Clinical tissues

The test was approved by Jing Men No. 2 People’s Hospital. Sixty-four BC tissues (BC-S: 30 cases were sensitive to tamoxifen; BC-R: 34 cases were resistant to tamoxifen) and normal tissue (n = 64) were collected from Jing Men No. 2 People’s Hospital. All volunteers wrote the informed consent. Besides, the clinicopathological features of these patients were shown in  Table 1.

Table 1 Relationship between circTRIM28 expression and clinicopathologic features of BC patientsCell lines

The BC cell lines MCF7 and MDA-MB-231 with MCF10A as control. MCF7 cells a human breast cancer cell line with estrogen, progesterone and glucocorticoid receptors. When grown in vitro, the cells are capable of forming domes and the epithelial-like cells grow in monolayers. MDA-MB-231 cells were chosen as an ER negative cell line with invasive phenotype in vitro, and have epithelial-like morphology and appears phenotypically as spindle-shaped cells. All cells were purchased from the American type culture collection (ATCC, Manassas, VA, USA) and used at between three and ten passages. Cells were grown in RPMI 1640 supplemented with 10% fetal bovine serum (FBS) and penicillin/streptomycin at 37 °C in a 5% CO2 and 95% humidified atmosphere at a frequency of 2–3 times/week. According to the https://web.expasy.org/cellosaurus/CVCL_0023 website, the cells have no cross-infection. MCF7 and MDA-MB-231 cells were treated with 0.1 μM 4-hydroxytamoxifen (4-OH-TAM) (Sigma, St. Louis, MO, USA) for at least 6 months to produce anti-TAM BC cells, named MCF7/R and MDA-MB-231/R. The TAM-resistant phenotype of BC cells was maintained using the same medium supplemented with 0.5 μM TAM.

QRT-PCR

Whole RNAs were utilized by Trizol reagent (Takara, Tokyo, Japan) and reverse transcription as cDNA. Afterward, the cDNA level was quantified by SYBR Green kit (Takara). Successively, the data were assessed by the 2-△△Ct method and standardized by U6 or GADPH. Primers were listed in Table 2.

Western blot

The method described by Hou et al. [23]. The antibodies were listed: anti- HMGA2 (ab207301; 1:1000 dilutions; 0.693 μg/mL; Abcam, Cambridge, MA, USA), anti-PCNA (ab92552; 1:1000 dilutions; 0.164 μg/mL; Abcam), anti-Cleaved-caspase 3 (ab32042; 1:500 dilutions; 3.088 μg/mL; Abcam), anti-MMP9 (ab76003; 1:10,000 dilutions; 0.186 μg/mL; Abcam), and anti-β-actin (ab8227; 1:1000 dilutions; 0.1 μg/mL; Abcam).

Cell transfection

The sh-circTRIM28 (sh-circTRIM28#1, sh-circTRIM28#2, sh-circTRIM28#3), the control (sh-NC), miR-409-3p mimics, miR-409-3p inhibitors (anti-miR-409-3p) and controls, Bio-NC, Bio-miR-409-3p, Oligo probe, circTRIM28 probe, HMGA2 overexpression (HMGA2) and control plasmid (vector) were also fabricated by Ribobio (Guangzhou, China). Lipofectamine 2000 (Sigma) was employed in transfection.

RNase R and act D assay

The circ_0047921 and CD226 mRNA were treated with RNase R (Sigma). Similarly, actinomycin D (Act D, 2 mg/mL) or DMSO (Sigma) was added to the medium as control, and the tests were performed respectively. Finally, the expression level of circ_0047921 and CD226 mRNA were uncovered by qRT-PCR.

CCK8 assay and cytotoxicity assay

After post-transfection, BC cells (2.0 × 103/well) were seeded in 96-well plates. The cells in each well were exposed to different doses of tamoxifen (0, 5, 10, 20, 30 or 40 μM) for 48 h. Next, CCK8 (Sigma) was supplemented and the half-maximal inhibitory concentration (IC50) with assessed. A similar model was enforced to measure cell viability.

Cell proliferation assay

BC cells were seeded into 96-well plates. Then, the EdU Apollo In Vitro Imaging Kit (Sigma) was employed along with the guide. Generally, tumor cells were incubated with 50 μM EdU for 2 h at 37 °C and fixed in 4% formaldehyde. After being stained with Apollo reaction cocktail and DAPI (identify the nuclei) for 30 min, the proliferation-positive cells were photographed and counted under a microscope.

Colony formation assay

In short, 1 × 103 BC cells were plated into 6-well plates and maintained for 2 weeks. Finally, the colonies were fixed with methanol for 20 min and dyed with 0.1% crystal violet (Sigma) at room temperature. Finally, cell colonies were counted and photographed.

Flow Cytometry assay

BC cells with transfection were seeded in 6-well plates. The apoptotic was assessed by.

Annexin V-FITC/PI kit (Sigma) with a flow cytometer. In short, tumor cells were trypsinized and washed three times in PBS, followed by re-suspending in binding buffer. After being dual-stained with 5 μL Annexin V-FITC and 10 μL PI solution in the dark for 10 min, the apoptotic cells were identified according to a flow cytometer.

Transwell assay

BC cells were plated into a transwell with a pore polycarbonate membrane (BD Bio-sciences, Bedford, MA, USA). In brief, 4 × 105 transfected BC cells were planted into the top chamber. Then the lower chamber of the transwell contains 500 μL of DMEM and 10% FBS. Succeeding, the inferior surface of the membrane cells was stained. The same method was enforced to detect the invasion ability, but the transwell chamber was precoated with matrigel (BD Biosciences). Eventually, a light microscope was performed to validate the count of cells.

Dual-luciferase reporter assay

The targeted combination relationship of miR-409-3p and circTRIM28 or HMGA2 was predicted by circbank (http://www.circbank.cn/), starbase (http://starbase.sysu.edu.cn), and circinteractome (https://circinteractome.nia.nih.gov). Then, the circTRIM28 and HMGA2 wild and mutant were synthesized by Ribobio (circTRIM28-WT, HMGA2–3’UTR-WT or circTRIM28-MUT, HMGA2–3’UTR-MUT). Finally, luciferase activity was quantified.

RNA pull-down

Bio-miR-409-3p and Bio-NC were synthesized from RiboBio. The RNA pull-down assay was applied as beforehand reported [24]. Finally, circTRIM28 and miR-409-3p contents were assessed.

RIP assay

The Magna RNA immunoprecipitation kit (Millipore, Billerica, MA, United States) was enforced to carry out RIP assay in accordance with the guide. To be brief, tumor cells at 80% confluency were harvested, followed by lysing in complete RIP lysis buffer. Then, the obtained cell lysates were incubated with anti-AGO2 or anti-IgG for 4 h at 4 °C before treating magnetic protein A/G beads for 2 h. Subsequently, the beads were washed, and the extracted total RNA was subjected to qRT-PCR analysis.

In vivo assay

All-female BALB/C nude mice (6-week-old, 18–22 g) were bought from Shanghai Laboratory Animal Company (SLAC, Shanghai, China) under a pathogen-free environment at a temperature of 25 °C and relative air humidity between 45 and 50% with a 12/12 h day/night cycle. The in vivo work was approved by the guidance of the Animal Care and Use Committee of Jing Men No. 2 People’s Hospital. MDA-MB-231/R cells (1 × 106) with or without tamoxifen (5 mg/kg) transfected with the circTRIM28 knockdown vector (sh-circTRIM28) or sh-NC were vaccinated into mice (n = 6/group) under a specific-pathogen-free environment. The volume (mm3) = length×width2/2. Mice were sacrificed on day 35, the tumors were used for deeper level research.

IHC assay

In short, formalin-fixed and paraffin-embedded mice tumor tissue specimens were cut into 4 μm sections, followed by staining with hematoxylin and eosin for histopathology under a microscope. To analyze cell proliferation, the Ki67 (ab92742; 1:1000 dilutions; Abcam) contents in the tumor were detected by IHC assay. The unambiguous assessment scheme as per the description of Ma et al. [25]. Ultimately, the slides were photographed.

Statistical assay

These statistics were assembled from no less than 3 groups of repeats. The Student’s t-test or ANOVA was enforced to measure the difference in GraphPad Prism 7. P-value < 0.05 was significant.