- Email: [email protected]
Demethylation of classical satellite 2 and 3 DNA with chromosomal instability in senescent human fibroblasts
Experimental Gerontology 37 (2002) 1005–1014 www.elsevier.com/locate/expgero
Demethylation of classical satellite 2 and 3 DNA with chromosomal instability in senescent human fibroblasts Toshikazu Suzuki, Michihiko Fujii, Dai Ayusawa* Kihara Institute for Biological Research, Yokohama City University, Maioka-cho 641-12, Totsuka-ku, Yokohama 244-0813, Japan Received 23 January 2002; received in revised form 15 April 2002; accepted 3 May 2002
Abstract Demethylation of genomic 5-methylcytosine is reported in aged human tissues and senesced human cells, although it is not understood to what extent this phenomenon contributes to replicative senescence. We examined methylation status of satellite 2 and 3 sequences during passages of normal human fibroblasts. These sequences are abundant in the juxtacentromeric heterochromatin of human chromosomes 1, 9 and 16, and heavily methylated in tissues of normal individuals. The decrease in DNA methylation level was two times faster in satellite 3 DNA than in satellite 2 and total DNA. Then we monitored appearance of micronuclei during the passages since they are indicative of heterochromatin decondensation or chromosome breakage. Concomitant with the DNA demethylation, micronuclei containing the heterochromatin of chromosomes 1, 9 or 16, appeared specifically. These results suggest that demethylation of heterochromatin has a role in replicative senescence through chromosome instability. q 2002 Published by Elsevier Science Inc. Keywords: Demethylation; Classical satellite DNA; Heterochromatin decondensation; Chromosomal instability; Cellular senescence
1. Introduction Normal human cells have a limited replicative potential in culture and stop dividing irreversibly, a phenomenon termed cellular senescence (Hayflick, 1965). Senescent cells are morphologically altered and express specific genes called senescence-associated genes (Campisi, 1997). The numbers of cell division depend on cell types and genetic backgrounds of donor cells under standard culture conditions (Stanulis-Praeger, 1987). Limitation of proliferative potential is thought to arise by an intrinsic mechanism that counts the number of cell * Corresponding author. Tel.: þ81-45-820-1906; fax: þ 81-45820-1901. E-mail address: [email protected] (D. Ayusawa).
division. Telomere shortening is proposed to be one candidate for such a mechanism (Harley, 1991). However, this hypothesis is not in agreement with the several observations. Mouse primary cells senesce much faster than human cells despite mouse cells having much longer telomeres (Holliday, 1996). In addition, there is little correlation between proliferative potential and telomerase activity or telomere length in stable clones expressing introduced telomerase (Franco et al., 2001; Ouellette et al., 2000). Gradual demethylation of 5-methylcyosine bases in DNA is thought to be another mechanism to count cell division (Hoal-van Helden and van Helden, 1989; Wilson and Jones, 1983). In human cells, 60 – 90% of the cytosine residues in CpG dinucleotides are methylated. Methylation of CpG islands is associated with condensation of chromatins enriched in
0531-5565/02/$ - see front matter q 2002 Published by Elsevier Science Inc. PII: S 0 5 3 1 - 5 5 6 5 ( 0 2 ) 0 0 0 6 1 - X
1006
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
hypoacetylated histones, thereby leading to transcriptional silencing (Henikoff, 2000; Razin, 1998). On the other hand, DNA methylation is shown to be essential for normal development, X chromosome inactivation, and genome imprinting (Constancia et al., 1998). The level of DNA methylation is shown to decrease in normal mammalian cells during passages (Fairweather et al., 1987; Wilson and Jones, 1983) and in tissues during ageing (Singhal et al., 1987; Wilson et al., 1987). Potent inhibitors of DNA methyltransferase, 5-azacytidine (aza-C) and 5-azadeoxycytidine (aza-dC), respectively, are shown to lead to shortening of in vitro lifespan (Fairweather et al., 1987; Holliday, 1986) and premature senescence followed by reduction in DNA methylation in normal human fibroblasts (Young and Smith, 2001). The premature senescence was not observed in p21-deficient fibroblasts, suggesting that it occurs by activation of cell cycle checkpoint pathway in response to DNA damage or lesions due to DNA demethylation, rather than transcriptional activation of particular genes by DNA demethylation. On the other hand, it is well known that a low dose of aza-C or aza-dC, dramatically inhibits condensation of the constitutive heterochromatins in human cells, especially in the pericentromeric regions of chromosomes 1, 9 and 16 and the q arm of chromosome Y in human lymphocytes (Haaf and Schmid, 2000). These regions are abundant in the classical satellite 2 and 3 sequences (Tagarro et al., 1994), and most prominently stained with antibody against 5-methylcytosine in metaphase spreads (Lubit et al., 1976). Further, demethylation of the satellite sequences is shown to cause chromosome breakage and rearrangements followed by decondensation of the heterochromatins (Hernandez et al., 1997; Kokalj-Vokac et al., 1993). Therefore, regarding cellular senescence, one of the specific targets for DNA demethylation may be the classical satellite DNA that constitutes heterochromatin. To date, decreases in DNA methylation in the satellite sequences during aging in vitro and in vivo have been reported in mice and bovines (Hornsby et al., 1992; Howlett et al., 1989), but not in humans. The aim of this study is to find a role of DNA demethylation in replicative senescence in normal human fibroblasts. We focused on methylation status of the classical satellite 2 and 3 sequences and
micronuclei containing these sequences. Although DNA demethylation was not specific to the satellite sequences, we found that micronuclei containing the constitutive heterochromatin of chromosomes 1, 9 or 16 preferentially appeared in human fibroblasts undergoing replicative senescence and premature senescence mediated by aza-C treatment. Thus, demethylation of heterochromatin is suggested to assist replicative senescence through chromosome instability.
2. Materials and methods 2.1. Materials Aza-C, distamycin A, 6-diamidino-2-phenylindole (DAPI), Taq DNA polymerase, and human placenta DNA (type XIII) were obtained from Sigma. Restriction endonucleases and T4 polynucleotide kinase were purchased from New England Biolabs. Tetramethylrhodamine-5-dUTP was from Roche Diagnostics. [g-32P]ATP and [a-32P]dCTP (3000 Ci/mmol) were obtained from ICN and HAS, respectively. The reagents used were of reagent grade. 2.2. Cell culture Human embryonic lung fibroblast line (TIG-7) was obtained from the Japanese Cancer Research Resources Cell Bank. Cells were cultured at 37 8C in 90-mm plastic Petri dishes containing Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum under 5% CO2 and 95% humidity. The cell culture was split at 1:4 ratio routinely, and at 1:2 ratio at late passages above 70 population doubling levels (PDLs). For aza-C treatment, 2.5 £ 105 cells were plated per dish, treated with 1 mM aza-C after culture for 24 h, and fed every 2 days with fresh medium containing aza-C for 1 week. Senescence-associated b-galactosidase was assayed as described previously (Nakabayashi et al., 1997). 2.3. Southern blot analysis Genomic DNA samples were prepared according
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
to the method of Sambrook et al. (Sambrook et al., 1989). The samples were digested with an appropriate restriction endonuclease, resolved by electrophoresis through agarose gel and transferred onto a nylon membrane as described previously (Nakabayashi et al., 1997). To determine methylation levels in the satellite sequences and genomic DNA, CpG methylation-sensitive Bst BI or Hpa II with CpG methylation-insensitive isoschizomer Msp I were used according to the previous report (Vilain et al., 2000; Xu et al., 1999). The membrane was hybridized at 37 8C with a probe in a mixture containing 0.43 M sodium phosphate (pH 7.2), 7% SDS and 20 mM EDTA for 20 h. The membrane was washed four times at 37 8C with 2X SSC (300 mM NaCl, and 30 mM trisodium citrate) containing 0.1% SDS and twice at 37 8C with 0.1X SSC containing 0.1% SDS for 20 min, and then exposed to an X-ray film at 280 8C. Radioactivity of the membrane was quantified using an imaging analyzer BAS2000 (Fuji Photo Film). Oligonucleotide probes for satellite 2 and 3 sequences were prepared based on the previous report (Tagarro et al., 1994). To determine terminal restriction fragments containing telomere sequences, oligonucleotide (TTAGGG)3 was used as a probe (Nakabayashi et al., 1997). Oligonucleotide probes were endolabeled with [g-32P]ATP using T4 polynucleotide kinase at 37 8C. Human placenta DNA and the intron 2 sequence of the glycelaldehyde 3phosphate dehydrogenase (GAPD) gene (Suzuki et al., 2001b) were labeled with [a-32P]dCTP using a random primed DNA-labeling kit (Mega-prime, Amersham-Pharmacia). 2.4. Determination of DNA methylation status DNA methylation status in the classical satellite sequences and total genomic DNA was determined by counting radioactivity of smears on a Southern blot according to the method of Vilain et al. (Vilain et al., 2000). For the classical satellite sequences, the radioactivities of the total lane (A ) and the region between the top and 2.9 kb (B ) were counted for each lane and the ratio (B/A ) was taken to represent a relative methylation level. For the total genomic DNA, the radioactivities of the total lane (A ) and the region between 2.9 and 1.5 kb
1007
(B) were counted for each lane and the ratio ðr ¼ B=AÞ was calculated for Hpa II (rH) and Msp I (rM) digestions. The ratio R ¼ rH =rM was calculated for each DNA sample, and the DNA methylation level was estimated as (1 2 R ), which theoretically ranged from 0 (fully demethylated) to 1 (fully methylated). In each case, the level was expressed relative to that of the 32 PDL cells. 2.5. Chromosome analysis Samples of cells or chromosomes were prepared according to the method of Zimonjic and Popescu (1994). The samples were stained with distamycin A/ DAPI according to the technique of Schweizer et al. (1978). The classical satellite sequences were located by primed in situ labeling (PRINS) analysis (Koch et al., 1989). Sample slides were heated at 95 8C on a thermal cycler (OmniGene, Hybaid) for 1 min, and added with a reaction mixture (30 ml) containing PRINS oligonucleotide primers (150 pmol), 50 mM each of dATP, dCTP, and dGTP, 5 mM dTTP, 1 mM tetramethylrhodamine5-dUTP, 10 mM Tris-HCl (pH 8.3), 1.5 mM MgCl2, 0.01% bovine serum albumin, and 2.5 units of Taq DNA polymerase. After covering samples with a cover slip, the slide was incubated at 95 8C for 2 min and then at 61 8C for 30 min, and immersed in a stop solution [50 mM NaCl, 50 mM EDTA (pH 8.0)] at 65 8C for 5 min. After washing out unlabeled materials with PBS containing 0.2% Tween 20 at 37 8C, the samples were counter stained with DAPI (20 ng/ml) at room temperature for 5 min. Fluorescence images were taken with a cooled CCD camera and analyzed with IPLAB SPECTRUM image software (Signal Analytics). The sequence information for PRINS oligonucleotides, the satellite 2 sequence in the pericentromeric heterochromatin of human chromosome 1q and 16q (PRINS 1 þ 16, Cat. No. 1 768 549) and the satellite 3 sequence in the proximal pericentromeric heterochromatin of human chromosome 9q (PRINS 9, Cat. No. 1 768 565), was provided by Roche Diagnostics (Seibl, 2000).
1008
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
Fig. 1. Morphology and expression of senescence-associated b-galactosidase in senesced and aza-C-treated young TIG-7 cells. Cell culture and b-galactosidase assay was performed as described in Section 2. (A) Young cells (32 PDLs); (B) senesced cells (76 PDLs); (C) young cells (33 PDLs) cultured in the presence of 1 mM aza-C for 1 week. Bar, 100 mm.
Fig. 2. Decrease in methylation levels in the classical satellite DNA in senescent and aza-C-treated TIG-7 cells. Genomic DNA samples were prepared from young (32 PDLs), senescent (76 PDLs), and aza-C-treated young (33 PDLs) TIG-7 cells. The samples were digested with a restriction endonuclease and subjected to Southern blot analysis using a 32P-labeled oligonucleotide. A, the samples were digested with Bst BI and probed with satellite 2 (Sat2) oligonucleotide. B, the samples were digested with Hpa II (H) or Msp I (M) and probed with satellite 3 (Sat3). C, the samples were digested with Hinf I and probed with (TTAGGG)3. In B and C, the membrane was re-hybridized with 32P-labeled GAPD sequence as a loading control.
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
1009
Fig. 3. Decrease in methylation levels in the classical satellite DNA during passages of TIG-7 cells. Genomic DNA samples were prepared from TIG-7 cells at the PDLs indicated and subjected to Southern blot analysis as in Fig. 2. A, the samples were digested with Bst BI and probed with Sat2 oligonucleotide. B, the samples were digested with Hpa II and probed with Sat3 oligonuleotide (upper panel) or GAPD sequence (lower panel). C, the samples were digested with Hinf I and probed with (TTAGGG)3 (upper panel) or GAPD sequence (lower panel).
3. Results 3.1. Inhibitors of DNA methyltransferase induce premature senescence in TIG-7 cells Recently, it has been reported that aza-dC induces premature senescence with demethylation of genomic 5-methylcytosine in human fibroblasts (Young and Smith, 2001). We used aza-C, another DNA methyltransferase inhibitor. When normal human fibroblasts (TIG-7 cells) were cultured in the presence of 1 mM aza-C, the cells stopped growing within 1 week (data not shown). These cells became flat and enlarged in shape and strongly induced senescence-associated bgalactosidase (Dimri et al., 1995) (Fig. 1). As no significant difference was found between aza-C and aza-dC in their effects on human fibroblasts, we used aza-C in subsequent experiments to induce DNA demethylation and premature senescence. 3.2. DNA methylation levels in the satellite sequences in senescent TIG-7 cells First, we determined the methylation status of
the classical satellites in normally senesced TIG-7 cells. We prepared genomic DNA samples from young (32 PDLs) and senescent (76 PDLs) cells, and prematurely senesced cells (33 PDLs) by azaC-treatment. Their methylation levels of 5-methylcytosine were examined by Southern blot analysis with a methylation-sensitive restriction endonuclease, Bst BI (for satellite 2) or Hap II (for satellite 3) (Xu et al., 1999). As shown in Fig. 2A, the smearing pattern with the satellite 2 sequence in the senescent cells migrated faster than that in the young cells, and was very similar to that in the aza-C treated cells (33, aza-C). Similar results were obtained with satellite 3 sequence (Fig. 2B). When digested with Msp I, methylation insensitive isoschizomer of Hpa II, the smearing patterns were much shorter than those digested with Hap II and showed no difference in the three DNA samples. These results demonstrate that both satellite 2 and 3 sequences were dramatically hypomethylated when the cells entered replicative senescence or premature senescence by treatment with aza-C. Telomere shortening was not observed in the cells treated with aza-C (Fig. 2C).
1010
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
3.3. DNA methylation levels in satellite sequences during passages of TIG-7 cells
Fig. 4. Methylation levels in the satellite 2, satellite 3, and the total genomic DNA during passages of TIG-7 cells. The methylation levels were determined and expressed relative to that of the cells at 33 PDLs as described in Section 2. The levels of the aza-C-treated cells at 33 PDLs are indicated as filled symbols. Circles, total genomic DNA; triangles, satellite 2 DNA; squares, satellite 3 DNA.
Next, we monitored methylation levels in the satellite 2 and 3 sequences during the passages by Southern blot analysis (Fig. 3). As PDLs increased, DNA fragments containing the sequences significantly decreased in size time-dependently. In a parallel experiment, we confirmed telomere shortening in the samples used (Fig. 3C). We also examined the methylation levels in the total genomic DNA during the passages by Southern blot analysis using human placenta DNA as a probe. And then, relative methylation levels in the total and the classical satellite DNA samples were calculated by the method of Vilain et al. (2000) and plotted against the PDLs (Fig. 4). In each case, the levels decreased nearly proportional to the PDLs. The rates of demethylation were quite similar in the total and satellite 2 DNA whereas the rate in the satellite 3 DNA was two times faster than those of the others. More than 70% of the methyl residues of
Fig. 5. Detection of micronuclei containing the satellite DNA sequences in senescent and aza-C-treated TIG-7 cells. Spreads of fixed cells were stained with distamycin A (DMA) and DAPI (A –C), or subjected to PRINS analysis using chromosome 1 þ 16 primers (D –F) or chromosome 9 primers (G –I). In the samples of the young cells (A, D, G), metaphase chromosomes and interphase nuclei are shown. The satellite DNA sequences are stained bright with DMA/DAPI or pink with PRINS analysis. Arrowheads indicate micronuclei.
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
1011
Table 1 Frequencies of cells containing micronuclei in TIG-7 cells Cell
Experiment
Cells with micronuclei/total cells (%)
Bright micronuclei/total micronuclei (%)
Young (32 PDLs)
Exp. 1 Exp. 2
53/10,000 (0.5) 29/6000 (0.5)
8/53 (15) 4/29 (14)
Old (71 PDLs)
Exp. 1 Exp. 2
266/5000 (5.3) 309/6000 (5.2)
125/266 (47) 143/309 (46)
aza-C (33 PDLs)
Exp. 1 Exp. 2
182/6000 (3.0) 199/6000 (3.3)
89/182 (49) 86/199 (43)
Bright micronuclei were stained with distamycin A/DAPI. For details, see text.
5-methylcytosine bases in the satellite 3 DNA were lost in the senescent cells. In contrast, treatment with aza-C decreased the levels of the three types of DNA to a level similar to those of the total and the satellite 2 DNA in the senescent cells. From these results, we concluded that the demethylation rate was higher in the satellite 3 sequence than in the satellite 2 sequence and total genomic DNA although the demethylation of DNA occurs in various regions of DNA during the passages. 3.4. Increase in micronuclei containing the constitutive heterochromatin in senescent TIG-7 cells The inhibitors of DNA methylation are shown to induce decondensation of the constitutive heterochromatin in human chromosomes 1, 9 and 16, which are abundant in the classical satellite 2 and 3 sequences (Haaf and Schmid, 2000). Further, the undermethylation in the satellite DNA is shown to be associated with decondensation of the constitutive heterochromatin in mitotic chromosomes in lymphocytes from
patients of the ICF syndrome that lacks DNA methyltransferase 3B (Dnmt3B) (Xu et al., 1999). With these results, we undertook cytogenetic analysis to detect decondensation of heterochromatin in the senescent TIG-7 cells. However, we were not able to do so as described previously in aza-C-treated human fibroblasts (Cimini et al., 1996). It is well known that decondensation of the heterochromatin results in micronuclei (Guttenbach and Schmid, 1994). Therefore, we examined micronuclei containing constitutive heterochromatin by staining with distamycin A/DAPI (Schweizer et al., 1978). As described previously (Guttenbach and Schmid, 1994), we observed two types of micronuclei with different staining patterns, bright ones (Fig. 5B and C) and dark ones (Fig. 5A) containing or not containing constitutive heterochromatin, respectively. The frequency of appearance of micronuclei was 10 and six times higher in the senescent and aza-Ctreated cells than in the young cells (Table 1). In the senescent and aza-C-treated cells, nearly half of the micronuclei were brightly stained, whereas only 15%
Table 2 Frequencies of micronuclei containing the satellite sequences of chromosome 1 þ 16 or 9 by PRINS analysis with chromosome specific oligonucleotide primers Micronuclei positive for PRINS 1 þ 16 primes (%)
Cell
Experiment
Young (32 PDLs)
Exp. 1 Exp. 2
4 (8) 2 (4)
1 (2) 1 (2)
Old (71 PDLs)
Exp. 1 Exp. 2
15 (30) 13 (26)
9 (18) 7 (14)
aza-C (33PDLs)
Exp. 1 Exp. 2
19 (38) 16 (32)
10 (20) 8 (16)
In each experiment, 50 micronuclei were examined under a fluorescence microscope.
Micronuclei positive for PRINS 9 primes (%)
1012
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
of the micronuclei were done so in the young cells. Therefore, chromosomes containing the constitutive heterochromatin were found to be preferentially released into micronuclei in the senescent and azaC-treated cells. We examined whether the bright micronuclei contained the classical satellite DNA (Koch et al., 1989) by PRINS analysis. The satellite DNA was amplified in situ on interphase nuclei or mitotic chromosomes with primers specific to the satellite 2 sequences of chromosome 1 and 16 (PRINS 1 þ 16) or the satellite 3 sequence of chromosome 9 (PRINS 9) (Fig. 5D – I). The frequencies of the micronuclei showing a positive signal for PRINS 1 þ 16 and PRINS 9 primers were about five and eight times higher in the senescent cells than in the young cells, respectively (Table 2). These results show that the chromosomes containing the classical satellite 2 or 3 sequences were preferentially released into micronuclei in the senescent as well as aza-C-treated cells.
4. Discussion Demethylation of genomic 5-methylcytosine has been reported in aged tissues and senescent human cells. It is, however, unknown whether demethylation in a specific locus or chromosomal region has some relevance to aging or cellular senescence. In this study, we monitored methylation levels in the classical satellite 2 and 3 sequences, which form constitutive heterochromatin, during passages of normal human fibroblasts. The demethylation seemed to proceed in any regions of the genomic DNA during replicative senescecne, although the rate was higher in the satellite 3 sequence than in the other sequences. With these observations, it may be difficult to find a role of the demethylation in cellular senescence. However, we found that the demethylation of the satellite sequences well correlated with the appearance of the micronuclei containing the satellite sequences in normal human fibroblasts undergoing senescence (Table 2). In addition, the fluorescence signals in interphase nuclei in PRINS analysis appeared to be more diffused in the senescent fibroblasts (Fig. 5E and H) than in young cells (Fig. 5D and G). These observations indicate that the constitutive heterochromatin is undercondensed not
only in aza-C treated fibroblasts but also in the normally senesced fibroblasts. Similarly, micronuclei were found to increase in vivo in lymphocytes of aged individuals (Bolognesi et al., 1999), as well as in lymphocytes treated with aza-C in vitro. It is known that heterochromatin decondensation induced by demethylating agents correlates well with breakage and rearrangements of chromosomes (Hernandez et al., 1997; Kokalj-Vokac et al., 1993). As the classical satellite sequences in chromosome 1 and 9 are abundant in the constitutive alkali-labile sites, DNA strand breaks or lesions are suggested to take place at these sites (Fernandez et al., 2001). Thus, DNA methylation and subsequent chromatin compaction with methylated DNA-binding proteins may protect DNA lesions from activation of the cell cycle checkpoint machinery; in other words, the demethylation of the satellite DNA and subsequent heterochromatin decondensation during passages may lead to cellular senescence by activating the cell cycle checkpoint machinery. In support of this, p212/2 human fibroblasts, which are defective in a final step of the machinery, were unable to enter replicative senescence (Brown et al., 1997) as well as premature senescence mediated by aza-dC (Young and Smith, 2001). Demethylation of the classical satellite DNA during the passages was also observed in EpsteinBarr virus-transformed lymphoblastoid cells (Vilain et al., 2000). However, they did not stop dividing or enter senescence even if the pericentromeric heterochromatin in chromosomes 1 and 16, which contain the satellite 2 sequence, were fully demethylated and undercondensed. Instead, non-clonal rearrangements of chromosomes were significantly induced in the regions. Further, the hypomethylation of the satellite 2 sequence was shown to correlate with loss or gain of all or most of the long arm of chromosomes 1 and 16, and tumor progression in some types of cancer (Qu et al., 1999; Wong et al., 2001). Therefore, the execution of senescence by DNA demethylation in normal cells may have a role in preventing tumorigenesity or malignancy due to chromosomal rearrangements. Recently, it was reported that Dnmt 3B, which leads to hypomethylation of satellite 2 and 3 DNA in the deficient lymphocytes (Xu et al., 1999), increased by three folds in senescent human fibroblasts
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
(Lopatina et al., 2002). Despite that, DNA methylation of the satellite 2 and 3 sequences significantly decreased during passages (Fig. 4). These observations seem to be contradictory, although it is not proven that Dnmt 3B specifically acts on the satellite DNA. Total DNA methyltransferase activity of senescent cells decreases to one-thirds of that of young cells (Lopatina et al., 2002; Young and Smith, 2001). Dnmt 3A and/or Dnmt 3B compensates for inefficient methylation of endogenous repetitive sequences by Dnmt 1 in mouse ES cells (Liang et al., 2002). Taken together, Dnmt 1 may be responsible for the decrease in DNA methylation in senescent human fibroblasts. However, the reason why demethylation preferentially proceeds in the satellite 3 sequence during senescence is not known. Other factors may be involved in this phenomenon because DNA methylation is regulated by various protein factors. In this study, we have clearly shown that the demethylation of the classical satellite DNA led to instability of specific chromosomes. In other experiments, we have shown that a transcript containing a satellite 3 sequence is induced in senescent human fibroblasts (Suzuki et al., 2001a). The release of the transcript from such silenced loci is thought to be a consequence of decondensation of heterochromatin. Taken together, it is reasonable to conclude that demethylation of 5-methylcytosine on heterochromatin is one of the specific events that lead to or accelerate cellular senescence in normal cells.
Acknowledgments This work was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan.
References Bolognesi, C., Lando, C., Forni, A., Landini, E., Scarpato, R., Migliore, L., Bonassi, S., 1999. Chromosomal damage and ageing: effect on micronuclei frequency in peripheral blood lymphocytes. Age Ageing 28, 393– 397. Brown, J.P., Wei, W., Sedivy, J.M., 1997. Bypass of senescence after disruption of p21CIP1/WAF1 gene in normal diploid human fibroblasts. Science 277, 831–834.
1013
Campisi, J., 1997. The biology of replicative senescence. Eur. J. Cancer 33, 703–709. Cimini, D., Tanzarella, C., Degrassi, F., 1996. Effects of 5azacytidine on the centromeric region of human fibroblasts studied by CREST staining and in situ hybridization on cytokinesis-blocked cells. Cytogenet. Cell Genet. 72, 219 –224. Constancia, M., Pickard, B., Kelsey, G., Reik, W., 1998. Imprinting mechanisms. Genome Res. 8, 881–900. Dimri, G.P., Lee, X., Basile, G., Acosta, M., Scott, G., Roskelley, C., Medrano, E.E., Linskens, M., Rubelj, I., Pereira-Smith, O., Peacocke, M., Campisi, J., 1995. A biomarker that identifies senescent human cells in culture and in aging skin in vivo. Proc. Natl Acad. Sci. USA 92, 9363–9367. Fairweather, D.S., Fox, M., Margison, G.P., 1987. The in vitro lifespan of MRC-5 cells is shortened by 5-azacytidine-induced demethylation. Exp. Cell Res. 168, 153 –159. Fernandez, J.L., Vazquez-Gundin, F., Rivero, M.T., Goyanes, V., Gosalvez, J., 2001. Evidence of abundant constitutive alkalilabile sites in human 5 bp classical satellite DNA loci by DBDFISH. Mutat. Res. 473, 163–168. Franco, S., MacKenzie, K.L., Dias, S., Alvarez, S., Rafii, S., Moore, M.A., 2001. Clonal variation in phenotype and life span of human embryonic fibroblasts (MRC-5) transduced with the catalytic component of telomerase (hTERT). Exp. Cell Res. 268, 14–25. Guttenbach, M., Schmid, M., 1994. Exclusion of specific human chromosomes into micronuclei by 5-azacytidine treatment of lymphocyte cultures. Exp. Cell Res. 211, 127 –132. Haaf, T., Schmid, M., 2000. Experimental condensation inhibition in constitutive and facultative heterochromatin of mammalian chromosomes. Cytogenet. Cell Genet. 91, 113–123. Harley, C.B., 1991. Telomere loss: mitotic clock or genetic time bomb? Mutat. Res. 256, 271 –282. Hayflick, L., 1965. The limited in vitro lifetime of human diploid cell strains. Exp. Cell Res. 37, 614 –636. Henikoff, S., 2000. Heterochromatin function in complex genomes. Biochim. Biophys. Acta 1470, O1–O8. Hernandez, R., Frady, A., Zhang, X.Y., Varela, M., Ehrlich, M., 1997. Preferential induction of chromosome 1 multibranched figures and whole-arm deletions in a human pro-B cell line treated with 5-azacytidine or 5-azadeoxycytidine. Cytogenet. Cell Genet. 76, 196– 201. Hoal-van Helden, E.G., van Helden, P.D., 1989. Age-related methylation changes in DNA may reflect the proliferative potential of organs. Mutat. Res. 219, 263 –266. Holliday, R., 1986. Strong effects of 5-azacytidine on the in vitro lifespan of human diploid fibroblasts. Exp. Cell Res. 166, 543 –552. Holliday, R., 1996. Endless quest. Bioessays 18, 3–5. Hornsby, P.J., Yang, L., Gunter, L.E., 1992. Demethylation of satellite I DNA during senescence of bovine adrenocortical cells in culture. Mutat. Res. 275, 13–19. Howlett, D., Dalrymple, S., Mays-Hoopes, L.L., 1989. Age-related demethylation of mouse satellite DNA is easily detectable by HPLC but not by restriction endonucleases. Mutat. Res. 219, 101 –106. Koch, J.E., Kolvraa, S., Petersen, K.B., Gregersen, N., Bolund, L.,
1014
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
1989. Oligonucleotide-priming methods for the chromosomespecific labelling of alpha satellite DNA in situ. Chromosoma 98, 259 –265. Kokalj-Vokac, N., Almeida, A., Viegas-Pequignot, E., Jeanpierre, M., Malfoy, B., Dutrillaux, B., 1993. Specific induction of uncoiling and recombination by azacytidine in classical satellite-containing constitutive heterochromatin. Cytogenet. Cell Genet. 63, 11–15. Liang, G., Chan, M.F., Tomigahara, Y., Tsai, Y.C., Gonzales, F.A., Li, E., Laird, P.W., Jones, P.A., 2002. Cooperativity between DNA methyltransferases in the maintenance methylation of repetitive elements. Mol. Cell Biol. 22, 480 –491. Lopatina, N., Haskell, J.F., Andrews, L.G., Poole, J.C., Saldanha, S., Tollefsbol, T., 2002. Differential maintenance and de novo methylating activity by three DNA methyltransferases in aging and immortalized fibroblasts. J. Cell. Biochem. 84, 324–334. Lubit, B.W., Pham, T.D., Miller, O.J., Erlanger, B.F., 1976. Localization of 5-methylcytosine in human metaphase chromosomes by immunoelectron microscopy. Cell 9, 503 –509. Nakabayashi, K., Ogata, T., Fujii, M., Tahara, H., Ide, T., Wadhwa, R., Kaul, S.C., Mitsui, Y., Ayusawa, D., 1997. Decrease in amplified telomeric sequences and induction of senescence markers by introduction of human chromosome 7 or its segments in SUSM-1. Exp. Cell Res. 235, 345–353. Ouellette, M.M., Liao, M., Herbert, B.S., Johnson, M., Holt, S.E., Liss, H.S., Shay, J.W., Wright, W.E., 2000. Subsenescent telomere lengths in fibroblasts immortalized by limiting amounts of telomerase. J. Biol. Chem. 275, 10072–10076. Qu, G., Dubeau, L., Narayan, A., Yu, M.C., Ehrlich, M., 1999. Satellite DNA hypomethylation vs. overall genomic hypomethylation in ovarian epithelial tumors of different malignant potential. Mutat. Res. 423, 91–101. Razin, A., 1998. CpG methylation, chromatin structure and gene silencing-a three-way connection. EMBO J. 17, 4905–4908. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: A Laboratory Manual, 2nd ed, Cold Spring Harbor Laboratory Press, Plainview, NY. Schweizer, D., Ambros, P., Andrle, M., 1978. Modification of DAPI banding on human chromosomes by prestaining with a DNAbinding oligopeptide antibiotic, distamycin A. Exp. Cell Res. 111, 327 –332. Seibl, R., 2000. Identification of chromosomes in metaphase spreads and interphase nuclei with PRINS oligonucleotide
primers and DIG- or rhodamine-labeled nucleotides. Nonradioactive In Situ Hybridization Application Manual, Roche Diagnostics Corporation, pp. 79– 87. Singhal, R.P., Mays-Hoopes, L.L., Eichhorn, G.L., 1987. DNA methylation in aging of mice. Mech. Ageing Dev. 41, 199–210. Stanulis-Praeger, B.M., 1987. Cellular senescence revisited: a review. Mech. Ageing Dev. 38, 1–48. Suzuki, T., Minagawa, S., Michishita, E., Ogino, H., Fujii, M., Mitsui, Y., Ayusawa, D., 2001a. Induction of senescenceassociated genes by 5-bromodeoxyuridine in HeLa cells. Exp. Gerontol. 36, 465 –474. Suzuki, T., Yaginuma, M., Oishi, T., Michishita, E., Ogino, H., Fujii, M., Ayusawa, D., 2001b. 5-Bromodeoxyuridine suppresses position effect variegation of transgenes in hela cells. Exp. Cell Res. 266, 53– 63. Tagarro, I., Fernandez-Peralta, A.M., Gonzalez-Aguilera, J.J., 1994. Chromosomal localization of human satellites 2 and 3 by a FISH method using oligonucleotides as probes. Hum. Genet. 93, 383 –388. Vilain, A., Bernardino, J., Gerbault-Seureau, M., Vogt, N., Niveleau, A., Lefrancois, D., Malfoy, B., Dutrillaux, B., 2000. DNA methylation and chromosome instability in lymphoblastoid cell lines. Cytogenet. Cell. Genet. 90, 93 –101. Wilson, V.L., Jones, P.A., 1983. DNA methylation decreases in aging but not in immortal cells. Science 220, 1055–1057. Wilson, V.L., Smith, R.A., Ma, S., Cutler, R.G., 1987. Genomic 5methyldeoxycytidine decreases with age. J. Biol. Chem. 262, 9948–9951. Wong, N., Lam, W.C., Lai, P.B., Pang, E., Lau, W.Y., Johnson, P.J., 2001. Hypomethylation of chromosome 1 heterochromatin DNA correlates with q-arm copy gain in human hepatocellular carcinoma. Am. J. Pathol. 159, 465 –471. Xu, G.L., Bestor, T.H., Bourc’his, D., Hsieh, C.L., Tommerup, N., Bugge, M., Hulten, M., Qu, X., Russo, J.J., Viegas-Pequignot, E., 1999. Chromosome instability and immunodeficiency syndrome caused by mutations in a DNA methyltransferase gene. Nature 402, 187–191. Young, J.I., Smith, J.R., 2001. DNA methyltransferase inhibition in normal human fibroblasts induces a p21-dependent cell cycle withdrawal. J. Biol. Chem. 276, 19610–19616. Zimonjic, D.B., Popescu, N.C., 1994. An improved procedure for chromosome preparations from solid tumors. Cancer Genet. Cytogenet. 72, 161.
Demethylation of classical satellite 2 and 3 DNA with chromosomal instability in senescent human fibroblasts Toshikazu Suzuki, Michihiko Fujii, Dai Ayusawa* Kihara Institute for Biological Research, Yokohama City University, Maioka-cho 641-12, Totsuka-ku, Yokohama 244-0813, Japan Received 23 January 2002; received in revised form 15 April 2002; accepted 3 May 2002
Abstract Demethylation of genomic 5-methylcytosine is reported in aged human tissues and senesced human cells, although it is not understood to what extent this phenomenon contributes to replicative senescence. We examined methylation status of satellite 2 and 3 sequences during passages of normal human fibroblasts. These sequences are abundant in the juxtacentromeric heterochromatin of human chromosomes 1, 9 and 16, and heavily methylated in tissues of normal individuals. The decrease in DNA methylation level was two times faster in satellite 3 DNA than in satellite 2 and total DNA. Then we monitored appearance of micronuclei during the passages since they are indicative of heterochromatin decondensation or chromosome breakage. Concomitant with the DNA demethylation, micronuclei containing the heterochromatin of chromosomes 1, 9 or 16, appeared specifically. These results suggest that demethylation of heterochromatin has a role in replicative senescence through chromosome instability. q 2002 Published by Elsevier Science Inc. Keywords: Demethylation; Classical satellite DNA; Heterochromatin decondensation; Chromosomal instability; Cellular senescence
1. Introduction Normal human cells have a limited replicative potential in culture and stop dividing irreversibly, a phenomenon termed cellular senescence (Hayflick, 1965). Senescent cells are morphologically altered and express specific genes called senescence-associated genes (Campisi, 1997). The numbers of cell division depend on cell types and genetic backgrounds of donor cells under standard culture conditions (Stanulis-Praeger, 1987). Limitation of proliferative potential is thought to arise by an intrinsic mechanism that counts the number of cell * Corresponding author. Tel.: þ81-45-820-1906; fax: þ 81-45820-1901. E-mail address: [email protected] (D. Ayusawa).
division. Telomere shortening is proposed to be one candidate for such a mechanism (Harley, 1991). However, this hypothesis is not in agreement with the several observations. Mouse primary cells senesce much faster than human cells despite mouse cells having much longer telomeres (Holliday, 1996). In addition, there is little correlation between proliferative potential and telomerase activity or telomere length in stable clones expressing introduced telomerase (Franco et al., 2001; Ouellette et al., 2000). Gradual demethylation of 5-methylcyosine bases in DNA is thought to be another mechanism to count cell division (Hoal-van Helden and van Helden, 1989; Wilson and Jones, 1983). In human cells, 60 – 90% of the cytosine residues in CpG dinucleotides are methylated. Methylation of CpG islands is associated with condensation of chromatins enriched in
0531-5565/02/$ - see front matter q 2002 Published by Elsevier Science Inc. PII: S 0 5 3 1 - 5 5 6 5 ( 0 2 ) 0 0 0 6 1 - X
1006
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
hypoacetylated histones, thereby leading to transcriptional silencing (Henikoff, 2000; Razin, 1998). On the other hand, DNA methylation is shown to be essential for normal development, X chromosome inactivation, and genome imprinting (Constancia et al., 1998). The level of DNA methylation is shown to decrease in normal mammalian cells during passages (Fairweather et al., 1987; Wilson and Jones, 1983) and in tissues during ageing (Singhal et al., 1987; Wilson et al., 1987). Potent inhibitors of DNA methyltransferase, 5-azacytidine (aza-C) and 5-azadeoxycytidine (aza-dC), respectively, are shown to lead to shortening of in vitro lifespan (Fairweather et al., 1987; Holliday, 1986) and premature senescence followed by reduction in DNA methylation in normal human fibroblasts (Young and Smith, 2001). The premature senescence was not observed in p21-deficient fibroblasts, suggesting that it occurs by activation of cell cycle checkpoint pathway in response to DNA damage or lesions due to DNA demethylation, rather than transcriptional activation of particular genes by DNA demethylation. On the other hand, it is well known that a low dose of aza-C or aza-dC, dramatically inhibits condensation of the constitutive heterochromatins in human cells, especially in the pericentromeric regions of chromosomes 1, 9 and 16 and the q arm of chromosome Y in human lymphocytes (Haaf and Schmid, 2000). These regions are abundant in the classical satellite 2 and 3 sequences (Tagarro et al., 1994), and most prominently stained with antibody against 5-methylcytosine in metaphase spreads (Lubit et al., 1976). Further, demethylation of the satellite sequences is shown to cause chromosome breakage and rearrangements followed by decondensation of the heterochromatins (Hernandez et al., 1997; Kokalj-Vokac et al., 1993). Therefore, regarding cellular senescence, one of the specific targets for DNA demethylation may be the classical satellite DNA that constitutes heterochromatin. To date, decreases in DNA methylation in the satellite sequences during aging in vitro and in vivo have been reported in mice and bovines (Hornsby et al., 1992; Howlett et al., 1989), but not in humans. The aim of this study is to find a role of DNA demethylation in replicative senescence in normal human fibroblasts. We focused on methylation status of the classical satellite 2 and 3 sequences and
micronuclei containing these sequences. Although DNA demethylation was not specific to the satellite sequences, we found that micronuclei containing the constitutive heterochromatin of chromosomes 1, 9 or 16 preferentially appeared in human fibroblasts undergoing replicative senescence and premature senescence mediated by aza-C treatment. Thus, demethylation of heterochromatin is suggested to assist replicative senescence through chromosome instability.
2. Materials and methods 2.1. Materials Aza-C, distamycin A, 6-diamidino-2-phenylindole (DAPI), Taq DNA polymerase, and human placenta DNA (type XIII) were obtained from Sigma. Restriction endonucleases and T4 polynucleotide kinase were purchased from New England Biolabs. Tetramethylrhodamine-5-dUTP was from Roche Diagnostics. [g-32P]ATP and [a-32P]dCTP (3000 Ci/mmol) were obtained from ICN and HAS, respectively. The reagents used were of reagent grade. 2.2. Cell culture Human embryonic lung fibroblast line (TIG-7) was obtained from the Japanese Cancer Research Resources Cell Bank. Cells were cultured at 37 8C in 90-mm plastic Petri dishes containing Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum under 5% CO2 and 95% humidity. The cell culture was split at 1:4 ratio routinely, and at 1:2 ratio at late passages above 70 population doubling levels (PDLs). For aza-C treatment, 2.5 £ 105 cells were plated per dish, treated with 1 mM aza-C after culture for 24 h, and fed every 2 days with fresh medium containing aza-C for 1 week. Senescence-associated b-galactosidase was assayed as described previously (Nakabayashi et al., 1997). 2.3. Southern blot analysis Genomic DNA samples were prepared according
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
to the method of Sambrook et al. (Sambrook et al., 1989). The samples were digested with an appropriate restriction endonuclease, resolved by electrophoresis through agarose gel and transferred onto a nylon membrane as described previously (Nakabayashi et al., 1997). To determine methylation levels in the satellite sequences and genomic DNA, CpG methylation-sensitive Bst BI or Hpa II with CpG methylation-insensitive isoschizomer Msp I were used according to the previous report (Vilain et al., 2000; Xu et al., 1999). The membrane was hybridized at 37 8C with a probe in a mixture containing 0.43 M sodium phosphate (pH 7.2), 7% SDS and 20 mM EDTA for 20 h. The membrane was washed four times at 37 8C with 2X SSC (300 mM NaCl, and 30 mM trisodium citrate) containing 0.1% SDS and twice at 37 8C with 0.1X SSC containing 0.1% SDS for 20 min, and then exposed to an X-ray film at 280 8C. Radioactivity of the membrane was quantified using an imaging analyzer BAS2000 (Fuji Photo Film). Oligonucleotide probes for satellite 2 and 3 sequences were prepared based on the previous report (Tagarro et al., 1994). To determine terminal restriction fragments containing telomere sequences, oligonucleotide (TTAGGG)3 was used as a probe (Nakabayashi et al., 1997). Oligonucleotide probes were endolabeled with [g-32P]ATP using T4 polynucleotide kinase at 37 8C. Human placenta DNA and the intron 2 sequence of the glycelaldehyde 3phosphate dehydrogenase (GAPD) gene (Suzuki et al., 2001b) were labeled with [a-32P]dCTP using a random primed DNA-labeling kit (Mega-prime, Amersham-Pharmacia). 2.4. Determination of DNA methylation status DNA methylation status in the classical satellite sequences and total genomic DNA was determined by counting radioactivity of smears on a Southern blot according to the method of Vilain et al. (Vilain et al., 2000). For the classical satellite sequences, the radioactivities of the total lane (A ) and the region between the top and 2.9 kb (B ) were counted for each lane and the ratio (B/A ) was taken to represent a relative methylation level. For the total genomic DNA, the radioactivities of the total lane (A ) and the region between 2.9 and 1.5 kb
1007
(B) were counted for each lane and the ratio ðr ¼ B=AÞ was calculated for Hpa II (rH) and Msp I (rM) digestions. The ratio R ¼ rH =rM was calculated for each DNA sample, and the DNA methylation level was estimated as (1 2 R ), which theoretically ranged from 0 (fully demethylated) to 1 (fully methylated). In each case, the level was expressed relative to that of the 32 PDL cells. 2.5. Chromosome analysis Samples of cells or chromosomes were prepared according to the method of Zimonjic and Popescu (1994). The samples were stained with distamycin A/ DAPI according to the technique of Schweizer et al. (1978). The classical satellite sequences were located by primed in situ labeling (PRINS) analysis (Koch et al., 1989). Sample slides were heated at 95 8C on a thermal cycler (OmniGene, Hybaid) for 1 min, and added with a reaction mixture (30 ml) containing PRINS oligonucleotide primers (150 pmol), 50 mM each of dATP, dCTP, and dGTP, 5 mM dTTP, 1 mM tetramethylrhodamine5-dUTP, 10 mM Tris-HCl (pH 8.3), 1.5 mM MgCl2, 0.01% bovine serum albumin, and 2.5 units of Taq DNA polymerase. After covering samples with a cover slip, the slide was incubated at 95 8C for 2 min and then at 61 8C for 30 min, and immersed in a stop solution [50 mM NaCl, 50 mM EDTA (pH 8.0)] at 65 8C for 5 min. After washing out unlabeled materials with PBS containing 0.2% Tween 20 at 37 8C, the samples were counter stained with DAPI (20 ng/ml) at room temperature for 5 min. Fluorescence images were taken with a cooled CCD camera and analyzed with IPLAB SPECTRUM image software (Signal Analytics). The sequence information for PRINS oligonucleotides, the satellite 2 sequence in the pericentromeric heterochromatin of human chromosome 1q and 16q (PRINS 1 þ 16, Cat. No. 1 768 549) and the satellite 3 sequence in the proximal pericentromeric heterochromatin of human chromosome 9q (PRINS 9, Cat. No. 1 768 565), was provided by Roche Diagnostics (Seibl, 2000).
1008
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
Fig. 1. Morphology and expression of senescence-associated b-galactosidase in senesced and aza-C-treated young TIG-7 cells. Cell culture and b-galactosidase assay was performed as described in Section 2. (A) Young cells (32 PDLs); (B) senesced cells (76 PDLs); (C) young cells (33 PDLs) cultured in the presence of 1 mM aza-C for 1 week. Bar, 100 mm.
Fig. 2. Decrease in methylation levels in the classical satellite DNA in senescent and aza-C-treated TIG-7 cells. Genomic DNA samples were prepared from young (32 PDLs), senescent (76 PDLs), and aza-C-treated young (33 PDLs) TIG-7 cells. The samples were digested with a restriction endonuclease and subjected to Southern blot analysis using a 32P-labeled oligonucleotide. A, the samples were digested with Bst BI and probed with satellite 2 (Sat2) oligonucleotide. B, the samples were digested with Hpa II (H) or Msp I (M) and probed with satellite 3 (Sat3). C, the samples were digested with Hinf I and probed with (TTAGGG)3. In B and C, the membrane was re-hybridized with 32P-labeled GAPD sequence as a loading control.
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
1009
Fig. 3. Decrease in methylation levels in the classical satellite DNA during passages of TIG-7 cells. Genomic DNA samples were prepared from TIG-7 cells at the PDLs indicated and subjected to Southern blot analysis as in Fig. 2. A, the samples were digested with Bst BI and probed with Sat2 oligonucleotide. B, the samples were digested with Hpa II and probed with Sat3 oligonuleotide (upper panel) or GAPD sequence (lower panel). C, the samples were digested with Hinf I and probed with (TTAGGG)3 (upper panel) or GAPD sequence (lower panel).
3. Results 3.1. Inhibitors of DNA methyltransferase induce premature senescence in TIG-7 cells Recently, it has been reported that aza-dC induces premature senescence with demethylation of genomic 5-methylcytosine in human fibroblasts (Young and Smith, 2001). We used aza-C, another DNA methyltransferase inhibitor. When normal human fibroblasts (TIG-7 cells) were cultured in the presence of 1 mM aza-C, the cells stopped growing within 1 week (data not shown). These cells became flat and enlarged in shape and strongly induced senescence-associated bgalactosidase (Dimri et al., 1995) (Fig. 1). As no significant difference was found between aza-C and aza-dC in their effects on human fibroblasts, we used aza-C in subsequent experiments to induce DNA demethylation and premature senescence. 3.2. DNA methylation levels in the satellite sequences in senescent TIG-7 cells First, we determined the methylation status of
the classical satellites in normally senesced TIG-7 cells. We prepared genomic DNA samples from young (32 PDLs) and senescent (76 PDLs) cells, and prematurely senesced cells (33 PDLs) by azaC-treatment. Their methylation levels of 5-methylcytosine were examined by Southern blot analysis with a methylation-sensitive restriction endonuclease, Bst BI (for satellite 2) or Hap II (for satellite 3) (Xu et al., 1999). As shown in Fig. 2A, the smearing pattern with the satellite 2 sequence in the senescent cells migrated faster than that in the young cells, and was very similar to that in the aza-C treated cells (33, aza-C). Similar results were obtained with satellite 3 sequence (Fig. 2B). When digested with Msp I, methylation insensitive isoschizomer of Hpa II, the smearing patterns were much shorter than those digested with Hap II and showed no difference in the three DNA samples. These results demonstrate that both satellite 2 and 3 sequences were dramatically hypomethylated when the cells entered replicative senescence or premature senescence by treatment with aza-C. Telomere shortening was not observed in the cells treated with aza-C (Fig. 2C).
1010
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
3.3. DNA methylation levels in satellite sequences during passages of TIG-7 cells
Fig. 4. Methylation levels in the satellite 2, satellite 3, and the total genomic DNA during passages of TIG-7 cells. The methylation levels were determined and expressed relative to that of the cells at 33 PDLs as described in Section 2. The levels of the aza-C-treated cells at 33 PDLs are indicated as filled symbols. Circles, total genomic DNA; triangles, satellite 2 DNA; squares, satellite 3 DNA.
Next, we monitored methylation levels in the satellite 2 and 3 sequences during the passages by Southern blot analysis (Fig. 3). As PDLs increased, DNA fragments containing the sequences significantly decreased in size time-dependently. In a parallel experiment, we confirmed telomere shortening in the samples used (Fig. 3C). We also examined the methylation levels in the total genomic DNA during the passages by Southern blot analysis using human placenta DNA as a probe. And then, relative methylation levels in the total and the classical satellite DNA samples were calculated by the method of Vilain et al. (2000) and plotted against the PDLs (Fig. 4). In each case, the levels decreased nearly proportional to the PDLs. The rates of demethylation were quite similar in the total and satellite 2 DNA whereas the rate in the satellite 3 DNA was two times faster than those of the others. More than 70% of the methyl residues of
Fig. 5. Detection of micronuclei containing the satellite DNA sequences in senescent and aza-C-treated TIG-7 cells. Spreads of fixed cells were stained with distamycin A (DMA) and DAPI (A –C), or subjected to PRINS analysis using chromosome 1 þ 16 primers (D –F) or chromosome 9 primers (G –I). In the samples of the young cells (A, D, G), metaphase chromosomes and interphase nuclei are shown. The satellite DNA sequences are stained bright with DMA/DAPI or pink with PRINS analysis. Arrowheads indicate micronuclei.
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
1011
Table 1 Frequencies of cells containing micronuclei in TIG-7 cells Cell
Experiment
Cells with micronuclei/total cells (%)
Bright micronuclei/total micronuclei (%)
Young (32 PDLs)
Exp. 1 Exp. 2
53/10,000 (0.5) 29/6000 (0.5)
8/53 (15) 4/29 (14)
Old (71 PDLs)
Exp. 1 Exp. 2
266/5000 (5.3) 309/6000 (5.2)
125/266 (47) 143/309 (46)
aza-C (33 PDLs)
Exp. 1 Exp. 2
182/6000 (3.0) 199/6000 (3.3)
89/182 (49) 86/199 (43)
Bright micronuclei were stained with distamycin A/DAPI. For details, see text.
5-methylcytosine bases in the satellite 3 DNA were lost in the senescent cells. In contrast, treatment with aza-C decreased the levels of the three types of DNA to a level similar to those of the total and the satellite 2 DNA in the senescent cells. From these results, we concluded that the demethylation rate was higher in the satellite 3 sequence than in the satellite 2 sequence and total genomic DNA although the demethylation of DNA occurs in various regions of DNA during the passages. 3.4. Increase in micronuclei containing the constitutive heterochromatin in senescent TIG-7 cells The inhibitors of DNA methylation are shown to induce decondensation of the constitutive heterochromatin in human chromosomes 1, 9 and 16, which are abundant in the classical satellite 2 and 3 sequences (Haaf and Schmid, 2000). Further, the undermethylation in the satellite DNA is shown to be associated with decondensation of the constitutive heterochromatin in mitotic chromosomes in lymphocytes from
patients of the ICF syndrome that lacks DNA methyltransferase 3B (Dnmt3B) (Xu et al., 1999). With these results, we undertook cytogenetic analysis to detect decondensation of heterochromatin in the senescent TIG-7 cells. However, we were not able to do so as described previously in aza-C-treated human fibroblasts (Cimini et al., 1996). It is well known that decondensation of the heterochromatin results in micronuclei (Guttenbach and Schmid, 1994). Therefore, we examined micronuclei containing constitutive heterochromatin by staining with distamycin A/DAPI (Schweizer et al., 1978). As described previously (Guttenbach and Schmid, 1994), we observed two types of micronuclei with different staining patterns, bright ones (Fig. 5B and C) and dark ones (Fig. 5A) containing or not containing constitutive heterochromatin, respectively. The frequency of appearance of micronuclei was 10 and six times higher in the senescent and aza-Ctreated cells than in the young cells (Table 1). In the senescent and aza-C-treated cells, nearly half of the micronuclei were brightly stained, whereas only 15%
Table 2 Frequencies of micronuclei containing the satellite sequences of chromosome 1 þ 16 or 9 by PRINS analysis with chromosome specific oligonucleotide primers Micronuclei positive for PRINS 1 þ 16 primes (%)
Cell
Experiment
Young (32 PDLs)
Exp. 1 Exp. 2
4 (8) 2 (4)
1 (2) 1 (2)
Old (71 PDLs)
Exp. 1 Exp. 2
15 (30) 13 (26)
9 (18) 7 (14)
aza-C (33PDLs)
Exp. 1 Exp. 2
19 (38) 16 (32)
10 (20) 8 (16)
In each experiment, 50 micronuclei were examined under a fluorescence microscope.
Micronuclei positive for PRINS 9 primes (%)
1012
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
of the micronuclei were done so in the young cells. Therefore, chromosomes containing the constitutive heterochromatin were found to be preferentially released into micronuclei in the senescent and azaC-treated cells. We examined whether the bright micronuclei contained the classical satellite DNA (Koch et al., 1989) by PRINS analysis. The satellite DNA was amplified in situ on interphase nuclei or mitotic chromosomes with primers specific to the satellite 2 sequences of chromosome 1 and 16 (PRINS 1 þ 16) or the satellite 3 sequence of chromosome 9 (PRINS 9) (Fig. 5D – I). The frequencies of the micronuclei showing a positive signal for PRINS 1 þ 16 and PRINS 9 primers were about five and eight times higher in the senescent cells than in the young cells, respectively (Table 2). These results show that the chromosomes containing the classical satellite 2 or 3 sequences were preferentially released into micronuclei in the senescent as well as aza-C-treated cells.
4. Discussion Demethylation of genomic 5-methylcytosine has been reported in aged tissues and senescent human cells. It is, however, unknown whether demethylation in a specific locus or chromosomal region has some relevance to aging or cellular senescence. In this study, we monitored methylation levels in the classical satellite 2 and 3 sequences, which form constitutive heterochromatin, during passages of normal human fibroblasts. The demethylation seemed to proceed in any regions of the genomic DNA during replicative senescecne, although the rate was higher in the satellite 3 sequence than in the other sequences. With these observations, it may be difficult to find a role of the demethylation in cellular senescence. However, we found that the demethylation of the satellite sequences well correlated with the appearance of the micronuclei containing the satellite sequences in normal human fibroblasts undergoing senescence (Table 2). In addition, the fluorescence signals in interphase nuclei in PRINS analysis appeared to be more diffused in the senescent fibroblasts (Fig. 5E and H) than in young cells (Fig. 5D and G). These observations indicate that the constitutive heterochromatin is undercondensed not
only in aza-C treated fibroblasts but also in the normally senesced fibroblasts. Similarly, micronuclei were found to increase in vivo in lymphocytes of aged individuals (Bolognesi et al., 1999), as well as in lymphocytes treated with aza-C in vitro. It is known that heterochromatin decondensation induced by demethylating agents correlates well with breakage and rearrangements of chromosomes (Hernandez et al., 1997; Kokalj-Vokac et al., 1993). As the classical satellite sequences in chromosome 1 and 9 are abundant in the constitutive alkali-labile sites, DNA strand breaks or lesions are suggested to take place at these sites (Fernandez et al., 2001). Thus, DNA methylation and subsequent chromatin compaction with methylated DNA-binding proteins may protect DNA lesions from activation of the cell cycle checkpoint machinery; in other words, the demethylation of the satellite DNA and subsequent heterochromatin decondensation during passages may lead to cellular senescence by activating the cell cycle checkpoint machinery. In support of this, p212/2 human fibroblasts, which are defective in a final step of the machinery, were unable to enter replicative senescence (Brown et al., 1997) as well as premature senescence mediated by aza-dC (Young and Smith, 2001). Demethylation of the classical satellite DNA during the passages was also observed in EpsteinBarr virus-transformed lymphoblastoid cells (Vilain et al., 2000). However, they did not stop dividing or enter senescence even if the pericentromeric heterochromatin in chromosomes 1 and 16, which contain the satellite 2 sequence, were fully demethylated and undercondensed. Instead, non-clonal rearrangements of chromosomes were significantly induced in the regions. Further, the hypomethylation of the satellite 2 sequence was shown to correlate with loss or gain of all or most of the long arm of chromosomes 1 and 16, and tumor progression in some types of cancer (Qu et al., 1999; Wong et al., 2001). Therefore, the execution of senescence by DNA demethylation in normal cells may have a role in preventing tumorigenesity or malignancy due to chromosomal rearrangements. Recently, it was reported that Dnmt 3B, which leads to hypomethylation of satellite 2 and 3 DNA in the deficient lymphocytes (Xu et al., 1999), increased by three folds in senescent human fibroblasts
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
(Lopatina et al., 2002). Despite that, DNA methylation of the satellite 2 and 3 sequences significantly decreased during passages (Fig. 4). These observations seem to be contradictory, although it is not proven that Dnmt 3B specifically acts on the satellite DNA. Total DNA methyltransferase activity of senescent cells decreases to one-thirds of that of young cells (Lopatina et al., 2002; Young and Smith, 2001). Dnmt 3A and/or Dnmt 3B compensates for inefficient methylation of endogenous repetitive sequences by Dnmt 1 in mouse ES cells (Liang et al., 2002). Taken together, Dnmt 1 may be responsible for the decrease in DNA methylation in senescent human fibroblasts. However, the reason why demethylation preferentially proceeds in the satellite 3 sequence during senescence is not known. Other factors may be involved in this phenomenon because DNA methylation is regulated by various protein factors. In this study, we have clearly shown that the demethylation of the classical satellite DNA led to instability of specific chromosomes. In other experiments, we have shown that a transcript containing a satellite 3 sequence is induced in senescent human fibroblasts (Suzuki et al., 2001a). The release of the transcript from such silenced loci is thought to be a consequence of decondensation of heterochromatin. Taken together, it is reasonable to conclude that demethylation of 5-methylcytosine on heterochromatin is one of the specific events that lead to or accelerate cellular senescence in normal cells.
Acknowledgments This work was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan.
References Bolognesi, C., Lando, C., Forni, A., Landini, E., Scarpato, R., Migliore, L., Bonassi, S., 1999. Chromosomal damage and ageing: effect on micronuclei frequency in peripheral blood lymphocytes. Age Ageing 28, 393– 397. Brown, J.P., Wei, W., Sedivy, J.M., 1997. Bypass of senescence after disruption of p21CIP1/WAF1 gene in normal diploid human fibroblasts. Science 277, 831–834.
1013
Campisi, J., 1997. The biology of replicative senescence. Eur. J. Cancer 33, 703–709. Cimini, D., Tanzarella, C., Degrassi, F., 1996. Effects of 5azacytidine on the centromeric region of human fibroblasts studied by CREST staining and in situ hybridization on cytokinesis-blocked cells. Cytogenet. Cell Genet. 72, 219 –224. Constancia, M., Pickard, B., Kelsey, G., Reik, W., 1998. Imprinting mechanisms. Genome Res. 8, 881–900. Dimri, G.P., Lee, X., Basile, G., Acosta, M., Scott, G., Roskelley, C., Medrano, E.E., Linskens, M., Rubelj, I., Pereira-Smith, O., Peacocke, M., Campisi, J., 1995. A biomarker that identifies senescent human cells in culture and in aging skin in vivo. Proc. Natl Acad. Sci. USA 92, 9363–9367. Fairweather, D.S., Fox, M., Margison, G.P., 1987. The in vitro lifespan of MRC-5 cells is shortened by 5-azacytidine-induced demethylation. Exp. Cell Res. 168, 153 –159. Fernandez, J.L., Vazquez-Gundin, F., Rivero, M.T., Goyanes, V., Gosalvez, J., 2001. Evidence of abundant constitutive alkalilabile sites in human 5 bp classical satellite DNA loci by DBDFISH. Mutat. Res. 473, 163–168. Franco, S., MacKenzie, K.L., Dias, S., Alvarez, S., Rafii, S., Moore, M.A., 2001. Clonal variation in phenotype and life span of human embryonic fibroblasts (MRC-5) transduced with the catalytic component of telomerase (hTERT). Exp. Cell Res. 268, 14–25. Guttenbach, M., Schmid, M., 1994. Exclusion of specific human chromosomes into micronuclei by 5-azacytidine treatment of lymphocyte cultures. Exp. Cell Res. 211, 127 –132. Haaf, T., Schmid, M., 2000. Experimental condensation inhibition in constitutive and facultative heterochromatin of mammalian chromosomes. Cytogenet. Cell Genet. 91, 113–123. Harley, C.B., 1991. Telomere loss: mitotic clock or genetic time bomb? Mutat. Res. 256, 271 –282. Hayflick, L., 1965. The limited in vitro lifetime of human diploid cell strains. Exp. Cell Res. 37, 614 –636. Henikoff, S., 2000. Heterochromatin function in complex genomes. Biochim. Biophys. Acta 1470, O1–O8. Hernandez, R., Frady, A., Zhang, X.Y., Varela, M., Ehrlich, M., 1997. Preferential induction of chromosome 1 multibranched figures and whole-arm deletions in a human pro-B cell line treated with 5-azacytidine or 5-azadeoxycytidine. Cytogenet. Cell Genet. 76, 196– 201. Hoal-van Helden, E.G., van Helden, P.D., 1989. Age-related methylation changes in DNA may reflect the proliferative potential of organs. Mutat. Res. 219, 263 –266. Holliday, R., 1986. Strong effects of 5-azacytidine on the in vitro lifespan of human diploid fibroblasts. Exp. Cell Res. 166, 543 –552. Holliday, R., 1996. Endless quest. Bioessays 18, 3–5. Hornsby, P.J., Yang, L., Gunter, L.E., 1992. Demethylation of satellite I DNA during senescence of bovine adrenocortical cells in culture. Mutat. Res. 275, 13–19. Howlett, D., Dalrymple, S., Mays-Hoopes, L.L., 1989. Age-related demethylation of mouse satellite DNA is easily detectable by HPLC but not by restriction endonucleases. Mutat. Res. 219, 101 –106. Koch, J.E., Kolvraa, S., Petersen, K.B., Gregersen, N., Bolund, L.,
1014
T. Suzuki et al. / Experimental Gerontology 37 (2002) 1005–1014
1989. Oligonucleotide-priming methods for the chromosomespecific labelling of alpha satellite DNA in situ. Chromosoma 98, 259 –265. Kokalj-Vokac, N., Almeida, A., Viegas-Pequignot, E., Jeanpierre, M., Malfoy, B., Dutrillaux, B., 1993. Specific induction of uncoiling and recombination by azacytidine in classical satellite-containing constitutive heterochromatin. Cytogenet. Cell Genet. 63, 11–15. Liang, G., Chan, M.F., Tomigahara, Y., Tsai, Y.C., Gonzales, F.A., Li, E., Laird, P.W., Jones, P.A., 2002. Cooperativity between DNA methyltransferases in the maintenance methylation of repetitive elements. Mol. Cell Biol. 22, 480 –491. Lopatina, N., Haskell, J.F., Andrews, L.G., Poole, J.C., Saldanha, S., Tollefsbol, T., 2002. Differential maintenance and de novo methylating activity by three DNA methyltransferases in aging and immortalized fibroblasts. J. Cell. Biochem. 84, 324–334. Lubit, B.W., Pham, T.D., Miller, O.J., Erlanger, B.F., 1976. Localization of 5-methylcytosine in human metaphase chromosomes by immunoelectron microscopy. Cell 9, 503 –509. Nakabayashi, K., Ogata, T., Fujii, M., Tahara, H., Ide, T., Wadhwa, R., Kaul, S.C., Mitsui, Y., Ayusawa, D., 1997. Decrease in amplified telomeric sequences and induction of senescence markers by introduction of human chromosome 7 or its segments in SUSM-1. Exp. Cell Res. 235, 345–353. Ouellette, M.M., Liao, M., Herbert, B.S., Johnson, M., Holt, S.E., Liss, H.S., Shay, J.W., Wright, W.E., 2000. Subsenescent telomere lengths in fibroblasts immortalized by limiting amounts of telomerase. J. Biol. Chem. 275, 10072–10076. Qu, G., Dubeau, L., Narayan, A., Yu, M.C., Ehrlich, M., 1999. Satellite DNA hypomethylation vs. overall genomic hypomethylation in ovarian epithelial tumors of different malignant potential. Mutat. Res. 423, 91–101. Razin, A., 1998. CpG methylation, chromatin structure and gene silencing-a three-way connection. EMBO J. 17, 4905–4908. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: A Laboratory Manual, 2nd ed, Cold Spring Harbor Laboratory Press, Plainview, NY. Schweizer, D., Ambros, P., Andrle, M., 1978. Modification of DAPI banding on human chromosomes by prestaining with a DNAbinding oligopeptide antibiotic, distamycin A. Exp. Cell Res. 111, 327 –332. Seibl, R., 2000. Identification of chromosomes in metaphase spreads and interphase nuclei with PRINS oligonucleotide
primers and DIG- or rhodamine-labeled nucleotides. Nonradioactive In Situ Hybridization Application Manual, Roche Diagnostics Corporation, pp. 79– 87. Singhal, R.P., Mays-Hoopes, L.L., Eichhorn, G.L., 1987. DNA methylation in aging of mice. Mech. Ageing Dev. 41, 199–210. Stanulis-Praeger, B.M., 1987. Cellular senescence revisited: a review. Mech. Ageing Dev. 38, 1–48. Suzuki, T., Minagawa, S., Michishita, E., Ogino, H., Fujii, M., Mitsui, Y., Ayusawa, D., 2001a. Induction of senescenceassociated genes by 5-bromodeoxyuridine in HeLa cells. Exp. Gerontol. 36, 465 –474. Suzuki, T., Yaginuma, M., Oishi, T., Michishita, E., Ogino, H., Fujii, M., Ayusawa, D., 2001b. 5-Bromodeoxyuridine suppresses position effect variegation of transgenes in hela cells. Exp. Cell Res. 266, 53– 63. Tagarro, I., Fernandez-Peralta, A.M., Gonzalez-Aguilera, J.J., 1994. Chromosomal localization of human satellites 2 and 3 by a FISH method using oligonucleotides as probes. Hum. Genet. 93, 383 –388. Vilain, A., Bernardino, J., Gerbault-Seureau, M., Vogt, N., Niveleau, A., Lefrancois, D., Malfoy, B., Dutrillaux, B., 2000. DNA methylation and chromosome instability in lymphoblastoid cell lines. Cytogenet. Cell. Genet. 90, 93 –101. Wilson, V.L., Jones, P.A., 1983. DNA methylation decreases in aging but not in immortal cells. Science 220, 1055–1057. Wilson, V.L., Smith, R.A., Ma, S., Cutler, R.G., 1987. Genomic 5methyldeoxycytidine decreases with age. J. Biol. Chem. 262, 9948–9951. Wong, N., Lam, W.C., Lai, P.B., Pang, E., Lau, W.Y., Johnson, P.J., 2001. Hypomethylation of chromosome 1 heterochromatin DNA correlates with q-arm copy gain in human hepatocellular carcinoma. Am. J. Pathol. 159, 465 –471. Xu, G.L., Bestor, T.H., Bourc’his, D., Hsieh, C.L., Tommerup, N., Bugge, M., Hulten, M., Qu, X., Russo, J.J., Viegas-Pequignot, E., 1999. Chromosome instability and immunodeficiency syndrome caused by mutations in a DNA methyltransferase gene. Nature 402, 187–191. Young, J.I., Smith, J.R., 2001. DNA methyltransferase inhibition in normal human fibroblasts induces a p21-dependent cell cycle withdrawal. J. Biol. Chem. 276, 19610–19616. Zimonjic, D.B., Popescu, N.C., 1994. An improved procedure for chromosome preparations from solid tumors. Cancer Genet. Cytogenet. 72, 161.