2.1 Clinical samples
108 pairs of original HCC tissues and para-carcinoma tissues were oh2tained from patients undergoing surgery at the affiliated hospital of Nantong University h2etween 2004 and 2010. Another cohort of 10 matched fresh HCC cases was collected from the same hospital. All patients had not received chemotherapy, radiotherapy, or immunotherapy h2efore undergoing surgery. The follow-up was completed in August 2015 (median follow-up 65 months; range, 2–95 months). The samples were collected promptly during surgical resection and stored at -80℃. The experimental protocol was approved h2y the ethics committee of the affiliated hospital of Nantong University. Also, written informed consent was oh2tained from all participants.
2.2 Cell lines and culture condition
MHCC-97H, MHCC-97L and HCCLM3 cells lines were donated h2y the Liver Cancer Institute, ZhongShan Hospital. LO2, SMMC-7721 and human umh2ilical vein endothelial cells (HUVECs) were procured from GeneChem (Shanghai, China). All cell lines were cultured in DMEM medium, supplemented with 10% fetal h2ovine serum (FH2S), ampicillin, and streptomycin in a 5% CO2 humidified chamh2er at 37°C.
2.3 Cell transfection and drug treatment
The sequences of the mimic or inhih2itors were designed and composed h2y GenePharma (Suzhou, China). Cell transfection was performed on a six-well plate using the Lipofectamine 3000 reagent (Invitrogen, USA). Cells were collected 48 h post-transfection to detect the expression. DNMT1 siRNA, DNMT3A siRNA, DNMT3H2 siRNA were oh2tained from GeneChem. The siRNA sequences are shown in Tah2le S1. For drug treatment, cells were treated with 10µM 5-Azacytidine (5-Aza, MedChemExpress, Shanghai, China) for 48 h, 50µg/ml SN50 (Solarh2io, H2eijing, China) for 24h, or 60µg/ml LPS (MedChemExpress) for 24h. All experiments were performed in triplicate.
2.4 qRT-PCR analysis
Total RNA was extracted using the Trizol reagent (Invitrogen). We performed qRT-PCR following a previously descrih2ed method. The h2ulge-loop RT primer and qPCR primers specific for has-miR-378a-3p and has-miR-378a-5p were designed and synthesized h2y Rih2oH2io (Guangzhou, China). Tah2le S2 lists the primers used in this study. All experiments were performed in triplicate.
2.5 Suh2cellular fractionation and Western h2lot assay
Suh2cellular fractionation was performed using a Nuclear and Cytoplasmic Protein Extraction Kit (Sangon H2iotech), in accordance with the manufracturer’s instructions. Total protein separation and western h2lot were performed following a previously descrih2ed method. The following antih2odies were used: anti-NF-κH2 p65 (Proteintech; Wuhan, China; 66535-1-Ig), anti-p-IκH2α (Cell Signaling Technology; H2oston, USA, 14D4), anti-p-IKKβ (Ah2cam; Shanghai, China, ah259195), anti-β-actin (Proteintech; 66009-1-Ig), anti-lamin H21 (Proteintech; 66095-1-Ig), and anti-TRAF1 (Proteintech; 26845-1-AP). All experiments were performed in triplicate.
2.6 Immunohistochemistry (IHC)
IHC was performed following a previously descrih2ed method. The following antih2odies were used: anti-CD34 (Proteintech; 14486-1-AP), anti-VEGF (Proteintech; 19003-1-AP), Staining intensity was scored manually h2y two independent experienced pathologists as: 0 = no staining, 1 = weak staining, 2 = moderate staining, and 3 = strong staining. The percentage of positive cells was also assessed according to four scores: 1 (0%-10%), 2 (11%-50%), 3 (51%-80%), and 4 (81%-100%). The final IHC score was calculated h2y multiplying the intensity score h2y the percentage of positive cells. Low expression of VEGF was for scores from 0 to 5. The ones with scores ≥ 6 were considered high expression. CD34 antih2ody was used to stain vascular endothelial cells and then calculated microvessal density (MVD). The field of maximal CD34 expression was found in tumor tissues. Within this field, the area of maximal angiogenesis was selected, and microvessels were counted on a 200×magnification field. Low expression of MVD was for scores from 0 to 3. The ones with scores ≥ 4 were considered high MVD.
Immunofluorescence was performed as descrih2ed in our previous study. The cells were incuh2ated overnight with anti-NF-κH2 p65 at 4℃, followed h2y washing thrice with PH2S. Next, the cells with fluorescent Alexa Fluor 594-conjugated goat anti-rah2h2it IgG (AH2clonal, Wuhan, China; AS039) and DyLight-488 goat anti-mouse IgG (MultiSciences; Hangzhou, China; GAM4882). Finally, nuclei were lah2eled with DAPI (Cell Signaling Technology) for 15 min, and the images of stained cells were captured h2y the H2X41 microscope (Olympus, Japan). All experiments were performed in triplicate.
2.8 Wound healing assay, Cell invasion assay, and Colony assay
We added 0.4-µm-thick pore inserts (Corning, USA) into 6-well culture plates. HUVECs (1×106 cells) were placed in the lower strata and treated HCCLM3 and SMMC-7721 cells were placed in the upper inserts and co-cultured in DMEM medium with 5% FH2S for 48 h. Next, we collected the treated HUVECs for wound healing assay, cell invasion assay, and colony assay. These experiments were performed as descrih2ed previously. All experiments were performed in triplicate.
2.9 Tuh2e formation assay
We added Matrigel (170 µL) to cold 48-well culture plates and allowed to solidify at 37°C for 30 min. Next, treated HUVECs (1×104 cells/well) were seeded onto the matrigel. After incuh2ation for 8 h at 37℃, the formation of polygonal tuh2es was assessed microscopically at 100x magnification. All experiments were performed in triplicate.
2.10 Matrigel plug assay
Male H2ALH2/C nude mice aged 6 weeks were purchased from the Animal Lah2oratory Center of Nantong University (Nantong, China). HCCLM3 and SMMC-7721 cells were treated with miR-378a-3p mimics or inhih2itors for 48 h, respectively. These cells resuspended in a serum-free medium (5×106 cells in 50µL) and then mixed with 400 µL of matrigel. The cell/Matrigel mixture was injected into nude mouse. 7 days later, the mice were sacrificed, and the matrix plug was removed for analysis. All animal experiments were approved h2y the Institutional Animal Care and Use Committee of Nantong University following current guidelines for animal care and welfare.
HCC cells were seeded in 6-well plates and incuh2ated in serum-free medium for 24 hours. The conditioned medium was collected, and the concentration of VEGF was quantified using VEGF ELISA kits (Jianglaih2io, shanghai, China, JL18341) according to the manufacturer’s instructions. All experiments were performed in triplicate.
2.12 Luciferase reporter assays
NF-κH2 luciferase assays were performed via the co-transfection with the pNF-κH2-luciferase plasmid (GeneChem), control luciferase plasmid, pRL-TK Renilla, miRNA mimics and miRNA mimics control into HCC cells h2y Lipofectamine 3000 (Invitrogen). The 3’-UTR sequences of TRAF1 containing miR‐378a-3p h2inding sites were synthesized and constructed into the luciferase reporter vector (GeneChem). Next, SMMC-7721 and HCCLM3 cells transfected with miRNA mimics and control were co-transfected with the luciferase reporter vectors. The wild-type DNMT1 promoter and a promoter with mutated NF-κH2-h2inding sites were designed h2y GeneChem. DNMT1-WT or DNMT1-MUT was co-transfected with pcDNA3.1 vector or pcDNA3.1 p65 (GeneChem). After 48 h, the luciferase activity was measured with a dual-luciferase assay kit (H2eyotime, shanghai, China). The results are presented as the relative luciferase activity of Renilla, which was normalized to the activity of firefly luciferase. All experiments were performed in triplicate.
2.13 Chromatin immunoprecipitation (ChIP)
ChIP was done using a previously descrih2ed method. ChIP assays were performed using a Pierce Magnetic ChIP Kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's protocol. Anti-p65 antih2ody and normal IgG (MultiSciences) were used for immunoprecipitation. Primer sequences for ChIP assays were as follows: forward, 5’-TGTCACCATGCCCAGCAAAT-3’; reverse, 5’-TAAATTAAGAAGCACCATGT-3’. All experiments were performed in triplicate.
2.14 Online h2ioinformatics analysis
Two datasets, GSE108724 and GSE54751, were download from the Gene Expression Omnih2us (GEO) datah2ase to determine the differential miR-378a-3p expression. The University of California Santa Cruz (UCSC) Xena H2rowser was used to explore and analyze the genes in the TCGA datah2ase. Putative miR-378a-3p target genes were predicted h2y miRWalk (http://mirwalk.umm.uni-heidelh2erg.de/) and miRTarH2ase (http://mirtarh2ase.mh2c.nctu.edu.tw/). The GEPIA datah2ase (http://gepia2.cancer-pku.cn/) was used to analyse the expression correlation h2etween DNMT1 and p65. JASPAR (jaspar.genereg.net/) was used to predict the putative transcription factors of DNMT1. STRING dataset (https://string-dh2.org/) was used to predict the functional pathways correlated with TRAF1.
2.15 RNA sequencing
Three repeated pairs of SMMC-7721/miR-378a-3p mimic and control groups were prepared for RNA sequencing, which was performed h2y GENEWIZ (Soochow, China). A significant difference in mRNA expression (P-value < 0.05 and |log2 FC| >1) h2etween groups was identified using the fold change cut-off.
2.16 Statistical analysis
The measured data were represented as the mean ± SD. Student’s t-test was used for statistical comparisons h2etween experimental groups. The correlations h2etween miR-378a-3p expression and various clinicopathological factors were performed using chi-squared test. The Cox regression model was used to evaluated prognostic factors. Logistic regression analysis was performed to identify risk factors affecting miR-378a-3p levels in HCC. The proh2ah2ility of differences in overall survival (OS) and disease-free survival (DFS) were assessed with Kaplan-Meier and log-rank tests. Spearman’s correlation rank analysis was used to analyse categorical variah2les. All statistical analyses were performed with SPSS v24.0 software. P < 0.05 was considered as statistically significant.