Patients blood sample and RNA-Sequencing analysis

The blood samples of HCC patients and healthy controls (HC) were obtained for RNA sequencing were all from the First Affiliated Hospital of Guangxi Medical University. The methods of blood collection and RNA-sequencing refer to the previous published articles [20]. Briefly, RNA of Plasma exosomes from HCC patients and healthy controls were extracted for RNA-sequencing. The raw read sequences were filtered to remove adapter sequences and low-quality reads (more than 20% bases with quality less than 20, or reads with more than 10% undefined nucleotides) by using Skewer and the FastX-Toolkit software. Reads were mapped with human reference rRNA sequences (GenBank and GENCODE v26); reads that did not map with rRNA (clean reads) were subjected to subsequent analysis. Clean reads were aligned to human reference genome assembly (GRCh38) to define mRNA and lncRNA profiles by HISAT2 aligner with the default parameters. Clean reads were also annotated to GENCODE (v26) and the transcript expression level was quantified by the Expectation-Maximization (RSEM) package. Expression levels were calculated by reads per kilobase of transcript per million mapped reads. The sequencing results of HCC patients and healthy controls were compared. When the |fold change|≥2.0 and P < 0.05 of the transcript, it could be considered that the expression of the transcript was different between HCC patients and healthy controls.

Cell culture

Normal hepatocyte (HL7702) and HCC cells (Hep3B and Huh7) were obtained from Chinese Academy of Sciences (Shanghai, China). Dulbecco’s modified Eagle’s medium (Invitrogen, USA) was used to culture HL7702 and Huh7 cells, and Minimum Essential medium (Invitrogen, USA) was used to culture Hep3B. The composition of the cell culture medium was as follows: 98% culture medium, 10% fetal bovine serum (Thermo Fisher Scientific, USA), and 1% penicillin-streptomycin (Thermo Fisher Scientific, USA). All cells were placed in a 37℃ humidified incubator with 95% air and 5% CO2.

Establishment of ethanol treated cell model

HL7702 (2.5 × 105 per well), Huh7 (2.5 × 105 per well) and Hep3B (3.0 × 105 per well) cells were plated in 6-well plates with 2 ml complete medium for 12 h, then remove the culture medium and added 2 ml of 200 mM ethanol (24 µl anhydrous ethanol to 2 ml complete medium) into each well and cultured for 48 h to establish ethanol treated cell model. The control group was added with 2 ml complete medium.

Cell transfection

Three small interfering RNAs of lnc171, inhibitor and mimics of mir-873-5p were designed and construction by Sangong Biotech (Table S1) (Shanghai, China). siRNA, mir-873-5p inhibitor and mimics were transfected by Lipofectamine 3000 (Thermo Fisher Scientific, USA). Before transfection, cells were plated in 6-well plates (HL7702 and Huh7 cells: 2.5 × 105 per well; Hep3B cells: 3.0 × 105 per well) with 2 ml complete medium for 12 h. Three si171, mir-873-5p mimics, mir-873-5p inhibitor, and negative control (NC) were transfected, respectively, into the appropriate cell group, then added with 2 ml Opti-MEM medium without SBF per well following the manufacturer’s instruction.

Transwell assay

Transwell chambers (pore size 8 μm; Coning, USA) were used to conduct transwell assay with or without matrigel (BD Biosciences, USA). After transfection or ethanol treatment, cells (HL7702 and HepG2 cells: 5 × 104 per well; Hep3B cells: 7 × 104 per well) were cultured in the dishes with 200 ml serum-free medium. The bottom chamber was added with 600 µl of completed cultured medium. After incubation for 48 h, cells that migrated or invaded to the bottom of the filter membrane were fixed with methanol, stained with 0.5% crystal violet solution then photographed. Three fields were randomly selected and the relative cell number was calculated.

Quantitative real-time PCR analysis

The process of total RNAs extraction, RNA reverse transcription, and qRT-PCR was according to our previously study [20]. The 2−ΔΔCt method was used to calculate the relative folding changes of gene expression. U6 andβ-actin were used as an endogenous control. Primer sequences were listed in Table S2.

Western blot assay

The process of obtain the cellular protein, protein quantitation, protein electrophoresis, and protein bands detection were according to our previously study [20]. Briefly, RIPA Buffer (Beyotime Biotechnology, Shanghai, China) was used to lyse cells. The concentrations of protein were quantified using BCA kit (Beyotime Biotechnology, Shanghai, China). Then separated the protein by sodium dodecyl sulfate (SDS)-PAGE on 10% gels and transferred to poly-vinylidene difluoride (PVDF) membranes (the membranes were cut to appropriate size prior to hybridisation with antibodies during blotting) (Sigma-Alcdrih, St. Louis, MO, USA). After overnight incubation with first antibodies, then incubated for 2 h in the presence of secondary antibody and washed 3 times with TBST for 5 min. Quantity-one software (Bio-Rad Laboratories, USA) using the ECL-chemiluminescent kit. The signal of protein bands was quantified by ImageJ software for Windows (NIH, USA). The first antibody was ZEB1 (200 kDa, ab203829, dilution 1:500, Abcam, UK), GAPDH (37 kDa, 14C10, dilution 1:1000, Cell Signaling Technology, USA). The secondary antibody was HRP-conjugated Goat anti-rabbit IgG (abs20040ss, dilution 1:1000, absin, China).

Luciferase reporter assay

The binding sites between lnc171 and mir-873-5p, and between mir-873-5p and ZEB1 were predicted by online website. The DNA fragments of wild-type-lnc171 (lnc171-WT), mutant-lnc171 (lnc171-MUT), Wild-type-ZEB1, and mutant-ZEB1 were cloned into the pmiRGLO vector (Promega, USA). The HEK-293T cells was selected to perform luciferase reporter assay and were seeded on 96-well plate, then mir-873-5p NC or mir-873-5p mimics were transfected with wild- or mutant-type luciferase plasmids using Lipofectamine 3000 for 48 h. After transfection, the fluorescence intensity of firefly and Renilla was measure by Luciferase reporter kit (Yisheng Bio-Technology, China) and Multifunctional enzyme labeling instrument (Tecan, CH).

Bioinformatics analysis

The target gene of lnc171 or mir-873-5p were predicted by miRDB (https://mirdb.org/), TargetScan (https://www.targetscan.org/), and MicroT-CDS (https://dianalab.e-ce.uth.gr/html/dianauniverse/index.php?r=microT_CDS). miRDB could show the binding sites of target genes. UALCAN (https://ualcan.path.uab.edu/index.html) was used to obtain expression level data in HCC patients of ZEB1 that link TCGA database. ONCOLNC (http://www.oncolnc.org/) was used to obtain survival data of ZEB1 which link TCGA database.

Statistical analysis

All data were analyzed using SPSS20.0 software. Using GraphPad Prism 5.0 software to draw graphs. All data were represented as the mean ± S.D. The comparison of the data between two experimental groups was conducted by using Student’s t-test, and the comparison of multiple groups was conducted by using one-way ANOVA. p < 0.05 (*), p < 0.005 (**) and p < 0.0005 (***), were significant statistical.