According to the telomere hypothesis of senescence, the progressive shortening of telomeres that occurs upon division of normal somatic cells eventually leads to cellular senescence. The immortalisation of human cells is associated with the acquisition of a telomere maintenance mechanism which is usually dependent upon expression of the enzyme telomerase. About one third of in vitro immortalised human cell lines, however, have no detectable telomerase but contain telomeres that are abnormally long. The nature of the alternative telomere maintenance mechanism (referred to as ALT, for Alternative Lengthening of Telomeres) that must exist in these telomerase-negative cells has not been elucidated. It has previously been shown that abnormal lengthening of yeast telomeres may occur due to mutations in the yeast telomerase RNA gene. That this is not the mechanism of the abnormally long telomeres in ALT cell lines was demonstrated by the finding that seven of seven ALT lines have wild-type human telomerase RNA (hTR) sequence, including a novel polymorphism that is present in 30% of normal individuals. We found that two ALT cell lines have no detectable expression of the hTR gene. This shows that the ALT mechanism in these cell lines is not dependent on hTR. Expression of exogenous hTR via infection of these cells with a recombinant hTR-adenovirus vector did not result in telomerase activity, indicating that their lack of telomerase activity is not due to absence of hTR expression. We conclude that the ALT mechanism is not dependent on the expression of hTR, and does not involve mutations in the hTR sequence.
Normal human cells have a finite lifespan in vitro, after which they cease dividing and undergo senescence (1 ). Tumour cells, however, will often divide indefinitely in vitro, and hence have overcome the normal limits to proliferation and are immortal. The molecular basis of senescence and immortalisation is not well understood, but one hypothesis for which there has recently been a large amount of evidence involves the shortening of telomeres. Human telomeres consist of large tracts of the 6 bp repeat TTAGGG (reviewed in ref. 2 ), which diminish in size as somatic human cells age, both in vitro and in vivo (3 -8 ). Germline cells, however, do not exhibit telomere shortening (9 -11 ), and have been shown to have active telomerase (13 ), an enzyme which adds telomeric repeats to telomeres (12 ). It has been proposed that the expression of telomerase leads to stabilisation of telomere length and is a requirement for the immortalisation of human cells (14 ,15 ). Using a very sensitive PCR-based assay (13 ,16 ) telomerase activity has been detected in most tumour-derived cell lines and 70-90% of human tumours of different types (17 ).
Many in vitro-immortalised human cell lines, however, lack telomerase activity and all have very long and heterogeneous telomeres, up to about 20-50 kb (13 ,18 -20 ), distinct from those of telomerase-positive lines. This suggests that their telomeres have been lengthened by a novel mechanism, which we have called ALT (Alternative Lengthening of Telomeres) (21 ). ALT has been identified in about one third of cell lines immortalised in vitro using simian virus 40 (SV40), human papillomavirus (HPV) and chemical carcinogens or spontaneously immortalised cells (21 ). ALT has also recently been detected in four tumour-derived cell lines and a small number of tumour samples, indicating that ALT is a phenomenon involved in in vivo tumour development, albeit at a much lower frequency than in in vitro-derived cell lines (T.M.Bryan, A.Englezou,L.Dalla-Pozza, M.Dunham and R.R.Reddel, manuscript in preparation).
The mechanism of telomere lengthening in ALT is currently unknown. It does not, however, appear to be due to a burst of telomerase activity that has subsequently been switched off since a gradual increase in telomere length in two ALT cell lines occurred in the absence of telomerase activity (19 ,20 ). It also does not appear to be a result of movement of transposons to telomeres, the telomere-maintenance mechanism utilised by Drosophila, since the telomeric sequences hybridise strongly to TTAGGG repeat probes (18 ,20 ).
Another possible explanation for ALT involves the RNA subunit of telomerase. Telomerase is a ribonucleoprotein, consisting of one RNA and at least two protein subunits, all of which are encoded by separate genes. The gene for the RNA subunit has been cloned from more than 20 ciliate species (22 -27 ), the yeasts Saccharomyces cerevisiae and Kluyveromyces lactis (28 ,29 ), mice (30 ) and humans (31 ). These RNAs all contain a sequence that is complementary to about 1.5 telomere repeats and has been shown to act as the template for the addition of nucleotides to the G-rich strand of the telomere. Two different mutations within this template region in the telomerase RNA of Tetrahymena thermophila resulted in an altered repeat being added to telomeres, which became longer and had a broader size distribution (32 ). Similarly, telomerase RNA mutations in K.lactis caused `runaway telomere elongation' of up to 100 times the normal telomere length (29 ). It is proposed that this elongation is caused by the inability of a negative regulator of telomerase to bind to altered telomere repeats, leading to unregulated telomerase action.
The similarity between these long and heterogeneous telomeres in ciliates and yeast and those in human ALT cell lines suggested that ALT may be due to a mutation in the gene for the human telomerase RNA, hTR. hTR was recently cloned and sequenced; it encodes an RNA of 445 nucleotides (nt) and includes 11 nt complementary to TTAGGG (31 ,33 ). We hypothesised that the putative mutation, as well as leading to dysregulated telomere lengthening, makes telomerase undetectable in the in vitro assay for telomerase activity. To test this, we isolated and sequenced hTR from a set of human cell lines with ALT activity.
We used PCR to amplify 598 base pairs (bp) of the hTR gene, using the primers indicated in Figure 1 . This segment of the gene contains the whole 445 nt transcribed region (33 ). The location of the 11 nt template region is also indicated in Figure 1 . The 598 bp segment was amplified from genomic DNA of six cell lines with telomerase activity and seven cell lines with no telomerase and the long telomeres characteristic of ALT (Table 1 ). Both strands of the PCR product were sequenced, using a total of seven internal hTR primers (see Materials and Methods).
The fact that we were able to amplify hTR from genomic DNA of all seven ALT cell lines indicates that this gene has not suffered a homozygous deletion in these cell lines. Furthermore, the gene has wt sequence in all seven cell lines, ruling out the possibility that a mutant hTR is responsible for ALT. We were therefore interested to determine whether this wt hTR is being expressed in these cell lines.
hTR was amplified by PCR from the cell line HeLa, purified and cloned into plasmid pGEM-T. This cloned hTR fragment was labelled with 32P and used as a probe on a Northern blot of RNA from two telomerase-positive cell lines (293 and HeLa), two normal fibroblast cell strains (HFF5 and WI38) and nine ALT cell lines (Fig. 3 and data not shown). The normal cell strains expressed a low level of hTR. Most of the immortal cell lines produced higher but varying amounts of hTR, regardless of telomerase status. However, two ALT cell lines (SUSM-1 and WI38 VA13/2RA) produced no hTR RNA that was detectable on a Northern blot (Fig. 3 ). WI38 is the cell strain from which WI38 VA13/2RA was derived; there has therefore been a loss of hTR expression during immortalisation of this cell line. These two cell lines lacked detectable hTR expression even when the amount of RNA on the blot was increased from 10 to 40 [mu]g (Fig. 4 , first four lanes), whereas 1 [mu]g of 293 RNA was sufficient to detect expressed hTR (Fig. 4 , last two lanes). A band smaller than hTR is visible on this Northern blot for all cell lines except WI38 VA13/2RA (the band in SUSM-1 is the same size as the others, although it appears larger due to a `smile' in the gel). The identity of this band is unknown, but may represent an artefact of the in vitro-transcribed RNA probe (the probe used in Figure 3 , in contrast, was a cloned DNA fragment).
To test whether it is possible to regain telomerase activity in SUSM-1 and WI38 VA13/2RA by restoring hTR expression, hTR was introduced into these cell lines via an adenovirus vector. The cells were infected at high multiplicity with the recombinant virus and tested for expression of introduced hTR by Northern blot (Fig. 4 ). Both cell lines expressed high amounts of hTR after infection.
The cells were then lysed and the extract used in the PCR-based assay for telomerase activity (TRAP). No activity was detected in either cell line (Fig. 5 ). TRAP activity has been detected in telomerase-positive cell lines infected with adenoviruses containing mutant hTRs, using substrate primers appropriate for the mutations (L.Marusic et al., manuscript in preparation). This indicates that it is possible to reconstitute in vitro telomerase activity by transient expression of exogenous copies of hTR if active telomerase protein subunits are present. It remains a possibility, however, that SUSM-1 and WI38 VA13/2RA do have wt telomerase protein subunits that are present in an inactive form.
GM847 and GM639 (SV40-immortalised skin fibroblasts) were a gift from Dr O. Pereira-Smith. WI38 (normal lung fibroblasts), WI-38 VA13/2RA (SV40-immortalised WI38), HeLa (cervical carcinoma cell line), T24 (bladder carcinoma cell line) and 293 (adenovirus 5-transformed human embryonic kidney) were obtained from the American Type Culture Collection. HFF5 is a normal fibroblast cell strain established from foreskin fibroblasts obtained from Ralph Böhmer, Ludwig Institute of Cancer Research, Melbourne. SUSM-1 are 4-nitroquinoline (4NQO)- immortalised liver fibroblasts (39 ). BET-3M, BET-3a and BET-3b were established by transfection of human bronchial epithelial cells with the SV40 early region plasmid, pRSV-T (40 ,41 ). MeT-4A is a human mesothelial cell line immortalised with pRSV-T (E.Duncan et al., unpublished data). IIICF cells are fibroblasts from an individual with Li-Fraumeni Syndrome (42 ). IIICF-T/A6, IIICF-T/B1 and IIICF-T/C3 were immortalised with pRSV-T (47 ). IIICF/c cells spontaneously immortalised in vitro (19 ).
IIICF cells were maintained in Roswell Park Memorial Institute (RPMI) 1640 medium with 10% foetal bovine serum (FBS) and gentamicin. All other fibroblasts and 293 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) with 10% FBS and gentamicin. Epithelial cells were maintained in Laboratory of Human Carcinogenesis-9 (LHC-9) medium in flasks coated with a matrix of collagen and fibronectin (43 ). Mesothelial cells were maintained in Laboratory of Human Carcinogenesis-Mesothelial medium (LHC-MM) (44 ), also in coated flasks. Cultures were maintained in a humidified 37oC incubator with 5% (fibroblasts) or 3.5% (epithelial and mesothelial cells) CO2 in air.
Primers used for amplification of hTR were A307 (5'-GGGTTGCGGAGGGTGGGC) and A308 (5'-CGACTTTGGAGGTGCCTTCA). Genomic DNA (500 ng) from each cell line or individual was added to 250 ng of each primer in a buffer containing final concentrations of 10 mM Tris-Cl pH 9.2, 1.5 mM MgCl2, 75 mM KCl, 0.2 mM dNTPs (Boehringer Mannheim), 15% glycerol and 2.5 U Taq polymerase (Boehringer Mannheim) in a final volume of 50 [mu]l. After placement in a hot (94oC) PCR machine, the reactions were cycled 35 times at 94oC for 30 s, 55oC for 30 s and 72oC for 1 min, with a final elongation step of 72oC for 2 min.
PCR products were purified with PCR Spinclean DNA columns (Progen). They were sequenced with the fmol DNA Cycle Sequencing kit (Promega), using 6 ng template and 9 ng [[gamma]-32P]ATP end-labelled primer, cycling the reactions for 30 cycles at 95oC for 30 s, 55oC for 30 s and 70oC for 1 min. Sequencing primers used were: A309 (5'-ACGTCTCCTGCCAATTTG), A310 (5'-ATGAACGGTGGAAGGCGG), A311 (5'-GACTCGCTCCGTTCCTCT), A319 (5'-AGGGTGGGCCTGGGAGGG), A493 (5'-CCTTCCACCGTTCATTCTAG), A494 (5'-CGAAGAGTTGGGCTCTGTC) and A495 (5'-GGACTCGGCTCACACATG). To eliminate `stops' (bands in all lanes), completed sequencing reactions were treated with 1 [mu]l of a solution containing 5 U/[mu]l terminal deoxytransferase (Boehringer Mannheim), 40 mM Tris-Cl pH 7.5, 20 mM NaCl, 20 mM MgCl2 and 1 mM dNTPs (Boehringer Mannheim) at 37oC for 60 min.
Spinclean-purified hTR PCR fragments (see above) were digested separately with the following restriction enzymes: SmaI, StuI (Boehringer Mannheim), BsrDI and EaeI (New England Biolabs). Digested DNA was electrophoresed on 3.5% or 2% MetaPhor agarose gels (FMC Bioproducts) at 4 V/cm for 16 h and stained with ethidium bromide.
Figure 3 . RNA was isolated from cells using a Total RNA Isolation Kit (Advanced Biotechnologies). RNA (5 [mu]g) was subjected to electrophoresis on a 1% agarose formaldehyde gel at 6 V/cm for 4 h, transferred to Hybond-N membrane (Amersham) by capillary transfer and UV cross-linked.
A 598 bp PCR fragment of HeLa hTR (see above) was ligated into plasmid pGEM-T (Promega) and propagated in the SURE strain of Escherichia coli (Stratagene). The insert was released by digestion with NcoI and NsiI (Boehringer Mannheim) and electrophoresed on a 3.5% NuSieve agarose gel (FMC Bioproducts). The hTR fragment was excised, purified with PCR Spinclean columns (Progen), labelled with [[alpha]-32P]dATP with the Gigaprime labelling kit (Bresatec) and used as a probe on Northern blots. Hybridisation and washing were carried out according to the membrane manufacturer's instructions, at 65oC. RNA loading was confirmed by hybridisation at 55oC with a [[gamma]-32P]ATP end-labelled 18S ribosomal RNA oligonucleotide probe (5'-GCATATGCTACTGGCAGGATCAACCAGGTA).Figure 4 . Total RNA was isolated with TRIZOL (Gibco/BRL), electrophoresed on 1.2% agarose formaldehyde gels and transferred to nylon membranes (BMC). Hybridisation was with a digoxigenin-labelled RNA probe according to BMC protocols and detection was by chemiluminescence (BMC).
Cells were infected with virus (100 p.f.u./cell) for 45 min at room temperature, refed with growth medium and harvested after 48 h. Permissive 293 cells were infected with 10 p.f.u./cell and harvested after 24 h.
Telomerase activity was detected by the PCR-based TRAP method (13 ) with the following modifications: cell extracts were prepared by three cycles of freezing and thawing in the absence of detergent (46 ); the TS oligonucleotide was end-labelled with [[gamma]-32P]ATP by T4 polynucleotide kinase; and the reverse primer and Taq polymerase were added after telomerase elongation and warming of the reactions to 92oC (hot start).
We thank Melissa Dunham and Anna Englezou for excellent technical assistance. We also thank Dr Cal Harley for informative discussions. These studies were supported by a fellowship from the NSW State Cancer Council (to R.R.R.), a National Health and Medical Research Council of Australia Biomedical Research Scholarship (to T.M.B.), and grants from the Medical Research Council and the National Cancer Institute of Canada (NCIC) and Geron Corporation (USA) (to S.B.). S.B. is a Terry Fox Cancer Research Scientist of the NCIC.
Human Molecular Genetics
Pages
Introduction
Results And Discussion
Sequence analysis of hTR
Northern blot analysis of hTR expression
Introduction of hTR into ALT cell lines with no detectable hTR expression
Materials And Methods
Cell lines and culture conditions
PCR and sequencing of hTR
Restriction enzyme analysis of sequence changes
Northern blot analysis of hTR
Infection of cell lines with recombinant adenovirus
Telomerase assay
Acknowledgements
References
REFERENCES
This page is maintained by OUP admin. Last updated Mon May 12 18:10:04 BST 1997. Part of the OUP Journals World Wide Web service. Copyright Oxford University Press, 1996