1. Introduction
Breast cancer is one of the most common causes of death in U.S. women [
1]. Subtypes of breast cancer, such as luminal A, luminal B, triple negative/basal-like, and HER2 type, are based on molecular characteristics including hormone receptor (estrogen receptor (ERα), progesterone receptor (PGR) and HER2/neu (V-Erb-B2 erythroblastic leukemia viral oncogene homolog 2)) status [
2]. An estimated 70% of all newly diagnosed breast cancer patients are classified as ERα+ and hormone responsive, relying on estrogen stimulation to maintain tumorigenesis [
3]. Alterations in estrogen signaling pathways in ERα+ breast tumors can arise from increased growth factor signaling, such as epidermal growth factor receptor (EGFR) and AKT/mammalian target of rapamycin (mTOR) pathways, and more recently to ERα isoform variation [
2,
4,
5,
6,
7,
8]. ERα-36 is a 36 kDa ERα isoform originating from within a conventional ERα transcript. Expression of ERα-36, which possesses DNA and ligand binding activity but lacks activation function-1 (AF-1) and -2 (AF-2) transactivation domains [
4,
8], has been observed in both ERα-positive and -negative breast tumors and cell lines [
4,
8,
9]. ERα-36, which is localized to the plasma membrane and induces rapid estrogen signaling through mitogen activated protein kinase (MAPK) and phosphoinositide 3-kinase (PI3K)/AKT signaling pathways [
8,
10], is linked to tamoxifen resistance in breast cancers and is capable of eliciting agonistic tamoxifen induced signaling [
8,
9,
10].
Recently, altered ERα signaling by microRNAs (miRNAs) has been observed, with some miRNAs directly repressing ERα and other miRNAs inducing resistance to endocrine therapies [
11,
12,
13]. Deregulated miRNA processing and miRNA expression in breast tumor subtypes has also been reported [
14,
15,
16,
17,
18]. In addition, the miRNA biogenesis associated proteins are associated with breast cancer progression. For example, Argonaute 2 (AGO2), a key component of the miRNA silencing complex and mRNA translational regulatory protein, is directly regulated by ERα signaling via the MAPK pathway [
14]. Among breast tumor types, increased AGO2 expression in ERα− breast cancer cell lines and tumor samples has been observed [
14]. A positive correlation between AGO2 expression levels and the ERα− phenotype in breast cancer cell lines and tumor samples has been previously reported with AGO2 expression being regulated by EGFR/MAPK signaling [
14]. In this study we further evaluate the expression of AGO2 across breast cancer tumor samples with different degrees of receptor status (ERα, PGR, and HER2) and tumor molecular subtype. In addition, we show that a subset of ERα+ breast tumor samples, characterized as luminal B and ERα+/PGR−, demonstrated high AGO2 levels, similar to those observed in the ERα− tumor types. AGO2 expression correlates with a poor clinical outcome in ERα+ breast tumor samples.
3. Discussion
Patients with luminal B and ER+/PGR− tumors have a poor response to endocrine therapies, including tamoxifen. The underlying mechanism appears to be deregulation in estrogen receptor signaling pathways [
2,
29], due to crosstalk of growth factor signaling pathways. PI3K/AKT/mTOR and the epidermal growth factor receptor (EGFR) crosstalk with ERα signaling to enhance pro-proliferative ERα regulated gene expression and suppress PGR gene expression [
2]. AGO2, a key component of miRNA induced gene silencing, is regulated by the EGFR/MAPK signaling cascade and high AGO2 expression levels in ERα− breast cancers have been reported [
14]. Here, we demonstrate for the first time, high AGO2 expression levels in a subset of ERα+ breast tumors (luminal B and ERα+/PGR−), similar to AGO2 expression observed in ERα− tumors. Furthermore, we show that high AGO2 gene expression levels in ERα+ tumors correlate with a poor prognosis, in contrast to a lack of correlation between AGO2 expression levels and clinical outcome for ERα− tumor types. Among ERα+ tumor types, loss of PGR expression correlates with aberrant estrogen signaling, and we show that AGO2 expression in an ERα+ breast cancer cell line can repress classical ERα signaling (loss of ERα expression and loss of E2 stimulation of PGR;
Figure 2). Additionally, we show that despite the loss of classical estrogen signaling in vitro, estrogen-stimulated tumorigenesis is increased and there is no significant change. As a mechanism for altered ERα signaling, we next evaluated ERα-36 expression levels in the MCF-7 AGO2 cell line. AGO2 overexpression enhanced the expression of ERα-36 both at the gene and protein level. ERα-36 is involved in rapid estrogen signaling and its expression correlates with endocrine resistance. Surprisingly, MCF-7 generated with acquired endocrine resistance to tamoxifen (MCF-7-TAMR) and ICI (MCF-7-F) did not have enhanced ERα-36 gene expression. This study suggests that there may be a greater need in evaluating the alterations in miRNA biogenesis and associated genes in breast cancers demonstrating endocrine resistance.
4. Materials and Methods
4.1. Cells and Reagents
MCF-7 human breast cancer cell line was purchased from American Type Culture Collection (Manassas, VA, USA). MCF-7 and MCF-7-AGO2 validation of authenticity is provided as supplemental (Text S1 and S2 respectively). The MCF-7-TAMR and MCF-7-F cell lines were generated as previously described [
27]. Cells lines were cultured as previously described [
30]. Liquid nitrogen stocks were made upon receipt and maintained until the start of study. ERE-luciferase and/or qPCR for ERα and PGR were used to confirm MCF-7 sustained estrogen responsiveness. Morphology and doubling times were also recorded regularly to ensure maintenance of phenotype for all cell lines. Cells were used for no more than 6 months in culture. Cells were maintained in 10% fetal bovine serum (FBS) Dulbecco’s modified Eagle’s medium (DMEM) as previously described [
31]. MCF-7 parental cells were thawed at passage 65 and were not used past passage 80. The MCF-7-AGO2 cell line was used at passage 7 to passage 25. 17β-Estradiol (E2) was purchased from Sigma-Aldrich (St. Louis, MO, USA).
4.2. Transfection of MCF-7 Cell Line
Parental MCF-7 cell line (passage 65) was stably transfected with pIRES-vector or pIRES-AGO2 plasmid (Addgene plasmid 10821 and 45567, Cambridge, MA, USA) with Lipofectamine 2000 per manufacturer’s protocol (Invitrogen, Grand Isles, NY, USA). Parental MCF-7 cells were grown in 100 mm dishes. The plasmid (5 μg) was added to 100 μL serum free opti-MEM followed by 15 μL Lipofectamine. After 30 min incubation, opti-MEM containing the plasmid was added. The following day, pIRES-transfected cells were treated with 200 ng/mL neomycin. Cells were grown in 10% DMEM and treated with 200 ng/mL neomycin every two days for 2 weeks. Colonies were pooled and verification of AGO2 overexpression was confirmed using qPCR. Stable pools of transfected cells were maintained in 10% DMEM as described above and were not used beyond passage 25.
4.3. RNA Extraction and Quantitative Real Time RT-PCR
MCF-7-pIRES-vector and MCF-7-AGO2 cells were harvested for total RNA extraction using Qiagen RNeasy ( Valencia, CA, USA). Quantity and quality of the RNA were determined by absorbance at 260 and 280 nm using the NanoDrop ND-1000. Total RNA (1 µg) was reverse-transcribed using the iScript kit (Bio-Rad Laboratories, Hercules, CA, USA) and qPCR was performed using SYBR-green and 300 ng cDNA (Bio-Rad Laboratories). β-Actin, PGR, ERα, and ERα-36 genes were amplified (n = 3) using the following primers: ERα forward 5′-GGCATGGTGGAGATCTTCGA-3′, ERα reverse 5′-CCTCTCCCTGCAGATTCATCA-3′, ERα-36 forward 5′-CAAGTGGTTTCCTCGTGTCTAAAGC-3′, ERα-36 reverse 5′-TGTTGAGTGTTGGTTCCAGG-3′, PGR forward 5′-TACCCGCCCTATCTCAACTACC-3′, PGR reverse 5′-TGCTTCATCCCCACAGATTAAACA-3′, β-actin forward 5′-TGAGCGCGGCTACAGCTT-3′, β-actin reverse 5′CCTTAATGTCACACACGATT3′. For E2 stimulation experiments, cells were grown in 5% DMEM for 48 h prior to 18 h of treatment with 1 nM E2. Data were analyzed by comparing relative target gene expression to β-actin. Relative gene expression was analyzed using 2−ΔΔCt method. qPCR for miRNA was as follows: total RNA was extracted using the Qiagen miRNeasy kit as per the manufacturer’s protocol, small RNA fraction was not selected, 1.5 μg of total RNA was reverse transcribed using the Qiagen miscript II kit and qPCR was performed using miscript SYBR green and primers for U6 purchased from Qiagen. Normalization was to U6.
4.4. Western Blot
MCF-7-vector and MCF-7-AGO2 cells were grown in 10% FBS DMEM. Cells were washed with phosphate-buffered saline (PBS) and lysed with M-Per lysis buffer supplemented with 1% protease inhibitor and 1% phosphatase inhibitors (I/II) (Invitrogen). Supernatant containing protein extracts was obtained through centrifugation at 12,000 rpm (5415, Eppendorf, Westbury, NY, USA) for 10 min at 4 °C. Protein extracted per sample was determined by absorbance at 260 and 280 nm. Proteins were heat denatured and loaded on Bis-Tris-nuPAGE gel (Invitrogen). Protein transfer to nitrocellulose through iBlot and iBlot transfer stacks was per the manufacturer’s protocol (Invitrogen). Nonspecific binding was blocked by incubation in 3% milk (in 1% Tris buffered saline-Tween (TBS-T)) for 1 h. Overnight incubation of membrane with primary antibody for AGO2 diluted 1:1000 (Cell Signaling Technology, Beverly, MA, USA) and ERα (Santa Cruz Biotechnology, Santa Cruz, CA, USA) diluted 1:250, and ERα-36 diluted 1:500 at 4 °C followed by 3 × 15 min washes in 1% TBS-T. Membranes were incubated for 1 h in secondary antibody 1:10,000 dilution (LiCor Bioscience, Lincoln, NE, USA) followed by 3 × 10 min washes in 1% TBS-T. Band density was determined by LiCor gel imager. Normalization was to Rho GDI-α (Santa Cruz Biotechnology).
4.5. Animal Studies
Ovariectomized SCID/Beige female mice (4–6 weeks old, Charles River Laboratories; Wilmington, MA, USA) were allowed a 2-week period of adaptation in a sterile and pathogen-free environment with food and water ad libitum. Cells were harvested in the exponential growth phase using a PBS/ethylenediaminetetraacetic acid (EDTA) solution and washed. Viable cells (5 × 106) in 50 µL of sterile PBS suspension were mixed with 100 µL Reduced Growth Factor Matrigel (BD Biosciences, Bedford, MA, USA). Injections were administered into the mammary fat pad using 27 ½ gauge sterile syringes. Animals were divided into treatment groups of five mice each: MCF-7 control vector, MCF-7 control vector plus E2, MCF-7 cells transduced to overexpress AGO2, MCF-7 cells transduced to overexpress AGO2 plus E2. Placebo or E2 pellets (0.72 mg of estradiol-17β, 60-day release; Innovative Research of America; Sarasota, FL, USA) were implanted subcutaneously in the lateral area of the neck using a precision trochar (10 gauge). All procedures in animals were carried out under anesthesia using a mix of isoflurane and oxygen. Tumor size was measured every 2–3 days using digital calipers. The volume of the tumor was calculated using the formula: 4/3π LS2 (L = larger radius; S = shorter radius). Animals were euthanized by cervical dislocation after exposure to CO2. Tumors were removed and frozen in liquid nitrogen or fixed in 10% formalin for further analysis. All procedures involving these animals were conducted in compliance with State and Federal laws, standards of the U.S. Department of Health and Human Services, and guidelines established by Tulane University Animal Care and Use Committee. The facilities and laboratory animals program of Tulane University are accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care. (Protocol #: 4299R Approval Date: 3/8/2016)
4.6. Immunohistochemistry (IHC)
IHC was performed on 5 µm thickness sections made from paraffin-embedded tumor samples that were fixed with formalin 10% neutral buffered as described previously [
32]. Briefly, slides with tumor sections were deparaffinized in xylene, dehydrated in ethyl alcohol, rinsed in water and antigen retrieval was done with Diva declocker for 30 min in a steamer and then incubated with 3% hydrogen peroxide for 5 min for quenching endogenous peroxides. The slides were rinsed with deionized water and PBS and then were blocked by incubation in 10% normal goat serum for 30 min. After blocking, the sections were incubated overnight with anti-ERα rabbit monoclonal primary antibody. The source of the primary antibody and the dilutions used for IHC are as follows: ERα (1:100; SP1 Thermo Scientific, Waltham, MA, USA). After overnight incubation with primary antibody, slides were washed with PBS followed by 30 min incubation with biotinylated secondary antibody (Vector Labs, Burlingame, CA, USA), rinsed in PBS and incubated with ABC reagent (Vector labs) for 30 min. Finally, 3,3-diaminobenzidine (DAB) was added to the sections and color was allowed to develop for 5 min and counterstained with hematoxylin for 30 s. Internal negative control samples incubated with either non-specific rabbit IgG, or 10% goat serum instead of the primary antibody showed no specific staining. Slides were dehydrated and mounted using two drops of Permount.
4.7. MicroRNA PCR Array
MCF-7 cells were plated at a density of 1 million cells in 25 cm2 flasks in normal culture media (10% DMEM) and allowed to adhere overnight at 37 °C. Cells were harvested in PBS, collected by centrifugation, and total RNA extracted using the miRNeasy kit (Qiagen) according to manufacturer’s protocol. Quantity and quality of RNA were determined by absorbance (260, 280 nm). SABiosciences Breast Cancer miRNA PCR array was used to detect changes in miRNA as per the manufacturer’s protocol and SABiosciences SYBR green (Qiagen).
4.8. Data Sources
TCGA research network breast cancer gene expression data (RNA-seq deep sequencing data) were viewed through the University of California, Santa Cruz (UCSC) Cancer Genomics Browser. The breast invasive carcinoma TCGA data set (total of n = 1032 tumor samples) was used and analyzed for gene expression aligned through the Illumina HiSeq system (Illumina, San Diego, CA, USA). Gene signatures were based on receptor status (ERα, PGR, and HER2) and molecular subtype (Luminal A, Luminal B, HER2-enriched, and basal-like). The linear relationships between AGO2 and ERα isoform RNA expression levels were measured in TCGA data sets for breast and invasive cancer. Pearson correlation coefficients and corresponding p values were calculated for each isoform with coefficients and −log10 transformed p values plotted on the X and Y axis, respectively.
Targeted analysis of prognostic gene expression for AGO2, amongst a cohort of breast tumor samples, was performed using the Breast Cancer Gene-Expression Miner v3.0.
Table 1 designates ERα status, nodal status, and patient number. Kaplan–Meier analysis was performed from the “pool” of cohorts, meaning all data sets were merged from all studies and converted to a common scale with normalization. The prognostic impact of AGO2 was evaluated through the univariate Cox proportional hazard model obtained through pooled data [
24,
25].
