Cell culture

Breast cancer cell lines, MCF7 and MDA-MB-231, and the HEK293T cell line were cultured in Dulbecco’s Modified Eagle Medium (GIBCO, Carlsbad, CA, USA). MDA-MB-361 cells were cultured in L-15 medium (GIBCO). ZR-75-30 cells were cultured in Dulbecco's Modified Eagle Medium: Nutrient Mixture F-12 (GIBCO). T47D cells were cultured in RPMI 1640 medium (GIBCO). All breast cancer cell lines and HEK293T were supplemented with 10% fetal bovine serum (FBS) (HyClone, Logan, UT, USA) and 1% penicillin–streptomycin solution (GIBCO). All cell lines were incubated at 37 °C in a humidified incubator with 5% CO2 under normoxia. To verify whether circAAGAB was oxygen-responsive, cells were cultured in a hypoxic chamber (InVivO2-200, Ruskinn Technology, Bridgend, UK) filled with 0.5% O2, 5% CO2, and 94.5% N2 for 24 h. After incubation under hypoxia, cells were moved to the humidified incubator with normoxic conditions for another 24 h to mimic re-oxygenation.

Cell line authentication

Cell experiments were performed on cells that were passaged less than 15 times and were routinely tested for mycoplasma using PCR Mycoplasma Detection Kit (ABM Inc., Vancouver, Canada). The cell lines were purchased and authenticated by the Bioresource Collection and Research Center, Food Industry Research and Development Institute (Hsinchu, Taiwan).

Plasmid construction, RNA interference, and miRNA overexpression

To overexpress circular RNA circAAGAB, the circAAGAB sequence was inserted into the DNA plasmid pCIRC2T7 to construct pCIRC2T7-circAAGAB (BioMed Resource Core (BMRC) of the 1st Core Facility Lab, College of Medicine, National Taiwan University). To check the interaction and binding site between circAAGAB and miR-378 h, plasmid pMIR-REPORT-circAAGAB (BMRC) was constructed by inserting the circAAGAB sequence behind the sequence of firefly luciferase. In addition, pMIR-REPORT-circAAGAB-mut (BMRC) was constructed by mutating the putative binding site on circAAGAB. To knock down the expression of FUS, MDA-MB-231 cells were transfected with 5 μM of small interfering RNA (siRNA) (Dharmacon, Lafayette, CO, USA) in 2 mL medium for 48 h. To overexpress miR-378 h, MDA-MB-231 cells were transfected with 5 μM of miR-378 h mimics (Dharmacon) in 2 mL medium for 48 h. After transfection, MDA-MB-231 cells were lysed to extract total RNA.

Genomic DNA extraction, RNA isolation, reverse transcription, and quantitative RT-PCR

Genomic DNA (gDNA) was extracted by QIAamp DNA Kits (Qiagen, Hilden, Germany). Cells were lysed by Nucleozol reagent (Machery-Nagel, Düren, Germany) and total RNA was purified according to the manufacturer’s protocols. Subsequently, total RNA was reverse-transcribed to complementary DNA (cDNA) using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Carlsbad, CA, USA). For reverse transcription of miRNA, SuperScript IV Reverse Transcriptase (Invitrogen, CA, USA) was used with the primers (Table 1). cDNA acted as the template for quantitative RT-PCR with OmicsGreen qPCR MasterMix (OmicsBio, New Taipei City, Taiwan), and the cycle threshold (Ct) value was detected by StepOnePlus Real-Time PCR System (Thermo Fisher, Waltham, MA, USA). The relative gene expression was evaluated using the 2−∆∆Ct method.

Table 1 The primers for quantitative RT-PCRWestern blotting

MDA-MB-231 cells were lysed in RIPA lysis buffer (Millipore, Billerica, MA, USA) with 0.1% SDS and sonicated. Subsequently, the proteins diluted by sample buffer were separated by sodium dodecyl sulfate‑polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred onto polyvinylidene difluoride (PVDF) membranes (Bio‑Rad, Hercules, California, USA). The membranes containing proteins were blocked in Lightning Blocking Buffer (ArrowTec Life Science, Taiwan) for 5 min. Afterwards, membranes were incubated with primary antibodies against FUS (cat. no. A5921; ABclonal, Woburn, MA, USA), EGR1 (cat. no. 4154; Cell Signaling, Danvers, MA, USA), ATF3 (cat. no. 18665; Cell Signaling), phospho-p38 MAPK-T180/Y182 (cat. no. AP0526; ABclonal), total p38 MAPK (cat. no. A4771; ABclonal), vimentin (cat. no. ARG66199; Arigo Biolaboratories, Hsinchu, Taiwan), E-cadherin (cat. no. 3195; Cell Signaling), phospho-histone H2AX-Ser139 (cat. no. 2577; Cell Signaling), caspase-3 (cat. no. MBS8811560; MyBioSource, San Diego, CA, USA), GAPDH (cat. no. 2118; Cell Signaling), and ACTB (β-actin; cat. no. 4968; Cell Signaling) overnight at 4 °C. After washing 3 times with Tris-buffered saline with 1% Tween-20, the membranes were hybridized with horseradish peroxidase-conjugated secondary antibodies at room temperature for 1 h. Protein expression was visualized by an enhanced chemiluminescence substrate (Millipore, Billerica, MA, USA) and imaged using a BioSpectrum Imaging System (UVP, Upland, CA, USA). The intensities of bands were analyzed using ImageJ 1.48v (National Institutes of Health, USA).

RNase R treatment

To confirm the circular structure of circAAGAB, total RNA was treated with 3 U RNase R (Lucigen, LGC Ltd, Teddington, UK) and 10X reaction buffer (Lucigen), and then incubated at 37 °C for 20 min. After RNA reverse transcription and PCR, PCR products were subjected to gel electrophoresis and visualized by a UVP Gel Solo system (Analytik Jena US, Upland, CA, USA).

RNA pull-down assay

A total of 1 × 107 MDA-MB-231 cells were lysed by cell lysis buffer (25 mM Tris–HCL pH 7.5, 150 mM NaCl, 1 mM EDTA, 0.5% NP-40) for each sample. Then, a Pierce Magnetic RNA–Protein Pull-Down Kit (Thermo Fisher) was used according to the manufacturer’s protocols. Magnetic beads conjugated with streptavidin were incubated with 3 μg biotinylated DNA oligo probe for 1 h. After the complex was formed, cell lysate was added into tubes containing the beads and incubated for 4 h. Subsequently, the complex was washed 4 times with wash buffer from the kit and 500 μL of cell lysis buffer. Finally, washed samples were measured by western blotting.

Actinomycin D treatment

At approximately 60% confluency, MDA-MB-231 cells were transferred into 6-well plates. Cells were treated with 5 μg/mL actinomycin D (Sigma, Saint Louis, MO, USA) dissolved in DMSO and collected at the indicated time points. Total RNA was purified after the cells were lysed. After treatment with actinomycin D, the RNA expression levels of circAAGAB were analyzed by quantitative RT-PCR.

Nuclear-cytoplasmic fractionation

For nuclear-cytoplasmic fractionation, RNAs in 8 × 105 cells were extracted by a Cytoplasmic & Nuclear RNA Purification Kit (Norgen Biotek Corp., Ontario, Canada) according to the manufacturer’s protocols. The isolated RNA was detected by quantitative RT-PCR and normalized to GAPDH (cytoplasmic control) or BCAR4 (nuclear control).

RNA fluorescence in situ hybridization (RNA-FISH)

MDA-MB-231 cells (6 × 105) were first seeded in cover glass chamber (80826, ibidi, Martinsried, Germany) overnight, washed with PBS once, and fixed with 4% paraformaldehyde for 10 min. After washing with PBS twice, cells were dehydrated with 70% EtOH for 2 h, and incubated with RNA probe (125 nM) at 37 °C overnight. The RNA labeling probes conjugated with 5ʹ modification 6-FAM (TTC CAA GGA TAT CAT TCT TCA TCA) were designed to target the back-splicing site of circ-AAGAB. Next, the cells were washed at 37 °C for 30 min, and mounted with Mounting Medium containing DAPI (ab104139, Abcam, Cambridge, UK). Finally, the images were acquired using a ZEISS LSM880 laser confocal microscopy (ZEISS, Heidelberg, Germany).

Luciferase reporter assay

HEK293T cells were cultured in 24-well plates with 4 × 104 cells per well. Cells were co-transfected with 50 μg of pMIR-REPORT-circAAGAB or pMIR-REPORT-circAAGAB-mut, 2 × 10–2 ng of miR-378 h mimics or mimic control, and 1 μg of Renilla luciferase vector as the transfection control. After transfecting for 48 h, cells were lysed by 100 μL luciferase lysis buffer (92.8 mM K2HPO4, 9.2 mM KH2PO4 and 0.2% Triton X-100 in ddH2O) on ice for 15 min and then centrifuged at 12,000xg relative centrifugal force for 2 min at 4 °C. Afterwards, the supernatant was isolated, and the luciferase signal was measured using the Dual-Glo Luciferase Assay System (Promega, Fitchburg, WI, USA).

Microarray analysis

MDA-MB-231 cells were transfected with 4 μg of pCIRC2T7-circAAGAB plasmids and total RNA was purified. Subsequently, microarray experiments were done through the service of the Core Instrument Center, National Health Research Institutes (Miaoli, Taiwan). Briefly, human single-stranded cDNA was generated from the amplified complementary RNA with the WT Plus cDNA Synthesis Kit (Affymetrix, Santa Clara, CA, USA), and then fragmented and labeled with the WT Terminal Labeling Kit (Affymetrix). RNA expression profiling was performed using the Clariom S Assay (Affymetrix). After scanning, the data from the Affymetrix microarray was normalized by robust multichip averaging. Visual representation of expression profiles was evaluated by principal component analysis (PCA) and hierarchical clustering by the Genesis 1.7.7 program (Graz University of Technology, Graz, Austria). Interactions between genes, biological functions, and pathways were analyzed by Ingenuity Pathway Analysis (IPA, Ingenuity Systems Inc., Redwood City, USA). The datasets generated during the current study are available in the Gene Expression Omnibus repository (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE158779).

Colony formation assay

MDA-MB-231 cells were cultured in 6-well plates with 2 × 105 cells per well and transfected with 4 μg of pCIRC2T7-circAAGAB plasmids for 24 h. Cells were reseeded into 6-well plates with 500 cell per well. After 3 weeks, cells were fixed by fixing buffer containing 75% methanol and 25% acetate (Sigma), and 0.1% crystal violet (Sigma) was added to dye the cells. Colonies of at least 50 cells were counted manually.

Cell migration and invasion assay

MDA-MB-231 cells were seeded in 6-well plates with 2 × 105 cells per well and transfected with pCIRC2T7-circAAGAB plasmids for 24 h. For cell migration, 6 × 105 cells were seeded into the transwell chambers and incubated for 24 h. For cell invasion, 1 × 106 cells were seeded into the transwell chambers coated with Matrigel and incubated for 48 h. After migration and invasion, cells were fixed by fixing buffer containing 75% methanol and 25% acetate (Sigma), and 0.1% crystal violet (Sigma) was added to stain the cells. Cells on the membrane were counted manually.

Ionizing radiation (IR) treatment

MDA-MB-231 cells were seeded and transfected with pCIRC2T7-circAAGAB plasmids for 24 h. Then, the cells were exposed to 10 Gy of γ-rays by an IBL-637 Cesium-137 γ-ray machine (Cis-Bio International, Ion Beam Applications, Belgium) and harvested at 24 h after irradiation. Finally, the cells were stained by propidium iodide (PI) and analyzed on a Beckman Coulter FC500 instrument (Beckman Coulter, Inc.) using CXP Analysis Software v2.3 (Beckman Coulter, Inc.).

Cell apoptosis and cell cycle analysis

MDA-MB-231 cells were seeded in 10 cm dishes with 1 × 106 cells per dish and transfected with pCIRC2T7-circAAGAB plasmids for 24 h. Cells for examining apoptosis were harvested and stained by using a FITC Annexin V Apoptosis Detection Kit I (PharMingen, BD Biosciences, NJ, USA) according to the manufacturer’s protocols. Cells for examining the cell cycle were harvested and fixed by 75% ethanol at − 20 °C overnight, then lysed with 0.5% Triton X-100 (Amersham, Little Chalfont, Buckinghamshire, UK), subjected to RNase A (Qiagen) treatment (20 ng RNase A/mL in PBS) and stained with PI (BD Biosciences, NJ, USA) solution (20 μg PI/mL in PBS) in the dark for 10 min. Afterwards, cell cycle and cell apoptosis were detected by a Beckman Coulter FC500 instrument (Beckman Coulter, Inc.) using CXP Analysis Software v2.3 (Beckman Coulter, Inc.).

Statistical analysis

All quantitative data are presented as the means ± standard deviations of data from at least three independent experiments, and an unpaired Student’s t-test was used to compare differences between groups. All analyses were performed using Microsoft Office Excel software, and a P-value < 0.05 was considered to be statistically significant.