Role of progerin-induced telomere dysfunction in HGPS premature cellular senescence

Hutchinson-Gilford progeria syndrome (HGPS) is a premature-aging syndrome which affects one in 4-8 million children with symptoms resembling normal adult aging that include very thin skin, loss of subcutaneous fat, alopecia, stiff joints, osteoporosis and heart disease (Hennekam, 2006; Merideth et al., 2008). The age of onset is within 2 years of life, with death at a mean age of 13 as a result of heart attack or stroke (Hennekam, 2006; Merideth et al., 2008). HGPS is caused by a silent G608G mutation within the LMNA gene that encodes lamin A (De Sandre-Giovannoli et al., 2003; Eriksson et al., 2003). This mutation exposes a cryptic splice site in exon 11 that leads to deletion of 50 amino acids required for normal lamin A processing, resulting in an aberrant and permanently farnesylated lamin A protein termed progerin (De Sandre-Giovannoli et al., 2003; Eriksson et al., 2003). Lamins have a structural role in supporting the nuclear envelope and have been implicated in many nuclear functions, including mitosis, DNA synthesis and repair, RNA transcription and processing, apoptosis, organization of chromatin structure and regulation of gene expression (Dechat et al., 2008; Goldman et al., 2002; Gruenbaum et al., 2005; Stuurman et al., 1998; Taddei et al., 2004).

Cellular defects associated with HGPS include a reduced lifespan in culture, irregular nuclear phenotypes such as blebbing of the nuclear envelope, altered chromatin organization, reduced telomere lengths, and a chronic DNA-damage response (Allsopp et al., 1992; Decker et al., 2009; Goldman et al., 2004; Huang et al., 2008; Liu et al., 2005; Liu et al., 2006; Shumaker et al., 2006). Previous studies have shown that the catalytic subunit of human telomerase (TERT) extends HGPS cellular lifespan and rescues proliferative defects associated with progerin (Kudlow et al., 2008; Ouellette et al., 2000; Wallis et al., 2004). However, progerin has been reported to induce increased DNA damage despite the presence of exogenous telomerase expression (Scaffidi and Misteli, 2008), suggesting that telomerase immortalization does not involve amelioration of the DNA-damage phenotype. We undertook the present studies in an effort to elucidate the mechanism by which telomerase overcomes the HGPS premature-senescence phenotype.

Previous studies have provided evidence that ectopic telomerase expression can extend the lifespan of HGPS fibroblasts (Kudlow et al., 2008; Ouellette et al., 2000; Wallis et al., 2004). To confirm and extend these findings, we infected HGPS fibroblasts that were near the end of their proliferative lifespan [two remaining population doublings (PDs)] with a retroviral construct expressing TERT. HGPS fibroblasts transfected with a control vector ceased proliferation within two additional PDs (Fig. 1A). However, HGPS fibroblasts expressing TERT propagated continuously for over 70 PDs without any evidence of a decline in their proliferative capacity (Fig. 1A). This proliferative ability was correlated with an increase in S-phase and decrease in G0-G1-phase fractions of TERT-expressing HGPS cells (Fig. 1B). We also analyzed cultures for senescence-associated β-galactosidase (SA-β-gal), an empirical marker of cellular senescence (Dimri et al., 1995). Although essentially all control HGPS fibroblasts were positive for SA-β-gal activity, more than 90% of HGPS fibroblasts expressing TERT were negative for SA-β-gal at 2 weeks after selection (Fig. 1C). Of note, exogenous TERT expression did not result in decreased progerin expression, and even after many PDs in the presence of TERT, progerin protein levels remained unchanged (Fig. 1D).

The p53 and Rb tumor-suppressor pathways have been implicated in the cellular senescence of normal fibroblasts and can be activated by DNA damage, including damage leading to telomere dysfunction (Campisi, 2001; Maslov and Vijg, 2009). Furthermore, the p53 pathway has been reported to be chronically activated in HGPS (Kudlow et al., 2008; Liu et al., 2005; Liu et al., 2006; Scaffidi and Misteli, 2006). Previous studies have indicated that HPV E6, but not E7, suppresses the growth inhibitory effects of ectopic progerin expression in normal diploid human fibroblasts (NDFs) (Kudlow et al., 2008). However, these studies did not resolve whether this effect was due solely to inactivation of p53 by HPV E6, or whether other E6 activities, such as TERT induction, were involved. Furthermore, lifespan extension in HGPS fibroblasts by direct interference with p53 or Rb pathways, or the effect of TERT on the activation of these pathways has not been studied. Therefore, we analyzed the activation of several cell cycle inhibitors known to be involved in these pathways (Vidal and Koff, 2000), and observed that HGPS fibroblasts transduced with TERT expressed lower levels of p53, p21 and p16 proteins than control fibroblasts, as well as higher levels of the phosphorylated (active) form of Rb (Rb-P) (Fig. 1E). Thus, TERT-induced extension of HGPS fibroblast lifespan was associated with decreased activation of p53 and Rb tumor-suppressor pathways.

To directly investigate the involvement of p53 and Rb pathways in the premature cellular senescence exhibited by HGPS fibroblasts, we infected HGPS fibroblasts that were near the end of their proliferative lifespan with retroviral constructs that block either Rb by CDK4 overexpression or p53 by DNp53 expression. By measuring the proliferative lifespan of these fibroblasts, we observed that CKD4 was able to extend HGPS lifespan by about 26 PDs, whereas DNp53 extended HGPS lifespan by about 6 PDs (Fig. 1F). Moreover, the combination of CDK4 and DNp53 extended HGPS cellular lifespan by over 68 PDs (Fig. 1F). Although the relative ability of DNp53 or CDK4 to extend lifespan varied in different experiments, the combination consistently extended lifespan for many PDs beyond that of either construct alone. We also observed similar lifespan extension with DNp53 and CDK4 in normal senescing fibroblasts (data not shown), indicating that these same effector pathways are involved in HGPS premature senescence, as well as normal senescence.

DNA damage triggers the phosphorylation and/or activation of proteins involved in the DNA-damage response. These include H2AX, a variant of the H2A histone, which is distributed throughout chromatin and becomes rapidly phosphorylated (γH2AX) at nascent double-strand breaks, and ATM, which is activated through autophosphorylation (ATM-P) and recruited to sites of double-strand breaks (Riches et al., 2008). These proteins form discrete foci in cells at sites of DNA damage, and thus are useful markers of such damage. HGPS fibroblasts have been reported to exhibit increased DNA-damage signaling (Liu et al., 2005; Liu et al., 2006). When we compared the effects of exogenous TERT expression on the number of γH2AX and ATM-P foci observed in these cells, we found that TERT expression resulted in a striking reduction in the number of such foci (Fig. 2A,B). Similarly, the total level of ATM-P detectable by immunoblot analysis was significantly reduced in HGPS fibroblasts expressing TERT (Fig. 2C). The effects of TERT expression in HGPS fibroblasts were rapid, with γH2AX levels significantly reduced as early as 7 days after selection, and this decrease correlated with the onset of increased proliferative capacity of the same cells (Fig. 2G).

To further investigate the effects of TERT on the HGPS DNA-damage signaling, we ectopically expressed progerin in human NDFs. In these cells, exogenous progerin was expressed at a level similar to endogenous progerin expression in HGPS fibroblasts (Fig. 2D). Whereas ectopic progerin expression in NDFs induced DNA-damage signaling (Fig. 2E,F), progerin failed to do so in NDFs previously infected with TERT (Fig. 2E,F; Fig. 3D). Thus, TERT expression was protective against DNA-damage signaling induced by progerin.

Accumulating evidence indicates that TERT has activities that are independent of its catalytic function required for telomere maintenance (Choi et al., 2008; Cong and Shay, 2008; De Semir et al., 2007; Lee et al., 2008; Park et al., 2009; Smith et al., 2003; Zhou et al., 2009). For example, both wild-type and catalytically inactive mutant TERT are recruited to the promoters of growth-controlling genes and function indistinguishably to modulate gene expression and enhance cell proliferation (Park et al., 2009; Zhou et al., 2009). Thus, we compared the abilities of wild-type TERT and either a catalytically inactive TERT mutant, D868A TERT, or a telomere-binding-deficient mutant, N125A+T126A TERT, which is catalytically active but fails to elongate the telomeres (Goldkorn and Blackburn, 2006), to extend lifespan and reduce DNA damage. Although all TERT constructs were expressed at similar levels (Fig. 3C), only wild-type TERT extended the proliferative lifespan of HGPS fibroblasts (Fig. 3A). Likewise, only wild-type TERT reduced the level of DNA-damage signaling in HGPS fibroblasts (Fig. 3B). These results demonstrate that amelioration of DNA-damage signaling and premature senescence in HGPS fibroblasts requires both the catalytic and the DNA-binding functions of TERT for telomere maintenance.

To further establish the specificity of telomerase with respect to amelioration of progerin-induced DNA-damage signaling, we investigated the effects of exogenous telomerase on DNA damage caused by doxorubicin (DOX), which induces DNA double-strand breaks through inhibition of topoisomerase II (Tewey et al., 1984). Under identical conditions in which TERT effectively blocked DNA-damage signaling induced by progerin, DOX treatment induced equivalent levels of DNA damage in NDFs with or without ectopic TERT expression (Fig. 3D). These findings further support the conclusion that the ability of TERT to ameliorate progerin-induced DNA-damage signaling and premature senescence is specific to its functions at the telomere.

We next sought to determine whether DNA-damage signaling induced by progerin expression was localized at telomeres. To do so, we examined progerin-expressing fibroblasts for the presence of telomere dysfunction-induced foci (TIFs), as defined by the co-localization of γH2AX and TRF1, which serves as a telomere marker (Takai et al., 2003). TRF2 is a telomeric DNA-binding protein that is essential for normal telomere protection, and a dominant-negative form of TRF2 (TRF2^ΔBΔM) has previously been shown to induce the formation of TIFs (Takai et al., 2003). Ectopic expression of progerin induced the formation of TIFs in NDFs in a similar manner to that observed with TRF2^ΔBΔM (Fig. 4A,B). By contrast, γH2AX foci detected in response to DNA damage induced by DOX showed no evidence of telomere specificity (Fig. 4A,B). Moreover, progerin-induced TIFs occasionally involved the co-localization of γH2AX with several overlapping TRF1 foci (telomere aggregates), which was not observed in any control cells analyzed (Fig. 4C). These findings further implicate the telomere as the specific target of progerin-induced DNA-damage signaling.

To further demonstrate that the DNA-damage signaling observed in response to progerin was localized to telomeres, we performed telomere chromatin immunoprecipitation (ChIP). Whereas progerin induced a 4.2-fold increase in the amount of γH2AX associated with a telomere repeat sequence compared with the control, the association of γH2AX to an internal Alu sequence increased only 1.2-fold under the same conditions (Fig. 5). As a positive control for telomere-specific DNA damage, TRF2^ΔBΔM induced a 6.2-fold increase in γH2AX associated with telomere repeats, whereas γH2AX associated with an Alu sequence showed only a 1.4-fold increase (Fig. 5). Our confocal imaging of DNA-damage signaling at telomeres, and ChIP data demonstrating increased γH2AX associated with telomere sequences, provide strong evidence that implicates progerin in the induction of telomere dysfunction.

TIF formation caused by TRF2^ΔBΔM is the result of dissociation of TRF2 from telomeric DNA, which leads to a loss of the 3′ overhang and DNA-damage signaling (Takai et al., 2003). We used the ChIP assay to compare the effects of progerin and TRF2^ΔBΔM on the association of TRF2 to telomeric DNA. As expected, TRF2^ΔBΔM expression markedly decreased the binding of TRF2 to telomeres while minimally affecting the binding of TRF1, another major regulator of telomere stability (Smogorzewska et al., 2000) (Fig. 6A,B). By contrast, progerin expression was associated with an increase in the binding of both TRF1 and TRF2 to telomeric DNA in two independent experiments (Fig. 6A-D). Binding of TRF1 and TRF2 to telomeric DNA was specific, because no binding was detected at Alu sequences (Fig. 6C). Thus, unlike telomere dysfunction caused by TRF2^ΔBΔM, progerin-induced TIFs do not result from a dramatic loss of TRF2 binding to telomeres.

The TIFs observed in response to progerin expression prompted us to investigate whether we could detect any chromosomal aberrations known to occur under conditions of telomere dysfunction (Bolzan and Bianchi, 2006; Davoli et al., 2010; Hande et al., 2001; Michishita et al., 2008). We performed telomere fluorescent in-situ hybridization (FISH) on metaphase spreads with progerin-expressing NDFs (IMR90), as well as vector controls. It was difficult to identify metaphase cells in progerin-expressing NDFs, which is consistent with their premature-senescence phenotype. However, in those complete or partial progerin metaphases observed, we found abnormalities in 3.5% of 1344 chromosomes. These included chromosomal and sister-chromatid fusions, sister-telomere losses, double telomeric signals on single chromatids (telomere doublets), chromosomal breaks, extra-chromosomal telomeric signals, as well as two metaphases containing diplochromosomes and a chromatin bridge containing telomeric signals that persisted after reformation of the nuclear envelope (Fig. 7A-H). By contrast, we observed abnormalities in only 0.8% of 1506 control chromosomes analyzed. These results indicate that progerin expression promotes chromosomal aberrations in the setting of telomere dysfunction. The fact that progerin induces premature senescence probably protects such cells from more-severe chromosomal instability.

Our present studies demonstrate that telomerase extends HGPS cellular lifespan by decreasing progerin-induced DNA-damage signaling and activation of p53 and Rb pathways that otherwise mediate the onset of premature senescence. We showed further that progerin-induced DNA-damage signaling was localized to telomeres and was associated with telomeric aberrations. Telomerase amelioration of DNA-damage signaling was relatively rapid, required both its catalytic and DNA-binding functions, and correlated in time with the acquisition by HGPS cells of the ability to proliferate. All of these findings establish that progerin-induced telomere dysfunction is responsible for the premature cellular senescence observed in HGPS fibroblasts as well as in normal human fibroblasts in response to exogenous progerin expression.

We observed telomere aggregates in progerin-expressing cells that were similar to those reported in cells ectopically expressing other mutant forms of lamin A and lamin B, and in normal senescent cells (Raz et al., 2008). Telomere aggregates have been suggested to represent telomere fusions and have been linked to genomic instability (Amiel et al., 2009; Louis et al., 2005). Consistent with this, we found several chromosomal aberrations in progerin-expressing cells, including a low level of fusions. In fact, other evidence of genomic instability in HGPS fibroblasts includes findings of lagging chromosomes, and binucleated cells, as well as aneuploid cells (Cao et al., 2007; Ly et al., 2000; Mukherjee and Costello, 1998). Such fusions are also consistent with an elevated level of non-homologous end-joining previously reported in HGPS cell lines (Liu et al., 2005). Genomic instability is also a hallmark of tumorigenesis. Thus, the absence of any reported increased incidence of malignancies in HGPS patients might reflect the severity of premature aging and early death associated with this syndrome.

We observed telomeric DNA-damage signaling within 5 days of exogenous progerin expression in NDFs. Thus, progerin induces telomere dysfunction rapidly, well before telomere attrition, as previously observed in HGPS fibroblasts (Allsopp et al., 1992; Decker et al., 2009; Huang et al., 2008), would be detectable. We also observed that exogenous TERT rapidly ameliorated DNA-damage signaling in HGPS fibroblasts. Our findings contrast with those of a recent report indicating that telomerase fails to protect against progerin-induced DNA damage (Scaffidi and Misteli, 2008). A possible explanation for these differences is the use of a GFP-progerin fusion protein in these latter studies, because GFP tags have been shown to alter protein function, localization and protein interactions (Limon et al., 2007; Skube et al., 2009; Wang et al., 2008).

There is controversy in the literature as to whether the abnormal nuclear shape observed in HGPS fibroblasts is ameliorated by telomerase (Huang et al., 2008; Kudlow et al., 2008; Scaffidi and Misteli, 2008). Several studies have shown that blocking progerin farnesylation by farnesyltransferase inhibitors (FTIs) reverses the aberrant nuclear morphology of HGPS fibroblasts (Capell et al., 2005; Columbaro et al., 2005; Toth et al., 2005), whereas FTIs have no effect on progerin-induced DNA-damage signaling (Liu et al., 2006). Thus, studies with FTIs support the concept that nuclear-envelope defects and telomere dysfunction are independent outcomes of progerin expression and might also explain why progerin-induced proliferation defects are only partially improved by FTIs (Candelario et al., 2008).

The mechanisms involved in progerin-induced telomere dysfunction and its rapid amelioration by telomerase remain to be resolved. Recently, it was shown that knockdown of SIRT6, a histone H3 lysine 9 (H3K9) deacetylase, which modulates chromatin structure at telomeres, induces an abnormal telomeric chromatin state that leads to premature cellular senescence and DNA-damage signaling at telomeres (Michishita et al., 2008), similarly to that observed in response to progerin expression in our present studies. Of note, telomere stabilization by the ectopic expression of TERT also reversed the premature senescence of SIRT6-knockdown cells (Michishita et al., 2008). A hallmark of HGPS cells is decreased heterochromatin (Goldman et al., 2004; Pegoraro et al., 2009; Shumaker et al., 2006), which might be a contributing factor to progerin-induced telomere dysfunction reversible by TERT. This is in contrast to the inability of TERT to rescue TRF2^ΔBΔM-induced telomere dysfunction (Takai et al., 2003). Also, unlike telomere damage caused by TRF2^ΔBΔM, progerin did not cause a decreased association of TRF2 with telomeres. These findings help to distinguish telomere dysfunction induced by progerin from that induced by TRF2^ΔBΔM.

Several mutations in the LMNA gene, which encodes lamins A and C, result in various laminopathies with overlapping phenotypes to those observed in HGPS patients (Raz et al., 2008). Some of these mutations have been reported to result in increased lamin binding to telomeres and altered telomere localization (Raz et al., 2008). In addition, fibroblasts from mice lacking LMNA show altered telomere localization and telomere dysfunction (Gonzalez-Suarez et al., 2009). Therefore, it is possible that the absence of lamins, or the presence of mutant lamins, might disrupt normal lamin-telomere interactions leading to telomere dysfunction by limiting or changing the ability of proteins that function in normal telomere maintenance or repair in a manner that can be overcome by TERT overexpression.

Our results demonstrating that telomere dysfunction, not just reduced telomere lengths, is involved in the premature senescence of HGPS fibroblasts adds HGPS to the list of premature-aging syndromes associated with telomere dysfunction such as Werner's, Bloom's and ataxia telangiectasia (Callen and Surralles, 2004; Crabbe et al., 2007; Puzianowska-Kuznicka and Kuznicki, 2005). Thus, accumulating evidence indicates that telomere maintenance is a common target of diverse genetic defects causing premature-aging syndromes and premature cellular aging in culture. Since increased levels of progerin expression through aberrant splicing have also been observed in normally aging cells (McClintock et al., 2007; Scaffidi and Misteli, 2006), our present findings suggest that progerin-induced telomere dysfunction also contributes to normal aging.

Fibroblasts from a HGPS patient (AG01972, Coriell Cell Repository) were obtained, and sequencing was performed to confirm a G608G LMNA mutation. Cultures were maintained in minimum essential medium (MEM; Invitrogen) supplemented with 0.2 mM non-essential amino acids (NEAA; Invitrogen), 15% heat-inactivated fetal bovine serum (FBS; Invitrogen) and 50 U/ml penicillin and streptomycin (Pen/Strep; Invitrogen). NDFs used were 501T (derived from adult skin) except where IMR90 (fetal lung, ATCC) fibroblasts were used as indicated. NDFs were grown in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 10% FBS and 50 U/ml Pen/Strep. All cells were cultured at 37°C in 5% CO₂. Cellular proliferative lifespan was measured by subculturing fibroblasts 1:4 or 1:2, depending on growth rate, at 90% confluence and recording total population doublings (PDs) and time in culture. PDs were calculated by the formula PD=log₂(1/split ratio) (Harley and Sherwood, 1997).

Full-length cDNA encoding progerin was obtained from total RNA from the AG01972 HGPS cell line by RT-PCR amplification with primers specific for LMNA (sense, including a GCCACC Kozak sequence, 5′-GCCACCATGGAGACCCCGTCCCAGC-3′; and antisense, 5′-GGTCCCAGATTACATGATGCTGC-3′). Progerin was expressed in a NSPI-derived lentiviral vector containing a puromycin selection marker (Akiri et al., 2009). Wild-type TERT, dominant-negative p53 (R248W) and wild-type CDK4 were expressed in pBabe-derived retroviral vectors containing puromycin, hygromycin and neomycin selection markers, respectively (Mahale et al., 2008). TERT mutants, N125A+T126A and D868A, in the pBABE retroviral backbone (kind gift from Elizabeth Blackburn, UCSF, CA) (Kim et al., 2003), along with wild-type TERT, were subcloned and expressed in a NSPI-derived lentiviral vector backbone with the blasticidin selection marker. TRF2^ΔBΔM was expressed in the pLPC retroviral vector with a puromycin selectable marker (Addgene plasmid 18008) (Karlseder et al., 2002). To create retroviral stocks, HEK293T cells were co-transfected with the appropriate retroviral expression vector and pCL-ampho packaging plasmid. To create lentiviral stocks, HEK293T cells were co-transfected with the appropriate lentiviral expression vector, pCMVΔR8.74 packaging vector and pMD2 VSVG envelope vector. Titers for each virus stock were determined by colony formation following marker selection in the same assay cell, HT1080, making it possible to compare results using similar amounts of virus in different experiments. Retroviral and lentiviral infections were carried out on mass populations of fibroblasts in the presence of 8 μg/ml polybrene (Sigma). Cells were subsequently selected for antibiotic resistance (2 μg/ml puromycin, 100 μg/ml hygromycin, 750 μg/ml neomycin and 5 μg/ml blasticidin) and expanded as mass populations. In all cases, similar MOIs were used.

Antibodies against the following proteins were used: Lamin A/C (Millipore, MAB3211), p53 (1801, Mount Sinai School of Medicine Hybridoma Center), p21 (BD Biosciences, 556431), p16 (Santa Cruz, sc-468), Rb (Cell Signaling, 9309), Ser139 γH2AX (Millipore, 05-636), Ser139 γH2AX (Millipore, 07-164), Ser-1981 phospho-ATM (Rockland, 600-401-400), TERT (abcam, ab32020), TRF1 (#370) (a kind gift from Titia de Lange, Rockefeller University, NY), TRF1 (Santa Cruz, sc-6165), TRF2 (Imgenex, IMG-124A), β-actin (Sigma, A5441) and IgG (Millipore, 12-371).

Cell-cycle analysis was performed using the CycleTEST Plus DNA reagent kit (Beckton Dickinson), according to the manufacturer's instructions. Detection of cellular γH2AX was performed using the γH2AX Phosphorylation Assay Kit for Flow Cytometry (Millipore, 17-344), according to the manufacturer's instructions. To combine cell-cycle analysis and γH2AX staining, cells already stained for γH2AX were incubated with 10 μg/ml propidium iodide (Trevigen, 4830-250-3) and 10 μg/ml RNase A (Invitrogen, 12091-021) for 10 minutes at 37°C. At least 10,000 stained cells were sorted by FACS (FACSCalibur, Beckton Dickinson) and analyzed with Cell Quest 3.2 software (Beckton Dickinson).

Cells were washed in PBS and fixed with 2% formaldehyde and 0.2% glutaraldehyde in PBS for 5 minutes at room temperature and then stained as previously described (Dimri et al., 1995).

Whole cell extracts were obtained by solubilizing cells in lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Triton X-100, supplemented with the following protease and phosphatase inhibitors: 5 mM EDTA, 50 mM sodium fluoride, 25 mM β-glycerophosphate, 1 mM sodium orthovanadate, 0.5 mM phenylmethylsulfonyl fluoride, and 10 μg/ml aprotinin) or SDS-lysis buffer [50 mM Tris-HCl, pH 8.1, 1% SDS, 10 mM EDTA, supplemented with the Complete Mini Protease Inhibitor Cocktail (Roche, 11836153001)]. Protein samples (50 μg) were subjected to SDS-PAGE, transferred onto an Immobilon-P or Immobilon-FL filter (Millipore) and probed with the indicated antibodies. Detection was carried out with either an ECL system (GE Healthcare) with horseradish peroxidase-conjugated secondary antibodies (GE Healthcare) or an Odyssey Infrared Imaging System (LI-COR Biosciences) with IR-dye-tagged secondary antibodies (LI-COR Biosciences).

Cells plated and grown on glass coverslips at least 48 hours were washed twice with PBS followed by fixation for 20 minutes with methanol at −20°C. Cells were then washed twice with PBS-T (PBS with 0.1% Tween 20), blocked for 1 hour in PBS-BSA-T (PBS with 1% BSA and 0.1% Tween 20) and incubated overnight at 4°C with primary antibodies. Samples were then washed three times with PBS-BSA-T and incubated with secondary antibodies for 1 hour. Anti-rabbit Cy3 (Jackson ImmunoResearch, 711-165-152) or anti-mouse Alexa Fluor 488 (Molecular Probes, A11029) secondary antibodies were used. In some experiments, doxorubicin (DOX) (Sigma, D1515) at 500 nM was added for 1 hour before fixation. Coverslips were mounted using Vectashield Mounting Medium with DAPI (Vector Laboratories, H-1200). Confocal imaging was performed with a Zeiss LSM 510 META confocal microscope (Carl Zeiss Microimaging) using the 63× oil objective. Images were cropped and combined in Adobe Photoshop. The amount of DNA damage was quantified by scoring the percentage of cells containing 0, 1, 2-5 or >5 γH2AX and ATM-P colocalized foci. At least 300 cells were scored for each variable. For TIF analysis, cells were scored as having 0-1, 2-5, 5-10 or >10 TRF1 foci that colocalized with γH2AX staining. At least 100 cells were scored for each variable.

The ChIP assay was performed using the EZ-Chip assay kit (Millipore, 17-371), according to the manufacturer's instructions. Sonication was done using a Sonicator 3000 (Misonix) under the following conditions: Amp=5.5 with six cycles of 20 seconds on and 20 seconds off. Immunoprecipitation was performed with either Ser139 γH2AX (Millipore, 07-164) or IgG (Millipore, 12-371) antibodies, and immunoprecipitated DNA was transferred to a Hybond-N (GE Healthcare, RPN2020N) membrane using a slot-blot apparatus. The membrane was then hybridized with a DIG-labeled telomeric (TTAGGG)₄ probe and detected with the TeloTAGGG Telomere Length Assay kit (Roche 12209136001). The membrane was then stripped by washing twice with 0.2 M NaOH and 0.1% SDS for 30 minutes at 52°C and re-hybridized with a DIG-labeled Alu (GGAGGCTGAGGCAGGAGAATTGCT) probe. DIG labeling was performed using the DIG Oligonucleotide 3′-End Labeling Kit (Roche, 03353575910). Quantification of the signal was performed with ImageJ software (NIH). The amount of telomeric and Alu DNA after ChIP was normalized to the total input signal for each condition.

Cells were treated with 0.1 μg/ml demecolcine (Sigma, D7385) for 16-22 hours, harvested by trypsinization, and swelled in 0.075 M KCl for 20 minutes at 37°C. Cells were fixed overnight at −20°C in 3:1 methanol:acetic acid, dropped onto humidified slides, and air-dried overnight. Cells were then rehydrated in PBS and fixed with 4% formaldehyde in PBS for 4 minutes at room temperature. After rinsing, cells were dehydrated in a 70%, 90% and 100% ethanol series, and air-dried. Hybridizing solution [70% formamide, 0.5% blocking reagent (Roche, 11096176001), 20 mM Tris-HCl, pH 7.5, and 500 nM FAM-OO-(CCCTAAA)3 PNA probe (Panagene)] was added, and the slides were heated for 3 minutes at 80°C, followed by incubation in the dark for 2 hours at room temperature. The slides were washed twice for 15 minutes each in wash solution 1 (70% formamide, 10 mM Tris-HCl, pH 7.5 and 0.1% blocking solution), and three times for 5 minutes each in wash solution 2 (100 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 0.08% Tween 20). Slides were mounted using Vectashield Mounting Medium with DAPI (Vector Laboratories, H-1200) and imaging was performed with a Zeiss Axioplan II microscope (Carl Zeiss Microimaging) using the 63× oil objective. Images were cropped and combined in Adobe Photoshop.

This work was supported by grant number PO1CA80058 from the NCI and the New York State Stem Cell contract #C024313. E.B. was supported by the NCI Training Program in Cancer Biology (T32 CA078207) and the NIGMS Training Program in Cellular and Molecular Biology (T32 GM008553). Confocal laser scanning microscopy was performed at the MSSM-Microscopy Shared Resource Facility, supported with funding from NIH-NCI shared resources grant (5R24 CA095823-04), NSF Major Research Instrumentation grant (DBI-9724504) and NIH shared instrumentation grant (1 S10 RR0 9145-01). We thank Elizabeth Blackburn (UCSF) for providing us with TERT mutant constructs, N125A+T126A and D868A, in the pBABE-puro backbone. We thank Titia de Lange (Rockefeller University) for the TRF1 (#370) antibody and for a TRF2^ΔBΔM construct obtained via Addgene. We also thank Bo Zhao and Cesar Munoz-Fontela for helpful discussions. Deposited in PMC for release after 12 months.