Evidence of epigenetic changes affecting the chromatin state of the retinoic acid receptor β2 promoter in breast cancer cells


Retinoic acid (RA)-resistance in breast cancer cells has been associated with irreversible loss of retinoic acid receptor β, RARβ, gene expression. Search of the causes affecting RARβ gene activity has been oriented at identifying possible differences either at the level of one of the RARβ promoters, RARβ2, or at regulatory factors. We hypothesized that loss of RARβ2 activity occurs as a result of multiple factors, including epigenetic modifications, which can pattern RARβ2 chromatin state. Using methylation-specific PCR, we found hypermethylation at RARβ2 in a significant proportion of both breast cancer cell lines and primary breast tumors. Treatment of cells with a methylated RARβ2 promoter, by means of the DNA methyltransferase inhibitor 5-Aza-2′-deoxycytidine (5-Aza-CdR), led to demethylation within RARβ2 and expression of RARβ indicating that DNA methylation is at least one factor, contributing to RARβ inactivity. However, identically methylated promoters can differentially respond to RA, suggesting that RARβ2 activity may be associated to different repressive chromatin states. This supposition is supported by the finding that the more stable repressive RARβ2 state in the RA-resistant MDA-MB-231 cell line can be alleviated by the HDAC inhibitor, trichostatin A (TSA), with restoration of RA-induced RARβ transcription. Thus, chromatin-remodeling drugs might provide a strategy to restore RARβ activity, and help to overcome the hurdle of RA-resistance in breast cancer.


Retinoic acid (RA) controls fundamental developmental processes, induces terminal differentiation of myeloid progenitors and suppresses cancer and cell growth (Smith et al., 1992; Gudas et al., 1994). RA activity is mediated by nuclear receptors, the retinoic acid receptors, RARs, that act as RA-dependent transcriptional activators in their heterodimeric forms with retinoid X receptors, RXRs (Chambon, 1996). RARs induce local chromatin changes at level of target genes, containing responsive RA elements (RAREs) by recruiting multiprotein complexes with histone acetyltransferase (HAT) activity and histone deacetylase (HDAC) activity, that dynamically pattern chromatin modification and regulate gene expression (see for review Chambon, 1996; Minucci and Pelicci, 1999).

RARs and RXRs, when disrupted, result in severe developmental defects and neoplastic transformation (Smith et al., 1992; Gudas et al., 1994; Chambon, 1996). In breast cancer cells, the expression of one member of the RARs family, RARβ is found consistently downregulated or lost (Roman et al., 1992; Shao et al., 1994; Swisshelm et al., 1994; Li et al., 1995; Widshwendtner et al., 1997; Xu et al., 1997; Liu et al., 1997). RARβ downregulation can be reversed by RA in estrogen receptor (ER)-positive, but not in ER-negative breast carcinoma cell lines, believed to represent more advanced forms of tumors (Liu et al., 1997). Loss of RA-induced RARβ expression is considered a crucial step in the development of RA-resistance in breast carcinogenesis. A complex regulatory region, with two promoters regulates RARβ gene expression. Only one promoter, RARβ2, containing several RA-response elements, including a canonical and an auxiliary RA response element, βRARE (de The' et al., 1990; Valcarel et al., 1994) is active in human mammary epithelial cells (HMEC). The transcription of the RARβ2 promoter is mediated by multiple RARs including, RARα and RARβ itself (Chiba et al., 1997) able to recruit coactivator and corepressor protein complexes with HAT/HDAC activities, respectively (Chambon, 1996). To understand why RARβ activity is downregulated, or lost, in breast cancer, intense search has been oriented at identifying possible alterations affecting either the RARβ2 promoter, or regulatory factors (Seewaldt et al., 1995; Widschwendtner et al., 1997; Xu et al., 1997; Tsou et al., 1998; Folkers et al., 1998).

DNA methylation is an epigenetic change that induces chromatin modifications and repression of transcription via a methyl CpG binding protein MeCP2, and recruitment of a Sin3A/HDAC corepressor complex (Nan et al., 1998; Wade et al., 1998; Razin, 1998; Ng and Bird, 1999; Jones and Wolffe, 1999). For this reason, we decided to investigate whether RARβ2 promoter was affected by DNA methylation. Indeed, we found hypermethylation at the RARβ2 promoter both in breast carcinoma cell lines, and a significant proportion of primary breast tumors. Treatment with the methyltransferase inhibitor 5-Aza-CdR partially reversed the DNA methylation state, and restored RARβ transcription, thus indicating that DNA methylation is at least one factor contributing to RARβ inactivity. However, the available data indicate that DNA methylation is only a component of the observed RARβ gene inactivity. Very likely, RA-inducibility of RARβ gene is influenced by modifications altering RARβ2 chromatin, produced by the nuclear receptors that act at βRARE (RARα and the same RARβ), as well as DNA methylation.


The RARβ2 promoter is methylated in breast cancer cell lines independently of their ER status and RA-inducibility

RARβ transcription was first tested in a panel of breast cancer cell lines grown in the absence of exogenous RA, by reverse transcriptase-PCR (RT–PCR), using primers encompassing exons 5 and 6 (de The' et al., 1990; van der Leede et al., 1992; Toulouse et al., 1997). Under these conditions, only one cell line, Hs578t, produced a detectable 256 bp RT–PCR product (Figure 4a). Thus, we confirmed previous reports that RARβ gene expression is down regulated/lost in breast cancer cell lines. Growing cells in the presence of RA can assess the distinction between downregulation and loss. As previously reported (Swisshelm et al., 1994; Liu et al., 1997; Shang et al., 1999), we observed induction of RARβ expression and growth inhibition in T47D, MDA-MB-435, MCF7 and ZR75-1 cell lines treated for 48 h with 1 μM RA, but not in the MDA-MB-231 and MDA-MB-468 cell lines.

Figure 4

Treatment with 5-Aza-CdR and TSA triggers re-expression of RARβ. (a) RT–PCR of mortal HMEC strains 48R and 172R and breast cancer cell lines. Brain RNA was used as a control. (b) RT–PCR of breast cancer cell lines treated for 3 days with 0.4 μM (lanes 2, 5, 8 and 11) and 0.8 μM (lanes 3, 6, 9 and 12) 5-Aza-CdR and untreated cells (lanes 1, 4, 7 and 10). (c) RT–PCR of MDA-MB-231 cells untreated (lane 2) in comparison with cells treated for 3 days with 1 μM RA (lane 3), 0.4 μM 5-Aza-CdR (lane 4), 0.4 μM 5-Aza-CdR+1 μM RA (lane 5). Hs578t used as positive control (lane 6). (d) RT–PCR of MDA-MB-231 cells untreated (lane 1), in comparison with cells treated for 48 h with 1 μM RA (lane 2), 100 ng/ml TSA (lane 3), 100 ng/ml TSA+1 μM RA (lane 4); solvent (lane 5)

To see whether the RARβ2 methylation status correlated with the ER status, we examined the methylation status at RARβ2 in a panel of ER-positive (MCF7, T47D, ZR75-1) and ER-negative (Hs578t, MDA-MB-231, MDA-MB-435, MDA-MB-468) cell lines.

By Southern blotting we analysed the CpG island of the RARβ2 promoter within a 7.5 kb XbaI DNA fragment encompassing the TATA box, the βRARE, the transcriptional start site (TS) and the 5′ untranslated region of exon 5 (Figure 1a). In this region we can identify nine HpaII sites (Shen et al., 1991; Baust et al., 1996). The DNA methylation status was analysed by using the methylation-sensitive enzyme, HpaII (Figure 1b). MspI, the isoschizomer of HpaII, insensitive to methylation, was used as a positive control. The PCR probe spans the βRARE and the TATA box regions (Figure 1a). The same 7.5 kb region was previously analysed in a colon carcinoma cell line, and the size of all the possible fragments relative to the most 3′HpaII site were reported (Cote' and Momparler, 1997). A representative blot is shown in Figure 1b. Genomic DNA from the ER-positive, RA-inducible cell line T47D is digested to completion, indicating that it is not methylated at any of the HpaII sites. In contrast, DNA from the ER-positive, RA-inducible ZR75-1 cell line and DNA from the ER-negative, RA-resistant MDA-MB-231 cell line showed to be differentially methylated at the methylation-sensitive sites. (Figure 1b). Using methylation-specific PCR (MSP), we further analysed a 616 bp long RARβ2 region from nucleotide 481 to nucleotide 1096 (Shen et al., 1991) in all the cell lines. MSP entails the modification of genomic DNA by sodium bisulfite that converts all unmethylated, but not methylated, cytosine to uracil (Herman et al., 1996). The distribution of CpGs expected after Na bisulfite modification and the four MSP primers (1–4) is reported in Figure 2a. The genomic DNAs from four breast cancer cell lines ZR751, MCF7, MDA-MB-231, MDA-MB-468 showed partial to complete methylation of the promoter region (Figure 2b). The human mammary epithelial cell (HMEC) strain 48R, expressing RARβ and three breast cancer cell lines, the RARβ-positive Hs578t and the RA-inducible MDA-MB-435 and T47D, revealed only the (U) unmethylated PCR products (Figure 2b).

Figure 1

Methylation sensitive Southern blotting of the RARβ2 promoter. (a) Genomic map of the RARβ2 promoter-exon 5 region indicating the position of HpaII sites (H) relative to βRARE, TATA, transcription start site (TS) and the ATG. (b) Southern analysis of : T47D and ZR571 DNAs digested with XbaI, HpaII (right) and MspI and MDA-MB-231 cells before, and after treatment with 0.8 μM 5-Aza-CdR for 3 days (left)

Figure 2

MSP analysis of DNA from cell lines and primary tumors. (a) Distribution of the methylated CpGs (filled circles) in the RARβ2 promoter region spanning nt 498 to nt 1096 and position of the MSP primers. (b) MSP analysis of a panel of breast carcinoma cell lines. U and M products amplified with the four sets of MSP primers in ER-positive and -negative cell lines and the mortal HMEC (48R) strain. (c) MSP analysis of two mortal (48R and 172R) and two immortal (184A1 and 184B5) HMEC strains. (d) MSP analysis of three breast tumors (T1–T3) and matching tumor cell free lymph nodes (N1–N3)

These results indicate that hypermethylation of the RARβ2 promoter occurs in breast cancer cell lines irrespective of the ER status, and can be detected in both RA-inducible, and RA-resistant breast cancer cells.

RARβ2 is unmethylated in both mortal and immortalized HMEC, but is methylated in primary breast tumors

Next, we asked whether hypermethylation of RARβ2 promoter in cell lines has correlates in clinical breast cancer. As a normal control we analysed the HMEC mortal strains (48R, 172R), that are the closest representation of normal mammary epithelial cells available. We also analysed two immortal mammary epithelial strains (184A1 and 184B5). The DNA of these strains was found to be unmethylated (Figure 2c). Consequently, methylation of RARβ2 may be an event in the progression of breast cancer, following immortalization. Genomic DNAs from three paraffinated samples of breast tumors, two ER-positive (T1, T2) and one ER-negative (T3), estimated to contain more than 90% tumor cells, were analysed with all MSP primer pairs, and shown to be partially methylated (Figure 2d). Both microdissected breast stroma, and microdissected normal epithelial cells were found unmethylated at RARβ2 (our unpublished observations), making it very likely that the U products in the tumor samples were amplified either from residual normal epithelial cells, or stromal cells mixed to tumor cells. DNAs from matching histologically tumor free lymph node samples (N1–N3), were similarly analysed and produced only the unmethylated PCR products (Figure 2d). The DNA of additional 21 tumors was performed using two sets of primer pairs (U3/M3 and U4/M4). Fifteen (7 ER-positive and 8 ER-negative) of the 24 tumors presented methylation at the RARβ2 promoter. With the same primer sets hypermethylation at RARβ2 was detected in the DNA of ten out of 39 primary breast tumors collected, and analysed independently, at the Johns Hopkins University.

The overall data indicate that hypermethylation at RARβ2 promoter occurs in approximately one third of primary breast tumors, and that the RARβ2 methylation state is independent of the ER status of the tumor.

5-Aza-CdR induces partial demethylation at the RARβ2 CpG island and reactivation of RARβ gene expression

In order to determine whether DNA methylation is affecting, at least in part, RARβ gene expression, we treated all the cell lines showing methylation at the RARβ2 promoter with the DNA methyltransferase inhibitor, 5-Aza-CdR. Treatment of cells with either 0.4 or 0.8 μM 5-Aza-CdR for 3 days, led to partial demethylation of the CpG rich RARβ2 region. This was evident both by Southern analysis in the MDA-MB-231 cell line (Figure 1b, left panel), and by MSP in all cell lines (Figure 3). Moreover, 5-Aza-CdR treatment resulted in reactivation of gene expression both in RA-inducible MCF7 and ZR75-1, and RA-resistant MDA-MB-231 and MDA-MB-468 cells (Figure 4b). We asked whether reactivation of RARβ expression by 5-Aza-CdR A-resistant cells could be enhanced by RA. By using non-quantitative RT–PCR, we could not appreciate a difference in the level of RARβ transcription in MDA-MB-231 cells treated with 0.4 μM 5-Aza-CdR alone, or in combination, with 1 μM RA (Figure 4c). In this experiment, 5-Aza-CdR alone, or in combination with RA, produced 63 and 96% growth inhibition respectively. In the same experiment, treatment with 1 μM RA alone produced a negligible effect on growth inhibition (<2%). A synergistic effect of the two drugs on cancer cells was previously reported (Cote' and Momparler, 1997; Bovenzi et al., 1999).

Figure 3

Treatment with 5-Aza-CdR induce partial de-methylation. MSP analysis of DNA of four breast cancer cell lines before and after treatment for 3 days with 0.8 μM 5-Aza-CdR

These data indicate that DNA methylation is, at least, one factor influencing the downregulation/loss of RARβ transcription in breast cancer cell lines with a methylated RARβ2 promoter. Cells treated with 5-Aza-CdR alone, or in combination with RA, showed re-expression of RARβ, which may have contributed, along with the toxic 5-Aza-CdR, to the observed growth inhibition.

The HDAC inhibitor TSA can reactivate RARβ expression in RA-resistant cells; demethylation of the RARβ2 promoter is not an absolute requirement for RARβ reactivation

The chromatin status at a given locus can be dynamically influenced by the degree of acetylation/deacetylation due to HAT/HDAC activities. Absence of RARβ regulatory factors, like RARα, as well as DNA-methylation, can contribute to pattern chromatin modifications at RARβ promoter in RA-resistant cell lines. One of these cell lines, MDA-MB-231, lacks RA-inducible RARα activity (Shao et al., 1994) and displays a RARβ2 methylated promoter. We decided to probe indirectly whether the level of HDAC at RARβ2 can influence RARβ expression, by testing the effect of TSA, a HDAC inhibitor on MDA-MB-231 cells (Yoshida et al., 1995). Cells were treated for 2 days, in the presence or absence of 100 ng/ml TSA alone, or in combination, with 1 μM RA. By using RT–PCR, it was clear that, unlike cells treated with RA alone, cells treated with a combination of RA and TSA re-expressed RARβ mRNA (Figure 4d). Under the same experimental conditions, 100 ng/ml TSA alone, or in combination with 1 μM RA, produced 77 and 92% growth inhibition, respectively. Treatment with 1 μM RA alone did not affect significantly growth inhibition (<2%). By MSP analysis, we could assess that RARβ expression was restored in the presence of a methylated RARβ2 promoter. The MSP profile obtained with primer set 3, spanning the βRARE region is reported in Figure 5. This finding indirectly shows that global alterations of HDAC activity, generated by TSA in MDA-MB-231 cells, involved RARβ2 resulting in RA-induced RARβ expression. Further, demethylation at RARβ2 did not seem to be an absolute requirement for RARβ gene expression in MDA-MB-231 cells. Noteworthy, persistence of methylation at RARβ2 was observed also in MCF7 cells where RARβ transcription could be restored in the presence of RA (data not shown). Growth inhibition was observed in cells treated with TSA alone, or in combination, with RA. Very likely, RARβ along with TSA, a drug known to induce growth inhibition (Yoshida et al., 1995), contributed to the massive growth inhibitory effect that we observed.

Figure 5

DNA methylation in RARβ2 promoter of MDA-MB-231 cells expressing RA-induced RARβ after TSA treatment. MSP analysis of MDA-MB-231 cells using primers 3 shows that there is no demethylation of the βRARE containing region in cells treated with 1 μM RA (lane 3), 100 ng/ml TSA (lane 4), 100 ng/ml TSA+1 μM RA (lane 5); in comparison with untreated cells (lane 2), or cells grown in the presence of solvent (lane 6). MSP of Hs578t was used as a control (lane 1)


RARβ2 promoter is methylated in breast cancer

In this study, we show evidence that, in breast cancer cells, RARβ2 promoter undergoes DNA hypermethylation, an epigenetic change known to induce chromatin modifications and influence gene expression (Razin, 1998; Ng and Bird, 1999; Jones and Wolffe, 1999). We detected methylation of the RARβ2 promoter region, both in breast carcinoma cell lines, and a significant proportion of primary breast tumors. RARβ2 methylation status did not correlate with the ER status of breast cancer cells and was observed both in in situ lesions and invasive tumors (our unpublished observations).

It is not clear when epigenetic changes occur during breast cancer progression. However, methylation of the promoter was not detected in both mortal, and immortal human mammary epithelial cell (HMEC) strains, as well as in normal microdissected breast epithelial cells (our unpublished observations). These results suggest that aberrant methylation of the RARβ2 CpG island may be a later event following immortalization. Treatment of breast cancer cells presenting with a methylated RARβ2, with the demethylating agent 5-Aza-CdR, induced partial DNA demethylation and restored RARβ gene expression. This evidence clearly indicates that DNA methylation is at least a component contributing to RARβ downregulation/loss.

RARβ2 methylation state and RA-inducibility

The correlation between RARβ2 methylation and RA-inducibility in different breast cancer cell lines, indicates that DNA methylation is not the only factor influencing RARβ silencing. Survey of different breast cancer cell lines shows that RARβ is downregulated, but can be reinduced by RA both in MDA-MB-435 and T47D cells, with unmethylated RARβ2 promoter and in MCF7 and ZR751 cells, with a methylated promoter. In contrast, in MDA-MB-231 and MDA-MB-468 cell lines the methylated RARβ2 promoter is indifferent to RA treatment. Apparently, different degrees of repression can affect RARβ2 promoter, and only in some cases, the ligand is sufficient to alleviate methyl-directed repression. Extinction of RARβ transcription must be determined by a stable repressive state in the chromatin structure determined by more than one mechanism, including DNA methylation.

DNA-methylation might be secondary to RARβ2 promoter inactivity

We hypothesize that low intracellular levels of RA in breast cancer cells may induce chromatin structure alterations at RARβ2, similar to the ones observed in the P19 embryonal carcinoma cell line (Bhattacharyya et al., 1997). Although the mechanism of chromatin structure alterations are not fully understood, current evidence indicates that local histone acetylation is a crucial factor (Razin, 1998). An altered chromatin environment may predispose to DNA methylation, a condition that might further affect histone deacetylation at RARβ2 (Razin, 1998; Ng and Bird, 1999; Jones and Wolffe, 1999). The first to propose that gene inactivity ‘invitesrsquo; de novo methylation was Bird (1986). The hypothesis was further refined, after the discovery of the mechanistic link between DNA methylation and chromatin conformation mediated by the MeCP2/Sin3A/HDAC corepressor complex (Nan et al., 1998; Wade et al., 1998). According to the revisited hypothesis, Ng and Bird (1999) propose that: ‘DNA methyltransferase – either independently or assisted by accessory proteins – may be capable of reading the histone acetylation pattern on the chromatin and its de novo methyltransferase activity can respond differentially to different states of chromatin modification. In this case, deacetylated chromatin would provoke de novo methylation. This self-reinforcing mechanism, supported by DNA methylation and histone deacetylation, could provide a stable state of inactive chromatin, unless overcome by other mechanisms'.

RARβ2 promoter in breast cancer might provide an ideal system to test this hypothesis, given the heterogeneous correlation between its methylation state and RA-inducibility in different breast cancer cells. Unmethylated, RA-inducible RARβ2 promoters are expected to be associated with, either an active chromatin state, or a mild repressive state. A methylated RARβ2 promoter is expected to be associated with a more repressive chromatin environment. As a consequence, transcription from a methylated promoter should be possible, either by recruiting consistent HAT activity, or by inhibiting excessive HDAC activity. These speculations are so far supported by compelling circumstantial evidence. Notably, RA can induce both RARα and RARβ in MCF7 cells from a RARβ2 methylated promoter (Shao et al., 1994; Shang et al., 1999; our unpublished observations). This suggests that RA may trigger recruitment of HAT activity at RARβ2, sufficient to override methylation-related chromatin constraints. On the contrary, in the MDA-MB-231 cells we saw that RA-induced RARβ transcription is possible after treatment with TSA, a HDAC inhibitor, already known to induce chromatin alterations at RARβ2 promoter in P19 cells (Minucci et al., 1997). Analysis of the DNaseI sensitivity pattern, in and around RARβ2, as well as the assessment of RARβ2 histone acetylation state (Keshet et al., 1986; Hebbes et al., 1994; Eden et al., 1998), in both MCF7 and MDA-MB-231 cells will give us an idea of the relation between chromatin environments and RARβ transcription. Moreover, these studies are expected to shed light on the relation of histone acetylation and methylation of the RARβ2 promoter. This issue is of particular interest since it is not yet completely clear whether DNA demethylation is indeed always required to restore transcription from genes with fully methylated promoters (Cameron et al., 1999; Ferguson et al., 1998; Razin, 1998; Ng and Bird, 1999).

In conclusion, we provide evidence that DNA-methylation at RARβ2 promoter in breast cancer cells is affecting, at least in part, RARβ transcription. We argue that DNA-methylation is secondary to the inactive state at RARβ2 promoter and may contribute to create a stable repressive RARβ2 environment and extinction of RARβ transcription. Further understanding of epigenetic changes and chromatin alteration at RARβ2 may have preventive and therapeutic implications. Changes altering RARβ2 chromatin structure and RARβ transcription in breast cancer might be prevented in the presence of supraphysiological levels of RA (Minna and Mangeldorf, 1997). Knowledge of RARβ2-methylation state of primary breast cancers might be useful to identify tumors that are more likely to respond to RA-therapy. Finally, the possibility to re-induce RARβ activity in RA-resistant breast cancer cells, using both TSA and RA, a combination proven to be effective for treating leukemia (Grignani et al., 1998; Guidez et al., 1998; He et al., 1998; Lin et al., 1998; Warrell et al., 1998), might have therapeutic implications also in the treatment of RA-resistant breast tumors.

Materials and methods

Cell cultures

Human epithelial mammary cells (HEMC) from reduction mammoplasty including three mortal strains, 184, 48R and 172R, and two immortal strains, 184A1 and 184B5, were obtained and cultured according to the protocols designed by Dr Martha Stampfer (see the HMEC Homepage, http://www.lbl.gov/mrgs/index.htlm) using Clonetics (Walkersville, MD, USA) reagents.

Human breast cancer cell lines were maintained in Dulbecco's modified Eagle's medium (GIBCO) (Hs578t, MCF-7, MDA-MB-231 and T47D) or IMEM medium (Biofluids) (MDA-MB-435, MDA-MB-468, ZR751) with 5% fetal calf serum (FCS). For drug treatments, exponentially growing cells were seeded in 10 cm2 plates at a density of 3×105 cells/plate or in 6-well plates at 1×105 cells/well. Cells were allowed to attach overnight before the addition of the appropriate concentration of 5-Aza-2′ deoxycytidine (5-Aza-CdR) (Sigma), Trichostatin A (TSA) (Sigma) or RA (Sigma). When reduction of retinoids was required, cells were treated in either medium with 0.5% FCS or charcoal-dextran stripped FCS (Hyclone). At the indicated time points, both attached and detached cells were harvested, counted with Trypan Blue (Life Technologies) and processed for DNA or RNA extraction. 5-Aza-CdR was dissolved in 0.45% NaCl containing 10 mM sodium phosphate (pH 6.8). Trichostatin A and all-trans-retinoic acid (RA) (Sigma) were reconstituted in absolute ethanol (solvent). The growth inhibition (%) was calculated as: (1-NT/NC)×100, where NT is the number of treated cells and NC is the number of control cells.

Tissue samples

Normal and tumor tissues were collected from existing tumor banks (Instituto per lo Studio e la Cura dei Tumori, Milan; the Cancer Center, Rotterdam, the Johns Hopkins Breast Cancer Program, Baltimore, MD, USA). All tumor samples were obtained from excess clinical specimens and institutional guidelines for the acquisition and maintenance of such specimens were followed.

DNA and RNA extraction

Extraction of DNA and RNA from breast cancer cell lines was performed by using DNAzol and Trizol respectively (Life Technologies) according to the manufacturer's instructions. Genomic DNA was further treated with 500 μg/ml proteinase K at 55°C, extracted with phenol-chloroform-isoamylic alcohol (24 : 24 : 1) (CIA) and ethanol precipitated. Extraction of DNA from paraffinated breast cancer and lymph node tissues was essentially performed as previously described (Formantici et al., 1999). One to three consecutive sections estimated to contain at least 90% tumor cells were incubated at 58°C overnight in 200 μl of extraction buffer (50 mM KCl, 10 mM Tris-HCl (pH 7.5), 2.5 mM MgCl2, 0.1 mg/ml gelatin, 0.45% NP-40, 0.45% Tween 20, and the solution was heated at 95°C for 15 min to inactivate the proteinase K and then centrifuged at 6000 r.p.m. The DNA in the supernatant was used for analysis.

Southern blotting

Genomic DNA (7 μg) was digested overnight with 15 U/μg of XbaI, HpaII and MspI enzymes, electrophoresis on a 0.8% agarose gel and transferred to Hybond-N filter. A 227 bp probe was amplified using the sense 5′-AGA GTT TGA TGG AGT TGG GTG GAG-3′ and antisense 5′-CAT TCG GTT TGG GTC AAT CCA CTG-3′ primers, gel purified and labeled with 32P-dCTP using the Megaprime DNA labeling system (Amersham). After hybridization the filters were washed and exposed to X-ray film at −80°C for autoradiography.

Methylation specific PCR (MSP)

Bisulfite modification of genomic DNA was essentially performed as described by Herman et al. (1996). Modified DNA was used immediately or stored in aliquots at −20°C. The PCR mixture contained 1×PCR buffer (16.6 mM ammonium sulfate, 67 mM Tris (pH 8.7), 1.5 mM MgCl2), dNTPs (each at 1.25 mM), primers (300 ng each per reaction), and bisulfite-modified DNA (50 ng) or unmodified DNA (50 ng). Reactions were hot started at 95°C before the addition of 2.5 U of Taq polymerase (Qiagen). Amplification was carried out in a Thermal Cycler 480 Perkin Elmer for 30 cycles (1 min at 94°C, 1 min at the annealing temperature (at) selected for each primer pair, 1 min at 72°C), followed by 4 min at 72°C. Twelve μl of the PCR reaction were electrophoresed onto 1.5% agarose gels, stained with ethidium bromide and visualized under UV. Two primer pairs, W3 sense 5′-CAGCCCGGGTAGGGTTCACC-3′, W3 antisense 5′-CCGGATCCTACCCCGACGG-3′, and W4 sense 5′-CCGAGAACGCGAGCGATCC-3′ and W4 antisense 5′-GGCCAATCCAGCCGGGGCG-3′, were designed on the human RARβ2 sequence (Shen et al., 1991) and used to control the Na bisulfite modification. The primer pairs selected to detect the unmethylated DNA were as follows: U1 sense 5′-GTG GGT GTA GGT GGA ATA TT-3′ and U1 antisense 5′-AAC AAA CAC ACA AAC CAA CA-3′ (at 55°C); U2 sense 5′-TGT GAG TTA GGA GTA GTG TTT T-3′ and U2 antisense 5′-TTC AAT AAA CCC TAC CCA-3′ (at 49°C); U3 sense 5′-TTA GTA GTT TGG GTA GGG TTT ATT-3′ and U3 antisense 5′-CCA AAT CCT ACC CCA ACA-3′ (at 55°C); U4 sense 5′-GAT GTT GAG AAT GTG AGT GAT TT-3′ and U4 antisense 5′-AAC CAA TCC AAC CAA AAC A-3′ (at 55°C); The sequences of the primers to detect the methylated DNA were: M1 sense 5′-AGC GGG CGT AGG CGG AAT ATC-3′ and M1 antisense 5′-CAA CGA ACG CAC AAA CCG ACG-3′ (at 63°C); M2 sense 5′-CGT GAG TTA GGA GTA GCG TTT C-3′ and M2 antisense 5′-CTT TCG ATA AAC CCT ACC CG-3′ (at 57°C); M3 sense 5′-GGT TAG TAG TTC GGG TAG GGT TTA TC-3′ and M3 antisense 5′-CCG AAT CCT ACC CCG ACG-3′ (at 64°C); M4 sense 5′-GTC GAG AAC GCG AGC GAT TC-3′ and M4 antisense 5′-CGA CCA ATC CAA CCG AAA CG-3′ (at 64°C).

The distrubution of the CpG methylated sites and the position of the primers is reported in Figure 2. M and U primers were designed in the same regions, with one or two nucleotide differences to meet annealing requirements. Fragment M3 (position 773–1007) contains the βRARE (792–808) and the transcription start site (position 844); fragment M4 (position 949–1096) contains an Sp1 element (position 1074–1081).


The exon 5 (sense primer 5′-GAC TGT ATG GAT GTT CTG TCA G-3′) and exon 6 (antisense primer 5′-ATT TGT CCT GGC AGA CGA AGC A-3′) were designed on the basis of published RARβ2 transcript (de The' et al., 1990; van der Leede et al., 1992) and used to amplify 50 ng of DNase treated total RNA using the Superscript One-Step RT–PCR System (Life Technologies). RT–PCR with actin primers (sense primer 5′-ACC ATG GAT GAT GAT ATC G-3′ and antisense primer 5′-ACA TGG CTG GGG TGT TGA AG-3′ was used as an internal RNA control.


  1. Baust C, Redpath L and Schwarz E . 1996 Int J Cancer 67: 409–416

  2. Bhattacharyya N, Dey A, Minucci S, Zimmer A, John S, Hager G and Ozato K . 1997 Mol Cell Biol 17: 6481–6490

  3. Bird A . 1986 Nature 321: 209–213

  4. Bovenzi V, Le NLO, Cote' S, Sinnet D, Momparler LF and Momparler RI . 1999 Anticancer Drugs 10: 471–476

  5. Cameron EE, Bachman KE, Myohanen S, Herman JG and Baylin SB . 1999 Nature Genet 21: 103–107

  6. Chambon P . 1996 FASEB J 10: 940–954

  7. Chiba H, Clifford J, Metzger D and Chambon P . 1997 Mol Cell Biol 17: 3013–3020

  8. Cote' S and Momparler RL . 1997 Anticancer Drugs 8: 56–61

  9. De The' H, del Mar Vivanco-Ruiz M, Tiollais P, Stunnenberg HG and Dejan A . 1990 Nature 343: 177–180

  10. Eden S, Hashimshony T, Keshet I, Cedar H and Torne AW . 1998 Nature 394: 842

  11. Ferguson AT, Lapidus RG and Davidson NE . 1998 Oncogene 17: 577–583

  12. Folkers GE, van der Burg B and van der Saag PT . 1998 J Biol Chem 273: 32200–32212

  13. Formantici C, Orlandi R, Ronchini C, Pilotti S, Ranzani GN, Colnaghi MI and Menard S . 1999 J Pathol 187: 424–427

  14. Grignani F, Dematteis S, Nervi C, Tomassoni L, Gelmetti V, Cioce M, Fanelli M, Ruthardt M, et al. 1998 Nature 391: 815–818

  15. Gudas LJ, Sporn MB and Roberts AB . 1994 In The Retinoids Biology, Chemistry and Medicine Raven Press New York pp 443–520

  16. Guidez F, Ivins S, Zhu J, Soderstrom M, Waxman S and Zelent A . 1998 Blood 91: 2634–2637

  17. He LZ, Guidez F, Tribioli C, Peruzzi D, Ruthardt M, Zelent A and Pandolfi PP . 1998 Nature Genet 18: 126–135

  18. Hebbes TR, Clayton AL, Thorne AW and Crane-Robinson C . 1994 EMBO J 13: 1823–1830

  19. Herman JG, Graff JR, Myohanen S, Nelkin BD and Baylin SB . 1996 Proc Natl Acad Sci USA 93: 9821–9826

  20. Jones PL and Wolffe AP . 1999 Semin Cancer Biol 9: 339–347

  21. Keshet I, Lieman-Hurwitz J and Cedar H . 1986 Cell 44: 535–543

  22. Li X-S, Shao Z-M, Sheikh MS, Eiseman JL, Sentz D, Jetten AM, Chen J-C, Dawson ML, Aisner S, Rishi AK, Gutierrez P, Schnapper L and Fontana JA . 1995 J Cell Physiol 165: 449–458

  23. Lin RJ, Nagy I, Inoue S, Shao W, Miller Jr WH and Evans RM . 1998 Nature 391: 811–814

  24. Liu Y, Lee M-O, Wang H-G, Li Y, Hashimoto Y, Klaus M, Reed JC and Zhang X-K . 1997 Mol Cell Biol 16: 1138–1149

  25. Minucci S, Horn V, Bhattacharyya N, Russanova V, Ogryzko VV, Gabriele L, Howard BH and Ozato K . 1997 Proc Natl Acad Sci USA 94: 11295–11300

  26. Minna JD and Mangeldorf DJ . 1997 J Natl Cancer Inst 89: 602–604

  27. Minucci S and Pelicci P . 1999 Semin Cell Dev Biol 2: 215–225

  28. Ng H-H and Bird A . 1999 Curr Opin Genet Dev 9: 158–163

  29. Nan X, Ng HH, Johnson CA, Laherty CD, Turner BM, Eisenman RN and Bird A . 1998 Nature 393: 386–389

  30. Razin A . 1998 EMBO J 17: 4905–4908

  31. Roman SD, Clarke CL, Hall RE, Alexander IE and Sutherland RL . 1992 Cancer Res 52: 2236–2242

  32. Seewaldt VI, Johnson BS, Parker MB, Collins SJ and Swisshelm K . 1995 Cell Growth Differ 6: 1077–1088

  33. Shao Z-M, Sheikh MS, Chen JC, Kute T, Aisner S, Schnaper L and Fontana JA . 1994 Int J Oncol 4: 849–853

  34. Shang Y, Baumrucker CR and Green MH . 1999 J Biol Chem 274: 18005–18010

  35. Shen S, Kruyt FA, den Hertog J, van der Saag FT and Kruijer W . 1991 DNA Seq 2: 111–119

  36. Smith MA, Parkinson Dr, Cheson BD and Friedman MA . 1992 J Clin Oncol 10: 839–864

  37. Swisshelm K, Ryan K, Lee X, Tsou HC, Peacoque M and Sager R . 1994 Cell Growth Differ 5: 133–141

  38. Toulouse A, Morin J, Pelletier M and Bradley WEC . 1997 BBA 1309: 1–4

  39. Tsou HC, Yao YJ, Xie XX, Ping XL and Peacocke M . 1998 Exp Cell Res 245: 221–227

  40. Valcarel R, Holz H, Garcia Jimenez C, Barettino D and Stunnenberg HG . 1994 Genes Dev 8: 3068–3079

  41. Van der Leede BJ, Folkers GE, Kruyt FA and van der Saag PT . 1992 Biochem Biophys Res Commun 188: 695–702

  42. Wade PA, Jones PL, Vermaak D, Veenstra GJ, Imhof A, Sera T, Tse C, Ge H, Shi YB, Hansen JC and Wolfee AP . 1998 Cold Spring Harb Symp Quant Biol 63: 435–445

  43. Warrell Jr, RP, He LZ, Richon V, Callega E and Pandolfi PP . 1998 J Natl Cancer Inst 90: 1621–1625

  44. Widschwendtner M, Berger J, Daxenbichler G, Muller-Holzner E, Widschwendtner A, Mayr A, Marth C and Zeimet AG . 1997 Cancer Res 17: 4158–4161

  45. Xu XC, Sneige N, Liu X, Nandagiri R, Lee JJ, Lukumanji F, Hortobagy G, Lippman SM, Dhingra K and Lotan R . 1997 Cancer Res 57: 4992–4996

  46. Yoshida M, Horinouchi S and Beppu T . 1995 Bioessays 17: 423–430

Download references


We thank Drs WEC Bradley (Montreal) and X-Q Zhang (La Jolla) for initial helpful discussions on the idea behind this work and reagents and Dr A Hoogeveen for critical suggestions; Dr M Stampfer for the gift of the HMEC strains; Dr A de Klein for providing DNA from breast tumors. Funding for this work were provided by Associazione Italiana Ricerca sul Cancro (AIRC) and by BC980803 (USA) to N Sacchi.

Author information

Correspondence to Silvia M Sirchia or Anne T Ferguson or Nicoletta Sacchi.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Sirchia, S., Ferguson, A., Sironi, E. et al. Evidence of epigenetic changes affecting the chromatin state of the retinoic acid receptor β2 promoter in breast cancer cells. Oncogene 19, 1556–1563 (2000) doi:10.1038/sj.onc.1203456

Download citation


  • breast cancer
  • DNA methylation
  • chromatin remodeling
  • retinoic acid receptor (RAR)β

Further reading