Tissue samples
A total of 36 matched pairs of liver cancer tissues and their corresponding noncancerous liver tissues were obtained from AJOU University Hospital and Keimyung University Hospital, a member of National Biobank of Korea. Written informed consent was obtained from each subject according to the Declaration of Helsinki, and the study was approved by the Institutional Review of Board (IRB) of the Songeui Campus, College of Medicine, the Catholic University of Korea (IRB approval number: MC18TESI0075, MC19TESI0016).
Cell culture
Human liver cancer cells (Hep3B, HepG2, Huh7, PLC/PRF/5, SK-HEP-1, SNU-182, SNU-354, SNU-368, SNU-387, SNU-423, SNU449, and SNU-475) were obtained from KCLB (Korean Cell Line Bank, Seoul, Korea). Normal liver cell line MIHA was kindly provided by Dr. Roy-Chowdhury (Albert Einstein College of Medicine, Bronx, NY). All of the cell lines were held in an RPMI-1640 or DMEM medium with 10% fetal bovine serum added and 100 units/ml of penicillin/streptomycin (GenDepot, Katy, TX). All cells were cultured at 37°C in a humidified incubator with 5% CO2.
Transfection and treatment
Small interfering RNAs (siRNAs) were synthesized by Genolution (Seoul, Korea) or purchased from BIONEER (Daejeon, Korea). The sequences of the siRNAs, miRNA mimics and antisense miRNAs are listed in Supplementary data, Table S5. Human ADAR1-p110, MSI2 and SLC38A4 expression plasmid, subcloning gene ORF sequence in pcDNA3.1+/C-(K)-DYK plasmid, was purchased from Genscript™ (Piscataway, NJ, USA). Transfections were carried out using Lipofectamine RNAiMAX or Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s instructions.
RNA, DNA extraction, RT-PCR and qRT-PCR
Total RNA from frozen tissues and cells were isolated using Trizol reagent (Invitrogen, Carlsbad, CA, USA). One microgram of total RNA was reverse transcribed into cDNA using Tetro cDNA Synthesis Kit (Bioline, London, UK) according to the manufacturer’s instructions. The quantitative real-time PCR (qRT-PCR) were performed with SensiFAST SYBR No-ROX Kit (Bioline, London, UK) and monitored in real-time by an iQ™-5 (Bio-Rad, Hercules, CA, USA). The average threshold cycle (Ct) value, from triplicate assays were used for further calculations. Normalized gene expression was determined using the relative quantification method. Results were expressed as the mean value of triplicate experiments. Genomic DNA from tissue and cells were isolated using DNAzol reagent (Invitrogen, Carlsbad, CA, USA), according to the manual. qRT-PCR was performed as described above and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as endogenous loading control. The sequences of primers used in qRT-PCR are listed in Supplementary data, Table S6.
FLAG immunoprecipitation
Cells were transfected with pcDNA3.1_ADAR1-p110 or pcDNA3.1_MSI2, which encodes a FLAG-tag. 48 hr after transfection, the cells were washed with phosphate-buffered saline (PBS) and lysed at 4°C in PBS, pH 7.2, containing 1.0% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 10mM NaF, 1.0 mM NaVO4, and 1.0% protease inhibitor cocktail (Sigma). The FLAG tag was immunoprecipitated with anti-FLAG DynaBeads (Invitrogen, Carlsbad, CA) during an overnight incubation. Immunoprecipitated proteins were eluted using 3X FLAG peptide (Sigma) and analyzed by Western blot, probing with anti-FLAG antibody (Cell Signaling). For primary-mir-3144 pulldown analysis using qRT-PCR, RNA was isolated and reverse-transcribed using a miScript II RT kit (Qiagen).
Western blot analysis
Cells were lysed using a lysis buffer (50 mM HEPES, 5 mM EDTA, 50 mM NaCl, 1% Triton X100, 50 mM NaF, 10 mM Na2P4O7, 1 mM Na3VO4, 5 ug/mL aprotinin, 5 ug/mL leupeptin, 1 mM PMSF and protease inhibitor cocktail). Lysates containing equal amounts of proteins were separated by SDS-PAGE and transferred onto a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA, USA). The blots were blocked with a 5% skim milk solution and incubated with the following antibodies: anti-ADAR1, anti-CTNNB1, anti-GAPDH, anti-MSI2, anti-MET, and anti-SLC38A4 (Santa Cruz Biotechnology, Dallas, TX, USA), and anti-Flag-Tag (Cell Signaling Technology, Danvers, MA, USA). The Immobilon™ western blot detection system (Millipore) was used to detect bound antibodies. The intensities of the western blot bands were quantified using LAS-4000 (Fuji Photo Film Co., Tokyo, Japan).
Cell growth assay
Cells were seeded in a 6-well plate for transfection. After transfection, cells were incubated with 0.5 mg/ml of MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] solution (Sigma) for 1 hr. The dark blue formazan products formed by viable cells were dissolved in dimethyl sulfoxide (DMSO; Sigma), and absorbance was measured using a VICTOR3 Multilabel plate reader (PerkinElmer, Waltham, MA).
Bromodeoxyuridine (BrdU) incorporation assay
Cells were seeded in a 24-well plate to 40% ~ 50% confluency. The assay was performed with a Bromodeoxyuridine (BrdU) cell proliferation assay kit (Millipore) in accordance with the manufacturer’s protocol every 24 hr.
Clonogenic assay
Cells were transfected with miRNA mimics or siRNA in 60 mm2 cell culture plates. After transfection for 24 hr, cells were re-seeded in 6-well plates and incubated for 2 weeks. Next, cells were washed with PBS and fixed with 1% paraformaldehyde for 30 min at room temperature. Fixed cells were stained with 0.5% crystal violet for 1 hr at room temperature. Colonies were counted using a clono-counter program.
Apoptosis assay
To measure levels of apoptosis, Annexin V-FITC Apoptosis Detection Kit I (BD Biosciences, San Jose, CA) was used. After transfection, liver cancer cells were washed with cold phosphate buffered saline (PBS) and resuspended in 1× binding buffer. Then, 1×10^5 cells were transferred to a 5 ml culture tube and mixed with 5 µl of annexin V-FITC and 10 ul of propidium iodide solution. After 20 min at room temperature in the dark, 400 µl of 1X binding buffer was added to each tube, and apoptotic fractions were measured by a FACS Calibur flow cytometer (BD Biosciences).
Cell cycle analysis
Liver cancer cells were transfected with miRNA mimics or siRNA in 60 mm dishes. After 48 h incubation, cells were harvested Trypsin, washed with cold PBS, fixed in 70% ethanol, resuspended in 200 ul PBS containing 1 mg/ml RNase and incubated in the dark for 30 min at 37o C. Nuclei were stained with 50 ug/ml propidium iodide (BD Biosciences). Stained cells fractions were measured using Cell-Quest FACS analysis software (BD Biosciences).
Migration and invasion assay
For in vitro cell migration and invasion assay, cell motility was measured by using a modified Boyden chamber assay. For invasion assay, Matrigel (BD Biosciences) was diluted to a concentration of 0.3 mg/ml with coating buffer. One hundred microliter aliquots of Matrigel were used to coat the upper surface of the Transwell cell culture inserts. After incubation for 2 h at 37°C, the inserts were ready to be seeded with cells. After preparation, cells were plated on the top surfaces of Transwell inserts, and the inserts were placed in a 24-well plate. The lower wells contained 2% FBS as a chemoattractant. The plate was incubated overnight and stained using a Diff-Quik staining kit (Sysmex, Kobe, Japan). The cell images were captured using an Axiovert 200 inverted microscope (Zeiss, Oberkochen, Germany) at × 200 magnification, and the number of cells was counted in three random image fields.
Wound healing assay
Cells were transfected and incubated for 24 hr in 60 mm2 cell culture plates. Then, cells were trypsinized, and 1 × 10^6 cells per well were seeded in a 6-well cell culture plate. After overnight incubation, cell monolayers were scraped with a sterile micropipette tip. Initial gap widths 0 hr after scratching and residual gap widths 24 hr after scratching were photographed using an IX71 photomicrograph (Olympus, Tokyo, Japan).
Mutagenesis
For mutagenesis of ADAR1-p110 adenosine deaminase activity, a QuickChange kit (Agilent Technologies, Palo Alto, CA, USA) was used according to the manufacturer’s instruction.
Mouse liver cancer model
The H-ras homozygous transgenic mice were kindly provided by Dr. Dae-Yeoul Yu (Laboratory of Human Genomics, Korea Research Institute of Bioscience and Biotechnology, Daejeon, Korea).13 Male mice spontaneously developed liver cancer beginning at approximately 15–18 weeks of age. Mouse livers were harvested at 24 weeks of age, and processed for the experiments. All procedure of animal research were provided in accordance with the Laboratory Animals Welfare Act, the Guide for Care and Use of Laboratory Animals and the Guidelines and Policies for Rodent experiment provided by the IACUC (Institutional Animal Care and Use Committee) in school of medicine, The Catholic University of Korea. (Approval number: CUMS-2019-0115-02)
Mfuzz clustering
The gene expression profiles were log-normalized and clustered using the c-mean algorithm by the Bioconductor Mfuzz package v. 2.30.0 as indicated by the author.14
Analyses of publicly available genomic data
To investigate differential gene expression of coding and non-coding RNAs in multi-stage liver disease, data were obtained from ‘The Cancer Genome Atlas’ liver cancer project (TCGA_LIHC), the International Cancer Genome Consortium liver Cancer–RIKEN, JP (ICGC_LIRI) and the Gene Expression Omnibus (GEO) database of the National Center for Biotechnology Information (NCBI) (Accession Numbers: GSE6764, GSE77314, GSE114564 and GSE174608). Level 3 mRNA expression data from TCGA-LIHC HTSeq-FPKM were log2 transformed [log2(fpkm + 1)] and used to assess gene expression.
MicroRNA target prediction
An in silico analysis was conducted to predict target candidates of miR-3144-3p and ED_miR-3144(3_A < G) by using the TargetScan algorithm (http://www.targetscan.org/). The sequence information of miR-3144-3p was obtained from miRbase database (http://www.mirbase.org).
Analysis of miRNA A-to-I editing using the catholic_mLIHC and TCGA_LIHC datasets
All bam files were converted to fastq format using bedtools bamtofastq. To remove adaptors, and low-quality and inadequate-length reads from the our human multi-stage liver cancer transcriptome data (Catholic_mLIHC) and TCGA_LIHC data, Cutadapt was used with the “-a adaptor_sequence --quality-base 33 -m 15 -M 28 -f fastq -O 3 -q 20” options. Next, Bowtie was used to align the filtered read to the human genome (hg19) with the "-n 1 –e 50 –a –m --best --strata --trim3 2” option for the best alignment, one mismatch per read, and no cross mapping. In the miRNA editing profile stage, all procedures were performed using the scripts reported in Wang et al. 15 Briefly, first, the binomial test was performed, and only results with a Bonferroni-corrected p-value of 0.1 or less were selected. Second, all SNPs present on mitochondria and in sites other than pre-miRNA were removed. Finally, in order to exclude all SNPs already reported, all mutations overlapping with information recorded from the dbSNP and gnomAD were removed. To obtain more meaningful results, miRNAs with an average editing rate of 5% or more and a TPM expression value of 1 or more were selected for the Catholic_mLIHC, and miRNAs with an editing rate of 5% or more in at least 10 samples or more were selected in the TCGA_LIHC dataset.
In vivo tumorigenesis study
Ras-Tg mice were intravenously injected with Invivofectamine 3.0 (Invitrogen, Carlsbad, CA, USA) containing 0.25 mg/kg of Adar1 and Msi2 siRNAs as previously described.16 Ras-Tg mice were also intravenously injected with 50 µg of pcDNA3.1_Slc38a4 using Turbofect in vivo Transfection Reagent (Invitrogen, Carlsbad, CA, USA). The ultrasonography images were taken at 17, 19, 21 and 23 weeks of age with an ultrasound machine (Philips, Amsterdam, Nederland) by the same medical imaging expert each time.
Statistical analyses
Survival curves were plotted using the Kaplan-Meier product limit method, and significant differences between survival curves were determined using the log-rank test. All experiments were performed at least three times, and all samples were analyzed in triplicate. The statistical significance of the difference between experimental groups was assessed by paired or unpaired Student’s t tests using GraphPad 7.0 software (GraphPad Software Inc.). Statistical significance was determined for p < 0.05. A chi-square test (two-sided) was used to determine associations between parameters. Receiver operating characteristic (ROC) curves for each candidate marker were analyzed to calculate sensitivity, specificity, and areas under the curve (AUC) with 95% confidence intervals.