Abnormal telomere dynamics of B-lymphoblastoid ... - Semantic Scholar

Oncogene (1997) 15, 1911 ± 1920  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Abnormal telomere dynamics of B-lymphoblastoid cell strains from Werner's syndrome patients transformed by Epstein ± Barr virus Hidetoshi Tahara1, Yoshiki Tokutake2, Shizuko Maeda1, Hiroshi Kataoka2, Taro Watanabe1, Misako Satoh2, Takehisa Matsumoto2, Minoru Sugawara2, Toshinori Ide1, Makoto Goto3, Yasuhiro Furuichi2 and Masanobu Sugimoto2 1

Department of Cellular and Molecular Biology, Hiroshima University School of Medicine, 1-2-3, Kasumi, Minamiku, Hiroshima 734; 2AGENE Research Institute, 200 Kajiwara, Kamakura, Kanagawa 247; 3Department of Rheumatology, Tokyo Metropolitan Otsuka Hospital, 2-8-1, Minami-Otsuka, Toshima-ku, Tokyo 170, Japan

The characteristics of B-lymphoblastoid cell strains transformed by Epstein ± Barr virus (EBV) from normal individuals and Werner's syndrome (WRN) patients were compared. We continuously passaged cell strains from 28 WRN patients and 20 normal individuals for about 2 years corresponding to over 160 population doubling levels (PDLs). First, the WRN mutation signi®cantly suppressed the immortalization: all the 28 cell strains from WRN patients, as well as 15 out of 20 cell strains from normal individuals, died out before 160 PDLs mostly without developing a signi®cant telomerase activity. The remaining ®ve cell strains from normal individuals became moderately/strongly telomerase-positive and, three of them were apparently immortalized with an in®nitively proliferating activity. Second, the monitoring of the telomere length of both normal and WRN cell strains during the culture period suggests that the WRN gene mutation causes abnormal dynamics of the telomere: (1) a signi®cant proportion of WRN cell strains showed drastic shortening or lengthening of telomere lengths during cell passages compared with normal cell strains, and (2) WRN cell strains terminated their life-span at a wide range of telomere length (between 3.5 and 18.5 Kbp), whereas normal cell strains terminated within a narrow telomere length range (between 5.5 and 9 Kbp). The chromosomal aberration characteristic of WRN cells, including translocation was con®rmed in our experiment. We discussed the correlation between the chromosomal instability, abnormal telomere dynamics and inability of immortalization of the WRN B-lymphobloastoid cell strains. Keywords: Werner's syndrome; B-lymphoblastoid cells; telomere; telomerase; EBT

Introduction Werner's syndrome (WRN) is an autosomal recessive disorder causing symptoms of premature aging (Epstein et al., 1996; Goto et al., 1981; Martin, 1978) accompanied by rare cancers (Goto et al., 1996). Human diploid ®broblasts have a ®nite proliferative life-span in vitro (Hay¯ick and Moorhead, 1961), which is considered as a model of cellular senescence. Interestingly, ®broblasts from WRN patients are Correspondence: Y Furuichi Received 16 April 1997; revised 23 June 1997; accepted 24 June 1997

known to have a shortened life-span in vitro (Martin et al., 1970). The WRN gene was mapped to chromosome 8 (Goto et al., 1992; Schellenberg et al., 1992). Recently, Yu et al. (1996) identi®ed a gene mutated in WRN patients that encodes a protein signi®cantly similar to RecQ-type DNA helicases. They described 15 mutations, and we found four additional mutations in this gene (Goto et al., 1997; Matsumoto et al., 1997; Oshima et al., 1996; Yu et al., 1996). At least 83% of chromosomes of WRN patients had one of these mutations, while the remaining 17% of chromosomes were considered to have unrecognized mutations. These facts together indicate that the mutation of the RecQ-type helicase gene is responsible for WRN symptoms, as well as the in vitro senescence of ®broblasts. An intriguing question that leads from this ®nding is the molecular mechanisms underlying the cellular senescence. The problem of using WRN ®broblast cell strains to clarify this question is that they are usually not easy to obtain and they have a very limited life-span. B-lymphoblastoid cell strains transformed by Epstein ± Barr virus (EBV) are the most readily available subjects from patients. Furthermore, they are very easy to manipulate. They are expected, therefore, to be useful to investigate WRN pathogenesis at molecular and cellular levels, provided that the WRN gene is expressed and has a biological function in the cells. Indeed, the following evidence supports this idea. First, chromosomal aberration, known as variegated translocation mosaicism (VTM), is a pathological phenotype of both ®broblasts and EBV-transformed B-lymphoblastoid cell strains (Salk et al., 1985a). Second, impaired S-phase transit has been reported in EBVtransformed B-lymphoblastoid cell strains from WRN patients (Poot et al., 1992). Third, hypermutable ligation of plasmid DNA ends has been reported in these cell strains of patients (Runger et al., 1994). Fourth, the WRN gene was shown to be expressed in these cell strains as shown by Northern blot and RT ± PCR analyses (Yu et al., 1996; Matsumoto et al., 1997). We lack, however, some fundamental knowledge about EBV-transformed B-lymphoblastoid cell strains to use them to investigate the molecular mechanisms of WRN pathogenesis: for example, whether the Blymphoblastoid cell strains are mortal or immortal is unknown. Usually, these cell strains are considered to be immortalized (Miller, 1990). On the other hand, fragmented results support that some of them are mortal: (1) two WRN patient-derived cell strains were reported to be mortal (Salk et al., 1985b), and (2) two

Abnormal telomere dynamics of WRN B-lymphoblasts H Tahara et al

1912

cell strains cloned from normal individuals were reported to be mortal provided that they lacked telomerase activity (Counter et al., 1994). To systematically clarify the mortality and immortality of the cells, we carried out a long-term cell passage of established EBV-transformed lymphoblastoid cell strains from 28 WRN patients and 20 normal individuals. Their telomerase activity and the length of telomere were assessed simultaneously, since both of them are considered to be intimately correlated with cellular senescence and immortalization (Blackburn, 1991; Counter et al., 1994). We found that WRN gene mutation caused abnormal dynamics of telomere length and suppressed the occurrence of immortalized cells with a strong telomerase activity.

Results Life-span More than 95% of the PBL samples from both normal individuals and WRN patients gave EBV-transformed B-lymphoblastoid cell strains, and the capability of transformation of the PBLs between them showed no dierence. The cell strains used in our experiment were con®rmed to possess Blymphoblastoid characteristics by staining the cells

Figure 1 Maximal PDLs of B-lymphoblastoid cell strains from WRN patients and normal individuals. Mortal cell strains from normal individuals (*); immortalized cell strains from normal individuals (*); mortal cell strains from the near relatives of patients (^); an immortalized cell strain of the brother of a patient (^); mortal cell strains of WRN patients (~). The asterisk indicates that the cell line showed a strong telomerase activity at later stages of PDLs

with anti-human immunoglobulins (Kataoka et al., manuscript in preparation). The EBV-transformed Blympho-blastoid cells were referred to simply as `Blymphoblastoid cells' in the text for convenience except where otherwise mentioned. We have continuously passaged the established B-lymphoblastoid cell strains from 28 WRN and 20 normal individuals for more than 2 years. At around 30 ± 50 PDLs, the cell strains from both groups began to stop proliferation one by one (Figure 1). All the 28

Table 1 Donors of B-lymphoblastoid cell lines and their mutations Donors

Age/Sex

Mutation

Maximal PDLs

WS6801 WS5801 WS6501 WS0101 WS2101 WS11001 WS23702 WS23703 WS10402 WS24002 WS9101 WS9801 WS10201 WS10501 WS10901 WS11201 WS9601 WS10101 WS10801 WS11301 WS11601 WS0701 WS0702 WS0801 WS6201 WS11801 WS11901 WS12301

45F 42M 50F 30M 49F 38F 45M 51M 39F 47F 52F 42M 43F 53F 60F 49F 48F 38F 38F 48F 38F 52M 50F 42F 42F 39F 35M 42M

?/? 1/? 6/? 1/4 4/4 6/6 4/6 4/6 4/4 4/4 4/4 4/4 6/6 6/6 4/? 10/10 4/4 4/4 4/4 1/4 4/4 4/4 4/4 4/4 7/7 4/4 4/? ?/?

69 67 57 52 95 143 124 105 86 99 90 111 157 55 123 93 63 83 141 148 142 56 43 80 97 99 115 103

N0102 N0103 N0104 N4702 N6803 N7401

27F 60M 60F 52M 50M 50M

1/7 1/7 4/7 7/7 7/7 ND

80 30 136 104 244* 75

N0001 N0002 N0003 N0004 N0005 N0006 N0007 N0008 N0009 N0010 N0011 N0012 N0013 N0014

100F 97F 95F 97F 99M 29M 35M 51M 54M 0M 0M 25F 53F 30M

ND ND ND ND ND 7/7 7/7 7/7 7/7 ND ND ND ND 7/7

80 56 185* 66 185* 150 111 149 58 75* 65 103 140* 125

The donors with `WS' are WRN patients with WRN mutation and those with `N' are normal individuals, among which donors from NO102 to N7401 are the parents or brothers and sisters of patients. For patients, mutations 1 to 10 were analysed. For normal inividuals, three major mutations (mutations 1, 4 and 6) were analysed and `7' means negative for these mutations (Matsumoto et al., 1997). Mutations 1 ± 4 are described in (Yu et al., 1996); mutation 6 is a C to T substitution at nucleotide 1366 (Yu et al., 1996); mutation 7 is a deletion of A at nucleotide 3677 (Yu et al., 1996); mutation 10 is a deletion of two as at nucleotides 733 and 734 (Matsumoto et al., 1997). Unrecognized mutation are indicated with `?'. ND, not done. *These cell strains continue proliferation at present


lymphoblastoid cell strains from WRN patients died out at below 157 PDLs: an average life-span was 95.2+6.3 PDLs, range between 39 and 157 PDLs. The type of mutation in the WRN gene showed no apparent correlation with the life-span (Table 1). Among the cell strains from the 20 normal individuals, including six of those who are parents or brothers and sisters of the patients, 15 cell strains were mortal with an average life-span of 92.5+3.7 PDLs, and a range from 30 ± 150 PDLs. The remaining ®ve cell strains, N0003, N0005, N6803, N0010 and N0013, behaved quite dierently. First, all of them developed a signi®cant telomerase activity at later cell passages, as described below. Second, the three cell strains, N0003, N0005 and N6803 continued to proliferate over 180 PDLs. Third, the three cell strains, N6803, N0010 and N0013, overcame a crisis

followed by active proliferation. These results indicate that the N0003, N0005 and N6803 strains are immortalized and that the N0010 and N0013 strains are also likely immortalized. To summarize: (1) WRN lymphoblastoid cell strains were essentially mortal (100%, 28/28), (2) although a large part of the normal cell strains were also mortal, a signi®cant proportion of them became immortal (15 ± 25%, 3/20 - 5/20), (3) the upper limit of the maximal PDLs of mortal cells is around 160 PDLs, and (4) the mortal cell strains showed no signi®cant dierence in life-span between normal and WRN lymphoblastoid cell strains (Figure 1). A statistical analysis demonstrated that the frequency of immortalized cells was signi®cantly high in normal cells than in WRN cells (if only three cells are immortalized, P50.05; if four or ®ve cells are immortalized, P50.01).

Table 2 The change of telomerase activity with the progress of PDLs Cell line WS6801 WS5801 WS6501 WS0101 WS2101 WS11001 WS23702 WS23703 WS10402 WS24002 WS9101 WS9801 WS10201 WS10501 WS10901 WS11201 WS9601 WS10101 WS10801 WS11301 WS11601 WS0701 WS0702 WS0801 WS6201 WS11801 WS11901 WS12301

1st* 2nd PDL Act.*** PDL Act. 10 37** 33 33 33 29 24 22 22 22 22 21 21 20 20 20 17 14

N0102 N0103 N0104 N4702 N6803 N7401

54

N0001 N0002 N0003 N0004 N0005 N0006 N0007 N0008 N0009 N0010 N0011 N0012 N0013 N0014

7 7 20

37 11 4

7 37 22 22 11 4

ND ND ND ND W W M N M W M W W M W W N N M W M ND ND ND ND ND ND ND

24 50 46 46 46 42 37 35 35 35 35 34 34 33 33 33 30 27 19 19 19

N ND M ND N M

67

S N M ND W N N W W W ND ND ND ND

20 16 38 20 26 50 35 35 24 10 10

50 70 17

ND M ND ND W W W W M W W N N W W N W W M M W W W W ND ND ND ND N ND N ND S N M W N W M N W M W M M ND ND ND

PDL

3rd

Act.

68 64 64 64 57 55 53 53 53 53 52 52 51 51 51 48 45 37 37 37 8 9 8 7

ND N ND ND N N N N N N N N N W N N N N N N N N N N W N W W

68 9 101 35

42

38 33 54 38 38 68 53 53 42 28 28 7 8

PDL

4th

Act.

53 24 25 24 23

ND N ND ND N N N N N N N N N ND N N ND N N M N N ND N W N N M

ND ND N N S N

84 25 117 51

ND ND N N S N

N W N W N N N N N N N N N ND

54 49 68 54 54 84 69 69 58 44 44 23 22 16

W W W N S N N N N W N N W W

58 84 80 80 80 73 71 69 69 69 68 68 67 67 64 61 53

PDL

5th

Act.

67 38 39 38 37

ND ND ND ND ND N W N W N N N N ND N W ND N N W W ND ND N W N N W

98 39 131 65

ND ND N N S N

94 94 94 87 85 83 83 83 82 82 81 81 78 77

68 93 68 98 83 83 58 58 37 64 30

N ND N ND S N N N ND S W N W N

PDL

119 119

108 107 107

106 103 100

63 64 63 62

123 64 173

110 93 123 108 108

62 89 55

6th

Act. ND ND ND ND ND N N ND ND ND ND N N ND N ND ND ND N W N ND ND ND N W W M ND ND N N S N ND ND N ND S N N N ND ND ND N N W

PDL

136 136

124 124

123 120 117

80 81 80 79

140 81 198

135 135 140 125

79 121 72

7th

Act. ND ND ND ND ND N N ND ND ND ND ND N ND N ND ND ND N N N ND ND ND N N N N ND ND N N S ND ND ND M ND S N ND N ND ND ND N N W

PDL

149

145

105

226

167 167

141 97

8th

Act. ND ND ND ND ND ND ND ND ND ND ND ND N ND ND ND ND ND ND M ND ND ND ND ND ND N ND ND ND ND ND S ND ND ND S ND S ND ND ND ND ND ND ND M N

*1st to 8th indicate the experimental numbers for the assay of telomerase activity. **The ®gure indicates PDLs of each cell line at which a telomerase assay was done. ***Act., telomerase activity; S, strong (130000 rlu5); M, moderate (800005130000 rlu); W, weak (30000580000 rlu); N, negative (30000 rlu4). See ®gure 2a. ND, not done

1913


1914

Telomerase activity

WS11901

198

WS11301

90 97

WS10201

N6803

135

N0014

135

N0013

N0005

PDLs

N0003

Telomerase activity was monitored at about 15 ± 30 PDL intervals during cell passages (Table 2). Figure 2 shows ITAS-TRAP examples which are classi®ed by HPA-TRAP into strong (S), moderate (M), weak (W) and negative (N) telomerase activities. Typical 6-base ladder was RNase-sensitive and primer-dependent. Discontinuous weak ladder was sometimes RNase

L

P

149 145 105

ITAS

resistant suggesting PCR artifact, and was de®ned as telomerase negative. Mixing experiments of telomerasenegative and telomerase-positive cell extracts indicated that telomerase negative cell extract did not contain a signi®cant activity inhibiting telomerase activity. At early stages of PDLs, more than half of the cell strains from both normal individuals and WRN patients showed a weak or moderate telomerase activity while the remaining cells showed no detectable activity. A large proportion of the cell strains tended to lose telomerase activity with the progress of PDLs, regardless of whether they were derived from normal individuals or WRN patients, except for a few cell strains including immortalized ones. The presence of a weak or moderate telomerase activity in pre-crisis Blymphoblastoid cells was consistent with the observation by other investigators (Avilion et al., 1996; Norrback et al., 1996). Four normal cell strains, N6803, N0003, N0005 and N0010 became strongly telomerase-positive between 11 and 69 PDLs, between 110 and 167 PDLs, between 38 and 54 PDLs, and between 44 and 58 PDLs, respectively, and the N0013 strain became moderately positive at 141 PDLs (Table 2). Telomere length

Cell No. 103 102 103 102 103 103 103 102 103 103 103 M

S

N

N S

N

M N N

103 S

Figure 2 Telomerase activity of WRN and normal cells as assessed by TRAP assay. Various degrees of telomerase activity. S, stong (130 k relative light units (rlu) or above); M, moderate (80 krlu or above and below 130 krlu); W, weak (30 krlu or above and below 80 krlu); N, negative (below 30 krlu). The values of 130 krlu, 80 krlu and 30 krlu are equivalent to the activities of 1000 cells, 100 cells and 10 cells of positive control cell line, L1210, respectively. The names of cell strains are shown at the top and PDLs at sampling are shown below them. 103 and 102: mean the number of cells applied to analysis. L: lysis buer (negative control), P: positive control cell lines L1210 (1000 cells)

The length of telomere was assessed by Southern blot analysis of the TRFs. Mean TRF size calculated from the highest density of smear pattern of Southern blot was only a rough approximation but was valuable enough to compare drastic changes. Figures 3 and 4 show changes in the TRF length during cell passage for the cell strains with longer life-spans of 100 PDLs or more. The telomere length did not necessarily shorten uniformly, and some cell strains restored, in average, a long telomere length after shortening in both WRN and normal cell strains. However, a comparison of the pro®les of telomere length clearly shows that the changes in telomere length of WRN cell strains were far more drastic than in normal cell strains (Figure 3). Especially, the three cell strains, WS10801, WS11301 and WS11601, showed typical irregular patterns with repeated gain and loss of telomere

Figure 3 Changes in telomere length with the progress of PDLs from WRN and normal cell strains. The changes in TRF length of mortal normal cell strains (left ®gure), mortal WRN cell strains (two middle ®gures) and normal immortalized cell strains (right ®gure) are shown. The mean TRF lengths were from the signals of Southern blot as described in Materials and methods. the (a) and (b) of WS11301 and (c) and (d) of WS11601 are the points where PRINS analyses were done. The ®gures in the parentheses indicate PDLs in these analyses. See Figure 6


length (Figure 3, WRN (2)). In the WS11301 cell line, the gain of TRF length between 30 and 64 PDLs was calculated as 378 bp/PDL while the loss between 120 and 146 PDLs was calculated as 340 bp/PDL. Most of other WRN cell strains also showed similar rapid shortening and lengthening of telomere lengths. Such abnormal dynamics of telomere length was not so obvious in mortal or immortalized cell strains from normal individuals. The analysis of the DNAs by gel electrophoresis and staining with ethidium bromide showed that this abnormal dynamics of telomere length was not due to an artifact such as a simple degradation of DNAs. Our observations are quite dierent from those by Schultz et al. (1996) in ®broblasts: (1) the loss of TRF length per PDL was 58 and 108 bp/PDL for normal and WRN ®broblast cell strains, respectively, and (2) the telomere length of ®broblasts decreased uniformly. Both of the observations were, however, similar in that the rate of TRF shortening was sharper in WRN cells than in normal cells. These results indicate that the rate of the loss of TRF length of WRN B-lymphoblastoid cell strains was much higher than the rate of WRN ®broblasts. The two immortalized cell strains, N6803 and N0005, showed relatively constant TRF lengths after they became strongly telomerase-positive between 54 and 69 PDLs (Table 2, Figure 3). Another immortalized cell line, N0003, uniformly shortened its telomere length to attain 86103 bp at 135 PDLs, when the telomerase activity obviously increased. Figure 4 shows the change in TRF length with the progress of PDLs in Southern blot analyses; the N4702 cell line showed a uniform decrease; the WS11301 and WS11601 cell strains showed a complete shortening and lengthening; the N0005 cell line at ®rst shortened its TRF length followed by keeping the length at around 60 PDLs after they became strongly telomerase-positive. Figure 5 shows the distribution of TRF length determined at the PDLs nearest to the end of their lifespan (terminal TRF length). The terminal TRF length of normal cell strains centered at a narrow range

PDLs (bp)

WS11301

WS11601

Analyses of chromosomes The chromosomal aberration called VTM is a wellknown cytopathological event of WRN cells (Salk et

N0005

25 39 64 81 17 30 64 103 120 146 14 27 61 78 100 117 20 54 68 93 110

N4702

between 5.5 and 9 Kbp. On the other hand, the terminal TRF length of WRN cell strains distributed widely from 3.5 to 18.5 Kbp. To investigate the localization of telomere at each chromosomal terminal, we also observed telomeres by staining cells with PRINS. Figure 6a and b show the PRINS pro®les of metaphase at 32 and 62 PDLs of the S11301 cell strain, and Figure 6c and d show the pro®le of metaphase at 58 and 82 PDLs of the WS11601 cell strain (Figure 3, WRN (2)). Although PRINS was not quantitative enough for determination of each TRF size and both cells showed small and medium-sized signals, larger signals were found only in the cells at the elongating phase of the TRF as observed by Southern blot (Figure 6b and d). The results of PRINS indicate that the size of telomere varied even in a single cell regardless of whether they were derived from normal or WRN cells and that the variation was larger in WRN cells at the TRF elongating phase. These EBV-transformed B-lymphoblastoid cells, especially those from WRN patients, likely correspond to the immortalized cells with heterogeneous telomeres derived from cloned cells with long homogeneous telomeres after cell passages, as described by Murnane et al. (1994).

23100 — 9420 — 6560 — 4360 —

2320 — 2030 —

Figure 4 Southern blot of TRF. The names of cell strains are shown at the top and the PDLs in the analyses are shown below them

Figure 5 TRF length near the maximal PDLs of mortal WRN and mortal normal cell strains. The TRF lengths determined at the PDL nearest to the maximal PDL were compared between normal and WRN cell strains. The dierence between the maximal PDL and the PDL at which the TRF was determind was 18.8+3.0 (s.e.) for normal cell strains and 17.6+2.4 in average for WRN cell strains

1915


1916

Figure 6 Telomere staining by PRINS of WRN cell strains. WS11301 cell strain at 32 (a) and 62 PDLs (b); WS11601 cell line at 58 (c) and 82 PDLs (d). See also WRN (2) of Figure 3 and its legend

Figure 7 Chromosomal abnormality of WRN and normal cell strains. Chromosomal analyses were made for 20 metaphase cells for each cell strain. Cells with abnormal chromosomes(s), such as translocation, deletion, addition, telomeric association and ring formation were, judged as abnormal. Mortal normal cell strains, (*); mortal WRN cell strains, (*); immortalized normal cell strains, (~). the PDLs were shown in the parentheses for the cell strains near at their maximal PDLs


al., 1985a). We monitored the VTM for normal and WRN cell strains at various PDLs by examining 20 metaphase cells in each cell strain. A cell with at least one aberrent chromosome was judged as abnormal (Figure 7). Most of the mortal normal cell strains showed a normal pattern of 46 XX/XY at the PDLs between 5 and 100, and the proportion of abnormal cells increased moderately toward the maximal PDLs. Three immortalized cell strains, N6803, N0003 and N0005, showed completely abnormal patterns, including 45 XX or 47 XX. On the other hand, one third of WRN cell strains (5 out of 15 cell strains) were shown to possess ®ve or more aberrant chromosomes per 20 chromosomes at the PDLs between 11 and 100, and the chromosomal abnormality increased drastically toward the maximal PDLs. Translocation, deletion, addition, telomeric association and ring formation were the major abnormalities, which may be linked with the augmentation of hyper-recombination of DNA. Thus, our results con®rmed and extended the previous observations that the VTM was more obvious in B-lymphoblastoid cells from WRN patients than from normal individuals throughout the cell passages including the early stages after EBV transformation. Discussion This study demonstrated that all 28 EBV-transformed lymphoblastoid cell strains from WRN patients were mortal. On the other hand, ®ve out of 20 normal cell strains are likely immortalized accompanied by a signi®cant telomerase activity, although the remaining 15 cell strains were mortal being unable to overcome a crisis. These results were consistent with the following previous observations. Salk et al. (1985a) preliminarily reported that among four B-lymphoblastoid cell strains, one normal cell line grew continuously over 180 PDLs, while one cell strain from a WRN brother (normal) and two WRN cell strains died out at 158, 54 and 88 PDLs, respectively. Counter et al. (1994) reported that two clonal telomerase-negative cell lines died out at 45 and 95 PDLs, while two telomerasepositive cell lines were immortalized. Bryan et al. (1995) reported that the in vitro transformed immortal ®broblasts showed, at high frequency, no telomerase activity along with aberrantly long telomeres. On the other hand, observations by ours as well as by Counter et al. (1994) strongly suggested an intimate correlation between the immortalization and development of strong telomerase activity. It could be either due to the dierences in the methods for transformation (EBV versus non-EBV) or in the types of cells (Blymphoblastoid cells versus ®broblasts) between these studies. In SV40-induced transformation of ®broblasts the cells are known to become immortalized after they experience the `crisis' (Girardi et al., 1965). In our experiment, among the three immortalized B-lymphoblastoid cell strains, the N6803 cell strain showed an apparent crisis phase accompanied by a decreased activity of cell growth, while the remaining two cell strains, N0003 and N0005, did not show an apparent crisis. In addition, two telomerase-positive cell strains, N0010 and N0013 not yet reaching the limit PDLs of

immortalization (160), also overcame a crisis. Taking into consideration that each culture bottle contained about 5 ± 106106 cells and that ®ve telomerasepositive immortalized cell strains arose out of 20 normal cell strains, the frequency of immortalization was calculated roughly as 2.5 to 5 in 108 crisis cells in the experimental system of this study. This frequency was not necessarily high compared with the frequency calculated for the SV40-induced immortalization of ®broblasts ranging from 1 in 105 to 1 in 109 crisis cells (Bryan and Reddel, 1997). Therefore, although the peripheral blood cells from virtually all normal individuals were transformed by EBV to a cell strain with a long-term growth (about 90 PDLs on average), most of them were clearly mortal in that they were usually not able to exceed 160 PDLs. A rare cell, 2.5 to 5 in 108, was considered to be immortalized and gain strong telomerase activity leading to continued proliferation beyond this limit. The immortalization in WRN cell strains are considered to be much rarer compared with normal cell strains. Why were the immortalized cells so rare in WRN cell strains? First, the increased chromosomal aberration in WRN cells may result in accumulation of chromosomal damage more quickly in WRN cells than in normal cells. If so, a rare immortalized cell positive for telomerase, if it appears, cannot expand due to an extensive chromosomal damage. In other words, this suggests that a normal WRN gene is essential for in®nite proliferation of immortalized cells that avoid the excess accumulation of chromosomal damage. The accumulation of chromosomal damage may be correlated to the observation by Poot et al. (1992) that the S-phase transit of cell cycle is impaired in EBV-transformed lymphoblastoid cell strains from WRN patients. Second, it is possible that TRF lengths of WRN cell strains are more variable than those of normal cell strains, and that the TRF length of a speci®c telomere shortens below a critical length leading to cell death while the average TRF length remains relatively long. In other words, a signi®cant proportion of WRN cells died, even when they, as a cell population, had relatively long TRFs (12 ± 18 Kbp) (Figure 5). This event of B-lymphoblastoid cell strains in our study was apparently consistent with the observation by Schulz et al. (1996) that WRN ®broblasts stopped proliferation at longer TRF than normal ®broblasts. The chromosomal instability of WRN cells with the resultant accumulation of chromosomal damage may also have a role in this phenomenon. Third, normal WRN helicase may be required to triggr telomerase activation, although we do not have any evidence either to support or negate this possibility. The B-lymphoblastoid cell strains cultured for longterms from WRN patients showed higher instability of telomere length than the corresponding normal cell strains: the telomeres showed a drastic shortening and lengthening. An immortal human ®broblasts derived from SV40-transformed cells have no telomerase activity and showed a dynamic change in telomere size (Murnane et al., 1994). Chemical or gamma-ray treated human ®broblasts also elongated TRF size during immortalization (Sugihara et al., 1996). A similar phenomenon was reported in yeast and in

1917


1918

immortalized T-cell and B-cell lines (Lundblad and Blackburn, 1993; Strahl and Blackburn, 1996). Lundblad and Blackburn (1993) proposed that telomere lengths in yeast and these lymphoid cell lines are determined by both telomerase and telomeraseindependent mechanisms. A similar hypothesis was proposed by Reddel et al. (Bryan et al., 1995; Bryan and Reddel, 1997). Since most of the long-term cell strains over 100 PDLs from both WRN patients and normal individuals had no or only weak telomerase activity, cells from WRN patients may extraordinarily activate telomerase-independent mechanisms which is activated only moderately in wild type cells. Since all the cell strains of this study were derived from mass culture, selection of cells with longer TRFs would occur during cell passage. This may partly explain the dominance of cell populations with longer TRFs in some cell strains, especially at their younger PDLs. This explanation, however, cannot be used to interpret the sharp lengthening of TRFs occurring at mid/later PDL phases, since the cells were diluted by a half at every passage. The analysis of cell surface immunoglobulin at later PDLs also support this speculation, since the cell population at later PDLs were relatively homogeneous (clonal) as judged by the analyses of surface immunoglobulin by ¯ow cytometer (Kataoka et al., manuscript in preparation). From these results, we suggest the possibility that non-reciprocal recombination between telomere repeats as shown in yeasts (Lundblad and Blackburn, 1993) results in an increase in telomere repeats in Blymphoblastoid cell strains. If this were the case, enhancement of this process by the WRN gene mutation would be possible, since recombination events seem to be activated in WRN cells as speculated from the increased occurrence of chromosomal aberration in WRN cells, including translocation (Figure 7; ref. Salk et al., 1985a). Guarente (1996) proposed the possibility that the WRN gene mutation indeed stimulates recombination between repetitive DNA sequences, since mutation in sgs1, a gene coding the RecQ-type helicase gene of yeast, is known to increase recombination within the hundreds of copies of ribosomal DNA that are tandemly repeated in the genome (Ganglo et al., 1994). Whether or not the WRN gene really increase the recombination at telomeres remained to be clari®ed.

Materials and methods Patient samples Blood samples were collected from normal individuals and WRN patients. WRN was diagnosed from at least three of the following four major signs and symptoms (Goto et al., 1981): (1) characteristic disposition and stature; (2) premature senescence; (3) scleroderma-like skin changes; and (4) endocrinological abnormalities. The presence or absence of WRN gene mutations 1 ± 10 in WRN patients and normal individuals was investigated as previously reported (Oshima et al., 1996; Yu et al., 1996; Goto et al., 1997; Matsumoto et al., 1997). Table 1 shows the blood donors of WRN patients and normal individuals, including the parents or brothers and sisters of the patients. Chromosomes from normal random individuals have none of these mutations (Matsumoto et al., 1997).

Establishment of B-lymphoblastoid cell strains, cellular passage and judgement of maximal PDLs Peripheral blood leukocytes (PBLs) of patients were puri®ed by a Lymphocyte Separation Medium (West Chester, PA) and were transformed by EBV propagated in marmoset cell line B95-8 (Miller, 1990) in the presence of 200 ng of Ciclosporin (Sandoz, Switzerland). Established cell strains were cultured in 25 cm2 bottles (Corning, NY) standing ¯at, using RPMI1640 medium (Nissui Pharmaceutical Co. Ltd., Tokyo) supplemented with 10% fetal calf serum, (Gibco, NY), Bath No. 30F6131C), 50 units/ml of penicillin G and 50 mg/ml of streptomycin sulfate (Sigma, St Louis), and 0.25 mg/ml of Fungison (Gibco BRL, NY at 378C in 7.5% CO2 atmosphere. Continuous cellular passages were carried out every 3 or 4 days (twice a week) by exchanging a half volume of cultivating cell suspension with a fresh medium. During the 3 ± 4 day culture period, the cells usually attained a maximal cell density of betwwen 1 to 26106/ml for both normal and WRN cell strains. All the cell strains were judged to be free of mycoplasma using Mycoplasma Primer Set (Stratagene, La Jolla, CA). The maximal population doubling levels (PDLs) were judged as follows. At a crisis phase, the culture medium obviously shifted to an alkaline pH and the maximal cell density decreased. When the maximal cell density decreased to below about 2 6 105/ml, the cells were transferred to a 6well culture plate, and the culture was continued by changing a half of the supernatant medium. Usually, the cultured cells were watched for three weeks to two months under this condition. If they did not show a recovery of proliferation activity, the PDLs, when the cells were transferred to the 6well plate culture were judged to be at their maximum. Occasionally, the cells showed a partial recovery after a crisis, but in all cases they stopped proliferation, with the exception of the N6803, N0010 and N0013 cell lines that will be mentioned in the Results section. At a crisis, the cells showed signi®cantly lowered proliferating activity as judged by the decreased incorporation of bromodeoxyuridine assessed by a ¯ow cytometer, Epics Elite ESP (Coulter Corporation, Miami) (Dolbeare et al., 1983). At every interval of about 20 to 30 PDLs, parts of all the cell strains were stored in a deep freezer at 71158C in fetal calf serum containing 10% dimethyl sulfoxide (Aldrich Chem Co, Inc., Milwaukee). Cells of appropriate PDLs were thawed and used for culture when necessary. Cytogenetic analyses Chromosomes were prepared by a conventional method (Freshney, 1987): cells were arrested in metaphase by colcemid, ®xed in methanol/acetic acid (3:1) and spread on acid/alcohol-cleaned slides. They were then stained by the trypsin/Giemsa banding method. For each sample, 20 metaphase cells were examined. If chromosomal aberration, such as telomeric association, ring formation, translocation, deletion and addition, was observed, the cell was judged as abnormal. The chromosome analyses were done completely blind in an independent institute (BML Co. Ltd., Saitama). Telomerase assay We used a PCR-based modi®ed telomeric repeat amplification protocol (TRAP) assay to detect telomerase activity (Kim et al., 1994). To measure the telomerase activity and to evaluate the inhibitor of Taq DNA polymerase in the cell extract, we used an internal telomerase assay standard (ITAS) for the TRAP assay (Wright et al., 1995). The cell pellet (16105) was suspended in 200 ml of cold TRAP lysis buer in an Eppendorf safe-lock 2.0 ml tube, incubated for 30 min on ice, and centrifuged at 15 000 g for 20 min at


48C. The supernatant (ca. 180 ml) was collected, and aliquots were placed into two 1.5 ml tubes, rapidly frozen in liquid nitrogen, and stored at 7808C. The extract equivalent of 103 cells (2 ml) was incubated with 48 ml of 20 mM Tris-HCl (pH 8.3), 1.5 mM MgCl2, 68 mM KCl, 0.05% Tween 20, 1 mM EGTA, 50 mM dNTP, 1 mg (0.5 mM) T4g32 protein, 5 mg BSA, 2 units Taq DNA polymerase (AmpliTaq; PerkinElmer/Cetus), a-[32P]dCTP (148 KBq/assay, 110 TBq/mmol) (Amersham Co., Ltd.), 5 attg ITAS DNA and 0.1 mg of extra puri®ed TS primer (5'AATCCGTCGAGCAGAGTT-3'; NipponGene Co. Ltd.) at 208C for 30 min and then heated to 908C for 3 min to terminate the telomerase reaction. During this step, 0.1 mg (2 ml) extra puri®ed CXII primer (5'-CCCTTACCCTTACCCTTACCCT-3'; NipponGene Co. Ltd.) was added, and the reaction mixture was subjected to 31 PCR cycles at 948C for 45 s, 508C for 45 s and 728C for 90 s (2 min for ®nal step). Ten ml of PCR product was separated by electrophoresis on a 12% non-denaturating polyacrylamide gel and the 6-base pair ladder and ITAS signal were visualized by autoradiography. In all the series of the TRAP assay, we used an extract of L1210 cells as a positive control and lysis buer as a negative control. The intensity of telomerase ladder and ITAS band were analysed by phosphoimage analyser. For the semi-quantitative assessment of telomerase activity, the intensity of the telomerase signal (TRAP ladder) was normalized by ITAS signal (ITAS-TRAP assay) (Wright et al., 1995), as well as by hybridization protection assay (HPA-TRAP assay) measuring telomerase productspeci®c HPA probe (Hirose et al., 1997). Results from ITASTRAP and HPA-TRAP were parallel. In HPA-TRAP assay, ¯uorescent intensity versus amount of cell extract showed linearity by semi-log plot from one cell through 1000 cell equivalent extracts of immortal tumor cell line. Telomerase activity in serially diluted extract from 1000 lymphoblastoid cells were compared with that in serially diluted extracts from a tumor cell line, L1210, as a reference standard of the activity. The determined telomerase activities from HPATRAP were classi®ed into four grades by approximately 10fold dierence of intensity; S, as strong as L1210; M, moderate between L1210 and 1/10 of it; W, weak between 1/ 10 and 1/100 of L1210; N, below than 1/100 of L1210 or negative. Examples of corresponding results from ITASTRAP were shown in Figure 2. Determination of the lengths of the terminal restriction fragments (TRF) Genomic DNA was puri®ed by DNA extraction kits (Stratagene, La Jolla, CA). Five mg of genomic DNA was digested with restriction endonuclease HinfI (50 U) overnight at 378C. The HinfI digested genomic DNA was precipitated with ethanol containing 1/10 volume of 3 M sodium acetate. The DNA was dissolved in 10 ml of Tris HCl-EDTA buer and 2 ml of 66gel-loading buer (0.25% bromophenol blue, 0.25% xylene cyanol, 30% glycerol) was added, followed by electrophoresis in 16Tris-borateEDTA buer on 0.7% agarose gels. After electrophoresis, DNA in the gel was denatured by soaking in 0.5 M sodiumhydroxide-1.5 M NaCl for 30 min, neutralized by soaking in 0.5 M Tris HCl (pH 7.5)-1.5 M NaCl for 30 min,

transferred to a nylon membrane Hybond N (Amersham Corp, UK), and baked at 808C for 2 h. Digestion and transfer of DNA were con®rmed by staining gel with ethidium bromide. The membrane was hybridized with 5'32P-labeled (TTAGGG)4 telomeric oligonucleotide probe in 56SSC with 1% SDS for 12 ± 18 h. After hybridization, the ®lter was washed twice with a mixture of 26SSC and 1% SDS for 5 min at 378C, twice with a mixture of 0.26SSC and 1% SDS for 5 min at 378C, and exposed to Fuji XR ®lm with an intensifying screen. The intensity of TRF signal was analysed on a Macintosh computer using the public domain NIH Image program (written by Wayne Rasband at the US National Institutes of Health and available from the Internet by anonymous ftp from zippy.nimh.nih.gov or on ¯oppy disk from NTIS, 5285 Port Royal Rd., Spring®eld, VA22161, part number PB93504868). Cycling oligonucleotide-primed in situ synthesis (cyclingPRINS) Metaphase chromosomes were prepared as described in the Cytogenetic and immunological analyses section. To study the localization of telomere sequences, cycling-PRINS was performed by the method of Gosden and Lawson (Gosden and Lawson, 1994) with some modi®cation, as we now brie¯y described. The oligonucleotide of 5'-(TTAGGG)4-3' (Sawady, Tokyo) was used as a primer. The reaction mixture was as follows: 1 ml of each of 10 mM dATP, 10 mM dGTP 1mM dTTP and 1 mM biotin-16-dUTP (Boehringer Mannheim, Mannheim); 5 ml of 106Taq buer 2 ml of 20 mM primer, 4 ml of 25 mM MgCl2, 0.5 ml of 20 mg/ml BSA, 5 ml of DMSO and 28.5 ml of distilled water. Then 5 units of Taq polymerase (Promega) was added, and the reaction was carried out with a Perkin Elmer Gene Amp In Situ PCR system 1000. First, the slides were heated at 948C for 1 min followed by 25 cycles at 948C for 1 min, at 608C for 1 min, at 758C for 5 min and ®nally at 758C for 10 min. The reaction was stopped with a solution containing 50 m M EDTA and 500 mM NaCl, and the slides were washed. Biotin-16-dUTP was detected with a biotin-detection system (ONCOR, Gaithersburg, MD) consisting of avidin-FITC and an anti-avidin antibody. All the slides were observed under a Nikon OPTIPHOTO-2 ¯uorescent microscopy and the images were stored by Mac Probe Version 3.2.1. (Perceptive Scienti®c Instruments, Inc., League City, Texas).

Acknowledgements We thank Professor T Sairenji, School of Life Science, Faculty of Medicine, Tottori University for his advice in the preparation of EBV-transformed B-lymphoblastoid cell strains, and Mr Yoshito Tsurukubo of BML Co. Ltd. for the chromosomal analysis. We thank also Ms Chie Itoh and Ms Kumiko Fujita for their technical assistance. This work was supported by the Drug Organization (The Organization for Drug ADR Relief, R&D Promotion and Product Review, supervised by the Ministry of Health and Welfare of the Japanese Government).

References Avilion AA, Paityszek MA, Gupta J, Shay JW, Bacchetti S and Greider CW. (1996). Cancer Res., 56, 645 ± 650. Blackburn EH. (1991). Nature, 350, 569 ± 573. Bryan TM, Englezou A, Gupta J, Bacchetti S and Reddel RR. (1995). EMBO J., 14, 4240 ± 4208. Bryan TM and Reddel RR. (1997). Eur. J. Can., in press.

Counter CM, Botelho FM, Wang P, Harley CB and Bacchetti S. (1994). J. Virol., 68, 3410 ± 3414. Dolbeare F., Gratzner H, Pallavicini MG and Gray JW. (1983). Proc. Natl. Acad. Sci. USA, 80, 5573 ± 5577. Epstein CJ, Martin GM, Schultz AL and Motulsky AG. (1966). Medicine, 45, 177 ± 221.

1919


1920

Freshney RI. (1987). Culture of Animal Cells. Alan R Liss, Inc.: New York, pp. 175 ± 179. Ganglo S, McDonald JP, Bendixen C, Arthur L and Rothstein R. (1994). Mol. Cell. Biol., 14, 8391 ± 8398. Girardi AJ, Jensen FC and Koprowski H. (1965). J. Cell. Comp. Physiol., 65, 69 ± 83. Gosden J and Lawson D. (1994). Hum. Mol. Genet., 3, 931 ± 936. Goto M, Imamura O, Kuromitsu J, Matsumoto T, Yamabe Y, Tokutake Y, Suzuki N, Mason B, Drayna D, Sugawara M, Sugimoto M and Furuichi Y. (1997). Human Genet., 99, 191 ± 193. Goto M, Miller RW, Ishikawa Y and Sugano H. (1996). Cancer Epidemiol. Biomarkers Prev., 5, 239 ± 246. Goto M, Rubenstein M, Weber J, Woods K and Drayna D. (1992). Nature, 355, 735 ± 738. Goto M, Tanimoto K, Horiuchi Y and Sasazuki T. (1981). Clin. Genet., 19, 8 ± 15. Guarante L. (1996). Cell, 86, 9 ± 12. Hay¯ick L and Moorhead PS. (1961). Exp. Cell Res., 25, 585 ± 621. Hirose M, Abe-Hashimoto J, Ogura K, Tahara H, Ide T and Yoshimura T. (1997). J. Cancer Res. Clin. Oncol., in press. Kim, NW, Piatyszek MA, Prowse KR, Harley CB, West MD, Ho PL, Coviello GM, Wright WE, Weinrich SL and Shay JW. (1994). Science, 266, 2011 ± 2015. Lundblad V and Blackburn EH. (1993). Cell, 73, 347 ± 360. Martin GM. (1978). Bergsma D and Harrison DE (eds). Alan R Liss, Inc.: New York pp. 3 ± 39. Martin GM, Sprague CA and Epstein CJ. (1970). Lab Invest, 23, 86 ± 92. Matsumoto T, Imamura O, Yamabe Y, Kuromitsu J, Tokutake Y, Shimamoto A, Suzuki N, Satoh M, Kitao S Ichikawa K, Kataoka H, Sugawara K, Thomas W, Mason B, Tsuchihashi Z, Drayna D, Sugawara M, Sugimoto M, Furuichi Y and Goto M. (1997). Hum. Genet., 100, 123 ± 130.

Miller G. (1990). Virology. Fields, BN and Knipe KM (eds). Raven Press: New York, pp. 1921 ± 1958. Murnane JP, Sabatier L, Marder BA and Morgan WF. (1994). EMBO J., 13, 4953 ± 4962. Norrback KF, Dahlenborg K, Carlsson R and Roos G. (1996). Blood, 88, 222 ± 229. Oshima J, Yu C-E, Piussan C, Klein G, Jabkowski J, Balci S, Miki T, Nakura J, Ogihara T, Ells J, Smith MdC, Melaragno MI, Fraccaro M, Scappaticci S, Matthews J, Ouais S, Jarzebowicz A, Schellenberg GD and Martin GM (1996). Hum. Mol. Genet., 5, 1909 ± 1913. Poot M, Hoehn H, Runger TM and Martin GM. (1992). Exp. Cell Res., 202, 267 ± 273. Runger TM, Bauer C, Dekant B, Moller K, Sobotta P, Czerny C, Poot M and Martin GM. (1994). J. Invest. Dermatol., 102, 45 ± 48. Salk D, Au K, Hoehn H and Martin GM. (1985a). Adv. Exp. Med. Biol., 541 ± 546. Salk D, Bryant E, Hoehn H, Johnston P and Martin GM. (1985b). Adv. Exp. Med. Biol., 305 ± 311. Schellenberg GD, Martin GM, Wijsman EM, Nakura J, Miki T and Ogihara T. (1992). Lancet, 339, 1002. Schulz VP, Zakian VA, Ogburn CE, McKay J, Jarzebowicz AA, Edland SD and Martin GM. (1996). Hum. Genet., 97, 750 ± 754. Strahl C and Blackburn EH. (1996). Mol. Cell. Biol., 16, 53 ± 65. Sugihara S, Mihara K, Marunouchi T, Inoue H and Namba M. (1996). Hum. Genet., 97, 1 ± 6 Wright WE, Shay JW and Piatyszek MA. (1995). Nucl. Acids Res., 23, 3794 ± 3795. Yu CE, Oshima J, Fu YH, Wijsman EM, Hisama F, Alisch R, Matthews S, Nakura J, Miki T, Ouais S, Martin GM, Mulligan J and Schellenberg GD. (1996). Science, 272, 258 ± 262.