Telomerase abrogates aneuploidy-induced telomere replication stress, senescence and cell depletion

The causal role of aneuploidy in cancer initiation remains under debate since mutations of euploidy-controlling genes reduce cell fitness but aneuploidy strongly associates with human cancers. Telomerase activation allows immortal growth by stabilizing telomere length, but its role in aneuploidy survival has not been characterized. Here, we analyze the response of primary human cells and murine hematopoietic stem cells (HSCs) to aneuploidy induction and the role of telomeres and the telomerase in this process. The study shows that aneuploidy induces replication stress at telomeres leading to telomeric DNA damage and p53 activation. This results in p53/Rb-dependent, premature senescence of human fibroblast, and in the depletion of hematopoietic cells in telomerase-deficient mice. Endogenous telomerase expression in HSCs and enforced expression of telomerase in human fibroblasts are sufficient to abrogate aneuploidy-induced replication stress at telomeres and the consequent induction of premature senescence and hematopoietic cell depletion. Together, these results identify telomerase as an aneuploidy survival factor in mammalian cells based on its capacity to alleviate telomere replication stress in response to aneuploidy induction.


Introduction
Although being present in the majority of human cancers, the functional role of aneuploidy in tumorigenesis is still poorly understood. In humans and mice, most autosomal monosomies and trisomies are not compatible with organismal survival (Magnuson et al, 1985;Munne et al, 2004), and it was shown that aneuploidy impairs cellular proliferation, survival, and spontaneous immortalization (Segal & McCoy, 1974;Castedo et al, 2004;Zhivotovsky & Kroemer, 2004;Williams et al, 2008). In contrast to its negative effects on cell viability and immortalization, aneuploidy is a hallmark of cancer cells (Boveri, 2008;Gordon et al, 2012). Recent bioinformatic analysis revealed evidence that the number of genes that drive carcinogenesis is higher than previously thought and aneuploidy appears to be instrumental to shape the complexity of cancer genomes affecting multiple cancer driver genes (Davoli et al, 2013).
Studies on genetic mouse models provided experimental evidence that mutation in genes that control chromosome segregation promotes cancer formation by increasing loss of heterozygosity , but the effect on carcinogenesis was context dependent and aneuploidy suppressed tumorigenesis in some mouse strains (Weaver et al, 2007;Schvartzman et al, 2010). Genetic studies in yeast identified genes that influence cell survival in response to aneuploidy (Torres et al, 2010). Aneuploidy survival factors in human cells and its role in tumor formation remain yet to be defined. It is known that proliferation of aneuploid tumor cells is limited by the activation of ATM/p53-dependent DNA damage checkpoints (Li et al, 2010). However, mechanistically, it remains poorly understood how aneuploidy triggers DNA damage responses. Here, we characterized DNA damage response induced by aneuploidy. The study shows that aneuploidy induces telomere replication stress resulting in DNA damage responses, premature senescence, and hematopoietic stem cell depletion in mice. Telomerase is identified as an aneuploidy survival factor that increases cell viability in response to aneuploidy induction by alleviating telomere replication stress.

Results
Telomerase expression rescues aneuploidy-induced senescence Current data suggest that a variety of cellular stress factors can lead to chromosome mis-segregation (Duijf & Benezra, 2013;Orr & Compton, 2013). Supported by recent iRNA screens on focus set libraries (Conery & Harlow, 2010), the number of aneuploidyinducing gene mutations is higher than previously thought and involves a variety of cellular signaling pathways. In iRNA screen, we identified a set of gene knockdowns that increase aneuploidy in the human cancer cell line U2OS (Meena, Günes and Rudolph in preparation). Here, we analyzed the cellular responses to the knockdown of the top 3 enriched candidate genes (GJB3, RXFP1-also known as LGR7-and OSBPL3) in two primary human fibroblast cell lines (BJ and IMR90) as well as in murine hematopoietic stem and progenitor cells.
The top 3 candidate genes have not been directly associated with aneuploidy induction, but associations with cancer formation and ploidy-affecting pathways were reported. Specifically, OSBPL3 (ORP3) was described as an R-Ras interaction partner regulating cell adhesion (Lehto et al, 2008). Knockdown of RXFP1 (relaxin family peptide receptor 1) was shown to suppress the proliferation of prostate cancer cells (Feng et al, 2010). In addition, RXFP1 activates a wide spectrum of signaling pathways to generate second messengers, including cAMP and nitric oxide-two known signaling pathways influencing ploidy control (see Bathgate et al, 2013). Mutations in the connexin gene Cx31 (encoding the GJB3 protein) have been associated with deafness and erythrokeratodermia variabilis (EKV), a skin disorder accompanied by hyperproliferation of the skin (Richard et al, 1998). A close homolog of GJB3, the GJB5 (Cx31.1), acts as a tumor suppressor in NSCLC (Zhang et al, 2012), and connexin 32 (Cx32) knockout mice exhibit increased chemical-and radiationinduced liver and lung tumor formation (King & Lampe, 2004a,b).
In addition to these three top candidate genes, two control genes with a known role in ploidy control were included: (i) STARD9, a recently identified regulator of spindle pole assembly (Torres et al, 2011) that was also found in our screen to aggravate aneuploidy of human cancer cells and (ii) MAD2L1 (referred to as MAD2 hereafter), a well-characterized component of the spindle-assembly checkpoint that prevents the onset of anaphase until all chromosomes are properly aligned at the metaphase plate (Li & Benezra, 1996). To exclude off-target effects, two shRNAs were used per gene and the knockdown efficiencies of the shRNAs were verified by qPCR ( Supplementary Fig S1A and B). The knockdown of the candidate genes in human primary cells led to a significant increase in chromosomal aberrations (Fig 1A, Supplementary Fig S1C and D). Chromosome banding analysis revealed that most of the aberrations were whole chromosome gains and losses (Fig 1B and Supplementary Table S1).
Aneuploidy induced by candidate gene-targeting shRNAs led to a premature loss of proliferation potential compared to control fibroblasts infected with a scrambled shRNA ( Fig 1C). Cessation of proliferation of fibroblasts infected with aneuploidy-inducing shRNAs was characterized by senescence morphology, senescence-associated b-galactosidase activity , phosphorylation of p53 and p38 (Fig 1D), and induction of p21 expression (Fig 1D and E,Supplementary Fig S2H and J). Combined inactivation of the p53 and Rb pathways by the SV-40 ST/LT antigens abrogated the induction of premature senescence and led to an increase in the number of chromosomal gains and losses in primary human cells that were targeted by aneuploidy-inducing shRNAs (Fig 1F and G).
Together, these results indicated that the knockdown of candidate genes from our screen led to aneuploidy in primary human fibroblasts, but this resulted in p53/Rb-dependent checkpoint activation and premature senescence-a phenomenon that was previously described as aneuploidy-induced senescence (AIS) (Humbert et al, 2010;Lentini et al, 2012). Of note, the knockdown of MAD2 induced a stronger phenotype characterized by a more rapid growth arrest ( Fig 1C) and multinucleated cells (Supplementary Fig S2A). These effects stand in line with previous reports (Lentini et al, 2012). Due to this rapid proliferation arrest, it was not possible to prepare metaphase spreads from MAD2-deficient primary human fibroblasts. Therefore, the subsequent analyses focused on the analysis of aneuploidy-inducing candidate shRNAs from our screen that induced aneuploidy but a delayed growth arrest.
Primary human fibroblasts lack telomerase expression, and telomere shortening limits the proliferative lifespan of human fibroblasts to 50-70 cell divisions. It was shown that telomerase expression is sufficient to immortalize human fibroblasts by stabilizing telomere length, thus preventing the induction of DNA damage responses at dysfunctional telomeres in late passages of human fibroblasts (Bodnar et al, 1998;d'Adda di Fagagna et al, 2003). To test whether telomerase expression would influence the induction of AIS, shRNAs against GJB3, RXFP1, OSBPL3 and STARD9 were infected in BJ and IMR90 fibroblast cell lines that over-expressed telomerase reverse transcriptase (hTERT)-the limiting component for telomerase activity in human cells (Weinrich et al, 1997). In both fibroblast cell lines, TERT expression fully prevented the induction of DNA damage checkpoints and AIS in response to the infection of aneuploidy-inducing shRNAs (Fig 1I and Supplementary Fig S2E,G,I and K). Of note, TERT-expressing human fibroblasts developed aberrations in chromosomal copy numbers to a similar extent as TERT-negative HFs in response to the knockdown of the candidate genes (Fig 1A versus H). The suppression of AIS in response to the expression of aneuploidy-inducing shRNAs in human cells was dependent on the enzymatic activity of telomerase as a catalytically dead mutant of TERT (Hahn et al, 1999) was not able to rescue growth arrest and senescence in the same setting (Supplementary Fig S2L).
Together, the above result indicated that single-gene mutation can lead to aneuploidy in primary human cells, but aneuploid cells can only survive and continue to proliferate in the presence of telomerase or when p53/Rb-dependent checkpoints are defective.

Telomerase activity suppresses aneuploidy-induced telomere replication stress
The primary function of telomerase is to compensate for the loss of telomeric DNA in response to cell proliferation (Bodnar et al, 1998).
The EMBO Journal Telomerase as an aneuploidy survival factor Jitendra K Meena et al It was also shown that telomerase activity prevents the accumulation of DNA damage and the induction of senescence in response to oncogene activation (Suram et al, 2012). To test whether aneuploidy induction would lead to accumulation of DNA damage and whether this would be influenced by telomerase gene status, the activation of DNA damage responses (DDR) was measured in telomerase-positive and telomerase-negative human fibroblasts infected with aneuploidy-inducing shRNAs. In agreement with the AIS phenotype ( Early passages of primary human fibroblasts (BJ and IMR90) or hTERT-transduced fibroblasts (BJ-hTERT and IMR90-hTERT) were infected with aneuploidy-inducing shRNAs or a scrambled shRNA.
A The dot plot shows chromosome numbers of metaphase spreads from BJ cells that were infected with shRNAs targeting the indicated candidate genes (n = 20 metaphases per construct, P-values were calculated against the control scramble construct). B Representative example of Giemsa-banding showing a gain of chromosome 11 in a metaphase of RXFP1 shRNA-infected BJ-hTERT cells. C Growth curves of BJ and IMR90 fibroblasts that were not infected or infected with shRNAs targeting the indicated genes. The graphs depict mean values AE SEM.
P-values were calculated by one-way ANOVA comparing pooled data for non-infected/scrambled shRNA-infected cells with pooled data of cells infected with shRNAs against candidate genes (three independent experiments per construct). D Western blot analysis of phospho-p38, phospho-p53, p21 and GAPDH in protein lysates from early passage (passage 2) BJ cells infected with shRNAs against the indicated candidate genes. E Immunofluorescence staining for p21. Representative images of BJ cells (scale bars: 50 lm). See Supplementary Fig S2H-K for quantification. F Growth curves of IMR90-ST/LT fibroblasts that were not infected or infected with shRNAs targeting the indicated genes. The graphs depict mean values AE SEM.
P-values were calculated by one-way ANOVA comparing pooled data for non-infected/scrambled shRNA-infected cells with pooled data of cells infected with shRNAs against candidate genes (two independent experiments per construct). G Chromosome numbers of metaphase spreads from IMR90-ST/LT cells that were infected with shRNAs targeting the indicated candidate genes (n = 20 metaphases per construct). H The dot plot shows chromosome numbers of metaphase spreads from BJ-hTERT cells treated and analyzed as described in (A). I Growth curves of BJ-hTERT and IMR90-hTERT fibroblast treated and analyzed as described in (C). To explore the involvement of telomeres in the accumulation of DNA damage in human fibroblasts in response to shRNA-mediated aneuploidy induction, telomere length was determined by quantitative fluorescence in situ hybridization (qFISH) and telomere stability was analyzed in metaphase spreads using a telomerespecific FISH probe (fluorescence-labeled TTTAGG 3 PNA probe). The EMBO Journal Telomerase as an aneuploidy survival factor Jitendra K Meena et al Telomere length was not affected by aneuploidy-inducing shRNAs in primary fibroblasts (Supplementary Fig S3E). However, increased numbers of fragile telomeres were observed in telomerase-negative BJ and IMR90 fibroblast cell lines in response to the transduction of aneuploidy-inducing shRNAs (Fig 2E and H). Specifically, all four selected candidate shRNAs led to an increase in the number of chromosome ends exhibiting multiple telomere signals-a marker for telomere replication stress (van Steensel et al, 1998;Sfeir et al, 2009). This increase in multi-telomeric signals was not observed in telomerase-positive human fibroblasts infected with the same aneuploidy-inducing shRNAs (Fig 2F and G). Anaphase bridges represent a hallmark feature of chromosomal fusions that form in response to: (i) telomere shortening (Rudolph et al, 2001), (ii) mutations in telomeric DNA (Kirk et al, 1997), (iii) the deletion of telomere-binding proteins (van Steensel et al, 1998), or (iv) telomere replication stress (Chan et al, 2009;Sfeir et al, 2009). Coating of anaphase bridges by the Bloom DNA helicase is indicative of DNA replication stress as a cause of chromosomal fusions. These ultrafine Bloom-containing bridges form in the context of replication stress by linking sister chromatids at fragile sites that carry intrinsic difficulties to replicate (Chan et al, 2009;Sfeir et al, 2009;Barefield & Karlseder, 2012). Our analysis revealed a significant increase in anaphase bridges (Supplementary Fig S3F and H) and in Bloom-coated anaphase bridges (Fig 2I-M) in telomerase-negative human fibroblasts expressing aneuploidy-inducing shRNAs compared to scrambled shRNA-infected cells. Of note, this phenotype was significantly ameliorated by telomerase expression (Fig 2J and K, Supplementary Fig S3G and H).
Phosphorylated RPA2 (S33-pRPA2), a target of ATR kinase, represents a hallmark feature of DNA replication stress (Vassin et al, 2009;Crasta et al, 2012). Immunofluorescence staining depicted a strong increase in nuclear pRPA2 foci in telomerasenegative human fibroblast infected with aneuploidy-inducing shRNAs compared to scrambled shRNA-infected fibroblasts (Fig 3A and C,Supplementary Fig S4A). Of note, this phenotype was completely rescued by telomerase expression (Fig 3B). The suppression of the pRPA2 foci was dependent on the enzymatic activity of telomerase since the expression of a catalytically inactive TERT mutant (DN-hTERT) did not rescue the induction of pRPA2 foci in HFs expressing aneuploidy-inducing shRNAs ( Supplementary Fig S4B).
Telomeres are known to represent fragile sites containing tertiary DNA structures that are difficult to replicate, for example, G-quadruplexes and T-loops (Sfeir et al, 2009;Paeschke et al, 2011); for review see Gunes and Rudolph (2013). Since it is known from studies in yeast and ciliates that telomerase itself enables unfolding of telomeric G-quadruplexes (Paeschke et al, 2008) and rescues telomere replication stress , it was tempting to speculate that aneuploidy of human cells induced replication stress at telomeres, which was rescued by telomerase expression. In agreement with this hypothesis, immuno-FISH staining (combining a telomere-specific FISH probe with a pRPA2 antibody) revealed that > 80% of pRPA2 foci that formed in response to aneuploidy induction in telomerase-negative HFs co-localized with telomeres ( Fig 3D  and E). Furthermore, quantification of telomeric DNA after chromatin immunoprecipitation (ChIP) with a pRPA2 antibody revealed a significant enrichment of telomeric DNA in telomerase-negative human fibroblasts infected with aneuploidy-inducing shRNAs compared to scrambled shRNA-infected cells (Fig 3F and H). The enrichment of pRPA2-bound telomeric DNA in human fibroblasts that were infected with aneuploidy-inducing shRNAs was completely suppressed by telomerase expression (Fig 3G and H).
Studies in ciliates showed that telomerase assists the opening of telomeric G-quadruplexes (required for telomere replication) by interacting with ciliate POT1/TPP1 complex (Paeschke et al, 2008). Interestingly, POT1 was one of the hits in one of our repeat screens on aneuploidy-inducing shRNAs (data not shown). Using two different shRNAs against POT1 confirmed that POT1 knockdown led to an induction of aneuploidy in primary human fibroblast (Supplementary Fig S5A and B). Similar to the results on the other aneuploidyinducing shRNAs (see above), POT1 knockdown led to telomeric DNA damage and AIS ( Supplementary Fig S5A, C, E, G and I). Telomerase overexpression rescued induction of telomeric DNA damage and AIS in response to POT1 knockdown ( Supplementary Fig S5D, F and H). These data suggest that telomerase can rescue telomere replication when POT1 levels become limiting. The data do not exclude the possibility that the remaining 15-30% of POT1 expression in our shRNA experiments (Supplementary Fig S5A) contribute to the telomerase-mediated rescue in telomere integrity by mediating telomerase binding to telomeres. Interestingly, studies on Pot1b null mice revealed that telomerase haploinsufficiency severely impairs the survival of Pot1b-deficient mice by impairing the maintenance of highly proliferating stem and progenitor cell compartments (He et al, 2009). These data suggest that telomerase can rescue telomere replication in a Pot1b-independent manner in mice.
Replication of human telomeres occurs at distinct time points during S-phase with a subset of telomeres replicating in late S-phase (Arnoult et al, 2010). We therefore determined the appearance of pRPA2 foci at early, mid and late S-phase by high-content microscopy analysis of synchronized cells (Fig 3I and J). These experiments revealed that pRPA2 foci accumulated in late S in telomerase-negative human fibroblasts infected with RXFP1-shRNA compared to scrambled shRNA-infected cells ( Fig 3K). Again, pRPA2 foci were suppressed by telomerase expression (Fig 3L). Together, these data indicated that telomere replication stress evolves in response to aneuploidy induction in late S-phase in telomerase-negative human fibroblast, but this phenotype is suppressed by telomerase expression.
Transient tetraploidy can be an initiating event leading to the induction of aneuploidy (Andreassen et al, 1996;Fujiwara et al, 2005). To examine whether tetraploidy would also provoke telomere replication stress and a senescence arrest as seen in AIS, primary human fibroblasts were treated with the cytokinesis inhibitor dihydrocytochalasin B (DCB). In accordance with previous studies (Ganem et al, 2014), DCB treatment resulted in rapid induction of tetraploidy ( Supplementary Fig S6A-D). DCB treatment induced a rapid proliferation arrest, which, however, was independent of telomerase gene status and not rescued by telomerase overexpression (Supplementary Fig S6E). Moreover, tetraploidy induction by DCB treatment did not result in an accumulation of replication stress as determined by the accumulation of pRPA2-foci (Supplementary Fig  S6E). These data indicate that tetraploidy-induced growth arrest involves different mechanisms than AIS. A recent study indicated that the Hippo pathway, specifically the phosphorylation of the transcription co-activator YAP1, contributes to the induction of proliferation arrest in response to tetraploidy (Ganem et al, 2014 Fig S6G).
Together, these data indicate that replication stress at telomeres is induced by aneuploidy but not by tetraploidy. A possible explanation indicates that imbalances in the expression of replication The EMBO Journal Telomerase as an aneuploidy survival factor Jitendra K Meena et al factors could contribute to aneuploidy-induced replication stress, which can be rescued by telomerase expression. Instead, tetraploidy-induced senescence involves a different mechanism (activation of the Hippo pathway) that is not associated with telomere replication stress and is telomerase independent. Mammalian stem and progenitor cells express endogenous levels of telomerase (Morrison et al, 1996). The finding from this study showing that that telomerase expression ameliorates AIS could therefore be relevant for aneuploidy-induced transformation of stem and progenitor cells. This question is of particular interest given that stem and progenitor cells were identified as the cell type of origin of cancer formation in cancer mouse models and cancer patients (Alcantara Llaguno et al, 2009;Barker et al, 2009;Jan et al, 2012;Welch et al, 2012;Genovese et al, 2014;Jaiswal et al, 2014). To directly test whether endogenous levels of telomerase were required for survival of stem cells in response to aneuploidy induction in an in vivo setting, freshly isolated hematopoietic stem and progenitor cells (HSPCs) from wild-type and first-generation telomerase knockout mice (TERC À/À ) were infected with a scrambled shRNA or aneuploidy-inducing shRNAs (targeting GJB3 or OSBPL3, Supplementary Fig S7A) and transplanted into lethally irradiated recipients. These mouse experiments confirmed that the knockdown of candidate genes induced aneuploidy in murine hematopoietic cells (Lin-negative) re-isolated from primary recipients (Fig 4A and B). Of note, in response to aneuploidy induction, the in vivo maintenance of hematopoietic cells from mTERC À/À mice was compromised compared to hematopoietic cells from mTERC +/+ mice as indicated by a decrease in peripheral blood cell chimerism 6 weeks after transplantation of targeted cells into lethally irradiated recipient (Fig 4C). These results indicated that endogenous telomerase expression improved the long-term survival of aneuploid hematopoietic stem and progenitor cells in vivo.
To directly test whether endogenous telomerase expression in HSPCs was sufficient and required to suppress telomere replication stress in response to aneuploidy induction, DNA combing experiments were combined with telomere fluorescence in situ hybridization  Representative images of DNA replication patterns occurring at telomeres. A fork arresting at a telomere was defined as an event in which the DNA replication signal stops within 2 kb from the telomeric border; partial telomere replication was defined as an event in which the DNA replication signal is only partially overlapping in length with the telomeric signal; complete telomere replication was defined as an event in which the DNA replication signal completely overlaps with the telomeric signal. Scale bar: 60 kb. 1: merge; 2: PNA-Telomeric Probe (red); 3: IdU (blue); 4: CldU (green). Freshly isolated wild-type (WT) or telomerase knockout (G1-TERC) mouse HSPCs carrying shRNA against Osbpl3 or, as a control, against firefly luciferase were labeled with thymidine analogues following a pulse and chase protocol. DNA was then isolated, combed and processed by immuno-FISH (see Supplementary Fig S7 for details). The EMBO Journal on freshly isolated HSPCs from recipient mice transplanted with mTERC +/+ of mTERC À/À HSPCs that were infected with shRNAs against the OSBPL3 candidate gene or a control shRNA against Luciferase. These experiments demonstrated that aneuploidy induction resulted in impaired telomeric DNA replication in telomerasenegative HSPCs, but endogenous telomerase expression in mTERC +/+ HSPCs was sufficient to suppress this phenotype (Fig 4D-F and Supplementary Fig S7B and C).

Discussion
It has previously been noted that aneuploidy induces ATM-dependent activation of DNA damage signals and senescence, also referred to as aneuploidy-induced senescence (AIS) (Humbert et al, 2010;Li et al, 2010). However, the mechanistic basis of the activation of this tumor suppressor checkpoint in response to aneuploidy remained elusive. This study demonstrates that aneuploidy induces replication stress at telomeres, and this represents the main cause of AIS, which can be bypassed by telomerase expression. The relevance of this new concept is supported by in vivo experiments demonstrating that endogenous telomerase expression in murine hematopoietic stem cells is sufficient to abrogate telomere replication stress and the depletion of stem cells in response to aneuploidy induction. Recently, it was reported that genome-wide DNA replication stress by itself can contribute to the generation of structural and numerical chromosome abnormalities in tumor cells (Burrell et al, 2013). However, as shown here, the outcome is different in primary cells where aneuploidy induces telomere replication stress limiting the proliferation of cells that lack telomerase expression. Given the increasing evidences that aneuploidy contributes to the transformation of cells during cancer formation (Davoli et al, 2013), it is tempting to speculate that telomere replication stress contributes to tumor suppression in somatic cells that lack telomerase expression. It will be of future interest to delineate molecular mechanisms that contribute to the induction of replications stress at telomeres in response to aneuploidy. Studies on DNA replication indicated that some components of the replication machinery represent limiting factors that need to be re-utilized for completion of DNA replication (Douglas & Diffley, 2012). It is conceivable that alteration in chromosome copy numbers results in imbalances in the amount of molecules required for replication and the amount of genomic DNA that needs to be replicated. Such imbalances would eventually lead to replication stress at chromosomal regions that replicate late during S-phase such as certain telomeres (Arnoult et al, 2010). In agreement with this assumption, the current study shows that replication stress in response to aneuploidy induction evolves in late S-phase (Fig 3I-L). In addition to the timing of replication, specificities in the tertiary structure of telomeres likely contribute to the concentration of replication stress at telomeres in response to aneuploidy induction. Telomeres form tertiary structures, for example, G-quadruplexes and T-Loops, that are difficult to replicate and can be a source of genomic instability (Sfeir et al, 2009;Paeschke et al, 2011). Experimental evidences indicate that specific DNA-helicases (Ribeyre et al, 2009;Paeschke et al, 2011Paeschke et al, , 2013, telomere-binding proteins (Sfeir et al, 2009), and telomerase itself (Paeschke et al, 2008) are required for the opening of telomere tertiary structures and for the faithful replication of these structures (for review see . The current study could extend our understanding of the role of telomerase in cancer formation. Studies in yeast provided the first evidence that the telomerase enzyme contributes to telomere replication by enabling the unwinding G-quadruplexes (Paeschke et al, 2008;Chang et al, 2009). Here, described experiments indicate that telomerase has a conserved function in alleviating replication stress at telomeres that occurs in response to aneuploidy induction in mammalian cells. The findings could be of particular importance for the transformation of tissue stem cells that are known to express high levels of telomerase (Morrison et al, 1996;Ritz et al, 2005) and represent an important cell type of origin for the evolution of cancer in both humans and mice (Alcantara Llaguno et al, 2009;Barker et al, 2009;Welch et al, 2012). It is tempting to speculate that telomerase expression although being essential to ensure prolonged proliferation of stem and The EMBO Journal Telomerase as an aneuploidy survival factor Jitendra K Meena et al progenitor cells also renders these cells more vulnerable to undergo transformation in response to aneuploidy-inducing gene mutations by abrogating the efficiency of aneuploidy-induced senescence (Fig 5). Of note, the increase in cancer incidence in humans with Down Syndrome predominantly affects stem cell system with intrinsically high levels of telomerase activity, for example, the germline and the hematopoietic system (Satge et al, 1998;Ayed et al, 2012).

Cell lines and growth curves
293T, BJ and IMR90 cells were purchased from ATCC. BJ-hTERT and IMR90-hTERT were generated by transduction with the pMSCV-Neo-hTERT retroviral vector. IMR90-ST/LT cells were generated by transduction with the pBABE-Bla-ST/LT vector. All cells were cultured in DMEM high glucose medium (Sigma), penicillin (100 unit/ml), streptomycin (100 lg/ml) (Sigma) and 10% FBS (Lonza) at 37°C and 5% CO 2 atmosphere. For growth curves, 5 × 10 5 cells were seeded in 6-well plates and counted using a hemocytometer and automated cell counter after reaching 70-80% confluency. Population doublings (PD) were calculated using n = (logN-logN0)/log2. N = number of cells after n divisions, N0 = number of cells at time point 0, n = PD difference.

DNA transfection and transduction
All plasmid DNAs were amplified using maxi-prep kit (Promega) in the E. coli DH5a strain. The calcium phosphate method was used for transfection of lentiviral or retroviral construct together with VSVG and GAG/POL (lenti or retro) into 293T cells to produce virus. Virus supernatant was collected at 36 and 48 h after transfection. For transduction, targeted cells were incubated with medium containing virus for 24 h. 8 lg/ml polybrene was added to the medium for enhanced transduction. Twenty-five hours later, appropriate antibiotic selection was started (puromycin 1 lg/ml, neomycin 500 lg/ml, blasticidin 2 lg/ml).

Microscopy
Microscopic analysis of bright-field, immunofluorescence and FISH stainings were carried out on Leica and Zeiss systems with the help of appropriate software. Fluorescence microscopy for DNA molecular combing was performed on an UltraVIEW VoX (Perkin-Elmer) spinning disk confocal unit, equipped an EclipseTi inverted microscope (Nikon), a C9100-50 emCCD camera (Hamamatsu) and driven by Volocity software (Improvision, Perkin-Elmer). Images were acquired using a 40×/0.95 NA dry objective. The right focal plane was maintained using the Nikon Perfect Focus system.
Metaphase countings in the mice experiments were performed with lineage-negative (Lin-) cells. For this experiment, lineage-negative mice HSCs were isolated, kept in culture over night and infected The EMBO Journal with lentiviral particle harboring the indicated shRNA-encoding vectors. After 48 h, cells were transplanted into lethally irradiated mice. Six weeks post-transplantation, mice were killed, lineagenegative cells were isolated, and metaphases were prepared as described above.
Analysis of DNA replication dynamics at telomeres by DNA combing and FISH DNA molecular combing coupled with FISH was performed as described in Suram et al (2012) with minor modifications. HSCs were sequentially labeled with 25 lM IdU for 1 h, followed by 200 lM CldU for one additional hour in the culture medium. Since mouse HSCs were maintained in suspension, no medium changes were performed between the two labelings, and a higher concentration of the second analogue was used in order to compete over the first one during incorporation into the replicating DNA. Cells were then harvested and centrifuged twice and incubated for 3 h in media without IdU/CldU. This procedure was repeated three times. Halogenated nucleotides were detected with specific primary antibodies (IdU: mouse anti-BrdU, Becton Dickinson; CldU: rat anti-BrdU, Abcam) and secondary antibodies (Alexa Fluor 647-conjugated goat anti-mouse, Molecular Probes; Alexa Fluor 488-conjugated goat antirat; Molecular Probes). Images were acquired with a spinning disk fluorescence microscope, and labeled DNA molecules were individually manually measured by ImageJ software and analyzed by Excel.

Chromatin immunoprecipitation (ChIP)
To crosslink protein and DNA, 1% formaldehyde (final concentration) was added to cell culture dishes containing medium and neutralized by adding 0.125 M glycine solution. Cells were washed in ice cold PBS and collected by scraping in PBS. Cell membrane was lysed by NP-40 lysis buffer and then treated with micrococcal nuclease at 37°C. Nuclei were lysed using nucleus lysis buffer, and digested chromatin was sonicated to get equal sizes of chromatin

RNA extraction and qPCR
RNA was isolated by Trifast (PeqGold, PeqLab) according to the recommendations of the supplier. For RT-PCR, cDNA was prepared from 1 lg total RNA using Go script reverse transcriptase (Promega; A5000). Quantitative PCR (qPCR) was performed using the iQ SYBR Green super mix (Bio-Rad; 170-8880) by Applied Biosystem 7300 real-time PCR system. All samples were normalized to GAPDH.
Sequences of hairpin (from human pGIPZ shRNA library) GJB3: TGCTGTTGACAGTGAGCGACCACTCTGAGTTCACTAAGTTT  AGTGAAGCCACAGATGTAAACTTAGTGAACTCAGAGTGGGTGCCT  ACTGCCTCGGA  and  TGCTGTTGACAGTGAGCGAAAACAAGCGAATTAACTATCTTAGTG  AAGCCACAGATGTAAGATAGTTAATTCGCTTGTTTCTGCCTACTGC  CTCGGA  RXFP1: TGCTGTTGACAGTGAGCGCCCACAAGACCATTTAAAGAAA  TAGTGAAGCCACAGATGTATTTCTTTAAATGGTCTTGTGGTTGCCT  ACTGCCTCGGA  and  TGCTGTTGACAGTGAGCGCACGAGACTCAATTCCTATTCATAGTGA  AGCCACAGATGTATGAATAGGAATTGAGTCTCGTTTGCCTACTGCC  TCGGA  OSBPL3: TGCTGTTGACAGTGAGCGCGGCACAAACCTGTGTTTATAT  TAGTGAAGCCACAGATGTAATATAAACACAGGTTTGTGCCATGCC  TACTGCCTCGGA  and  TGCTGTTGACAGTGAGCGCTCCGGTTCTTGGAGAAACATATAGTGA  AGCCACAGATGTATATGTTTCTCCAAGAACCGGATTGCCTACTGCC  TCGGA  STARD9: TGCTGTTGACAGTGAGCGACGAGTTATGACGAAACTTAT  TTAGTGAAGCCACAGATGTAAATAAGTTTCGTCATAACTCGCTGC  CTACTGCCTCGGA  and  TGCTGTTGACAGTGAGCGACCTCAACAGCTCAGTCAGCAATAGTG  Telomere amounts were determined essentially as described previously (Gonzalez et al, 2014). In brief, amplification plot of telomeres or HBG was determined for the linear range according to the telomere or HBG standard. Signal in the ChIPed sample was subtracted by telomere signal of IgG control and corrected for DNA amount by HBG. Finally, values obtained with the indicated shRNAs were calculated relative to the values obtained with the scrambled shRNA and plotted.

Dot blot
A total of 50 ng ChIPed DNA was denatured incubating for 10 min at 99°C with 0.4 N NaOH, and 10 mM EDTA. Denatured DNA samples were applied to hybond N + (Amersham; RPN119B) membrane using a dot-blot apparatus connected to a vacuum source and washed with 2× SSC once. DNA was denatured on a membrane using 1.5 M NaCl/0.5 N NaOH for 10 min and then neutralized using 1.5 N NaCl/0.5 M Tris pH 7. DNA was UV crosslinked using a UV Stratagene crosslinker. Membrane was prehybridized by rapid hybridization buffer (Amersham; RPN1635) for 1 h and incubated over night with radiolabeled telomere (TTAGGG) 3 or Alu probe (5 0 -CGGGAAGCAGAGGTTGTAGTGAGCC). Reaction mixture was 1 ll telomere restriction fragment, 13.5 ll H 2 O, 2 ll T4 PNK buffer, 2.5 ll 32 P-cATP and 1 ll T4-polynucleotide kinase (New England Biolabs) at 37°C. The membrane was washed the next day once with 0.1% SDS/2× SSC buffer and twice with 0.1% SDS/1× SSC. Radiolabeled membranes were exposed to a phosphorimager (Fuji), and signal intensities were quantified after 3 days exposure.

SDS-PAGE and immunoblotting
Same numbers of cells were used to extract proteins in 1× passive lysis buffer (Promega; E194A), and protein supernatant was collected after centrifugation at 16,200 g at room temperature. Proteins were denatured in 1× sample buffer (10% glycerol, 0.002% bromophenol blue, 0.075 M Tris, 2% SDS and 5% betamercaptoethanol) by boiling at 99°C for 10 min and centrifuged at 16,200 g for 10 min. Western blotting was performed as described (Kumar et al, 2013). Following antibodies were used for immunoblot analyses: anti-phospho p53 (s15) (Cell Signaling; #9284),