of random and clonal structural chro-mosome anomalies [51]. Complicated more by high frequencies oftelomere dysfunction–driven polyploidization and extensive chromo-some gains or losses [38,52], the ALT karyotype provides an ideal

context to study massive tumor genome alterations occurring in a“fast-forward” mode.In this first comparative detailed molecular cytogenetic analysis

of human ALT cancer and immortalized cell lines, we identified arelatively large cohort of recombinant chromosomes suitable for sta-tistical analyses concerning distribution of imbalances, breakpoints,and clonality potential of cancerous rearrangements under extremetelomere dysfunction.Telomere dysfunction has been related to polyploidization through

genome endoreduplication as well as to chromosome losses throughB/F/B cycles and anaphase lagging [53,54]. Our results suggest thatextreme telomere deprotection might also be capable of triggering sin-gle chromosomal non-disjunctions, or ALT immortalized cells maygain chromosomes either by whole chromosome reduplication orwidespread hyperpolyploidy reduction [55].Despite the increase of genomic content because of polyploidization,

the great extent of large-scale deletions and unbalanced translocationsidentified in this study confirmed that the ALT genome suffers fromextensive dosage depletion of large segments of genomic material, asdescribed in a recent SNP array report from 22 human ALT cell lines(including four of the lines of our panel: U-2OS, GM847, SaOS-2,and VA13) [29]. The rarity of the ALT pathway in neoplasia andthe tightly regulated stochastic activation of ALT-like mechanisms inembryonic stem (ES) and induced pluripotent stem (iPS) cells[56,57] imply that a number of critical suppressor pathways are abro-gated during massive engagement of ALT in cancer. Hence, severalALT suppressors may exist and might be recurrently depleted in theALT cell lines of our panel.We have shown in the past that, in addition to the expected fre-

quent telomeric vulnerability, ALT cells present also high rates ofpericentromeric instability [24]. Centromere mitotic recombination isincreased in cells lacking the DNA (cytosine-5) methyltransferase 3a(DNMT3a) and DNMT3b [58], suggesting an altered state of ALTcentromeric heterochromatin that facilitates local recombination ac-tivity. Our current data expand these previous observations, propos-ing that, in the ALT context, every human telomere, or centromere,can be involved in clonal rearrangements. Clustering of breakpointsat specific genomic regions might correspond to peculiarities of localchromosome architecture or represent selection-adaptation processesfavoring ALT critical genome alterations.Although the presence of one or a few dysfunctional telomeres has

been associated with initiation of CIN [16], little is known about thedegree of involvement of every specific telomere of the human genomein dysfunction-driven chromosome rearrangements. Our results revealthat virtually all telomeres of the ALT human karyotype can be affectedby random or clonal recombinatorial events. However, specific chromo-some termini were found more fusion-prone, and some were resistant.This may suggest that specific human chromosome termini are morevulnerable to generalized telomere protection failure than others. How-ever, selection and adaptation of marker chromosomes that conferoncogenic advantages to the ALT cells should not be disregarded. Forexample, our results indicate high telomeric vulnerability of 9qter,but this region contains important cancer-related genes such as Notch1,TRAF2, and COBRA1 [59–61], depletion, duplication, or alteration ofwhich might be important for ALT.To become clonal and to contribute into critical genome imbalances

that are responsible for cancerous growth, CIN-generated aberrantchromosomes with blunt chromosome ends get stabilized throughextensive telomere healing [21,62]. In the absence of telomerase activ-


ity, clonal perpetuation of terminally deleted chromosomes impliesseeding of telomeres at the terminal breakpoint. Telomere healingcan be achieved through homologous recombination, NHEJ, or byreplication fork stalling and template switching [21]. We identified219 genomic regions at which microscopically detectable terminaldeletions were mitotically stabilized by telomere healing so to becomeclonally maintained. Virtually all pericentromeric regions were capableto acquire telomeric repeats and to ensure mitotic perpetuation ofstructurally altered chromosomes. CFSs and ITRs were also frequentsites of telomere healing. Hence, our results demonstrate that, in theALT pathway, telomere healing occurs in a massive scale and mayinvolve both NHEJ and homologous recombination (HR).Centromeric inactivation renders human meiotic or mitotic, recom-

binant, poly-centric chromosomes to become functionally mono-centric by a largely unknown process of epigenetic modifications thatameliorate binding of centromere-specific CENP proteins and restrictchromatid cohesion that shapes primary constriction of metaphasechromosomes despite the presence of alphoid DNA repeats [46].The high frequencies of clonal pseudo-polycentric chromosomes inthe ALT pathway are described for the first time. They reflect thewell-established elevated rates of B/F/B cycles, owing to inert con-tinuous stochastic telomere failure [22,53], but may also represent apeculiar ALT heterochromatin epigenetic status that is highly per-missive for the inactivation of excessive centromeres. Loss of functionof DNMT3a and DNMT3b in mice was associated with defects intelomeric and subtelomeric heterochromatin structure, altered telomerelength homeostasis, and activation of ALT [63].Along the clonal pseudo-polycentrics, recombinant chromosomes

with large ITRs, due to increased telomere recombinogenicity, mayalso represent typical cytogenetic markers of the ALT karyotype.Similar to variable in size, species-specific microscopically visibleITRs, observed in hamsters, pigs, and muntjacs [64–66], ALT ITRsremained relatively stable for several PDs and did not impair clonalperpetuation of ITR-carrier chromosomes.Oncogene-induced replication stress affects DNA integrity at

numerous CFSs within the cancer cell genome [11,67]. The greatextent of recombinogenicity at CFS, observed in the cell lines of ourpanel, indicates that in ALT cells extreme replication stress affects thewhole genome. Another issue might be the implication of CFS genes inALT critical genomic rearrangements. Several CFS regions harbor genesthat may be involved in ALT, such as the CENPA-CAD, localized inclose proximity with FRAXB at Xq22.1 [68]. CENPA-CAD encodesa complex recruited to centromeres to facilitate incorporation ofnewly synthesized CENP-A that entails centromeric functionality[69]. Deficiency of CENPA-CAD may explain the increased propen-sity for the formation of ALT clonal pseudo-polycentric chromosomesdemonstrated in this study. Other important genes may also be affectedby recurrent losses of chromosome X, whole-arm deletions, or trans-locations. For example, mutations in ATRX, sited at Xq21.1 (recurrentlyaffected in our study) have been considered capable to activate ALT[29,68,70]. Interestingly, ATRX deficiency also disturbs sister chromatidcohesion and causes mitotic aberrations [71].Partial or full monosomies of the short arm of chromosome 1 have

been frequently reported in human malignant tumors that, in greatmajority, express telomerase activity [72,73]. The ALT pathway seemsto take advantage of negatively unbalanced genome material from botharms of this chromosome that is also proved highly recombinogenic[[29] and this study]. However, chromosome 1qter might harborimportant genes for ALT, because its tip showed a remarkable re-

sistance to clonal telomere dysfunction. Indeed, EXO1 is located at1q42-q43 [68]. EXO1 is important for generating both types of telo-merase knockdown survivors (I and II) in budding yeast [74]. Terminaldeletions, affecting a large portion of 3q material including the hTERCgene at 3q26.2, were identified in 55.5% of our ALT cell lines in-cluding VA-13 and U2-OS. In study of Christodoulidou et al., weverified the lack of hTERC expression in both VA-13 and U2-OS [38].Four of nine ALT cell lines showed large terminal deletions involv-

ing 4q31. At least three genes that might associate with ALT con-tinuous growth are located here: the transcriptional modulator ofcell growth and apoptosis SMAD1 [68,75]; SMARCA5, an SWItch/sucrose non fermentable (SWI/SNF)-related matrix-associated actin-dependent regulator of chromatin [68,76]; and the anaphase promotingcomplex subunit 10 [68]. The TERT gene is located at the tip of theshort arm of chromosome 5 (p15.33) [68,77]. Interestingly, six of ninecell lines showed variable size terminal deletions of 5p. The TERT geneis silenced in all ALT cell lines of our panel [3,38,78–80].The gene for the CENP-C binding protein, DAXX, is located at

6p21.3 [68]. DAXX has been shown to be a negative regulator ofALT [29,81]. Consistent with Lovejoy et al., deletions or unbalancedtranslocations of 6p21-22 were found in 55% of the ALT cell linesof our panel (five of nine) [29]. In proximity to DAXX, at 6p22[68], the DNA replication inhibitor Geminin (GMNN) resides, deple-tion of which might be associated to increased rates of polyploidizationthrough whole-genome endoreduplication observed in ALT [[38,52]and this study]. The same region (6p21.2) harbors RNF8 [68], anE3 ubiquitin–protein ligase gene, involved in DDR. RNF8 mediatesthe ubiquitination of histones H2A and H2AX, promoting theformation of TP53BP1 and BRCA1 ionizing radiation–induced foci[82]. The long arm of chromosome 6 displayed a recurrent break-point hotspot at q21-q23.3. This region harbors ASF1, an H3/H4chaperone that is required for the formation of senescence associatedheterochromatin foci, as well as the DNA replication licensing factorMCM9 [83]. Recurrent deletions/translocations affecting 7q22.1-7q22.3 were found in four of nine cell lines. This is the site for MLL5gene, responsible for a histone methyltransferase that specifically mono-methylates and dimethylates “Lys-4” of histone H3 (H3K4me1 andH3K4me2) [68,84].Losses of whole chromosome 8 and/or 8p arm deletions were

observed in seven of nine ALT cell lines. The WRN helicase geneis located at 8p12 [68]. Despite its involvement in APBs, WRN isdispensable for ALT [22]. Another gene that might be related tothe formation of ALT pseudo-polycentrics is the ESCO2, locatedat 8p21.1, that is important for sister chromatid cohesion [68,85].PIN2/TERF1 interacting, telomerase inhibitor 1 (PINX1) at 8p23.1 is amicrotubule-binding protein essential for faithful chromosome segrega-tion [68,86]. PINX1mediates telomeric repeat binding factor 1 (TRF1)and TERT accumulation in nucleoli and enhances TRF1 binding totelomeres [87]. Tankyrase, a telomere interacting factor related tomitotic fidelity and telomere integrity, is also located at 8p23.1 [68,88].The telomeric region of 9p was involved in the structural rearrange-

ments in four of nine cell lines. The SMARCA gene that encodesa protein participating in the large ATP-dependent chromatin re-modeling complex SNF/SWI is located at 9p24.3 [68,89]. Dele-tions and/or translocations affecting the short arm of chromosome 9clustered between bands p21.2 and p21.3. This is the genomic sitefor the tumor suppressor cyclin-dependent kinase inhibitor 2A(p16) [68,90]. The cyclin-dependent kinase inhibitor 2A locus regu-lates the two tumor-suppressive pathways (Rb and p53) that are


most commonly disrupted in a wide range of human malignancies[90,91]. The product of p16 induces cell cycle arrest by preventingphosphorylation of pRb. Loss of function of p16 is reported tobe involved in immortalization as an alternative to inactivation ofpRb [91–93].The tip of the short arm of chromosome 11 was involved in clonal

translocations or fusions in 66% of the nine lines. The HRAS onco-gene is located in close proximity to 11pter [68]. This region harborsalso insulin-like growth factor 2 growth. It is an imprinted gene, ex-pressed only from the paternal allele, and epigenetic changes at thislocus are associated with rhabdomyosarcomas [90,94,95]. The longarm of chromosome 11 showed recurrent involvement in deletionsor translocations at bands q23.1 to q23.3; the MLL1 gene is locatedhere [68]. MLL1 is another histone methyltransferase implicated intelomere metabolism [92]. MLL1 depletion in human diploid fibro-blasts resulted in reduced levels of telomere H3/K4 methylation, theinduction of a DDR at telomeres, a p53-dependent growth arrest,and cellular senescence [92].Almost 50% of retinoblastomas use the ALT pathway to maintain

their telomeres [91]. The Rb1 gene localizes at 13q14.2 [68]. Recurrentlosses of whole chromosome 13 and/or deletions/translocationsinvolving 13q14.2 were evident in all cell lines of our panel. Fibro-blast-deriving human ALT cell lines immortalized either sponta-neously or by physical or chemical carcinogens, lack expression ofp16INK4a, and display hyperphosphorylation of pRb [91].The region 15q13.1 was recurrently involved in translocations

affecting different chromosomes, such as 12, 18, and 20, and ter-minal deletions. The NDNL2 gene that resides at 15q13.1 is a partof the SMC5-SMC6 complex, involved in homologous recombinationDNA repair [68,93]. The complex is required for telomere main-tenance through recombination in ALT cell lines. The same regionhosts the gene HERC2, responsible for an E3 ubiquitin–protein ligasethat regulates ubiquitin-dependent retention of DDR proteins ondamaged chromosomes [96].The great majority of known ALT cell lines is impaired in the p53

pathway, either due to the expression of viral oncoproteins or througha p53 mutation [97]. Immortalized cell lines, derived from Li-Fraumenibreast fibroblasts carrying a germline mutation in p53, use ALT for telo-mere maintenance [29,98]. Deletions and/or negatively unbalancedtranslocations affecting the region of the TP53 gene at 17p13.1 wereobserved in five of nine cell lines of our panel. In contrast, 17q mightharbor important genes for ALT because 17q was found to be unaffectedin eight of nine cell lines. The APB component BRCA1 is located at17q21.31, RECQL5 at q25, MAP3K3 at q23.3, MAPKK6 at q24.3, aswell as the topoisomerase TOP2A located at q21.2 [68,99–103].Gains of genomic material from 18p and pericentromeric rearrange-

ments of both the p- and q-arms of this chromosome were frequent inthe ALT cell lines of our study and might implicate genes such as trans-forming growth factor β–induced factor homeobox 1 and methyltransferase-like 4 at 18p11.3, whereas the 18q11.2 region harbors the BRCA1modulator retinoblastoma binding protein 8 (RBBP-8) [68,104]. Severalgenes located at 18p such as CENTRIN1, Yamaguchi sarcoma viraloncogene homolog 1 (YES1), methyltransferase-like 4, transforming growthfactor β–induced factor homeobox 1, establishment of sister chromatid cohe-sion N-acetyltransferase 1 (ESCO1), GATA-binding factor 6 (GATA6 ),or cutaneous T-cell lymphoma-associated antigen 1 (CTAGE1) may conferadvantages to the ALT phenotype [68,103].Losses of 19p and gains of 19q were observed in all lines of our

panel. The DNMT1 located at 19p13.2 may be an ALT repressor

[68]. DNA methylation is enriched at pericentromeric, centromeric,and subtelomeric regions and is regarded to undergo late replicationand to prevent frequent recombination events [105]. The product ofthe ZSCAN4 gene at the highly recombinogenic tip of 19q is responsiblefor the stochastic engagement of an ALT-like phenotype in ES and iPScells [56,64].Amplification of the long arm of chromosome 20 was observed in

44.4% of our cell line panel. The gene responsible for the immuno-deficiency, centromeric region instability, and facial anomalies syn-drome (ICF) syndrome, the DNMT3B, localizes at 20q11.21.Hypomethylation of subtelomeric and epigenetic modifications attelomeric regions was shown to affect the formation of telomere sisterchromatid exchange [106].TOP1 (20q12) and AURKA (20q13.2) mayalso be candidates for the list with important genes for ALT continuousgrowth [68,107,108].The 21q22.1-3 was involved in structural aberrations in 50% of our

cell line panel. Specific genes, like RUNX1, ERG, TMPRSS2, and TFF,located in this region have been involved in tumorigenesis [114].Moreover, Pericentrin, a cell cycle regulation protein, necessary forcentrosomal function, is located at the tip of 21q (q22.3) [68,109].Extensive loss of genomic material from chromosome 22 was verified

in six of nine lines. Chromosome 22 hosts, among many others, CHK2(q12.1), Merlin (q12.2), and EP300 (q13.2) [68]. Mutations of theCHK2 gene involved in DDR and the control of the cell cycle arefound in osteosarcomas [110]. Deletions in Merlin are responsible forneurofibromatosis 2, associated with predisposition to the develop-ment of the nerve sheath tumors schwannomas [111]. Interestingly,telomerase activity could not be detected in all schwannomas examined,implying that they might use ALT to sustain continuous growth [111].The most consistent cytogenetic change reported in benign menin-giomas is partial del(22)(q12) or total deletion of chromosome 22[112]. Many meningiomas were found to use ALT [81]. Moreover,cells without the p300 protein cannot effectively restrain growth anddivision, allowing cancerous tumors to develop and grow [81].In the highly aberrant ALT karyotypes, the identification of struc-

tural chromosomal rearrangements that might be harboring importantgenes for the ALT pathway is hampered by stochastic events driven bytelomeric, pericentromeric, or CFS vulnerability. Notably, several ofthe above recurrent breakpoints or imbalances correspond very wellto recently published SNP array data from 22 ALT cell lines (four ofthose are included in our panel) [29]. Hence, despite the relativelylower throughput and resolution, compared to whole-genome com-parative genomic hybridization (CGH) approaches, spectral karyotyping(SKY)/M-FISH/DAPI banding analysis is capable of identifying withgreat accuracy major ALT imbalances and recurrent genomic recom-binatorial hotspots.Exogenous reconstitution of telomerase activity reduced significantly

the rates of B/F/B cycles and suppressed overall CIN in the ALT VA-13cells [this study and [31,38]]. Telomeres, centromeres, and subtelo-meric regions are considered to be highly sensitive to B/F/B-induceddouble-strand breaks [113]. Our results show that a proportion ofALT pericentromeric instability may be driven by mechanical forcesoperating during B/F/B cycles. However, replication stress persists evenafter stable restoration of telomerase activity, because the proportionof random rearrangements affecting CFS in VA-13TA cells remainedsimilar or even increased compared to the isogenic ALT VA-13.In contrast to many hematologic malignancies [114,115], and to a

small number of solid tumors [114,116], our results did not reveal clearevidence for recurrent translocation-driven gene fusions in the ALT


pathway. However, a thorough examination of specific breakpointssuch as those at 15q and 17q12 that were found consistently involvedin recurrent but variable translocations may—in the near future—reveal if such entities exist.Our data are consistent to previous studies [29,117] suggesting that

the major oncosuppressor pathway abrogated in ALT cells is p53/Rb/INK4. In addition, the ALT cells take advantage of high rates ofendogenous CIN and undergo selection adaptation processes thatshape their genome toward the depletion of several regulators of theepigenome such as ATRX and DAXX [29,117].Hence, the ALT karyotype is shaped by massive stochastic chromo-

some mutation events that are selected to cluster in particular genomicsites. In this context, clonality of cancerous recombinant chromosomesis preserved through high rates of telomere seeding and frequent de-activation of extra centromeres of fused chromosomes. Our resultsshed light into the dynamics of clonal perpetuation of cancerous re-combinant chromosomes under extreme telomere dysfunction, unveil-ing, in parallel, specific genomic sites that may harbor genes critical forALT cancerous cell growth.

AcknowledgmentsWe thank J. Shay for providing the VA-13+hTERC+hTERT cells.We also thank all cell line donators, as indicated in the Materials andMethods section, as well as G. Grigoriadis for continuous supportand valuable discussions.

References[1] Halazonetis TD, Gorgoulis VG, and Bartek J (2008). An oncogene-induced

DNA damage model for cancer development. Science 319, 1352–1355.[2] Kazda A, Zellinger B, Rössler M, Derboven E, Kusenda B, and Riha K

(2012). Chromosome end protection by blunt-ended telomeres. Genes Dev 26,1703–1713.

[3] Gagos S and Irminger-Finger I (2005). Chromosome instability in neoplasia:chaotic roots to continuous growth. Int J Biochem Cell Biol 37, 1014–1033.

[4] Roschke AV and Kirsch IR (2010). Targeting karyotypic complexity andchromosomal instability of cancer cells. Curr Drug Targets 11, 1341–1350.

[5] Eckert CA, Gravdahl DJ, and Megee PC (2007). The enhancement of pericentro-meric cohesin association by conserved kinetochore components promotes high-fidelity chromosome segregation and is sensitive to microtubule-based tension.Genes Dev 21, 278–291.

[6] Silva P, Barbosa J, Nascimento AV, Faria J, Reis R, and Bousbaa H (2011).Monitoring the fidelity of mitotic chromosome segregation by the spindle assemblycheckpoint. Cell Prolif 44, 391–400.

[7] Greaves M and Maley CC (2012). Clonal evolution in cancer. Nature 481,306–313.

[8] Stratton MR (2011). Exploring the genomes of cancer cells: progress and promise.Science 331, 1553–1558.

[9] Durkin SG and Glover TW (2007). Chromosome fragile sites. Annu Rev Genet41, 169–192.

[10] Dillon LW, Burrow AA, and Wang YH (2010). DNA instability at chromosomalfragile sites in cancer. Curr Genomics 11, 326–337.

[11] Sideridou M, Zakopoulou R, Evangelou K, Liontos M, Kotsinas A, RampakakisE, Gagos S, Kahata K, Grabusic K, Gkouskou K, et al. (2011). Cdc6 expressionrepresses E-cadherin transcription and activates adjacent replication origins.J Cell Biol 195, 1123–1140.

[12] Debatisse M, Le Tallec B, Letessier A, Dutrillaux B, and Brison O (2012).Common fragile sites: mechanisms of instability revisited. Trends Genet 28,22–32.

[13] Ma K, Qiu L, Mrasek K, Zhang J, Liehr T, Quintana LG, and Li Z (2012).Common fragile sites: genomic hotspots of DNA damage and carcinogenesis.Int J Mol Sci 13, 11974–11999.

[14] Blasco MA (2005). Telomeres and human disease: ageing, cancer and beyond.Nat Rev Genet 6, 611–622.

[15] Bryan TM, Englezou A, Gupta J, Bacchetti S, and Reddel RR (1995). Telomereelongation in immortal human cells without detectable telomerase activity.EMBO J 14, 4240–4248.

[16] Sabatier L, Ricoul M, Pottier G, and Murnane JP (2005). The loss of a singletelomere can result in instability of multiple chromosomes in a human tumorcell line. Mol Cancer Res 3, 139–150.

[17] Kim NW, Piatyszek MA, Prowse KR, Harley CB, West MD, Ho PL, CovielloGM, Wright WE, Weinrich SL, and Shay JW (1994). Specific associationof human telomerase activity with immortal cells and cancer. Science 266,2011–2015.

[18] Rizki A and Lundblad V (2001). Defects in mismatch repair promote telomerase-independent proliferation. Nature 411, 713–716.

[19] Henson JD and Reddel RR (2010). Assaying and investigating AlternativeLengthening of Telomeres activity in human cells and cancers. FEBS Lett 584,3800–3811.

[20] Shay JW, Reddel RR, and Wright WE (2012). Cancer and telomeres—anALTernative to telomerase. Science 336, 1388–1390.

[21] Daniel A, St Heaps L, Sylvester D, Diaz S, and Peters G (2008). Two mosaicterminal inverted duplications arising post-zygotically: Evidence for possibleformation of neo-telomeres. Cell Chromosome 7, 1.

[22] Cesare AJ and Reddel RR (2010). Alternative lengthening of telomeres: models,mechanisms and implications. Nat Rev Genet 11, 319–330.

[23] Nabetani A and Ishikawa F (2011). Alternative lengthening of telomeres pathway:recombination-mediated telomere maintenance mechanism in human cells.J Biochem 149, 5–14.

[24] Gagos S, Chiourea M, Christodoulidou A, Apostolou E, Raftopoulou C,Deustch S, Jefford CE, Irminger-Finger I, Shay JW, and Antonarakis SE(2008). Pericentromeric instability and spontaneous emergence of human neo-acrocentric and minute chromosomes in the alternative pathway of telomerelengthening. Cancer Res 68, 8146–8155.

[25] Gisselsson D (2005). Mitotic instability in cancer: is there method in the mad-ness? Cell Cycle 4, 1007–1010.

[26] Gorringe KL, Chin SF, Pharoah P, Staines JM, Oliveira C, Edwards PA, andCaldas C (2005). Evidence that both genetic instability and selection contribute tothe accumulation of chromosome alterations in cancer. Carcinogenesis 26, 923–930.

[27] Meeker AK, Hicks JL, Iacobuzio-Donahue CA, Montgomery EA, Westra WH,Chan TY, Ronnett BM, and De Marzo AM (2004). Telomere length abnor-malities occur early in the initiation of epithelial carcinogenesis. Clin Cancer Res10, 3317–3326.

[28] Mitelman F, Johansson B, and Mertens F (2007). The impact of translocationsand gene fusions on cancer causation. Nat Rev Cancer 7, 233–245.

[29] Lovejoy CA, Li W, Reisenweber S, Thongthip S, Bruno J, de Lange T, De S,Petrini JH, Sung PA, Jasin M, et al. (2012). Loss of ATRX, genome instability,and an altered DNA damage response are hallmarks of the alternative lengthen-ing of telomeres pathway. PLoS Genet 8, e1002772.

[30] Dessain SK, Yu H, Reddel RR, Beijersbergen RL, and Weinberg RA (2000).Methylation of the human telomerase gene CpG island. Cancer Res 60, 537–541.

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

[32] Capo-Chichi CD, Smith ER, Yang DH, Roland IH, Vanderveer L, Cohen C,Hamilton TC, Godwin AK, and Xu XX (2002). Dynamic alterations of theextracellular environment of ovarian surface epithelial cells in premalignant trans-formation, tumorigenicity, and metastasis. Cancer 95, 1802–1815.

[33] ISCN (2009). An International System for Human Cytogenetic Nomenclature.S. Karger, Basel, Switzerland.

[34] Campbell LJ (2011). Cytogenetic analysis and reporting. Methods Mol Biol730, 259–268.

[35] Bailey SM, Cornforth MN, Kurimasa A, Chen DJ, and Goodwin EH (2001).Strand-specific postreplicative processing of mammalian telomeres. Science 293,2462–2465.

[36] Fabarius A, Hehlmann R, and Duesberg PH (2003). Instability of chromosomestructure in cancer cells increases exponentially with degrees of aneuploidy.Cancer Genet Cytogenet 143, 59–72.

[37] Li R, Hehlman R, Sachs R, and Duesberg P (2005). Chromosomal alterationscause the high rates and wide ranges of drug resistance in cancer cells. CancerGenet Cytogenet 163, 44–56.

[38] Christodoulidou A, Raftopoulou C, Chiourea M, Papaioannou GK, HoshiyamaH, Wright WE, Shay JW, and Gagos S (2013). The roles of telomerase in thegeneration of polyploidy during neoplastic cell growth. Neoplasia 15, 156–168.


[39] Cancer Genomics Program. SKY Karyotypes and FISH Analysis of Epithelial CancerCell Lines Database. Available at: http://www.path.cam.ac.uk/~pawefish/index.html.

[40] Kawai K, Viars C, Arden K, Tarin D, Urquidi V, and Goodison S (2002).Comprehensive karyotyping of the HT-29 colon adenocarcinoma cell line.Genes Chromosomes Cancer 34, 1–8.

[41] Makridakis M, Gagos S, Petrolekas A, Roubelakis MG, Bitsika V, StravodimosK, Pavlakis K, Anagnou NP, Coleman J, and Vlahou A (2009). Chromosomaland proteome analysis of a new T24-based cell line model for aggressive bladdercancer. Proteomics 9, 287–298.

[42] Macville M, Schröck E, Padilla-Nash H, Keck C, Ghadimi BM, Zimonjic D,PopescuN, and Ried T (1999). Comprehensive and definitive molecular cytogeneticcharacterization of HeLa cells by spectral karyotyping. Cancer Res 59, 141–150.

[43] Vitre BD and Cleveland DW (2012). Centrosomes, chromosome instability(CIN) and aneuploidy. Curr Opin Cell Biol 24, 809–815.

[44] Desmaze C, Soria JC, Freulet-Marriere MA, Mathieu N, and Sabatier L (2003).Telomere-driven genomic instability in cancer cells. Cancer Lett 194, 173–182.

[45] Azzalin CM, Nergadze SG, and Giulotto E (2001). Human intrachromosomaltelomeric-like repeats: sequence organization and mechanisms of origin.Chromosoma 110, 75–82.

[46] Stimpson KM, Matheny JE, and Sullivan BA (2012). Dicentric chromosomes:unique models to study centromere function and inactivation. Chromosome Res20, 595–605.

[47] Black BE and Cleveland DW (2011). Epigenetic centromere propagation andthe nature of CENP-a nucleosomes. Cell 144, 471–479.

[48] Sfeir A, Kosiyatrakul ST, Hockemeyer D, MacRae SL, Karlseder J, SchildkrautCL, and de Lange T (2009). Mammalian telomeres resemble fragile sites andrequire TRF1 for efficient replication. Cell 138, 90–103.

[49] Gerlinger M, Rowan AJ, Horswell S, Larkin J, Endesfelder D, Gronroos E,Martinez P, Matthews N, Stewart A, Tarpey P, et al. (2012). Intratumor hetero-geneity and branched evolution revealed by multiregion sequencing. N Engl J Med366, 883–892.

[50] Podlaha O, Riester M, De S, and Michor F (2012). Evolution of the cancergenome. Trends Genet 28, 155–163.

[51] Scheel C, Schaefer KL, Jauch A, Keller M, Wai D, Brinkschmidt C, van ValenF, Boecker W, Dockhorn-Dworniczak B, and Poremba C (2001). Alternativelengthening of telomeres is associated with chromosomal instability in osteo-sarcomas. Oncogene 20, 3835–3844.

[52] Davoli T, Denchi EL, and de Lange T (2010). Persistent telomere damageinduces bypass of mitosis and tetraploidy. Cell 141, 81–93.

[53] Gisselsson D, Pettersson L, Höglund M, Heidenblad M, Gorunova L, Wiegant J,Mertens F, Dal Cin P, Mitelman F, and Mandahl N (2000). Chromosomalbreakage-fusion-bridge events cause genetic intratumor heterogeneity. Proc NatlAcad Sci USA 97, 5357–5362.

[54] Martínez AC and van Wely KH (2010). Are aneuploidy and chromosomebreakage caused by a CINgle mechanism? Cell Cycle 9, 2275–2280.

[55] Erenpreisa J and Cragg MS (2010). MOS, aneuploidy and the ploidy cycle ofcancer cells. Oncogene 29, 5447–5451.

[56] Zalzman M, Falco G, Sharova LV, Nishiyama A, Thomas M, Lee SL, Stagg CA,Hoang HG, Yang HT, Indig FE, et al. (2010). Zscan4 regulates telomereelongation and genomic stability in ES cells. Nature 464, 858–863.

[57] Wang F, Yin Y, Ye X, Liu K, Zhu H, Wang L, Chiourea M, Okuka M, Ji G,Dan J, et al. (2012). Molecular insights into the heterogeneity of telomerereprogramming in induced pluripotent stem cells. Cell Res 22, 757–768.

[58] Jaco I, Canela A, Vera E, and Blasco MA (2008). Centromere mitotic recom-bination in mammalian cells. J Cell Biol 181, 885–892.

[59] Klinakis A, Lobry C, Abdel-Wahab O, Oh P, Haeno H, Buonamici S, van DeWalle I, Cathelin S, Trimarchi T, Araldi E, et al. (2011). A novel tumour-suppressor function for the Notch pathway in myeloid leukaemia. Nature 473,230–233.

[60] Gonzalvez F, Lawrence D, Yang B, Yee S, Pitti R, Marsters S, Pham VC,Stephan JP, Lill J, and Ashkenazi A (2012). TRAF2 sets a threshold for extrinsicapoptosis by tagging caspase-8 with a ubiquitin shutoff timer. Mol Cell 48,888–899.

[61] Aiyar SE, Cho H, Lee J, and Li R (2007). Concerted transcriptional regulationby BRCA1 and COBRA1 in breast cancer cells. Int J Biol Sci 3, 486–492.

[62] Chabchoub E, Rodriguez L, Galán E, Mansilla E, Martínez-Fernandez ML,Martínez-Frías ML, Fryns JP, and Vermeesch JR (2007). Molecular charac-terisation of a mosaicism with a complex chromosome rearrangement: evidencefor coincident chromosome healing by telomere capture and neo-telomereformation. J Med Genet 44, 250–256.

[63] Gonzalo S, Jaco I, Fraga MF, Chen T, Li E, Esteller M, and Blasco MA (2006).DNA methyltransferases control telomere length and telomere recombination inmammalian cells. Nat Cell Biol 8, 416–424.

[64] Ji G, Liu K, Chen C, Ruan W, Glytsou C, Yang Y, Okuka M, Song W, GagosS, Li N, et al. (2012). Conservation and characterization of unique porcineinterstitial telomeric sequences. Sci China Life Sci 55, 1029–1037.

[65] Solovjeva LV, Demin SJ, Pleskach NM, Kuznetsova MO, and Svetlova MP(2012). Characterization of telomeric repeats in metaphase chromosomes andinterphase nuclei of Syrian Hamster Fibroblasts. Mol Cytogenet 5, 37.

[66] Lee C, Sasi R, and Lin CC (1993). Interstitial localization of telomeric DNAsequences in the Indian muntjac chromosomes: further evidence for tandemchromosome fusions in the karyotypic evolution of the Asian muntjacs. CytogenetCell Genet 63, 156–159.

[67] Tsantoulis PK, Kotsinas A, Sfikakis PP, Evangelou K, Sideridou M, Levy B, MoL, Kittas C, Wu XR, Papavassiliou AG, et al. (2008). Oncogene-inducedreplication stress preferentially targets common fragile sites in preneoplasticlesions. A genome-wide study. Oncogene 27, 3256–3264.

[68] GeneCards: The Human Gene Compendium. Available at: www.genecards.org.[69] Carroll CW, Silva MC, Godek KM, Jansen LE, and Straight AF (2009).

Centromere assembly requires the direct recognition of CENP-A nucleosomesby CENP-N. Nat Cell Biol 11, 896–902.

[70] Khuong-Quang DA, Buczkowicz P, Rakopoulos P, Liu XY, Fontebasso AM,Bouffet E, Bartels U, Albrecht S, Schwartzentruber J, Letourneau L, et al.(2012). K27M mutation in histone H3.3 defines clinically and biologically dis-tinct subgroups of pediatric diffuse intrinsic pontine gliomas. Acta Neuropathol124, 439–447.

[71] Ritchie K, Seah C, Moulin J, Isaac C, Dick F, and Bérubé NG (2008). Loss ofATRX leads to chromosome cohesion and congression defects. J Cell Biol 180,315–324.

[72] Dong Z, Pang JS, Ng MH, Poon WS, Zhou L, and Ng HK (2004). Identifica-tion of two contiguous minimally deleted regions on chromosome 1p36.31-p36.32 in oligodendroglial tumours. Br J Cancer 91, 1105–1111.

[73] Idbaih A, Marie Y, Pierron G, Brennetot C, Hoang-Xuan K, Kujas M, MokhtariK, Sanson M, Lejeune J, Aurias A, et al. (2005). Two types of chromosome 1plosses with opposite significance in gliomas. Ann Neurol 58, 483–487.

[74] Maringele L and Lydall D (2004). EXO1 plays a role in generating type I andtype II survivors in budding yeast. Genetics 166, 1641–1649.

[75] Massagué J, Seoane J, and Wotton D (2005). Smad transcription factors. GenesDev 19, 2783–2810.

[76] Gigek CO, Lisboa LC, Leal MF, Silva PN, Lima EM, Khayat AS, Assumpção PP,Burbano RR, and Smith Mde A (2011). SMARCA5 methylation and expressionin gastric cancer. Cancer Invest 29, 162–166.

[77] Shay JW and Wright WE (2000). Implications of mapping the human telomerasegene (hTERT) as the most distal gene on chromosome 5p. Neoplasia 2, 195–196.

[78] Hoare SF, Bryce LA, Wisman GB, Burns S, Going JJ, van der Zee AG, andKeith WN (2001). Lack of telomerase RNA gene hTERC expression in alterna-tive lengthening of telomeres cells is associated with methylation of the hTERCpromoter. Cancer Res 61, 27–32.

[79] Atkinson SP, Hoare SF, Glasspool RM, and Keith WN (2005). Lack of telo-merase gene expression in alternative lengthening of telomere cells is associatedwith chromatin remodeling of the hTR and hTERT gene promoters. Cancer Res65, 7585–7590.

[80] Zhang Y, Cai L, Wei RX, Hu H, Jin W, and Zhu XB (2011). Different expres-sion of alternative lengthening of telomere (ALT)-associated proteins/mRNAsin osteosarcoma cell lines. Oncol Lett 2, 1327–1332.

[81] Heaphy CM, Subhawong AP, Hong SM, Goggins MG, Montgomery EA,Gabrielson E, Netto GJ, Epstein JI, Lotan TL, Westra WH, et al. (2011).Prevalence of the alternative lengthening of telomeres telomere maintenancemechanism in human cancer subtypes. Am J Pathol 179, 1608–1615.

[82] Peuscher MH and Jacobs JJ (2011). DNA-damage response and repair activitiesat uncapped telomeres depend on RNF8. Nat Cell Biol 13, 1139–1145.

[83] Zhang R, Chen W, and Adams PD (2007). Molecular dissection of formationof senescence-associated heterochromatin foci. Mol Cell Biol 27, 2343–2358.

[84] Sebastian S, Sreenivas P, Sambasivan R, Cheedipudi S, Kandalla P, Pavlath GK,and Dhawan J (2009). MLL5, a trithorax homolog, indirectly regulates H3K4methylation, represses cyclin A2 expression, and promotes myogenic differentia-tion. Proc Natl Acad Sci USA 106, 4719–4724.

[85] Whelan G, Kreidl E, Wutz G, Egner A, Peters JM, and Eichele G (2012).Cohesin acetyltransferase Esco2 is a cell viability factor and is required forcohesion in pericentric heterochromatin. EMBO J 31, 71–82.


[86] Yuan K, Li N, Jiang K, Zhu T, Huo Y, Wang C, Lu J, Shaw A, Thomas K,Zhang J, et al. (2009). PinX1 is a novel microtubule-binding protein essentialfor accurate chromosome segregation. J Biol Chem 284, 23072–23082.

[87] Yonekawa T, Yang S, and Counter CM (2012). PinX1 localizes to telomeresand stabilizes TRF1 at mitosis. Mol Cell Biol 32, 1387–1395.

[88] Ha GH, Kim HS, Go H, Lee H, Seimiya H, Chung DH, and Lee CW (2012).Tankyrase-1 function at telomeres and during mitosis is regulated by Polo-likekinase-1-mediated phosphorylation. Cell Death Differ 19, 321–332.

[89] Sudarsanam P, Cao Y, Wu L, Laurent BC, and Winston F (1999). The nucleo-some remodeling complex, Snf/Swi, is required for the maintenance of tran-scription in vivo and is partially redundant with the histone acetyltransferase,Gcn5. EMBO J 18, 3101–3106.

[90] Krskova L, Augustinakova A, Drahokoupilova E, Sumerauer D, Mudry P, andKodet R (2011). Rhabdomyosarcoma: molecular analysis of Igf2, MyoD1 andMyogenin expression. Neoplasma 58, 415–423.

[91] Tsutsui T, Kumakura S, Yamamoto A, Kanai H, Tamura Y, Kato T, Anpo M,Tahara H, and Barrett JC (2002). Association of p16INK4a and pRb inactiva-tion with immortalization of human cells. Carcinogenesis 23, 2111–2117.

[92] Caslini C, Connelly JA, Serna A, Broccoli D, and Hess JL (2009). MLL asso-ciates with telomeres and regulates telomeric repeat-containing RNA trans-cription. Mol Cell Biol 29, 4519–4526.

[93] De Piccoli G, Cortes-Ledesma F, Ira G, Torres-Rosell J, Uhle S, Farmer S,Hwang JY, Machin F, Ceschia A, McAleenan A, et al. (2006). Smc5-Smc6mediate DNA double-strand-break repair by promoting sister-chromatid recom-bination. Nat Cell Biol 8, 1032–1034.

[94] Rezvani G, Lui JC, Barnes KM, and Baron J (2012). A set of imprinted genesrequired for normal body growth also promotes growth of rhabdomyosarcomacells. Pediatr Res 71, 32–38.

[95] Martins AS, Olmos D, Missiaglia E, and Shipley J (2011). Targeting the insulin-like growth factor pathway in rhabdomyosarcomas: rationale and future perspec-tives. Sarcoma 2011, 209736.

[96] Bekker-Jensen S, Danielsen RendtlewJ, Fugger K, Gromova I, Nerstedt A,Lukas C, Bartek J, Lukas J, and Mailand N (2010). HERC2 coordinatesubiquitin-dependent assembly of DNA repair factors on damaged chromosomes.Nat Cell Biol 12, 80–86; sup. pp. 1–12.

[97] Henson JD, Neumann AA, Yeager TR, and Reddel RR (2002). Alternativelengthening of telomeres in mammalian cells. Oncogene 21, 598–610.

[98] Rogan EM, Bryan TM, Hukku B, Maclean K, Chang AC, Moy EL, Englezou A,Warneford SG, Dalla-Pozza L, and Reddel RR (1995). Alterations in p53 andp16INK4 expression and telomere length during spontaneous immortalizationof Li-Fraumeni syndrome fibroblasts. Mol Cell Biol 15, 4745–4753.

[99] Couch FJ,Wang X,McGuffog L, Lee A,Olswold C,Kuchenbaecker KB, Soucy P,Fredericksen Z, Barrowdale D, Dennis J, et al. (2013). Genome-wide associationstudy in BRCA1 mutation carriers identifies novel loci associated with breast andovarian cancer risk. PLoS Genet 9, e1003212.

[100] Ohhata T, Araki R, Fukumura R, Kuroiwa A, Matsuda Y, and Abe M (2001).Cloning, genomic structure and chromosomal localization of the gene encodingmouse DNA helicase RECQL5β. Gene 280, 59–66.

[101] Marchiò C, Natrajan R, Shiu KK, Lambros MB, Rodriguez-Pinilla SM, Tan DS,Lord CJ, Hungermann D, Fenwick K, Tamber N, et al. (2008). The genomic

profile of HER2-amplified breast cancers: the influence of ER status. J Pathol216, 399–407.

[102] Mao X, Lillington D, Child F, Russell-Jones R, Young B, and Whittaker S(2002). Comparative genomic hybridization analysis of primary cutaneousB-cell lymphomas: identification of common genomic alterations in diseasepathogenesis. Genes Chromosomes Cancer 35, 144–155.

[103] National Center for Biotechnology Information (NCBI), Gene Database. Availableat: www.ncbi.nlm.nih.gov/gene.

[104] Höglund M, Gorunova L, Jonson T, Dawiskiba S, Andrén-Sandberg A,Stenman G, and Johansson B (1998). Cytogenetic and FISH analyses ofpancreatic carcinoma reveal breaks in 18q11 with consistent loss of 18q12-qterand frequent gain of 18p. Br J Cancer 77, 1893–1899.

[105] Schoeftner S and Blasco MA (2009). A ‘higher order’ of telomere regulation:telomere heterochromatin and telomeric RNAs. EMBO J 28, 2323–2336.

[106] Yehezkel S, Segev Y, Viegas-Péquignot E, Skorecki K, and Selig S (2008).Hypomethylation of subtelomeric regions in ICF syndrome is associated withabnormally short telomeres and enhanced transcription from telomeric regions.Hum Mol Genet 17, 2776–2789.

[107] Sillars-Hardebol AH, Carvalho B, Tijssen M, Belien JA, de Wit M, Delis-vanDiemen PM, Pontén F, van de Wiel MA, Fijneman RJ, and Meijer GA(2012). TPX2 and AURKA promote 20q amplicon-driven colorectal adenomato carcinoma progression. Gut 61, 1568–1575.

[108] Smith DH, Christensen IJ, Jensen NF, Markussen B, Romer MU, Nygard SB,Muller S, Nielsen HJ, Brunner N, and Nielsen KV (2013). Mechanisms oftopoisomerase I (TOP1) gene copy number increase in a stage III colorectalcancer patient cohort. PLoS One 8, e60613.

[109] Delaval B and Doxsey SJ (2010). Pericentrin in cellular function and disease.J Cell Biol 188, 181–190.

[110] Miller CW, Ikezoe T, Krug U, HofmannWK, Tavor S, Vegesna V, Tsukasaki K,Takeuchi S, and Koeffler HP (2002). Mutations of the CHK2 gene are foundin some osteosarcomas, but are rare in breast, lung, and ovarian tumors. GenesChromosomes Cancer 33, 17–21.

[111] Chen HJ, Cho CL, Liang CL, Lu K, and Lin JW (2002). Implication oftelomere length as a proliferation-associated marker in schwannomas. J SurgOncol 81, 93–100; discussion 100.

[112] Bayani J, Pandita A, and Squire JA (2005). Molecular cytogenetic analysis inthe study of brain tumors: findings and applications. Neurosurg Focus 19, E1.

[113] Murnane JP (2010). Telomere loss as a mechanism for chromosome instabilityin human cancer. Cancer Res 70, 4255–4259.

[114] Shugay M, Ortiz de Mendibil I, Vizmanos JL, and Novo FJ (2012). Genomichallmarks of genes involved in chromosomal translocations in hematologicalcancer. PLoS Comput Biol 8, e1002797.

[115] Mills KD, Ferguson DO, and Alt FW (2003). The role of DNA breaks ingenomic instability and tumorigenesis. Immunol Rev 194, 77–95.

[116] Nambiar M, Kari V, and Raghavan SC (2008). Chromosomal translocationsin cancer. Biochim Biophys Acta 1786, 139–152.

[117] Bower K, Napier CE, Cole SL, Dagg RA, Lau LM, Duncan EL, Moy EL, andReddel RR (2012). Loss of wild-type ATRX expression in somatic cell hybridssegregates with activation of Alternative Lengthening of Telomeres. PLoS One7, e50062.


alternative lengthening of telomeres: recurrent ... · cancer genome evolution that shapesthe...

Documents