Oncogene (2012) 31, 1–12 & 2012 Macmillan Publishers Limited All rights reserved 0950-9232/12 www.nature.com/onc ORIGINAL ARTICLE Anthracyclines disrupt maintenance by through inducing PinX1 ubiquitination and degradation

B Zhang1,5, D Qian1,5, H-H Ma2, R Jin2, P-X Yang2, M-Y Cai1, Y-H Liu3, Y-J Liao1, H-X Deng1, S-J Mai1, H Zhang2, Y-X Zeng1, MC Lin4, H-F Kung1,4, D Xie1 and J-J Huang2

1State Key Laboratory of Oncology in South China, Cancer Center, Sun Yat-Sen University, Guangzhou, China; 2Laboratory of Tumor and Molecular Biology, Beijing Institute of Biotechnology, Beijing, China; 3Department of Pathology, Guangdong Provincial People’s Hospital, Guangzhou, China and 4State Key Laboratory of Oncology in South China, The Chinese University of Hong Kong, Hong Kong, China

Telomere maintenance is essential for cancer growth. responsible for the hallmark immortalization phenotype Induction of telomere dysfunction, for example, by inhibition of cancer cells; this is because of its effect on telomere of telomeric proteins or telomerase, has been shown to maintenance. Consequently, targeting the pathway of strongly enhance cancer cells’ sensitivity to chemotherapies. telomere maintenance by telomerase represents a novel However, it is not clear whether modulations of telomere approach for anti-cancer therapy. Understanding how maintenance constitute cancer cellular responses to che- this carcinogenic pathway is regulated in cancer cells is motherapies. Furthermore, the manner in which anti-cancer highly desirable for this purpose. drugs affect telomere function remains unknown. In this Telomere is the terminal DNA component of the linear study, we show that anthracyclines, a class of anti-cancer . Owing to the incapacity of the DNA drugs widely used in clinical cancer treatments, have an active polymerase to fully replicate the very end sequences of role in triggering telomere dysfunction specifically in double-stranded , telomeric DNA would telomerase-positive cancer cells. Anthracyclines interrupt shorten at a constant rate in each cell division (De Lange, telomere maintenance by telomerase through the down- 1998; Blackburn, 2000). When telomere shortens to a regulation of PinX1, a protein factor responsible for targeting critical point, it would trigger DNA-damage responses telomerase onto , thereby inhibiting telomerase (DDRs) at the chromosomal ends and force normal cells association with telomeres. We further demonstrate that into senescence. By contrast, telomerase, which is anthracyclines downregulate PinX1 by inducing this protein suppressed in normal somatic cells, is reactivated to degradation through the ubiquitin–proteasome-dependent counteract telomere shortening in nearly all human cancer pathway. Our data not only reveal a novel action for cells, and these malignant cells enjoy telomere-length anthracyclines as telomerase functional inhibitors but also maintenance and unlimited proliferative capacity (Kim provide a clue for the development of novel anti-cancer drugs et al., 1994; Autexier and Greider, 1996; Collins, 2000). based on telomerase/telomere targeting, which is actively Telomerase is a ribonucleoprotein reverse transcrip- investigated by many current studies. tase specialized for de novo synthesis of telomeric DNA Oncogene (2012) 31, 1–12; doi:10.1038/onc.2011.214; at chromosomal ends after each cell cycle. The core published online 6 June 2011 of this enzyme from all species is minimally composed of a protein-catalytic subunit and a single-stranded Keywords: anthracyclines; telomere dysfunction; PinX1; RNA called TERT and TR, respectively (Nugent and telomerase; cancer Lundblad, 1998; Collins, 2006). The expression of TERT has been demonstrated as a rate-limiting step in human cellular telomerase activation, and it is sup- pressed by multiple tumor-suppressor pathways in Introduction normal human cells but is activated by numerous oncogenic transcription factors (Meyerson et al., 1997; Telomerase reactivation has been identified as a key Liu et al., 2000; Cong et al., 2002; Lin and Elledge, 2003; step in human cellular tumorigenesis, and it is also Goueli and Janknecht, 2004). After its expression, regulations of the subcellular transportation of the Correspondence: Dr D Xie or Dr H-F Kung, State Key Laboratory of TERT protein contribute to telomere homeostasis Oncology in South China, Cancer Center, Sun Yat-Sen University, maintenance in telomerase-positive cancer cells. It has Room 634, 651, Dongfeng Road East, 510060 Guangzhou, China or Dr J-J Huang, Laboratory of Tumor and Molecular Biology, been shown that the expressed TERT undergoes a Beijing Institute of Biotechnology, 27, Taiping Road, Hai Dian highly programmed subcellular transportation before it District, Beijing 100850, China. reaches its final functional site at the telomere E-mail: [email protected] or (Wong et al., 2002; Tomlinson et al., 2006). Moreover, [email protected] or [email protected] 5These authors contributed equally to this work. regulations of the TERT subcellular transportation may Received 16 December 2010; revised and accepted 27 April 2011; have implications in the two central issues of telomerase/ published online 6 June 2011 telomere biology, that is, telomerase biogenesis and Anthracyclines disrupt telomere maintenance by telomerase B Zhang et al 2 telomerase telomere recruitment in vivo. Moreover, findings not only suggest that there is a PinX1-mediated alterations of TERT subcellular distributions have been pathway for telomere maintenance in cancer cells but observed in cancer cellular responses to certain cyto- they also indicate that the functional maintenance of toxic-damaging stimuli, such as DNA damages and this pathway may have an important impact on cancer oxidative stresses, emphasizing that modulations of cells’ survival under DNA-damaging stresses. Therefore, TERT transportation, thereby regulating the pathway it is worthwhile to further investigate whether modula- of telomere maintenance by telomerase, may have an tions of this pathway are involved in cancer cellular important impact on cancer cell survival under anti- responses to anti-cancer drug treatments. cancer treatments. However, how TERT transporta- Anthracyclines rank among the most-effective anti- tions are regulated in vivo is still poorly understood. cancer drugs ever developed. It has generally been Undoubtedly, elucidating these questions will facilitate accepted that anthracyclines function as a topoisome- the development of novel anti-cancer therapies based on rase poison and that they kill cancer cells through telomerase/telomere biology. inducing DNA double-stranded break (DSB) damage Telomere maintenance by telomerase requires telo- (Minotti et al., 2004). In this study, we show that the merase interplay with telomeric DNA. In cells, telomeric anthracyclines doxorubicin (DOX) and its analog DNA is coated with a set of specialized telomeric epirubicin (EPI) also have an effect on inducing proteins. Six telomere-constitutive protein components telomere dysfunction, specifically in telomerase-positive have been identified, including Pot1, TPP1, TIN2, cancer cells. We further demonstrate that both DOX TRF1, TRF2 and Rap1 (de Lange, 2005). These six and EPI perform this novel effect by triggering PinX1 telomeric proteins form a complex called at into a proteasome-dependent protein degradation path- telomere ends, and they function as a whole to protect way, thereby disrupting the telomere maintenance by telomeric DNA from being recognized as DNA broken telomerase in affected cancer cells. Therefore, we for the by the DNA damage-repair machinery but, meanwhile, first time provide the direct evidence that modulations of they also prevent telomerase from having access to its PinX1-dependent telomere maintenance by telomerase functional substrate telomeric DNA (de Lange, 2005; constitute cancer cellular responses to certain anti- Palm and de Lange, 2008). Therefore, communication cancer drug treatments, such as anthracyclines, and with the shelterin complex is an essential step for the that this may have important consequences for their telomerase to carry out its telomere maintenance action. anti-cancer effects. These novel findings may also Recent studies have revealed that in addition to provide a clue for designing new anti-cancer drugs participating in the shelterin formation, the Pot1-TPP1 based on targeting telomere maintenance by telomerase. heterodimers have crucial roles in recruiting telomerase function on telomeres (Xin et al., 2007). The studies suggest that the Pot1-TPP1 dimer may function as the Results telomerase receptor at telomeres for guiding this enzyme’s access to its functional default substrates. Inducible telomere dysfunction is specifically associated However, given that the Pot1-TPP1 dimer has been with anthracycline treatments in telomerase-positive considered as a constitutive telomeric component, how cancer cells the telomerase ribonucleoprotein complex, which is The induction of DNA DSB damage is a common clearly assembled within nuclei outside the telomere method for the action of most anti-cancer drugs to treat ends, is targetedly delivered onto the Pot1-TPP1 unit tumors. It has been shown that telomere maintenance still remains unknown. has great impact on cancer cellular responses to DSBs. The nucleolar protein PinX1 was originally found as Therefore, it is interesting to ask whether modulations an interacting protein of both the telomeric protein of telomere maintenance constitute cancer cellular TRF1 and the telomerase catalytic protein subunit responses to these anti-cancer drug treatments. One of TERT (Zhou and Lu, 2001). Although PinX1–TRF1 the major functions of telomeres is to protect chromo- interaction has not been functionally characterized, the somal ends from being recognized as DNA DSBs. binding of PinX1 to TERT was experimentally shown to Consequently, telomere dysfunction evokes the activa- have a telomerase activity (TA)-inhibitory effect, sug- tion of DDRs at chromosomal ends, represented by foci gesting that this protein might function as a cellular accumulation of activated DNA damage-responding telomerase inhibitor. However, it is intriguing that the protein factors, such as ATM, 53BP1 and gH2AX, at PinX1 expression appears not to be affected in most telomeres. This particular event is defined as TIFs human cancer cells (Akiyama et al., 2004; Chang et al., (telomere dysfunction-induced DNA-damaging foci). 2004; Hawkins et al., 2004). We recently found that TIFs can be judged by co-localizations of DNA- depletion of the PinX1 expression has an effect in damaged foci (such as the gH2AX foci) with telomere suppressing telomerase association with the Pot1-con- markers (such as telomeric DNA or telomeric protein taining telomeric protein complex, and it significantly components) (Takai et al., 2003; Denchi and de Lange, shortens telomere lengths in telomerase-positive cancer 2007). We examined TIFs in the two telomerase-positive cells. Moreover, silencing PinX1 makes these affected cancer cell lines of HepG2 and MCF7 during the courses cancer cells more sensitive to DNA damage-induced of their responses to different DSB-inducing anti-cancer telomere dysfunction and to in telomerase- drugs, including etoposide (Et) and the anthracyclines positive cancer cells (Zhang et al., 2009). These novel DOX and its analog EPI. As shown in Figure 1a, all

Oncogene Anthracyclines disrupt telomere maintenance by telomerase B Zhang et al 3

Figure 1 The anthracyclines DOX and EPI induce telomere dysfunction in telomerase-positive cancer cells. (a) Telomerase-positive MCF7 and HepG2 human cancer cells were treated with different DNA-damaging anti-cancer drugs, including etoposide (Et, 20 mg/ ml), doxorubicin (DOX, 1 mg/ml) and epirubicin (EPI, 1 mg/ml). Sixteen hours after the treatments, cells were double-stained with anti- gH2AX for DNA damage detection (green) and anti-TRF1 for telomere signals (red). Nuclei were stained with DAPI (blue). Merged images showing DNA damages at telomeres were presented in the anthracyclines DOX- and EPI-treated cells but were not significantly detected in Et-treated cells. Enlarged images were for clearly showing the co-localizations of gH2AX and TRF1 signals in the corresponding merged images. Arrows indicate co-localization events. A total of 400 cells were counted per experiment, and cells with >5 spots of gH2AX and TRF1 co-localization signals were considered as TIF positive. Column, mean of triplicate experiments. Error bars, s.d. (b) TIF incidences (percentage of TIF-positive cells) in HepG2 cells after different concentrations of DOX and EPI treatments. (c) TIF incidences in telomerase-negative BJ normal fibroblasts before and after EPI and DOX treatments. NT, untreated cells.

Oncogene Anthracyclines disrupt telomere maintenance by telomerase B Zhang et al 4 these drugs induced robust DDRs both in MCF7 and (Figure 2d). Like cases observed in Flag-TERT-stable HepG2 cancer cells, as indicated by foci formations of HepG2 cells, anthracycline treatments also did not gH2AX. Notably, we found that there were substantial affect the expressions of the endogenous TERT protein fractions of the gH2AX foci co-localized with the but significantly reduced its association with Pot1 in detecting signals for the telomeric constitutive protein non-transfected HepG2 cells upon the treatments component TRF1 in both DOX- and EPI-treated cells (Figures 2e and f). Therefore, the anthracyclines- (Figure 1a); this suggests that telomere dysfunction is induced telomere dysfunction is associated with the present in the courses of cancer cellular responses to reduction of telomerase association with telomeres in these two drug treatments. Moreover, both DOX and cancer cells. EPI could induce telomere dysfunction in a dose- dependent manner, with this being more obvious for Anthracyclines block telomerase association with EPI treatment (Figure 1b). However, incidences of telomeres by downregulating PinX1 expression telomere dysfunction were only occasionally observed We recently demonstrated that the nucleolar protein in Et-treated cells (Figure 1a). In fact, we found that PinX1 has an effect of mediating telomerase association TIF incidences were also not present in tested cancer with the Pot1-containing telomeric protein complex in cells, which had been treated with another kind of DSB- cancer cells, and that the disruption of this pathway by inducing anti-cancer drug camptothecin (data not PinX1 silencing renders telomeres prone to dysfunction shown). The effect of anthracyclines on inducing in telomerase-positive cancer cells under DNA-dama- telomere dysfunction was also examined in telomerase- ging stress conditions (Zhang et al., 2009). We thus negative BJ normal human fibroblasts. Interestingly, we tested whether anthracyclines-induced telomere dysfunc- found that, in the ranges of the indicated dosages, tion is related to suppressing this PinX1-dependent neither DOX nor EPI caused obvious TIFs in these telomere maintenance by telomerase. Western blotting normal human cells (Figure 1c). In fact, we also found assays indicated that an obvious inducible downregula- that both DOX and EPI have no effects on increasing tion of the PinX1 protein expression could be detected in TIF incidences in telomerase-negative U2OS human Flag-TERT-stable HepG2 cells as early as 6 h after cancer cells (data not shown). Thus, these results suggest DOX and EPI treatments and that it could be sustained that anthracyclines have an effect on inducing telomere for up to 24 h of observation (Figure 3a). However, the dysfunction in cancer cells through a telomerase- expression levels of the PinX1 protein remained un- dependent pathway. changed in the same cells before and after Et treatment (Figure 3a). To make sure that this inducible PinX1 Anthracycline treatments suppress telomerase association downregulation is responsible for the effect of anthra- with telomeres without affecting telomerase expression in cyclines on blocking telomerase association with telo- cancer cells meres and for inducing telomere dysfunction in cancer The findings that the anthracyclines DOX and EPI cells, full-length PinX1 was introduced into Flag-TERT- cause telomere dysfunction specifically in telomerase- expressing HepG2 cells through a recombinant lentiviral positive cancer cells indicated that this class of anti- infection, and the overexpression of PinX1 in Flag- cancer drug might affect a mechanism involved in the TERT-stable cells was confirmed by western blotting telomere maintenance by telomerase. Therefore, we first assay (Figure 3c). Interestingly, the expression levels of analyzed the telomerase enzymatic activities and their exogenous PinX1 could also be partially downregulated expression levels in cancer cells before and after by EPI treatment (Figure 3c). However, the remaining treatments of these reagents. As indicated in Figures PinX1 protein expression levels seemed to be sufficient 2a and b, we found that neither DOX nor EPI caused to suppress the inducible telomere dysfunction in these significant alterations in total telomerase enzymatic infected cells by means of EPI treatment. Incidences of activity (assayed by the standard TRAP) or in the the EPI-induced telomere dysfunction were significantly TERT expression (by reverse transcriptase–PCR) in the decreased in PinX1-overexpressing cells than in unin- tested MCF7 and HepG2 cancer cell lines. We next fected control cells (Figure 3e). In addition, as had been detected whether DOX and EPI affected the telomerase predicted, PinX1 overexpression significantly attenuated association with telomeres, a prerequisite step for the the effect of EPI on inhibiting the telomerase/telomere action of telomerase on telomere maintenance. The association (Figure 3d). We also conducted analysis to telomere associated with telomerase was monitored by discover whether silencing the endogenous PinX1 detecting the amounts of the telomerase protein TERT expression could enhance EPI-induced telomere dys- in Pot1-containing telomeric protein complex. Flag- function. A PinX1 short-hairpin RNA (shRNA) was TERT-stable HepG2 cells were first used for testing this introduced into Flag-TERT-stable HepG2 cells through hypothesis. As shown in Figure 2c, the expression levels a recombinant lentiviral infection, and significant of Flag-TERT in stable cells did not change either suppression of endogenous PinX1 expression by shRNA before or after drug treatments. However, the amounts was confirmed by immunoblotting assay (Figure 3c). of Flag-TERT in the Pot1 immunoprecipitates were Acutely silencing the endogenous PinX1 expression did significantly reduced upon EPI and DOX treatments not cause obvious telomere dysfunction in affected cells (Figure 2d). As the control, we showed that the amounts under their normal cultural conditions (Figure 3e), but of the Flag-TERT protein in the Pot1 immunoprecipi- incidences of the inducible telomere dysfunction by EPI tates were not affected by Et treatment in the same cells were increased remarkably in shRNA-infected cells than

Oncogene Anthracyclines disrupt telomere maintenance by telomerase B Zhang et al 5 HepG2 MCF7 HT NT EPI DOX NT EPI DOX DOX (1ug/ml) EPI (1ug/ml) Et (20ug/ml) (ng extracts) - ++- - + 500 500 100 20 500 100 20 500 100 20 500 100 20 500100 20 500 100 20 Flag-TERT

Pot1

IP: anti-Pot1

DOX (1ug/ml) EPI (1ug/ml) IC -++ - 14 14 12 NT 12 NT EPI TERT 1 EPI 1 DOX DOX 0.8 0.8 0.6 0.6 0.4 0.4 α-tubulin 0.2 0.2 0 0 500 100 20 (ng) 500 100 20 (ng) Relative Telomerase Activity Telomerase Relative Protein Extract Protein Extract

HepG2 MCF7

NT EPI DOX NT EPI DOX NT EPI DOX NT TERT (1ug/ml) (1ug/ml)

TERT GAPDH

Pot1 DOX (1ug/ml) EPI (1ug/ml) Heavy chain - + - + IP: anti-Pot1 Control IgG Flag-TERT Pot1

Input α-tubulin

Figure 2 Anthracyclines DOX and EPI stimulations decrease TERT association with telomeres without affecting its expression. (a) TRAP assays indicate that DOX (1 mg/ml) and EPI (1 mg/ml) treatments do not cause significant alterations of telomerase enzymatic activities in the tested MCF7 and HepG2 cancer cells. HT, heat-inactivated samples as a negative control. IC, 36-bp internal control. Column, quantitation of telomerase activity (see the ‘Materials and methods’ section). Results of three independent experiments. (b) After DOX (1 mg/ml) and EPI (1 mg/ml) treatments, expression levels of endogenous TERT mRNAs in both MCF7 and HepG2 cancer cells were detected by RT–PCR assays. (c) Flag-TERT-stable HepG2 cells were treated with indicated dosages of DOX and EPI for 24 h, respectively. IB assays with the anti-Flag antibody. Tubulin as control to show equal amounts for loading samples in each of the lanes. (d) Nuclear extracts were prepared from Flag-TERT-stable HepG2 cells before and after 24 h of treatment with indicated dosages of DOX, EPI and Et, respectively. Endogenous telomeric protein Pot1 was immunoprecipitated with a specific antibody from each of the indicated samples. The amounts of Flag-TERT associated with the Pot1 immunoprecipitates were monitored by IB with anti-Flag antibody. IB Pot1 as control to show equal amounts of the Pot1 immunoprecipitates for each detected sample. (e) After indicated treatments, endogenic TERT was detected by western blotting with Anti-TERT antibody. (f) The amounts of endogenic TERT associated with the Pot1 immunoprecipitates. Immunoprecipitates were performed with the Pot1 antibody. IB assays were performed with the TERT antibody. Control IgG was used to show the specificity of immunoprecipitates with the Pot1 antibody and immunoblots with the hTERT antibody. Lower panel indicated the equal input lysates. NT, untreated cells. in control cells (Figure 3e). Next, we examined the human cells. However, unlike the case observed for effects of DOX and EPI on PinX1 expression in telomerase-positive cancer cells, PinX1 overexpression telomerase-negative BJ normal fibroblasts and in or silencing PinX1 expressions in telomerase-negative BJ U2OS cancer cells, and we found that both can be cells (Figure 3f, left panel) and in U2OS cells (data not induced to exhibit a dose-dependent downregulation of shown) has no effect in inducing telomere dysfunction PinX1 expression upon DOX and EPI treatments with and without the anthracyclines (Figure 3f, right (Figure 3b), thereby indicating that induction of a panel). Thus, these results may suggest that the effect of PinX1 decrease is a general effect of anthracyclines in anthracyclines on inducing PinX1 downregulation to

Oncogene Anthracyclines disrupt telomere maintenance by telomerase B Zhang et al 6

Figure 3 Anthracyclines block TERT association with telomeres and trigger telomere dysfunction by downregulating PinX1 expression. (a) Flag-TERT-stable HepG2 cells were treated with Et (20 mg/ml), DOX (1 mg/ml) and EPI (1 mg/ml). At the indicated time points, the endogenous PinX1 expression was detected by immunoblotting assay with specific PinX1 antibodies. Tubulin as control to show equal loading amounts for each detected sample. (b) Telomerase-negative BJ normal fibroblasts and U2OS cancer cells were treated with the indicated doses of drugs for 24 h and PinX1 expression was examined by immunoblotting assay. (c) A PinX1 plasmid for PinX1 overexpression and a PinX1 shRNA plasmid to knock down endogenous PinX1 expression were delivered into Flag-TERT HepG2 cells through a lentiviral infection. PinX1 expression was examined by immunoblotting assay before and after the EPI treatment. Tubulin as control to show equal loading samples. Mock, uninfected control cells. (d) IP assays show that the effect of EPI on inhibiting TERT association with the Pot1-containing telomeric protein complex is relieved in PinX1-overexpressing cells. (e) TIFs assays in HepG2 cells after the indicated treatments. Left, representative images of gH2AX (red) and TRF1 (green) and DAPI (blue) staining. Arrows indicate co-localization events. Right, results are mean of triplicate experiments; bars, s.d. (f) BJ cells infected with lentiviruses for PinX1 overexpression and PinX1 knockdown were confirmed by western blotting (left panel). TIFs assays in BJ cells after the indicated treatments (right panel).

Oncogene Anthracyclines disrupt telomere maintenance by telomerase B Zhang et al 7 trigger telomere dysfunction is dependent on whether prompted us to consider whether these reagents might the affected cell is addicted to telomerase for its telomere affect this protein stability. Supporting this suggestion, maintenance. Therefore, we propose that anthracyclines reverse transcriptase–PCR assays indicated that the induce a PinX1 decrease to evoke telomere dysfunction expression levels of the PinX1 mRNA transcripts did in telomerase-positive cancer cells, but not in telomer- not show detectable changes in HepG2 cells before and ase-negative human cells, through disturbing a PinX1- after EPI and DOX treatments (Figure 4a). However, dependent pathway for recruiting telomerase to main- when these cells were pretreated with MG132, a reagent tain telomeres. that can prevent protein degradation through the proteasome pathway, the effect of EPI and DOX on Anthracyclines downregulate PinX1 expression inducing PinX1 downregulation was significantly by inducing this protein degradation through blocked (Figure 4b). Interestingly, MG132 treatment the ubiquitin–proteasome pathway had only a minor effect on PinX1 protein accumulation The finding that anthracycline stimulation can also in HepG2 cells without EPI and DOX stimulation downregulate ectopic-expressed PinX1 protein levels (Figure 4b); this indicates that cellular expressed PinX1

Figure 4 Anthracyclines downregulate PinX1 expression by inducing this protein degradation through the ubiquitin–proteasome- dependent pathway. (a) HepG2 cells were treated with DOX and EPI for the indicated periods. PinX1 mRNA expression levels were detected by RT–PCR assays. GAPDH as control for total RNA was used in the RT–PCR assays. (b) HepG2 cells were pretreated with ( þ ) or without (À) the proteasome inhibitor MG132 (10 mM) for 8 h before they were exposed to the indicated dosages of EPI and DOX, respectively. After 16 h of EPI and DOX treatments, nuclear extracts were prepared and subjected to IB assays with PinX1- specific antibodies. Lower panels showed the ratios of PinX1 to tubulin from densitometric analysis. The results were representative of the three experiments. (c) HepG2 cells were transfected with a plasmid encoding Flag-PinX1 or a vector control, and transfected cells were pretreated with MG132 (10 mM) for 4 h before they were exposed to DOX (1 mg/ml) and EPI (1 mg/ml), respectively. Twelve hours after anthracyclines treatments, ectopic-expressed Flag-PinX1 was immunoprecipitated from nuclear extracts by the anti-Flag M2 beads. The immunoprecipitates were subjected to IB assay with an anti-ubiquitin antibody. IB with the anti-Flag antibody shows equal amounts of the precipitated PinX1 proteins for each detecting samples.

Oncogene Anthracyclines disrupt telomere maintenance by telomerase B Zhang et al 8 proteins were not constantly undergoing proteasome- enzyme was characterized, but as yet a specific inhibitor mediated degradation but could be induced into this has not been discovered. On the other hand, researches pathway upon both EPI and DOX treatments. As have revealed that the process of telomere maintenance ubiquitination has been established as the major signal by telomerase in cancer cells involves multiple regula- for directing proteins into the proteasome-mediated tory steps, such as biogenesis of the telomerase degradation pathway, we further detected whether both ribonucleoprotein complex and recruitment of telomer- drugs have the effect of inducing PinX1 ubiquitination. ase onto telomeres, suggesting that interferences of these As shown in Figure 4c, robust ubiquitination signals events, without the need of wiping out TA per se, can could be detectable in the PinX1 protein immunopreci- also display the same outcome as do the telomerase pitates from MG132-pretreated HepG2 cells in the inhibitors (Collins, 2006; Bianchi and Shore, 2008; presence of EPI and DOX, but only faint ubiquitination Gallardo et al., 2008). Most recently, we have demon- signals were observed in the PinX1-immunoprecipitated strated that the nucleolar protein PinX1 has a role in products from MG132-pretreated HepG2 cells without mediating telomerase association with telomeres in EPI and DOX treatments. Thus, we propose that both cancer cells, and that disruption of this PinX1-depen- EPI and DOX trigger a response of PinX1 ubiquitina- dent telomerase/telomere recruitment pathway, for tion, which in turn targets PinX1 degradation through example, by PinX1 silencing, significantly shortens the proteasome pathway. telomere lengths, inhibits tumorigenesis and also en- hances their sensitivity to the cytotoxic effects of DNA- Suppression of PinX1-dependent telomere maintenance damaging reagents in telomerase-positive cancer cells. by telomerase contributes to the cancer-killing effects These results thus indicate that maintenance of this of anthracyclines PinX1-dependent telomerase/telomere recruiting path- So far, we have revealed a novel effect for anthracyclines way is critical for telomere stability in cancer cells under in inducing telomere dysfunction by suppression of both their normal and damaging growth conditions, PinX1-dependent telomere maintenance by telomerase. consequently with medical implications in anti-cancer We proceeded to analyze whether blocking this pathway therapy. In this study, we provide evidence that contributes to the cytotoxic effects of anthracyclines on inducible downregulation of PinX1, thereby causing cancer cells. We used annexin staining and flow acute telomere dysfunction, contributes to the cancer- cytometry as a sensitive, quantitative assay to detect killing effects of the chemotherapeutic anthracyclines. cells both in early and late apoptosis. As shown in These findings further suggest that this PinX1- Figure 5a, EPI-induced cell apoptotic death incidences dependent telomere maintenance by a telomerase in HepG2 cells were significantly increased by PinX1 pathway can provide novel druggable targets for silencing, but were decreased remarkably by PinX1 the currently pursued telomere/telomerase-based anti- overexpression. Accordingly, EPI-induced PARP (poly cancer therapies. (ADP-ribose) polymerase) cleavage, a molecular marker Anthracyclines represent a class of chemotherapeutics for cell apoptosis, was significantly enhanced in the that have long been widely used in clinical cancer PinX1 knockdown HepG2 cells but was robustly treatments. DOX is one of the first generation of blocked in PinX1-overexpressing cells (Figure 5b). anthracyclines. By different chemical modifications, Therefore, we conclude that induction of PinX1 DOX can be derived to more than hundreds of analogs, degradation, thereby suppressing telomere maintenance including the EPI used in this study. This property by telomerase, constitutes an anti-cancer action of makes anthracyclines a potential with a still-improvable anthracyclines. therapeutic index. Therefore, they may be regarded as the first evergreen anti-cancer drugs to be developed. The induction of DNA DSBs through the inhibition of topoisomerases has been identified as a major function Discussion of anthracyclines in killing cancers. Interestingly, we showed in this study that anthracyclines also have an Telomere maintenance by telomerase has been identified active role in inducing telomere dysfunction in telomer- as an essential pathway towards human cancer forma- ase-positive cancer cells, and that this may constitute tion. Numerous studies have demonstrated that the one of the anti-cancer effects of these drugs. This interruption of this pathway, either by telomerase activity seems to be unrelated to their known DNA- inhibition or by targeting the telomere structure per se, damaging effect because it is not present in other not only significantly inhibits cancer growth but also topoisomerase poisons, such as camptothecin and Et. greatly enhances cancer sensitivity to the cytotoxic Instead, we propose that this effect is caused by a effects of most currently used anti-cancer chemother- property of anthracyclines which downregulates the apeutics (Nakamura et al., 2005; Shay and Wright, 2006; PinX1 protein expression, which in turn disrupts the Harley, 2008; Uziel et al., 2010). Therefore, it can be interplay between telomerase and telomeres, therefore expected that the development of these telomere/ affecting telomere stability under DNA-damaging telomerase-targeting drugs will greatly improve the stress. Again, such an activity of downregulating the outcome of current anti-cancer therapies. Many efforts PinX1 expression is not found in other kinds of have been made to develop specific telomerase enzy- DNA damage inducers. Therefore, we identify a novel matic inhibitors. Two decades have passed since this effect for anthracyclines as an inhibitor of telomere

Oncogene Anthracyclines disrupt telomere maintenance by telomerase B Zhang et al 9

Figure 5 Modulations of PinX1 expressions affect the cytotoxic effects of EPI in HepG2 cancer cells. (a) Full-length PinX1 for PinX1 overexpression and PinX1 shRNA for PinX1 knockdown were introduced into HepG2 cells through lentiviral infection. Cells were then treated with indicated doses of EPI for 24 h, and EPI-induced cell apoptotic death events were monitored by Alexa Fluor 488 annexin V/PI staining and flow cytometry assays. Viable cells (annexin V and PI negative, D3) are located in the left lower box. Cells undergoing early apoptosis (annexin V positive and PI negative, D4) are located in the right lower box. Cells in the late stage of apoptosis and dead cells (annexin V and PI positive, D2) are located in the right upper box. (b) After treatment, immunoblotting was performed with PARP, PinX1 and tubulin antibody. Lower panels show the ratios of PinX1 to tubulin from densitometric analysis. Mock, uninfected control cells. maintenance by telomerase through suppressing the through two independent mechanisms: (1) by binding to PinX1 expression. the G-quanduplex, and increasing its melting tempera- It is notable that the nature of anthracyclines binding ture, these drugs may hamper telomerase access to its to the G-quanduplex of telomeres may also contribute functional sites at telomeres; and (2) while, by disrupting to the effect of these drugs in inducing telomere telomerase/telomere recruitment through downregulat- dysfunction (Mergny and Helene, 1998). We consider ing the PinX1 expression, anthracyclines inhibit telo- that anthracyclines may cause telomere dysfunction merase association with telomeres, thereby displaying a

Oncogene Anthracyclines disrupt telomere maintenance by telomerase B Zhang et al 10 synergetic effect, which causes telomere dysfunction in telomerase/telomere-delivering pathway can provide cancer cells. novel druggable targets for development of telomerase/ How anthracyclines regulate PinX1 expression still telomere-based anti-cancer drugs. The finding that this remains to be elucidated. Our data suggest that functional pathway can be specifically disrupted by anthracyclines might affect PinX1 protein stability by DOX and EPI not only elucidates a new mechanism for inducing this protein into the ubiquitin–proteasome- the anti-tumor activity of anthracyclines but also mediated degradation pathway. The fact that PinX1 can provides a new clue for development of the new be regulated at the level of protein stability has also been generation of anti-cancer drugs, based on targeting the reported recently. Wang et al. (2010) showed that polo- pathway of telomere maintenance by telomerase. As like kinase 1 can bind and phosphorylate PinX1, thereby anthracyclines can derivate into thousands of analogs by signaling PinX1 degradation through the ubiquitin– different chemical modifications, it is interesting to find proteasome pathway. However, because we found that whether these modifications could produce a de novo suppression of polo-like kinase 1 by small-interfering anthracycline that acts as a potent telomerase functional RNA did not affect anthracycline-induced PinX1 down- inhibitor. regulation (data not shown), we exclude that anthracy- cline stimulates PinX1 degradation through the reported polo-like kinase 1 pathway. It was reported that Materials and methods anthracyclines can activate some kinds of protein kinases, which may phosphorylate different proteins Cell lines and reagents and resulting in their degradation (Laurent and Jaffre- MCF7, HepG2, BJ and U2OS cells were maintained in a zou, 2001; Small et al., 2003). As PinX1 contains many Dulbecco’s modified Eagle’s medium supplemented with 10% potential phosphorylation sites, we speculate that the fetal bovine serum and 1% penicillin–streptomycin at 37 1Cin degradation of PinX1 stimulated by anthracyclines 5% CO2. Cells were treated with Et (Sigma, St Louis, MO, might be through kinase pathways other than the USA), DOX (Sigma) or EPI (Sigma) at the indicated polo-like kinase 1 pathway. Clearly, further works are concentrations. required to clarify the mechanisms of degradation of Antibody sources are as follows: Rabbit polyclonal antibodies against PinX1 were purchased from Proteintech PinX1 by anthracyclines in detail. Group Inc. (Chicago, IL, USA). Rabbit polyclonal antibodies Although we suggest that the inducible downregula- against Pot1 and rabbit mAB against TERT were purchased tion of PinX1, thereby interrupting telomerase binding from Abcam (Cambridge, UK). Rabbit mAB against to telomeres, contributes to the effect of anthracyclines ubiquitin, mouse mAB against PARP, rabbit mAB against on inducing telomere dysfunction in telomerase-depen- phospho-histone H2AX AB were purchased from Cell dent cancer cells, it should be noted that this inducible Signaling Technology (Beverly, MA, USA). Mouse mAB downregulation of PinX1 per se appears not to be the against FLAG were purchased from Sigma. Rabbit polyclonal cause of acute telomere dysfunction induced by anthra- antibodies against tubulin were purchased from Santa cyclines; this is because acute depletion of PinX1 by Cruz Biotechnology (California, CA, USA). Mouse mAB PinX1 shRNA virus infection does not cause obvious against TRF1 were purchased from GeneTex (San Antonio, TX, USA). telomere dysfunction in affected cancer cells under their normal cultural conditions (Figure 3e). This is unlike our previous observations (Zhang et al., 2009) that Construction of PinX1 cDNA and small-hairpin RNA lentiviral chronic suppression of PinX1 eventually results in vector telomere dysfunction in PinX1 shRNA-stable cells, A plasmid-containing PinX1 cDNA was cloned into a pCDH cDNA expression lentivector (System Biosciences, Mountain presumably by telomere shortening. However, in both View, CA, USA), and then co-transfected into 293 cells with a cases, PinX1 depletion renders telomere prone to packaging plasmids mix to generate a recombinant lentivirus instability in the presence of chromosomal DNA in accordance with the manufacturer’s instructions. The damages. One possibility is that internal DNA damages construction of PinX1–shRNA lentivirus vector was reported in chromosomes might elicit a signal to affect telomere previously (Zhang et al., 2009). stability and that this could be antagonized by the presence there of telomerase. Alternatively, telomere- Reverse transcription–PCR associated telomerase might have the effect of suppres- The expression of PinX1 and hTERT mRNA was analyzed sing the activation of DDRs at chromosomal ends by reverse transcriptase–PCR assay. The total RNA (2 mg), under cancer cellular DNA-damaging stresses. No which was extracted using TRIZOL (Invitrogen, Carlsbad, matter what the case is, it is certain that telomerase CA, USA), was used for cDNA synthesis with MMLV maintains telomeres in cancer cells by more than its (Moloney murine leukemia virus) reverse transcriptase known action on counteracting telomeric DNA short- (Promega, Madison, WI, USA). cDNA was subjected to PCR with primers for PinX1 (forward, 50-ATGTCTATGCTG ening. Characterizing these still-unknown functions for 0 0 this enzyme at telomeres will permit new insights into GCTGAACGTCGG-3 and reverse, 5 -GCTCTTCTTCTTG GCCACTCG-30), hTERT (forward, 50-GTATGGCTGCGT how telomerase facilitates cancer development and how GGTGAA-30 and reverse, 50-TCAGTCCAGGATGGTCTT these functions might be used as targets for anti-cancer GAA-30) and GAPDH (forward, 50-GTTCGACAGTCAG therapy. CCGCATCT-30 and reverse, 50-CCTGCAAATGAGCCC In summary, the results here, together with our CAGCCT-30). Amplification consisted of 30 cycles of 30 s at previous findings, suggest that the PinX1-mediated 94 1C, 30 s at 55 1C and 60 s at 72 1C.

Oncogene Anthracyclines disrupt telomere maintenance by telomerase B Zhang et al 11 Immunoprecipitation and immunoblotting considered as TIF positive (Takai et al., 2003; Denchi and de Cells (2  107) were lysated and sonicated in a cell lysis buffer Lange, 2007). (20 mM Tris (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 10 mg/ml protease inhibitor cocktail) on ice. Protein concen- TA assay trations were determined using the BCA kit (Pierce, Rockford, Telomerase activity was measured using a TRAPeze telomer- IL, USA). Equal amounts of cell lysates precleared with ase detection kit (Chemicon, Temecula, CA, USA) in proteinA/G agarose (Calbiochem, Darmstadt, Germany) were accordance with the manufacturer’s instructions. The samples incubated with the primary antibody with gentle rocking were electrophoresed on a 12.5% non-denaturing PAGE overnight at 4 1C. Immune complexes were then precipitated in 0.5  Tris-borate-EDTA buffer and visualized by SYBG by incubating them with proteinA/G agarose for 2 h at 4 1C. green (Invitrogen). Quantification was performed using the Immunoprecipitates were washed five times with a cell lysis Quantity-One software for Bio-Rad Image analysis systems (Bio-Rad Laboratories, Hercules, CA, USA). The amount of buffer. After boiling in 20 ml2 SDS sample buffer, the samples were analyzed by western blotting. Western blot the telomerase product was calculated by using the following analysis followed the standard procedures and was repeated at formula: TPG (U) ¼ (XÀB)/C:(rÀB)/Cr  100, where TPG is least thrice for each protein tested. the total product generated, X sample signal, B background signal intensity, C the 36-bp internal PCR control and r the TSR8 quantification control. Relative TA was calculated by In vivo ubiquitination assay dividing TA at certain point to baseline TA. Cells were lysed in SDS containing a buffer (50 mM Tris, pH 7.5, 0.5 mM EDTA, 1% SDS and 1 mM dithiothreitol), and boiled for 10 min to denature the proteins. Lysates were Annexin V-propidium iodide staining and flow-cytometry clarified by centrifugation at 14 000 r.p.m. on a microcentrifuge analysis for 10 min. Equal amounts of protein were diluted 10 times To determine the amount of apoptosis cells, Alexa Fluor 488 with an immunoprecipitation buffer (50 mM Tris, pH 7.5, 2 mM annexin V/propidium iodide apoptosis assay kit (Invitrogen) EDTA, 150 mM NaCl and 0.5% NP-40 and protease inhibitor was used. Cells treated by the indicated doses of drugs were cocktail). The diluted lysates were applied to immunoprecipi- harvested, rinsed with cold phosphate-buffered saline and tation with Flag M2 beads (Sigma) and immunoblotting with incubated with Alexa Fluor 488 annexin V/propidium iodide an anti-ubiquitin antibody. according to the manufacturer’s instructions. Cells were analyzed by flow cytometry (Becton Dickinson, San Jose, CA, USA). Immunofluorescence Cells grown on coverslips were fixed in 4% paraformaldehyde. Fixed cells were permeabilized with 1% Triton-X100 and blocked with 10% normal goat serum. Cells were then Conflict of interest incubated with primary antibody for 2 h at 37 1C in a humid chamber, washed in phosphate-buffered saline and were then The authors declare no conflict of interest. incubated with the secondary antibodies for 1 h. The secondary antibodies, including DyLight 549-conjugated Goat anti-Mouse IgG (Thermo Scientific, Rockford, IL, USA) and Alexa Fluor 488-conjugated goat anti-rabbit IgG (Invitrogen), Acknowledgements were used at dilutions of 1:300. Nuclei were stained with DAPI (40,60-diamidino-2-phenylinodole, Sigma). Immunofluores- This work is supported by grants from the Chinese NSFC cence images were captured using the Olympus FluoView (30901769), Guangdong National Science Foundation confocal microscopes and analyzed with FV10-ASW viewer (9451008901002371) and the China Postdoctoral Science software (Olympus, Tokyo, Japan). TIFs were evidenced by Foundation (20090450900) to B Zhang, and also partially by co-localizations of the DNA-damage factor gH2AX with the grants from the National Basic Research Program (973 constitutive telomere protein component TRF1. A total of 400 program) of China (No. 2010CB912802) to H-F Kung and cells were counted for each observation, and cells with >5 (No. 2010CB529404) to J-J Huang, the Chinese NSFC spots of gH2AX and TRF1 co-localization signals were (30771074) to H Zhang.

References

Akiyama Y, Maesawa C, Wada K, Fujisawa K, Itabashi T, Noda Y et al. Collins K. (2006). The biogenesis and regulation of telomerase (2004). Human PinX1, a potent telomerase inhibitor, is not involved in holoenzymes. Nat Rev Mol Cell Biol 7: 484–494. human gastrointestinal tract . Oncol Rep 11: 871–874. Cong YS, Wright WE, Shay JW. (2002). Human telomerase and its Autexier C, Greider CW. (1996). Telomerase and cancer: revisiting the regulation. Microbiol Mol Biol Rev 66: 407–425. telomere hypothesis. Trends Biochem Sci 21: 387–391. De Lange T. (1998). Telomeres and senescence: ending the debate. Bianchi A, Shore D. (2008). How telomerase reaches its end: Science 279: 334–335. mechanism of telomerase regulation by the telomeric complex. de Lange T. (2005). Shelterin: the protein complex that shapes and Mol Cell 31: 153–165. safeguards human telomeres. Dev 19: 2100–2110. Blackburn EH. (2000). Telomere states and cell fates. Nature 408: Denchi EL, de Lange T. (2007). Protection of telomeres through 53–56. independent control of ATM and ATR by TRF2 and POT1. Nature Chang Q, Pang JC, Li J, Hu L, Kong X, Ng HK. (2004). Molecular 448: 1068–1071. analysis of PinX1 in medulloblastomas. Int J Cancer 109: 309–314. Gallardo F, Olivier C, Dandjinou AT, Wellinger RJ, Chartrand P. Collins K. (2000). Mammalian telomeres and telomerase. Curr Opin (2008). TLC1 RNA nucleo-cytoplasmic trafficking links telomerase Cell Biol 12: 378–383. biogenesis to its recruitment to telomeres. EMBO J 27: 748–757.

Oncogene Anthracyclines disrupt telomere maintenance by telomerase B Zhang et al 12 Goueli BS, Janknecht R. (2004). Upregulation of the catalytic Nugent CI, Lundblad V. (1998). The telomerase reverse transcriptase: telomerase subunit by the transcription factor ER81 and oncogenic components and regulation. Genes Dev 12: 1073–1085. HER2/Neu, Ras, or Raf. Mol Cell Biol 24: 25–35. Palm W, de Lange T. (2008). How shelterin protects mammalian Harley CB. (2008). Telomerase and cancer therapeutics. Nat Rev telomeres. Annu Rev Genet 42: 301–334. Cancer 8: 167–179. Shay JW, Wright WE. (2006). Telomerase therapeutics for cancer: Hawkins GA, Chang BL, Zheng SL, Isaacs SD, Wiley KE, Bleecker challenges and new directions. Nat Rev Drug Discov 5: 577–584. ER et al. (2004). Mutational analysis of PINX1 in hereditary Small GW, Somasundaram S, Moore DT, Shi YY, Orlowski RZ. prostate cancer. Prostate 60: 298–302. (2003). Repression of mitogen-activated protein kinase (MAPK) Kim NW, Piatyszek MA, Prowse KR, Harley CB, West MD, Ho PL phosphatase-1 by anthracyclines contributes to their antiapoptotic et al. (1994). Specific association of human telomerase activity with activation of p44/42-MAPK. J Pharmacol Exp Ther 307: 861–869. immortal cells and cancer. Science 266: 2011–2015. Takai H, Smogorzewska A, de Lange T. (2003). DNA damage foci at Laurent G, Jaffrezou JP. (2001). Signaling pathways activated by dysfunctional telomeres. Curr Biol 13: 1549–1556. daunorubicin. Blood 98: 913–924. Tomlinson RL, Ziegler TD, Supakorndej T, Terns RM, Terns MP. Lin SY, Elledge SJ. (2003). Multiple tumor suppressor pathways (2006). Cell cycle-regulated trafficking of human telomerase to negatively regulate telomerase. Cell 113: 881–889. telomeres. Mol Biol Cell 17: 955–965. Liu Y, Snow BE, Hande MP, Yeung D, Erdmann NJ, Wakeham A Uziel O, Beery E, Dronichev V, Samocha K, Gryaznov S, Weiss L et al. (2000). The telomerase reverse transcriptase is limiting et al. (2010). Telomere shortening sensitizes cancer cells to selected and necessary for telomerase function in vivo. Curr Biol 10: cytotoxic agents: in vitro and in vivo studies and putative 1459–1462. mechanisms. PLoS One 5: e9132. Mergny JL, Helene C. (1998). G-quadruplex DNA: a target for drug Wang C, Yu J, Yuan K, Lan J, Jin C, Huang H. (2010). Plk1-mediated design. Nat Med 4: 1366–1367. mitotic phosphorylation of PinX1 regulates its stability. Eur J Cell Meyerson M, Counter CM, Eaton EN, Ellisen LW, Steiner P, Caddle Biol 89: 748–756. SD et al. (1997). hEST2, the putative human telomerase catalytic Wong JMY, Kusdra L, Collins K. (2002). Subnuclear shuttling of subunit , is up-regulated in tumor cells and during immortaliza- human telomerase induced by transformation and DNA damage. tion. Cell 90: 785–795. Nat Cell Biol 4: 731–736. Minotti G, Menna P, Salvatorelli E, Cairo G, Gianni L. (2004). Xin H, Liu D, Wan M, Safari A, Kim H, Sun W et al. (2007). TPP1 is a Anthracyclines: molecular advances and pharmacologic develop- homologue of ciliate TEBP-b and interacts with POT1 to recruit ments in antitumor activity and cardiotoxicity. Pharmacol Rev 56: telomerase. Nature 445: 559–562. 185–229. Zhang B, Bai YX, Ma HH, Feng F, Jin R, Wang ZL et al. (2009). Nakamura M, Masutomi K, Kyo S, Hashimoto M, Maida Y, Silencing PinX1 compromises telomere length maintenance as well Kanaya T et al. (2005). Efficient inhibition of human telomerase as tumorigenicity in telomerase-positive human cancer cells. Cancer reverse transcriptase expression by RNA interference sensitizes Res 69: 75. cancer cells to ionizing radiation and chemotherapy. Hum Gene Zhou XZ, Lu KP. (2001). The Pin2/TRF1-interacting protein PinX1 is Ther 16: 859–868. a potent telomerase inhibitor. Cell 107: 347–359.

Oncogene