MCF-7 cells were purchased from Bioresource Collection and Research Center (Hsinchu, Taiwan), and HEK-293T cells were provided by Dr. Shau-Ping Lin (Institute of Biotechnology, National Taiwan University, Taiwan). All cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) (Gibco, Thermo Fisher, CA, USA) containing 1% penicillin-streptomycin (Gibco) and 10% foetal bovine serum (FBS) (HyClone, Logan, UT, USA). The culture plates were maintained at 37°C in a humidified atmosphere of 5% CO₂ and were routinely tested for Mycoplasma sp. using Mycoplasma polymerase chain reaction (PCR) Detection Kit (ABM Inc., Vancouver, BC, Canada).

MCF-7 cells were seeded at 5 ×10³ cell into 96-well plates 1 day before lidocaine or bupivacaine (Sigma, St. Louis, MO, USA) treatment at the indicated concentrations. LAs and the controls were diluted in 200 μL DMEM for each well. For the thiazolyl blue tetrazolium bromide (MTT) colorimetric assay, 5 mg/mL of MTT (Sigma) were added to each well and incubated at 37°C for 4 h. The supernatants were removed and replaced with dimethyl sulfoxide (DMSO) (Sigma). The absorbance was measured using a microtiter plate reader (BioTek, Winooski, VT, USA) at 570 nm.

MCF-7 cells were seeded at a density of 2.5 ×10⁴ cells/well in a medium containing 10% FBS and incubated overnight. The culture inserts were removed and an image of the gap area at 0 h was captured. The cells were further incubated at 37°C in a 5% CO₂ incubator, and the images were captured at 12, 24, and 36 h, respectively. The cell migration ability in the gap area was quantified using ImageJ v1.8.0 (National Institutes of Health, USA).

Migration assays were performed using SPLInsert^TM Hanging plate (SPL Life Sciences, Pocheon, South Korea). The upper chamber of the transwell unit was seeded at 4 ×10⁴ cells/well in 0.2 mL serum-free DMEM. The lower chambers were loaded with 0.6 mL DMEM containing 10% FBS as chemo-attractant. The MCF-7 cells were incubated at 37°C for 36 h. A methanol: acetic acid (3:1) mixture was added to the lower chamber for fixing the cells by incubating at 25°C for 10 min. The non-migrating cells on the inner transwell membrane were removed carefully, and the membranes were stained with 1 mL of 0.5% crystal violet. The stained cells were solubilized with 10% acetic acid, imaged, and analysed using ImageJ v1.8.0.

The proteins were extracted from the cells lysed with radioimmunoprecipitation assay buffer (RIPA) Lysis Buffer (EMD Millipore, Billerica, MA, USA) and the concentration was determined using Coomassie Protein Assay Reagent (Thermo Fisher, Waltham, MA, USA). The extracted protein samples (20 μg) were separated using 10% sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene fluoride (PVDF) membranes (GE healthcare, Chicago, IL, USA). The membranes were incubated with primary rabbit polyclonal antibodies against c-MYB (Proteintech, Rosemont, IL, USA) and β-actin (GeneTex, Irvine, CA, USA), at 4°C overnight; the samples were hybridised with horseradish peroxidase (HRP) conjugated anti-rabbit IgG (GeneTex) at 25°C for 1 h. The blotted proteins were detected using an enhanced chemiluminescence (ECL) system (Millipore, Billerica, MA, USA) equipped with a BioSpectrum Imaging System (UVP, Upland, CA, USA). The images were analysed using ImageJ v1.8.0.

The RNA was extracted using TRIZOL (Invitrogen, Carlsbad, CA, USA) and reverse-transcribed using High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Carlsbad, CA, USA). The RNA quality was detected by a spectrophotometer (NanoDrop; Thermo Scientific) with the A260/A280 ratio between 1.8~2.0. Samples were kept in RNase free water buffer at neutral pH. All RNA samples were stored at −80°C. One μg of total RNA was reverse-transcribed by High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Carlsbad, CA, USA). SuperScript IV Reverse Transcriptase (Invitrogen) was used for reverse transcription of the miRNAs in accordance with the manufacturer's instructions. Two point five percent of each reaction was used as template for quantitative PCR with OmicsGreen qPCR MasterMix (OmicsBio, New Taipei City, Taiwan) and the reactions were performed on StepOnePlus Real-Time PCR System (Thermo Fisher, Waltham, MA, USA). The primers were checked for only generating a single melting curve peak and detailed information are provided in Supplementary Table 1. The relative expression was normalised to that of 18S rRNA or U6 snRNA, using the 2^−ΔΔCt method.

The total RNA was extracted using TRIZOL reagent (Invitrogen). The small RNA library was constructed from total RNA (2 μg) using the TruSeq Small RNA Library Prep Kit (Illumina Inc., San Diego, CA, USA), according to the manufacturer's instructions. A total of 75 single-read nucleotides were obtained from each miRNA library using NextSeq500 (Illumina Inc., CIC bioGUNE, Bilbao, Spain). The sequencing data were submitted to the Gene Expression Omnibus (GEO) (accession number: GSE171282).

The raw Illumina FASTQ sequencing reads were mapped to the miRBase mature miRNA reference (http://www.mirbase.org/) using Partek Flow™ v5.0 (Partek Inc., St. Louis, MO, USA) with Bowtie alignment algorithm. The Bowtie alignment method was generally thought to be optimal for sequences less than 50 bp. The adapter sequences were first removed from each FASTQ reads, and the remaining bases were trimmed from the 3′ end with a minimum Phred quality score of 20. A minimum read-length philtre of 15-bases in length was used in this work. A minimum seed length of 10 that was consistent with the standard setting provided by the Partek Flow pipeline was chosen for the Bowtie aligner. The remaining reads were aligned to the human genome reference (RefSeq Hg19). Raw miRNA expression reads were normalised by scale normalisation and then processed by log₂ transformation. Then, miRNAs with extremely low expression (log₂ <5 in all samples) were removed from further analysis. Then, differentially expressed miRNA profiles between groups were identified by log₂ fold-change value ≧2X or < -1X and P-value <0.05 using a two-sided Student's t-test. Principal component analysis using the expression values of total miRNAs was used to visualise the similarity of different groups.

MCF-7 cells were cultured in an antibiotic-free medium to 70–80% confluence in 6-well plates (2.5 ×10⁵ cells/well) and transfected with the 0.025 nmol of miR-187-5p mimic (GenePharma, Shanghai, China) or negative control (GenePharma), using Lipofectamine 2000 as the transfection reagent (Invitrogen), following the manufacturer's instructions.

The MCF-7 cells were transfected with 0.5 nmol of pcDNA-DANCR or the empty vector (Dharmacon, New Taipei city, Taiwan) using jetPRIME (Polyplus-transfection, New York, NY, USA) reagent, following the manufacturer's instructions. After 4 h, the medium was replaced with fresh medium containing serum. The cellular RNA expression after 48 h was determined using qRT-PCR.

To determine promoter activity by luciferase assay, the luciferase expression plasmid pMIR-REPORT-c-MYB 3′UTR and pMIR-REPORT-DANCR was purchased from the BioMed Resource Core of the 1st Core Facility Lab, National Taiwan University College of Medicine (Taipei, Taiwan). Briefly, c-MYB 3′UTR region encompassing 2,493–3,684 bp relative to the transcription start site was amplified from the human genomic DNA by PCR. The c-MYB 3′UTR was subcloned into pMIR-REPORT vector, and the final construct was called pMIR-REPORT-c-MYB 3′UTR. DANCR encompassing 1–915 bp relative to the transcription start site was amplified from the human genomic DNA by PCR and the final construct was called pMIR-REPORT-DANCR.

miR-187-5p binding sites of c-MYB 3′UTR were located at 2,710–2,734 and 3,634–3,656 bp. Besides, miR-187-5p binding sites of DANCR were located at 134–157 and 380–408 bp. To determine binding activity of miR-187-5p by luciferase assay, the luciferase expression plasmid pMIR-REPORT-c-MYB 3′UTR S1, pMIR-REPORT-c-MYB 3′UTR S2, pMIR-REPORT-c-MYB 3′UTR S12, pMIR-REPORT-DANCR S1, pMIR-REPORT-DANCR S2, and pMIR-REPORT-DANCR S12 was purchased from the BioMed Resource Core of the 1st Core Facility Lab, NTU-CM (Taipei, Taiwan). In addition, the mutated sequences were validated by sequencing.

Luciferase assays were performed with HEK-293T cells using the Dual-Glo® luciferase reporter assay system (Promega, Fitchburg, WI, USA). Cells (5 ×10⁴/well) were transfected with 100 ng reporter constructs containing the wild-type or mutant-type of MYB, the wild-type or mutant-type of DANCR and 2 ng Ranilla luciferase plasmid (BioMed Resource Core of the 1st Core Facility Lab, National Taiwan University College of Medicine, Taiwan), using the jetPRIME (Polyplus-transfection) reagent. The cells were additionally transfected with 0.025 nmol of miR-187-5p (GenePharma). After 24 h, the cells were lysed and the Renilla luciferase signals were used for normalisation, according to manufacturer's protocol.

The experimental data are presented as the mean ± standard deviation (SD) of at least three independent experiments. The genes with statistically significant expression patterns in the different samples were determined by a two-tailed Student's t-test using GraphPad Prism v5 (GraphPad Software, Inc., La Jolla, CA, USA). p-values < 0.05 were considered statistically significant.

To determine the effects of lidocaine and bupivacaine on MCF-7 cells, the cell viability following treatment with different concentrations of lidocaine or bupivacaine for 6 h was determined. Lidocaine and bupivacaine inhibited the growth of MCF-7 cells in a dose-dependent manner (Figure 1A). The median lethal dose (LD₅₀) of lidocaine (4 mM) and bupivacaine (1 mM) were used as the concentrations for the following experiments, respectively. Wound healing and transwell invasion assays were performed using MCF-7 cells treated with lidocaine or bupivacaine for 12, 24, and 36 h. Both the LAs significantly reduced the migration and invasive ability of MCF-7 cells (Figures 1B,C), suggesting suppression of malignancy in the MCF-7 cells.

To identify the miRNAs that mediated the effects of the LAs on MCF-7 cells, next-generation sequencing was used for identifying the differentially expressed miRNAs, following lidocaine or bupivacaine treatment. Principal component analysis (PCA) was performed for the samples (lidocaine, bupivacaine, and control groups, nine dots in total) using the data for 2,522 differentially expressed miRNAs in the three groups (Figure 2A). The miRNAs in the same group showed similar expression profiles, while the expression profiles in the lidocaine- and bupivacaine-treated groups were distinct from those of the control.

The miRNAs that were differentially expressed in the lidocaine- and bupivacaine-treated groups were filtered using the criteria: fold change ≥ 2.0X and p-value ≤ 0.05. The expression of 1,228 and 133 miRNAs was significantly upregulated in the lidocaine (Figure 2B) and bupivacaine groups (Figure 2C), respectively (red dots). The expression of 25 and 36 miRNAs was significantly downregulated (fold change < -1X and p-value ≤ 0.05) in the lidocaine (Figure 2B) and bupivacaine groups (Figure 2C), respectively (green dots). The genes that were not differentially expressed are depicted in grey. Among the upregulated miRNAs, 131 were common to both the lidocaine and bupivacaine groups (Figure 2D). The number of downregulated miRNAs common to both the groups was low; therefore, only the upregulated miRNAs were selected for subsequent studies. The expression of the selected miRNAs is depicted using a heatmap (Figure 2E).

The data from next-generation sequencing was validated by selecting the miRNAs whose expression was significantly higher and for which experimentally validated data were available. The expression of only miR-187-5p and miR-223-5p was significantly (p < 0.05) higher than that of the control. The expression of the other miRNAs examined using qRT-PCR was not statistically different between the lidocaine and bupivacaine groups (Figure 2F).

To identify the genes regulated by miR-187-5p and ascertain their functional relevance, we used the TargetScan tool (http://www.targetscan.org/vert_72/) for identifying the downstream target genes. The MYB transcription factor, a proto-oncogene, was a predicted target gene of miR-187-5p. In vitro and in vivo studies indicated that MYB plays an important role in different cancers. MYB overexpression and dysregulation is observed in almost all breast tumours, and is associated with poor prognosis (20, 21). MYB also regulates cell proliferation, differentiation, and angiogenesis (22), and therefore, was selected for further studies.

qRT-PCR and western blotting were used to investigate the mechanism underlying the influence of miR-187-5p on cellular function. The mRNA and protein levels of MYB were significantly downregulated in the LA-treated groups (Figures 3A,B). The levels of MYB mRNA were significantly reduced when miR-187-5p was overexpressed (Figure 3C). Using bioinformatics analysis, we validated MYB as a target gene of miR-187-5p; two binding sites (2,710–2,732 nt and 3,634–3,656 nt) were predicted in the 3′-UTR of MYB, which were complementary to the seed sequence of miR-187-5p (Figure 3D). The direct interactions between miR-187-5p and MYB were examined using luciferase reporter plasmids containing the wild-type MYB (MYB-wt) and binding site mutants (MYB-site1-mut, MYB-site2-mut, and MYB-site1+2-mut). The luciferase activity of MYB-wt decreased significantly, with the transfection of miR-187-5p mimics; however, they had no influence on MYB-site1, MYB-site2, and MYB-site1+2 (Figure 3E), indicating that MYB is a direct target gene of miR-187-5p. The LAs, lidocaine and bupivacaine, upregulated the expression of miR-187-5p, directly downregulating the expression of MYB.

LncRNAs act as ceRNAs and influence mRNA levels by sequestering miRNAs that target both lncRNA and mRNAs. To identify the ceRNAs of miR-187-5p, we used DIANA-LncBase v2 (23) for predicting the lncRNAs that bind to miR-187-5p. Five possible lncRNAs were predicted and validated using qRT-PCR (Figure 4A). The expression of DANCR was downregulated by the LAs, while that of the HIF1A-AS2 and SNHG1 lncRNAs was upregulated; however, these lncRNAs did not function as ceRNAs of miR-187-5p. In addition, the expression of DANCR significantly decreased when the expression of miR-187-5p increased (Figure 4B).

Luciferase reporter assays were performed for validating the association between DANCR and miR-187-5p. Prediction revealed two potential binding sites of miR-187-5p at 134–157 nt and 380–408 nt of DANCR, which were individually positioned upstream of the luciferase reporter genes (Figure 4C). The binding site mutants (sites 1-mut, 2-mut, and 1+2-mut) were constructed and transfected into HEK-293T cells. The luciferase activity of DANCR-wt was significantly reduced by the miR-187-5p mimics; however, they had no effect on sites 1-mut, 2-mut, and 1+2-mut (Figure 4D), indicating that the DANCR lncRNA sponges miR-187-5p by direct binding.

The overexpression of DANCR suppressed miR-187-5p expression, while the downregulation of DANCR increased miR-187-5p expression (Figure 4E). These results demonstrated a reciprocal inhibitory relationship between DANCR and miR-187-5p, and that miR-187-5p could mediate the inhibitory effect of the LAs on MCF-7 cell viability and migration.