Differential Regulation and Targeting of Estrogen Receptor α Turnover in Invasive Lobular Breast Carcinoma

Abstract

Invasive lobular breast carcinoma (ILC) accounts for 10% to 15% of breast cancers diagnosed annually. Evidence suggests that some aspects of endocrine treatment response might differ between invasive ductal carcinoma (IDC) and ILC, and that patients with ILC have worse long-term survival. We analyzed The Cancer Genome Atlas dataset and observed lower levels of ESR1 mRNA (P = 0.002) and ERα protein (P = 0.038) in ER+ ILC (n = 137) compared to IDC (n = 554), and further confirmed the mRNA difference in a local UPMC cohort (ILC, n = 143; IDC, n = 877; P < 0.005). In both datasets, the correlation between ESR1 mRNA and ERα protein was weaker in ILC, suggesting differential post-transcriptional regulation of ERα. In vitro, 17β-estradiol (E2) decreased the rate of degradation and increased the half-life of ERα in ILC cell lines, whereas the opposite was observed in IDC cell lines. Further, E2 failed to induce robust ubiquitination of ERα in ILC cells. To determine the potential clinical relevance of these findings, we evaluated the effect of 2 selective estrogen receptor downregulators (SERDs), ICI 182,780 and AZD9496, on ERα turnover and cell growth. While ICI 182,780 and AZD9496 showed similar effects in IDC cells, in ILC cell lines, AZD9496 was not as effective as ICI 182,780 in decreasing ERα stability and E2-induced proliferation. Furthermore, AZD9496 exhibited partial agonist activity in growth assays in ILC cell lines. Our study provides evidence for a distinct ERα regulation by SERDs in ILC cell lines, and therefore it is important to include ILC models into preclinical and clinical testing of novel SERDs.

Keywords: breast cancer; endocrine response; estrogen receptor; invasive lobular carcinoma.

Figure 1.

Correlation of ESR1 mRNA and ERα protein levels in ER+ ILC and IDC tumors. A: ESR1 mRNA and RPPA ERα protein levels (B) in ER+ invasive lobular carcinoma (ILC) and ER+ invasive ductal carcinoma (IDC) samples analyzed from the TCGA data set. C:  PGR mRNA and PGR protein levels (D) from the TCGA data set. E: Correlation between RPPA ERα protein and ESR1 mRNA levels between ILC and IDC samples analyzed from the TCGA data set using Pearson’s (r) correlation. P = 5.9e-6 for Wilcoxon rank-sum test comparison of Pearson correlations. F:  ESR1 mRNA levels in IDC and ILC tumor samples as analyzed by qRT-PCR in the UPMC cohort. G: Immunohistochemical semiquantitation of ER in IDC and ILC tumor samples determined using a modified H-score. H:  PGR mRNA by qRT-PCR and immunohistochemical semiquantitation of PGR (I) in IDC and ILC tumor samples determined using a modified H-score in the UPMC cohort. J: Correlation between ER IHC H-score and ESR1 mRNA levels between ILC and IDC samples as analyzed using Pearson’s coefficient (r) correlation. P = 0.0017 for Wilcoxon rank-sum test comparison of Pearson correlations.

Figure 2.

ERα protein levels and estrogen response in ILC and IDC cell line models. A: Expression of ESR1 mRNA in IDC and ILC cell lines. Cell lines were grown in standard culture conditions and ESR1 mRNA levels were quantified by qRT-PCR. Data are shown as mean ± SEM from 3 biological replicates. *P < 0.05; ****P < 0.0001; calculated by 2-way ANOVA followed by Dunnett’s multiple comparison test; comparing the ESR1 mRNA levels to that of MCF-7. B: Expression of ERα protein in IDC and ILC cell lines. Western blot analysis of ERα in total protein extracts from IDC and ILC cell lines. The protein quantity is expressed as average fold change versus MCF-7 cells. Results represent the mean ± SEM of 3 independent experiments. C: Pelleted, fixed, and paraffin-embedded cells were immunostained for ERα. Representative images taken at 40x objective are shown (scale bar = 100 μM). D: Effect of 17β-estradiol (E2) on growth of IDC and ILC cell lines. Hormone-deprived cell lines were treated with Vehicle (Veh, 0.01% DMSO) or increasing doses of E2 (10^–14 to 10^–7 M) for 6 (MCF-7, T47D) or 7 (ZR-75-1, BCK-4, MDA-MB-134-VI, SUM44) days and proliferation assessed by FluoReporter® Blue Fluorometric dsDNA Quantitation Kit. Data are shown as fold growth versus Veh control. Points represent the mean of 6 biological replicates; error bars denote SD. E: Heat maps depicting gene expression changes (log₂ fold change [FC] vs Veh) after treatment with E2 ± ICI 182,780 in IDC and ILC cell lines. Cell lines were hormone-deprived and treated with Veh or 1 nM E2 ± 1 μM ICI 182,780 for 6 hours. GREB1 and PGR mRNA levels were quantified by qRT-PCR. Data expressed as log₂ FC versus Veh from 3 biological replicates.

Figure 3.

17β-estradiol (E2)-induced changes in ERα protein levels, ubiquitination, and turnover in IDC and ILC cell lines. Cells were hormone-deprived and treated with Vehicle (V, 0.01% DMSO) or 1 nM E2 for varying time points (0, 3, 6, 24, 48 hours) (A) or increasing doses of E2 (0.01, 0.1, 1, 10, 100 nM) for 24 hours (B). ERα protein levels were assessed by immunoblotting. Protein levels are expressed as a percentage of ERα remaining as compared to the corresponding Veh-treated controls. Results represent the mean ± SEM of 3 experiments. *P < 0.05; **P < 0.01; ****P < 0.0001; calculated by 1-way ANOVA followed by Dunnett’s multiple comparison test. C: Half-life of ERα protein calculated by cycloheximide (CHX) chase assay. Cells were treated with CHX (50 µg/ml) in complete growth media or in combination with Veh or E2 (1 nM) after hormone deprivation. ERα protein bands were normalized to β-actin and then to the time (0 hour) control. Half-life of ERα protein was calculated based on 1-phase decay. *P < 0.05; **P < 0.01 calculated by 2-way ANOVA followed by Bonferroni multiple comparison test. D: Effect of E2 on ERα ubiquitination. MCF-7 and MDA-MB-134-VI cells were pretreated with 10 μM MG132 for 30 minutes followed by 5-hour treatment with Veh, or E2 (1 nM). ERα was immunoprecipitated (IP) from the total protein lysates, and ubiquitination was evaluated by IB using a ubiquitin (Ub)-specific antibody. Input lanes represent 5% of the amount of protein lysates used for IP. The bottom panel shows the blots re-probed with ERα specific antibody.

Figure 4.

Oral SERD AZD9496 activity in blocking ILC and IDC growth, and ability to degrade ERα. A: Cells were hormone-deprived and treated with Vehicle (Veh, 0.01% DMSO) or increasing doses (10^–12 to 10^–6M) of SERDs ICI 182,780 (ICI) or AZD9496 (AZD) for 24 hours. ERα protein levels were assessed by immunoblotting. Protein levels are expressed as a percentage of ERα, remaining as compared to the corresponding Veh-treated controls. Results represent the mean ± SEM of 2 to 3 experiments. B: Half-life of ERα protein calculated by cycloheximide (CHX) chase assay. Cells were treated with CHX (50 µg/ml) in complete growth media or in combination with Veh, or ICI 182,780 (100 nM) or AZD (100 nM) after hormone deprivation. Cells were harvested after 0, 3, 6, 12, 24, and 48 hours. Protein lysates were prepared and ERα protein levels were assessed by Western blotting. ERα protein bands were normalized to β-actin and then to the time (0 hour) control. Half-life of ERα protein was calculated using GraphPad Prism software based on 1-phase decay. C, D: Hormone-deprived cells were treated with Veh, 100 nM ICI 182,780 or AZD with (C) or without (D) 100 pM E2 for 0, 3, 5, and 7 days, and proliferation was assessed by FluoReporter® Blue Fluorometric dsDNA Quantitation Kit. Data are shown as fold growth versus Day 0. Points represent mean of 6 biological replicates; error bars denote SD. **P < 0.01; ***P < 0.001; ****P < 0.0001; comparing the growth rate between treatment groups.