Telomerase activity is necessary and sufficient for immortality in many cells and hence represents a prime target for antitumor strategies. Here, we show that a hammerhead ribozyme cleaves human telomerase (hTERT) mRNA in vitro. Stable transfection in clones of the human breast tumor line MCF-7 and the immortal breast cell line HBL-100 results in expression of the ribozyme, diminishes the abundance of hTERT mRNA, and inhibits telomerase activity. This led to shortened telomeres, inhibition of net growth, and induction of apoptosis. In HBL-100 mass cultures infected with a ribozyme-expressing adenovirus diminution of hTERT mRNA, attenuation of telomerase activity, inhibition of net growth, and induction of apoptosis was found as well. Attenuation of telomerase activity increased the sensitivity of HBL-100 and MCF-7 clones specifically to inhibitors of topoisomerase. Likewise, expression of exogenous telomerase in originally telomerase-negative human fibroblasts decreased their sensitivity to topoisomerase poisons but not to a number of other cytotoxic drugs. The data validate a ribozyme approach for telomerase inhibition therapy in cancer and suggest that it might be combined advantageously with topoisomerase-directed chemotherapy.
Maintenance of telomeres is necessary and sufficient for cellular immortality in a number of human cell types (1) . The vast majority of tumors are made of immortal cells, and most of them maintain their telomeres by activation of the enzyme, telomerase. Telomerase is strongly repressed in most human somatic tissues. Accordingly, inhibition of telomerase has been suggested as a specific means to “remortalize” tumor cells (2) .
Specific inhibition of telomerase in living human cells leading to telomere shortening, and eventually, cell death has been demonstrated in a still limited number of recent papers. Some studies used modified antisense oligonucleotides directed against the template RNA molecule (hTR) of the telomerase complex (3, 4, 5, 6, 7, 8) . Recently, the successful inhibition of telomerase in cancer cells by expression of a dominant-negative version of the catalytic subunit of the human telomerase (hTERT) was reported (9 , 10) . Very few attempts exploited the potential of the ribozyme technology to inhibit telomerase. Some papers reported the inhibition of telomerase in cell lysates, but not in living cells, by anti-hTR ribozymes (11 , 12) . One recent report showed inhibition of telomerase by expression of an anti-hTR ribozyme in endometrial carcinoma cells but did not demonstrate significant growth inhibition or cell death (13) . To our knowledge, there have been no successful attempts to inhibit telomerase by ribozymes directed against the catalytic subunit, hTERT.
In a significant number of model systems studied, a long delay between inhibition of telomerase and cessation of cell growth and/or massive induction of apoptosis was observed. As expected, this delay was especially prominent in cell systems with long telomeres (3 , 7 , 8 , 10) . Telomeres are highly sensitive to oxidative damage (14) , and telomere shortening can be significantly accelerated by oxidative stress (15, 16, 17) and, possibly, by other DNA-damaging treatments as well. There are few data in the literature indicating that the cytotoxic or cytostatic effects of telomerase inhibition and treatment with DNA-damaging drugs might be cumulative. Inhibition of telomerase increased the susceptibility of human glioblastoma cells to cisplatin-induced apoptosis (4) . It enhanced the rate of apoptosis induced by either staurosporine, amyloid β peptide, or oxygen-free radicals produced by Fe2+ (18) . Human ovarian and melanoma cell lines with long telomeres were found to be less sensitive to DNA-damaging drugs, especially cisplatin (19) .
We show here that a hammerhead ribozyme directed against a sequence within the hTERT mRNA T-motif specifically cleaved its target sequence in vitro and diminished the target RNA in vivo. It attenuated telomerase activity in stable transfected clones of the immortal, telomerase-positive human breast epithelial cell line HBL-100 and the breast cancer cell line MCF-7 and in adenovirus-infected mass cultures of HBL-100. Apoptosis was induced, and net growth was strongly attenuated and in some clones abolished. Clones with reduced telomerase activity showed an increased sensitivity to inhibitors of topoisomerase. Accordingly, human fibroblasts expressing an exogenous hTERT gene became less sensitive to three different topoisomerase poisons but not to a number of other cytotoxic drugs acting via topoisomerase-independent mechanisms.
MATERIALS AND METHODS
The immortal human breast epithelial cell line HBL-100 and the human breast adenocarcinoma line MCF-7 were obtained from the American Type Culture Collection. Cells were grown in RPMI 1640 (HBL 100) or DMEM (MCF-7) plus 10% FCS (PAA Laboratories, Cölbe, Germany). Parental BJ human foreskin fibroblasts and hTERT-expressing BJ clones BJ-5te and BJ-6te (1) were grown in DMEM plus 10% FCS. Selection of hTERT-expressing lines was performed using 10 μg/ml hygromycin B.
In Vitro Ribozyme Construction and Testing.
Hammerhead ribozymes were designed against four GUC sequences in the hTERT mRNA in the region encoding the T-motif (20) . Using seven antisense nucleotides on each side of the catalytic core, they were designed to cleave 3′ of the Cs at positions 1714, 1717, 1731, and 1744 (GenBank accession no. AF015950). The ribozymes for in vitro testing were obtained as synthetic ribooligonucleotides (MWG Biotech, Ebersberg, Germany) with the following sequences: R1, 5′-GCUCGACCUGAUGAGUCCGUGAGGACGAAACGUACACA-3′; R2, 5′-GCAGCUCCUGAUGAGUCCGUGAGGACGAAACGACGUAC-3′; R3, 5′-AAAGAAACUGAUGAGUCCGUGAGGACGAAACCUGAGCA-3′; and R4, 5′-UCUCCGUCUGAUGAGUCCGUGAGGACGAAACAUAAAAG-3′.
For production of the ribozyme substrate used in cell-free assays, we used the plasmid pGRN121 containing the cDNA of hTERT (21) . A 224-nucleotide fragment of the cDNA including the T-motif region was linked to the T7 promoter sequence via PCR using the primers: P+, 5′-TAATACGACTCACTATAGGGAGAGCACCG-3′; and P−, 5′-CGTCTGCAGCTTCCGACAGCT-3′.
The PCR fragment was transcribed in vitro using 200 ng of DNA, 40 units of RNAsin, 10 nmol of each of the nucleotide triphosphates, 0.37 MBq P-32-UTP, 40 units of T7 RNA polymerase (Boehringer), and 5 μl of T7 transcription buffer (Boehringer) in a volume of 50 μl for 1.5 h at 37°C. The product was separated on a 8% PAGE containing 7 m urea, eluted in 0.5 m NH4 acetate, 0.1% SDS, and 1 mm EDTA overnight at 37°C, and precipitated twice in liquid nitrogen.
In vitro cleavage conditions were optimized with respect to temperature, magnesium concentration, and ribozyme:substrate ratio as described (22) , and labeled substrate (0.5 pmol) was coincubated with 1 pmol of ribozyme at 42°C in a buffer containing 40 mm Tris (pH 8), 12 mm MgCl2. As control, the substrate was incubated in the same buffer without ribozyme. Incubation was done in parallel experiments for either 1 or 3 h. RNA was separated on a 8% PAGE containing 7 m urea and analyzed with a phosphorimager (Bio-Rad Laboratories, Hercules, CA). Cleavage activity was estimated as the ratio of signal intensities in the two cleaved fragments divided by the sum of intensities in the cleaved plus uncleaved fragments.
Plasmid Construction for Ribozyme R4.
Double-stranded synthetic DNA oligonucleotides encoding R4 were obtained from the single-stranded oligonucleotides (BioTez Berlin): R4-A, 5′-AATTCTCTCCGTCTGATGAGTCCGTGAGGACGAAACATAAAAGG-3′; and R4-B, 5′-GATCCCTTTTATGTTTCGTCCTCACGGACTCATCAGACGGAGAG-3′.
To create a negative control plasmid encoding the catalytically inactive mutR4, nucleotides G5 in helix I and A14 in helix III were mutated: mutR4-A, 5′-AATTCTCTCCGTCTTATGAGTCCGTGAGGACGACACATAAAAGG-3′; and mutR4-B, 5′-GATCCCTTTTATGTGTCGTCCTCACGGACTCATAAGACGGAGAG-3′.
Annealing of complementary fragments resulted in DNA oligonucleotides with EcoRI and BamHI protruding ends. Oligonucleotides were cloned between the EcoRI and BamHI sites of the pCDNA3.1(−) vector (Invitrogen). The successful insertion of the ribozyme was verified by sequencing.
The ribozyme containing expression vector pCDNA3.1(R4) was stably transfected into HBL-100 and MCF-7 cells by electroporation using a Gene Zapper 450/2500 (Kodak). Forty μg DNA were used to transfect 10 million cells at 250 V and 950 kOhm in a 0.4-cm (1-ml) cuvette. Clones were selected under 0.7 μg/μl G418.
Construction of a Recombinant Adenovirus Expressing Ribozyme R4.
Vector plasmids pHVad1 and pHVad2 were constructed as follows. pHVad1 has the Ad 3 type 5 genome with a deletion in the E3 region (bp 28133–30818) originating from plasmid pBHG10 (23) . pHVad1 was generated in Escherichia coli BJ5183 by homologous recombination of the SalI-linearized plasmid pTG3602 (complete Ad type 5 genome flanked by a PacI site; Ref. 24 ) with a 8741-bp HpaI/NotI fragment of pBHG10. The unique PacI site in pBHG10 was deleted prior to cloning. The ClaI site in pHVad1 was changed into a meganuclease site. The shuttle plasmid pHVad2 has the left end of the Ad type 5 genome (bp 1–341 and bp 3524–5790) with an E1 deletion and a multiple cloning site, which were derived from plasmid pΔE1sp1A (23) . pHVad2 was generated in E. coli BJ5183 by homologous recombination of the SalI-linearized plasmid pTG9530 (24) with a SgrA1/BstEII fragment of pΔE1sp1A. pHVad2-R4 was constructed by insertion of a 977-bp SalI PCR fragment of pCDNA3.1(R4) comprising the R4 expression cassette into the SalI site of pHVad2. The plasmid pHVad-R4 for rescue of Ad-R4 virus was generated by homologous recombination in E. coli BJ5183 after cotransformation of the meganuclease-linearized vector backbone of pHVad1 with a PacI/BstEII fragment of pHVad2-R4. pHVad-R4 was amplified by transformation into E. coli HB101.
Virus rescue and preparation was performed in 293 cells as described (25 , 26) . The virus suspension was subjected to CsCl density gradient centrifugation (27) . CsCl was removed by dialysis, and virus aliquots were stored at −80°C in storage buffer containing 100 mm NaCl, 10 mm Tris (pH 7.4), 0.1% BSA, and 50% glycerol. The total number of viral particles was determined spectrophotometrically, and infectious units (pfu) were determined by plaque assays on 293 cells.
To estimate the suitable multiplicity of infection, HBL-100 cells and BJ-5te clones were infected with the bacterial β-galactosidase expressing recombinant human Ad Ad-lacZ (28) . Cytotoxicity was still not evident after infection with 150 pfu/cell in HBL-100 or 500 pfu/cell in BJ-5te, respectively, but >90% of the cells showed strong expression of the transgene.
Characterization of Treated Cells.
hTERT mRNA abundance was measured by either RNase protection assay (RPA III; Ambion, Austin, TX) or by real-time RT-PCR (Lightcycler hTERT Quantification kit; Roche Molecular Biochemicals, Mannheim, Germany). The hTERT T-motif was used as probe for RNase protection, and glyceraldehyde-3-phosphate dehydrogenase was used as control. Real-time RT-PCR was performed according to the recommendations of the manufacturer using porphobilinogen deaminase as control.
Telomerase assays used the semiquantitative TRAP assay (Intergen). Lysates equivalent to 2000 cells were analyzed in a telomerase reaction for 20 min, followed by 27 PCR cycles (30 s 95°C, 45 s 60°C). Gels were scanned in a phosphorimager (Bio-Rad). Relative telomerase activities were obtained in comparison with the signal from parental cells, which was always measured on the same gel. Between two and five parallel measurements were performed from each clone.
Telomere length was measured as described (14 , 29) . Briefly, DNA was embedded in agarose plugs, and plugs were digested with proteinase K. HinfI-restricted DNA was run in a 1% agarose gel in a CHEFIII pulsed field gel apparatus (Bio-Rad) for 15 h at 3 V/cm. After Southern blotting, hybridization was done with a 18-mer telomeric probe directly coupled to alkaline phosphatase (Promega). Blot lanes were scanned in an imaging densitometer, and the average telomere length was calculated as the weighted mean (2) .
The frequency of apoptotic cells within R4-transfected clones was estimated by either TUNEL assay (In situ Cell Death Detection kit; Boehringer Mannheim, Mannheim, Germany) by counting at least 1000 cells in 10 different randomly chosen fields of view per sample or by flow cytometry (PAS; Partec GmbH, Münster, Germany). For flow cytometric assessment of apoptosis, 105 cells were stained with 4′,6-diamidino-2-phenylindole and the intensity of the sub-G1 peak was measured after UV excitation. In Ad-R4-infected HBL-100 cultures, the small, granular apoptotic cells were discriminated from viable cells by their lower forward and higher sideward light scatter (30) . Cells (105–106) were stained with propidium iodide to exclude cells with damaged plasma membranes, and forward/sideward scattergrams of the gated, propidium iodide-negative cells were taken using blue light excitation in the flow cytometer.
Cell survival under treatment with cytotoxic drugs was measured by XTT assay (Cell Proliferation kit; Boehringer Mannheim). Cells (2 × 103) were seeded/well of a microtiter plate and were treated for 4 days with the cytotoxic agent before the XTT reaction was carried out and quantitated in an ELISA plate reader. The LD50 was calculated as the dose that resulted in a 50% reduction of the absorbance at 490 nm.
Ribozyme R4 Cleaves the hTERT T-Motif in Vitro and Decreases hTERT mRNA Abundance in Vivo.
The telomerase catalytic subunit differs from other reverse transcriptases by the presence of a conserved motif, the T-motif (20) , which is essential for catalytic activity (21) . This motif contains five GTC sequences, and we designed hammerhead ribozymes against four of them. The cleavage activity of the ribozymes was tested in an in vitro assay using a 224-nucleotide substrate in a 1:2 ratio to the ribozymes (Fig. 1A) ⇓ . Ribozymes R1 and R2 displayed no significant catalytic activity, even under optimized conditions. However, ribozyme R3 cleaved ∼30% of the substrate within 3 h, and ribozyme R4 displayed even higher cleavage activity in vitro. About 40% of the substrate was cleaved within 1 h, and >70% was fragmented after 3 h incubation with the ribozyme R4. The obtained fragment sizes were concordant with the expected ones. If R4 was mutated in its catalytic core to produce the catalytically inactive mutR4, no cleavage activity was seen (not shown).
To test the ribozyme R4 cleavage in vivo, hTERT-overexpressing human BJ-5te foreskin fibroblasts (1) were infected with 500 pfu Ad-R4, and the abundance of hTERT mRNA was measured by RNase protection assay. BJ-5te fibroblasts express the hTERT transgene under the control of the MPSV promoter, resulting in strong telomerase activity, telomere elongation, and life span extension (1) . The expression level of hTERT in these cells is so high that it can easily be detected by RNase protection. Fig. 1B ⇓ demonstrates a transient attenuation of the hTERT mRNA abundance after infection with the R4-expressing adenovirus. Four days after infection, hTERT mRNA levels decreased to <10% of the control level but returned to normal with further dilution of the viral DNA at 8 days after infection. Ad-LacZ infection had no influence on hTERT levels (not shown). It should be mentioned that hTERT levels in BJ-5te cells at 4 days after infection are still much higher than those found in HBL-100 or MCF-7 cells under control conditions without ribozyme treatment (data not shown). Accordingly, Ad-R4 infection did not decrease telomerase activity or telomere length and did not inhibit growth in BJ-5te fibroblasts.
Degradation of hTERT mRNA by Stable Expression of R4 Attenuates Telomerase Activity and Shortens Telomeres in Breast Epithelial Cell Clones.
R4 and mutR4 were cloned in pCDNA3.1 under the control of the cytomegalovirus promoter and transfected into the human breast cancer cell line MCF-7 and in HBL-100, a human immortal breast cell line with high intrinsic telomerase activity. After selection for stable transfection, 28 MCF-7-R4 clones, 20 HBL-100-R4 clones, and 10 each of the mutR4 clones were analyzed. All clones expressed the ribozyme as established by RT-PCR, followed by Southern blotting (Fig. 2A) ⇓ . Quantitative differences between clones could not be detected reproducibly with this technique. The average telomerase activity of all HBL-100-mutR4 clones was 91 ± 22% (mean ± SD) of the parental cells, and that of MCF-7- mutR4 clones was 82 ± 24%. These values are not significantly below those in the parental cells, indicating that antisense effects do not contribute to the action of the ribozyme. On the contrary, telomerase activity was significantly reduced in HBL-100-R4 clones (Fig. 2B) ⇓ and MCF-7-R4 clones (Fig. 2C) ⇓ . Average activities from all analyzed clones were 38 ± 33% (HBL-100-R4) and 54 ± 38% (MCF-7-R4) of the parental cells. Telomerase activity correlated well with the remaining hTERT mRNA expression level as measured by real-time RT-PCR in HBL-100 cells; the hTERT mRNA level in clones 5 and 7, which were <50% of the telomerase activity in parental cells, was about one-third of that in the parental line, whereas both telomerase activity and hTERT mRNA content were undetectable in clone 3, for instance (Fig. 2B) ⇓ . hTERT mRNA levels were lower in MCF-7 cells, amounting to an average of 27.5 ± 5.5% of that measured in K562 cells in 3 mutR4 clones and to 11.2 ± 2.2% in 8 R4 clones.
Clones were classified by the average level of telomerase inhibition as estimated in triplicate experiments. Group I was most strongly inhibited (TRAP activity <25% of parental cells) and had 9 MCF-7-R4 and 6 HBL-100-R4 clones; group II (50–75% inhibition) had 10 MCF-7-R4 and 11 HBL-100-R4 clones; and group III (<50% inhibition) had 9 MCF-7-R4 and 3 HBL-100-R4 clones. In fact, telomerase activity in most of the group I clones was <10% of that in parental cells. No mutR4 clone with a telomerase activity corresponding to group I or II was seen.
After clonal expansion (corresponding to a minimum of ∼20 population doublings), telomeres in group I clones were shorter than in mutR4-transfected clones (Fig. 2D) ⇓ . The average telomere length in the parental MCF-7 and in MCF-7-mutR4 clones used in our laboratory was only ∼2.3 kbp, much shorter as reported by others (31) . This restricted the possible range of shortening greatly. Still, the average telomeric restriction fragment length was significantly shortened in group I clones from both HBL-100 and MCF-7 cells (Fig. 2E) ⇓ . It should be taken into account that telomere length could not been measured in three of the most strongly inhibited MCF-7 clones, because inhibition of growth occurred in these clones before enough cells for a telomere Southern blot could be obtained.
Clones with Attenuated Telomerase Activity Show Slow Net Growth and Apoptosis.
After clonal expansion, net growth rate was measured in all group I clones as well as in arbitrarily chosen group II, III, and mutR4 clones (Fig. 3A) ⇓ . Strong attenuation of telomerase activity (group I) was associated with a significant reduction of the net growth rate in both MCF-7 and HBL-100 clones. Five of nine group I MCF-7 clones and two of six HBL-100 clones were unable to reach confluence within 1 month. This was accompanied by morphological signs of senescence and/or crisis, i.e., an abundance of enlarged, often multinuclear cells as well as rounded, apparently apoptotic cells detaching from the culture dish. Group I MCF-7-R4 clones especially had greatly enlarged, flattened cells reminiscent of senescence at the outer rim of colonies, indicating the exhaustion of replicative capacity within a few population doublings. A significant increase in the percentage of apoptotic cells in MCF-7-R4 group I clones after expansion was confirmed by both TUNEL staining and by measurement of the frequency of sub-G1 nuclei by flow cytometry (Fig. 3B) ⇓ .
Infection of HBL-100 Cells with Ad-R4 Attenuates Telomerase Activity and Induces Apoptosis.
Only small numbers of clones could be generated by stable transfection with the R4 expression vector, especially in HBL-100. To exclude the influence of clone selection artifacts, the ribozyme expression cassette was cloned in an adenoviral vector, and mass cultures of HBL-100 cells were infected with 150 pfu of Ad-R4 or Ad-lacZ, respectively. This titer resulted in >90% infection of HBL-100 cells without any recognizable cytotoxic effect attributable to the viral infection itself. Ad-LacZ infection reduced the telomerase activity of the cultures slightly, as did Ad-R4 infection at day 1 after the treatment. However, a significant and specific inhibition of telomerase occurred in the ribozyme-infected cultures between days 4 and 9 after infection, followed by a recovery to the parental levels at ∼2 weeks after a single infection (Fig. 4A) ⇓ . The maximum inhibition of telomerase activity occurred at day 9 after infection with activities well below 10% of that in parental cells. At the same day the average hTERT mRNA abundance/cell was decreased to 48% of the control (Fig. 4A) ⇓ . The measurement was performed by real-time RT-PCR using exclusively RNA from well adherent, apparently viable cells. In the same cell population, no significant shortening of telomeres could be measured over a period of 2 weeks. Both the average telomere length and the size distribution of terminal restriction fragments remained constant within this period (not shown). However, infection with the recombinant virus resulted in a significant diminution of the net growth rate between about 3 and 14 days after infection, together with the appearance of both large, multinuclear, senescence-like cells and of cells with an apoptotic morphology (Fig. 4B) ⇓ . This reduction of net growth is to a large extent attributable to the induction of apoptosis within the same period of time (Fig. 4C) ⇓ . The reversion of the phenotype at ∼2 weeks after infection is probably caused by dilution and loss of the adenoviral vector attributable to residual turnover of the nonapoptotic cells.
Attenuation of Telomerase Sensitizes Mammary Epithelial Cells to Doxorubicin.
Although inhibition of telomerase is a strong inducer of growth arrest and apoptosis in the breast epithelial cells examined here as well as in other cellular models, especially those with short telomeres (5 , 6 , 9) , it is obvious that this approach alone cannot completely reverse or inhibit tumor cell growth in every case. Therapeutic opportunities would be greatly improved if a synergism between telomerase inhibition and established antitumor strategies could be shown. It was demonstrated before that telomeres are preferentially vulnerable to DNA-damaging treatments (14, 15, 16, 17 , 29) . Inhibition of telomerase was shown to increase the sensitivity of different tumor cells to at least one DNA-damaging drug, cisplatin (4 , 32) . Therefore, we measured whether the sensitivity of MCF-7 and HBL-100 cells to the two cytotoxic drugs, cisplatin and doxorubicin, might be modified by inhibition of telomerase.
Doxorubicin is an intercalating dye, inhibits topoisomerases, alkylates DNA, generates ROS, and can lead to DNA double-stranded breaks (33) . It is widely used for adjuvant chemotherapy of breast cancer. Cisplatin binds covalently to DNA bases, especially G, and cross-links them (34) .
An increase in sensitivity to doxorubicin with decreased telomerase activity was found for both HBL-100 and MCF-7 clones as measured in an XTT assay. In group I clones, the LD50 for doxorubicin is ∼50% (HBL-100) to 25% (MCF-7) of that necessary to kill parental cells or mutR4-transfected clones (Fig. 5A) ⇓ . Apoptosis accompanied doxorubicin-induced cell killing as measured by DNA flow cytometry (Fig. 5B) ⇓ . Even a reduction of telomerase activity to 25–50% of controls (group II clones) resulted in a decrease in doxorubicin LD50, which was significant for MCF-7 (Fig. 5A) ⇓ . On the other hand, the LD50 for cisplatin was not significantly different among parental cells, mutR4, and group II and I clones of either HBL-100 or MCF-7 (not shown). Neither cisplatin nor doxorubicin treatment caused any measurable loss of telomeric DNA in MCF-7 or HBL-100 cells, which were still adherent after 1 week of treatment (data not shown).
Within each group, there was no correlation between proliferation rate and LD50 of the clones. In previous experiments (35) , telomerase activity in MCF-7 cells has been inhibited by transfection with an antisense construct (3) directed against bases 1–185 of hTR. This transfection reduced telomerase activity levels to values <25% of controls but did neither induce telomere shortening nor did it attenuate the growth rate (35) . LD50s for doxorubicin were reduced from >150 ng/ml in vector-transfected controls to 41 ± 3, 43 ± 3, and 79 ± 8 ng/ml in the antisense clones tested, in good agreement with the data given in Fig. 5A ⇓ .
Expression of hTERT Decreases the Sensitivity of Human Fibroblasts to Doxorubicin and Other Inhibitors of Topoisomerase II.
The facts that the LD50 for doxorubicin, but not for cisplatin, was decreased in telomerase-inhibited clones and that the sensitivity to doxorubicin did not depend on the net growth rate suggested a certain degree of specificity for the correlation between telomerase activity and sensitivity to doxorubicin. To test this suggestion, we compared the LD50 for this and a number of other drugs in hTERT immortalized skin fibroblasts to that in parental BJ fibroblasts. Although parental BJ fibroblasts are devoid of any telomerase activity, expression of the hTERT transgene in BJ-5te and BJ-6te clones is sufficient to generate strong telomerase activity (1) . BJ-5te and BJ-6te clones maintain unlimited growth, despite normal checkpoint control (36) . Parental cells and hTERT-expressing clones are similar in terms of average telomere length (about 10–12 versus 7–8 kb) and are equal with respect to growth rate. There is no significant number of apoptotic cells in either line (data not shown).
Telomerase-positive BJ-5te and BJ-6te clones and telomerase-negative parental BJ mass cultures were subjected to different DNA-damaging drugs, and their LD50 was measured (Fig. 5C) ⇓ . Bleomycin generates DNA strand breaks via site-specific formation of metal-bound oxyl radicals (37) . Hydrogen peroxide generates ROS both extra- and intracellularly. Both mitoxantrone and etoposide inhibit topoisomerase II (38) . In addition, mitoxantrone inhibits RNA and DNA synthesis, whereas etoposide produces ROS. Fig. 5C ⇓ shows that the LD50 for doxorubicin was increased 2-fold in telomerase-positive BJ fibroblasts, whereas the sensitivity to the guanine-cross-linking drug, cisplatin, was not dependent on hTERT expression, in concordance with the data obtained using breast epithelial cells. Moreover, telomerase-expressing fibroblasts display a decreased sensitivity to mitoxantrone and etoposide, which also inhibit topoisomerase II. The sensitivity to peroxidation (H2O2) or to bleomycin-generated ROS is not dependent on telomerase activity. Together, these data suggest that expression of telomerase in the cell lines tested selectively decreases the sensitivity to drugs that inhibit topoisomerase II.
The present report demonstrates the use of an anti-hTERT ribozyme to inhibit telomerase activity in immortal human breast epithelial and tumor cell clones as well as mass cultures. The ribozyme R4 cleaved hTERT mRNA within the T-motif 3′of the C at position 1744 in vitro and diminishes the hTERT mRNA abundance in telomerase-positive human fibroblasts and breast epithelial cells in vivo. Expression of the ribozyme in HBL-100 or MCF-7 cells attenuated significantly the telomerase activity in the transfected clones. This effect was not observed after inactivation of the catalytic core of the ribozyme (mutR4), strongly suggesting specific cleavage activity in the cell culture experiments. A positive correlation was found between the degree of attenuation of telomerase activity, the amount of telomere shortening, the inhibition of net cellular growth, and the induction of apoptosis in the expanded clones. Likewise, the infection of HBL-100 cells in mass culture with an adenovirus expressing ribozyme R4 resulted in diminution of the hTERT mRNA amount, attenuation of telomerase activity, inhibition of net growth, and induction of apoptosis. Although these effects were transient after a single infection, it can be envisaged that multiple infections with the adenovirus, say at weekly intervals, should result in a sustained cytotoxic effect.
In contrast to the situation after clonal expansion, telomeres in HBL-100 mass culture did not detectably shorten after Ad-R4 infection, despite massive induction of apoptosis. This result was unexpected in the light of the idea that inhibition of telomerase triggers cell cycle arrest and apoptosis exclusively via critical shortening of telomeres. We did not measure telomere length in apoptotic cells. However, there must be a significant number of preapoptotic cells among those subjected to telomere length measurements and yet, neither a change in the average telomere length nor a shift of the length distribution of telomeres could be found.
Telomerase has been shown to stabilize short telomeres in yeast (39) and to extend the life span of certain human cells without net telomere lengthening (40) . It was suggested that telomerase might have a function in the capping of telomeres (41) . In addition to structural constraints (42) , active telomerase is one possible factor to physically shield the telomeric G-rich single-stranded overhang. The presence of free G-rich single-stranded telomeric DNA within the nucleus was found sufficient to trigger cell cycle arrest in U87 glioblastoma cells and in human fibroblasts (43) . One might speculate that inhibition of telomerase might increase the probability that at some point in the cell cycle a free telomeric overhang becomes exposed to the nucleoplasm and could trigger cell cycle arrest or apoptosis, depending on the cellular context. Further data will be necessary to establish firmly whether capping of telomeres by telomerase might contribute to clonal stability independently of a net change in telomere length. However, it is interesting to note that although the telomere/telomerase complex seems to be involved in the modification of the sensitivity of cells to topoisomerase poisons, this involvement again does not include a measurable shortening of telomeres.
There is a fast-growing list of enzymes with functions in DNA structure remodeling and DNA repair that have been localized to the telomeres or integrated into telomeric functions. This list includes, for instance, poly(ADP-ribose)polymerases like the telomere-specific poly(ADP-ribose) polymerase, tankyrase (44) , and the common poly(ADP-ribose) polymerase 1 (45) , double-strand break repair proteins Ku80 and DNA-PK (46) , the Rad50/hNRE11/NBS1 complex (47) , or the two members of the RecQ helicase family, WRN (48, 49, 50) and BLM (51) . There are indications that topoisomerases need to be added to this list as well. Both the yeast RecQ helicase, SGS1, and their human homologues WRN and BLM interact with topoisomerases (52, 53, 54, 55) , although it is not known whether these interactions are related to the telomeric functions of the helicases. Topoisomerase II was found to be associated with telomeric DNA in HeLa cells after treatment with etoposide (56) . Etoposide treatment of human pancreatic cancer cells up-regulated telomerase (57) . We have shown here that telomerase activity and/or abundance modulates the sensitivity of cells to topoisomerase poisons. Inhibition of telomerase translation in breast tumor and immortal cells increases the sensitivity to the topoisomerase inhibitor, doxorubicin, whereas activation of telomerase by expression of an hTERT transgene decreases the sensitivity of human fibroblasts to different topoisomerase inhibitors. It is not clear whether these effects are mediated via changes in the activity of telomerase upon telomere length or whether they are primarily dependent on the concentration of the enzyme as, for instance, protein-protein interactions or capping of telomeres by telomerase. In conclusion, our data do not yet prove but surely do suggest a functional interplay of topoisomerases with telomeres and telomerase. They validate a ribozyme approach for telomerase inhibition therapy in cancer and suggest that telomerase inhibition might advantageously be combined with topisomerase-directed chemotherapy.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 Supported by grants from the Deutsche Forschungsgemeinschaft, the Berliner Krebshilfe e.V., and Verum e.V.
↵2 To whom requests for reprints should be addressed, at Department of Gerontology, Wolfson Research Centre, Newcastle General Hospital, Westgate Road, Newcastle upon Tyne NE4 6BE, United Kingdom. Phone: 44191-256-3310; Fax: 44191-219-5074; E-mail:
↵3 The abbreviations used are: Ad, adenovirus; pfu, plaque-forming unit; RT-PCR, reverse transcription-PCR; TRAP, telomeric repeat amplification protocol; XTT, 2,3-bis[2-methoxy-4-nitro-5-sulfophenyl]-2H-tetrazolium-5-carboxanilide inner salt; ROS, reactive oxygen species.
- Received August 17, 2000.
- Accepted January 29, 2001.
- ©2001 American Association for Cancer Research.