|Home | About | Journals | Submit | Contact Us | Français|
Expression of hTERT is the major limiting factor for telomerase activity. We previously showed that methylation of the hTERT promoter is necessary for its transcription and that CTCF can repress hTERT transcription by binding to the first exon. In this study, we used electrophoretic mobility shift assay (EMSA) and chromatin immunoprecipitation (ChIP) to show that CTCF does not bind the methylated first exon of hTERT. Treatment of telomerase-positive cells with 5-azadC led to a strong demethylation of hTERT 5′-regulatory region, reactivation of CTCF binding and downregulation of hTERT. Although complete hTERT promoter methylation was associated with full transcriptional repression, detailed mapping showed that, in telomerase-positive cells, not all the CpG sites were methylated, especially in the promoter region. Using a methylation cassette assay, selective demethylation of 110bp within the core promoter significantly increased hTERT transcriptional activity. This study underlines the dual role of DNA methylation in hTERT transcriptional regulation. In our model, hTERT methylation prevents binding of the CTCF repressor, but partial hypomethylation of the core promoter is necessary for hTERT expression.
Telomeres are nucleoprotein complexes that ‘cap’ and stabilize the termini of linear chromosomes. They are also involved in chromosome replication, maintenance of nuclear architecture, chromosome stability, gene expression, aging and cell division (1,2). In normal cells, each division is associated with the loss of 50–100 base pairs of telomere length. This shortening of the telomeres acts as a mitotic counter and limits life span. Telomerase, a complex consisting of a reverse transcriptase bound to its own RNA template, allows the maintenance of the telomere length (3). In most somatic cells, telomerase activity is not detectable (4). In contrast, telomerase is expressed in highly proliferative cells, such as germ cells and stem cells, and in the cells of about 85% of cancers (5). In vitro, two components are absolutely essential for telomerase activity: the catalytic subunit, hTERT, containing the reverse transcriptase activity, and the RNA component, hTERC, containing a complementary template for the telomeric DNA sequence (TTAGGG) (6,7).
Following the characterization of the genomic sequence of hTERT and the elucidation of the organization of the gene (8–11), many studies have shown that expression of hTERT represents the limiting factor for telomerase activity, and that the regulation of hTERT expression occurs primarily at the transcriptional level. Transient transfection experiments have identified a minimal promoter encompassing the 283bp region upstream of the ATG initiation codon (9–11). The 5′-hTERT regulatory region contains numerous binding sites for transcription factors. Activators of hTERT transcription include c-Myc, Sp1, hALP, Hif-1, Mbi-1, USF1/2 and estrogen response element. Repressors for hTERT have also been identified and include the tumor suppressor protein p53, Mad1, myeloid-specific zinc finger protein 2 (MZF-2), TGF-β, Wilms' Tumor 1 (WT1) and CTCF. (12–20). Recently, Horikawa et al. described an E-box binding sequence located upstream of the transcriptional start site, as inhibitor of hTERT but not in mTERT, suggesting a differencial regulation of TERT in human and mouse (21). We previously showed that 5′ exonic sequences might be involved in the regulation of the hTERT gene through an inhibitory effect on transcriptional activity of the minimal hTERT promoter (22). Moreover, this region inserted immediately downstream of CMV and CDKN2A promoters repressed their activity in normal, immortalized and cancer cells. We observed that the 11-zinc finger factor CTCF binds this region in telomerase-negative cells, but not in telomerase-positive cells (18). Since CTCF represses transcriptional activity when it binds DNA downstream of transcriptional start sites (23,24), these findings suggest that CTCF might function as a transcriptional repressor for hTERT in normal cells. However, the mechanisms that might prevent CTCF from binding to the hTERT gene in telomerase-positive cells have not been determined. Therefore, it is essential to note that the identification of these activators and repressors of hTERT do not take in account the methylation profile of the hTERT promoter described in numerous studies (25–27).
The fact that the hTERT promoter is located within a CpG island suggests that transcription of the gene might be regulated by DNA methylation. Previous studies showed that the hTERT promoter is hypermethylated in most telomerase-positive tumors and hypomethylated in telomerase-negative normal tissues (25–27). These observations contrast with the general association between promoter methylation and gene silencing (28), which prompted us to study how DNA methylation of the hTERT promoter can lead to its expression.
To this end, we initiated a detailed evaluation of the methylation pattern of each CpG site in the areas that appear to contribute to the transcriptional regulation of the hTERT gene. Based on our previous analyses (22), two regions of the hTERT CpG island seemed essential: one is the core promoter that is necessary to its expression; the other is the 5′ exonic region where the CTCF repressor binds. The hTERT region at −441 to −218 from the ATG translational start site examined in our previous studies was located immediately upstream of the regions we considered now as essential (29).
Although the first two exons play a role in the repression of the hTERT transcription by CTCF, which binds to two sites, one in the first exon and the other at the beginning of the second (18); in the present study, only the effect of CTCF on the first exon was investigated. The presence of an intron between these two CTCF sites complicate the analyses since, in transient transfections, splicing of this intron is only partial.
In the present report, we tested the hypothesis that hTERT methylation prevents binding of CTCF inhibitor and that a partial methylation of the hTERT promoter region can result in some level of transcriptional activity. The results have allowed us to identify an unexpected dual role for DNA methylation in the transcriptional regulation of the hTERT promoter.
The human tumor cell lines HeLa (cervical adenocarcinoma), SW480 (colorectal adenocarcinoma) and normal BJ fibroblasts were obtained from the ATCC. All these cells were grown in the medium recommended by the ATCC. The HLF/hTERT cells were kindly provided by Dr. Joachim Lingner (ISREC, Epalinges, Switzerland), and were cultured in DMEM supplemented with 10% heat inactivated fetal bovine serum (Invitrogen, Basel, Switzerland). HeLa and SW480 cell lines are telomerase positive, whereas BJ is telomerase negative. HLF/hTERT cells, obtained through stable transfection of HLF cells with an MSCV-hTERT retrovirus, constitutively express hTERT and have high telomerase activity.
Tumor tissues (breast, bladder and cervix) were obtained from the Tissue Bank of the Institute of Pathology of Lausanne. Microdissection and DNA extraction were performed as described previously (30). Briefly, frozen tissue sections (7-μm thickness) were stained with toluidin blue. Then, normal cells were removed by scratching. Verification that the remaining cells were indeed tumor was performed before the cells were harvested.
For stable transfection, hTERT sequences and firefly luciferase gene were extracted from the pTERT-297/ex2 vector, containing the hTERT minimal promoter and the 1071bp downstream of the ATG (22), by digestion with Asp718I and BamHI. Then, this fragment was cloned into the pcDNA5/FRT vector (Invitrogen), previously deleted of the CMV promoter, in order to create the pTERT-297/ex2/FRT that is used in stable transfection experiments with or without in vitro methylation of all the CpG sites with the SssI methylase (Promega, Madison, WI, USA). Full methylation was confirmed by digestion with MspI (Promega, Madison, WI) and HpaII (Amersham Biosciences Buckinghamshire, England) restriction enzymes.
pCpG-LacZ, a CpG-free vector (InvivoGen, San Diego, CA, USA), was used to study the effect of hTERT methylation on transcriptional activity. hTERT minimal promoter and exon 1 were generated by PCR, using primers containing either the SdaI or the XbaI sites. These fragments were cloned into pCpG-LacZ to produce pCpG-Tm (hTERT promoter) and pCpG-Tmex1 (hTERT promoter and the first 80bp of the exon 1).
ChIP assays were performed as previously described (18). Then, purified DNA was analyzed by PCR with specific primers for the co-amplification of the first exon of the hTERT gene to generate a 171-bp fragment (TERT-FW 5′-CTGCTGCGCACGTGGGAAGCC-3′ and TERT-REV 5′-GTCCCCGCGCTGCACCAGCC-3′) or the H19 gene as a 149-bp control (H19-FW 5′-CACCGCCTGGATGGCACGGAATTG-3′ and H19-REV 5′-TGCGACGCGTGGCTTGGGTGAC-3′). Amplification was performed with the following PCR conditions: 94°C for 30s, 62°C for 45s and 72°C for 60s. The cycle number and the amount of template were varied to ensure that results were within the linear range of the PCR. PCR products were analyzed on 2% agarose gel. ChIP experiments were carried out in triplicate to ensure reproducibility.
DNA was extracted from culture cells and modified with sodium bisulfite, as previously described (31). After bisulfite modification, PCR were performed with primers specifically designed to amplifiy bisulfite-modified DNA sequence of the promoter and first exon of hTERT 5′-CTACCCCTTCACCTTCCAA-3′ and 5′-GTTAGTTTTGGGGTTTTAGG-3′. The amplified region corresponded to nucleotide positions 3791-4105 of the unmodified hTERT gene sequence (GenBank AF097365). Amplification was performed using the master mix (Promega), with the following PCR conditions: 40 cycles of 94°C for 30s, 57°C for 45s and 72°C for 50s. PCR products were cloned into the pGEM-T vector using the pGEM-T vector system II (Promega). DNA extracted from bacterial clones (QIAprep Spin Miniprep Kit, Qiagen) was analyzed by sequencing with the M13 forward primer (5′-GTAAAACGACGGCCAG-3′), using a Big Dye Terminator Cycle Sequencing Kit, and an ABI Prism 3100 sequencer (Applied Biosystems, Foster City, CA).
To verify the methylation status of the hTERT stable transfectant, DNA from cells were extracted with the DNeasy extraction kit (Qiagen) and analyzed by MS-SSCA (Methylation-Sensitive Single Strand Conformation Analysis) (31). The forward Primer used was specific to the plasmid sequence (5′-TGTATGAAGAATTTGTTTAGGGTTA-3′), and the reverse primer was located within the second exon of hTERT (5′-AACCACCAACTCCTTCAAA-3′).
The F1 fragment, which contains part of the first exon of hTERT, was synthesized by PCR using the plasmid pTERT-297/ex2 as template, previously sequenced and published (22) and with the following primers previously used (18): 5′-CCTGCTGCGCACGTGGGAAGCC-3′ and 5′-GGCAGCACCTCGCGGTAGTGG-3′. Purified fragments were verified by sequencing. Five μg were methylated with 2 units/μg SssI methyltransferase in the presence of 180μM S-adenosyl-L-methionine for 16h at 37°C. Following termination of methylation reaction by heating at 65°C for 15min, the methylation status of plasmid constructs was analyzed by digesting overnight with an excess amount of BstUI. EMSA was then performed as previously described (18).
Stable transfection experiments were performed with the Flp-In system, allowing integration of the gene of interest, always at the same site, in the genome of a mammalian cell, following the manufacturers′ recommendations (Invitrogen). A HeLa Flp-In host cell line containing a single FRT site was created with this system. Positive clones were selected using 500μg/ml of Zeocin antibiotic. The vector containing the gene of interest was integrated into the genome via Flp-recombinase-mediated DNA recombination at the FRT site. Integration confers hygromycin resistance and Zeocin sensitivity. Cells were selected with 600μg/ml of Hygromycin antibiotic.
Transient transfection assays were performed with cells seeded at a concentration of either 200000/3.8cm2 for HeLa and SW480, or 50000 cells/3.8cm2 for BJ. Cells were then cultured overnight. Transient transfection of reporter plasmids (0.75μg/well) was carried out in triplicate using JetPEI Cationic Polymer Transfection reagent (4μl/well) (Polyplus-transfection, Illkirch, France). All experiments were performed at least twice and in triplicate. Analysis of LacZ reporter plasmids was performed with a β-galactosidase assay. Briefly, protein was extracted from cells using 500μl of the 5X passive lysis buffer (Promega). In a 96-wellplate, 50μl of cell lysate was analyzed with 100μl of the β -galactosidase substrate buffer (NaH2PO4 0.2M, pH 8, β-mercaptoethanol 0.1M, MgCl2 2mM, ONPG 1.33mg/ml). The kinetics of the β-galactosidase activity was followed with a colorimeter at 414nm during 15min. The kinetics defined a linear curve, and the mean was calculated with values taken at fixed time points. β-galactosidase activity of the different constructs was compared to the level of the pCpG-LacZ vector containing the synthetic LacZΔCpG gene under the control of a mammalian promoter (combination of the CMV enhancer, the human elongation factor 1 alpha core promoter and 5′UTR containing a synthetic intron) and to the level of the pCpG-basic vector, corresponding to that of the pCpG-LacZ vector without its promoter region.
A methylation cassette assay was used to determine the effect of hypomethylation of a specific regulatory region on hTERT transcription. Two restriction sites were introduced by site-directed mutagenesis using the QuickChange XL Site-Directed Mutagenesis Kit (Stratagene, CA, USA) in pCpG constructs. The mut1 primers set (FW-mut1 5′-ACCTTCCAGCTCCGGATCCTCCGCGCGGAC-3′ and REV-mut1 5′-GTCCGCGCGGAGGATCCGGAGCTGGAAGGT-3′) created the upstream site BamHI (indicated in bold) and the mut2 primers set (FW-mut2 5′-CCGCCCTCTCCTCGACGGCGCGAGTTTCAGG-3′ and REV-mut2 5′-CCTGAAACTCGCGCCGTCGAGGAGAGGGCGG-3′) created the downstream site TaqI (indicated in bold). The generated plasmids, pCpG-Tm Δ (containing the hTERT minimal promoter) and pCpG-Tmex1 Δ (containing the hTERT minimal promoter and the first 80bp of exon 1), were methylated in vitro with the SssI methylase following the manufacturers instructions (Promega). The methylated and unmethylated cassette BamHI-TaqI 110bp fragment (−183 to −74bp of the hTERT sequence) were extracted through enzymatic digestion and ligated back into the methylated or unmethylated vector, using the Ligafast Rapid Ligation System (Promega). The efficiency of the ligation reaction was assessed by analysis of an aliquot of the ligation reaction mixture on a 1% agarose gel. The methylated, unmethylated and partially methylated constructs were confirmed by digestion with HpaII restriction enzyme. The remainder of the ligation reaction mixtures was then transfected into HeLa cells. β-galactosidase activities were measured as described.
Treatment with the demethylating agent, 5′-aza-2′-deoxycytidine (5-azadC), was performed on HeLa and SW480 cell lines. Cells were treated 24h after seeding with 3μM of 5-azadC (Flucka, Buchs, Switzerland) in 48-h intervals and for four passages. Cells were then collected and treated for ChIP or used for DNA and RNA extractions (as described below).
RNA was extracted from cells before and after 5-azadC treatment, using the TRIzol LS Reagent (Invitrogen). RNA was then converted to cDNA using random primers and Superscript III reverse transcriptase (Invitrogen, Carlsbad, CA). Real-time PCR analysis was preformed using the TaqMan® Universal PCR Master Mix (Applied Biosystems, Foster city, CA) and using the Applied Biosystems 7900HT Real-Time PCR system. CTCF expression was determined using the following primers and probe: 5′-TGACACAGTCATAGCCCGAAAA-3′ (FW), 5′-TGCCTTGCTCAATATAGGAATGC-3′ (REV) and 6FAM-TGATTTGGGTGTCCACTTGCGAAAGC-MGB (probe). Human Glycerhaldehyde-3-phosphate dehydrogenase (hGAPDH: 99999905) and human Telomerase Reverse transcriptase (hTERT: 00162669) primers/probe mixtures were purchased as Predevelopped Assay (Applied Biosystems, Foster city, CA).
To evaluate binding of CTCF to sequences within the first exon of hTERT, ChIP assays were performed on telomerase-positive tumor cell lines (HeLa, SW480), normal telomerase-negative cells (BJ) and telomerase-positive cells immortalized by hTERT transfection (HLF/hTERT). CTCF binding to the ICR of H19 gene was used as a positive control for the efficiency of the experimental protocol. PCR reactions were performed under conditions in which the negative control samples (no antibody) always showed negligible levels of background amplification (data not shown). ChIP experiments were performed in a way that the positive control, H19, was coamplified with the hTERT exon 1 in a single reaction and PCR fragments were resolved on the same gel. Results show that CTCF bound to the first exon of the endogenous hTERT gene of hTERT-negative BJ and HLF/hTERT cells but not of the tumor cell lines tested (Figure 1A). These results support our previous data (18) and are more convincing because positive controls and experimental fragments were amplified in the same reaction.
As methylation of CpG dinucleotides within the CTCF recognition sequence can influence its binding (32), we examined the methylation status of CpG sites within the hTERT minimal promoter and the first exon sequences from −200 to +100 from the ATG using bisulfite modification, PCR amplification, and subsequent cloning and sequencing of the PCR products. Clones from two tumor cell lines, HeLa and SW480, two normal cell lines, BJ and HLF/hTERT, as well as tumor tissues from breast, bladder and cervical cancer were sequenced and are presented in Figure 2. In tumor tissues and tumor cell lines, almost all CpGs between −100 and +100, including the CTCF binding site (Region B) were methylated (75 to 100% methylated). By comparison, most of the CpGs from approximately −165 to −100 (Region A) were unmethylated (0 to 55% methylated). In BJ and HLF/hTERT, very few CpGs in the entire region were methylated (3 to 23%). These findings suggested that methylation of CpGs in Region B might inhibit CTCF binding while leaving unexplained the selective hypomethylation of Region A.
To test whether methylation might interfere with CTCF binding, we used SssI methylase to modify a PCR fragment encompassing the first exon of the hTERT gene, previously derscribed to be bound by CTCF (18), and compared the native (F1) and modified sequences (F1-met) for their ability to bind CTCF in EMSAs. These studies revealed a marked preference of CTCF for the unmethylated site (Figures 1B and C).
We also tested the effect of CpG methylation on the affinity of CTCF binding in vivo. HeLa cells were stably transfected with an hTERT construct, pTERT –297/ex2/FRT (Figure 1D) that was methylated in vitro before transfection. Genomic DNA of stable transfectants was extracted after 30 population doublings (PD) and the methylation status was analyzed by Methylation-Sensitive Single-Strand Conformation Analysis (MS-SSCA). To differentiate endogenous hTERT gene sequences from the transfected sequence, we used a specific forward primer located in the reporter vector (Figure 1D). The transfected hTERT sequence remained methylated even after 30 PDs, whereas the unmethylated control transfectant stayed unmethylated (Figure 1E). ChIP assays were performed on methylated and unmethylated stable transfectants. CTCF bound to the hTERT exon1 region of the vector sequences only when this region was unmethylated (Figure 1F). Taken together, these results indicated that methylation of the first exon of hTERT prevented binding of CTCF both in vitro and in vivo.
As was shown in previous studies (33–35), treatment of telomerase-positive cells with the demethylating agent 5-azadC inhibits hTERT transcription and reduces telomerase activity. We hypothesize that demethylation of the first exon of hTERT restores CTCF binding. To test this hypothesis, we treated HeLa and SW480 cells with 5-azadC for 4 weeks. Methylation of hTERT as determined by bisulfite sequencing showed that CpG methylation in Region B was reduced from that in untreated cells (54 and 64% methylated after 5azadC treatment versus 93 and 90% methylated in untreated HeLa and SW480 cells, respectively) (Figure 2 and and3A).3A). ChIP analyses revealed significant CTCF binding to hTERT first exon sequences from treated cells (Figure 3B). In addition, real-time PCR analyses of hCTCF mRNA levels showed no significant differences before and after 5azadC treatment. In contrast, the hTERT mRNA level, which is high in HeLa and SW480 before 5azadC treatment, cannot be detected after 5azadC treatment (Figure 3C).
To evaluate the importance of methylation for hTERT activation, HeLa cells were transiently transfected with CpG-free plasmids containing a hTERT promoter that was previously methylated in vitro. The hTERT minimal promoter did not show any activity if all the CpG sites were methylated (Figure 4A). Likewise, no transcriptional activity was observed in a construct with methylated promoter and the first exon of hTERT. These results showed that transcription of hTERT was completely repressed by full methylation of the promoter. Sequencing of bisulfite-modified DNA from telomerase-positive tumor cell lines (HeLa and SW480) and tumor tissues (breast, bladder and cervix) showed partial demethylation of the promoter, between −160 to −80bp (Figure 2, Region A). A recent study has also described partial methylation of the promoter region in telomerase-positive cell lines (35). This hypomethylation, just upstream of the transcriptional start sites, might permit the low transcriptional activity of the hTERT promoter observed in tumor cells (36,37). To test the role of methylation of this region in the regulation of hTERT transcriptional activity, we performed a methylation cassette assay (38). A β-galactosidase reporter plasmid was generated, containing a cassette with sequences from −183 to −73bp, encompassing 12 CpG sites and representing the region A described in Figure 2. The methylated or unmethylated cassette was excised and ligated back into the vector, which was then transfected into HeLa cells. Methylation of the cassette reduced the activity of the hTERT promoter to the background levels, a result similar to that obtained with the fully methylated promoter (Figure 4B). In contrast, the unmethylated cassette, ligated to the methylated vectors, reduced transcriptional activity of the minimal promoter nine-fold compared to the activity obtained with the same unmethylated reporter. When exon 1 is present in the construct, the activity is reduced only two-fold. Moreover, if we compare fold difference between the minimal promoter activity and the minimal promoter + exon1 in cassette methylation experiments, the difference is very significant when the plasmids are unmethylated; but when the exon1 is methylated, there is almost no difference (Figure 4C). These results suggested that the hypomethylation of the region A (Figure 2), which contains three of the four Sp1 binding sites, was permissive for low levels of hTERT expression, comparable to those seen in vivo (36,37), and that the methylation of the exon1 allowed the transcriptional level to be likely the same than with the minimal promoter only. These results also suggested that the regions surrounding the region A in the minimal promoter might be sites for a strong activator that might have binding sites sensitive to methylation.
The aim of the present study was to define more precisely the role of CTCF and DNA methylation in transcriptional regulation of the hTERT gene. CTCF binds preferentially to GC-rich DNA regions, exerts an inhibitory effect when bound downstream of a transcriptional start site (39) and may be released when its binding site is methylated (32). In a previous study, we showed that CTCF binds to GC-rich regions within exons 1 and 2 of the hTERT gene, irrespective of the cell type or the promoter used. In the hTERT CpG island, binding of CTCF might be influenced by the methylation status of its sites. In the present study, ChIP assays revealed that CTCF binds to the first exon of hTERT when the hTERT CpG island is not methylated. In contrast, CTCF no longer binds its recognition sequence when this site is methylated. Hypermethylation of its binding site can therefore abolish CTCF repressor activity. This was confirmed by inducing demethylation with 5-azadC, which allowed CTCF to bind to the first exon region and repress hTERT expression. Therefore, the main role of hTERT methylation in tumor cells is probably to prevent binding of the CTCF repressor and, as a consequence, to allow transcription of the hTERT gene.
The observation that 5-azadC treatment results in hTERT gene repression was confirmed by other groups (33–35). Nevertheless, these results are apparently in marked contrast to two other reports showing activation of hTERT following treatment of telomerase-negative ALT cells upon 5-azadC (25,26). These seemingly contrasting effects might be explained as a result of the balance between demethylation of the hTERT first exon region allowing binding of CTCF and, possibly, demethylation of genes necessary for hTERT transcription. Indeed, in luciferase reporter gene assays, it was shown that the hTERT promoter is much less active in ALT cells (e.g. U2-os) than in telomerase-positive tumor cell lines (21). Further studies of ALT cells should be undertaken to develop a better understanding of this phenomenon.
Promoter methylation is an epigenetic process most commonly associated with transcriptional repression. Methylation of DNA helps to stabilize chromatin in an inactive configuration, thereby inhibiting gene expression (28). As expected, our results clearly show that methylation of all the CpG sites within the hTERT promoter CpG island resulted in complete transcriptional repression, in spite of the inability of CTCF repressor to bind to the hTERT gene. In this condition, hTERT does not provide a real exception to the general model of gene silencing by promoter methylation. However, even if hTERT is generally hypermethylated in telomerase-positive tumor cells (29), we showed in the present study that a region between −165 to −80bp upstream of the translational start site was hypomethylated in tumor tissues and cell lines. This region contains three of the four Sp1 sites present in the hTERT core promoter. We used cassette methylation assay to demonstrate the importance of this region for the transcriptional activity of the hTERT promoter. A recent study also showed that hTERT transcription might require partial methylation of its promoter (35). However, our data clearly indicate that selective partial demethylation of 12 CpGs limited to the small region upstream the transcription start site, defined as region A, significantly activates hTERT promoter in a reporter plasmid. Nevertheless, the activity of the promoter in these conditions is significantly lower than the activity obtained with the unmethylated minimal promoter. These results suggested that the regions surrounding the minimal promoter region A might be the target of transcriptional factors other than CTCF, whose binding are sensitive to methylation. The low expression level of the hTERT showed with this partial demethylation of the core promoter corresponded rather well to the level of hTERT expression observed in vivo. Indeed, the hTERT mRNA levels detected in telomerase-positive cell lines are very low, 0.2 to 6 copies per cell (36,37). This is in stark contrast with the high level of transcriptional activity obtained in transient transfection of the hTERT core promoter in telomerase-positive cell lines, a level comparable to that induced by strong SV40 early promoter (8).
Methylation cassette experiments shows that the expression levels of the construct with partially methylated hTERT minimal promoter and methylated exon 1 in transient transfection is very close to the levels of the endogenous hTERT in human cancers showing similar methylation pattern. The methylation of the exon1 of hTERT seems to be sufficient to allow the transcription from the minimal promoter, and does no more exert its inhibitory effect as seen in umethylated constructs. Moreover, a CTCF site located at the beginning of exon 2 also plays an important role in the downregulation of the constructs with unmethylated hTERT (18). However, as previously mentioned, the presence of an intron between the two CTCF sites complicates the interpretation of the transfection assays (22). Nevertheless, in our proposed model for the transcriptional regulation of the hTERT gene (Figure 5), we took the effect of the two CTCF sites into account.
Although promoter methylation might be one of the main mechanisms involved in hTERT regulation in tumor tissues and cell lines (27,40,41), methylation-independent mechanisms have also been identified (25,26,41–43). In normal tissues, we observed that the hTERT gene was not methylated in testis, a site where telomerase is highly expressed (unpublished data). It is interesting to note that expression of an hTERT transgene in mice remained high, especially in testis, regardless of whether the transgene comprised only an 8-kbp region of the hTERT promoter or the entire hTERT gene promoter (21,44). According to Horikawa et al., the use of the entire hTERT promoter in mouse leads to a human-like pattern of hTERT expression, suggesting that specific regulatory sequences, rather than mouse background, determines differences in hTERT and mTERT expression. Together, these data suggest that tissue-specific factors present in testis and in some tumor cells could prevent binding of the CTCF repressor to the first two exons of hTERT and thus lead to hTERT expression. According to previous data, the testis-specific factor BORIS, a paralog of CTCF (45), might be a good candidate, and further studies will be necessary to investigate its potential role in the hTERT regulation.
In summary, hTERT expression is induced when the hTERT CpG island is sufficiently hypermethylated to avoid binding of the CTCF repressor and when a small part of the core promoter region is hypomethylated to allow the transcription complex to be formed. Under these conditions, additional regulators can come into play to effectively induce hTERT expression. Figure 5 illustrates the dual role of DNA methylation in the transcriptional regulation of the telomerase hTERT gene. The complexity of hTERT regulation might hamper the development of anticancer therapies targeting telomerase and calls for further studies.
The authors are grateful to Dr. Herbert C. Morse III for critical reading of the manuscript. This work was funded mainly by a grant from the Swiss National Science Foundation (grant number: 3100AO-101732) and partially by NIAID intramural funding. Funding to pay the Open Access publication charge was provided by xxxx.
Conflict of interest statement. None declared.