We first examined the kinetics of ERα protein turnover in MCF-7 cells following treatment with estradiol (E2), two SERMs (4-hydroxytamoxifen or OHT and RU39,411 or RU39) and two SERDs (ICI 182,780 or ICI and RU58,668 or RU58). It has been proposed that ligand dependent ERα regulation may result from the presence a long aliphatic side chain on steroid core. Thus in this study we selected RU39 and RU58 which are derivatives of 17β-estradiol but with different side chains. RU39 has a dimethyl-amino-ethoxy-phenyl side chain similar to the one in tamoxifen, while RU58 has a bulky hydrophobic side chain similar to the one in Fulvestrant (ICI) (Figure 1A). ERα protein levels in E2, ICI and RU58 treated MCF-7 cells rapidly decreased (Figure 1B). Time course experiments showed that 1 h after E2 induction, the detected amount of ERα protein accounted for only 40% of ERα levels before treatment; after 4 h, ERα levels were as low as 20% of the quantity of ERα present in untreated cells, and after 16 h ERα protein remained at a level equivalent to the one observed 1 h after addition of E2. Treatment with SERDs (ICI and RU58) resulted in >70% reduction of ERα protein levels after 1 h, 4 h and even 16 h reaching 95% after 1 h exposure to ICI (Figure 1B). Treatment of MCF-7 cells with OHT or RU39 (Figure 1B), two compounds classified as SERM, reduced from 40% to 50% of ERα protein levels at the initial 1 h time-point and about 20% after 16 h and 4 h treatment with OHT and RU39, respectively. In addition, ERα protein levels were almost equivalent to the ones detected in untreated cells after 4 h or 16 h culture in the presence of OHT or RU39. Hence, ERα protein levels are stabilized by SERMs.

To assess whether changes in protein levels reflect variations of ERα protein stability or of transcription rates of the ESR1 gene, we quantified ERα mRNA accumulated following 16 h treatment with the different compounds (Figure 1C). ESR1 mRNA expression was greatly reduced after treatment with ERα ligands. In the presence of E2, only 40% of ESR1 mRNA could be recovered. Similarly, treatment with SERMs and SERDs repressed ESR1 mRNA transcription by 45%-60% relative to untreated MCF-7 cells. Despite the fact that a reduction in ERα protein levels was readily detectable after 1 h and significant after 16 h (Figure 1B), we observed that E2 induced a 7 to 10 fold increase in mRNA levels of the ERα-target gene GREB1 compared to mock treated cells (Figure 1C). GREB1 transcription was inhibited by SERMs and SERDs (> 40% reduction compared to untreated cells; Figure 1C). These results were expected since E2 is known to activate this ERα target gene, while SERMs and SERDs are antiestrogens and thus repress GREB1 transcription in ERα positive mammary tumour cells 20.

Thus, in MCF-7 cells, variations in ERα protein levels do not necessarily correlate with ESR1 transcription in the presence of ligands. We note that the decrease in ERα protein levels is more pronounced after treatment with SERDs than after addition of E2, while the effect of hormone and SERDs on ESR1 mRNA accumulation was comparable. These results indicate that reduction of ERα protein levels following treatment with SERDs cannot be solely attributed to decreased ESR1 mRNA levels. SERDs apparently act both on transcription of the ESR1 gene and on ERα protein turnover. In contrast, ERα protein levels appear stable after 16 h treatment with SERMs (data not shown) despite reduced ESR1 expression levels. This suggests that binding to SERMs stabilizes the ERα.

SERMs and SERDs can be distinguished based on molecular mechanisms 9. To unambiguously determine localization of the estrogen receptor and its intracellular trafficking in response to treatment with various ligands we established a MCF-7 cell line stably expressing GFP-ERα from a CMV promoter. It was previously shown that transiently expressed GFP-ERα is functional using an estrogen response element driven luciferase reporter gene 24. Expression of GFP-ERα in MCF-7 cells did reportedly not alter cell cycle progression and GFP-ERα participated in estrogen target gene regulation similarly to endogenous ERα 25. We tagged the N-terminus of the human ERα with the S65T variant of GFP for transfection and stable integration in MCF-7 cells. Several clones were recovered and screened for total GFP-ERα protein content after treatment with E2, OHT or ICI using fluorescence microscopy and western blots. Here, we selected a MCF-7 derived clone (SK19) expressing GFP-ERα in which changes in endogenous ERα protein levels in response to a 4 h treatment with E2, OHT and ICI were identical to the ones observed in MCF-7 cells (compare lanes labeled ERα in MCF-7 and SK19 cells in Figure 2A). In addition, mRNA expression levels of some ERα target genes, ESR1, TFF1/pS2, GREB1 and PGR, were verified in the selected clone SK19 and compared to gene expression levels in MCF-7 cells (Figure 2B). mRNA levels of the progesterone receptor gene (PGR) and GREB1 increased rapidly after addition of 10 nM E2 to cells grown in steroid free medium to reach 2.2 to 2.8 fold (after 2 h) for both genes, and after 16 h to reach from 3.8 to 4.6 fold for PGR and 6.3 to 7.0 fold for GREB1 gene, in MCF-7 and SK19 cells respectively. TFF1 mRNA also accumulated after 16 h E2 treatment to reach 1.5 fold in both cell lines. As expected, ESR1 transcription was reduced in the presence of E2. The RPLPO gene is not a target of ERα and its expression levels were insensitive to hormone addition. Expression levels of all tested genes were similar in SK19 and MCF-7 cells. Thus the presence of GFP-ERα does not alter hormone-responsiveness at the transcriptional level.

In SK19 cells, GFP-ERα protein accounted for 50% of total ERα (GFP-ERα and endogenous ERα) in untreated cells. In the presence of E2 both GFP-ERα and endogenous ERα protein levels are reduced (Figure 2A). The CMV promoter being insensitive to E2 and antiestrogens, GFP-ERα protein levels are unlikely to be transcriptionally regulated. This observation together with the results shown in Figure 1 provides evidence that ERα protein turnover is regulated directly by binding of the receptor to ligands and its subsequent degradation.

Previous studies have demonstrated that GFP-ERα resides predominantly in the nucleus in transiently transfected mammary tumour cell lines 24, Hela cells 2627 and in MCF-7 cells expressing GFP-ERα from an inducible promoter 25. These microscopy based observations largely contradict results based on cellular fractionation which suggest that large amounts of ERα, in the absence or the presence of ligands, associate with the cytoplasmic fraction 19. It has been proposed that the relative amount of cytoplasmic ERα is indicative of the mechanism of action of certain antiestrogens 19. Commonly used cell fractionation protocols include a detergent based extraction step. Importantly, ERα and other nuclear receptors such as the glucocorticoid receptor (GR) are easily extracted from the nucleus in the presence of low concentrations of detergents such as NP40 (V. Marsaud and H. Richard-Foy, unpublished observations). As a consequence, apparent enrichment of ERα or GR in the cytoplasm likely results from the extraction protocol rather than a specific behavior of nuclear receptors. Here, we used a digitonin based cell fractionation protocol to determine the distribution of unbound and ligand-bound ERα and GFP-ERα in different cellular compartments (Figure 2C). Effectiveness of the fractionation protocol (for details see Methods) was confirmed using lamin A for the nuclear insoluble fractions, cytokeratin 18 for the nuclear and cytoplasmic fractions, and α-tubulin for the cytoplasmic fraction (Figure 2C). Treatment of cells with E2 and various antiestrogens did not affect cellular distribution of these proteins. We found that endogenous ERα associates predominantly with the nuclear fraction in the SK19 cells. In untreated cells, the part of ERα retained in the cytoplasm corresponded to ~20% of total endogenous ERα detected using the HC-20 antibody (Figure 2C, lanes 1 and 13). Similarly, the bulk of GFP-ERα, detected using either the HC-20 antibody or an antibody directed against GFP, was found in the nucleus. Following addition of E2, we note an overall decrease in ERα protein levels that could mainly be attributed to a reduction in nuclear ERα (Figure 2C, lanes 4-6). Treatment of SK19 cells with SERDs, ICI or RU58 leads to a decrease in overall ERα protein levels as shown for MCF-7 cells in Figure 1B. Notably, the remaining ERα was concentrated in the nuclear insoluble fraction which corresponded to ≥40% of total ERα in the presence of either ICI or RU58 (Figure 2C, lanes 12 and 21) suggesting that the nuclear soluble fraction was rapidly degraded. In contrast, we found that treatment with OHT and RU39 (Figure 2C, lanes 7-9 and 16-18) resulted in a cellular distribution similar to the one observed in untreated cells (Figure 2C, lanes 1-3) where at least 50% of ERα protein remained in a soluble nuclear compartment. Our cellular fractionation protocol is robust since the effects of various ligands are reproducible inside each category: OHT and RU39 induce the same effect on ERα protein distribution and this effect is distinct from the one of ICI and RU58. In addition, we show that occupation of different cellular compartments by GFP-ERα reflected the localization of endogenous ERα as detected by fluorescence imaging (see below).

GFP-ERα can be visualized in SK19 cells using conventional wide-field microscopy. SK19 cells were cultured on conventional glass microscopy coverslips in phenol-red free media for 3 days. Culture conditions were identical to conditions used for cell fractionation, immunoblotting or RNA extraction prior to RT-qPCR. Figure 3A shows representative images of SK19 cells treated or not with E2, SERMs and SERDs. We note that in the SK19 cell line GFP-ERα was excluded from the nucleoli, as previously observed for the cellular distribution of endogenous ERα in MCF-7 cells 2829 and of transiently transfected GFP-ERα 242630, under all conditions tested. Exposure times were identical for all conditions examined by fluorescence microscopy.

In untreated cells, ERα was uniformly distributed in the nucleus (compare GFP-ERα fluorescence (Figure 3Ab), to the DAPI (4',6-diamidino-2-phenylindole) nuclear stain in Figure 3Aa). A linear scan across the entire field including cytoplasm and nucleus (Figure 3Ac) shows that the cytoplasmic GFP-ERα fluorescence was barely above background (~15% of the maximum fluorescence intensity detected in the nucleus) which correlates with observations from cell fractionation experiments (Figure 2C, lane 1). In the presence of E2, GFP-ERα rapidly relocalized to accumulate in numerous foci scattered throughout the nucleoplasm (Figure 3Ae). In E2 treated cells, no GFP-ERα fluorescence could be detected in the cytoplasm (see linescan Figure 3Af). In contrast, after 1 h treatment with SERMs, OHT or RU39, we did not observe any intranuclear reorganization of GFP-ERα compared to untreated cells. This observation also correlates with our fractionation experiments (Figure 2C, lanes 1-3 compare to lanes 7-8 and 16-18). GFP-ERα staining remained diffuse with fluorescence intensity comparable to mock cells (Figure 3A b, n, r and corresponding linescans Figure 3A c, o, s). However, again no cytoplasmic GFP-ERα could be detected.

The distribution of the intensity of the fluorescent signals was determined within nuclei excluding the nucleolus. The frequency of pixels with respect to their intensity allows to calculate a coefficient of variation (CV). In cells treated with SERMs the CV was comparable to the one in control cells while the CV was 2 to 3 fold higher in cells exposed to E2 or SERDs (Table 1). This quantitive measure strengthens our observation that ERα accumulates in intranuclear foci when bound to E2 or SERDs but not in the presence of SERMs.

Upon exposure to SERDs, both ICI and RU58, GFP-ERα accumulated at numerous sites, reminiscent of the ones observed in the presence of E2 (Figure 3A h, k and 3Ae). We ascertained that the fluorescent foci detected in SK19 cells correspond to an accumulation of endogeneous ERα using immuno-electron microscopy of MCF-7 cells. Several immunogold labeled ERα molecules were frequently detected within ~100 nm distance from each other in 80 nm thin sections of E2 or ICI treated cells (Figure 3B).

In addition, in SK19 cells, the maximum fluorescence intensity measured after E2 and ICI treatments decreased by 20-40% as compared to untreated cells consistent with degradation of GFP-ERα (Figure 3Af and 3Ai compare to 3Ac). The effects of ICI and RU58 were indistinguishable suggesting that both molecules operate via similar molecular mechanisms despite significant structural differences 19.

ERα is a short-lived protein (half-life of >3 h for unbound ERα and ~ 1-3 h for ligand-bound ERα) 1031. ERα degradation occurs in presence of natural ligands (E2) or pure antiestrogens such as ICI in a proteasome dependent manner 131432.

The 26S proteasome is a large protein complex (1500-2000 kDa) present in the cytoplasm and nucleus of eukaryotic cells. The catalytic core of this multi-subunit complex, described as the 20S proteasome, contains α and β subunits 33. We visualized GFP-ERα and the 20S proteasome subunit α2 in SK19 cells. SK19 cells grown on glass coverslips and treated as described were fixed, permeabilized and subjected to indirect immunofluorescence using a monoclonal anti-20S proteasome subunit α2 primary antibody. Images acquired on an Olympus inverted wide-field microscope in 3 D and subjected to deconvolution revealed punctuate nuclear staining of proteasome subunits throughout the nucleus (Figure 4A and 4D, center panels). We did not observe any cytoplasmic staining of this proteasome subunit under our culture conditions. In the presence of E2, GFP-ERα accumulated at numerous nuclear sites that colocalized at least partially with proteasome foci (Figure 4A panel f).

Next we used a double-immuno-nanogold labelling approach in MCF-7 cells to characterize the extent of ER/α2 colocalization. Upon exposure to E2, at least four nuclear clusters per nuclear sections were detected. In the majority of clusters more than 3 gold particles for each protein were present (as indicated by circles in Figure 4B). Endogenous ERα (represented by 10 nm gold particles) colocalized with the 20S proteasome subunit α2 (represented by 6 nm gold particles) in nuclear microdomains of about 100 nm in diameter (Figure 4B, inset).

We then determined the effect of LMB (Leptomycin B), an inhibitor of the nuclear export receptor CRM1, and of ALLN (acetyl-leucyl-leucyl-norleucinal), an inhibitor of the proteasome, on SERD-dependent degradation of ERα in SK19 cells. SK19 cells were pretreated with 10 nM LMB or 100 μM ALLN for 30 min. Figure 4C shows that LMB did not block E2, ICI or RU58 induced ERα degradation suggesting that SERD-bound ERα is degraded in the nucleus. In the presence of E2, but not ICI or RU58, degradation was slightly less pronounced in cells pretreated with LMB suggesting that a fraction of E2 bound ERα is also degraded by the cytoplasmic proteasome. Furthermore, ALLN inhibited E2, ICI and RU58 induced degradation of ERα confirming that SERD-ERα complexes were degraded by the nuclear proteasome (Figure 4C, lanes 6, 9 and 12). Note that at the protein level, GFP-ERα is degraded to a lesser extent than endogenous ERα which is likely to be a consequence of reduced transcription of ESR1 in the presence of E2 and SERDs. GFP-ERα transcription is under the control of a CMV promoter which insensitive to antiestrogens.

Finally, we investigated the distribution of GFP-ERα and the 20S proteasome subunit α2 in SK19 cells treated with ICI or RU58 (Figure 4D). GFP-ERα foci also significantly overlapped with accumulation sites of the 20S proteasome subunit α2 throughout the nucleus (Figure 4D, insets of panels c and i). On average we observed larger and more frequent GFP-ERα-proteasome complexes in the presence of SERDs than in the presence of E2 consistent with the fact that ERα is readily degraded when ERα is bound to SERDs. As a consequence of the ALLN treatment, contacts between GFP-ERα and proteasome foci were largely abolished (Figure 4D panels f and l).

Interestingly, in a few cells treated with either E2 or SERDs we observed a single very large site of accumulation of the 20S proteasome α2 subunit (data not shown). These sites, also called clastosomes, were reported to colocalize with the c-jun and c-fos proteins 34, very unstable proteins with half lives of less than 90 min. In our cells, clastosomes did not colocalize with GFP-ERα foci which may indicate that E2 bound ERα is more stable than c-jun and/or c-fos proteins.

17β-estradiol (E2), 4-hydroxytamoxifen (OHT) and Leptomycine B (LMB) were purchased from Sigma-Aldrich (St. Louis, MO). ICI 182,780 (ICI) was purchased from Zeneca Pharmaceuticals. RU39,411 (RU39) and RU58,668 (RU58) were kindly provided by Dr. J.M. Renoir (Paris, France). Stock solutions of E2, OHT, ICI, RU39 and RU58 were prepared in ethanol. Stock solution of LMB was prepared in methanol. The solution of proteasome inhibitor acetyl-leucyl-leucyl-norleucinal (ALLN) was purchased from Calbiochem.

Rabbit polyclonal anti-ERα (HC-20), rabbit polyclonal anti-lamin A (H-102), rabbit polyclonal anti-cytokeratine 18 (H-80) were purchased from Santa Cruz Biotechnology, Inc. Mouse monoclonal anti-GAPDH (MAB374) was purchased from Chemicon International, mouse monoclonal anti-GFP from Roche, mouse monoclonal anti-α-tubulin (clone DM1A) from Sigma-Aldrich. Mouse monoclonal anti-20S proteasome subunit α2 (clone MCP21) was gift from Dr. M.P. Bousquet (IPBS, Toulouse, France).

All cell culture products were obtained from Invitrogen.

Human breast cancer cell lines were maintained in Dulbecco's modified Eagle's medium F-12 (DMEM F-12) with Glutamax containing 50 μg/ml gentamicin, 1 mM sodium pyruvate and 10% heat-inactivated fetal calf serum. All cells were grown at 37°C in a humidified atmosphere containing 5% CO₂. The stably transfected GFP-ERα reporter SK19 cell line was generated from ERα-positive breast cancer MCF-7 cells (ATCC). 2nd passage cells were transfected with a GFP-ERα expression vector (pEGFP-C2-hERα) using FuGENE^® HB Transfection Reagent (Roche Applied Science) and G418 resistant clones were selected at the concentration 1 mg/ml. GFP-ERα expressing clones were isolated, ERα protein expression in response to estradiol and to anti-estrogens was quantified using fluorescence microscopy and western blot. Expression of ERα-regulated genes was tested by qRT-PCR and compared to gene-expression regulation in MCF-7 cells. The clone SK19 in which GFP-ERα behavior was comparable to endogenous ERα was selected for further investigation.

To study the effects of estrogens and antiestrogens, cells were grown for 3 days in medium containing phenol red-free DMEM F-12 supplemented with 5% charcoal-stripped fetal calf serum, without gentamicin and sodium pyruvate. Cells were subsequently treated or not with 10 nM E2, 1 μM ICI, 1 μM OHT, 1 μM RU39, 1 μM RU58 for the indicated times. To study ERα degradation by the proteasome, cells were pre-treated 30 min with 100 μM ALLN, a proteasome inhibitor, or 10 nM LMB, a nuclear export inhibitor.

MCF-7 cells grown in 6-well plates were treated as indicated, washed with ice-cold PBS and collected by centrifugation. Total cell lysates were prepared by resuspension of cells in 100 μl lysis buffer (50 mM Tris pH = 6.8, 2% SDS, 5% glycerol, 2 mM EDTA, 1.25% β-mercaptoethanol, 0.004% Bromophenol blue). The samples were boiled for 20 min at 95°C and cleared by centrifugation at 12 000 × g for 10 min. Protein concentration was determined by an Amido Schwartz assay when the samples contained SDS. Samples were subjected to SDS-PAGE and proteins transferred onto nitrocellulose membranes. Western blot analysis was performed as previously described 48 using ERα and GAPDH antibodies and quantified using the TINA PC-Base Software from FUJI.

Total RNAs were extracted using TRIzol reagent (Invitrogen) following the manufacturer's protocol. 1-5 μg of total RNA was reverse transcribed in a final volume of 20 μl using SuperScript™III Reverse Transcriptase (Invitrogen). cDNA was stored at -80°C. All target transcripts were detected using quantitative RT-PCR (SYBRGreen SuperMix, Invitrogen) assays on a Mastercycler Realplex device (Eppendorf) using TBP or RPLP0 genes as endogenous control for normalization of the data. The following primer pairs were used for amplification:

TBP: 5'-CGGCTGTTTAACTTCGCTTTC-3'

5'-CCAGCACACTCTTCTCAGCA-3'

ESR1: 5'-TGGAGATCTTCGACATGCTG-3'

5'-TCCAGAGACTTCAGGGTGCT-3'

GREB1: 5'-GTGGTAGCCGAGTGGACAAT-3'

5'-AAACCCGTCTGTGGTACAGC-3'

RPLP0: (Fwd) 5'-TGGCAGCATCTACAACCCTGAA-3'

(Rev) 5'-ACACTGGCAACATTGCGGACA-3'

TFF1: (Fwd) 5'-CCCCTGGTGCTTCTATCCTAAT-3'

(Rev) 5'-CAGATCCCTGCAGAAGTGTCTA-3'

PGR: (Fwd) 5'-CTTAATCAACTAGGCGAGAG-3'

(Rev) 5'-AAGCTCATCCAAGAATACTG-3'

The results were analyzed using Mastercycler Realplex and qBASE software.

Three hours after incubation with ERα ligands, SK19 cells were washed with ice-cold PBS, scraped and centrifuged at 1,500 rpm for 5 min at 4°C. The pellets were resuspended in 150 μl digitonin lysis buffer containing 1% digitonin and 1 mM EDTA in PBS, immediately centrifuged at 13,000 rpm for 20 min at 4°C, to obtain the cytosolic fraction (C). The pellets were resuspended in 150 μl HEPES lysis buffer containing 1% Triton X-100, 10% glycerol, 10 μg/ml leupeptin, 5 μg/ml aprotinin, 1 mM PMSF, 1 mM Na₃VO₄ and 50 mM NaF in HEPES buffer (25 mM HEPES, 0.3 M NaCl, 1.5 mM MgCl₂, 20 mM β-glycerol-phosphate, 2 mM EDTA, 2 mM EGTA and 1 mM DTT), kept 15 min on ice and centrifuged at 13,000 rpm for 15 min at 4°C, to obtain the soluble nuclear fraction (SN). The pellets from the previous step were resuspended in 100 μl of a third buffer containing 95% Laemmli buffer and 5% β-mercaptoethanol and incubated 5 min on ice and boiled for 20 min at 95°C to obtain the insoluble nuclear fraction (IN). The different fractions were stored at -80°C until use. Protein concentrations were determined using the Bio-Rad Protein Assay (Bio-Rad).

For indirect immunofluorescence experiments, SK19 cells were grown for 3 days on coverslips in DMEM without phenol red, containing 5% charcoal-stripped fetal serum. After 3 days, cells were treated for 1 h with the following ligands: 10 nM E2, 1 μM ICI, 1 μM OHT, 1 μM RU39, 1 μM RU58. Cells were then washed twice with PBS, fixed in 4% paraformaldehyde/PBS for 10 min at room temperature, subsequently permeabilized with 0.5% Triton X-100 in PBS for 15 min at room temperature, counterstained with DAPI (4',6-diamidino-2-phenylindole) and mounted on microscopy slides.

To study co-localization of ERα and proteasome by immunofluorescence, SK19 cells were grown for 3 days on coverslips in DMEM without phenol red, containing 5% charcoal-stripped fetal serum and next treated for 3 h with drugs as indicated above. To block proteasome-mediated ERα degradation, the cells were incubated 30 min with 100 μM ALLN prior to treatment with ICI or RU58. Before immunostaining, cells were fixed in 4% paraformaldehyde/PBS for 30 minutes at room temperature, washed three times in PBS, quenched in 75 mM NH₄Cl containing 20 mM glycine and permeabilized with 0.5% Triton X-100 in PBS for 30 minutes. Next, cells were washed with PBS, blocked for 1 h at room temperature in 5% dry milk in TBS-T (20 mM TRIS-HCl, 150 mM NaCl, 0.1% Tween 20, pH = 7.4) and incubated overnight at 4°C with anti-20S proteasome antibody at a final concentration 2 μg/ml in 5% dry milk in TBS-T followed, after washing, by incubation with the Alexa Fluor^® 647 goat anti-mouse secondary antibody (1:1000, Invitrogen, Molecular Probes) for 90 min in the dark at room temperature. Finally, cells were washed with TBS-T, counterstained with DAPI and mounted on microscopy slides.

Cells were examined by fluorescence microscopy using an Olympus IX-81 microscope, equipped with a CoolSNAP HQ camera (Roper Scientific) and imaged through an Olympus oil-immersion objective 100x PLANAPO NA1.4. Images were recorded and deconvolved using Metamorph software (Universal Imaging). All images were processed for presentation using Adobe Photoshop 9.0.2.

MCF-7 cells were grown and treated as described above. For immune-electron microscopy cells were fixed with 4% paraformaldehyde in Na cacodylate buffer (pH 7.4), dehydrated in a graded series of ethanol and embedded in acrylic resin (LRWhite). 80 nm ultrathin sections were mounted on Nickel grids, incubated with 2% BSA/PBS and incubated overnight at 4°C with a mixture of primary antibodies (anti -20S proteasome antibody at final concentration 2 μg/ml and anti-ERα antibody (dilution 1/500)) in 2% BSA/PBS, washed 5 times for 5 mins in 1% BSA/PBS and then labeled for 1 h with 6 nm goat anti-mouse and 10 nm goat anti-rabbit gold conjugated particles in 1% BSA/PBS. Grids were finally washed 4 times for 5 mins in 1% BSA/PBS, incubated for 15 mins in 1% glutaraldehyde/PBS, washed 2 times for 5 mins in PBS, 3 times in distilled water and dried at room temperature. The samples were visualized using 120 kV Jeol electron microscope at 80 kV and images were captured using a digital camera AMT.