Once immortalized, human cells are susceptible to transformation by introduction of an oncogene such as ras. Several lines of evidence now suggest that the maintenance of telomere length is a major determinant of replicative lifespan in human cells and thus of the immortalized state. The majority of human tumor cells acquire immortality through expression of the catalytic subunit of telomerase (hTERT), whereas others activate an alternative mechanism of telomere maintenance (ALT) that does not depend on the actions of telomerase. We have examined whether ALT could substitute for telomerase in the processes of transformation in vitro and tumorigenesis in vivo. Expression of oncogenic H-Ras in the immortal ALT cell line GM847 did not result in their transformation. However, subsequent ectopic expression of hTERT in these cells imparted a tumorigenic phenotype. Indeed, this outcome was also observed after introduction of a mutant hTERT that retained catalytic activity but was incapable of maintaining telomere length. These studies indicate that hTERT confers an additional function that is required for tumorigenesis but does not depend on its ability to maintain telomeres.

Cultured primary human cells inevitably enter a state of replicative senescence for which the specific molecular trigger is unknown. We show that the single-strand telomeric overhang, a key component of telomere structure, is eroded at senescence. Expression of telomerase prevents overhang loss, suggesting that this enzyme prevents senescence by maintaining proper telomere structure. In contrast, progressive overhang loss occurs in cells that avoid senescence through the inactivation of p53 and Rb, indicating that overhang erosion is the result of continuous cell division and not a consequence of senescence. We thus provide evidence for a specific molecular alteration in telomere structure at senescence and suggest that this change, rather than overall telomere length, serves to trigger this state. Fig. 1 Measurement of the telomeric overhang using T-OLA. a, Schematic of the T-OLA method. Total genomic DNA is incubated with a radioactively labeled oligonucleotide of the sequence (CCCTAA) 4 . The oligonucleotide hybridizes to accessible G-rich telomeric overhangs. A ligase is added to the reaction, resulting in ligation of oligonucleotides that hybridized to adjacent positions along the overhang. This creates a collection of DNA fragments with sizes increasing by units of 24 nt that are denatured and separated on a denaturing polyacrylamide gel. A population of telomeres carrying a longer average telomeric overhang length is expected to produce a stronger total T-OLA signal and a greater maximal product length. b, T-OLA analysis of DNA fragments of 200 bp and 350 bp representing 10 7 , 10 8 and 10 9 telomeres (7,8,9) on the background of yeast genomic DNA. Analysis was also done with the same fragments in the absence of denaturation (no den) and with a 550-bp DNA fragment of random sequence (random). Product sizes are indicated. c, Analysis of genomic DNA from BJ foreskin fibroblasts at PD 40 pretreated with different enzymes as indicated. Bal31 is a single-strand endonuclease, ExoI is a single-strand 3′ specific exonuclease, ExoIII degrades recessed and blunt 3′ ends, T7 Gene 6 (T7G6) exonuclease degrades 5′ ends and Klenow DNA polymerase fills in 5′ overhangs but not 3′ overhangs. To verify equal amounts of input DNA, quantitative PCR was done on each T-OLA sample using primers directed against the genomic GAPD locus. Results of these PCR reactions are displayed below each T-OLA panel. d, Analysis using oligonucleotides complementary to the G-rich strand (CCCTAA), to the C-rich strand (TTAGGG) or to neither (CCCTTA). e, T-OLA dose response using genomic DNA extracted from GM847 cells. 2, 4, 10 and 20 µg of genomic DNA were analyzed.