TNBC tissues and the corresponding adjacent normal tissues were collected from 40 patients (aged ~26–58 years) with TNBC who underwent tumor resection at Tangshan People's Hospital (Tangshan, China) between April 2019 and July 2020, upon obtaining written informed consents from the donors. ADM-resistant tissue samples (n=15) and ADM-sensitive tissue samples (n=25) were acquired from patients who met the Response Evaluation Criteria in Solid Tumors (15). The present study was approved (approval no. TNBC20190304) by the Ethics Committee of Tangshan People's Hospital (Tangshan, China).

Normal mammary epithelial cell line MCF-10A (CRL-10317), and TNBC cell lines MDA-MB-231 (CRM-HTB-26), MDA-MB-453 (HTB-131), MDA-MB-157 (HTB-24) and BT549 (HTB-122) were obtained from American Type Culture Collection (ATCC). ADM-resistant cell line MDA-MB-231/ADM (IMD-003) was purchased from Xiamen Immocell Biotechnology Co., Ltd. DMEM (cat. no. SNM-002C; Sunncell) supplemented with 10% fetal bovine serum (FBS; cat. no. SNS-002; Sunncell) was used for the culture of all cells at 37°C with 5% CO₂ as previously described (6). Subsequently, reverse transcription-quantitative polymerase chain reaction (RT-qPCR) was carried out to quantify the expression of circRNA_0044556 in these cells.

To determine the role circRNA_0044556 plays in TNBC, MDA-MB-231 and MDA-MB-231/ADM cells were subjected to the transfection. CircRNA_0044556-overexpression plasmid was synthesized and obtained from BersinBio, the empty plasmid as the control, and then these plasmids were transfected at 37°C and for 48 h into the MDA-MB-231 cells which were used to establish the groups of circRNA_0044556 and negative control (NC), respectively, while the cells without any transfection or treatment served as the control. Likewise, small interfering (si)RNA against circRNA_0044556 (50 pmol, si-circRNA_0044556, 5′-AGCCACAAAGAGTCTACATGTCT-3′) and its negative control (50 pmol, si-NC, 5′-UUCUCCGAACGUGUCACGU-3′), purchased from Shanghai GenePharma Co., Ltd., were independently transfected into MDA-MB-231/ADM cells, and the Control group was established as well. In the further investigation on the molecular interplay, 100 nM microRNA (miR)-145 mimic (M, miR10000437-1-5, 5′-GUCCAGUUUUCCCAGGAAUCCCU-3′; Guangzhou RiboBio Co., Ltd.) or 100 nM mimic control (MC, 5′-UUCUUCGAACGUGUCACGUTT-3′; Guangzhou RiboBio Co., Ltd.) was transfected into the parental MDA-MB-231 cells, and 100 nM miR-145 inhibitor (I, miR20000437-1-5, 5′-AGGGAUUCCUGGGAAAACUGGAC-3′; Guangzhou RiboBio Co., Ltd.) or 100 nM inhibitor control (IC, 5′-CAGUACUUUUGUGUAGUACAA-3′) was transfected into MDA-MB-231/ADM cells at 37°C for 48 h. All transfections on the cells were carried out using Lipofectamine^® 2000 (cat. no. 11668019; Invitrogen; Thermo Fisher Scientific, Inc.), and the transfected cells were assigned to the groups as follows: circRNA_0044556 group, si-circRNA_0044556 group, NC group, si-NC group, circRNA_0044556+M group, si-circRNA_0044556+I group, circRNA_0044556+MC group, si-circRNA_0044556+IC group, si-NC+I group, NC+M group, NC+MC group and si-NC+IC group. At 48 h after transfection, cells were collected for subsequent experiments.

Total RNA was separated from tissue samples and cells by TRIzol reagent (cat. no. 15596026; Invitrogen; Thermo Fisher Scientific, Inc.). Then, the extracted RNA was reversely transcribed into complementary DNA (cDNA) using a reverse transcription kit (cat. no. 18090010; Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocols. Subsequently, the expression levels of circRNA_0044556, miR-145 and NRAS proto-oncogene, GTPase (NRAS) were evaluated by RT-qPCR using qPCR SYBR Green Master Mix (cat. no. Q121-02; Vazyme Biotech Co., Ltd.) on an ABI 7500 system (Applied Biosystems; Thermo Fisher Scientific, Inc.). The thermocycling conditions were set as follows: Predenaturation at 95°C for 5 min, 40 cycles of 95°C for 10 sec and 60°C for 35 sec. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and U6 were used for the normalization of the expression levels of circRNA_0044556, NRAS, and miR-145. The expression levels of circRNA_0044556/NRAS (relative to GAPDH) and miR-145 (relative to U6) were calculated by the 2^−ΔΔCq method (16). Primer sequences used in this experiment were as follows: circRNA_0044556 forward, 5′-TGACGAGACCAAGAACTGCC-3′ and reverse, 5′-GCACCATCATTTCCACGAGC-3′; miR-145 forward, 5′-CAGTCTTGTCCAGTTTTCCCAG-3′ and reverse, 5′-TATGCTTGTTCTCGTCTCTGTGTC-3′; NRAS forward, 5′-ATGACTGAGTACAAACTGGTGGT-3′ and reverse, 5′-CATGTATTGGTCTCTCATGGCAC-3′; U6 forward, 5′-ATTGGAACGATACAGAGAAGATT-3′ and reverse, 5′-GGAACGCTTCACGAATTTG-3′; and GAPDH forward, 5′-GAAGGTGAAGGTCGGAGTC-3′ and reverse, 5′-GAAGATGGTGATGGGATTTC-3′.

To explore the role of circRNA_0044556 in the viability of TNBC cells to ADM, Cell Counting Kit-8 (CCK-8) assay was applied. Firstly, parental and ADM-resistant MDA-MB-231 cells (2×10⁴) were seeded into a 96-well plate and treated with different concentrations of ADM (0, 0.1, 0.2, 0.4, 0.8, 1.6, 3 and 6 µg/ml) at 37°C for 24 h, followed by the addition of 10 µl CCK-8 solution (cat. no. CK04; Dojindo Molecular Technologies, Inc.) to treat cells in each well at 37°C with 5% CO₂ for 2 h. A microplate reader (Infinite M200; Tecan Group, Ltd.) was used to measure the optical density (OD) value at a wavelength of 490 nm, and then the half maximal inhibitory concentration (IC₅₀) value was calculated.

Annexin V/FITC apoptosis detection kit (cat. no. AD10; Dojindo Molecular Technologies, Inc.) was used for assessing the apoptotic capacity of TNBC cells undergoing the indicated treatments (parental and ADM-resistant MDA-MB-231 cells transfected with circRNA_0044556, si-circRNA_0044556, miR-145 mimic, inhibitor and their negative control). According to the description in the manual, cell suspension (1×10⁶) was prepared using 1X Annexin V Binding Solution, and was then incubated with 5 µl Annexin V/FITC solution and 5 µl PI solution at room temperature for 15 min in the dark. Next, the stained samples were analyzed using an Accuri C6 flow cytometer with CFlow software (v. 1.32; BD Biosciences) to determine the apoptosis.

Transwell inserts (8 µm; product number 351184; Corning, Inc.) were used to conduct the Transwell assay. Following transfection, parental and ADM-resistant MDA-MB-231 cells were resuspended at a density of 4×10⁴ cells/well with 5 µl serum-free medium in the upper chamber of the insert, whereas 500 µl culture medium containing 10% FBS was added in the lower chamber at the same time. After 48 h of incubation at 37°C, the residual cells on the upper chamber were removed with a cotton swab, and the membrane was subjected to fixation with 4% paraformaldehyde (cat. no. E672002; Sangon Biotech Co., Ltd.) at 4°C for 30 min and staining was performed using 0.1% crystal violet for 15 min at room temperature (cat. no. G1064; Beijing Solarbio Science and Technology Co., Ltd.). Finally, a light microscope (×250, magnification; CX23; Olympus Corporation) was used to observe the migration of cells on the lower side of the membrane.

The targeting relationship among circRNA_0044556, miR-145, and NRAS in TNBC cells was predicted by circInteractome (https://circinteractome.irp.nia.nih.gov/) and StarBase v2.0 (http://starbase.sysu.edu.cn/index.php), respectively.

For verifying the targeting relationship through dual luciferase reporter assay, vectors (cat. no. E1330; Promega Corporation) were used to construct wild-type (wt) reporter plasmid of circRNA_0044556 (circRNA_0044556-wt; 5′-GGUGCCAAGGGUCUGACUGGAAG-3′), mutant (mut) plasmid of circRNA_0044556 (circRNA_0044556-mut; 5′-GGUGCCAAGGGUCUGCCUCGUAG-3′), NRAS-wt reporter plasmid (5′-CAAACCCUUUACCAUGACUGGAA-3′) and NRAS-mut reporter plasmid (5′-CAAACCCUUUACCCUGUCUCGUA-3′). Next, MDA-MB-231 cells were co-transfected with miR-145 mimic (5′-GUCCAGUUUUCCCAGGAAUCCCU-3′) or its miR-NC (5′-UUCUUCGAACGUGUCACGUTT-3′) and the reporter plasmids (circRNA_0044556-wt, circRNA_0044556-mut, NRAS-wt or NRAS-mut) using Lipofectamine^® 2000 reagent (Invitrogen; Thermo Fisher Scientific, Inc.). After 48 h, the relative luciferase activity was determined by a Dual Luciferase Reporter Gene Assay Kit (cat. no. 11402ES60; Shanghai Yeasen Biotechnology Co., Ltd.). The luciferase activity was normalized to the Renilla luciferase activity.

RIP kit (cat. no. KT102-01), obtained from Guangzhou Saicheng Biotechnology Co., Ltd. was used to determine the connectivity between circRNA_0044556 and miR-145. Briefly, the transfected cells (4×10⁷ cells) were collected with phosphate-buffered saline (cat. no. C0221A; Beyotime Institute of Biotechnology) and lysed with cell lysis buffer. G/A beads (100 µl) were prepared and divided into the groups argonaute2 (Ago2) and immunoglobulin G (IgG) (4). Then, the antibodies against Ago2 (product code ab32381, 1:1,000) and IgG (product code ab133470, 1:1,000; both from Abcam) were added into the lysates of cells in the indicated groups at 4°C. After 6 h of incubation, the antibody-tagged beads were incubated with the lysates at 4°C overnight. Subsequently, the beads were washed with RIP buffer for five times by centrifugation at 1,500 × g for 2 min at 4°C. Proteinase K was added and incubated for 45 min at 65°C to remove the protein prior to RNA isolation, and RNA was isolated using TRIzol^® reagent (Thermo Fisher Scientific, Inc.) as aforementioned. RNA was purified and precipitation according to the manufacturer's protocols. The immunoprecipitated RNA was subjected to RT-qPCR to determine the enrichment of circRNA_0044556 in MDA-MB-231 cells with miR-145 overexpression.

All data were analyzed using GraphPad Prism 8.0 software (GraphPad Software, Inc.), and measurement data were expressed as the mean ± standard deviation. The differences between two independent samples or two paired samples were analyzed by independent samples t-test or paired samples t-test. One-way analysis of variance with Tukey's post hoc test was applied for single-factor differences between multiple groups, while two-way analysis of variance for the two-factor differences between multiple groups. The correlation among circRNA_0044556, miR-145 and NRAS in TNBC was analyzed with Pearson's correlation test. P<0.05 was considered to indicate a statistically significant difference.

RT-qPCR was first performed to quantify the expression of circRNA_0044556 in tissues (TNBC tissues with or without ADM resistance and the adjacent tissue) and cells (mammary epithelial cell line MCF-10A, TNBC cell lines and MDA-MB-231/ADM cells). As revealed in Fig. 1A-D, circRNA_0044556 was highly expressed in TNBC tissues, ADM-resistant TNBC tissues and TNBC cells (P<0.001). As the most significantly high expression of circRNA_0044556 was detected in MDA-MB-231 cells among the four other TNBC cell lines, MDA-MB-231 cells were selected for subsequent experiments.

Subsequently, MDA-MB-231 cells were transfected with circRNA_0044556 overexpression plasmid, and MDA-MB-231/ADM cells were transfected with si-circRNA_0044556. RT-qPCR was used to estimate the transfection efficiency and it was determined that the expression of circRNA_0044556 was significantly promoted by the overexpression plasmid of circRNA_0044556, yet it was suppressed by si-circRNA_0044556 (Fig. 2A and B; P<0.001). It was indicated in the results of the CCK-8 assay that after the treatment of ADM in gradient concentrations, overexpressed circRNA_0044556 enhanced the viability of parental MDA-MB-231 cells and increased the IC₅₀ value (Fig. 2C and D; P<0.05), whereas the silencing of circRNA_0044556 significantly reduced the viability of MDA-MB-231/ADM cells and decreased the IC₅₀ value (Fig. 2E and F; P<0.01).

It was demonstrated in the apoptotic assay (Fig. 3A and B) that the overexpression of circRNA_0044556 decreased the percentage of apoptotic parental TNBC cells (P<0.001), whereas the silencing of circRNA_0044556 promoted apoptosis of the ADM-resistant cells (P<0.001). As demonstrated in Fig. 3C and D, a distinctly increased number of migratory TNBC cells were observed after the overexpression of circRNA_0044556 compared with those in the NC group (P<0.001), whereas the migratory capacity of MDA-MB-231/ADM cells was significantly suppressed after the inhibition of circRNA_0044556 as compared with the si-NC group (P<0.001).

The binding site between the 3′-untranslated region (UTR) of circRNA_0044556 and the seed region of miR-145 (Fig. 4A) was predicted by bioinformatics analysis. Subsequently, the dual-luciferase reporter assay in Fig. 4B indicated that miR-145 mimic significantly suppressed the luciferase activity of circRNA_0044556-wt (P<0.001). The binding between circRNA_0044556 and miR-145 in TNBC cells was further determined via RIP assay, and it was revealed that circRNA_0044556 was enriched in Ago2-tagged beads compared with the control group of IgG (Fig. 4C; P<0.001). Notably, detection with RT-qPCR revealed that the expression of miR-145 was downregulated in TNBC tissues, and ADM-resistant TNBC tissues in particular (Fig. 4D and E; P<0.001). Additionally, according to Pearson's correlation analysis (Fig. 4F), miR-145 was negatively correlated with circRNA_0044556 in TNBC patients with ADM resistance (r=−0.6168, P=0.0143). Correspondingly, a lower expression of miR-145 in TNBC cell lines and MDA-MB-231/ADM cells was also identified (Fig. 4G and H; P<0.001).

In order to understand the interplay between circRNA_0044556 and miR-145 in TNBC, miR-145 mimic and circRNA_0044556-overexpressing plasmid or miR-145 inhibitor and si-circRNA_0044556 were co-transfected in parental and ADM-resistant MDA-MB-231 cells, and the transfection efficiency was evaluated by RT-qPCR. As revealed in Fig. 5A, the overexpression of circRNA_0044556 downregulated miR-145 expression and miR-145 mimic induced upregulation of miR-145 compared with the NC+MC group, the trends of which were both overturned in the circRNA_0044556+M group (P<0.001). Conversely, si-circRNA_0044556 upregulated miR-145 expression and miR-145 inhibitor led to decreased miR-145 expression compared with the si-NC+IC group, and these trends were both reversed in the si-circRNA_0044556+I group (Fig. 5B; P<0.01). Subsequently, the results of rescue experiments indicated that the high expression of miR-145 induced the apoptosis, decreased the migration of parental MDA-MB-231 cells, and abrogated the effects of overexpressed circRNA_0044556 on the apoptosis and migration of parental cells (Fig. 6A and C; P<0.05). In ADM-resistant MDA-MB-231 cells, the knockdown of miR-145 dampened apoptosis, facilitated migration, and reversed the effects of circRNA_0044556 silencing (Fig. 6B and D; P<0.05).

Following analysis by StarBase, NRAS was predicted to contain the binding site for miR-145 (Fig. 7A), the existence of which was then evidenced in the dual-luciferase reporter assay, where a decrease of luciferase activity in TNBC cells co-transfected with miR-145 mimic and NRAS-wt was identified (Fig. 7B; P<0.001). In addition, NRAS was revealed to be overexpressed in TNBC tissues and cells with or without ADM treatment, according to the results of RT-qPCR (Fig. 7C-F; P<0.001). Furthermore, NRAS was negatively correlated with miR-145 (Fig. 7G; r=−0.6552, P=0.008), and positively correlated with circRNA_0044556 in TNBC cells (Fig. 7H; r=0.6391, P=0.0103). circRNA_0044556 overexpression promoted the expression of NRAS in MDA-MB-231 cells, and silencing of circRNA_0044556 inhibited the expression of NRAS in ADM-resistant MDA-MB-231 cells (Fig. 7I and J; P<0.001).