Cell culture and clinical specimens

The human breast cancer cell lines (BT-549, T-47D, MDA-MB-468, MDA-MB-231, MCF-7), the noncancerous breast epithelial cell line HBL-100, and the human embryonic kidney (HEK) 293T cells were purchased from the American Type Culture Collection (ATCC). All cell lines were cultured following the instructions of ATCC. The DOX resistant MCF-7 was purchased from shanghai Meixuan Biotechnology Co., LTD. The DOX resistant MDA-MB-231 was constructed by our laboratory using conventional continuous exposure to DOX with dose-stepwise increment. DOX resistant MCF-7 (MCF-7/DOX) were cultured in RPMI-1640 with 10% FBS containing 430 nM DOX. DOX resistant MDA-MB-231 (MDA-MB-231/DOX) were cultured in DMEM with 10% FBS containing 350 nM DOX.

Blood samples were obtained from the Central Hospital of Wuhan. Eighty breast cancer patients and 80 healthy controls were included in the study. Twenty paired blood samples were collected, including preoperative and postoperative blood from breast cancer patients. The newly diagnosed breast cancer patients had received no chemotherapy or radiotherapy before sample collection. All healthy subjects did not have any diseases or injuries. Peripheral blood samples from participants were collected in tubes containing EDTA and were processed within 4 h by centrifugation at 1000 g for 15 min at 4 °C. The supernatant was then gently transferred to a fresh RNase-free 1.5-ml tube and cryopreserved at − 80 °C immediately. The study was approved by the Ethical and Scientific Committees of the Central Hospital of Wuhan, and all participants provided informed consent during enrollment.

RNA isolation and qRT-PCR

RNA was extracted by TRIzol reagent (Invitrogen, Carlsbad, CA, USA). BCAR4 was measured by SYBR Green (Applied Biosystems, Carlsbad, CA, USA) in StepOne™ Real-Time PCR System (Applied Biosystems). MiR-644a was measured with a bulge-loop miRNA real-time PCR kit (RiboBio, Guangzhou, China). GAPDH was used as the endogenous control for BCAR4 and U6 for miR-644a. The relative expression of BCAR4 and miR-644a was calculated by 2–△△Ct in breast cancer cells and by 2–△Ct for that in plasma. The primers for BCAR4 were as follows: sense, ATA CAA TGG CGT AAT CAT AGC, and antisense, AGA CAT TCA GAG CAA GAC A. ABCB1: sense, CAG GCT TGC TGT AAT TAC CCA, and antisense, TCA AAG AAA CAA CGG TTC GG. GAPDH: sense, GGG AGC CAA AAG GGT CAT, and antisense, GAG TCC TTC CAC GAT ACC AA.

Cell transfection

MiR-644a mimic, or miR-644a inhibitor, small interfering RNAs (siRNAs) targeting BCAR4, and their respective negative controls were all obtained from RiboBio (Guangzhou, China). The sequence of siRNAs targeting BCAR4 were as follows: si-1#: CCT GTG ATG TGT TGA TAA A, si-2#: CAC AAT TGA TGT TCT CTA A, and si-3#: GAG TTC TGC AAT CCA CAA T. The full-length of BCAR4 was cloned into BamH1 and Xho1 sites in a pcDNA-3.1 vector ((Invitrogen), which was confirmed by sequencing. Transfection was performed using ExFect® Transfection reagent (Vazyme, Nanjing, China) according to the manufacturer’s guidelines.

Colony formation assay

MCF-7 or MDA-MB-231 cells were seeded into 12-well plates with 500 cells/well and incubated for ten days. Plates were then gently washed with PBS and stained with 0.1% crystal violet (Beyotime, Nantong, China). A colony containing more than 50 cells was counted as one colony.

Wound healing assay

For cell mobility assay, MCF-7 or MDA-MB-231 cells were seeded into 6-well plates (4 × 105 cells/well). An artificial wound was made by scratching the confluent cell monolayer with a sterile 200 μl pipette tip 24 h after transfection. Then, cells were incubated in a medium containing 1% FBS. After scratching for 0 h and 48 h, wound healing images were photographed at magnification 100 × using an IX81 microscope (Olympus, Tokyo, Japan). The migration rate was determined by ImageJ software.

Transwell assay

To further investigate breast cancer cell migration and invasion, transwell chambers (Corning, NY, USA) were used. MCF-7 or MDA-MB-231 cells were detached and suspended in 200 μL serum-free medium 24 h after transfection. A total of 2 × 104 cells were added into the upper chamber for migration assay or a total of 3 × 104 cells were added into the upper chamber coated with Matrigel (BD Biosciences, CA, USA) for invasion assay. The bottom chambers were filled with the 600 μL complete culture medium. After being cultured for 48 h, cells that migrated or invaded the lower filter surfaces were fixed and then stained with 0.1% crystal violet (Beyotime). Images (magnification, × 100) were taken from each membrane, and the numbers of migrated or invaded cells were counted using an IX81 microscope (Olympus).

Drug sensitivity assay and CCK8 assay

The DOX (Selleck Chemicals) sensitivity of cells was detected using the CCK8 kit (Beyotime) by measuring the IC50 value (the concentration of DOX resulted in 50% decline of the absorbance compared with control). MCF-7 or MDA-MB-231 cells were seeded into 96-well plates with 5000 cells/well and incubated with different concentrations of DOX (5 nM, 10 nM, 30 nM, 100 nM, 200 nM, 500 nM, 1000 nM and 2000 nM). After treatment for 48 h, CCK8 solution was then added to wells and incubated for another 4 h. The absorbance was recorded at 450 nm by the EnSpire multimode reader (PerkinElmer, Waltham, MA).

Western blot assay

Western blot analysis was performed according to the description previously [20]. The following primary antibodies were used: rabbit polyclonal anti-ABCB1 (1: 500, Proteintech, Rosemont, USA), rabbit monoclonal anti-CCR7 (1:10,000, Abcam, Cambridge, MA), mouse monoclonal anti-ERK1/2 (1:2000, Proteintech), mouse monoclonal anti-P38 MAPK (1: 2000, Proteintech). rabbit monoclonal anti-phospho-p44/42 MAPK(Erk1/2)(Thr202/Tyr204, 1: 1000, Cell Signaling Technology, Danvers, MA, USA), rabbit monoclonal anti-phospho-p38 MAPK (Thr180/Tyr182, 1: 1000, Cell Signaling Technology), mouse monoclonal anti-E-cadherin (1:2000, Proteintech), mouse monoclonal anti-Vimentin(1:5000, Proteintech), rabbit polyclonal anti-VEGF(1: 1000, Proteintech). Rabbit polyclonal anti-GAPDH (1:5000, Proteintech) was used as an internal control.

Luciferase reporter assay

For the BCAR4 luciferase reporter assay, pRL-TK-BCAR4 or pRL-TK-BCAR4 mutant vectors and pGL3 control (Promega, Madison, WI) were transfected into HEK-293T cells along with miR-644a mimic or miR-644a inhibitor using ExFect® Transfection reagent (Vazyme) according to the instructions of the manufacturer. For miR-644a target gene CCR7 luciferase reporter assay, pRL-TK-CCR7-3΄UTR or pRL-TK- CCR7-3΄UTR mutant vectors and pGL3 control (Promega) were transfected into HEK-293T cells along with miR-644a mimic or miR-644a inhibitor using ExFect® Transfection reagent (Vazyme). The luciferase activity was measured 48 h after transfection using the Dual-Glo luciferase reporter assay system (Promega) according to the guidelines of the manufacturer. Luminescence values were recorded by a Victor X2 multilabel reader (PerkinElmer).

In vivo assay

Female athymic four-week-old BALB/c nude mice were purchased from HFK Bio-Technology Co. Ltd. (Beijing, China). A total of 5 × 106 MDA-MB-231 cells, resuspended in Opti-MEM, were injected into the mammary fat pad. Tumor volume was analyzed as V = D × d2 × 0.5 (D, the longer diameter; d, the shorter diameter). When the tumor reached about 60 mm3, the mice were divided into a si-control group (n = 5) and a si-BCAR4 group (n = 5) randomly. The in vivo siRNA for BCAR4 knockdown was synthesized by RiboBio (Guangzhou, China). The sequence of the in vivo BCAR4 knockdown was as follows: si-BCAR4: CAC AAT TGA TGT TCT CTA A. In vivo si-BCAR4 or the control (2.5 nmol / 20 g body weight) was directly administered into the implanted tumor two times per week for three weeks. One month after the first injection of si-BCAR4 or the control, all mice were anesthetized with 1% pentobarbital sodium. The tumors, lungs, and livers were obtained and then fixed in formalin for hematoxylin and eosin (HE) staining or immunohistochemistry. The rabbit polyclonal anti-CCR7 (1:200, Proteintech) was used to determine CCR7 expression. The mouse monoclonal anti-E-cadherin (1:1000, Proteintech), mouse monoclonal anti-Vimentin (1: 1000, Proteintech), rabbit polyclonal anti-VEGF (1:200, Proteintech) was used to detect E-cadherin, Vimentin and VEGF expression respectively.

Statistical analysis

All statistical analyses were performed using SPSS 22.0 software. The one-way analysis of variance (ANOVA), χ2 test or the student’s t-test was carried out for comparison between groups. All experiments were independently repeated at least three times. P value < 0.05 represents statistically significance. *P < 0.05, ** P < 0.01, ***P < 0.001.