Telomerase can inhibit the recombination-based pathway of telomere maintenance in human cells.

Telomere length can be maintained by telomerase or by a recombination-based pathway. Because individual telomeres in cells using the recombination-based pathway of telomere maintenance appear to periodically become extremely short, cells using this pathway to maintain telomeres may be faced with a continuous state of crisis. We expressed telomerase in a human cell line that uses the recombination-based pathway of telomere maintenance to test whether telomerase would prevent telomeres from becoming critically short and examine the effects that this might have on the recombination-based pathway of telomere maintenance. In these cells, telomerase maintains the length of the shortest telomeres. In some cases, the long heterogeneous telomeres are completely lost, and the cells now permanently contain short telomeres after only 40 population doublings. This corresponds to a telomere reduction rate of 500 base pairs/population doubling, a rate that is much faster than expected for normal telomere shortening but is consistent with the rapid telomere deletion events observed in cells using the recombination-based pathway of telomere maintenance (Murnane, J. P., Sabatier, L., Marder, B. A., and Morgan, W. F. (1994) EMBO J. 13, 4953-4962). We also observed reductions in the fraction of cells containing alternative lengthening of telomere-associated promyelocytic leukemia bodies and extrachromosomal telomere repeats; however, no alterations in the rate of sister chromatid exchange were observed. Our results demonstrate that human cells using the recombination-based pathway of telomere maintenance retain factors required for telomerase to maintain telomeres and that once the telomerase-based pathway of telomere length regulation is engaged, recombination-based elongation of telomeres can be functionally inhibited.

Telomeres shorten due to the inability of lagging C-strand synthesis to replicate the extreme ends of linear DNA (2) as well as nuclease processing of the parental C-strand following leading strand synthesis (3,4). Telomeres decrease in length, ultimately leading to proliferative failure in most dividing nor-mal human cells (5)(6)(7). However, in vitro immortalized human cells and most cancer cells escape this growth arrest signal by activating a cellular mechanism of telomere length maintenance (8). Telomere length maintenance is most often performed by the ribonucleoprotein telomerase that adds blocks of telomere repeats onto the ends of chromosomes (9). However, a recombination-based pathway to maintain telomere length often termed the alternative lengthening of telomeres (ALT) 1 pathway can also be found in some human in vitro immortalized and cancer-derived cells (10,11).
Telomerase-independent mechanisms of telomere length maintenance also exist in other organisms. Drosophila and mosquitoes use transposable elements to maintain telomeres (12,13). In mutant yeast lacking telomerase, survivors use one of two independent pathways of telomere length maintenance that require either Rad50 or Rad51 (14,15). Type I survivors contain multiple tandem copies of the YЈ element and very short terminal tracts, whereas Type II survivors (10% of the total) require rad50p and have long and heterogeneous telomeres with terminal tracts of 12 kb or longer (15,16). The Type II survivors are thought to maintain telomere length in a similar fashion as that found in human cells using the recombination-based pathway of telomere maintenance (17). Embryonic stem cells derived from telomerase knockout mice can also activate a telomerase-independent mechanism to maintain telomere length (18). Because telomerase-independent mechanisms of telomere length maintenance exist, it has been suggested that telomerase inhibition therapy for telomerasepositive human cancers may result in resistant cells that have activated an alternative mechanism of telomere length regulation. However, several studies demonstrating that telomerase inhibition leads to cell death in human cancer cells have not reported the emergence of escapees that use the recombinationbased pathway of telomere maintenance (19 -21).
Several characteristics can be used to identify human cells using the recombination-based pathway of telomere length maintenance. First, they have no detectable telomerase activity and lack expression of the catalytic protein component, hTERT, or in some cases, they lack both hTERT and the integral RNA component, hTR (22). Second, cells using the recombination-based pathway to maintain telomeres have very long and heterogeneous telomeres ranging in length from less than 2 kb to 50 kb (1). Third, they contain extrachromosomal telomere repeats (ECTRs) that may be linear double-stranded fragments of telomeric DNA (23). Fourth, cells using the recombination-based pathway to maintain telomeres have a novel type of promyelocytic leukemia (PML) nuclear bodies called ALT-associated PML bodies (AA-PBs) that contain PML protein, TRF1, TRF2, replication factor A, Rad51, and Rad52 (24). It is not clear how characteristics associated with the recombination-based pathway of telomere maintenance are generated, and the molecular mechanisms involved also remain unclear.
Various hypotheses for the mechanism that causes this apparently dysregulated telomere function in cells using the recombination-based pathway of telomere maintenance have been made (1). Telomere elongation could occur in a cell using the ALT pathway when telomere-binding proteins can no longer associate with the telomere, at which time recombination enzymes may cause the observed rapid elongation. It has also been proposed that cells using the recombination-based pathway of telomere maintenance may contain abnormal telomere end-binding proteins that under certain conditions cause rapid loss of telomere repeat sequences. It appears as though cells that use telomerase to maintain telomeres repress the recombination-based pathway of telomere maintenance (19 -21), and cells that use the recombination-based pathway of telomere maintenance repress the expression of components of telomerase (22). At present, there is no experimental evidence demonstrating that ALT and telomerase pathways co-exist under nonexperimental circumstances.
Murnane et al. (1) have carefully followed telomere dynamics at a single telomere marked by a plasmid-induced chromosomehealing event. They showed a general pattern of progressive telomere shortening at about 50 bp/division until a critical short size was reached, at which time rapid telomere elongation events occurred, producing subclones with widely different telomere lengths. This observation has several important implications. It suggests that telomeres in cells using the recombination-based pathway of telomere maintenance are not continuously undergoing lengthening events, but rather that it is a signal generated by a "too short telomere" that initiates a recombination-based process. In addition, it suggests that these cells should be in a frequent state of crisis due to a "too short telomere." In the extreme case where one ignores deletion events producing rapid shortening, it would take a maximum of 1000 population doublings for a 50-kb telomere to lose all its repeats at 50 bp/division. Most cells using the recombination-based pathway of telomere maintenance have 140 -180 telomeres (70 -90 chromosomes). If the telomeres are not synchronized into a coordinated shortening/ lengthening cycle, then one of their telomeres should be critically short about every 6 or 7 divisions (1000 divisions/150 telomeres ϭ 1 short telomere/6.7 divisions). Given that deletion events greatly increase the rate of shortening (see "Results"), cells using the recombination-based pathway of telomere length maintenance are likely to be in a relatively continuous state of crisis, in which one or a few of their telomeres are critically short and either recombining and regenerating telomeric sequences or fusing and producing chromosomal abnormalities.
In the present study, we hypothesized that expressing an exogenous telomerase in cells using the recombination-based pathway of telomere maintenance might prevent telomeres from becoming critically short and that telomerase should thus prevent both the initiation of recombination-based elongation events and ongoing chromosomal abnormalities. We tested these predictions by expressing hTR and hTERT in the SV40 large T antigen immortalized normal human lung fibroblast cell line VA13, which uses the recombination-based pathway of telomere maintenance. Some clones of telomerase-expressing ALT cells exhibited homogeneous short telomeres that persisted during longterm passage. This change in telomere length represents a rapid telomere length reduction of about 500 bp/population doubling that is 10-fold faster than expected for the rates of shortening observed in normal telomerase-negative cells. In addition, characteristics associated with the ALT pathway are specifically diminished. Our results indicate that once telomerase is controlling telomere maintenance, the recombination-based pathway of telomere maintenance can be functionally inhibited.

EXPERIMENTAL PROCEDURES
TRF and TRAP Analysis-Telomere length was determined using TRF analysis. First, cell pellets were suspended in 100 mM NaCl, 100 mM EDTA (pH 8.0), and 10 mM Tris (pH 8.0), using 30 l/10 6 cells. Triton X-100 (1% final concentration) and proteinase K (2 mg/ml final concentration) were added. After digestion at 55°C for 12 h and proteinase K inactivation at 70°C for 30 min, the samples were dialyzed overnight against 10 mM Tris, pH 8.0, and 0.1 mM EDTA. After dialysis, 1 g of each sample was digested with a mixture of six enzymes (AluI, CfoI, HaeI, HinfI, MspI, and RsaI) and run on a 0.7% agarose gel overnight at 70 V. The gel was denatured for 20 min in 0.5 M NaOH and 1.5 M NaCl, rinsed for 10 min in distilled water, dried for 1 h at 55°C, neutralized in 1.5 M NaCl and 0.5 M Tris, and hybridized with 32 P-labeled oligonucleotide (TTAGGG) 4 . After washing once with 2ϫ SSC (1ϫ ϭ 0.15 M sodium chloride and 0.015 M sodium citrate, pH 7.0) for 15 min at room temperature and twice (10 min each time) in 0.1ϫ SSC ϩ 0.1% SDS, the gel was exposed to a Phosphor screen and analyzed using a Storm 860 Phosphor-Imager (Molecular Dynamics). The mean TRF length was calculated using the TEL-ORUN program as described previously (25).
The TRAP reactions were performed according to the manufacturer's recommendations (Intergen).
Cell Growth and Manipulation-Cells were grown at 37°C in a 4:1 mixture of Dulbecco's modified Eagle's medium:Medium 199 (Life Technologies, Inc.) plus 10% cosmic calf serum (HyClone Laboratories). To make retroviral vector particles, 30 g of vector plasmid (pBabePuro-hTERT) containing the cDNA of the human telomerase catalytic protein component was transfected into the packaging cell line PhoenixE using FUGENE6 (Roche). Two days later, the supernatants from these cells were used to infect the amphotropic packaging cell line PA317. After 10 days of puromycin (3 g/ml) selection, the supernatants of the PA317 cells were filtered and used to infect VA13 cells. After the infection, cells were selected using puromycin (750 ng/ml) for 5 days. VA13 cells were then transfected with a plasmid that allows expression of human telomerase RNA (pU1hTR), which has been described previously (26). After transfection and selection, the VA13 cell population that was now telomerase positive was plated at clonal density. Clones were picked, expanded, and maintained in continuous culture.
Fluorescence in Situ Hybridization and Immunofluorescence-VA13 cells and their clones were incubated with Colcemid (Life Technologies, Inc.) for 12 h and then harvested using trypsin. Cells were then treated with hypotonic KCl buffer (0.075 M) for 25 min at 37°C, fixed, washed several times with methanol/acetic acid (3:1), suspended, and dropped onto slides. After washing with PBS for 15 min, slides were fixed in formaldehyde (4% in PBS) for 2 min, washed three times with PBS (5 min each time), and treated with 1 mg/ml pepsin for 8 min at 37°C. Formaldehyde fixation and washing steps were repeated an additional time. The slides were dehydrated by a 2-min incubation in 70% ethanol, a 2-min incubation in 90% ethanol, and a 2-min incubation in 100% ethanol and air-dried. The slides were incubated with a hybridization mixture (20 l) containing 70% formamide, 3Ј-Cy3-conjugated (CCCTAA) 3 2Ј-deoxyoligonucleotide N3Ј-P5Ј phosphoramidate probe (27), 0.25% (w/v) blocking reagent (1096 176; Roche Molecular Biochemicals), and 5% MgCl 2 in 10 mM Tris (pH 7.2) for 3 min at 78°C. The slides were then incubated for 3 h at room temperature and washed with 70% formamide, 0.1% bovine serum albumin, and 10 mM Tris (pH 7.2) four times (15 min each time). The slides were then washed with 0.15 M NaCl, 0.05% Tween-20, and 0.05 M Tris four times (5 min each time); dehydrated by a 2-min incubation in 70% ethanol, a 2-min incubation in 90% ethanol, and a 2-min incubation in 100% ethanol; air dried in the dark; mounted with Vectashield containing DAPI (Vector Laboratories); and imaged using a Zeiss Axioplan 2 microscope.
For fluorescence labeling and identification of AA-PBs, 50,000 parental VA13 cells and clones 1 and 2 expressing both hTR and hTERT were grown on coverslips in a 12-well tissue culture dish. After 1-2 days in culture, the cells were washed twice with 1ϫ PBS, fixed with 4% paraformaldehyde/PBS for 5 min, and washed three times (5 min each time) with PBS. Cells were permeabilized by exposure to 0.1% Triton X-100/PBS for 5 min and washed three times (5 min each time) with PBS. The cells were blocked with a solution of 3% bovine serum albumin/PBS for 1 h and then incubated with mouse anti-PML and rabbit anti-hTRF2 for 1 h. The cells were then washed briefly with PBS three times and incubated with secondary antibodies (fluorescein-conjugated goat antimouse IgG and rhodamine-conjugated donkey anti-rabbit IgG) diluted in PBS for 1 h. The cells were then washed in PBS and mounted with Vectashield containing DAPI (Vector Laboratories). Image acquisition and overlay analysis were done using a Leica confocal microscope.
Sister Chromatid Exchange-Cells treated with 100 M bromodeoxyuridine for 72 h were incubated with Colcemid (15210-040; Life Technologies, Inc.) for 12 h, trypsinized, and treated with hypotonic KCl buffer (0.075 M) for 25 min at 37°C. They were then fixed and washed several times with methanol/acetic acid (3:1) until a clean white cell pellet was obtained. The pellet was suspended in methanol/acetic acid  (VA13hTel clones 1-3). All three VA13hTel clones had ϳ10-fold greater telomerase activity than the reference tumor cell line H1299. B, analysis of telomere lengths using the TRF protocol of VA13 parental cells and VA13hTel clones 1-3 that were run on a 1.0% agarose gel using a Fige Mapper (Bio-Rad). VA13 and VA13 clone 3 expressing telomerase have telomeric DNA that migrates above 19 kb, VA13hTel clone 1 has a mean telomere length of 6 kb, and VA13hTel clone 2 has a mean telomere length of 3 kb. C shows TRF analysis of serial-passaged VA13-hTel clones 1 and 2 run on a 1.0% agarose gel using a Fige Mapper (Bio-Rad) alternating from 180 V forward to 180 V backwards in decreasing intervals of 0.1-0.8 s at 14°C.
ALT (telomerase-negative) cell line to reconstitute telomerase activity. After telomerase was reconstituted, clones were picked and expanded. Although the mixed population of telomerase-expressing VA13 cells continued to have very long telo-meres, we noted a significant increase in the fraction of shorter telomeres. We thus picked individual clones to assess whether this was due to clonal variation in telomere sizes or to a greater fraction of small telomeres within individual clones. Fig. 1   FIG. 3. Chromosome aberrations are reduced in VA13 cells expressing telomerase. A, metaphase spread from a VA13 cell showing multiple chromosomal aberrations. The metaphase spread was stained for telomeres (red) using a Cy3-conjugated (CCCTAA) 3 phosphoramidate probe, for centromeres (aqua) using a probe provided by Geron Corp., and for DNA using DAPI (blue). A detailed protocol is described under "Experimental Procedures." The arrows point to a fused chromosome and chromosomes with internal telomeric repeats. Enlarged chromosomes containing internal telomeres and chromosome fusions are shown to the right. B, metaphase spread of VA13hTel clone 1 showing a homogeneous telomere length and reduced number of chromosome aberrations. The metaphase spread was stained for telomere sequence (red) using a Cy3-conjugated (CCCTAA) 3 phosphoramidate probe and for DNA using DAPI (blue).

FIG. 4. Reduction in ALT-associated characteristics in VA13 cells expressing telomerase. A, VA13 cells and
VA13hTel clones 1-3 were grown on glass slides in a 12-well dish as described under "Experimental Procedures." The cells were then fixed, permeabilized, blocked, and incubated with antibody against TRF1 and TRF2 (we used antibody to both TRF1 and TRF2 to increase the sensitivity of the telomere stain) and PML protein (PG-M3; Santa Cruz Biotechnology). Cells were treated with secondary antibodies, washed, mounted using Vetashield containing DAPI, and analyzed using a Leica confocal microscope. An example of a cell containing AA-PBs is shown in the three panels at the top. Quantitation of the total number of cells containing AA-PBs is listed below A. B, the number of ECTRs is decreased in metaphase spreads of VA13hTel clone 1 and VA13hTel clone 2. Metaphase spreads were stained for DNA using DAPI and stained for telomeres using the Cy3-conjugated (CCCTAA) 3 phosphoramidate probe obtained from Geron Corp. Some extrachromosomal telomeric repeats have white circles drawn around them. Quantification of the frequency of ECTRs in the different cell types is shown below B.
shows that both processes are occurring. Clones of parental telomerase-negative VA13 cells have no increase in shorter telomeres or clones that lose all long telomeric signal (Fig. 1A). However, almost all telomerase-expressing clones exhibited an accumulation of some shorter telomeres, whereas a few clones had completely eliminated long telomeres and only contained short ones (Fig. 1B). Clones that have lost all long telomeric signal are marked with a star. Thus, the changes in telomere length observed in VA13 cells by telomerase expression are specific. The TRAP activity for three VA13 clones expressing telomerase is shown in Fig. 2A. The average telomere length declined after 40 population doublings from 25 kb to 6.2 kb for clone 1 and 2.6 kb for clone 2 (Fig. 2B). This rate of shortening of about 500 bases/population doubling is 10-fold faster than the rate of telomere loss observed for normal human cells. This rapid loss is consistent with the observed rapid deletion of long telomeres that occurs in yeast and human cells using ALT (1,28). Telomeres in VA13 cells expressing human telomerase (hTel) clone 1 and clone 2 became more homogeneous in length compared with the parental VA13 telomerase-negative ALT cells. The distribution pattern of the telomere lengths observed in VA13hTel clones 1 and 2 is similar to the distributions found in most cancer or in vitro immortalized cells using telomerase to maintain telomeres. In contrast, clone 3 had significant amount of telomerase activity but still contained the long telomeric signal. Nevertheless, clone 3 does appear to have an increase in smaller telomere sequence compared with parental VA13 cells, suggesting that telomerase is functioning on chromosome ends in these cells, perhaps in competition with ALT, and may be in the process of converting to a short-telomere phenotype (see Fig. 2B). Because the reformation of long heterogeneous telomeres was not observed (even after 100 population doublings) in clones 1 and 2 (Fig. 2C), we conclude that the recombination-based pathway of telomere maintenance can be functionally inhibited by the telomerase pathway once it fully controls telomere length maintenance.
If telomerase were preventing telomeres from becoming critically short, one might expect a reduction in apoptosis and chromosomal instability. Apoptosis rates were determined by analyzing the presence of nuclear fragmentation by DAPI staining. The fraction of cells containing apoptotic bodies was 18% in the parental VA13 cells, 5% in VA13hTel clone 1, 8% in VA13hTel clone 2, and 6% in VA13hTel clone 3. Because apoptosis can be induced by chromosome fusions, we analyzed the chromosomes of the parental VA13 telomerase-negative cells and of VA13hTel clones 1-3. The parental VA13 cells had multiple chromosomal rearrangements including end-to-end associations, chromosomes with internal chromosomal telolomere repeat DNA, and some chromosome ends with no detectable telomeres (Fig. 3, arrows; data not shown). The telomeres with no detectable telomere signals may represent terminally short telomeres that are substrates for either elongation or end-fusion events. VA13hTel clones 1-3 had no detectable end-to-end associations (see Fig. 3B). Thus, VA13hTel clones 1-3 have decreased rates of apoptosis and greater chromosomal stability resulting from telomerase expression.
We next tested whether VA13hTel clones 1-3 had AA-PBs and ECTRs. The structures containing both PML protein and telomere-binding proteins TRF1 and TRF2 as determined by overlaying composites were scored as AA-PBs. Fig. 4A shows a representative picture of a VA13 cell containing AA-PBs, in which immunofluorescence of PML protein and TRF1 and TRF2 overlap. Although VA13hTel clones 1-3 contained some AA-PBs, clones 1 and 2 showed a 20-fold reduction in the percentage of cells containing AA-PBs as compared with the parental VA13 cells (see Fig. 4A, percentage of cells containing AA-PBs). Fig. 4B shows a representative metaphase spread of a VA13 cell containing ECTR DNA as determined by fluorescence in situ hybridization (white circles). ECTR DNA was abundant in the parental VA13 cells, whereas VA13hTel clones 1-3 had significantly fewer ECTRs (Fig. 4B, see ECTR/metaphase). Clone 2, which has the shortest telomeres, showed a 50-fold reduction in the amount of ECTR DNA. Interestingly, the number of both ECTRs and AA-PBs was greater in clone 3 (which still has large telomeres) as compared with clone 1 or 2. To test whether a non-telomere-specific recombination event was affected in VA13hTel clones 1-3, we analyzed the average number of sister chromatid exchanges (SCEs) that occurred per chromosome. Fig. 5 shows a representative analysis of SCE in the parental VA13 cell. The parental VA13 and VA13hTel clones 1-3 all had an average of 0.6 exchanges/chromosome, which is a typical number for a cell that has been immortalized with SV40 T antigen (Fig. 5) (29). Blooms cells were used as a positive control in Fig. 5. This indicates that although telomerase specifically diminishes AA-PBs and ECTRs, it has not changed the recombination pathway used to exchange sister chromatid. However, because we have only analyzed mitotic recombination by determining the average number of SCEs per chromosome, we cannot assess whether other recombination pathways have been affected by the expression of telomerase.

DISCUSSION
This is the first report in human cells demonstrating that telomerase can functionally inhibit the recombination-based pathway of telomere length maintenance. We demonstrate this by expressing telomerase in a cell line that uses the recombination-based pathway of telomere maintenance and show that telomerase expression can specifically change telomere length and diminish additional characteristics found in cells that use the recombination-based pathway of telomere maintenance. The ability of telomerase to block the recombination-based pathway of telomere length regulation could be achieved in several ways. First, the telomerase ribonucleoprotein may directly sequester proteins required for the initiation of the recombination-based pathway of telomere maintenance. Second, telomerase may simply change the telomere structure. This change in telomere structure could be sufficient to prevent the recombination pathway of telomere maintenance from functioning. Third, telomerase may cap the ends of telomeres and prevent reinitiation of the recombination-based pathway of telomere maintenance (30). We favor the hypothesis that telomerase prevents the telomeres from becoming critically short and inhibits the activation of the rapid elongation events by the recombination machinery. The progressive disappearance of overly long telomeres in turn inhibits the generation of rapid deletion events and the production of ECTRs. Interestingly, the population of ALT cells expressing telomerase can be converted to cells containing short telomeres more rapidly by growing them in 40% oxygen (data not shown). Hypoxia induces telomere breakage (31) and would accelerate the loss of long telomeres. Clone 3, which had long telomeres even after 100 population doublings in 20% oxygen, developed short telomeres when grown in 40% oxygen for only 10 doublings.
ECTRs are reduced upon expression of telomerase in VA13hTel clones 1-3. ECTRs may be created by the cleavage of the long telomeres during rapid deletion events, and such events may not occur in shorter telomeres. This suggests they are a byproduct of the very long telomeres generated by the ALT pathway. ECTRs have been identified in ATMϪ/Ϫ mouse cells, establishing that ECTRs are not uniquely present in cells using the ALT pathway of telomere length regulation (32). Our prediction is that ECTR DNA is produced by telomere instability. We also observed a decrease in the number of AA-PBs in VA13hTel clones 1-3. These structures contain several proteins that could be important for telomere maintenance in the ALT pathway, including TRF1, TRF2, replication factor A, Rad51, and Rad52 (24). These proteins could be involved in forming new telomeres in cells using the recombination-based pathway of telomere maintenance due to their associated functions; however, it is not clear whether their sequestration at PML bodies is indirect or important for the ALT pathway.
The level of telomerase activity in these cells was much greater than that observed in our reference tumor cell line H1299 ( Fig.  2A). In our experience, this high level of activity would normally be sufficient to elongate telomeres to greater than 10 -15 kb in most cell types (33). The construct we used to express hTR produces an RNA in which the integral telomerase RNA is flanked by U1 RNA sequences on both ends. We suspect that these U1 RNA sequences are interfering with the efficiency with which telomerase is recruited to act on the telomeres, so that the maintenance telomere length in clones 1 and 2 is only a few kilobases. Using constructs that produced a telomerase maintenance length that was short rather than long greatly facilitated the ability to distinguish telomeres being maintained by telomerase from the very long ALT telomeres.
We demonstrate that telomerase expression in the VA13 ALT cells leads to increased chromosome stability. In cells using the recombination-based pathway of telomere maintenance and some cancer cells, occasional telomeres can shorten to the point that telomere sequence can no longer be detected (1,34). The frequency of appearance of terminally short telomeres may be one of the factors that influence the growth properties of a particular cancer cell. In the parental VA13 cells that use the recombination-based pathway of telomere maintenance, telomeres appear to undergo fusion-breakage-fusion cycles much more frequently compared with VA13 cells that use telomerase to maintain telomere length ( Fig. 3; data not shown). The constant generation of critically short telomeres could explain the increase in cell death and chromosome abnormalities in these ALT cells. This would suggest that cancer cells using the recombinationbased pathway of telomere length maintenance might not divide as fast as cancer cells using the telomerase-based pathway to maintain telomere length. Because PML bodies are thought to be DNA repair factories, the ability of telomerase to prevent telomeres from becoming critically short may be directly related to the observed reduction in AA-PBs. In conclusion, the VA13 cells that use the recombination-based pathway of telomere maintenance appear to be continuously faced with a state of crisis. The expression of telomerase prevents telomeres from becoming so short that they initiate recombination events. Once telomerase has "trapped" the bulk of the telomeres in a normal size range, the observable characteristics of ALT pathway (heterogeneous telomeres, ECTRs, AA-PBs, and genomic instability) disappear.