methylation of the gene and alterati - Oxford University Press

Download PDF
15 downloads 0 Views 674KB Size Report
Comment
We thank Ms Yoko Sekine for her excellent technical assistance, Dr Osamu. Goto for his helpful support, Dr Masakazu Toi for providing breast cancer cell lines ...
Carcinogenesis vol.21 no.12 pp.2193–2201, 2000

Distinct mechanisms of loss of estrogen receptor α gene expression in human breast cancer: methylation of the gene and alteration of trans-acting factors

Takashi Yoshida1,5, Hidetaka Eguchi1, Kei Nakachi1, Keiji Tanimoto3, Yasuhiro Higashi2, Kimito Suemasu2, Yuichi Iino4, Yasuo Morishita5 and Shin-ichi Hayashi1,6 1Hormone-Associated

Cancer Research Group, Laboratory of Cancer Diagnosis and Therapy, Saitama Cancer Center Research Institute and of Breast Surgery, Saitama Cancer Center Hospital, 818 Komuro, Ina, Saitama 362-0806, Japan, 3Department of Cell and Molecular Biology, Medical Nobel Institute, Karolinska Institute, S-171 77, Stockholm, Sweden and 4Department of Critical Care Medicine and 5Department of Second Surgery, Gunma University School of Medicine, 3-39-22 Showa, Maebashi, Gunma 371-8511, Japan 2Department

6To

whom correspondence should be addressed Email: [email protected]

We have previously shown that the distal promoter (promoter B) of the estrogen receptor α (ERα) gene is responsible for the enhanced expression of the ERα gene seen in human breast cancer and that a novel trans-acting factor, estrogen receptor promoter B associated factor 1 (ERBF-1), is required for transcription from promoter B in breast cancer cells. In development of breast cancer, loss of ERα gene expression is one of the most important steps in acquiring hormone resistance, though the mechanisms are poorly understood. Recent studies have reported that methylation of the ERα gene promoter A and exon 1 was inversely associated with ERα gene expression in human breast cancer and cell lines. The methylation status of the promoter B region, which is responsible for overexpression of ERα protein in cancer tissue, has not been investigated. In this report, we found that the methylation status of promoter B, as well as that of promoter A, was inversely associated with ERα gene expression in human breast cancer and cell lines. Specific methylation of ERα gene promoters in vitro directly decreased transcription of the ERα gene in a reporter assay. Demethylating treatment induced transcription of ERα mRNA from promoter B in ZR-75-1 cells, which showed no transcription from promoter B, despite weak ERBF-1 expression, but not in ERα-negative MDA-MB-231 and BT-20 cells, which lack ERBF-1. ZR-75-1 cells showed promoter activity equal to that of MCF-7 cells in a reporter assay. Our results indicate that methylation of promoter B of the ERα gene is important for loss of ERα gene expression in human breast cancer, and methylation of the promoters can directly modulate ERα gene expression. However, loss of critical transcriptional factors such as ERBF-1 may also be involved in some ERα-negative cases. Introduction As a mediator of the effects of estrogen on target cells, estrogen receptor α (ERα) plays an important role in regulating growth Abbreviations: 5-aza-dC, 5-aza-2⬘-deoxycytidine; ER, estrogen receptor; ERBF-1, estrogen receptor promoter B associated factor 1; ERF-1, estrogen receptor factor; GAPDH, glyceraldehyde phosphate dehydrogenase; RT–PCR, reverse transcription–polymerase chain reaction; TK, thymidine kinase. © Oxford University Press

and differentiation of normal breast epithelium as well as in the development and progression of breast cancer (1,2). Approximately two-thirds of breast cancers express ERα protein, as assayed by ligand binding and immunohistochemistry; breast cancers lacking ERα (approximately onethird of breast cancers) rarely respond to endocrine therapy and are associated with a lower grade of histological differentiation, higher growth fraction and worse clinical outcome (3). Thus, loss of ERα gene expression is one of the most important steps in acquiring hormone resistance; in ERα-negative breast cancers, no significant genetic alterations, such as insertion, deletion, rearrangement, or point mutation within the ERα gene, have been reported (3). Transcription of the human ERα gene occurs from at least two different promoters (4–6), with the distal promoter (promoter B) located 2 kb upstream of the proximal one (promoter A). The resultant transcripts from the two promoters differ only in the non-coding region at the 5⬘ end, and both types of ERα mRNA encode the same protein. We have previously shown that the levels of expression of total ERα mRNA and of transcript from promoter B correlated well with the amount of ERα protein in human primary breast cancer, indicating that promoter B plays an essential role in the regulation of ERα gene expression in human breast cancer (7). We have also identified a novel trans-acting factor, estrogen receptor promoter B associated factor 1 (ERBF-1), which is critical for the transcription activity of promoter B in ERαpositive breast cancer cell lines (8). Loss of ERα gene expression may be explained by either epigenetic modifications of the gene or loss of trans-acting factors which are essential for expression of the ERα gene. One mechanism that may block transcription is methylation of cytosine/guanine-rich areas, termed CpG islands, which are located in the 5⬘ regulatory regions of housekeeping genes and numerous tissue-specific genes (9). Cancer cells often display anomalous patterns of DNA methylation, with sitespecific hypermethylation in CpG islands and hypomethylation of bulk genomic DNA (10). In breast cancer, aberrant hypermethylation has been reported with respect to some genes, such as the E-cadherin (11), BRCA1 (12,13) and HIC-1 genes (14). The ERα gene has a typical CpG island within its promoters and first exon. The CpG island is unmethylated in both normal breast tissue (15) and non-tumorigenic breast epithelium cell lines that do not express ERα (16). Recent studies have reported that methylation of the CpG island in ERα gene promoter A and exon 1 was associated with loss of ERα gene expression in human breast cancer tissue and cell lines (15,17–19) and that ERα-negative breast cancer cells MDAMB-231 treated with the demethylating reagent, 5-aza-2⬘deoxycytidine (5-aza-dC) induced expression of ERα mRNA, resulting in production of functional ERα protein (20). However, until now there have been no reports on the methylation status of the promoter B region, which is responsible for overexpression of ERα protein: Direct evidence of suppression 2193

T.Yoshida et al.

of ERα gene expression by methylation, and the molecular mechanism of this suppression, have not been addressed so far. In this report, we show that the methylation status of promoter B of the ERα gene, along with the methylation status of promoter A, was inversely associated with expression of this gene in human breast cancer and in breast cancer cell lines. Specific methylation of ERα gene promoters directly decreased the transcription of this gene in a transient transfection experiment. However, demethylating treatment of ERα-

negative breast cancer cells induced ERα gene expression in BT-20 cells but not in MDA-MB-231 cells, and MDA-MB231 cells showed little activity in transfection of reporter constructs containing unmethylated promoter B. Our results indicated that methylation of the two promoter regions of the ERα gene can directly modulate expression of this gene, though some other mechanism, such as loss of critical transcriptional factors, may also be involved in some cases of ERα-negative breast cancer. Materials and methods Tissue samples Samples of human primary breast cancer tissue from 22 female patients were obtained from Saitama Cancer Center Hospital, Japan. Tumor samples obtained in surgery were immediately dissected to remove residual normal tissue. DNA and RNA were then extracted from these samples. Cells and culture Human breast cancer cells (MCF-7, T-47-D, ZR-75-1, MDA-MB-231 and BT-20) were cultured in RPMI 1640 medium supplemented with 5% or 10% fetal calf serum (FCS), 2 mM L-glutamine and 2 µg/ml gentamicin at 37°C in a humidified atmosphere of 5% CO2 in air. Southern blot analysis Ten to 15 µg of genomic DNA was digested first with 10 units/µg of SacI (methylation-insensitive; TaKaRa, Tokyo, Japan) in an appropriate buffer at 37°C for 8 h, followed by either 15 units/µg of HhaI or HpaII (methylationsensitive; TaKaRa) at 37°C overnight. The DNA fragments were separated in 1% agarose gels, blotted on to Hybond-N nylon membrane (Amersham, Little Chalfont, UK), and fixed with UV radiation. Either a 1.4 kb fragment of promoter B (–3284 to –1864) or a 1.3 kb fragment of promoter A (–1168 to ⫹190) of the ERα gene, which was synthesized from genomic DNA using PCR, was labeled with [α-32P]dCTP (3000 Ci/mmol) using a Rediprime II DNA labeling kit (Amersham). Hybridization was carried out in ExpressHyb Hybridization Solution (Clontech, Palo Alto, CA) at 65°C for 1 h. The hybridized bands were visualized by autoradiography with a Fuji Bio-Image Analyzer BAS2000 (Fuji Film, Tokyo, Japan). Reverse transcription–polymerase chain reaction (RT–PCR) Total RNAs were prepared from cultured cells (1–5 ⫻ 106 cells) and ~0.1 g of human breast cancer tissue according to the method of Chomczynski and Sacchi (21). RT–PCR, used to detect total ERα mRNA, mRNA transcribed from promoter A (type A) and mRNA from promoter B (type B), together with internal control GAPDH mRNA, was performed as previously reported (7). For ZR-75-1 cells, three more PCR cycles were used to detect ERα mRNAs than for other cells. ERα protein assay ERα enzyme immunoassay for tumor samples was performed with ERα enzyme immunoassay kits (Abbott, Abbott Park, IL). ERα values of ⬍5 fmol/mg protein were considered negative.

Fig. 1. (A) Organization of the ERα gene promoter region. (Middle) Arrows with numbers indicate the positions of the two major transcriptional start sites. The position of the proximal transcriptional start site is designated as ⫹1. The distal promoter (promoter B) is located 2 kb upstream of the proximal one (promoter A). The primary transcript originating from promoter B is spliced as shown. The translational start site is shown by a vertical arrow with ATG. (Above) Map of CpG sites of ERα gene. Vertical bars indicate CpG nucleotides. Horizontal bars indicate the probes for hybridization, a 1.4 kb fragment of promoter B and a 1.3 kb fragment of promoter A. (Below) Map of restriction enzyme sites used in this study within the promoter regions of ERα gene. SacI is a methylation-insensitive enzyme, while HpaII and HhaI are methylation-sensitive. (B) Methylation status of ERα gene promoter B (above) and A regions (below), together with expression levels of ERα mRNA and protein in human breast cancer tissue. Genomic DNA from breast cancer tissue was digested with SacI alone (lane 1) or SacI and HhaI (lanes 2–11). Total, type B and type A mRNA expression was examined by RT–PCR. Expression levels of mRNA was categorized as 3⫹ (strong expression), ⫹ (expression) or – (no detectable expression); n.d., not determined.

2194

Mechanisms of loss of ERα gene expression Reporter plasmid constructions The 1.3 kb fragment (–1168 to ⫹190) of promoter A and the 1.4 kb fragment (–3284 to –1864) of promoter B were ligated to pGL2-Basic vector (Promega, Madison, WI) at the XhoI restriction site to give plasmids pGL2-ProA and pGL2-ProB, as described previously (8). Next, pGL2-ProB was digested with

Fig. 2. Methylation status of the ERα gene promoter region B and ERα mRNA expression in human breast cancer cell lines. (A) Methylation status of the ERα gene promoter B region by Southern blot analysis in MCF-7 (lanes 1–3), T-47-D (lanes 4–7), ZR-75-1 (lanes 7–9), MDA-MB-231 (lanes 10–12) and BT-20 cells (lanes 13–15). SacI-digested samples are in lanes 1, 4, 7, 10 and 13, SacI- and HpaII-digested ones in lanes 2, 5, 8, 11 and 14 and SacI- and Hha I-digested ones in lanes 3, 6, 9, 12 and 15. (B) Schematic diagrams of methylation status of HpaII and HhaI sites within the ERα gene promoter regions are shown as follows: s, unmethylated; n, partially methylated; ⫻, methylated. (C) ERα mRNA expression detected by RT–PCR in MCF-7 (lane 1), T-47-D (lane 2), ZR-75-1 (lane 3), MDAMB-231 (lane 4) and BT-20 cells (lane 5).

KpnI and HindIII and the promoter B region was ligated to pGL3-Basic vector (Promega); the resultant plasmid was named pGL3-ProB. Methylation of promoter regions of the ERα gene in reporter plasmid constructs The promoter region in reporter plasmid constructs was methylated essentially as described (22). pGL2-ProA and pGL2-ProB were digested with KpnI and HindIII. The fragments of the promoter A and B regions were purified using a Geneclean III kit (Bio 101, Vista, CA), and methylated in vitro with 3 units of SssI methylase (New England Biolabs, Beverly, MA) per µg of DNA in the presence of 160 µM S-adenosyl-methionine at 37°C for 3 h. For the unmethylated control, the same fragments were incubated in the same conditions without the methylase. Complete methylation of the fragments was confirmed by HpaII and HhaI digestion. pGL3-Enhancer vector (Promega) was digested with KpnI and HindIII, and ligated with an equimolar concentration of methylated or unmethylated promoter fragments at 16°C for 30 min at a DNA concentration of 20 µg/ml. The resultant plasmids were ethanol-precipitated and named pGL3-E.-ProA and pGL3-E.-ProB. Four micrograms of ligated DNA was used for transfection experiments. Sodium bisulfite genomic sequencing Sodium bisulfite genomic sequencing was performed essentially as described (23–26). Ten micrograms of genomic DNA from cell lines was digested with XbaI in 90 µl of reaction mixture, denatured with 10 µl of 3 M NaOH at 37°C for 15 min and then incubated with 1040 µl of freshly prepared 3.6 M sodium bisulfite, pH 5.0 (Wako, Osaka, Japan) and 50 µl of freshly prepared 10 mM hydroquinone (Wako) for 16 h at 50°C. After bisulfite treatment, the DNA was desalted using a Geneclean III kit, denatured with 22 µl of 3 M NaOH at 37°C for 15 min, precipitated with 140 µl of 5 M ammonium acetate pH 7.0 and 900 µl of ethanol, and resusupended in 100 µl of MilliQ water. First-round PCR amplification was performed in 50 µl of reaction mixture containing 2 µl of bisulfite-treated genomic DNA, 0.5 µM each of primers, 200 µM dNTPs, 1.5 mM MgCl2 and 1.25 units of AmpliTaq Gold polymerase (Perkin–Elmer, Foster City, CA) under the following conditions: 95°C for 30 s, 58°C for 1 min, 72°C for 1 min for 40 cycles. To obtain products for sequencing, a second round of PCR was performed with nested primers

Fig. 3. Effects of in vitro methylation of ERα gene promoters on transcriptional activity. Methylated (black bars) or unmethylated (white bars) pGL3-E.-ProA and pGL3-E.-ProB were constructed and transfected into MCF-7 cells as described in Materials and methods. Luciferase activity is shown relative to that of unmethylated pGL3-E.-ProA or -ProB, which were taken as 100%. All the measured luciferase activities of plasmid constructs were normalized against the renilla luciferase activity of pRL-TK. The results are averages of three transfection experiments. Bars show standard errors.

2195

T.Yoshida et al. in 50 µl of reaction mixture containing 10 ng of the PCR product, 0.5 µM each of primers, 200 µM dNTPs, and 2.5 units of Pyrobest DNA polymerase (Takara) under the following conditions: 98°C for 10 s, 52°C for 1 min, 72°C for 1 min for 30 cycles. Modified primers were as follows: ERBM1U (5⬘-TTAAGGGTAGGGGTAAAGGGGTTGGGGTTT-3⬘) and ERBM1D (5⬘ATACCCCCATAAAAAACAACAATCCTCATC-3⬘) for first-round PCR, ERBF-b (5⬘-TGGTTGTGTTATATTGTTTT-3⬘) and ERBM2D (5⬘-TCTCCCTACTAAAATACAAACACTAACCAA-3⬘) for second-round PCR. The resultant PCR products were purified using a Geneclean III KIT (BIO 101) and subcloned into pGEM-7Zf(⫹) vector (Promega). Sequencing was performed using an ABI PRISM 310 Genetic Analyzer (Perkin–Elmer). 5-Aza-dC treatment of cell lines The demethylating reagent, 5-aza-dC (Sigma, St Louis, MO), was freshly prepared in MilliQ water. Human breast cancer cells (5–10 ⫻ 105 cells) were plated on to 100 mm plastic culture dishes in RPMI 1640 with 10% FCS. Twenty-four hours later, cells were treated with 10–6 M 5-aza-dC. The medium was changed 24 h after treatment and then every 3 days. DNA and RNA were prepared 7 days after treatment. Transient transfection and luciferase assay MCF-7 cells were transiently transfected using SuperFect (Qiagen, Hilden, Germany) and ZR-75-1 and MDA-MB-231 cells using Lipofectamine (Gibco– BRL), according to the manufacturers’ instructions. All the transfection experiments were done in triplicate. Transient transfection was performed essentially as described previously, with a slight modification (27). Briefly, culture cells were grown on 35 mm plastic culture dishes in RPMI medium with 5% or 10% FCS to 30–50% confluence. A plasmid cocktail with an internal control vector, pRL-TK (Promega), was mixed with the transfection reagent and added to the culture. After 48 h of incubation, the cells were lysed in a lysis buffer and the luciferase activity of plasmid constructs and pRL-TK, which encodes a renilla luciferase reporter gene driven by a thymidine kinase (TK) promoter, was measured using a dual luciferase assay

Fig. 4. Legend on facing page

2196

kit (Promega). All the measured luciferase activity of plasmid constructs was normalized against the renilla luciferase activity of pRL-TK.

Results Methylation status of ERα gene promoter regions is associated with expression of ERα mRNA and protein in human breast cancer tissue The methylation status of the promoter B region of the ERα gene was evaluated in human breast cancer tissue from 22 patients. Genomic DNA was digested with SacI (methylationinsensitive) and HhaI (methylation-sensitive) and subjected to Southern blot analysis. When HhaI site (a) within promoter B region was unmethylated, a 1.0 kb fragment and additional fragments of 1.3, 1.9 and/or 2.1 kb were expected to hybridize with the 1.4 kb probe of promoter B; when this site was methylated, fragments of 2.3, 2.9 and/or 3.1 kb would be observed (Figure 1A). Southern blot analysis in 10 representative patients are shown in Figure 1B, part 1, where expression levels of ERα protein, total and type B mRNA, are also presented. In patients with high levels of ERα (lanes 8–11), there was a major 1.0 kb band and an additional 1.3 kb one, indicating that HhaI site (a) was unmethylated. Thus, the promoter B region of the ERα gene was rarely methylated in these tissue samples. On the other hand, together with a 1.0 kb band, additional longer bands of 2.3, 2.9 and 3.1 kb were observed in the samples from patients with no ERα

Mechanisms of loss of ERα gene expression

(lanes 2–4) or low ERα (lanes 5–7). Thus, the promoter B region in these tissue samples seems to be partially methylated. The methylation status of the promoter A region of the ERα gene was also evaluated using the same human breast cancer tissue (Figure 1B, part 2). The promoter A region in the samples from patients with high levels of ERα was rarely methylated, and partial methylation was observed for the samples from patients with little or no ERα. Use of HpaII in combination with EcoRI to detect methylation in the promoter region showed essentially the same results as those obtained using HhaI (data not shown); similar results were obtained with the other 12 patients (data not shown). These results indicate that methylation of the promoter B region as well as the promoter A region of the ERα gene is inversely associated with expression of this gene.

Fig. 4. Methylation status of promoter B of the ERα gene by sodium bisulfite genomic sequencing. (A) Above: representative sequences containing nine CpG sites (vertical arrows indicate the positions of CpG sites). Unmethylated cytosine residues were converted to uracils by treatment with sodium bisulfite and sequenced as thymine, while methylated cytosine residues remain unchanged. Below: frequency of conversion of cytosines to thymines in CpG sites. (B) More than 14 positive clones were isolated and sequenced as described in Materials and methods. The number of methylated cytosines at a specific CpG site was divided by the total number of analyzed clones, to yield the methylation frequency. CpG1 (cross-hatched bars) and CpG3 (white bars) are located within the HhaI (a) and HpaII (y) sites, respectively; CpG9 (black bars) is within the ERBF-1 binding element. ERBF-b and ERBM2D (horizontal arrows) indicate the positions of primers used for the second-round PCR. The arrow with a number indicates the position of the transcriptional start site.

Methylation status of the ERα gene promoter regions in human breast cancer cell lines We next analyzed the methylation status of ERα gene promoter regions in human breast cancer cell lines. The methylation status of the promoter B region was evaluated in terms of SacI⫹HhaI digestion and of SacI⫹HpaII digestion (Figure 2A). Among three ERα-positive cell lines, HhaI site (a) in T47-D cells was completely unmethylated (lane 6). In MCF-7 cells, together with a 1.0 kb band, additional longer bands were observed (lane 3), indicating that HhaI site (a) was partially methylated. Importantly, ZR-75-1 cells did not show a 1.0 kb band (lane 9), which seems to signify complete methylation of HhaI site (a). Likewise, HhaI site (a) of ERαnegative BT-20 cells was completely methylated (lane 15). Surprisingly, other ERα-negative MDA-MB-231 cells showed a 1.0 kb band with a 3.1 kb band (lane 12), so apparently HhaI site (a) was partially methylated. The methylation status of the promoter A region was also evaluated as above (data not shown); results of Southern blot analysis of the methylation status of the promoter A and B regions are summarized in Figure 2B. We also measured the amounts of total, type A and type B ERα mRNAs for these cell lines by RT–PCR (Figure 2C): MCF-7 (lane 1) and T-47-D (lane 2) showed high expression of these mRNAs. ZR-75-1 cells use only promoter A in the transcription of ERα mRNA (7); in accordance with this, we found that ZR-75-1 cells (lane 3) showed low expression of total and type A mRNA, and no expression of type B mRNA. MDA-MB-231 (lane 4) and BT-20 (lane 5) cells did not express total, type B or type A mRNA.

Fig. 5. Effects of 5-aza-dC on ERα gene expression in human breast cancer cell lines. Breast cancer cells were treated with 10–6 M of 5-aza-dC for 7 days. Total mRNA and type A and type B mRNAs of the ERα gene were detected by RT–PCR as described in Materials and methods. Lanes 1 and 2, ZR-75-1; lanes 3 and 4, MDA-MB-231; lanes 5 and 6, BT-20. 5-Aza-dCtreated cells are in lanes 2, 4 and 6; untreated cells are in lanes 1, 3 and 5.

2197

T.Yoshida et al.

These results indicate that methylation of the promoter B region as well as the promoter A region is inversely associated with expression of the ERα gene in human breast cancer cell lines. Effects of methylation of ERα gene promoters on its transcription We investigated whether methylation of the ERα gene promoters can directly modulate the transcriptional activity of this gene. First, we performed a luciferase reporter assay in MCF-7 cells by transient transfection using overall methylated pGL2-ProA or -ProB, in which all CpG sites present in these vectors were methylated in vitro. The luciferase activity of methylated pGL2-ProA and -ProB fell to 20.9% and 34.8% respectively, of that of the non-methylated ones (data not shown). Since methylation of the coding region of the luciferase gene has also been shown to influence the activity (28), we next performed transfection experiments using promoter region-specific methylated constructs. In vitro methylation of promoters A and B decreased luciferase activity to 49.7% and 60.7% respectively, of that of the unmethylated ones (Figure 3). Thus, methylation of the ERα gene promoters directly suppresses transcription of the ERα gene. Methylation status of promoter B of the ERα gene analyzed by sodium bisulfite genomic sequencing Next, we examined the methylation status of nine CpG sites around the transcriptional start site of promoter B, which is responsible for overexpression of ERα protein in human breast cancer, in ERα-positive MCF-7 and T-47-D cells, and ERα-

Fig. 6. Functional analysis of ERα gene promoter B in breast cancer cell lines. pGL3-ProB was transiently transfected into MCF-7 (white bar), ZR-75-1 (cross-hatched bar) and MDA-MB-231 cells (black bar) as described in Materials and methods. All the measured luciferase activities of plasmid constructs were normalized against the renilla luciferase activity of pRL-TK. The normalized luciferase activity is shown as relative luciferase activity, with the activity of pGL3-ProB in MCF-7 cells taken to be 100%. The results are averages of three transfection experiments. Bars show standard errors.

2198

negative MDA-MB-231 and BT-20 cells, using sodium bisulfite genomic sequencing. In this method, unmethylated cytosine residues are converted to uracils by treatment with sodium bisulfite, and thus are sequenced as thymines, while methylated cytosine residues are unchanged. To examine the reliability of this method, we performed a control experiment using demethylated genomic DNA treated with SssI methylase. Both SssI-treated and untreated DNA were modified with sodium bisulfite and then amplified with PCR. The fragments were subcloned and the recombinants sequenced. Representative sequences are shown in Figure 4A. All cytosine residues outside the CpG sites were sequenced as thymines after treatment with sodium bisulfite, confirming that the method was working successfully. The frequency of conversion from cytosine to thymine at CpG sites was then compared between untreated and SssI-treated recombinants (Figure 4A). SssI treatment resulted in low conversion (艋20%) at CpG sites 1, 2, 3, 5 and 7 and modest conversion (40–70%) at CpG sites 4, 6, 8 and 9. This result implies that sodium bisulfite genomic sequencing may underestimate the degree of methylation, specifically at CpG sites 4, 6, 8 and 9, possibly due to sequence-specific PCR bias as will be discussed later. Chromosomal DNA from MCF-7, T-47-D, MDA-MB-231 and BT-20 cells was subjected to sodium bisulfite genomic sequencing: More than 14 positive clones in recombinants from each cell line were isolated and sequenced. The number of methylated cytosines at a specific CpG site was divided by the total number of analyzed clones to yield the methylation frequency (Figure 4B). In T-47-D cells (Figure 4B, panel b), which showed no methylation at HhaI (a) and HpaII (y) sites in Southern blotting (Figure 2B), the methylation frequency of CpG1 and CpG3 within these sites and other CpG sites in the region was low, indicating that this region is poorly methylated in these cells. In contrast, CpG1 and CpG3 were highly methylated in BT-20 cells (Figure 4B, panel d), showing a good correlation with the result of Southern blotting. The CpG sites were not uniformly methylated: sites 2, 5 and 7 were highly methylated wherease sites 4, 6, 8 and 9 were poorly methylated. In MCF-7 (Figure 4B, panel a) and MDAMB-231 cells (Figure 4B, panel c), which showed partial methylation at HhaI (a) and HpaII (y) in Southern blotting, the frequency of methylation of CpG sites 1 and 3 was intermediate. Interestingly, CpG9, which is located within the ERBF-1 binding element, was poorly methylated in ERαnegative BT-20 and MDA-MB-231 cells, while ERα-positive MCF-7 cells showed a higher frequency of CpG9 methylation. Since ERα-negative MDA-MB-231 cells showed partial methylation of CpGs in the analysed promoter region, which is similar to the methylation status in ERα-positive MCF-7 cells, methylation alone may not be the sole mechanism of the suppression of ERα gene expression. Effects of 5-aza-dC on ERα gene expression in human breast cancer cell lines In order to determine whether methylation of the ERα gene promoters suppresses expression of the ERα gene, we next examined expression of this gene in breast cancer cells treated with 1.0 µM 5-aza-dC for 7 days. This demethylating treatment has been reported to be sufficient to maintain a physiological level of ERα protein in MDA-MB-231 cells (20, 29). The methylation status of genomic DNA from treated cells was examined by Southern blot analysis; expression of ERα mRNAs was monitored by RT–PCR. The promoter B region

Mechanisms of loss of ERα gene expression

was partially demethylated in treated ZR-75-1 cells, and in treated MDA-MB-231 and BT-20 cells (data not shown). ERα-negative BT-20 cells treated with demethylating reagent induced expression of ERα mRNA, in particular type A mRNA (Figure 5, lane 6). In ZR-75-1 cells, where promoter B was inactive, treatment of the cells (lane 2) enhanced expression of both type A and B ERα mRNA compared with untreated cells (lane 1). Thus, demethylation of these cells altered expression of the ERα gene. On the other hand, treatment of MDA-MB-231 cells did not affect expression of ERα mRNA (lane 4). Although we treated MDA-MB-231 cells with 5-aza-dC at different concentrations (0.1, 0.5, 1.0, 2.0 and 4.0 µM) for 7 days, and performed five more PCR cycles than used in Figure 5 for the cells, no expression of ERα mRNA was observed (data not shown). These data indicate that loss of ERα gene expression may be regulated by different mechanisms in two ERα-negative breast cancer cells. Functional analysis of ERα gene promoter B in human breast cancer cell lines Untreated ZR-75-1 cells did not express type B ERα mRNA, but those treated with demethylating reagent did express this mRNA, indicating that unmethylated promoter B can function in these cells. To confirm this possibility, we transiently transfected ZR-75-1, MCF-7 and MDA-MB-231 cells with pGL3-ProB (Figure 6). ZR-75-1 cells and MCF-7 cells showed equal promoter activity. Thus, an unmethylated promoter can function in ZR-75-1 cells, and specific methylation of promoter B region in the cells may alter the usage of the promoters. On the other hand, MDA-MB-231 cells, in which ERα mRNA was not expressed even after demethylating treatment, showed little promoter activity. Thus, loss of ERα gene expression in the cells may be caused by loss of promoter B-specific transacting factors. Discussion We examined two possible mechanisms that could be responsible for loss of ERα gene expression in breast cancer: (i) the ERα gene promoters may be selectively methylated and inaccessible to the existing transcriptional activators; and/ or (ii) essential activators for ERα transcription may not be available or transcriptional suppressors may predominate. We therefore investigated the role of DNA methylation in regard to loss of ERα gene expression: We found that methylation of the promoter regions of the ERα gene directly suppressed transcription of this gene in vitro. We show that there is an excellent correlation between methylation of the promoter B region of the ERα gene and suppression of ERα gene expression. We also showed that loss of ERα gene expression is not always caused by methylation of promoter regions: In some cases, it may be due to loss of transcriptional activators. We have previously shown that levels of transcription from promoter B correlated well with total ERα mRNA levels and with ERα protein levels in human breast cancer, indicating that transcription from promoter B, but not from promoter A, is responsible for the enhanced ERα gene expression found in human breast cancer (7). Thus, we paid a great attention to the methylation of promoter B region which may be important in the process of loss of ERα gene expression in human breast cancer. In this paper, the methylation status of the promoter B region, as well as the promoter A region, correlated well with the expression of ERα in breast cancer tissue,

indicating that methylation of both promoter regions may occur simultaneously during the loss of ERα gene expression. DNA methylation has been attributed to increased activity of cytosine DNA methyltransferase during carcinogenesis (16,17,30,31). However, the molecular determinants of specific CpG sites, ‘hot spots’ for methylation by DNA methyltransferase, are not known. Therefore, we examined the methylation status of nine CpG sites around the transcription start site of promoter B, using the sodium bisulfite genomic sequencing method. Our control experiment showed that cytosine residues within specific CpG sites, namely CpG 4, 6, 8, and 9, appeared to be converted to thymine (40–70%), though the DNA had been treated with SssI methylase, implying that methylation of these CpG sites may be underestimated by this method. This may be due to the PCR bias observed for several genes: DNA templates containing certain thymines instead of cytosines at CpG sites are preferentially amplified in PCR (32). Methylation of CpG sites 1, 2, 3, 5 and 7 was, however, reliably detected by sodium bisulfite genomic sequencing. Results of sodium bisulfite sequencing agreed well with those of Southern blotting: CpG sites 1 and 3, which correspond to HhaI (a) and HpaII (y), respectively, were highly methylated in BT-20 cells, poorly methylated in T-47-D cells and partially methylated in MCF-7 and MDA-MB-231 cells. Next, we examined ERα gene expression in breast cancer cells treated with the demethylating reagent 5-aza-dC; MDAMB-231 cells did not express ERα mRNA by this treatment. ERBF-1 is highly expressed in MCF-7 and T-47-D cells, much less so in ZR-75-1 cells and not at all in MDA-MB-231 and BT-20 cells (8). In fact, our MDA-MB-231 cells showed little luciferase activity in a reporter assay using unmethylated pGL3-ProB. Thus, loss of ERBF-1 in MDA-MB-231 cells may be critical for suppression of ERα gene. This result differed from that reported by Ferguson et al. (20). However, expression of ERα mRNA in MDA-MB-231 cells was not affected by demethylating treatment, even when the number of PCR cycles and the concentration of 5-aza-dC were increased. This discrepancy in MDA-MB-231 cells may be due to a change in the nature of cells during the wide distribution. In ERα-negative BT-20 cells, expression of ERα total and type A mRNA was induced. It has been reported that ERα mRNA is very faintly expressed in BT-20 cells and is transcribed from promoter A alone (33). It has also been reported that estrogen receptor factor 1 (ERF-1), a member of the AP2 transcription factor family, is important for transcriptional regulation from promoter A in T-47-D cells (34–36); ERF-1 expression has been found in MCF-7, T-47-D and BT-20 cells, but not in MDA-MB-231 cells (34). Considering all of these results, promoter A-specific induction of ERα gene expression caused by the demethylating reagent may be explained by the fact that ERF-1 is present in BT-20 cells, while ERBF-1 is not. In vitro methylation of promoter regions decreased promoter activity to ~50–60% of that of unmethylated ones. Thus, we found that methylation of the promoters of the ERα gene directly suppressed its transcription, although not completely. Recently, it has been shown that CpG methylation exerts its inhibitory effect indirectly through recruitment of methylated CpG-binding proteins such as MeCP1 and MeCP2 (37,38), with MeCP2 being more abundant than MeCP1 in cells and more tightly bound to DNA containing methyl-CpG pairs in the nucleus (39). Furthermore, it has been reported that MeCP2 2199

T.Yoshida et al.

forms a co-repressor complex containing a transcriptional suppressor, mSin3, and histone deacetylases, and that this complex indirectly suppresses transcription by modulating chromatin structure (40). Thus, not only methylation of the promoter region but also other mechanisms, such as alteration of the chromatin structure of the gene, would be required for complete suppression of transcription. In conclusion, we have shown that that the methylation status of the ERα gene promoters correlated well with the suppression of the expression of this gene in human breast cancer tissue as well as in human breast cancer cell lines. Methylation of ERα gene promoters directly affected ERα activity in breast cancer cells, though other mechanisms, such as loss of the critical transcriptional factors ERBF-1 and ERF-1, may also be at work in some cases.

Acknowledgements We thank Ms Yoko Sekine for her excellent technical assistance, Dr Osamu Goto for his helpful support, Dr Masakazu Toi for providing breast cancer cell lines and Dr Atsuo Kuramasu for his valuable suggestions and advice. We are grateful to Dr Etsuro Ogata for his valuable suggestions and advice. This study was supported in part by Grants-in-Aid for Scientific Research on Priority Areas, Cancer, from the Ministry of Education, Science, Sports and Culture, Japan, for Scientific Reserch Expenses for Health and Welfare Programs and the Foundation for the Promotion of Cancer Research, and by Second-Term Comprehensive 10-Year Strategy for Cancer Control.

References 1. McGuire,W.L. (1978) Hormone receptors: their role in predicting prognosis and response to endocrine therapy. Semin. Oncol., 5, 428–433. 2. Russo,J. and Russo,I.H. (1994) Toward a physiological approach to breast cancer prevention. Cancer Epidemiol. Biomarkers Prev., 3, 353–364. 3. Lapidus,R.G., Nass,S.J. and Davidson,N.E. (1998) The loss of estrogen and progesterone receptor gene expression in human breast cancer. J. Mamm. Gland Biol. Neoplasia, 3, 85–94. 4. Keaveney,M., Klug,J., Dawson,M.T., Nestor,P.V., Neilan,J.G., Forde,R.,C. and Gannon,F. (1991) Evidence for a previously unidentified upstream exon in the human estrogen receptor gene. J. Mol. Endocrinol., 6, 111–115. 5. Piva,R., Bianchi,N., Aguiari,G.R. and Senno,L.D. (1993) Sequencing of an RNA transcript of the human estrogen receptor gene: evidence for a new transcriptional event. J. Steroid Biochem. Mol. Biol., 46, 531–538. 6. Grandien,K.F.H., Berkenstam,A., Nilsson,S. and Gustafsson,J.-A. (1993) Localization of DNase hypersensitive sites in the human oestrogen receptor gene correlates with the transcriptional activity of two differentially used promoters. J. Mol. Endocrinol., 10, 269–277. 7. Hayashi,S.-I., Imai,K., Suga,K., Kurihara,T., Higashi,Y. and Nakachi,K. (1997) Two promoters in expression of estrogen receptor messenger RNA in human breast cancer. Carcinogenesis, 18, 459–464. 8. Tanimoto,K., Eguchi,H., Yoshida,T., Hajiro-Nakanishi,K. and Hayashi, S.-I. (1999) Regulation of estrogen receptor gene mediated by promoter B responsible for its enhanced expression in human breast cancer. Nucleic Acids Res., 27, 903–909. 9. Bird,A.P. (1986) CpG-rich islands and function of DNA methylation. Nature, 321, 209–213. 10. Counts,J.L. and Goodman,J.I. (1995) Alterations in DNA methylation may play a variety of roles in carcinogenesis. Cell, 83, 13–15. 11. Hiraguri,S., Godfrey,T., Nakamura,H. et al. (1998) Mechanisms of inactivation of E-cadherin in breast cancer cell lines. Cancer Res., 58, 1972–1977. 12. Dobrovic,A. and Simpfendorfer,D. (1997) Methylation of the BRCA1 gene in sporadic breast cancer. Cancer Res., 57, 3347–3350. 13. Catteau,A., Harris,W.H., Xu,C.-F. and Solomon,E. (1999) Methylation of the BRCA1 promoter region in sporadic breast and ovarian cancer: correlation with disease characteristics. Oncogene, 18, 1957–1965. 14. Fujii,H., Biel,M.A., Zhou,W., Weitzman,S.A., Baylin,S.B. and Gabrielson,E. (1998) Methylation of the HIC-1 candidate tumor suppressor gene in human breast cancer. Oncogene, 16, 2159–2164. 15. Lapidus,R.G., Ferguson,A.T., Ottaviano,Y.L., Parl,F.F., Smith,H.S., Weitzman,S.A., Baylin,S.B., Issa,J.-P. and Davidson,N.E. (1996)

2200

Methylation of estrogen and progesterone receptor gene 5⬘ CpG islands correlates with lack of estrogen and progesterone receptor gene expression in breast tumors. Clin. Cancer Res., 2, 805–810. 16. Pilat,M.J., Schwab,E.D., Yao,K.-L. and Pienta,K.J. (1998) Examination of the DNA methylation properties in nontumorigenic and tumorigenic breast epithelial cell lines. Anticancer Res., 18, 2575–2582. 17. Ottaviano,Y.L., Issa,J.P., Parl,F.F., Smith,H.S., Baylin,S.B. and Davidson,N.E. (1994) Methylation of the estrogen receptor CpG island marks loss of estrogen receptor expression in human breast cancer cells. Cancer Res., 54, 2552–2555. 18. Chen,Z., Ko,A., Yang,J. and Jordan,V.C. (1998) Methylation of CpG island is not a ubiquitous mechanism for the loss of oestrogen receptor in breast cancer cells. Br. J. Cancer, 77, 181–185. 19. Lapidus,R.G., Nass,S.J., Butash,K.A., Parl,F.F., Weitzman,S.A., Graff,J.G, Herman,J.G. and Davidson,N.E. (1998) Mapping of ER gene CpG island methylation by methylation-specific polymerase chain reaction. Cancer Res., 58, 2515–2519. 20. Ferguson,A.T., Lapidus,R.G., Baylin,S.B. and Davidson,N.E. (1995) Demethylation of the estrogen receptor gene in estrogen receptor-negative breast cancer cells can reactive estrogen receptor gene expression. Cancer Res., 55, 2279–2283. 21. Chomczynski,P. and Sacchi,N. (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate–phenol–chloroform extraction. Anal. Biochem., 162, 156–159. 22. Kuramasu,A., Saito,S., Suzuki,S., Watanabe,T. and Ohtsu,H. (1998) Mast cell-/basophil-specific transcriptional regulation of human L-histidine decarboxylase gene by CpG methylation in the promoter region. J. Biol. Chem., 273, 31607–31614. 23. Clark,S.J., Harrison,J., Paul,C.L. and Formmer,M. (1994) High sensitivity mapping of methylated cytosines. Nucleic Acids Res., 22, 2990–2997. 24. Frommer,M., McDonald,L.E., Millar,D.S., Collis,C.M., Wat,F., Grigg,G.W., Molley,P.L. and Paul,C.L. (1992) A genomic sequencing protocol that yields a positive display of 5-methylcytosine residues in individual DNA strands. Proc. Natl Acad. Sci. USA, 89, 1827–1831. 25. Herman,J.G., Graff,J.R., Myo¨ ha¨ nen,S., Nelkin,B.D. and Baylin,S.B. (1996) Methylation specific PCR: a novel PCR assay for methylation status of CpG islands. Proc. Natl Acad. Sci. USA, 93, 9821–9826. 26. Tanaka,H., Shimada,Y., Harada,H., Shinoda,M., Hatooka,S., Imamura,M. and Ishizaki,K. (1998) Methylation of the 5⬘ CpG Island of the FHIT gene is closely associated with transcriptional inactivation in esophageal squamous cell carcinomas. Cancer Res., 58, 3429–3434. 27. Eguchi,H., Hayashi,S.-I., Watanabe,J., Gotoh,O. and Kawajiri,K. (1994) Molecular cloning of the human AH receptor gene promoter. Biochem. Biophys. Res. Commun., 203, 615–622. 28. Hsieh,C.L. (1997) Stability of patch methylation and its impact in regions of transcriptional initiation and elongation. Mol. Cell. Biol., 17, 5897–5904. 29. Bender,C.M., Pao,M.M. and Jones,P.A. (1998) Inhibition of DNA methylation by 5-Aza-2⬘-deoxycytidine suppresses the growth of human tumor cell lines. Cancer Res., 58, 95–101. 30. El-Deiry,W.S., Nelkin,B.D., Celano,P., Yen,R.W., Falco,J.P., Hamilton,S.R. and Baylin,S.B. (1991) High expression of the DNA methyltransferase gene characterizes human neoplastic cells and progression stages of colon cancer. Proc. Natl Acad. Sci. USA, 88, 3470–3474. 31. Issa,J.-P., Vertino,P.M., Wu,J., Sazawal,S., Celano,P., Nelkin,B.D., Hamilton,S.R. and Baylin,S.B. (1993) Increased cytosine DNA– methyltransferase activity during colon cancer progression. J. Natl Cancer Inst., 85, 1235–1240. 32. Rein,T., DePamphilis,M.L. and Zorbas,H. (1998) Identifying 5methylcytosine and related modifications in DNA genomes. Nucleic Acids Res., 26, 2255–2264. 33. Weigel,R.J., Crooks,D.L., Iglehart,J.D. and de Coninck,E.C. (1995) Quantitative analysis of the transcriptional start sites of estrogen receptor in breast carcinoma. Cell Growth Differ., 6, 707–711. 34. de Coninck,E.C., McPherson,L.A. and Weigel,R.J. (1995) Transcriptional regulation of estrogen receptor in breast carcinomas. Mol. Cell. Biol., 15, 2191–2196. 35. McPherson,L.A., Baichwal,V.R. and Weigel,R.J. (1997) Identification of ERF-1 as a member of the AP2 transcription factor family. Proc. Natl Acad. Sci. USA, 94, 4342–4347. 36. Tang,Z., Treilleux,I. and Brown,M. (1997) A transcriptional enhancer required for the differential expression of the human estrogen receptor in breast cancers. Mol. Cell. Biol., 17, 1274–1280. 37. Meehan,R.R., Lewis,J.D., McKay,S., Kleiner,E.L. and Bird,A.P. (1989) Identification of a mammalian protein that binds specifically to DNA containing methylated CpGs. Cell, 58, 499–507.

Mechanisms of loss of ERα gene expression 38. Lewis,J.D., Meehan,R.R., Henzel,W.J., Maurer-Fogy,I., Jeppesen,P., Klein,F. and Bird,A.P. (1992) Purification, sequence, and cellular localization of a novel chromosomal protein that binds to methylated DNA. Cell, 69, 905–914. 39. Meehan,R.R., Lewis,J.D. and Bird,A.P. (1992) Characterization of MeCP2, a vertebrate DNA binding protein with affinity for methylated DNA. Nucleic Acids Res., 20, 5085–5092.

40. Nan,X., Ng,H.H., Johnson,C.A., Laherty,C.D., Turner,B.M., Elisenman, R.N. and Bird,A.P. (1998) Transcriptional repression by the methyl-CpGbinding protein MeCP2 involves a histone deacetylase complex. Nature, 393, 386–389. Received January 6, 2000; revised August 7, 2000; accepted August 18, 2000

2201