E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

ava i l ab l e a t www.sc i enced i rec t . com

www.e l sev i e r. com/ loca te /yexc r

Research Article

SNEV overexpression extends the life span of humanendothelial cells

Regina Voglauera, Martina Wei-Fen Changa, Brigitta Dampierb, Matthias Wiesera,Kristin Baumanna, Thomas Sterovskya, Martin Schreiberb,Hermann Katingera, Johannes Grillaria,c,⁎aInstitute of Applied Microbiology, Department of Biotechnology, University of Natural Resources and Applied Life Sciences Muthgasse 18,A-1190 Vienna, AustriabDepartment of Obstetrics and Gynecology, Medical University of Vienna, A-1090 Vienna, AustriacBMT Research Vienna, Austria

A R T I C L E I N F O R M A T I O N

⁎ Corresponding author. Fax: +43 1 36 97 615.E-mail address: j.grillari@iam.boku.ac.at (J

0014-4827/$ – see front matter © 2005 Elsevidoi:10.1016/j.yexcr.2005.11.025

A B S T R A C T

Article Chronology:Received 5 May 2005Revised version received12 November 2005Accepted 21 November 2005Available online 4 January 2006

In a recent screening for genes downregulated in replicatively senescent human umbilicalvein endothelial cells (HUVECs), we have isolated the novel protein SNEV. Since then SNEVhas proven as a multifaceted protein playing a role in pre-mRNA splicing, DNA repair, andthe ubiquitin/proteosome system. Here, we report that SNEVmRNAdecreases in various celltypes during replicative senescence, and that it is increased in various immortalized celllines, as well as in breast tumors, where SNEV transcript levels also correlate with thesurvival of breast cancer patients.Since thesemRNA profiles suggested a role of SNEV in the regulation of cell proliferation, theeffect of its overexpression was tested. Thereby, a significant extension of the cellular lifespan was observed, which was not caused by altered telomerase activity or telomeredynamics but rather by enhanced stress resistance. When SNEV overexpressing cells weretreated with bleomycin or bleomycin combined with BSO, inducing DNA damage as well asreactive oxygen species, a significantly lower fraction of apoptotic cells was found incomparison to vector control cells. These data suggest that high levels of SNEVmight extendthe cellular life span by increasing the resistance to stress or by improving the DNA repaircapacity of the cells.

© 2005 Elsevier Inc. All rights reserved.

Keywords:SNEVPrp19Endothelial cellsStressDNA damageTelomerase activityTelomere length

Introduction

After a defined number of population doublings in vitro,normal human cells enter a state of irreversible growtharrest termed replicative senescence [1]. Although thisgrowth limit on the one hand has been reported to play animportant role as a tumor suppressive mechanism [2], the

. Grillari).

er Inc. All rights reserved

phenotypic changes associated with aged cells on the otherhand are discussed to contribute to aging and age-relateddiseases, such as several types of cancer [3]. Therefore, thecharacterization of senescent cells is of utmost importance.Since cellular phenotype and physiology are determined bythe repertoire of expressed and translated genes, severalstudies have been performed in order to isolate genes and

.

747E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

proteins that are differentially expressed during aging ofnormal human cells [4–6]. Similarly, we have reported theidentification of several differentially expressed genes,among which we found a protein that we termed SNEV for“Senescence Evasion Factor” [7], showing sequence similarityto the yeast Prp19/Pso4 protein, which has been character-ized as essential yeast splicing factor [8]. Recently, we havebeen able to show that SNEV is indeed the human orthologueof Prp19 and as such also essential for the human pre-mRNAsplicing reaction. SNEV forms at least homodimers, and, byinterfering with this oligomerization, the splicing reaction isblocked in vitro [9]. Supporting its role as a human splicingfactor, SNEV has been identified as nuclear matrix proteinNMP200 [10] and as core member of the CDC5L associatedprotein complex by various proteomic approaches to char-acterize the spliceosome or related complexes [11–13].

In addition to its role as a splicing factor, SNEV has alsobeen identified as a U-box protein showing ubiquitin E3 ligaseactivity in vitro [14]. Although its activity as E3 ligase has notbeen proven in vivo so far and the substrate is still unknown, apossible E3 activity in living cells is supported by our recentfinding that SNEV directly interacts with the proteosome,without being an in vitro substrate [15].

Another highly interesting report on the function of SNEVwas published recently, showing a possible role in DNA doublestrand break repair [16], a function that is supported by thenotion that also yeast Prp19 is involved in DNA repair [17].Especially this function inspired our choice to further inves-tigate the role of SNEV during cellular aging, since uncapped,shortened telomeres associate with many DNA damageresponse proteins, and replicative senescence resembles thegrowth arrest induced by activation of DNA damage check-point response pathways [18]. Furthermore, besides telomereshortening and dysfunction, DNA damage per se, induced byexogenous or endogenous stress, may cause the entry intosenescence [19]. However, the cellular response to stresssituations can be variable, including induction of DNA damageresponse pathways leading to senescence or apoptosis as wellas transient cell cycle arrest and DNA damage repair.Whichever the trigger for induction is, entry into irreversiblegrowth arrest is executed by the p53/p21 and/or the pRb/p16pathways [3].

Here, we report the follow up of our previous observationof differential regulation of SNEV mRNA in replicativelysenescent endothelial cells. We show that less SNEVtranscript is found in senescent cells of different tissueorigins, while it is higher expressed in tumor cell lineswhen compared to normal cells and therefore seems tocorrelate with the replicative potential of cells. Therefore,we investigated the effect of increased amounts of SNEV ongrowth behavior of endothelial cells by generating stableSNEV overexpressing cell lines. These cell lines showed anextended life span in vitro that was not due to telomerasereactivation or altered telomere dynamics but rather toincreased resistance against stress and lower levels of DNAdamage. This potential DNA protective activity prompted usto analyze SNEV mRNA levels in tissue samples of breastcancer patients, and we found that high levels of SNEVpositively correlated with the overall survival of the cancerpatients.

Materials and methods

Cells and culture conditions

HUVECs were isolated from umbilical veins as describedpreviously [20] and cultivated in gelatin (1% in PBS) pre-coated culture flasks in M199 medium (Biochrom) supple-mented with 4 mM L-glutamine, 15% fetal calf serum (FCS,Hyclone), 200 μg/ml endothelial cell growth supplement and170 U/ml heparin. HUVECs with extended life span (ESV)have been established using the SV40 early region [21].Human diploid fibroblasts were isolated from skin biopsies[22] and were grown in DMEM/Ham's-F12 (1:1) supplementedwith 4 mM L-glutamine and 10% FCS. Normal human diploidproximal tubular epithelial cells were isolated from renaltissue biopsies [22] and were grown under serum-free cellculture conditions. Immortalization of kidney epithelial cells(HK42/1) was achieved by introduction of the SV40 earlyregion. Confluent adherent cultures were detached using0.1% trypsin and 0.02% EDTA and were passaged with anappropriate split ratio of 1:2 or 1:4 once or twice a weekdepending on confluence and population doubling level (PD).Subsequently, cumulative PDs were calculated as a functionof passage number and split ratio [6]. Normal lymphocyteswere isolated from whole blood using Ficoll gradient andwere activated by phytohemagglutinin.

All other cell lines were obtained from ATCC if not statedotherwise and were cultivated according to the suppliers'specification: HL-60 (human promyelotic leukemia), Molt 3(human acute lymphoblastic leukemia, T-cell), AA-2 (WIL-2human spleen EBV+ B-lymphoblastoid cell line, NationalInstitute of Allergy and Infectious diseases), MT-2 (humanT-cell leukemia, National Institute of Allergy and Infectiousdiseases), A-498 (human kidney carcinoma), HEK293 (humantransformed embryonic kidney). Additionally, 15 humanbreast cancer cell lines (Hs578T, AU565, SK-BR3, ZR-75-1,BT-474, Hcc1143, Hcc1937, MDA-MB-453, HTB131, Kpl-1,MCF-7, T47D, MDA-MB-468, Cal51, Cama-1, MDA-MD-231,MDA-MB-435) as well as 4 normal human breast cell lines(MCF-10A, MCF-10F, HMEC, Hs578Bst) were used in thisstudy.

Growth arrest of quiescent normal cells was induced bycontact inhibition. Therefore, the cells were grown undernormal cell culture conditions and harvested 1 week afterhaving reached confluence.

RNA extraction and Northern blotting

Total RNA from senescent, quiescent or exponentially growingcells was isolated using 1 ml TRIzol® reagent (Life Technolo-gies)/5 × 10

6cells, and mRNA was isolated using paramagnetic

oligodT beads (Dynal) using standard procedures. SNEV cDNAwas DIG labeled by PCR according to the manufacturer'sprotocol (Roche) and then used as probe on Northern blots. 10ng probe/ml high SDS hybridization buffer was used, andhybridization was carried out overnight at 50°C. After strin-gent washing conditions, chemiluminescent signals weredetected using CDP-star measured by the LumiImager(Roche). The Northern blots contained 12 μg total RNA/lane.

748 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

To confirm that equal amounts of RNA were loaded onto thegel, blots were also hybridized to GAPDH.

Quantitative real-time PCR

Total RNA from exponentially growing breast cell lines andhomogenized breast tumor tissue was extracted with TRIr-eagent (Sigma) according to the manufacturer's instructions.Total RNA (10 μg) was reverse transcribed in a final volume of100 μl, using the high-capacity cDNA Archive Kit (AppliedBiosystems). Quantitative PCR was performed on the ABIPrism® 7000 sequence detection system (Applied Biosys-tems). A gene expression assay specific for SNEV (Cat-#Hs00205622_m1; Assay-on-Demand™, Applied Biosystems)suitable for this system was used according to the manufac-turer's instructions [23]. For each data point, duplicatesamples of 3 μl each of template cDNA (equivalent to 100ng of total RNA) were analyzed in parallel in a final reactionvolume of 20 μl, containing 10 μl TaqMan Universal PCRMaster Mix (Applied Biosystems). Duplicate assays of serialdilutions of cDNA from cultured normal breast epithelial cellswere included in the reaction as a standard. Transcript levelswere normalized to those of β-actin, and relative expressionlevels were calculated using the ΔΔCT-method [23].

Plasmid construction, generation of recombinant retrovirusesand cell line establishment

SNEV cDNA was amplified by PCR and ligated into theretroviral plasmid pLXSN. The negative control vectorcontained only the neomycin resistance gene. Retroviralparticles were generated according to the manufacturer'sprotocol (Clontech). For cell line establishment, HUVECs (I38,PD14) were infected with retroviral particles according to themanufacturer's protocol (Clontech). Thereafter, transfectantswere selected using 20 μg geneticin sulfate (G418)/ml. Arisingcell clones (similar number in all experiments) were grown asmass culture designated I38-SNEV-1 and 2 (SNEV overexpres-sing) and I38-neo-1 and 2 (vector control). PDs were calculatedstarting with the first passage post infection (PDpI).

Senescence associated β-galactosidase assay (SA-β-gal)

Cells were fixed with 2% formaldehyde/0,2% glutaraldehyde,and SA-β-Gal activity was measured as described previously[24].

SDS-PAGE and Western blotting

Cells were washed with PBS, detached by applying 50 mMEDTA and centrifuged (10 min, 170 g, room temperature).Thereafter, the cells were lysed in a buffer containing 30 mMTris pH 8.5, 2 M thiourea, 7 M urea and 4% CHAPS. Proteinconcentrationwas determined using 2DQuantKit (Amersham-Biosciences), and the lysates were separated on a NuPAGE 4–12% Bis/Tris polyacrylamide gel (Invitrogen). Electrophoresisand blotting to Protran nitrocellulose membrane (Schleicherand Schuell) were performed in accordance to the manufac-turer's protocol. After blocking with 3% skim milk powder inPBS, the membranes were incubated with the primary

antibodies (rabbit anti-human SNEV antibody 866, kindlyprovided by Paul Ajuh; goat anti-human CD31 antibody,Santa Cruz; mouse anti-p16 antibody, Santa Cruz; mouseanti-p21 antibody, USB; mouse anti-p53 antibody, Sigma;mouse anti-γH2AX antibody, Upstate; mouse anti-β-actinantibody clone AC-15, Sigma; mouse anti-GAPDH antibody,Sigma), followed by incubation with the appropriate peroxi-dase conjugated secondary antibody (all purchased by Sigma).Detection was performed using ECL-Plus (AmershamBios-ciences) and the LumiImager.

Indirect immunofluorescent staining

For immunodetection of SNEV and von Willebrand factor(vWF), cells were grown on cover slides to subconfluence, fixedwith 70% ethanol and stained with the appropriate primaryantibody (rabbit anti-human vWF antibody, Dako Cytomation;rabbit anti-human SNEV antibody 867, kindly provided by PaulAjuh) and a secondary antibody (goat anti-rabbit IgG/FITC-conjugated, Sigma) using standard procedures [21]. Cell nucleiwere counterstained with 4′,6-diamidino-2-phenylindole(DAPI, Molecular Probes). The fluorescence signal was visual-ized using a Leica TCS SP2 confocal microscope.

Matrigel assay—in vitro angiogenesis

The in vitro angiogenesis assay onmatrigel (BectonDickinson)was performed in 24-well cell culture plates according to themanufacturer's instruction using 5 × 104 cells/cm2. Cells weremicroscopically examined 6 h post-inoculation onto matrigel.

Determination of telomere length

Telomere lengths of the established cell lines were deter-mined using a modified flow-FISH (fluorescence in situhybridization and flow cytometry) assay as described previ-ously [25], and the measured mean fluorescence values wereconverted into molecular equivalents of soluble fluorochrome(MESF) using Quantum™24 FITC-labeled standard beads(Bangs Laboratories, Inc.).

Analysis of telomerase activity

Telomerase activity was determined by amodified TRAP assayas described previously [26]. Modifications included theomission of the T4 gene protein, and the primers used wereCXa [27] and TSmodIV (5′ AATCCGTCGAGAACAGTT). Aliquotsof cell extracts comprising 2 × 104 endothelial cells and 1 × 103

HEK293 cells, which were used as a positive control, wereanalyzed. Cell extracts incubated with DNAse-free ribonucle-ase A (10 μg/μl) for 15 min and lysis buffer alone served asnegative controls. Total reaction volume was 50 μl, 15 μl ofwhich was resolved on a 10% polyacrylamide gel (AA/BAA19:1). Gels were stained with SYBR-Green I, and fluorescencewas detected with the LumiImager.

Comet assay

Single- and double-stranded DNA damage in I38-SNEV-1 andI38-neo-1 cells wasmeasured by single-cell gel electrophoresis

Table 1 – Characteristics of the 39 breast cancer patientsanalyzed in this study

Number ofpatients

%

All patients 39 100.0

Histological tumor typeInvasive ductal carcinoma 32 82.1Infiltrating lobular carcinoma 7 17.9

Estrogen receptor statusER+ 18 46.2ER− 21 53.8

Lymph node involvementN0 14 35.9N+ 25 64.1

Tumor gradeG1 1 2.6G2 18 46.2G 2.5 3 7.7G3 17 43.6

Tumor stageT1 13 33.3T2 22 56.4T3 2 5.1T4 2 5.1Median age at time of diagnosis

(range)45.9 (29.2–80.3)

Median overall survival (range) 6.7 (0.1–12.2)Median follow up of patients still

alive11.1 (0.1–12.3)

749E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

(comet assay) under alkaline conditions as described previ-ously [28]. At least 100 randomly selected cells per slide wereexamined for the presence or absence of comets usingfluorescence microscopy. The cells were scored as containingeither no detectable DNA damage (no tail, Cat. I), only low(small tail, Cat. II) or high DNA damage (large tail, Cat. III).

Treatment with bleomycin and BSO and measurement ofapoptosis

Cells were seeded into 12-well cell culture plates 1 day beforetreatment with increasing concentrations of bleomycin(Sigma) for 24 h. To enhance the sensitivity of cells tobleomycin-induced apoptosis, cells were cultivated in thepresence of 1 mM DL-buthionine-sulfoximine (BSO, Sigma) for48 h before treatment with increasing concentrations ofbleomycin for further 24 h. Thereafter, the cells were detachedusing 50 mM EDTA and stained with Annexin V–FITC and PI(Roche) according to themanufacturer's instructions. Analysisof the percentages of apoptotic and necrotic/late-apoptoticcells was performed using a FACS-Calibur and the CellQuestsoftware (Becton Dickinson).

Detection of ROS in endothelial cells

The cells were washed once with PBS and incubatedwith 1 μMchloromethyl-2,7-dichlorodihydrofluorescein diacetate (CM-H2-DCFDA, Molecular Probes) in PBS at room temperature for30 min. After washing twice with PBS, cells were trypsinized,and the oxidation product of CM-H2-DCFDA (DCF) wasanalyzed on a FACS-Calibur. The absolute intracellular oxidantcapacity of the analyzed cell populationwas determined usingthe median fluorescence intensity on the FL1 detector.

Breast cancer study population

53 tumor and lymph node metastasis specimens from 39female breast cancer patients were analyzed retrospectivelyunder protocols approved by the institutional review boardof the Medical University of Vienna (Table 1). 32 patients hadinvasive ductal and 7 infiltrating lobular mammacarcinomas.21 tumors were estrogen receptor (ER) negative (ER−), and18 ER positive (ER+). 25 patients had lymph node metastases(N+), whereas the lymph nodes of 14 patients were notaffected (N0). 1 tumor was grade 1 (G1), 18 were grade 2 (G2),3 were grade 2.5 (G2.5), and 17 were grade 3 (G3).Furthermore, 13 tumors were tumor stage T1, 22 were T2,2 were T3, and 2 were T4. The median age was 45.9 years,the median overall survival was 6.7 years, and the medianfollow up for the patients still alive was 11.1 years (Table 1).RNA from 4 normal breast tissue specimen was obtainedfrom Stratagene #735044 (single donor, 64 years), Ambion#7952 (single donor, 60 years), BioChain Inst, inc. #R1234086(single donor, 34 years) and BD #64037-1 (pool from 3 donors;20, 21 and 28 years).

All statistical analyses were performed with the opensource statistical programming environment R (http://www.r-project.org/). To discriminate the poor and good prognosisgroups, maximally selected log rank statistics was performed[29], and the P value was determined as described [30].

Kaplan–Meier plots were determined with the R survivalpackage, defining overall survival as the time from diagnosisto time of breast cancer-related death.

Results

SNEV mRNA expression levels correlate with the cellulargrowth potential

In order to test if the changes in SNEV mRNA levels are onlyobserved in early passage and senescent HUVECs [7], or if thisis a more general phenomenon, we tested various normal aswell as continuously growing cell lines for their levels of SNEVmRNA. By Northern blotting, we observed that SNEV mRNAwas down-regulated by about 2-fold in senescent HUVECs(normal S) as compared to early passage (referred to as“young”, normal Y) or to early passage quiescent HUVECs(normal Q) (Fig. 1A). These results indicate that SNEV down-regulation is not a mere growth arrest phenomenon but isconnected to the entry into replicative senescence. Addition-ally, SNEV mRNA was upregulated by 1.5-fold in SV40 earlyregion transfected HUVECs that display an extended life span(ESV). A similar 2-fold downregulation of SNEV mRNA wasfound in senescent vs. early passage kidney epithelial cells aswell as human diploid skin fibroblasts. Higher SNEV levelswere also found in SV40 early region immortalized cells (HK

Fig. 1 – SNEV expression level is decreased in senescent and increased in immortalized and transformed cells when comparedto early passage cells. (A) Northern blots using 10 μg of total RNA/lane were analyzed for SNEV and GAPDH expression withDIG-labeled probes. Representative pictures of normal endothelial cells (HUVECs), kidney epithelial cells, skin fibroblasts atdifferent points of serial passaging are shown (early passage/normal Y, senescent/normal S, quiescent/normal Q). Furthermore,various immortalized and transformed cell lines from similar tissue origins were analyzed for their SNEV expression levels:HUVECs with extended life span after SV40 early region introduction (ESV), SV40 early region immortalized (HK42-1) kidneyepithelial cells, transformed kidney cell lines A-498 and HEK293. The lymphoblastoid cell lines AA-2, Molt-3, MT-2 and HL-60were compared to normal activated lymphocytes. Signals of 3 independent experiments were quantitated by the LumiImager.Relative SNEVmRNA levels were calculated as the ratios of SNEV to GAPDHmRNA normalized to the young or normal cells. (B)Quantitative real-time PCR revealed that SNEV mRNA expression is increased in breast cancer cell lines when compared tonormal breast epithelial cell lines. The elevation of SNEV mRNA in tumor cell lines is highly significant (fold change = 2.6,t = −3.26, P = 0.006). (C)Western blot of early passage, quiescent and senescent endothelial cells stained for SNEV. Equal loadingwas again controlled by probing for GAPDH.

750 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

42-1), A-498 and HEK293 when compared to normal youngepithelial cells (Fig. 1A). The observed results demonstratethat SNEV regulation during senescence is not restricted to aspecific cell type and that continuously growing cells containhigher amounts of SNEV mRNA.

Comparison of freshly isolated and activated lympho-cytes to four different leukemia cell lines showed an up to12-fold higher amount of SNEV mRNA in the immortal celllines (Fig. 1A).

Additionally, comparison of the mean SNEV expressionlevels in a panel of 16 human breast cancer cell lines to 4normal breast epithelial cell lines by quantitative real-timePCR showed a highly significant, 2.6-fold upregulation(P = 0.006) in the cancer cell lines (Fig. 1B). Of note, theHs578T cell line was derived from a breast tumor and theHs578Bst cell line from adjacent normal breast tissue of the

same patient, and SNEV expression levels in Hs578T cells were∼2.5-fold higher than in Hs578Bst cells.

Furthermore, a Western blot of early passage, quiescentand senescent human endothelial cells demonstrated that theobserved downregulation of SNEV onmRNA level is also foundon protein level (Fig. 1C).

Taken together, the observed results suggest that SNEVexpression level is correlated with the replicative potential ofvarious human cells in vitro.

SNEV overexpression does not change endothelial celltype-specific markers

The pattern of SNEV regulation described so far prompted usto test if the level of SNEV expression plays a direct role in cellgrowth and immortalization. Therefore, we generated

751E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

retroviral vectors and infected HUVECs two times indepen-dently with retroviruses encoding either SNEV or the neomy-cin resistance gene alone, which resulted in stable SNEVoverexpressing cell lines (I38-SNEV-1, I38-SNEV-2) and in thecorresponding vector control cell lines (I38-neo-1, I38-neo-2).A Western blot revealed that the protein level of SNEV in I38-SNEV-1 cells was increased by about 3-fold as compared to I38-neo-1 cells (Fig. 2A). A similar result was obtained by indirectimmunofluorescent staining for SNEV, where a higher SNEV-specific fluorescent signal in the nucleus was observed in I38-SNEV-1 (Fig. 2B).

In order to characterize changes accompanying SNEVoverexpression, we analyzed expression and localization oftwo endothelial cell-specific marker proteins as well as theneoangiogenic potential. Indirect immunofluorescence forvWF showed a similar cellular distribution and storage inWeibel–Palade bodies in I38-SNEV-1 and I38-neo-1 as well asthe parental I38 cells (Fig. 2C). Furthermore, Western blotanalysis showed similar protein expression levels of CD31 (Fig.2D). Another characteristic of endothelial cells is the potentialto form neoangiogenic webs when inoculated onto matrigel.I38-SNEV-1, I38-neo-1 and I38 cells behaved similarly in thisassay (Fig. 2E). These results suggest that SNEV overexpressingHUVECs had maintained their highly differentiated endothe-lial phenotype.

Overexpression of SNEV in endothelial cells extends thereplicative life span of HUVECs

In order to observe any influence of SNEV on cellular life span,we had a closer look on the proliferative behavior of the newlyestablished cell lines I38-SNEV-1 and I38-SNEV-2 and the

Fig. 2 – SNEV overexpression does not change endothelial ceexpression levels in I38-SNEV-1 and I38-neo-1 using 10 μg of prI38-SNEV-1 after retroviral infection. Blots were also stained for Gfor SNEV shows amore intense fluorescent signal in I38-SNEV-1 nlines did not show changes in endothelial cell type-specific charcorresponding counterpart I38 showed similar staining for vWFchanges in CD31 protein levels by Western Blot analysis. (E) Furtinoculated onto matrigel.

corresponding vector control cell lines. During the first 40 dayspost-infection, the proliferation rates of SNEV overexpressingand vector control cells were identical (Fig. 3A) and did notdiffer from the growth rate of normal endothelial cells at asimilar age (data not shown). However, SNEV overexpressionled to a significant extension of the cellular life span whencompared to the empty vector control transfectants in twoindependent experiments.

I38-neo-1 cells entered growth arrest after around 8 PDspost-infection, where the cells showed the typicalmorphologyof senescent cells and stained positive for SA-β-Gal (Fig. 3B). Incontrast, I38-SNEV-1 cells at the same PD were characterizedby a young morphology with no positive SA-β-Gal staining(Fig. 3B) and entered replicative senescence around PD18pI(Fig. 3C). This corresponds to more than doubling of thereplicative potential after infection with SNEV. Pictures of theparental I38 cells at early passage and senescence are shownas a reference (Fig. 3D).

SNEV overexpression does not alter telomerase activity ortelomere dynamics

Having observed an extension of the replicative life span afterSNEV overexpression, we were further interested in themechanism responsible for this phenomenon. Since telomer-ase reactivation has been reported to extend the cellular lifespan of human cells in vitro [31,32], we asked whetheroverexpression of SNEV induces telomerase activity or influ-ences telomere dynamics. Although very low levels oftelomerase activity were detected in I38 cells at early passage(I38-normal Y; lane 7), neither I38-SNEV-1 (PD6pI; lane 5) norI38-neo-1 (PD6pI; lane3) showed any activity as analyzed using

ll type-specific markers. (A) Western blot analysis of SNEVotein per lane confirming the overexpression of SNEV inAPDH as a loading control. (B) Indirect immunofluorescenceuclei as compared to I38-neo-1. (C) The newly established cellacteristics since I38-SNEV-1, I38-neo-1 as well as the(green, cell nuclei were stained blue with DAPI) and (D) nohermore, the transfectants formed angiogenic webs when

Fig. 3 – SNEV overexpression in HUVECs results in extended replicative life span. (A) The growth curves of the cell linesI38-SNEV-1/-2 and I38-neo-1/-2 demonstrated the extension of the cellular life span after SNEV overexpression. (B) I38-neo-1cells showed the typical morphology of senescent human cells and stained positive for SA-β-Gal at around PD8pI. In contrast,I38-SNEV-1 cells at the same PD have a young phenotype. (C) I38-SNEV-1 entered senescence not before PD18pI. (D) Themorphology of young and senescent normal, non-transfected parental I38 cells is presented for comparison.

752 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

a modified TRAP assay (Fig. 4A), demonstrating that the lifespan extension is not associatedwith telomerase reactivation.

Furthermore, we determined the telomere lengths usingflow-FISH, an assay that is well suited to compare telomerelengths of different cells and cell lines, although it does not

Fig. 4 – The life span extension in I38-SNEV-1 is not due to reactactivity was determined using a TRAP assay, with HEK293 cell lycontrols. I38-neo-1 as well as I38-SNEV-1 cells did not show telodetected in young endothelial cells. (B) Flow-FISH analysis for thkMESF) was performed. While relative telomere lengths of youncontinuous cell propagation of both transfectants led to telomereI38-neo-1 and I38-SNEV-1 was detectable. Error bars indicate the

give absolute values. These experiments revealed that telo-meres had shortened in both established cell lines upon invitro cultivation, since early passage endothelial cells (I38normal Y) had significantly longer telomeres. However, nosignificant difference between I38-SNEV-1 (PD7pI) and I38-

ivation of telomerase or telomere elongation. (A) Telomerasesates as standard and RNAse-treated cell lysates as negativemerase activity, while very slight telomerase activity wase determination of the relative telomere length (expressed asg cells (I38-normal Y) were found to be around 58 kMESF,shortening (∼35 kMESF). No significant difference betweenstandard deviations of three independent experiments.

753E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

neo-1 (PD7pI) cells was observed (Fig. 4B). Therefore, weconclude that the delayed onset of replicative senescencewas not due to changed telomere dynamics.

SNEV reduces the accumulation of spontaneous DNA damage

Since telomere biology seems not to be involved in theextension of the replicative potential of HUVECs, we reasonedthat the observed effect might be a result of SNEV'sinvolvement in DNA damage repair processes [16], whichseem to be linked to replicative senescence [18]. Therefore,we further analyzed the level of spontaneously accumulatedDNA damage of I38-neo-1 cells approaching replicativesenescence (PD7pI) and I38-SNEV-1 cells at the same PD. Onthe basis of the described classifications for DNA damage insingle cells (see Materials and methods), at least 100 cellswere examined and assigned to Cat. I, II or III according to theforms of the DNA comets (Fig. 5A). 65% of I38-SNEV-1 cells didnot show DNA damage and were therefore assigned to Cat. I,whereas only 40% of I38-neo-1 cells were classified Cat. I.Accordingly, lower percentages of I38-SNEV-1 cells withaccumulated DNA damage were observed. Whereas 21% ofI38-SNEV-1 cells were assigned to Cat. II (low DNA damage),30% of I38-neo-1 cells were found to lie within this group.Furthermore, a 2-fold difference in the occurrence of highDNA damage (Cat. III) was observed: only 14% of I38-SNEV-1cells showed a high degree of DNA damage while 30% of I38-neo-1 cells were classified as highly damaged at the DNAlevel (Fig. 5B). These data indicate that a high SNEV levelcorrelates with a decrease of accumulated DNA damageduring in vitro cultivation.

Fig. 5 – SNEV overexpression significantly reduces the basal levethe comet assay, and cells were classified into 3 groups. (A) Cat. Itail), and Cat. III has accumulated high levels of DNA damage. (Bcategorized, and the results clearly demonstrated higher percentato I38-neo-1 at PD7pI (Cat. I), while less cells were classified as Cindependent experiments.

SNEV overexpressing cells are more resistant against stress

In addition to the basal levels of DNA damage, we furtherinvestigated the cellular response after treatment with agentsinducing genotoxic and/or oxidative stress. For this purpose,we used bleomycin, which has been reported to cause single-and double-stranded DNA breaks [33] as well as to activateoxidative stress pathways by producing ROS [34]. In order toset up the experimental conditions for “stressing,” weincubated normal endothelial cells with bleomycin for 24h (Fig. 6A; left picture/−BSO) and measured the percentages ofapoptotic and necrotic cells. Surprisingly, bleomycin up to 75μg/ml only slightly induced apoptosis and even higherconcentrations did not result in percentages of apoptoticcells higher than 16% vs. 7% of the untreated population. Thelevels of necrotic cells remained constantly low (around 5%)in all samples analyzed, indicating low acute cytotoxicity ofbleomycin within the time frame analyzed. These resultsindicate that higher concentrations of bleomycin mightactivate mechanisms regulating the cellular defense againstthe applied stress by inducing γ-glutamyl cysteine synthaseas reported previously [35]. Thus, the ROS level might be keptlow as observed (Fig. 6A; left picture/−BSO). In order tocounteract this phenomenon, we decided to pre-treat thecells with BSO, a pro-oxidative substance that depletes theglutathione stores by inhibiting γ-glutamyl cysteinesynthase. No significant changes in necrosis or apoptosis upto a concentration of 1 mM BSO were observed, although ROSlevels were increased around 4-fold (data not shown).Combined BSO and bleomycin treatment increased the levelsof ROS after 48 h similarly to BSO alone and a bleomycin

l of DNA damage. DNA damage was analyzed by performingshows no damaged DNA (no tail), Cat. II displays a low (small) Analysis of at least 100 individual cells/cell line wasges of undamaged I38-SNEV-1 cells at PD7pI, when comparedat. II and Cat. III. Similar results were obtained in two

Fig. 6 – SNEV overexpressing HUVECs are more resistant against stress induced by BSO and bleomycin. (A) Normal HUVECswere incubated with increasing amounts of bleomycin for 24 h (left panel/−BSO) and apoptosis as well as necrosis weremeasured by flow cytometry using Annexin V–FITC/PI staining. In addition to cell death, the relative amount of intracellularROS was analyzed using CM-H2-DCFDA. After pre-incubation of I38 cells with 1 mM BSO for 48 h, cells were treated withincreasing amounts of bleomycin (right panel/+BSO). These resultswere reproduced in at least three independent experiments.(B) Cultivation of I38-neo-1 and I38-SNEV-1 cell lines in the presence of bleomycin for 24 h (left panel/−BSO) and BSOpre-incubated I38-neo-1 and I38-SNEV-1 cells in the presence of bleomycin (right picture/+BSO). (C) Western blot analysis ofproteins involved in cell cycle regulation and DNA damage response pathways before and after combined BSO and bleomycintreatment was performed, β-actin was used as loading control.

754 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

dose-dependent increase in apoptosis from 16% to 33% wasobserved (Fig. 6A; right picture/+BSO). BSO treatment resultedalso in a higher degree of necrotic cells of around 13%, whichwas, however, independent of bleomycin addition. Theseexperimental conditions were then used to analyze the stressresponse of our established cell lines. Interestingly, aftertreatment of I38-SNEV-1 cells with bleomycin, we did not

observe any induction of apoptosis, since the percentages ofapoptotic cells remained constantly low up to 150 μg/mlbleomycin (Fig. 6B; left picture/−BSO). In contrast, I38-neo-1cells showed apoptosis induction similar to the parental I38cells (11% vs. 6% of the untreated cells). Moreover, the basallevel of apoptosis (7%) was slightly, but significantly higherwhen compared to I38-SNEV-1 cells (3%). No differences in

755E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

necrosis were observed (data not shown). Combined BSO andbleomycin treatment resulted in even higher differences instress resistance (Fig. 6B; right picture/+BSO). While thepercentages of apoptotic cells in I38-neo-1 cells significantlyrose from 8% (cells only treated with BSO) to 25% in aconcentration-dependent manner, percentages of apoptoticcells in I38-SNEV-1 cells did not differ from the negativecontrol that was treated with BSO only. The levels of necrosisin these experiments did not change with increasingbleomycin concentrations (data not shown). Additionally,consistent with the experiment without BSO (Fig. 6B; leftpicture/−BSO), a slight but significant lower rate of sponta-neous apoptosis was detected in I38-SNEV-1 cells treatedwith BSO alone (Fig. 6B; right picture/+BSO; 0 μg/ml bleomy-cin) as compared to I38-neo-1 cells. Taken together, theseresults indicate that SNEV overexpression exerts a protective,anti-apoptotic effect on the cells under the conditions testedand might also slightly reduce the spontaneous apoptoticlevel of the cell population.

In order to test the expression levels of important cellcycle regulators before and after BSO/bleomycin treatment,lysates from I38-neo-1, I38-SNEV-1 and the corresponding“stressed” cells (all at PD5pI) were analyzed by Westernblotting. As a marker for DNA damage, γ-H2AX was alsoanalyzed. As shown in Fig. 6C, “unstressed” cells show lowand similar levels of p53. Upon BSO/bleomycin treatment, p53levels increased, however, p53 induction or stabilization wasmuch more prominent in SNEV overexpressing cells. Thelevel of p21 was lower in untreated I38-SNEV-1 cells ascompared to untreated vector control cells. Surprisingly, p21,a known target gene of p53, was increased after incubationwith BSO/bleomycin, but to the same extent in both cell lines,and did not correlate with the pronounced increase of p53 inI38-SNEV-1. No difference in the basal level of the cell cycleinhibitor p16 was observed between I38-neo-1 and I38-SNEV-1 cells. However, BSO/bleomycin treatment resulted in asignificant upregulation of p16 only in SNEV overexpressingcells. γ-H2AX, a protein phosphorylated in response to DNAdouble-strand breaks [36], was indeed detectable only in BSO/bleomycin treated cells. However, I38-SNEV-1 cells showedsignificantly lower levels of γ-H2AX when compared to I38-neo-1 cells. This is in keeping with the lower percentage ofI38-SNEV-1 cells bearing DNA damage as observed by cometassay (see above). It is intriguing to think that higher levels ofSNEV might therefore enhance the DNA damage repaircapacity and velocity of the cells. Furthermore, we alsocontrolled the SNEV expression during these experimentsand observed increased SNEV levels upon BSO/bleomycintreatment in I38-SNEV-1. Since in these cells both, p53 andp16, were increased upon stress, SNEV might be involved ineither p53 or p16 signaling pathways or in feedback loops.Further experiments are necessary to address thesequestions.

SNEV mRNA levels are a predictor of survival of breast cancerpatients

If higher SNEV levels lead to increased DNA damage repaircapacity of cells, we hypothesized that SNEV might be ofrelevance either for tumorigenesis in vivo or for diagnostic

analyses. Therefore, we analyzed SNEV mRNA levels inprimary tumor and metastasis tissue samples of 39 breastcancer patients by real-time PCR. We have chosen this type ofcancer because of the significantly altered SNEV mRNA levelsin cell lines derived from breast cancer vs. normal mammaryepithelial cells (see above) as well as the overall frequency ofthis tumor type. Consistent with the results of the cell lines,these analyses revealed a 2-fold higher average SNEV tran-script level in the tumor tissues as compared to normal breasttissues with a significance level of P = 0.102 (Fig. 7A). Samplesfrom lymph node metastasis showed lower levels of SNEVthan primary tumors (Fig. 7A). Since we observed greatdifferences in the individual expression levels, the patientswere classified into two subgroups by log rank tests, one withhigh and one with low SNEV mRNA. When we compared theincidence of metastasis in these two patient subgroups, wefound a 1.4-fold higher incidence in the low expression groupfor distant metastasis (57% compared to 40%) as well as forlymph node metastasis (78% vs. 55%) as seen in Fig. 7B.Surprisingly, high mRNA expression of SNEV correlatedsignificantly (P b 0.03) with a higher probability of overallsurvival (good prognosis), while patients with low SNEVexpression levels have a poor prognosis as calculated byKaplan–Meier analysis (Fig. 7C).

Since the tumor tissue samples used for real-time PCRwerenot prepared by microdissection, we were not able to excludethe possibility that the surrounding stromal cells rather thanthe tumor cells themselves overexpress SNEV as a kind ofsurvival factor. Therefore, we tested tissue arrays of thesepatients by immuno-histochemistry and did not detect higherSNEV levels in surrounding cells (data not shown).

Taken together, these data indicate that SNEV might havepotential as prognostic marker for human breast cancer, andthat SNEV transcript levels negatively correlate with theaggressiveness of breast tumors.

Discussion

During the last years, replicative senescence andmechanismsthat help to bypass this growth arrest have been a focus ofintense research, since these processes are suspected to playcritical roles in both aging and tumor suppression [3,37]. Inthis context, especially oncogenes and tumor suppressorgenes have been studied intensively since the developmentand progression of most tumors are associated with geneticalterations of these important cell cycle regulatory proteins[38]. Furthermore, recent data support the fundamentalinvolvement of proteins playing a role in DNA damageresponse and repair in tumor pathogenesis [39].

After the observation that SNEV mRNA expression levelscorrelate with the growth potential of a large variety ofdifferent normal as well as continuously growing tumor celllines, we have analyzed in detail the role of this protein in theregulation of the cellular life span of human cells. Therefore,SNEV was stably overexpressed in normal human endothelialcells, and the cellular phenotype was characterized. Interest-ingly, SNEV overexpression, although not changing endothe-lial cell type-specific characteristics, led to a significantincrease of the life span when compared to the vector control

Fig. 7 – SNEVmRNA expression is a significant predictor of survival in breast cancer patients. (A) SNEVmRNAwas increased inmalignant breast tumors as compared to normal breast epithelial tissues. Quantitative RT-PCR of breast tumor and controltissues showed an increase of SNEV expression (fold change = 2.0, t = −1.87, P value = 0.102). Additionally, SNEV expressionwashigher in breast tumors than in lymph node metastases (fold change = 1.5, t = 1.70, P = 0.096). (B) Breast tumors with low SNEVexpression exhibited a higher metastatic potential. While 78% of the patients in the low expression group had developedlymph node metastases (N+) at the time of diagnosis, lymph node metastases were detectable in only 55% of the highexpression group patients (left panel). 57% of the patients from the low expression group, but only 40% of the high SNEVexpression group developed distant metastases (right panel; M1: distant metastasis, M0: no metastasis). (C) High SNEVexpression was correlated with a good prognosis in human breast cancer as established by Kaplan–Meier analysis. TheKaplan–Meier curve divides the patients into two groups according to SNEV mRNA level. “Low expression” of SNEV (dashedline) causes a poor prognosis for overall survival as compared to “high expression” (solid line).

756 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

cells. Since SNEV overexpressing cells did neither showenhanced telomerase activity nor elongated telomeres, wehad a closer look at differences between I38-neo-1 and I38-SNEV-1 cells in regard to DNA damage. Indeed, the vectorcontrol cells, at the end of their replicative life span, hadaccumulated higher levels of single- and double-strandedDNA damage as analyzed using comet assay, when comparedto SNEV overexpressing cells of the same age. Furthermore,combined treatment of the established cell lines with bleo-mycin and BSO clearly showed that SNEV overexpressioncorrelates with higher stress resistance. Interestingly, elevat-

ed expression levels of p16 and p53 and lower levels of DNAdamage (lower levels of phosphorylated H2AX) in “stressed”I38-SNEV-1 cells indicate that the DNA repair machinery andthe activation of cell cycle checkpoints are tightly connectedprocesses. Such interactions, which already have been de-scribed [40,41], probably induce a transient cell cycle arrest [42]that allows the cells to repair DNA damage before startingreplication.

But why is SNEV upregulated only in I38-SNEV-1 cells uponstress? The most likely explanation for the moment seems tobe that increased basal levels of SNEV decrease the proneness

757E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

of cells to undergo apoptosis by improving the basal “DNAhealth status” of the cells and thus shift the stress response toincreased p16 levels that in turn cause arrest of the cell cycle.This p16 induced cell cycle arrest could be used for repair andas a downstream target SNEV might be strongly induced as afactor necessary for DNA repair. A similar result has beendescribed byMahajan et al. although in this case, the dramaticincrease in SNEV protein upon stress treatment goes back tonormal levels within 4 h [16].

Although further experiments have to be performed inorder to characterize the role of SNEV within theseinterwoven processes in detail, the observed higher stressresistance might help to explain the extended life span ofI38-SNEV-1 cells. Cellular DNA is constantly subjected todamage by extrinsic and intrinsic factors, and DNA double-strand breaks (DSB) have been reported to accumulate withincreasing age in senescent cells [43]. Furthermore, replica-tive senescence resembles the activation of DNA damagecheckpoint response [18]. Therefore, we hypothesize thatSNEV overexpressing cells probably cope better with con-stant culture stress, since DNA damage is repaired moreefficiently. This hypothesis is further strengthened by thefact that the basal levels of apoptosis as well as thepercentages of untreated cells inflicted with DNA damagein I38-SNEV-1 cells are significantly lower than those of thecontrol cells.

A further report supporting our conclusions has beenpublished during the preparation of this manuscript. TheSNEV containing CDC5L spliceosome associated complexdirectly and physically interacts with WRN [44], a DNAhelicase of the RecQ family [45]. This interaction is importantfor repair of DNA interstrand cross-links [44]. A mutation inthe WRN gene is also responsible for the Werner syndrome,and fibroblasts derived from patients of this premature agingsyndrome show a markedly decreased replicative potential[46]. Therefore, an intriguing link between SNEV, DNA repairand the replicative life span has been created.

If SNEV counteracts DNA damage in vitro, then we wouldexpect that it might be involved in tumor pathogenesis invivo. In order to test this idea, we measured SNEV expressionlevels in tissue samples of breast cancer patients. Since thepatient history of the 39 cases tested was followed up for amedian of N10 years after diagnosis, we were able tocorrelate increased SNEV levels to a significantly longeroverall survival. In accordance with the data obtained invitro, more efficient DNA repair might in turn increasegenomic stability of the tumor cells and thus slow or stopfurther mutations that lead to highly aggressive phenotypes.A similar effect was reported for telomerase, a key regulatorof cellular life span extension, which is reactivated invirtually all tumors and thought to be essential for tumor-igenesis [26,47]. However, by association with telomeres,telomerase enhances genomic stability and DNA repair [48].Furthermore, Bartkova et al. reported recently that earlyprecursor lesions of different human tumors commonlyexpress markers of activated DNA damage response path-ways [49]. These findings allow the hypothesis that DNAdamage checkpoints, activated at early stages of tumorigen-esis, may lead to cell cycle blockade and therefore areessential in the protection against tumor progression.

However, SNEV is an obviously multitalented regulator ofthe replicative potential, since several other cellular functionsare connected with SNEV as mentioned in the introduction.Although the data acquired so far suggest that SNEV's role asDNA damage repair factor might be crucial for life spanextension of HUVECs, we cannot exclude that one of the otherfunctions of SNEV is involved in the complex regulatorynetwork of replicative senescence. A few hints are available,that pre-mRNA splicingmight be of importance, since Guo andcoworkers describe differential mRNA levels of splicing factors[50] during cellular aging in vitro. Similarly, alternativesplicing is implicated in hTERT regulation [51,52]. However,SNEV is the first splicing factor that seems to be functionallylinked to replicative senescence. In regard to SNEV's catalyticactivity of a ubiquitin E3 ligase, one other E3 enzyme hasrecently been identified to influence cellular aging: Smurf-1 isan E3 ligase that upon overexpression reduces the life span offibroblasts. However, the functional domain for this effect isnot the one that is responsible for Smurf-1's ubiquitin ligaseactivity [53].

Further analysis of SNEV and its functional domains in thecontext of replicative senescence will help to get detailedinsights into themechanistic role of SNEV and its influence oncellular aging.

Acknowledgments

This work was generously supported by the Polymun Scien-tific Immunbiolog. Forschungs GMBH, Vienna, and NFNproject S093-06 of the Austrian Science Fund. Work in M.S.'slaboratory is funded within the Austrian Genome ResearchProgram GEN-AU by funds of the Austrian Ministry ofEducation, Science, and Culture. We want to thank StefanGross and Doris Hofer for the excellent technical support.

R E F E R E N C E S

[1] L. Hayflick, P.S. Moorhead, The serial cultivation of humandiploid cell strains, Exp. Cell Res. 25 (1961) 585–621.

[2] J. Campisi, Aging, tumor suppression and cancer: highwire-act! Mech. Ageing Dev. 126 (2005) 51–58.

[3] J. Campisi, Senescent cells, tumor suppression, andorganismal aging: good citizens, bad neighbors, Cell 120 (2005)513–522.

[4] T. Kumazaki, T. Sakano, T. Yoshida, K. Hamada, H. Sumida, Y.Teranishi, M. Nishiyama, Y. Mitsui, Enhanced expression ofmitochondrial genes in senescent endothelial cells andfibroblasts, Mech. Ageing Dev. 101 (1998) 91–99.

[5] D.N. Shelton, E. Chang, P.S. Whittier, D. Choi, W.D. Funk,Microarray analysis of replicative senescence, Curr. Biol. 9(1999) 939–945.

[6] M.W. Chang, J. Grillari, C. Mayrhofer, K. Fortschegger, G.Allmaier, G. Marzban, H. Katinger, R. Voglauer, Comparison ofearly passage, senescent and hTERT immortalizedendothelial cells, Exp. Cell Res. 309 (2005) 121–136.

[7] J. Grillari, O. Hohenwarter, R.M. Grabherr, H. Katinger,Subtractive hybridization of mRNA from early passage andsenescent endothelial cells, Exp. Gerontol. 35 (2000) 187–197.

[8] W.Y. Tarn, K.R. Lee, S.C. Cheng, Yeast precursor mRNAprocessing protein PRP19 associates with the spliceosomeconcomitant with or just after dissociation of U4 small

758 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

nuclear RNA, Proc. Natl. Acad. Sci. U. S. A. 90 (1993)10821–10825.

[9] J. Grillari, P. Ajuh, G. Stadler, M. Löscher, R. Voglauer, W. Ernst,J. Chusainow, F. Eisenhaber, M. Pokar, K. Fortschegger, M.Grey, A. Lamond, H. Katinger, SNEV is an evolutionarilyconserved splicing factor whose oligomerization is necessaryfor spliceosome assembly, Nucleic Acids Res. 33 (2005)6868–6883.

[10] J. Gotzmann, C. Gerner, M. Meissner, K. Holzmann, R. Grimm,W. Mikulits, G. Sauermann, hNMP 200: a novel humancommon nuclear matrix protein combining structural andregulatory functions, Exp. Cell Res. 261 (2000) 166–179.

[11] P. Ajuh, B. Kuster, K. Panov, J.C. Zomerdijk, M. Mann, A.I.Lamond, Functional analysis of the human CDC5L complexand identification of its components by mass spectrometry,EMBO J. 19 (2000) 6569–6581.

[12] E.M. Makarov, O.V. Makarova, H. Urlaub, M. Gentzel, C.L. Will,M. Wilm, R. Luhrmann, Small nuclear ribonucleoproteinremodeling during catalytic activation of the spliceosome,Science 298 (2002) 2205–2208.

[13] O.V. Makarova, E.M. Makarov, H. Urlaub, C.L. Will, M. Gentzel,M. Wilm, R. Luhrmann, A subset of human 35S U5 proteins,including Prp19, function prior to catalytic step 1 of splicing,EMBO J. 23 (2004) 2381–2391 (electronic publication 2004 Jun.03).

[14] S. Hatakeyama, M. Yada, M. Matsumoto, N. Ishida, K.I.Nakayama, U-Box proteins as a new family ofubiquitin-protein ligases, J. Biol. Chem. 276 (2001)33111–33120.

[15] M. Loscher, K. Fortschegger, G. Ritter, M. Wostry, R. Voglauer,J.A. Schmid, S. Watters, A.J. Rivett, P. Ajuh, A.I. Lamond, H.Katinger, J. Grillari, Interaction of U-box E3 ligase SNEV withPSMB4, the beta7 subunit of the 20 S proteasome, Biochem. J.388 (2005) 593–603.

[16] K.N. Mahajan, B.S. Mitchell, Role of human Pso4 inmammalian DNA repair and association with terminaldeoxynucleotidyl transferase, Proc. Natl. Acad. Sci. U. S. A. 100(2003) 10746–10751.

[17] M. Grey, A. Dusterhoft, J.A. Henriques, M. Brendel, Allelism ofPSO4 and PRP19 links pre-mRNA processing withrecombination and error-prone DNA repair in Saccharomycescerevisiae, Nucleic Acids Res. 24 (1996) 4009–4014.

[18] F. d'Adda di Fagagna, P.M. Reaper, L. Clay-Farrace, H. Fiegler,P. Carr, T. Von, G. Zglinicki, N.P. Saretzki, S.P. Carter, A DNAdamage checkpoint response in telomere-initiatedsenescence, Nature 426 (2003) 194–198 (electronic publication2003 Nov. 5).

[19] A. Di Leonardo, S.P. Linke, K. Clarkin, G.M.Wahl, DNA damagetriggers a prolonged p53-dependent G1 arrest and long-terminduction of Cip1 in normal human fibroblasts, Genes Dev.8 (1994) 2540–2551.

[20] M.A. Gimbrone, E.J. Shefton, S.A. Cruise, Isolation andprimary culture of endothelial cells from human umbilicalveins, TCA Manual 4 (1978) 813–817.

[21] O. Hohenwarter, E. Zinser, C. Schmatz, F. Ruker, H. Katinger,Influence of transfected SV40 early region on growth anddifferentiation of human endothelial cells, J. Biotechnol. 25(1992) 349–356.

[22] A. Doyle, J.B. Griffiths, D.G. Newell, Cell and Tissue Culture:Laboratory Procedures, John Wiley and Sons, Ltd., WestSussex, 1996.

[23] U. Vinatzer, B. Dampier, B. Streubel, M. Pacher, M.J. Seewald,C. Stratowa, K. Kaserer, M. Schreiber, Expression of HER2 andthe co-amplified genes GRB7 and MLN64 in Human BreastCancer: Quantitative Real-Time RT-PCR as a DiagnosticAlternative to IHC and FISH, Clin. Cancer Res. 11 (2005)8348–8357.

[24] G.P. Dimri, X. Lee, G. Basile, M. Acosta, G. Scott, C. Roskelley,E.E. Medrano, M. Linskens, I. Rubelj, O. Pereira-Smith, et al.,

A biomarker that identifies senescent human cells in cultureand in aging skin in vivo, Proc. Natl. Acad. Sci. U. S. A. 92(1995) 9363–9367.

[25] R. Voglauer, J. Grillari, K. Fortschegger, M. Wieser, T.Sterovsky, P. Gunsberg, H. Katinger, R. Pfragner,Establishment of human fibroma cell lines from a MEN1patient by introduction of either hTERT or SV40 early region,Int. J. Oncol. 26 (2005) 961–970.

[26] N.W. Kim, M.A. Piatyszek, K.R. Prowse, C.B. Harley, M.D. West,P.L. Ho, G.M. Coviello, W.E. Wright, S.L. Weinrich, J.W. Shay,Specific association of human telomerase activity withimmortal cells and cancer, Science 266 (1994) 2011–2015.

[27] M.L. Falchetti, A. Levi, P. Molinari, R. Verna, E. D'Ambrosio,Increased sensitivity and reproducibility of TRAP assay byavoiding direct primers interaction, Nucleic Acids Res. 26(1998) 862–863.

[28] M. Wojewodzka, I. Buraczewska, M. Kruszewski, A modifiedneutral comet assay: elimination of lysis at high temperatureand validation of the assay with anti-single-stranded DNAantibody, Mutat. Res. 518 (2002) 9–20.

[29] B. Lausen, M. Schumacher, Maximally selected rank statistics,Biometrics 48 (1992) 73–85.

[30] T. Hothorn, B. Lausen, On the exact distribution of maximallyselected rank statistics, Comput. Stat. Data Anal. 43 (2003)121–137.

[31] A.G. Bodnar, M. Ouellette, M. Frolkis, S.E. Holt, C.P. Chiu, G.B.Morin, C.B. Harley, J.W. Shay, S. Lichtsteiner, W.E. Wright,Extension of life-span by introduction of telomerase intonormal human cells [see comments], Science 279 (1998)349–352.

[32] J. Yang, E. Chang, A.M. Cherry, C.D. Bangs, Y. Oei, A. Bodnar, A.Bronstein, C.P. Chiu, G.S. Herron, Human endothelial cell lifeextension by telomerase expression, J. Biol. Chem. 274 (1999)26141–26148.

[33] S.M. Hecht, Bleomycin: new perspectives on the mechanismof action, J. Nat. Prod. 63 (2000) 158–168.

[34] B.I. Sikic, Biochemical and cellular determinants of bleomycincytotoxicity, Cancer Surv. 5 (1986) 81–91.

[35] R.M. Day, Y.J. Suzuki, J.M. Lum, A.C. White, B.L. Fanburg,Bleomycin upregulates expression ofgamma-glutamylcysteine synthetase in pulmonary arteryendothelial cells, Am. J. Physiol.: Lung Cell. Mol. Physiol.282 (2002) L1349–L1357.

[36] I.M. Ward, J. Chen, Histone H2AX is phosphorylated in anATR-dependent manner in response to replicational stress,J. Biol. Chem. 276 (2001) 47759–47762.

[37] J. Campisi, Suppressing cancer: the importance of beingsenescent, Science 309 (2005) 886–887.

[38] B. Vogelstein, K.W. Kinzler, Cancer genes and the pathwaysthey control, Nat. Med. 10 (2004) 789–799.

[39] N. Motoyama, K. Naka, DNA damage tumor suppressor genesand genomic instability, Curr. Opin. Genet. Dev. 14 (2004)11–16.

[40] M. Giannattasio, F. Lazzaro, M.P. Longhese, P. Plevani, M.Muzi-Falconi, Physical and functional interactions betweennucleotide excision repair and DNA damage checkpoint,EMBO J. 23 (2004) 429–438.

[41] T.M. Runger, I. Vergilis, P. Sarkar, R.A. DePinho, N.E. Sharpless,Howdisruptionof cell cycle regulating genesmightpredisposeto sun-induced skin cancer, Cell Cycle 4 (2005) 643–645.

[42] D. Decraene, P. Agostinis, A. Pupe, P. de Haes, M. Garmyn,Acute response of human skin to solar radiation: regulationand function of the p53 protein, J. Photochem. Photobiol., BBiol. 63 (2001) 78–83.

[43] M. Chevanne, R. Caldini, D. Tombaccini, A. Mocali, G. Gori, F.Paoletti, Comparative levels of DNA breaks and sensitivity tooxidative stress in aged and senescent human fibroblasts: adistinctive pattern for centenarians, Biogerontology 4 (2003)97–104.

759E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 7 4 6 – 7 5 9

[44] N. Zhang, R. Kaur, X. Lu, X. Shen, L. Li, R.J. Legerski, The PSO4MRNA splicing and DNA repair complex interacts with WRNfor processing of DNA interstrand cross-links, J. Biol. Chem.(2005) 40559–40567.

[45] M.D. Gray, J.C. Shen, A.S. Kamath-Loeb, A. Blank, B.L. Sopher,G.M. Martin, J. Oshima, L.A. Loeb, The Werner syndromeprotein is a DNA helicase, Nat. Genet. 17 (1997) 100–103.

[46] S. Goldstein, C.B. Harley, In vitro studies of age-associateddiseases, Fed. Proc. 38 (1979) 1862–1867.

[47] J.W. Shay, S. Bacchetti, A survey of telomerase activity inhuman cancer, Eur. J. Cancer 33 (1997) 787–791.

[48] G.G. Sharma, A. Gupta, H. Wang, H. Scherthan, S. Dhar, V.Gandhi, G. Iliakis, J.W. Shay, C.S. Young, T.K. Pandita,hTERT associates with human telomeres and enhancesgenomic stability and DNA repair, Oncogene 22 (2003)131–146.

[49] J. Bartkova, Z. Horejsi, K. Koed, A. Kramer, F. Tort, K. Zieger,P. Guldberg, M. Sehested, J.M. Nesland, C. Lukas, T. Orntoft,

J. Lukas, J. Bartek, DNA damage response as a candidateanti-cancer barrier in early human tumorigenesis, Nature434 (2005) 864–870.

[50] S. Guo, Z. Zhang, T. Tong, Cloning and characterization ofcellular senescence-associated genes in human fibroblasts bysuppression subtractive hybridization, Exp. Cell Res. 298(2004) 465–472.

[51] G.A. Ulaner, J.F. Hu, T.H. Vu, L.C. Giudice, A.R. Hoffman,Tissue-specific alternate splicing of human telomerasereverse transcriptase (hTERT) influences telomere lengthsduring human development, Int. J. Cancer 91 (2001)644–649.

[52] X. Yi, J.W. Shay, W.E. Wright, Quantitation of telomerasecomponents and hTERT mRNA splicing patterns in immortalhuman cells, Nucleic Acids Res. 29 (2001) 4818–4825.

[53] H. Zhang, S.N. Cohen, Smurf2 up-regulation activatestelomere-dependent senescence, Genes Dev. 18 (2004)3028–3040.