Gain of DNA methylation is enhanced in the absence of CTCF at the human retinoblastoma gene promoter

HeLa cells were cultured in DMEM medium; K562 cells were cultured in ISCOVE medium; human glioblastoma-astrocytoma U87MG and U373MG cells, neuroblastoma SHSY-5Y cells, and human glioblastoma T98G cells were cultured in RPMI-1640 medium; all media contained 10% (v/v) fetal bovine serum (FBS) and 1% penicillin/streptomycin. Culture media were purchased from Invitrogen. Cells were maintained at 37°C in a humidified 5% CO₂-containing atmosphere.

The Rb promoter genomic DNA (positions 1634-2020, GenBank accession number L11910) was subcloned into the XhoI/HindIII sites of pGL3-Basic plasmid (Promega) to generate the pGLRb plasmid. The pERb and pERbmutE plasmids containing the GFP reporter gene under the control of the Rb wild-type and mutated promoters, respectively, were previously generated and described by De La Rosa-Velazquez et al., 2007 [8]. The integrity of all plasmid constructs was verified by DNA sequencing.

Linearized pERb and pERbmutE were used to generate stable cell lines. After selection, neomycin-resistant clones were isolated and analyzed by fluorescence-activated cell sorting (FACS). Clones were subsequently cultured in the absence of neomycin in medium for up to 23 weeks. The integrity of the transgene was tested by Southern blotting. For reactivation experiments, stable cell lines were treated for 3 days with 5-aza-2'-deoxycytidine (30 μmol/ml), Trichostatin-A (15 ng/ml), or both. After treatment, the cells were harvested in FACS flow (BD Science) and analyzed by FACS. The percentage of reactivated cells was plotted in the corresponding graph.

Bisulfite analysis was performed as described previously [9]. The PCR primers for stable cell lines were designed specifically against the plasmid sequences to avoid amplification of the endogenous promoter. Nested PCR was done with the EGFPbis1-EGFPbis2 primer pair, and the second round of PCR amplification was done with the EGFPbis3-EGFPbis4 primer pair. The product from the second PCR was gel-purified, and a third PCR was done with A3res-A4res primers specific for the Rb promoter (see below). PCR products were cloned in the pGEM-11zf vector (Promega) and sequenced using the T7 primer. Primers used were EGFPbis1: 5'-TTTGGTTTTTTGTTGGTT TTTTGT-3' and EGFPbis2: 5'-AAATAAACCAAAACACCAACAAC-3'; EGFPbis3: 5'-CG GGATCCTTTTTTTTGTGTTATTTTTTG-3' and EGFPbis4: 5'-CGGGATCCAAATCAACT TACCCTAAATAAC-3'; A3res: 5'-CGGGATCCTTAGGTTTTTTAGTTTAATTTTTTATGA T-3'; A4res: 5'-CGGGATCCAACTATAAAAAAACCCCAAAAAAAAC-3'; GFPRb-Fw: 5'-GGGATTTAGATTTTTTGTATAGTT-3' and GFPRb-Rev: 5'-CAAATAAACTTCAAAA TCAACTTAC-3'.

The following antibodies were used in this study: acH3, acH4 antibodies, CTCF (N-17), MBD2 (N-18), MeCP2 (07-013) and Kaiso (clone 6F) from Millipore (Millipore); H3K27me3 antibody was kindly provided by Dr. Thomas Jenuwein (Max-Planck Institute of Immunobiology and Epigenetics, Freiburg, Germany). The rabbit anti-Rb (C-15) antibody (sc 12370) was purchased from Santa Cruz Biotechnology and horseradish peroxidase (HRP) linked to anti-rabbit immunoglobulin (81-6120) from Zymed. Mouse anti-human actin antibody was kindly provided by Dr. Alejandro Zentella Dehesa (Instituto de Investigaciones Biomédicas, UNAM, México).

For semiquantitative RT-PCR, total RNA was isolated using Trizol reagent from Invitrogen Corporation (GIBCO), and 5 μg RNA from each preparation was used as an RT template in a reverse transcription system (Promega). PCR was performed using the following specific primer pairs: Ex27RBF2: 5'-GGTATGTAACAGCGACCGTGTG-3' and Ex27RBRev: 5'-CTCTTCCTTG TTTGAGGTATCC-3'; β-actin: 5'-CGTACCACTGGCATCGTG-3' and 5'-GGTAGTCAGTC AGGTCCC-3'. Total protein concentration was determined using the commercial Bradford reagent assay (Bio-Rad). 10 μg of total protein was used for the detection of Rb. Samples were first boiled in sample buffer (125 mM Tris-HCl pH 6.8, 1% v/w SDS, 10% v/v glycerol, 0.1% bromophenol blue, 2% v/v 2β-mercaptoethanol) for 5 min and resolved by 10% SDS-PAGE. The proteins were transferred to PVDF membranes (Bio-Rad) using a Trans-Blot Cell system (Bio-Rad) in transfer buffer (25 mM Tris, 190 mM glycine, 10% methanol) at 40 V overnight. The following day, the membrane was probed for 1 hr with rabbit anti-Rb (C-15) antibody diluted 1:2000 in TBS buffer (150 mM NaCl, 20 mM Tris, 0.1% Tween, 1% BSA, 5% non fat milk, pH 7.5). After extensive washing, the membrane was incubated for 1 hr with HRP anti-rabbit immunoglobulin. The signals were detected by enhanced chemiluminescence using the supersignal system (Pierce) and quantified by densitometry. As a control, actin was simultaneously detected using a mouse anti-human actin antibody diluted 1:750, and then developed using horseradish peroxidase linked to anti-mouse immunoglobulin and the same chemiluminescence system.

ChIP assays were performed as described previously [8]. Briefly, U87MG glioma cells (2 × l0⁷ cells) were fixed with 1% formaldehyde and then neutralized by adding 0.125 M glycine. Cells were collected and lysed in Cell Lysis buffer (5 mM EDTA pH 8.0, 50 mM de TRIS-HCl pH 8.1, 1% SDS, protease inhibitor cocktail). The nuclei lysate was sonicated to obtain soluble chromatin with an average length of 500 bp. After 1:10 dilution with dilution buffer (20 mM Tris-Cl, pH 8.1, 2 mM EDTA, 150 mM NaCl, 1% Triton X-100), chromatin solutions were precleared and then incubated with or without 4 μg anti-CTCF, anti-acH3, anti-acH4 and anti-H3K27me3 antibodies, then the mixture was incubated at 4°C overnight on a rotating platform. The same immunocomplexes were recovered with protein A-Sepharose beads. After extensive washing, the bound DNA fragments were eluted, and the resulting DNA was subjected to PCR reactions using the following primer pairs: RbPromFw/RbPromRev for the endogenous Rb promoter and RbPromFw/Hec02r for the stable transgene. As a negative control primers RTRb-F/RTRB-R were used to amplify exon 27 of the Rb gene. PCR products were separated by gel electrophoresis on a 2% agarose gel and visualized. Primer used: RbPromFw: 5'-CCAGACTCTTTGTATAGCC-3'; HEC02R: 5'-ACCATGGTGGCGACC-3'; RTRb-F: 5'-AAGTACCCATCTAGTACT-3'; RTRb-R: 5'-AAGTTACAGCATCTCTAAA-3'; Igf2-CTCF-Fwd: 5'-CAGGCTCCCCCAAAATCTA-3'; Igf2-CTCF-Rvs: 5'-GGGAACATAGAGAAAGAGG-3'; Retinoblastoma Exon 27: RTRb-F 5'-AAGTACCCATCTAGTACT-3' and RTRb-R: 5'-AAGTTACAGCATCTCTAAA-3' [8].

With the aim of further understanding the DNA methylation dynamics over the Rb promoter, we performed sodium bisulfite genomic DNA transformation and sequencing of the core promoter at 130 and 170 days of continuous cell culture (Figure 1A). The DNA methylation status was analyzed over a transgene that was stably integrated in HeLa cells [8]. Using the same cell line we had previously demonstrated a modest gain of DNA methylation (at 100 days in culture) with severe consequences on gene expression [8]. Here we found that after additional cell divisions there was a gradual but not dramatic increase in DNA methylation (Figure 1B). After 130 days, 14% DNA methylation was reached, and after 170 days, 22%. At this last time point we found an increase in the methylated CpG dinucleotides located preferentially over the CTCF, RBF-1 and E2F binding motifs (Figure 1B). These results suggest that after prolonged cell culture and many cell divisions there is a gain of DNA methylation, which reinforces epigenetic silencing.

To confirm the contribution of CTCF to the protection of tumor suppressor gene promoters against epigenetic silencing, we decided to generate a large set of independent stably integrated constructs carrying the Rb promoter fused to the GFP reporter gene, where the promoter had a mutated CTCF site. The mutation (pERmutE) was previously defined and confirmed by gel-shift assays using probes carrying different mutations [8]. Nineteen independent lines with the intact promoter and 18 independent lines with the mutant version of the promoter were generated in the erythroleukemic K562 cell line (Figure 2A and 2B). Transgene copy-number and integrity was confirmed by Southern blotting (Figure 2C). For the transgene in which the CTCF binding site was mutated, we found an approximately 2-fold increase in the number of stable cell lines with extinguished transgene expression after 100 days of continuous cell culture (Figure 2). This transgene silencing was also clearly seen in FACS profiles at different time points (see Additional file 1, Figure S1). We conclude that CTCF binding to the Rb promoter is critical for its epigenetic integrity and transcriptional activity. This is in agreement with our previous CTCF knockdown experiments, where CTCF depletion caused a loss of Rb gene expression [9].

Taking advantage of the K562 stable cell lines carrying a mutated Rb promoter (pERbmutE: lines 1109, 1112, 2102 and 2111), we performed reactivation experiments using the histone deacetylase inhibitor Trichostatin-A (TSA), the DNA methylation inhibitor 5-aza-2'-deoxycytidine (5-azadC), and their combination (Figure 3A). Reactivation conditions were previously established (Additional file 2, Figure S2) and transgene activity was followed at 30 and 100 days of continuous cell culture. Interestingly, we found that a significant reactivation was observed at early time points of cell culture (30 days) using each of the inhibitors or their combination. A more pronounced reactivation was observed when the 5-azadC inhibitor was tested compared to TSA treatment (Figure 3B). Of note, single-copy integrants were more efficiently reactivated in comparison to multi-copy integrants (Figure 3B). This observation may be related to evidence demonstrating that multi-copy integrants can be recognized as repetitive sequences that induce the formation of heterochromatin [17‐19]. Unexpectedly, at 100 days in cell culture the reactivation became modest. The DNA methylation status of the reactivated cell line 1112 and 2111, and a new and independent cell line 2102, were assessed (Additional file 3, Figure S3). Thus, in these series of reactivation experiments, using the 5-azadC inhibitor, DNA methylation was partially lost indirectly demonstrating that DNA methylation is one of the causes of silencing.

These results could appear contradictory to our previous observations where we observed 60% reactivation using 5-azadC in a HeLa cell line [8], but this seems not to be the case since that cell line, which carried an intact CTCF sequence, was previously sorted from a clone that was partially silenced after 100 days in continuous cell culture [8]. Additionally in the same previous work, we found similar reactivation levels at early time points, as observed here in the context of the mutant CTCF sites, which are more rapidly and robustly silenced. These results exemplify two different processes leading to the establishment of an epigenetic silencing conformation over the Rb core promoter. We believe that at early time points (30 days), DNA methylation is progressively and actively incorporated without reaching critical densities or resulting in the methylation of specifically located CpG dinucleotides. This is likely because this sort of intermediate state of DNA methylation can be efficiently reverted. In contrast, at later times of cell culture (100 days or more), histone deacetylation appears to be irreversible, but DNA methylation is partially erased suggesting that other repressive histone marks and nuclear factor complexes have been established. One attractive possibility is the gain of the histone H3K27me3 mark, a repressive mark incorporated by members of the Polycomb group repressor proteins.

Next, we assessed the DNA methylation status of two of the stable cell lines generated with the mutated CTCF binding site. These cell lines were previously tested in the reactivation experiments (Figure 4; cell lines 1112 and 2111). For comparison we evaluated the DNA methylation over a transgene in a K562 stable line in which the Rb promoter was intact (Figure 4A). In agreement with previous observations, we found that there was no significant DNA methylation at day 30 of continuous cell culture even though transgene expression started to decay (see Figure 4B). Instead, after 100 days in culture the 1112 and 2111 cell lines became DNA hypermethylated (Figure 4C). This result is consistent with the inability of CTCF to bind to the transgene Rb promoter. This was further confirmed by bisulfite sequencing of the K562 endogenous Rb promoter, where we found less than a 2-fold gain of DNA methylation (Additional file 4, Figure S4). Taken together, these results demonstrate that CTCF is a critical factor that contributes to the epigenetic protection of the Rb core promoter, which in the absence of CTCF binding, is prone to DNA hypermethylation.

To further assess the contribution of CTCF to the protection of the Rb promoter against DNA methylation, we analyzed a series of glioma cell lines (Figure 5). Based on our previous observations in which CTCF was not bound to the p53 promoter in vivo, we asked if in glioma cells the lack of Rb gene expression could correlate with CTCF absence at the Rb promoter [9]. We first evaluated the relative concentration of Rb by Western blotting in four different glioma cell lines (SHSY-5Y, T98G, U87MG and U373MG) compared to human K562 cells (Figure 5A). In this experiment we noticed that T98G and U87MG cells had the lowest amount of Rb protein compared to the other transformed cell lines. Next, we determined Rb mRNA levels by semi-quantitative duplex-PCR (Figure 5B). Rb transcription was significantly diminished in the U87MG glioma cell line. Based on these results we performed sodium bisulfite sequencing and sequencing of the endogenous Rb core promoter in the U87MG cell line (Figure 5C). In agreement with the Rb gene expression levels, we found that the Rb promoter reached 29% DNA methylation in U87MG cells. Subsequently, we performed a chromatin immunoprecipitation assay employing primers to amplify the Rb core promoter and a set of antibodies against different histone marks, CTCF and methyl-CpG-binding proteins. The results revealed no decrease in histone acetylation, gain of the H3K27me3 histone repressive mark, incorporation of the methyl-CpG-binding proteins MBD2 and MeCP2, and importantly, we confirmed the absence of CTCF bound to the Rb promoter in U87MG cells (Figure 5D). As a positive control for the in vivo binding of CTCF, we performed a ChIP assay in U87MG glioma cells using primers from the Imprinting Control Region of the human Igf2/H19 imprinted locus, which is known to have several CTCF binding sites [20, 21]. In summary, these results support our in vitro data and demonstrate that in a tumoral cell line, the lack of CTCF binding to the Rb promoter is, in part, responsible for its epigenetic silencing.