Oncogene (2002) 21, 8262 – 8271 ª 2002 Nature Publishing Group All rights reserved 0950 – 9232/02 $25.00 www.nature.com/onc Telomere dysfunction and telomerase reactivation in human leukemia cell lines after telomerase inhibition by the expression of a dominant-negative hTERT mutant Franc¸ois Delhommeau1, Antoine Thierry1, Danie` le Feneux2, Evelyne Lauret1, Edwige Leclercq3, Marie He´ le` ne Courtier3, Franc¸oise Sainteny1, William Vainchenker1 and Annelise Bennaceur-Griscelli*,1,3 1INSERM U362, PR-1, Institut Gustave Roussy, 39/53 rue Camille Desmoulins, 94805 Villejuif, France; 2Service d’He´matologie et de Cytoge´ne´tique, Hoˆpital de Biceˆtre, 78 rue du Ge´ne´ral Leclerc, 94270 Le Kremblin Biceˆtre, France; 3Department of Clinical Biology, Service d’He´matologie Biologique, Institut Gustave Roussy, 39/53 rue Camille Desmoulins, 94805 Villejuif, France As activation of telomerase represents a key step in the chromosomes consisting of TTAGGG repeats (de malignant transformation process, experimental models Lange et al., 1990) and telomere-binding proteins to develop anti-telomerase drugs provide a rational basis (Broccoli et al., 1997; Chong et al., 1995). Telomeres for anticancer strategies. We analysed the short and long- protect the chromosome ends from degradation, term efficacy of a stably expressed dominant-negative recombination and DNA repair activities (Bailey et mutant (DN) of the telomerase catalytic unit (hTERT) in al., 1999; d’Adda di Fagagna et al., 1999; Samper et UT-7 and U937 human leukemia cell lines by using an al., 2000). In addition, telomere length is progressively IRES-e-GFP retrovirus. As expected, telomerase inacti- reduced with cell divisions (Blasco et al., 1997; Harley vation resulted in drastic telomere shortening, cytogenetic et al., 1990), due to the ‘end-replication problem’ instability and cell growth inhibition in all e-GFP positive (Hastie et al., 1990) and the putative exonuclease DN clones after 15 – 35 days of culture. However, despite activity in the CA-rich strand (Makarov et al., 1997). this initial response, 50% of e-GFP positive DN clones Telomere shortening in aging cells induces replicative with short telomeres escaped from crisis after 35 days of senescence (Morin, 1997). When telomeres reach a culture and recovered a proliferation rate similar to the critical size, chromosomes become unstable and under- control cells. This rescue was associated with a go end-to-end fusions, DNA fragmentation, and telomerase reactivation inducing telomere lengthening. mutations (Blasco et al., 1997; Gisselsson et al., We identified two pathways, one involving the loss of the 2001). Conversely, telomere length is stable in cells DN transgene expression and the other the transcrip- with long-lived replicative life spans, such as germ line tional up-regulation of endogenous hTERT with and embryonic cells, and in immortalized and tumor persistence of the DN transgene expression. Although cells (Wright et al., 1996). this second mechanism appears to be a very rare event Telomerase activation is the major mechanism that (one clone), these findings suggest that genomic instability maintains telomere integrity (Morin, 1989). The induced by short telomeres after telomerase inhibition human telomerase reverse transcriptase (hTERT) is might enhance the probability of activation or selection of an RNA-dependent DNA polymerase which synthe- telomere maintenance mechanisms dependent on hTERT sizes telomeric DNA using the human telomerase transcription. RNA (hTR) as a template (Colgin and Reddel, 1999; Oncogene (2002) 21, 8262 – 8271. doi:10.1038/sj.onc. Feng et al., 1995; Morin, 1989; Nugent and 1206054 Lundblad, 1998). Unlike germline and embryonic cells, somatic cells do not have telomerase activity, Keywords: telomere; telomerase; leukemia; retrovirus; with the exception of regenerative tissues such as cell death; therapy hematopoietic stem cells (Chiu et al., 1996; Morrison et al., 1996) or lymphocytes (Hiyama et al., 1995). In contrast, most malignant cells are characterized by an increased telomerase activity (Meyerson et al., 1997; Introduction Raymond et al., 1996). Enhanced telomerase activity represents an important step in the transformation Telomeres are dynamic DNA-protein complexes process of human cells, as the combined expression of (Blackburn, 2001) that cap the ends of linear hTERT, SV40 large T antigen and H-Ras results in direct conversion from normal to tumor cells (Hahn et al., 1999a). In addition, acquisition of telomerase *Correspondence: A Bennaceur-Griscelli; E-mail: [email protected] activity may be associated with escape from senes- Received 3 July 2002; revised 13 September 2002; accepted 17 cence (Bodnar et al., 1998; Counter et al., 1998a; September 2002 Pendino et al., 2001). These findings validate Telomerase inhibition in human leukemia cell lines F Delhommeau et al 8263 telomerase as an important drug target for cancer therapy. Telomerase inhibitors are considered as promising agents for a wide variety of human malignancies. Impairing telomerase activity in human solid tumors by peptide nucleic acid or 2’-O-MeRNA (Herbert et al., 1999), chemical compounds (Damm et al., 2001), anti-sense oligonucleotides (Kondo et al., 1998), mutant hTR (Kim et al., 2001), hTERT dominant- negative (DN) mutants (Hahn et al., 1999b; Zhang et Figure 1 Schematic representation of the Mig-R and Mig-R-DN al., 1999), or by G-quadruplex DNA ligands (Riou et constructs al., 2002) results in cell growth inhibition and loss of tumorigenicity. Surprisingly, ‘revertant’ clones were selected in one murine tumor cell line after telomerase inhibition by a DN mutant, suggesting differences in per well. A dramatic decrease in telomerase activity telomerase regulation between human and murine cells was observed in 77% of U937 DN clones (31 out of (Sachsinger et al., 2001). Moreover, lessons learned 40) and 63% of UT-7 DN clones (47 out of 75), from knock-out mice have to be considered since compared to control vector clones (Figure 2a). The carcinogenesis is accelerated when both murine incidence of telomerase negative clones in DN clones telomerase RNA and p53 genes are inactivated (Chin was not different between clones with low and high et al., 1999). In addition, the possibility of resistance expression of e-GFP (data not shown). Consequences emerging from indirect telomerase-based therapies has of telomerase inhibition on telomere length were been recently observed in acute promyelocytic leukemia evaluated by Southern blot and Q-FISH in transduced cells treated by retinoic acid (Pendino et al., 2002). U937 and UT-7 clones after 25 – 35 and 20 – 25 PDs Thus it is possible that genetic instability induced by respectively (Figure 2b). Telomere length of the clones short telomeres might promote the selection of resistant transduced by the empty vector was not significantly sub-clones in human tumor cells subjected to a long different from uninfected parental cell lines (Figure 2b, term anti-telomerase strategy. As the possibility that lanes 1, 2, 3). During expansion, all DN U937 and UT- tumor cells might overcome telomerase inhibition has 7 clones exhibiting initially a reduced telomerase not been carefully evaluated in human tumors, we activity remained e-GFP positive and TRAP negative investigated this hypothesis in this work by studying (data not shown) indicating the stability of the the susceptibility of telomerase positive human leuke- transgene expression. Progressive telomere shortening mia cell lines to growth and death after short and long- was observed in all these clones, reaching values term telomerase inactivation by a DN-hTERT mutant around 2 kb (Figure 2b, lanes 4, 5, 6). Q-FISH (Hahn et al., 1999b). We studied the effects of analyses confirmed these data as a net decrease in telomerase inactivation in the myelomonocytic telomere fluorescence intensity was observed in DN U937 cell line (Harris and Ralph, 1985) and in the clones when compared to control clones (Figure 2c). UT-7 megakaryocytic cell line (Komatsu et al., 1991). We show that after an initial response defined by a DN affects cell growth and triggers cytogenetic instability critical telomere shortening, growth arrest and massive and cell death cell death, long-term culture results in the emergence of surviving clones, which escape from crisis by telomer- Clonogenicity (i.e. frequency of proliferating clones ase reactivation. Two mechanisms are responsible for from initial seeded cells) of DN cells was similar to this reactivation: the first one, occurring in both cell control cells in U937 and UT-7 cell lines (25 out of 60 lines, involves the loss of DN transgene expression control vector clones versus 51 out of 120 DN clones following telomere dysfunction and the second one, and 102 out of 180 control vector clones versus 154 out exclusively observed in one U937 DN clone, involves of 240 DN clones respectively), indicating the absence the increase in endogenous hTERT transcription with of cytotoxicity of DN in first mitoses. After 2 weeks of the persistence of the DN transgene expression. culture, 82% of DN UT-7 clones had stopped growing compared to 55% of control UT-7 clones. In U937 clones, a delayed effect relative to initial longer Results telomeres in this cell line (5.5 kb compared to 3.5 kb in UT-7 cell line) was observed. The loss of viability DN inhibits telomerase activity and induces telomere induced by telomere shortening was confirmed by an shortening in leukemia cell lines increased rate of dead cells when analysing propidium U937 and UT-7 cells were transduced by the control iodide incorporation in both U937 and UT-7 DN vector Mig-R and Mig-R DN-hTERT retroviruses clones (Figure 3a). These clones displayed senescent (Figure 1). After three days of culture, e-GFP positive features such as enlarged cells and granular morphol- cells were sorted by flow cytometry and seeded at one ogy (data not shown). Since genetic instability is a cell per well. Telomerase activity was measured in direct consequence of shortened telomeres’ dysfunction, proliferating clones when they reached 1006103 cells we performed cytogenetic analyses.