RPMI-1640 Medium, Dulbecco's modified Eagle's medium (DMEM) and fetal bovine serum (FBS) were obtained from Gibco (Grand Island, NY, USA). Antibodies against phospho-ERK (Thr202/Tyr204; #4370, 1:1,000), ERK (#4695, 1:1,000 WB), phospho-MEK (Ser217/221; #9154, 1:2,000), MEK (#9122, 1:2,000), were purchased from Cell Signaling Technology (Danvers, MA, USA). Antibodies against tubulin (#66240-1-Ig, 1:5,000), β-actin (#60008-1-Ig, 1:5,000), were purchased from Proteintech. The NCOA2 (ab1877, 1:5,000) antibody was purchased from Abcam (Cambridge, UK). Secondary antibodies (1:5,000) conjugated to horseradish peroxidase (HRP) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

To analyze copy number variation and mRNA expression of the indicated genes in breast cancers, we obtained different dataset by using the cBioPortal (www.cbioportal.org). The copy number of NCOA2 were analyzed in four independent dataset [METABRIC, Nat Commun 2016; BRCA, INSERM 2016; TCGA Pancancer Atlas; The Metastatic Breast Cancer (MBC) Project]. The mRNA expression of NCOA2 was available in the TCGA provisional dataset. To show the expression levels of NCOA2 in multiple cancer types, we determined the NCOA2 expression in the Ramaswamy Multi-cancer Statistics by using the Oncomine™ online system (www.oncomine.org). To study the effect of NCOA2 expression on the prognosis of patients with breast cancer, we generated Kaplan-Meier survival curve of breast cancer patients with low or high expression of NCOA2 by using the Gene Expression Profiling Interactive Analysis (GEPIA) online tool (http://gepia.cancer-pku.cn). GEPIA was also used to analyze the correlation (Pearson correlation coefficient) between NCOA2 and RASEF gene expression.

MDA-MB-231, and MCF7 cells were purchased from ATCC. T47D was purchased from Cell Bank of Shanghai Institutes for Biological Sciences of Chinese Academy of Sciences. Cells were cultured in RPMI-1640 Medium or DMEM supplemented with 10% FBS at 37°C in a humidified atmosphere with 5% CO₂. All constructs used in this study were generated via standard cloning strategy. Briefly, a FLAG-tagged NCOA2 expression vector was created by ligation of NCOA2 open reading frame (ORF) sequence into the lentiviral vector pHAGE (Addgene, MA, USA). Short interfering (si)RNA sequences were as follows: siNCOA2, GGG CTG TTA ACA TTA GC AA; si-negative control (NC), TTC TCC GAA CGT GTC AC GT; siRASEF-1, GCC TTT CTT CAG AGT GAG TTA; siRASEF-2, GCC AAG ATT AAT TCA GCC ATA. For short hairpin (sh)RNA cloning, the oligonucleotides encoding the above indicated siRNA sequences were synthesized with a loop sequence separating the complementary domains, and then cloned into the pLK0.1 vector (Sigma Aldrich, St. Louis, MO, USA). For lentivirus packaging, the recombinant plasmids or control pLKO.1 plasmids were then co-transfected into HEK293T cells with two helper plasmids, psPAX2 and pMD2G (Addgene). Forty-eight hours after transfection, the supernatants containing lentiviruses were harvested and filtered through a 0.45-μm-mesh filter. After lentiviral infection, the infected cells were selected with puromycin (Sigma Aldrich; 4 μg/mL) for 2 days.

After lentiviral infection, breast cancer cells were trypsinized and seeded in 6-well plates (1,500 cells/well). After 10 days (when colonies had formed), the plates were rinsed with PBS and the cells were fixed with 4% paraformaldehyde. After 15 min, the cells were stained with crystal violet for 10 min. After washing with PBS for three times, the stained colonies were observed and photographed.

After lentiviral infection, breast cancer cells were trypsinized, seeded in 96-well plates (3 × 10³ cells/well) and grown at 37°C. At the end of the indicated incubation time, cell proliferation was assessed using an MTS colorimetric assay (Promega, Madison, WI, USA). Upon media removal, MTS (20 μL) was added into the complete culture medium (100 μL) for 30 min. After incubation, the absorbance of the formazan product was recorded at 490 nm using the Spectra MAX 190 microplate reader (Molecular Devices, CA, USA).

At the indicated time points, the cells were trypsinized and washed with ice-cold PBS. For cell cycle distribution analysis, the cells were fixed with 70% ethanol and stained with PI/RNase/PBS (100 μg/mL PI and 10 μg/mL RNase A) buffer for 30 min at room temperature (20–25°C) in dark. For apoptosis analysis, the cells were stained using a PI and AnnexinV-FITC staining kit (BD Biosciences) for 30 min at room temperature in the dark. The stained single cell suspension was analyzed on a BD LSRFortessa SORP flow cytometer (BD Biosciences). The ModFit LT software (Verity Software House, Topsham, ME, USA) was used to analyze the flow cytometry data. For apoptotic analysis, the cells in the lower right quadrant are in early apoptosis and those in the upper right quadrant are in mid and late apoptosis. Generally, the two quadrants are scored as total (25).

Total cell lysates were prepared from breast cancer cells using RIPA lysis buffer (Applygen Technologies Inc., Beijing, China) supplemented with a protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany). Equal amounts of proteins were separated in SDS-polyacrylamide gels and transferred to a polyvinylidene difluoride (PVDF) membrane (Merck Millipore, Massachusetts, USA). The PVDF membrane was incubated with blocking buffer (5% non-fat dry milk in TBST) for 1 h and then with indicated primary antibodies at 4°C overnight. After three washes with TBST, the membrane was incubated with HRP-conjugated secondary antibodies for 1–2 h. Immunoblots were washed and visualized using the SuperSignal™ West Pico Chemiluminescent Substrate (Thermo Fisher Scientific, San Jose, CA, USA) and the ImageQuant LAS500 (GE Healthcare Bio-Sciences AB). The expression of tubulin or β-actin was used as a loading control.

For reverse transcription, 1 μg total RNA was reverse transcribed using the M-MLV Reverse Transcriptase Kit (Promega) following the manufacturer's protocol. Then following primers were used: Actin-F, 5′- GCA TCC CCC AAA GTT CAC AA-3′, Actin-R, 5′-AGG ACT GGG CCA TTC TCC TT-3′; NCOA2-F, 5′-TGG GGC CTA TGA TGC TTG AG-3′, NCOA2-R, 5′-GGT TTT TGA CAA ATT CCG TGT GG-3′; RASEF-F, 5′-TTC CCC TCA ACC TCT AGG CTA-3′, RASEF-R, 5′-CAA CTT CAC AAT TTG TCC TCT GC-3′. The PCR reactions (20 μL volume) contained: 2 × SYBR® premix Ex Taq™, 10 μL; forward and reverse primers (10 μM), 0.5 μL; cDNA, 4 μL; ddH2O, 5 μL. qPCR was performed on a CFX96 real-time detection system (Bio-Rad) according to the manufacturer's protocol. Each sample was run intriplicate.

Whole-transcriptome sequencing (RNA-Seq) was performed as described before (26). Briefly, total RNA was isolated from MDA-MB-231 cells with the QIAGEN RNeasy Plus kit (Qiagen, Valencia, CA, USA). The poly(A)-containing mRNAs were purified and enriched by PCR to create the final library according to the Illumina TruSeq RNA protocols. The libraries were sequenced on an Illumina High HiSeq 2000 with paired-end 100 base pair long reads. The raw sequencing data were examined using the FastQC software (http://www.bioinformatics.babraham.ac.uk/projects/fastqc). The reads were aligned to human genome sequences using TopHat2 (https://ccb.jhu.edu/software/tophat/index.shtml). Mapped reads were assembled with the Cufflinks (http://cole-trapnell-lab.github.io/cufflinks/). Read count per RNA was computed using HTSeq (https://htseq.readthedocs.io/en/release_0.10.0/). Log2 transformations were performed on normalized read counts. Differential expression analyses were performed using DESeq2 (https://bioconductor.riken.jp/packages/2.14/bioc/html/DESeq2.html). For functional annotation, the KEGG pathways were determined by analyzing dysregulated mRNAs using DAVID (https://david.ncifcrf.gov). Gene Set Enrichment Analysis (GSEA) was performed to evaluate significant enrichment of genes using GSEA software (http://software.broadinstitute.org/gsea/index.jsp).

For statistical analysis, the Student's t-test was used for parametric variables. All experiments were performed at least three times, and a P < 0.05 was considered statistically significant.

The copy number of the NCOA2 gene was analyzed by using the cBioPortal online tool (www.cbioportal.org) and found to be frequently amplified in four independent breast cancers datasets [METABRIC, Nat Commun 2016; BRCA, INSERM 2016; TCGA Pancancer Atlas; The Metastatic Breast Cancer (MBC) Project] (27–29), varying from 5 to 14% amplification (Figure 1A). Not surprisingly, the mRNA levels of NCOA2 were also upregulated in 118 (11%) of 1,082 sequenced cases/patients (TCGA provisional dataset; Figure 1B). Moreover, by cross-comparing the results obtained from the Ramaswamy Multi-cancer Statistics using the online Oncomine tool (www.oncomine.org), we found that the expression of NCOA2 in breast cancer was relatively higher than that in other cancer types, including prostate cancer, bladder cancer, lung cancer, and lymphoma among others (Figure 1C) (30), suggesting an important role for NCOA2 in regulating breast cancer development. Importantly, as revealed in Figure 1D, higher NCOA2 expression was significantly correlated with poor overall survival of patients with breast cancer, with a hazard ratio of 1.8 (log-rank p = 0.017).

To evaluate the biological role of NCOA2 in human breast cancer, we applied lentivirus-mediated RNA interference toward NCOA2 in three breast cancer cell lines (MDA-MB-231, ERα-, PR–, and HER2–; T47D, ERα+, PR+, and HER2–; MCF7, ERα+, PR+, and HER2+/–). Both colony formation (Figure 2A) and MTS cell proliferation (Figures 2C,D) assays indicated that NCOA2 knockdown significantly suppressed cell proliferation in different breast cancer cell lines. To rule out the unwanted toxicity induced by the off-target activity of the interfering RNAs (31) we generated a NCOA2-overexpressing construct with a silent mutation that made NCOA2 mRNA resistant to RNA interference. Ectopic expression of shRNA-resistant NCOA2 (dashed line, green) completely rescued the inhibitory effect of shNCOA2 treatment in MDA-MB-231 cells, demonstrating that the inhibition of cell growth resulted from NCOA2 depletion (Figure 2B).

Both NCOA2-depleted MDA-MB-231 and T47D cells showed a G2/M arrest, as demonstrated by cell cycle analysis (Figures 3A,C). At the same time, we detected significant apoptosis in these cells (Figures 3B,D). These results indicate that NCOA2 knockdown hampers breast cancer cells proliferation primarily by inducing cell cycle arrest and apoptosis.

To characterize the functional effects of NCOA2 knockdown, we assessed the gene expression profiles of NCOA2-depleted MDA-MB-231 cells by RNA-Seq. We identified 696 upregulated and 1,421 downregulated genes in NCOA2-depleted cells, as compared with the control group. All differentially expressed genes were subjected to pathway analysis using the DAVID online tool. Our data indicated that NCOA2 is implicated in the regulation of several essential pathways, including the apoptotic, p53, MAPK, and Ras pathways among others (Figure 4A). Next, GSEA analysis revealed that genes dysregulated upon NCOA2 depletion were enriched in the MAPK signaling pathway (Figure 4B). Furthermore, WB analysis confirmed that when NCOA2 was depleted in breast cancer cells, the activity of the MAPK/ERK signaling was strongly inhibited, as represented by the decreased protein levels of p-MEK and p-ERK (Figure 4C).

The Ras superfamily of small GTPases consists of more than 100 members, which can be divided into several principal families: Ras, Rho, Rab, Arf (ADP-ribosylation factors), and Ran (32). Studies have shown that many Ras superfamily members are involved in the regulation of the MAPK/ERK signaling pathway, which is an oncogenic signaling cascade essential for cancer cell growth (33–35). Interestingly, we identified that the expression of Ras and EF-hand domain containing (RASEF) was significantly downregulated upon NCOA2 depletion, as shown by RNA sequencing (Figure 4A) and qPCR analysis (Figure 5A). RASEF, a member of the Rab GTPase protein family, has been shown to positively regulate ERK signaling cascade in lung cancer (36). We confirmed that RASEF depletion could remarkably suppress the activity of the MAPK/ERK signaling cascade by using two independent shRNA sequences targeting RASEF (Figure 5B). Further, analysis of breast cancer tissue data from the TCGA database showed that RASEF and NCOA2 levels were strongly and positively correlated (R = 0.51, P < 0.0001; Figure 5C). Further, to determine whether RASEF downregulation accounted for the anti-proliferative effect of NCOA2 depletion, we assessed breast cancer cell growth after RASEF knockdown in MTS cell proliferation (Figure 5D) and colony formation (Figure 5E) assays. We found that RASEF depletion significantly inhibited the growth of MDA-MB-231 cells. Therefore, our data suggested that RASEF expression is controlled by NCOA2, and decreased RASEF levels might contribute to ERK deactivation and to the blockade of cell growth.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.