Glucocorticoid receptor modulation decreases ER-positive breast cancer cell proliferation and suppresses wild-type and mutant ER chromatin association

MCF-7 and T-47D cells were purchased from ATCC and cultured in DMEM supplemented with 10% FBS (Gemini Bio-Products, West Sacramento, CA) and 1% penicillin/streptomycin (Invitrogen, Waltham, MA) at 37 °C and 5% CO₂. MCF-7 HA-WT, HA-Y537S, and HA-D538G cells were a kind gift of S. Chandarlapaty (MSKCC) and were cultured in DMEM phenol-red free supplemented with 5% FBS, 1% Pen/Strep (Invitrogen, Waltham MA), 100 μg/mL Geneticin (Gibco, Gaithersburg, MD), and 100 μg/mL hygromycin B (Gibco, Gaithersburg, MD) at 37 °C and 5% CO₂ [16]. For MCF-7 HA-WT, HA-Y537S, and HA-D538G cells, MCF7 Tet-ON cells (Clontech, Mountain View, CA) were infected with retroviral vectors containing either doxycycline-inducible HA-tagged ER wild-type (WT) or Y537S or D538G mutants. For forty-eight-hour post-infection, the infected cells were selected with 500 μg/mL of hygromycin for a period of 14 days, in which afterwards, hygromycin concentration was lowered to 100 μg/mL for regular passaging of the stable cell lines [16, 17]. For all experiments, cells were seeded in normal growth medium. When cells reached ~ 60–80% confluence, they were placed in 2.5% charcoal-stripped serum (CSS) in phenol-red free DMEM for 48–72 h prior to hormone treatment. For hormone treatments, cells were treated with vehicle (Veh, ETOH), 10 nM E2 (Sigma-Aldrich, St. Louis, MO), 100 nM dexamethasone (Dex, Sigma-Aldrich, St. Louis, MO), 1 μM CORT125134 (C134, Corcept Therapeutics, Menlo Park, CA), 1 μM CORT118335 (C335, Corcept Therapeutics, Menlo Park, CA), or 1 μM CORT108297 (C297, Corcept Therapeutics, Menlo Park, CA). Final ETOH concentration did not exceed 0.2%. For HA-tagged cells, expression of the HA-tagged wild type or Y537S or D538G was induced following 0.5 μg/mL doxycycline (Sigma-Aldrich, St. Louis, MO) when cells were placed in CSS containing media. Cells regularly tested negative for mycoplasma using the Universal Mycoplasma Detection Kit (ATCC, Manassas, VA).

Cells were cultured in phenol red-free DMEM supplemented with 2.5% CSS and 1% Pen/Strep (Invitrogen, Waltham, MA) for 48 h, and cells were lysed with RIPA lysis buffer with phosphatase and protease inhibitors (Roche Diagnostics USA, Indianapolis, IN). Protein was quantified using Pierce BCA Protein Assay (Thermo Scientific, Waltham, MA) per manufacturer’s instructions. Protein (50 μg) was loaded per sample and resolved with SDS-PAGE. Membranes were blocked with 5% milk (Roche Diagnostics USA, Indianapolis, IN) or 5% BSA (Sigma-Aldrich, St. Louis, MO) in TBST. Membranes were immunoblotted with anti-GR (1:500, 41/GR, BD Biosciences, San Jose, CA), anti-ER (1:500, F10, Santa Cruz Biotechnology, Dallas, TX), anti-PR (1:1000, D8Q2J, Cell Signaling, Danvers, MA), anti-HA (1:1000, C29F4, Cell Signaling, Danvers, MA), anti-Cyclin D1 (1:100,000, EPR2241, Abcam, Cambridge, MA), anti-β-actin (1:1000, 8H10D10, Cell Signaling, Danvers, MA), or α-tubulin (1:5000, DM1A, Millipore, Burlington, MA). Densitometry analysis was performed using ImageJ version 1.52a. The intensity of Cyclin D1 and β-actin bands were quantified, and results are reported as a ratio of cyclin D1 band intensity/β-actin band intensity for each treatment condition.

MCF-7 and T-47D cells (2.5 × 10⁴) were seeded in 12-well plates. Cells were cultured in phenol red-free DMEM supplemented with 2.5% CSS and 1% Pen/Strep for 48 h and then treated with Veh (ETOH), 100 nM Dex/V, 10 nM E2/V, Dex/E2, 1 μM C335/E2, 1 μM C134/E2, or 1 μM C297/E2. Cells were harvested, and total live and dead cells were counted 0–8 days post-treatment using trypan blue exclusion. Experiments were repeated n = 3 times.

MCF-7 cells were seeded in 10-cm dishes. Cells were cultured in phenol red-free DMEM supplemented with 2.5% CSS and 1% Pen/Strep for 72 h. Cells were synchronized for 72 h in 0% CSS containing medium. Cells were treated with Veh (ETOH), 10 nM E2/V, 100 nM Dex/E2, 1 μM C335/E2, or 1 μM C134/E2 and fixed with 70% ETOH. Cells were stained with 0.02 mg/mL propidium iodide and examined by FACS. Cell cycle populations were determined using FlowJo Software (Ashland, OR).

The methylene blue proliferation assay was derived from Oliver and colleagues [18]. MCF-7 HA-tagged wild-type (WT), Y537S, or D538G cells were seeded in 96-well plates with 2.5% CSS containing medium and treated with vehicle (ETOH), 100 nM Dex/V, 10 nM E2/V, Dex/E2, 1 μM C335/E2, or 1 μM C134/E2 in triplicate or sextuplicate. Cells were fixed with methanol and stained with 0.05% methylene blue as described in [19].

MCF-7 cells were seeded in 10-cm dishes. Cells were cultured in phenol red-free DMEM supplemented with 2.5% CSS and 1% Pen/Strep for 48 h and treated with Veh (ETOH), 10 nM E2/V, 100 nM Dex/E2 for 4 h. Total RNA was harvested with RNeasy Kit (Qiagen, Germantown, MD) and examined by Affymetrix HG U133+2.0 microarray. The arrays were processed in R; and fold-changes were determined using a 1.3-fold cutoff (see Additional files 1 and 2) and examined by Ingenuity Pathway Analysis (IPA, Qiagen, Germantown, MD).

MCF-7 cells were incubated in phenol red-free DMEM containing 2.5% CSS for 4 days total (with a media change after 48 h) and treated with either ETOH (60 min), 100 nM dexamethasone (60 min), or 100 nM E2 (75 min) or pretreated with E2 for 15 min, followed by co-treatment with dexamethasone (i.e., 75 min E2 and 60 min dexamethasone). These hormone concentrations were used previously in GR ChIP-seq experiments [9]. DNA and associated proteins were crosslinked with 1% formaldehyde, and lysates were sonicated and immunoprecipitated as described previously [9]. ChIP experiments were conducted using the ChIP Assay Kit and the manufacturer’s protocol (EMD Millipore). Three-microgram ChIP-grade anti-ER HC-20 (sc-543x, Santa Cruz Biotechnology, Dallas, TX or sc-8002x, Santa Cruz, Dallas, TX) was used for immunoprecipitation. We also used normal rabbit IgG (#2729, Cell Signaling, Danvers, MA) as a negative control. Eluted ChIP DNA was purified using the PCR Purification Kit (Qiagen, Beverly, MA). ER ChIP-seq was performed on the Illumina HiSeq platform, generating raw reads for analysis (see Additional file 1). The sequence alignment and identification of peaks is described briefly. Sequence quality was assessed and aligned to the human genome (version hg19), and peaks were detected using MACS2 v2.1.1.20160309 [20].

For directed anti-GR or anti-ER ChIP experiments, MCF-7 cells were treated as described above. MCF-7 HA-tagged Y537S cells were incubated in phenol-red free medium with 2.5% CSS and 0.5 μg/mL doxycycline for at least 48 h and treated with vehicle (ETOH), 100 nM Dex, 1 μM C134, or 1 μM C335 for 15, 30, or 60 min. Cells were lysed as described above, and 3 μg ChIP-grade rabbit monoclonal anti-GR (GRXP, #3660, Cell Signaling, Danvers, MA), anti-ER (F10, sc-8002x, Santa Cruz Biotechnology, Dallas, TX), and anti-HA (F-7, sc-7392x, Santa Cruz Biotechnology, Dallas, TX) were used for immunoprecipitation. Normal rabbit IgG (#2729, Cell Signaling, Danvers, MA) or normal mouse IgG2a (E5Y6Q, #61656, Cell Signaling, Danvers, MA) was used for negative controls. Eluted ChIP DNA was purified using the PCR Purification Kit (Qiagen, Beverly, MA) and quantified by Qubit. GR and ER ChIP-seq peaks for specific genes (CCND1, CDK2, and CDK6) were visualized using the Integrative Genomics Viewer (The Broad Institute, Cambridge, MA). Primers for enh2 (CCND1) were previously published [21] and primers for CDK2 (CDK2-F 5′-CAGACTGCCTTCTATCCCAGA-3′; CDK2-R 5′-AGTGGCTTCTGGGAAAGGAA-3′) and CDK6 (CDK6-F: 5′-AGCTTAGCGCCTGAGAGATG; CDK6-R: CAGAGGCATCTGTTCTGCAA) putative enhancers were designed using Primer3 [22]. qPCR was carried out using PerfeCTa SYBR Green FastMix (Quanta Biosciences, Beverly, MA) and fold changes were calculated relative to vehicle-treated cells. FKBP5 [7] and TFF1 [7] enrichment was used as a control for GR or ER ChIP, respectively.

All studies were carried out in accordance with and approval of the Institutional Animal Care and Use Committee of The University of Chicago and the Guide for the Care and Use of Laboratory Animals. One week prior to cell implantation, female mice between 5 and 7 weeks of age began a diet of doxycycline-containing chow (TD.01306, Envigo, NJ). MCF-7 HA-Y537S cells (10 × 10⁶) were implanted into the 2nd thoracic mammary fat pad of SCID mice (Taconic, Rensselaer, NY) and allowed to grow. Mammary tumors were measured twice weekly by caliper, and tumor volume was calculated as described previously [23]. Mice were randomized into treatment groups, and when tumors reached ~ 200mm³, mice were treated with vehicle (1 ETOH:9 sesame oil, Spectrum, Gardena, CA), 20 mg/kg C134, or 20 mg/kg C335 intraperitoneally (i.p.) thrice weekly until the end of the study. Tumors were snap frozen for RNA analysis or fixed in 10% neutral buffered formalin for immunohistochemistry.

Xenograft tumors were homogenized in Blue Bullet Blender tubes (Next Advantage, Troy, NJ) with Buffer RLT (Qiagen, Germantown, MD), and RNA was extracted per manufacturer’s instructions. qRT-PCR was performed as described above.

For MCF-7 Y537S tumor xenograft IHC analysis, tissues were fixed in 10% neutral-buffered formalin and embedded in paraffin. Sections (5 μm) were stained with anti-Ki67 (1:300, SP6, Thermo Scientific, Waltham, MA), anti-Cyclin D1 (1:100, EPR2241, Abcam, Cambridge, MA), or anti-ER (1:50, RM-9101, Thermo-Fisher, Waltham, MA). Anti-Ki-67, anti-Cyclin D1, and anti-ER IHC immunoreactivity were scored by our pathologist (RL) blinded to the treatment conditions. Anti-Ki-67 IHC staining was evaluated by counting the percentage of positive cells (any intensity) in a full tumor section. Anti-ER IHC staining was evaluated by an H-score, which accounts for intensity and percent positivity and yields a score of 0–300 [24]. For anti-Cyclin D1 staining assessment, because the percentage of positive tumor cells was uniform > 95%, slides were scored as a binary high- versus medium/low-intensity staining and significance was determined by Fisher’s exact test. Differences in staining intensity were considered significant if p < 0.05. Data are presented as mean ± standard deviation (SD). Representative images were taken using a Nikon Eclipse Ti2 microscope at × 200 total magnification using identical settings for each slide.

Previous observations suggested that GR agonists inhibit ER-driven BC cell proliferation [3, 7, 9, 25, 26]. To validate these findings in our GR+/ER+ BC cell models (Additional file 3: Figure S1A), we treated GR+/ER+ MCF-7 and T-47D cells with vehicle (ETOH), dexamethasone (Dex, 100 nM), estradiol (E2, 10 nM), or Dex/E2 (Fig. 1a). Following co-treatment with Dex/E2, we indeed observed a significant reduction in cell proliferation compared to E2 treatment alone. However, GR liganding with the synthetic glucocorticoid Dex in the absence of E2 was not growth suppressive suggesting that GR specifically reduces E2-mediated proliferation (Fig. 1a).

Because long-term glucocorticoid treatment is not well-tolerated in patients due to metabolic side effects [27], we tested the effect of novel selective GR modulators (SGRMs) on ER-mediated BC cell proliferation. GR liganding with Dex results in an upregulation of canonical GR target genes (e.g., FKBP5) while the SGRMs C134 and C335 are GR antagonists with respect to canonical GR target gene expression (Additional file 3: Figure S1B). Surprisingly, we found that SGRMs [28‐30], similarly to Dex, resulted in significant inhibition of E2-mediated proliferation (Fig. 1b), suggesting that specific GR ligand binding per se results in slowed E2-mediated proliferation. The reduced cell number observed following several days of highly GR-selective liganding was not accompanied by increased cell death (Additional file 3: Figure S1C). However, cell cycle analysis showed slowing of progression from S to G2/M at 18 h in MCF-7 cells as reflected by increased S-phase accumulation (Fig. 1c). In the GR-negative T47D-Y cell line [31] (Additional file 4: Figure S2A), we observed that neither Dex nor the SGRMs CORT108297 [32] or C335 decreased E2-driven cell proliferation (Additional file 4: Figure S2B), suggesting a GR-specific mechanism of action. Furthermore, a radioactive E2 competition assay using recombinant ER ligand-binding domain (LBD) showed no radioactive E2 displacement by Dex, C134, or C335 (Additional file 4: Figure S2C), confirming that SGRMs are not interacting directly with ER LBD. Taken together, these data suggest that either a pure GR agonist (Dex) or a SGRM can inhibit ER-driven BC cell proliferation through a GR-mediated effect on ER activity.

To begin to determine how GR inhibits ER-mediated cell proliferation, we examined global gene expression in MCF-7 cells following E2, Dex, or E2/Dex co-treatment. Using Ingenuity Pathway Analysis (IPA), we found proliferative gene expression pathways to be significantly activated by E2; this activation was decreased by Dex co-treatment (Fig. 2a). Within these proliferation pathways, we consistently found three common E2-induced cell cycle genes (CCND1, CDK2, and CDK6). Moreover, the mRNA expression for these three ER-target genes was also consistently decreased by treatment with Dex or a SGRM. For example, in MCF-7 cells, we found that E2-induced CCND1 and CDK2 gene expression was significantly suppressed at 24 h following treatment with Dex, C134, or C335 (Fig. 2b, left panel). In T-47D cells, GR-selective ligands similarly resulted in suppressed expression of all three cell cycle genes at 24 h (Fig. 2b, right panel). Cyclin D1 protein expression by densitometry was also measured in Western blotting and found to be reduced by the addition of C335 (1.03 versus 1.20 for E2 alone, Fig. 2c). C134 addition did not show uniform reduction of Cyclin D1 amount. In T47-D cells, Cyclin D1 protein expression following the addition of either C134 (1.16) or C335 (1.21) was lower than Cyclin D1 expression after treatment with E2 alone (1.39, Fig. 2c, Western images representative of three independent experiments). Given the biological importance of CCND1, CDK2, and CDK6 in ER+ BC cell proliferation [33‐35], we next focused on examining ER-associated enhancer regions of these three genes in the context of GR modulation by SGRMs [21, 36, 37].

It has been reported that GR activation with Dex can cause chromatin remodeling resulting in differential ER chromatin association [8, 9]. We therefore determined whether in our model system GR and ER co-activation caused significant ER chromatin remodeling. Indeed, we observed both ER chromatin loss (n = 16,181 peaks) as well as gain of new sites (n = 10,265 peaks) at 60 min following GR liganding with only a relatively small number of ER peaks conserved (n = 2637) (Fig. 3a). This suggests rapid and highly dynamic global changes in ER association with DNA following GR liganding as previously reported [8, 9]. We next examined whether GR liganding influenced ER-chromatin association at specific enhancers for CCND1, CDK2, and CDK6 that had been previously identified using E2-activated ER ChIP [21] and enhancer (e)-RNA detection [36, 37], and further confirmed in our experiments at 60 min of E2 treatment (Fig. 3b). Motif analysis (Possum, https://​zlab.​bu.​edu/​~mfrith/​possum/​) of these three enhancer regions was also performed (Fig. 3c). The CCND1 enhancer (enhancer 2, enh2) demonstrated a previously known critical FOXA1 site [21] as well as several GR response elements (GREs) and AP1 REs, but no estrogen response elements (EREs) (Fig. 3c). Similarly, for CDK2 and CDK6 enhancer regions previously demonstrated to associate with activated ER, there were several GREs, FOXA1 REs, AP1 REs, and only a single ERE found in the CDK2 enhancer. These analyses are consistent with the current model of ER predominantly binding to chromatin indirectly via cooperative transcription factors, e.g. FOXA1, to regulate enhancer activity [38].

To determine whether ER-chromatin association is altered following GR liganding at enhancer regions for CCND1 [21], CDK2 [36], and CDK6 [36], we performed directed ER and GR ChIP. For the CCND1 enhancer [21], GR did not bind significantly in the absence of a GR ligand (Fig. 3d, top-left panel). However, in the context of ER activation, GR associated with chromatin as early as 15 min following C134 treatment and at 30 min following Dex or C335 treatment (Fig. 3d, left panel). In parallel, directed ER ChIP experiments revealed significantly increased ER bound to this CCND1 enhancer region at 60 min following E2 alone [21, 38] (Fig. 3e, left panel). However, ER binding was reduced in the presence of co-liganded GR (Fig. 3e, left panel). These findings suggest that GR chromatin association inhibits activated ER binding to the CCND1 enhancer. We next evaluated GR chromatin binding at the CDK2 enhancer (27 kb downstream of the TSS). In this case, GR association increased transiently at 15 min following E2 stimulation (Fig. 3d, middle panel), but was not found at later time points. As with the CCND1 enhancer, GR association following the addition of C134 peaked at 15 min (Fig. 3d, middle graph) and at 30 min following Dex or C335. As with CCND1, ER association with the CDK2 enhancer peaked at 60 min with E2 alone, but was relatively reduced in the presence of GR liganding (Fig. 3e, middle panel). Finally, we found that GR association with the known CDK6 enhancer (225 kb downstream from the TSS) showed a similar pattern in the context of ER activation; GR association peaked at 15 min following C134 treatment and at 30 min following Dex or C335 treatment (Fig. 3d, right panel). ER binding to the CDK6 enhancer was similarly strongest without GR liganding (Fig. 3e, right panel). Interestingly, we noted that C134 induced more rapid maximal GR chromatin association (15 min) compared with Dex or C335 (30 min) at the same compound concentrations (Fig. 3d, right panel), suggesting that different GR ligands produce similar anti-proliferative phenotypes with slightly different kinetics. In summary, for these three pro-proliferative gene enhancer regions, GR chromatin binding induces relatively suppressed ER chromatin occupancy and is associated with subsequently decreased ER-mediated CCND1, CDK2, and CDK6 gene expression.

Activating ER mutations are found in up to 40% of ER+ BCs that develop resistance to aromatase inhibitor therapy; ER Y537S and D538G are the most commonly observed [16, 39, 40]. Having discovered that GR inhibited estradiol-activated gene expression and proliferation in wild-type ER+ BC cells, we next examined whether GR similarly inhibited constitutively-active mutant ER activation of CCND1, CDK2, and CDK6. Several labs recently reported that both E2-activated wild-type and non-liganded mutant ER target many of the same enhancer regions [39, 41, 42]. We therefore hypothesized that GR modulation could suppress mutant ER transcriptional activity and resultant cell proliferation. Mutant ER (D538G and Y537S)-expressing MCF-7 cells (Additional file 5: Figure S3A) demonstrated significantly increased proliferation compared to wild-type ER+ MCF-7 cells (Additional file 5: Figure S3B) [16, 17]. MCF-7 HA-WT cells showed a significant increase in proliferation following E2 treatment and a significant reduction of E2-mediated proliferation following GR liganding with Dex or SGRMs (Additional file 5: Figure S3C, left panel). We then examined proliferation of MCF-7 cells expressing Y537S or D538G ER following GR-selective liganding and found a relative reduction in cell proliferation (HA-Y537S, Fig. 4A and HA-D538G, Additional file 5: Figure S3C, right panel). We also identified a significant decrease in CCND1 gene expression following C134 or C335 treatment at 4 and 6 h in MCF-7 HA-Y537S cells (Fig. 4b, left panels). CDK2 gene expression was significantly decreased following C335 at 4 h and all three ligands at 6 h (Fig. 4b, middle panels). For CDK6, gene expression was significantly reduced following Dex treatment at both 4 and 6 h (Fig. 4b, right panels).

For MCF-7 HA-D538G cells, CCND1 steady-state mRNA was also significantly reduced following Dex and C134 treatment at 4 h (Additional file 5: Figure S3D, left panel). Similarly, CDK2 steady-state mRNA expression was significantly decreased following all three ligands at 6 h (Additional file 5: Figure S3D, middle panel). CDK6 steady-state mRNA expression was only significantly reduced following C134 treatment in the MCF-7 HA-D538G cells (Additional file 5: Figure S3D, right panel) demonstrating target gene-specific variability of SGRM activity with respect to the D538G ER mutation.

We confirmed robust Y537S ER chromatin association at the same CCND1, CDK2, and CDK6 enhancer regions previously identified with wild-type ER ChIP (Fig. 3b and Additional file 6: Figure S4) [39]. To determine the effects of GR-selective liganding on mutant ER chromatin binding at these enhancer regions, we performed directed GR and mutant ER (HA-ER) ChIP in MCF-7 Y537S ER cells. For CCND1, GR bound to the same regulatory region in mutant ER-expressing cells as in wild-type ER cells with chromatin association peaking at 30 min following Dex and C335 treatment (Fig. 5a, left panel). Similarly, GR bound to the CDK2 enhancer region and chromatin association peaked at 30 min following Dex and C335 treatment (Fig. 5a, middle panel). At the CDK6 enhancer, GR liganding resulted in GR chromatin association peaking at 15 min following Dex, 30 min following C134, and 60 min following C335 treatment (Fig. 5a, right panel). Mutant ER binding was more variable (Fig. 5b). For the CCND1 enhancer, mutant ER binding was reduced with all GR ligands at 60 min, while CDK2 and CDK6 enhancer occupancy was not affected (Fig. 5b). Taken together, these results suggest that GR-selective liganding is associated with suppression of mutant (Y537S) ER chromatin association at the CCND1 enhancer; while CDK2 and CDK6 gene expression downstream of GR activation occurs, it appears be through a mechanism that does not involve loss of mutant ER by 60 min.

Having discovered that GR liganding suppresses mutant HA-Y537S ER chromatin association with the CCND1 enhancer, we next wished to determine whether SGRM treatment could also reduce mutant Y537S ER tumor growth in vivo. SCID mice were fed doxycycline-containing chow for 1 week prior to implantation of MCF-7 HA-Y537S cells in the mammary gland as previously described [16]. When tumors reached ~ 200 mm³ (Additional file 7: Figure S5A), mice were treated with vehicle, 20 mg/kg C134 or C335 thrice weekly. After 18 days of SGRM treatment (when vehicle-treated control animals had achieved maximally allowed tumor growth), we measured tumor volumes in C134 and C335 SGRM-treated mice (Fig. 6a). We found a significant decrease in average tumor volumes of SGRM-treated animals suggesting decreased cell proliferation of activating mutant ER tumor cells (Fig. 5a). In addition, tumor growth over time was significantly slowed in C134- and C335-treated mice compared to vehicle-treated mice (Additional file 7: Figure S5B). There were no overt signs of animal toxicity during SGRM treatment and body weights remaining relatively stable (Additional file 7: Figure S5C-D). SGRMs also led to significantly increased progression-free survival times (PFS) as defined by tumor volume reaching > 1000 mm³ (28 days for SGRMs versus 18 days for control animals, Fig. 6b). All tumor treatments resulted in similarly strong tumor ER intensity and percentage positivity by IHC staining (Additional file 8: Figure S6A, C). Tumor RNA isolation showed significantly decreased steady-state expression of CCND1 following either C134 (p = 0.0055) or C335 treatment (p = 0.0105, Fig. 6c). While IHC analysis revealed that all tumors had > 90% of cells with some Cyclin D1 positivity, staining intensity differed between treatment groups. In C335-treated mice, 5/5 tumors exhibited significantly lower intensity Cyclin D1 staining compared to vehicle-treated mice (p = 0.0476, Fig. 6D and Additional file 8: Figure S6C). For C134-treated mice, Cyclin D1 IHC staining was of relatively low intensity in 3/5 tumors (p value comparison with vehicle-treated tumors (1/5) was not statistically significant p = 0.52, Fig. 6d). Tumors from C134- and C335-treated mice also showed a trend toward a decreased percentage of Ki-67-positive cells (p = 0.384 for C134 vs. vehicle and p = 0.142 for C335 vs vehicle, Additional file 8: Figure S6B-C). Taken together, these data demonstrate that GR-selective liganding can decrease activating mutant (Y537S) ER tumor growth in association with reduced tumor CCND1 mRNA expression.