First found in equine urine(Reference Marrian and Haslewood1), equol is a non-steroidal oestrogen. Many years after its discovery, it was found that the soya isoflavone daidzein was a precursor to equol(Reference Axelson, Kirk and Farrant2), and that soya consumption increased the excretion of equol in some, but not all, adults. Studies have shown that gut microflora was responsible for the conversion of daidzein to S-equol(Reference Setchell, Borriello and Hulme3). More recently, particular bacteria capable of this conversion were even isolated(Reference Setchell and Clerici4). Multiple studies have shown that equol producers were more frequent in Asian countries than in Western countries, which led researchers to ask themselves whether particular diets would not favour equol-producing microflora(Reference Setchell and Cole5). Equol is a chiral molecule and two forms can coexist: R- and S-equol. Distinction between these two forms and purification of one of them is complex, so many studies have worked on the effects of racemic equol. Today, it has been shown that only S-equol is synthesised by gut bacteria(Reference Setchell, Brzezinski and Brown6) and S-equol is commercially available, leading to studies on the effect of S-equol alone.
Breast cancer is the most frequent cancer in women, with 1·38 million new cases and 458 000 deaths in 2008(Reference Jemal, Bray and Center7). The incidence of breast cancer is high in Western countries, and low in Asia. This difference has been attributed, at least in part, to the Asian traditional diet, containing larger amounts of soya than the Western diet. Particular chemicals in soya, namely phyto-oestrogens, are supposed to have protective effects on breast cancer, mainly because of their similarity of structure with 17-β-oestradiol, the natural human oestrogen, allowing them to bind and activate oestrogen receptors (ER)(Reference Kuiper, Carlsson and Grandien8–Reference Pfitscher, Reiter and Jungbauer10), with, contrarily to 17-β-estradiol, a higher affinity for ERβ(Reference Kuiper, Lemmen and Carlsson9, Reference Takeuchi, Takahashi and Sawada11). This is also the case for S-equol(Reference Setchell, Brzezinski and Brown6, Reference Muthyala, Ju and Sheng12). More recently, special attention has been paid to S-equol, as some studies have shown that equol had a greater affinity for ER than its precursor, daidzein(Reference Muthyala, Ju and Sheng12).
Many studies have worked on the potential protective effect of soya over breast cancer, but mixed results have been found(Reference Satih, Rabiau and Bignon13). Studies on the effects of S-equol on breast cancer risk have led to the same mixed results. As epigenetic mechanisms are implied in cancer, a growing number of studies have investigated the effect of soya phyto-oestrogens on those mechanisms, particularly DNA methylation(Reference Fang, Jin and Wang14). In normal tissues, oncogenes and repeated sequences are globally methylated while oncosuppressors are hypomethylated, particularly at the level of cytosine phosphate guanine (CpG) islands found in the promoters of these genes(Reference Das and Singal15). In cancer, an inversion of this methylation profile is found, so it has been stated that soya phyto-oestrogens could have protective effects on cancer by reverting this methylation profile. Moreover, protective effects of breast cancer are observed in women consuming moderate amounts of soya since their childhood but not in women starting soya consumption after the menopause(Reference Guha, Kwan and Quesenberry16, Reference Korde, Wu and Fears17). This observation could be the result of a protective epigenetic effect with expression changes of genes implicated in the early events of carcinogenesis. Some studies have shown a demethylating action of genistein and daidzein on oncosuppressors in cancer cells(Reference Fang, Jin and Wang14). To our knowledge, only one study showed an effect of equol on DNA methylation: Lyn-Cook et al. (Reference Lyn-Cook, Blann and Payne18) showed that high doses of equol caused the hypermethylation of the c-H-ras proto-oncogene in the pancreas cells of neonatal rats. Here, we investigated the effects of equol on the methylation of two major breast cancer oncosuppressors: BRCA1 and BRCA2. The breast cancer susceptibility gene 1 (BRCA1) and the breast cancer susceptibility gene 2 (BRCA2) are the major high-penetrance genes in which mutations increase susceptibility to breast cancer. Mutations in these genes account together for 2–3 % of all breast cancers and about 30–40 % of all familial breast cancers(Reference Wooster and Weber19). The BRCA1 gene is located on chromosome 17q12-21. BRCA1 is involved in many transcriptional activation or transcriptional repression processes(Reference Cable, Wilson and Calzone20). It also plays a role in apoptosis, genomic stability maintenance, and DNA recognition and repair(Reference Jhanwar-Uniyal21). The BRCA2 gene is located on chromosome 13q12-13. The gene codes for proteins involved in DNA repair, cell-cycle control and transcription(Reference Kerr and Ashworth22), and may have a function in the terminal differentiation of breast epithelial cells(Reference Vidarsson, Mikaelsdottir and Rafnar23).
Although somatic mutations of these genes are rarely found in sporadic breast cancers(Reference Kerr and Ashworth22–Reference Venkitaraman26), methylation of the promoter of BRCA1 coupled with a decrease in mRNA(Reference Rice, Ozcelik and Maxeiner27) or lower BRCA1 protein(Reference Matros, Wang and Lodeiro28, Reference Tapia, Smalley and Kohen29) can be found. BRCA2 promoter methylation has also been reported in sporadic breast cancer cases(Reference Cucer, Taheri and Ok30).
As a growing number of studies have shown the effects of soya phyto-oestrogens on DNA methylation(Reference Lyn-Cook, Blann and Payne18, Reference Day, Bauer and DesBordes31–Reference Majid, Dar and Ahmad35), the protective effects of soya isoflavones on breast cancer could be due, at least in part, to an effect on DNA methylation.
We undertook the present study to examine changes in DNA methylation of the CpG islands in the promoters of BRCA1 and BRCA2 in breast cancer cells following exposure to S-equol at physiological doses during 3 weeks.
Materials and methods
Cell lines
MCF-7 and MDA-MB-231 breast tumour cell lines came from a pleural effusion of patients with invasive breast carcinoma(Reference Cailleau, Young and Olive36, Reference Soule, Vazguez and Long37). The MCF-10a cell line was established from the breast tissue of patients with fibrocystic breast disease(Reference Soule, Maloney and Wolman38). All three human cell lines were provided by the American Type Culture Collection. MCF-7 were cultured in Roswell Park Memorial Institute (RPMI)-1640 medium supplemented with 2 mm-l-glutamine (Invitrogen), gentamycin (20 μg/ml; Panpharma), 10 % fetal bovine serum (Invitrogen) and insulin (1·4 μg/ml; Novo Nordisk) in a humidified atmosphere at 37°C containing 5 % CO2. This cell line has a positive ER status (ERα+/ERβ+). MCF-10a cells were maintained in Dulbecco's modified Eagle's medium F12 (Invitrogen) containing 10 % horse serum (Invitrogen), 2 mm-l-glutamine, gentamycin (20 μg/ml; Panpharma), epidermal growth factor (20 ng/ml; Sigma), cholera toxin (100 ng/ml; Sigma), insulin (10 μg/ml; Novo Nordisk) and hydrocortisone (0·5 μg/ml; Sigma) held at 37°C with 5 % CO2. This cell line has a negative oestrogen receptor status (ERα − /ERβ − ). MDA-MB-231 cells were grown in Leibovitz L-15 medium with 15 % fetal bovine serum (Invitrogen), gentamycin (20 μg/ml; Panpharma) and 2 mm-l-glutamine in a 37°C humidified atmosphere without CO2. This cell line has a negative ER status (ERα − /ERβ+).
The ER status of the three cell lines has previously been confirmed by immunohistochemistry(Reference Vissac-Sabatier, Bignon and Bernard-Gallon39).
Cell treatments
Cells (1 × 106 per T75 flask) were seeded in the medium and treated with 2 μm-S-equol provided by the ENITA Unité Micronutriments-Reproduction-Santé and dissolved in dimethyl sulfoxide. As controls, the cell lines were also conditioned in the medium with the solvent dimethyl sulfoxide.
During the 3 weeks, each 48 h and just before 80 % confluence, cells were trypsinised and cell number scored on a Malassez cell using Trypan blue, and then they were passed into three flasks and the treatments were added again.
DNA extraction
DNA was extracted using Millipore's non-organic DNA extraction kit as follows: after recovering the cells, 9 ml of wash buffer 1 × were added to resuspend the pellet. After 15 min of incubation at room temperature, the cells were centrifuged at 1000 g for 20 min. The supernatant was discarded and the cells were resuspended in 3 ml of suspension buffer I 1 × . Lysis buffer I (800 μl) and 50 μl of protein-digesting enzyme were added to the suspension. The samples were incubated for 2 h at 50°C. After adding 1 ml of a protein-precipitating agent, a 15 min centrifugation at 1000 g was carried out. The supernatant thus obtained was mixed with two volumes of absolute ethanol. The precipitated DNA was recovered using an inoculating needle, dried for 5 min at room temperature, and dipped in 5 ml of 70 % ethanol. DNA was resuspended in 300 μl of suspension buffer II. After vortexing them for 5 min, the samples were left in incubation overnight at 50°C. The quantity of DNA collected as well as the quality of the extraction was then determined by spectrometry using a NanoDrop™ 8-sample spectrophotometer (ND-8000, NanoDrop Technologies®).
Bisulfite treatment and quantitative analysis of methylated alleles
Conversion of unmethylated cytosines to uracil(Reference Frommer, McDonald and Millar40), leaving methylated cytosines unaltered, was achieved using the methylSEQr™ Bisulfite Modification Kit (Applied Biosystems) following the manufacturer's instructions. We measured the methylation of oncosuppressor promoters with the real-time PCR-based quantitative analysis of methylated alleles (QAMA) assay previously described by Zeschnigk et al. (Reference Zeschnigk, Bohringer and Price41) and adapted here by Bosviel et al. (Reference Bosviel, Michard and Lavediaux42). PCR was performed using a ninety-six-well optical tray with optical adhesive film at a final reaction volume of 20 μl. Samples contained 10 μl of TaqMan® Universal PCR Master Mix II, No AmpErase® UNG (uracil-N-glycosylase), 8 μl of bisulfite-treated DNA, an additional 5 U of FastStart Taq DNA Polymerase (Roche), 2·5 μm each of the primers and 150 nm of the fluorescently labelled methylated and unmethylated BRCA1 or methylated and unmethylated BRCA2 probes. Initial denaturation at 95°C for 10 min to activate DNA polymerase was followed by forty cycles of denaturation at 95°C for 15 s and annealing and extension at 60°C for 1 min (7900HT, Real-Time PCR System; Applied Biosystems). Primer and probe sequences were selected with the help of Primer Express software (ABI). PCR primers were designed to amplify the bisulfite-converted sense strand of the CpG island BRCA1 promoter sequence or the antisense strand of the CpG island BRCA2 promoter sequence, lacking any known nucleotide polymorphisms. The software designs primers with a melting temperature (T m) of 58–60°C and probes with a T m value of 68–69°C. The T m of both primers should be equal. The amplicon sizes were 79 bp for BRCA1 (located at chromosome 17: 41278096–41278175 on the Ensembl GRCh37/hg19 assembly) and 87 bp for BRCA2 (located at chromosome 13: 32889345–32889428). Primer and probe sequences are as follows: for BRCA1, forward primer – 5′-GGAGTTTGGGGTAAGTAGTTTTGTAAG-3′; reverse primer – 5′-TTCCCCTACCCCAAACAAATT-3′; methylated probe – 5′-VIC-ACTACGTCCCCGCAAA-MGBNFQ-3′; unmethylated probe – 5′-6FAM-ACTACATCCCCACAAAC-MGBNFQ-3′; for BRCA2, forward primer – 5′-GTTGGAGTAAAAAGAAAGGGATGG-3′; reverse primer – 5′-CCTTAAAAATCCCAAACCACCC-3′; methylated probe – 5′-VIC-AAACCGCCCCTATAC-MGBNFQ-3′; unmethylated probe – 5′-6FAM-AAAACCACCCCTATACC-MGBNFQ-3′. The primer binding sites lack CpG dinucleotides and, therefore, the nucleotide sequences in the methylated and unmethylated DNA are identical after the bisulfite treatment. Consequently, it is possible to amplify both alleles in the same reaction tube with one primer pair. Methylation discrimination occurs during probe hybridisation by the use of two different MGB Taqman® probes. The binding site of the BRCA1 and BRCA2 MGB Taqman® probes both cover two CpG dinucleotides. We used a VIC-labelled MGB Taqman® probe that specifically hybridises to the sequence derived from the methylated allele, and a 6-carboxyfluorescein (FAM)-labelled MGB Taqman® probe that binds to the sequence generated from the unmethylated allele. The amount of FAM and VIC fluorescence released during the PCR was measured by the real-time PCR system and is directly proportional to the amount of the PCR product generated. The cycle number at which the fluorescence signal crosses a detection threshold is referred to as C T and the difference of both C T values within a sample (ΔC T) is calculated (ΔC T = C T − FAM − C T − VIC). All samples were measured in duplicate using the mean for further analysis. For a precise quantification of the ratio of methylated:unmethylated alleles, the ΔC T value is determined and compared with a standard curve that exhibits a sigmoid shape with a linear part in the range of 10–90 % of methylated DNA (Fig. 1). To set up the curve, we mixed bisulfite-treated and methylated control human DNA (EpiTect, ref. 59 655; Qiagen) with defined ratios of bisulfite-treated and unmethylated control human DNA (EpiTect, ref. 59 665; Qiagen) implemented in each run. From this, we deduced an algorithm to calculate the methylation ratio of an unknown sample from its ΔC T value by the Mathematica software package version 5.2 from Wolfram Research (http://www.wolfram.com). Student's t test was performed using the data obtained with QAMA, and P < 0·05 was considered to be statistically significant compared with the cells treated with the solvent dimethyl sulfoxide.
Fig. 1 Example of a standard curve for breast cancer susceptibility gene 2 (BRCA2) quantitative analysis of methylated alleles. ΔC T values obtained for standard samples were plotted against their defined methylation ratio. The methylation ratio of the tested samples was found by plotting the ΔC T values obtained onto this standard curve. In the case where only one fluorescence signal crossed the threshold, indicating a relative absence of the opposite target, the methylation percentage was set to 0 or 100 %, depending on the nature of the fluorescence.
Western blotting
Proteins were extracted from the cells with lysis buffer containing 20 mm-Tris (pH 8), 50 mm-EDTA, 0·8 % NaCl, 0·1 % Triton X-100 and 1 % glycerol. Protease inhibitors (1 %, Protease Inhibitor Cocktail; Sigma) and phosphatase inhibitors (1 %, Phosphatase Inhibitor Cocktail 2; Sigma) were added to the basic buffer extemporaneously (1 % each). Then, 50 μg proteins were electrophoresed on a SDS-polyacrylamide gel and transferred onto a nitrocellulose membrane. After 1 h blocking in Tris-Buffered Saline Tween 0·1 % containing 5 % milk, membranes were incubated overnight at 4°C with anti-BRCA1 (1:150 Mouse (Ab-1); Calbiochem), anti-BRCA2 (1:50 Rabbit (H-300); Santa Cruz Biotechnology®) or anti-actin (1:120,000 Mouse (Ab-1); Calbiochem) antibodies. The membranes were then washed three times in Tris-buffered saline Tween and incubated for 1 h with alkaline phosphatase-conjugated secondary antibody (1:2000 goat anti-mouse IgG (H&L) AP conjugate or 1:2000 goat anti-rabbit IgG (Fc) AP conjugate; Promega). Detection was then performed with the Western Blue detection system (Promega). Relative quantification of immunoblotted proteins was achieved using Quantity One software (Bio-Rad) with the local background subtraction method. A ratio between the intensity of the protein of interest and a reference protein (actin) was then calculated. The relative ratio was then calculated between each condition and the reference condition (dimethyl sulfoxide-treated cells).
Immunohistochemistry
For immunohistochemical analysis, 4 μm alcohol–formalin–acetic acid-fixed and paraffin-embedded sections of MCF-7, MDA-MB-231 and MCF-10a cell pellets were cut using a microtome. They were mounted on silanised glass slides (Starfrost; Duiven) and dried overnight at 37°C. Slides were processed on an automated Benchmark XT immunohistochemical instrument (Ventana). In particular, sections were deparaffinised and rehydrated using EZ Prep (Ventana), and heat-induced antigen retrieval using CC1 (Ventana) was performed for 30 min. The slides were then incubated at 37°C for 44 min with anti-BRCA1 (1:20 mouse (8F7); GeneTeX®) or anti-BRCA2 (1:20 mouse (Ab-1); Calbiochem®) primary antibodies. For detection, we used the UltraView universal DAB detection kit (Ventana). Signal was amplified using the Ventana amplification kit. The slides were then counterstained with haematoxylin for 3 min, rinsed in distilled water and coverslipped with an aqueous Faramount mounting media (DAKO). The primary polyclonal antibody was omitted and replaced with PBS as a negative control.
Results
Effect of S-equol on BRCA1 and BRCA2 CpG promoter methylation
QAMA was used to study the effects of S-equol on BRCA1 and BRCA2 CpG islands. We showed a significant decrease in the methylation of the CpG islands in the promoters of BRCA1 and BRCA2 following the 2 μm-S-equol treatment during 3 weeks in MDA-MB-231 and MCF-7 cells compared with the control (Fig. 2(a) and (b), respectively). This demethylation was not significant in MCF-10a cells (Fig. 2(c)).
Fig. 2 Breast cancer susceptibility genes (BRCA1 and BRCA2) methylation in (a) MDA-MB-231, (b) MCF-7 or (c) MCF-10a cells treated for 3 weeks with 2 μm-S-equol compared with the dimethyl sulfoxide (DMSO) control. BRCA1 and BRCA2 methylation were decreased significantly following the S-equol treatment in the MDA-MB-231 and MCF-7 cells (P < 0·05).
Effect of S-equol on BRCA1 and BRCA2 protein expression
Western blotting was used to study the effects of S-equol on BRCA1 and BRCA2 protein expression. We showed an increase in BRCA1 and BRCA2 proteins following the 2 μm-S-equol treatment for 3 weeks in MCF-7, MDA-MB-231 and MCF-10a cell lines (Fig. 3(a)–(c), respectively). An extensive increase in BRCA1 staining was found by immunohistochemistry in the nuclei, the cytoplasm and nucleoli in MCF-7, MDA-MB-231 and MCF-10a cell lines after 2 μm-S-equol exposure for 3 weeks. For BRCA2, the increase in staining was exhibited preferentially in the cytoplasm (Fig. 4). The results of immunohistochemistry are compiled in Table 1.
Fig. 3 Western blots with breast cancer susceptibility genes (BRCA1 and BRCA2) and actin proteins extracted from (a) MCF-7, (b) MDA-MB-231 and (c) MCF-10a cells. Ratios shown correspond to relative ratios of optical densities of the bands (measured with Quantity One software; Bio-Rad) from interest proteins over actin, relatively to the control condition (dimethyl sulfoxide (DMSO)-treated cells). Cells were treated for 3 weeks with DMSO (control condition) or 2 μm-S-equol (E).
Fig. 4 Immunoperoxidase staining of MDA-MB-231 human breast cancer cell lines on paraffin-embedded sections (60 × ). (a) Cytoplasmic, nuclear and nucleolar staining were exhibited with 1:20 breast cancer susceptibility gene 2 (BRCA2) monoclonal antibody (Ab1), shown by arrowheads in untreated cells. (b) The BRCA2 staining after 2 μm-S-equol treatment was considerably increased. N, nucleus; Cyt, cytoplasm; NU, nucleoli.
Table 1 Effects of S-equol on breast cancer susceptibility genes (BRCA1 and BRCA2) expression in MCF-7, MDA-MB-231 and MCF-10a cell lines*
Cyt, cytoplasm; N, nucleus; Nu, nucleoli; DMSO, dimethyl sulfoxide.
* Cells were treated during 3 weeks with S-equol (2 μm). Cells were also treated with DMSO, the solvent in which S-equol was diluted. Then, the cells were immunostained with MoAb anti-BRCA1 (8F7) or anti-BRCA2 (Ab-1). Staining: negative ( − ); intermediate (+/ − ); less intensive (+); intensive (++); very intensive (+++).
Discussion
A growing number of studies have revealed the importance of DNA methylation in cancer, with a global hypomethylation of DNA and the hypermethylation of CpG islands of oncosuppressors, leading to chromosomic instability and loss of the expression of oncosuppressors. In breast cancer, hypermethylation of the BRCA1 and BRCA2 genes has been found, associated with a decrease in mRNA expression for BRCA1. S-equol, an intestinal bacterial metabolite of daidzein, is a putative protective molecule for breast cancer. The present study sustains the idea that this protective effect could pass through epigenetic modulation of BRCA1 and BRCA2 expression. The mechanism for this effect is not yet clearly known, although studies have shown that S-equol can bind and activate ER.
As more and more studies have shown the effects of soya phyto-oestrogens on DNA methylation, we decided to study the effects of S-equol in breast cancer cell lines on BRCA1 and BRCA2 methylation and consequent protein expression. We studied the effects of S-equol on the expression of the BRCA1 and BRCA2 genes that interact together in two human breast cancer cell lines (MCF-7 and MDA-MB-231) and in a fibrocystic cell line (MCF-10a). We chose an exposure of 3 weeks to S-equol, because this treatment has been shown to increase the number of cells blocked in the S phase(Reference Choi and Kim43), and BRCA1 and BRCA2 reach their maximal level in the late G1 and S phases in normal and tumour-derived breast epithelial cells(Reference Bertwistle and Ashworth44).
An important point in the design of the present study is the use of physiological doses of S-equol, in the same order of magnitude as plasma concentrations found in post-menopausal women(Reference Mathey, Lamothe and Coxam45, Reference Bennetau-Pelissero, Arnal-Schnebelen and Lamothe46). Long exposures were carried out to point out an eventually weak effect due to the use of such doses. The effects observed in the present study thus have better chances to be representative of real-life exposure.
We provide evidence that S-equol demethylates the promoters of the BRCA1 and BRCA2 genes in MDA-MB-231 and MCF-7 breast cancer cell lines, but not in the MCF-10a cell line. We also showed an increase in the expression of the BRCA1 and BRCA2 proteins in the studied cell lines following the S-equol treatment. The fact that demethylation occurred in MDA-MB-231 and MCF-7 cell lines but not in the MCF-10a cell line whereas protein expression increased in all the three cell lines could suggest that DNA methylation was not the only mechanism regulating BRCA1 and BRCA2 expression that can be modulated by S-equol, and thus studies on histone mark status following the S-equol treatment could be interesting. Indeed, many studies have shown the effects of soya phyto-oestrogens on histone modifications, and S-equol could have similar effects(Reference Majid, Dar and Ahmad35, Reference Jawaid, Crane and Nowers47–Reference Majid, Kikuno and Nelles50). Hong et al. (Reference Hong, Nakagawa and Pan51) also showed that equol stimulates ER-mediated histone acetyl transferase activity. ER status and, more particularly, ERβ status may play a role in the action of S-equol on DNA methylation, as the MCF-10a cell line lacks the ERβ receptor. To our knowledge, only one study has reported an effect of equol on DNA methylation, showing a rise in the methylation of the proto-oncogene c-H-ras in rat pancreatic cells(Reference Lyn-Cook, Blann and Payne18), while more data are found for other soya phyto-oestrogens(Reference Fang, Jin and Wang14, Reference Day, Bauer and DesBordes31, Reference Dolinoy, Weidman and Waterland32, Reference King-Batoon, Leszczynska and Klein34, Reference Majid, Dar and Ahmad35, Reference Majid, Dar and Shahryari49, Reference Berner, Aumuller and Gnauck52–Reference Wang and Chen58). Such effects on oncosuppressors could help prevent cancer by restoring their expression similar to the protein expression of BRCA1 and BRCA2 in the present experiment. Studies on whether this demethylating effect is limited to the CpG islands in the promoter of oncosuppressors or whether it also acts on the methylation of other CG sites could be interesting, as demethylating effects on global methylation and, more particularly, repeated elements or transposable elements would be a counter effect for cancer prevention(Reference Watanabe and Maekawa59).
In summary, the present study shows that S-equol has a demethylating effect on the CpG islands in the promoters of BRCA1 and BRCA2 genes. This effect might be linked with the presence of ER but the increase in subsequent protein expression is independent of this parameter. Thus, we suppose that other mechanisms can also be implied, such as effects on histone modifications.
Acknowledgements
We thank Nicolas Sonnier and Christelle Picard for helpful technical assistance. R. B. is the recipient of a grant from the Auvergne Regional Council/CPER 2008+ FEDER no. 32316 – 0930FDBG – 106NL. R. B., P. D., Y.-J. B. and D. B.-G. contributed to the experimental design. R. B. and J. D. were responsible for performing the experiments. R. B., J. D. and P. D. contributed to the data analysis. R. B. and D. B.-G. contributed to manuscript preparation. The authors declare that they have no conflict of interest.
