Telomere dynamics in human mesenchymal stem cells after exposure to acute oxidative stress

DNA Repair 11 (2012) 774–779

Contents lists available at SciVerse ScienceDirect

DNA Repair journal homepage: www.elsevier.com/locate/dnarepair

Brief report

Telomere dynamics in human mesenchymal stem cells after exposure to acute oxidative stress Maria Harbo a,b,c,∗ , Steen Koelvraa a,b,c , Nedime Serakinci a,b , Laila Bendix a,c,d a

Department of Clinical Genetics, Vejle/Lillebaelt Hospital, Kabbeltoft 25, 7100 Vejle, Denmark Institute of Regional Health Research, University of Southern Denmark, J.B. Winsloews Vej 19, 3., 5000 Odense C, Denmark Danish Aging Research Center, University of Southern Denmark, J.B. Winsloews Vej 9B, 5000 Odense C, Denmark d Department of Social Medicine, Institute of Public Health, University of Copenhagen, PO Box 2099, 1014 Copenhagen K, Denmark b c

a r t i c l e

i n f o

Article history: Received 9 February 2012 Received in revised form 14 June 2012 Accepted 14 June 2012 Available online 9 July 2012 Keywords: Telomere length Universal STELA TRF assay Oxidative stress Cellular senescence Human mesenchymal stem cells

a b s t r a c t A gradual shortening of telomeres due to replication can be measured using the standard telomere restriction fragments (TRF) assay and other methods by measuring the mean length of all the telomeres in a cell. In contrast, stress-induced telomere shortening, which is believed to be just as important for causing cellular senescence, cannot be measured properly using these methods. Stress-induced telomere shortening caused by, e.g. oxidative damage happens in a stochastic manner leaving just a few single telomeres critically short. It is now possible to visualize these few ultra-short telomeres due to the advantages of the newly developed Universal single telomere length assay (STELA), and we therefore believe that this method should be considered the method of choice when measuring the length of telomeres after exposure to oxidative stress. In order to test our hypothesis, cultured human mesenchymal stem cells, either primary or hTERT immortalized, were exposed to sub-lethal doses of hydrogen peroxide, and the short term effect on telomere dynamics was monitored by Universal STELA and TRF measurements. Both telomere measures were then correlated with the percentage of senescent cells estimated by senescence-associated ␤-galactosidase staining. The exposure to acute oxidative stress resulted in an increased number of ultra-short telomeres, which correlated strongly with the percentage of senescent cells, whereas a correlation between mean telomere length and the percentage of senescent cells was absent. Based on the findings in the present study, it seems reasonable to conclude that Universal STELA is superior to TRF in detecting telomere damage caused by exposure to oxidative stress. The choice of method should therefore be considered carefully in studies examining stress-related telomere shortening as well as in the emerging field of lifestyle studies involving telomere length measurements. © 2012 Elsevier B.V. All rights reserved.

1. Introduction In culture, most human cells can undergo only a limited number of cellular divisions before they enter senescence. This characteristic is known as the Hayflick limit [1,2]. Telomeres are the outmost ends of linear chromosomes and have been strongly linked to this phenomenon. Together with the protein complex, Shelterin, telomeres form a structure that protects chromosomes against being

Abbreviations: hMSCs, human mesenchymal stem cells; PD, population doubling; SA-␤-gal, senescence-associated ␤-galactosidase; STELA, single telomere length assay; TRF, telomere restriction fragments. ∗ Corresponding author at: Department of Clinical Genetics, Vejle/Lillebaelt Hospital, Kabbeltoft 25, 7100 Vejle, Denmark. Tel.: +45 7940 6556; fax: +45 7940 6871. E-mail addresses: [email protected] (M. Harbo), [email protected] (S. Koelvraa), [email protected] (N. Serakinci), [email protected] (L. Bendix). 1568-7864/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.dnarep.2012.06.003

recognized as double stranded DNA breaks [3,4]. Telomeres shorten gradually with each cellular division [5–7] due to what is known as the end replication problem [8,9]. A more stochastic and extensive shortening of telomeres is induced by various stimuli, including oxidative damage [10–13]. If a single telomere becomes too short to protect its chromosome end, a DNA damage pathway is initiated, ultimately leading to cellular senescence [14]. In accordance with this, Brandl et al. demonstrated that a prolonged culture of mesenchymal stem cells under low chronic stress increased the shortening rate of the mean telomere length [15]. They also showed that short-term, sub-lethal doses of the oxygen radical hydrogen peroxide (H2 O2 ) caused significantly shortened telomeres, but that this happened gradually with time and not immediately after exposure. This lead the authors to suggest that oxidative damage may be retained in the telomeres in the form of oxidized bases, abasic sites or nucleotide gaps [15]. We suggest that sub-lethal doses of H2 O2 actually do cause immediate damage to the telomeres such as single- and

M. Harbo et al. / DNA Repair 11 (2012) 774–779

double-stranded breaks, but at the time of exposure it only results in the formation of a few short telomeres. Such damaged, excessively shortened telomeres may still be long enough to protect the chromosome end and may therefore not immediately affect cell growth. However, as the cell proliferates, the telomeres will at an earlier time reach a critically short length, causing “premature” senescence [16]. Since the telomere restriction fragments (TRF) assay has been the standard method for measuring telomere length [17,18], such “scars” in the cells in the shape of a few unusually short telomeres may very well have gone undetected. This is due to the fact that the TRF assay measures the mean length of all telomeres in hundred thousands of cells. The method is therefore much better suited for estimating replication-associated shortening where telomere loss is more or less synchronized on all chromosomes. In contrast, telomere shortening due to oxidative damage can occur on a single telomere in a cell, leaving only this particular telomere ultra-short and therefore not affecting the mean telomere length considerably. The TRF assay as well as other methods focusing only on mean telomere length like quantitative PCR [19] and the original single telomere length assay (STELA) [20] are therefore not suitable for the detection of such a single, detrimental [21–25], ultra-short telomere. In accordance with this hypothesis, we have observed a strong correlation between senescent cell markers and the load of ultrashort telomeres in fibroblast cell cultures with mean lengths far above what is generally believed to induce senescence [22]. The load of ultra-short telomeres was here estimated using our recently developed method, Universal STELA, which exclusively measures the short telomeres within a cell population [22]. Furthermore, we have shown that irradiation causes the load of ultra-short telomeres to increase even in highly telomerase positive human mesenchymal stem cells (hMSCs) with very long mean telomere lengths [26]. We, therefore, believe that Universal STELA has the potential to detect even a very minor fraction of ultra-short telomeres generated by both very recent and earlier exposure to DNA damaging agents. Recently, we have demonstrated a strong correlation between load of ultra-short telomeres, cellular senescence and osteoarthritis in articular cartilage from human knees [27]. This suggests a role of stress-induced telomere shortening in the development of osteoarthritis, since cartilage is considered a post-mitotic tissue where cell proliferation is virtually absent. To further investigate this hypothesis and to verify whether Universal STELA provides a way to distinguish replication-associated shortening from the more stochastic shortening of single telomeres caused by cellular stressors, we have in the present communication exposed cultured hMSCs to sub-lethal doses of oxidative stress and followed the short term effect on telomere dynamics in culture and its correlation with cellular senescence. hMSCs are the progenitor for chondrocytes and are believed to be recruited to the site of cartilage damage at some point in the development of osteoarthritis. We further included immortalized cells transduced with hTERT [28] to elucidate whether the presence of extremely long telomeres influenced oxidative stress-induced telomere damage.

2. Materials and methods 2.1. Cell cultures Telomerase-negative primary hMSCs (mean telomere length around 5 kb) and telomerase-positive hMSC-telo1 (mean telomere length around 18 kb) [28] were cultivated in high glucose (4.5 g/L) Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum, 100 U/mL of penicillin and streptomycin and 2 mM

775

of L-glutamine (GIBCO, Invitrogen, Denmark) as described by Serakinci et al. [28]. A total of five experiments were performed: three with primary hMSCs (P1, P2 and P3) and two with hMSC-telo1 (T1 and T2). Four cultures were established per cell line. At time zero two of these cultures were harvested for DNA purification. The two remaining cultures were incubated another 10 min at 37 ◦ C; one in media added H2 O2 to a final concentration of 4.5 mM, and one in standard media (control). 4.5 mM H2 O2 for 10 min was found to cause temporary growth arrest in a large proportion of cells of both cell types, but with the main part of the cells remaining attached and seemingly viable. After 10 min both cultures were washed in PBS, trypsinized and counted. One half of each culture was washed again in PBS and stored at −20 ◦ C for DNA purification, while the other halves of the cultures were re-seeded in fresh media. After approximately one cell population doubling (PD), the cells were again harvested, counted and split in two: one for DNA purification, and one for re-seeding. This procedure was repeated four or five times for both treated and non-treated cultures. For parallel staining of the cells with senescence-associated ␤-galactosidase (SA-␤-gal) [29] 10,000 of the counted cells were seeded in slide flasks every time the cultures were split. P1 and T1 were performed without ␤-gal staining and without a non-treated culture as control. 2.2. DNA purification Cell samples were purified according to a common desalting procedure. Briefly, cells were lysed overnight at 55 ◦ C in lysis buffer (0.01 M Tris, 1 mM EDTA, 0.15 M NaCl and 0.5% SDS in sterile water, pH 10.5) containing 200 ␮g/mL Proteinase K (Fisher Scientific, Denmark). 6 M NaCl was added in the ratio of 1:3, and the mix was centrifuged 15 min at 4 ◦ C. DNA was precipitated with isopropanol and washed in 70% ethanol followed by dilution in TE-buffer. 2.3. SA-ˇ-gal measurements of senescence Cells cultured in slide flasks were washed in PBS, fixed for 3–5 min in 3.7% paraformaldehyde and washed again in PBS followed by incubation for 15–17 h at 37 ◦ C with fresh ␤-gal solution containing 1 mg of 5-bromo-4-chloro-3-indolyl ␤-D-galactoside (X-Gal; Sigma–Aldrich, Denmark) per mL, 40 mM citric acid (pH 6.0), 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 150 mM NaCl, 2 mM MgCl2 . Following incubation, cells were washed in PBS and evaluated using a standard light microscope. The number of SA-␤-gal positive cells were counted and expressed as percentage of total number of cells. 2.4. TRF measurements of mean telomere length The TRF assay was performed with the TeloTAGGG Telomere Length Assay Kit (Roche Applied Science, Denmark) according to the manufacturer’s manual. 3 ␮g of isolated DNA was digested with the restriction enzymes HinfI and RsaI (Roche Applied Science, Denmark). The digested DNA was separated by gel electrophoresis overnight on a 0.70% TBE Seakem agarose gel and detected by Southern blotting. TRF lengths were calculated based on a DIG-labeled molecular weight marker (Roche Applied Science, Denmark) using VisionWorksLS Acquisition and Analysis Software (UVP, AH Diagnostics, Denmark) and presented as mean telomere lengths. 2.5. Universal STELA measurements of ultra-short telomeres Universal STELA was performed as described by Bendix et al. [22]. In short, purified DNA was digested by a 1:1 mixture of the

776

M. Harbo et al. / DNA Repair 11 (2012) 774–779

A

B

β-gal staining - primary hMSC

β-gal staining - hMSC-telo1

% β-gal+

% β-gal+

100

100

90

90 80

80 70

P2 control

60

P2 treated

70 60

T2 control

50

50 40

P3 control

40

30

P3 treated

30

20

20

10

10

T2 treated

0

0 pre

10 min

1 PD

2 PD

3 PD

4 PD

pre

10 min 1 PD

2 PD

3 PD

4 PD

5 PD

Fig. 1. Effects of H2 O2 exposure on frequency of senescent cells in (A) primary hMSC cultures and (B) hMSC-telo1 cultures. Senescence is expressed as percentage of cells staining positive for ␤-gal. pre, before exposure to H2 O2 ; P, primary hMSC; T, hMSC-telo1; control, non-treated cell culture; treated, cell culture treated with H2 O2 .

restriction enzymes MseI and NdeI (NEB, Medinova, Denmark). After digestion, a double-stranded synthetic oligonucleotide with a sticky end corresponding to MseI and NdeI digests was annealed and subsequently ligated to the proximal end of the telomeric fragments. Next, a single-stranded oligonucleotide with a part of its sequence complementary to the telomeric overhang was ligated to the distal end of the telomeric fragment. PCR was performed using the two ligated oligonucleotides as targets for the PCR primers. Subsequently, detection and size calculation of the telomeric products was carried out as described for the TRF assay. The number of bands at a length below 1500 bp were counted and calibrated with regard to the PCR template concentration and presented as the number of telomeres below a length of 1500 bp per genome equivalent of template DNA. Since CV for Universal STELA is approximately 10%, samples were always run on the same gels. This has shown to limit the variation [22]. 2.6. Statistical analysis Correlation analysis between pairs of data was performed with Spearman’s correlation test. P ≤ 0.05 was considered significant. Stata 11 was used for the analysis (Stata Corp, College Station, Texas, USA).

occurred parallel with, but prior to a similar increase in control samples. 3.2. Effects of H2 O2 exposure on mean telomere length We did not notice a significant difference in mean telomere length between treated and non-treated cells (Fig. 2A and B). Both treated and non-treated cultures of primary hMSC and hMSC-telo1 slightly declined in mean telomere length over time. 3.3. Effects of H2 O2 exposure on frequency of ultra-short telomeres In both primary hMSCs and hMSC-telo1 cultures a rapid increase in number of ultra-short telomeres was observed within the first 10 minutes of H2 O2 treatment (Fig. 2C and D). In primary hMSCs this number dropped again to baseline within one PD, but increased again in the following PD to remain at a level above baseline (Fig. 2C). The number of ultra-short telomeres remained higher in the treated hMSC-telo1 cultures compared to the non-treated cultures up until PD 5 where it reached baseline again (Fig. 2D). 3.4. Correlations between senescence and mean telomere length as well as number of ultra-short telomeres

3. Results Data points were collected before exposing the cells to H2 O2 , 10 min after exposure and after each of four or five subsequent PDs. Measurements of mean telomere length and number of ultra-short telomeres were obtained for each data point, and in three experiments the percentage of senescent cells were estimated. Missing measurements are due to samples that were inappropriate for analysis due to either too low levels of DNA or the DNA being too damaged. DNA integrity was checked by gel electrophoresis as recommended by Kimura et al. [30].

It is evident from Fig. 3A that there was no correlation between percentage of senescent cells and mean telomere length neither in primary hMSCs (r = −0.43, P = 0.29) nor in hMSC-telo1 (r = 0.46, P = 0.13). Contrary to this, a strong correlation between percentage of senescent cells and the number of ultra-short telomeres was demonstrated for both primary hMSCs (r = 0.73; P = 0.0009) and hMSC-telo1 (r = 0.79; P = 0.0021) (Fig. 3B). Mean telomere length and the number of ultra-short telomeres correlated in hMSC-telo1 (r = 0.70, P = 0.0016), but not in primary hMSCs (r = 0.18, P = 0.58) (Fig. 3C).

3.1. Effect of H2 O2 exposure on frequency of senescent cells

4. Discussion

In both primary hMSCs and hMSC-telo1 cultures a substantial increase in senescent cells was observed immediately after exposure to H2 O2 (Fig. 1A and B). The H2 O2 exposure affected both cell population doubling time, and the number of survived cells (data not shown). In treated hMSC-telo1 cultures the initial rise in senescent cells after exposure rapidly faded in the following PDs (Fig. 1B), whereas in primary hMSCs the initial rise decreased only slightly and was followed by a second increase after PD 2 (Fig. 1A). This

Exposure of cells to oxidative stress such as H2 O2 induces various types of DNA damage. Thus low doses of stress generate oxidized bases, abasic sites and single-stranded breaks, whereas intense acute stress is believed to generate double-stranded breaks [31]. It has been demonstrated that telomeres are particularly sensitive to genotoxic stress [32–34]. This is believed to be due to the fact that the telomeric repeats contain guanine triplets highly sensitive to oxidation [33,35]. Furthermore, telomeric DNA

M. Harbo et al. / DNA Repair 11 (2012) 774–779

A

777

B

TRF - primary hMSC

kb

kb

6

20

5

TRF - hMSC-telo1

16

4

P2 control

T2 control

12 P1 treated

3

T1 treated 8

P2 treated

2

T2 treated 4

1 0

0 pre

C

10 min

1 PD

2 PD

3 PD

pre

D

Universal STELA - primary hMSC

10 min

1 PD

2 PD

3 PD

4 PD

5 PD

Universal STELA - hMSC-telo1

Load

Load 1

0.5

0.8

0.4 P2 control

0.6

T2 control 0.3

P2 treated 0.4

T2 treated 0.2

P1 treated

0.2

T1 treated 0.1

0

0 pre

10 min

1 PD

2 PD

3 PD

4 PD

pre

10 min

1 PD

2 PD

3 PD

4 PD

5 PD

Fig. 2. Effects of H2 O2 exposure on mean telomere length in (A) primary hMSC and (B) hMSC-telo1, and effects of H2 O2 exposure on load of ultra-short telomeres in (C) primary hMSC and (D) hMSC-telo1. kb, kilobases; load, number of telomeres below a length of 1500 bp per genome equivalent of template DNA; pre, before exposure to H2 O2 ; P, primary hMSC; T, hMSC-telo1; control, non-treated cell culture; treated, cell culture treated with H2 O2 .

damage is less efficiently repaired compared with the bulk of the genome [32,34,36]. Telomere shortening due to oxidative damage is stochastic by nature and can occur to any single telomere in a cell without affecting the mean length considerably, but nonetheless be detrimental to the cell since one ultra-short telomere alone is believed to be able to trigger senescence [21–25]. We tested this hypothesis by monitoring the relative frequency of ultra-short telomeres and mean telomere length, as well as the number of senescent cells in a series of cultured human mesenchymal stem cells exposed to H2 O2 . Data points were collected before and after exposure to H2 O2 . As expected, we found an immediate and very pronounced increase in the percentage of senescent cells after exposure to H2 O2 , which levelled off during the first few PDs (Fig. 1A and B). The increase in senescent cells could be due exclusively to damage elsewhere in the cells. Yet the fact that extremely long telomeres in hMSC-telo1 seem partly to protect against senescence suggests that the effect is, at least to some extent, mediated through telomeres. This is in compliance with evidence showing that telomeres are more sensitive to H2 O2 mediated oxidations than the bulk of the genome [32–34]. In treated hMSC-telo1 cultures, the percentage of senescent cells returned to almost undetectable levels after 4–5 PDs (Fig. 1B), whereas a second increase in senescent cells was observed in the primary hMSCs (Fig. 1A). This rise in frequency most likely reflects the phenomenon that these cells are close to the Hayflick limit, since the same rise is seen 1–2 PDs later in the non-treated cells. This time difference further suggests that telomere damage has been retained in the treated cells. Overall, the data in Fig. 1A and B suggests that damage to the telomeres is being inflicted by exposure to H2 O2 . The fact that

exposure does not notably affect mean telomere length in the short term is therefore striking (Fig. 2A and B). When looking at the number of ultra-short telomeres, a completely different, but very consistent picture emerges (Fig. 2C and D). It is noticeable that the number of ultra-short telomeres increases so shortly after exposure. The changes in the number of ultra-short telomeres correspond with the changes in the percentage of senescent cells at any given time. This is further confirmed by the strong correlation demonstrated between load of ultra-short telomeres and percentage of senescent cells (Fig. 3B). The correlation curves for hMSC-telo1 and primary hMSCs are highly coincidental, suggesting that the load of ultra-short telomeres necessary to induce senescence is similar in the two cell types and irrespective of mean telomere length. Moreover, a correlation between mean telomere length and the percentage of senescent cells is absent (Fig. 3A), further indicating that the number of oxidative lesions sufficient to cause extensive senescence is too low to affect the mean telomere length. These results are compatible with the findings of Brandl et al. [15], supporting the hypothesis that the suggested scarring of telomeres is in fact the occurrence of a few ultra-short telomeres. This scarring seems to disappear in the highly telomerase-positive hMSC-telo1, probably due to telomerase preferentially elongating the shortest telomeres [37] or simply due to selection. Most likely, selection could also explain the short term drop in the number of ultra-short telomeres observed in primary hMSCs after one PD (Fig. 2C), since it corresponds with an equivalent drop in the percentage of senescent cells (Fig. 1A). In the present communication we only report the oxidative damaging effects on hMSCs. It is possible that some cell strains are more sensitive to genomic damage than to telomeric damage [38].

778

M. Harbo et al. / DNA Repair 11 (2012) 774–779

A

β-gal staining and TRF kb 20 18 16 14 12

hMSC-telo1

10

biological tissues, these findings have major implications for studies examining the pathways from the exposure of cells to oxidative stress to the inducement of cellular senescence, since a telomeric role might be missed if a method insensitive to changes in single telomeres is applied. The choice of method should therefore be considered carefully in studies examining stress-induced telomere shortening as well as in the emerging field of lifestyle studies involving telomere length measurements.

Primary hMSC 8

Conflict of interest statement

6 4

The authors declare that there are no conflicts of interest.

2 % β-gal+

0 0

20

40

B

60

80

100

Authors’ contributions

β-gal staining and Universal STELA

All authors have contributed to the idea, the designing and planning of the studies as well as interpretation and discussion of the results. MH carried out the experiments and acquisition of data assisted by LB. NS provided the cells. All authors have contributed to, read and approved the final manuscript.

Load 1 0.9 0.8 0.7 0.6

hMSC-telo1

0.5

Acknowledgements

Primary hMSC 0.4 0.3 0.2 0.1 % β-gal+

0 20

0

40

60

80

The authors thank Marianne Mose Hansen for technical assistance and Sara Louise Harbo for proofreading the manuscript. The Danish Aging Research Center is supported by a grant from the VELUX FOUNDATION.

100

References

C

Universal STELA and TRF kb 20 18 16 14 12

hMSC-telo1

10 Primary hMSC 8 6 4 2 0

Load 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Fig. 3. Relationship between (A) percentage of senescent cells and mean telomere length, (B) percentage of senescent cells and load of ultra-short telomeres, and (C) load of ultra-short telomeres and mean telomere length. Kb, kilobases; load, number of telomeres below a length of 1500 bp per genome equivalent of template DNA; pre, before exposure to H2 O2 .

However, our first applications of Universal STELA were in fact on primary fibroblast cultures, showing that the stochastic onset of senescence in cell cultures under high (20%) oxygen levels strongly correlated with the load of ultra-short telomeres [22]. Furthermore, in our most recent study we investigated the role of chondrocyte telomere shortening in the development of osteoarthritis in human knees and found that the load of ultra-short telomeres exhibited a more significant association to chrondrocyte senescence than mean telomere length did [27]. 5. Conclusion The present study supports that single critically short telomeres are the prevailing products of stress-induced telomere shortening caused by exposure to oxidative stress, and that Universal STELA is superior to TRF in detecting these. Whether in cell cultures or in

[1] L. Hayflick, P.S. Moorhead, The serial cultivation of human diploid cell strains, Exp. Cell Res. 25 (1961) 585–621. [2] L. Hayflick, The limited in vitro lifetime of human diploid cell strains, Exp. Cell Res. 37 (1965) 614–636. [3] T. de Lange, Shelterin: the protein complex that shapes and safeguards human telomeres, Genes Dev. 19 (2005) 2100–2110. [4] J.D. Griffith, L. Comeau, S. Rosenfield, R.M. Stansel, A. Bianchi, H. Moss, T. de Lange, Mammalian telomeres end in a large duplex loop, Cell 97 (1999) 503–514. [5] R.C. Allsopp, H. Vaziri, C. Patterson, S. Goldstein, E.V. Younglai, A.B. Futcher, C.W. Greider, C.B. Harley, Telomere length predicts replicative capacity of human fibroblasts, in: Proceedings of the National Academy of Sciences of the United States of America, vol. 89, 1992, pp. 10114–10118. [6] J. Campisi, The biology of replicative senescence, Eur. J. Cancer 33 (1997) 703–709. [7] C.B. Harley, A.B. Futcher, C.W. Greider, Telomeres shorten during ageing of human fibroblasts, Nature 345 (1990) 458–460. [8] A.M. Olovnikov, A theory of marginotomy. The incomplete copying of template margin in enzymic synthesis of polynucleotides and biological significance of the phenomenon, J. Theor. Biol. 41 (1973) 181–190. [9] J.D. Watson, Origin of concatemeric T7 DNA, Nat New Biol 239 (1972) 197–201. [10] C. Martin-Ruiz, G. Saretzki, J. Petrie, J. Ladhoff, J. Jeyapalan, W. Wei, J. Sedivy, T. von Zglinicki, Stochastic variation in telomere shortening rate causes heterogeneity of human fibroblast replicative life span, J. Biol. Chem. 279 (2004) 17826–17833. [11] I. Rubelj, Z. Vondracek, Stochastic mechanism of cellular aging—abrupt telomere shortening as a model for stochastic nature of cellular aging, J. Theor. Biol. 197 (1999) 425–438. [12] N.S. Vidacek, A. Cukusic, M. Ivankovic, H. Fulgosi, M. Huzak, J.R. Smith, I. Rubelj, Abrupt telomere shortening in normal human fibroblasts, Exp. Gerontol. 45 (2010) 235–242. [13] T. von Zglinicki, G. Saretzki, W. Docke, C. Lotze, Mild hyperoxia shortens telomeres and inhibits proliferation of fibroblasts: a model for senescence? Exp. Cell Res. 220 (1995) 186–193. [14] F. d’Adda di Fagagna, P.M. Reaper, L. Clay-Farrace, H. Fiegler, P. Carr, T. Von Zglinicki, G. Saretzki, N.P. Carter, S.P. Jackson, A DNA damage checkpoint response in telomere-initiated senescence, Nature 426 (2003) 194–198. [15] A. Brandl, A. Hartmann, V. Bechmann, B. Graf, M. Nerlich, P. Angele, Oxidative stress induces senescence in chondrocytes, J. Orthop. Res. 29 (2011) 1114–1120. [16] O. Toussaint, E.E. Medrano, T. Von Zglinicki, Cellular and molecular mechanisms of stress-induced premature senescence (SIPS) of human diploid fibroblasts and melanocytes, Exp. Gerontol. 35 (2000) 927–945. [17] R.C. Allshire, M. Dempster, N.D. Hastie, Human telomeres contain at least three types of G-rich repeat distributed non-randomly, Nucleic Acids Res. 17 (1989) 4611–4627.

M. Harbo et al. / DNA Repair 11 (2012) 774–779 [18] R.C. Allsopp, E. Chang, M. Kashefi-Aazam, E.I. Rogaev, M.A. Piatyszek, J.W. Shay, C.B. Harley, Telomere shortening is associated with cell division in vitro and in vivo, Exp. Cell Res. 220 (1995) 194–200. [19] R.M. Cawthon, Telomere measurement by quantitative PCR, Nucleic Acids Res. 30 (2002) e47. [20] D.M. Baird, J. Rowson, D. Wynford-Thomas, D. Kipling, Extensive allelic variation and ultrashort telomeres in senescent human cells, Nat. Genet. 33 (2003) 203–207. [21] P. Abdallah, P. Luciano, K.W. Runge, M. Lisby, V. Geli, E. Gilson, M.T. Teixeira, A two-step model for senescence triggered by a single critically short telomere, Nat. Cell Biol. 11 (2009) 988–993. [22] L. Bendix, P.B. Horn, U.B. Jensen, I. Rubelj, S. Kolvraa, The load of short telomeres, estimated by a new method, Universal STELA, correlates with number of senescent cells, Aging Cell 9 (2010) 383–397. [23] R. Capper, B. Britt-Compton, M. Tankimanova, J. Rowson, B. Letsolo, S. Man, M. Haughton, D.M. Baird, The nature of telomere fusion and a definition of the critical telomere length in human cells, Genes Dev. 21 (2007) 2495–2508. [24] M.T. Hemann, M.A. Strong, L.Y. Hao, C.W. Greider, The shortest telomere, not average telomere length, is critical for cell viability and chromosome stability, Cell 107 (2001) 67–77. [25] Y. Zou, A. Sfeir, S.M. Gryaznov, J.W. Shay, W.E. Wright, Does a sentinel or a subset of short telomeres determine replicative senescence? Mol. Biol. Cell 15 (2004) 3709–3718. [26] R. Christensen, J. Alsner, S.F. Brandt, F. gnaes-Hansen, S. Kolvraa, N. Serakinci, Transformation of human mesenchymal stem cells in radiation carcinogenesis: long-term effect of ionizing radiation, Regen. Med. 3 (2008) 849–861. [27] M. Harbo, L. Bendix, A.C. Bay-Jensen, J. Graakjaer, K. Soe, T.L. Andersen, P. Kjaersgaard-Andersen, S. Koelvraa, J.M. Delaisse, The distribution pattern of critically short telomeres in human osteoarthritic knees, Arthritis Res. Ther. 14 (2012) R12. [28] N. Serakinci, R. Christensen, J. Graakjaer, C.J. Cairney, W.N. Keith, J. Alsner, G. Saretzki, S. Kolvraa, Ectopically hTERT expressing adult human mesenchymal stem cells are less radiosensitive than their telomerase negative counterpart, Exp. Cell Res. 313 (2007) 1056–1067.

779

[29] G.P. Dimri, X. Lee, G. Basile, M. Acosta, G. Scott, C. Roskelley, E.E. Medrano, M. Linskens, I. Rubelj, O. Pereira-Smith, A biomarker that identifies senescent human cells in culture and in aging skin in vivo, Proc. Natl. Acad. Sci. U.S.A 92 (1995) 9363–9367. [30] M. Kimura, R.C. Stone, S.C. Hunt, J. Skurnick, X. Lu, X. Cao, C.B. Harley, A. Aviv, Measurement of telomere length by the Southern blot analysis of terminal restriction fragment lengths, Nat. Protoc. 5 (2010) 1596–1607. [31] J.F. Passos, G. Saretzki, T. von Zglinicki, DNA damage in telomeres and mitochondria during cellular senescence: is there a connection? Nucleic Acids Res. 35 (2007) 7505–7513. [32] P.J. Rochette, D.E. Brash, Human telomeres are hypersensitive to UV-induced DNA damage and refractory to repair, PLoS Genet. 6 (2010) e1000926. [33] S. Petersen, G. Saretzki, T. von Zglinicki, Preferential accumulation of singlestranded regions in telomeres of human fibroblasts, Exp. Cell Res. 239 (1998) 152–160. [34] G. Hewitt, D. Jurk, F.D. Marques, C. Correia-Melo, T. Hardy, A. Gackowska, R. Anderson, M. Taschuk, J. Mann, J.F. Passos, Telomeres are favoured targets of a persistent DNA damage response in ageing and stress-induced senescence, Nat. Commun. 3 (2012) 708. [35] S. Oikawa, S. Tada-Oikawa, S. Kawanishi, Site-specific DNA damage at the GGG sequence by UVA involves acceleration of telomere shortening, Biochemistry 40 (2001) 4763–4768. [36] E.S. Henle, Z. Han, N. Tang, P. Rai, Y. Luo, S. Linn, Sequence-specific DNA cleavage by Fe2+ -mediated fenton reactions has possible biological implications, J. Biol. Chem. 274 (1999) 962–971. [37] B. Britt-Compton, R. Capper, J. Rowson, D.M. Baird, Short telomeres are preferentially elongated by telomerase in human cells, FEBS Lett. 583 (2009) 3076–3080. [38] B. Britt-Compton, F. Wyllie, J. Rowson, R. Capper, R.E. Jones, D.M. Baird, Telomere dynamics during replicative senescence are not directly modulated by conditions of oxidative stress in IMR90 fibroblast cells, Biogerontology 10 (2009) 683–693.