Data collection

Gene expression data and corresponding clinical information were obtained from two public databases, The Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/) and GEO (https://www.ncbi.nlm.nih.gov/geo/). Furthermore, the cohort of Chinese patients with hepatitis B virus (HBV) infection (CHCC-HBV) included 159 patients with liver cancer who underwent primary radical resection at Zhongshan Hospital (Fudan University, Shanghai, China) between 2010 and 2014 [21]. The dataset details are shown in Additional file 1: Table S1, and the clinicopathologic characteristics of patient cohorts was calculated the significance using Table 1 (https://benjaminrich.github.io/table1/vignettes/table1-examples.html) R package are presented in Additional file 1: Table S2.

Cell culture

HEK 293T cells, Huh7, MHCC-97H, and HepG2 HCC cell lines were obtained from the Chinese Academy of Sciences Cell Bank in Shanghai, China. The cells were cultured in DMEM (Gibco, USA) supplemented with 10% fetal bovine serum (FBS, Sigma, USA) and 1% Penicillin/Streptomycin (P/S, Gibco, USA) at 37°C in the presence of 5% CO2.

Knockdown and overexpression of hnRNPA2B1

We constructed sgRNA targeting human hnRNPA2B1 and PCK1 plasmid, and non-sense scrambled control plasmid with lenti-CRISPR-V2 vector (Addgene #52,961) based on CRIPSR-Cas 9 strategy using the following primers presented in the Additional file 1: Table S2. In addition, we constructed overexpression plasmid of human hnRNPA2B1 and PCK1 with pCDH-CMV-GFP-Puro vector (Addgene #91,892). Empty vector pCDH-CMV-GFP-Puro was used as control. Lentiviral particles were produced in HEK293T cells with hnRNPA2B1/PCK1 sgRNA plasmid or hnRNPA2B1/PCK1 overexpression plasmid together with two package plasmids psPAX2 (Addgene #12,260) and pMD2.G (Addgene #12,259). HCC cell lines were infected with the viral supernatants. After 24 h of infection, 5 μg/ml puromycin was added for cell screening to establish a stable knockout or overexpression cell lines. Finally, the expression of hnRNPA2B1 and PCK1 were confirmed by RT-PCR and western blot. The sequences of the cloning primers and sgRNA are listed in Additional file 1: Table S3.

Quantitative real-time PCR

Total RNA was extracted using RNA Isolation Kit (OMEGA, Norcross, USA) according to the manufacturer's instructions and reverse transcribed into cDNA using HiScript II Q Select RT SuperMix Kit (Vazyme, Nanjing, China). Quantitative real-time PCR (RT-PCR) was performed using the ChamQ Universal SYBR qPCR Master Mix Kit (Vazyme, Nanjing, China). The relative amount of each gene was quantified by the 2 − ΔΔCt method and normalized by the expression of β-actin. All quantitative RT-PCR experiments were performed in at least three independent experiments for culture cells. Primer sequences are listed in Additional file 1: Table S3.

Western blotting

The cells were lysed with RIPA Lysis buffer (Thermo Fisher Science, USA) that contained a protein inhibitor cocktail (Merck, USA). The protein concentration was assessed using the BCA kit (Thermo Fisher, USA). The PVDF membranes were blocked with non-fat milk and then incubated with primary antibodies indicated as follows: hnRNPA2B1 rabbit antibody (1:1000, #A1162, Abclonal, China), PCK1 rabbit antibody (1:1000, #A22172, ABclonal, China), β-Actin Mouse antibody, (1:1000, #AC004, ABclonal, China), Glycolysis Antibody Sampler Kit (#8337, Cell Signaling Technology, USA). After being washed with TBST, the membranes were incubated with corresponding HRP-conjugated secondary antibodies. The membranes were visualized with the BioRad VersaDoc 4000 imaging system, and densitometric analysis of proteins was quantified by Quantity One software (BioRad, CA, USA).

Immunohistochemistry staining

The liver cancer tissues from the mice were fixed, embedded in paraffin, and cut into 5 μm thick sections for immunohistochemical (IHC) staining. The tissue sections were blocked with 10% donkey serum for 1 h and incubated with primary antibodies against hnRNPA2B1or PCK1 (1:200, #A1162, #A22172, ABclonal, China) or Ki67 (1:1000, #ab15580, Abcam, UK), and then incubated with biotinylated anti-rabbit IgG secondary antibodies. Finally, the sections were visualized with a Vectastain ABC kit (Vector) and developed with 3,3′-diaminobenzidine (DAB) (Genetech, China). Tissue array chips containing HCC tissues and adjacent normal liver tissues were purchased from Servicbio (ZL-LivHCC961, China) and immunohistochemistry was performed with hnRNPA2B1. The staining score was calculated based on staining intensity and the percentage of positive cells.

CCK-8 assay and colony formation assay

Cell proliferation was tested by CCK-8 kit (Dojondo, Japan) and the absorbance was measured at 450 nm at different time points (0, 24, 48, and 72 h) to determine the proliferation rate. Experiments were independently repeated three times. To evaluate the ability of colony formation of HCC cells, the cells were seeded at a low density of 1000 cells per well of 6-well plates and allowed to grow for approximately 14 days to form visible colonies. The cells were then fixed with 4% paraformaldehyde (PFA) about 15 min at room temperature and stained with 0.2% crystal violet (Solarbio, China) for 30min and photographed.

Transwell migration and invasion assay

Cell migration and invasion were determined by Transwell chambers with 8μm pore size (Corning, USA). As for migration assay, HCC cells (4 × 104 cells) were seeded into the upper chamber of the transwell in serum-free DMEM, while DMEM containing 10% FBS was added to the lower chamber. 48 h later, non-migrating cells on the upper chamber were removed and migrated cells attached to the lower side of the membrane were fixed with 4% PFA for 15 min and stained with 0.2% crystal violet. The migration cells were visualized using an inverted microscope (Leica, Germany) and were counted. Prior to the start of the invasion assay the upper chamber of the transwell plates were coated with Matrigel (Matrigel diluted 1:8 with serum-free medium) at 37 ℃ for 2 h, then HCC cells were seeded into the upper chamber. The other steps were the same as for the migration assay.

Xenograft mouse models

Five-week-old male BALB/c nude mice were purchased from Shanghai SLAC Co. Ltd (Shanghai, China) and used in subcutaneous tumor-formation experiments. The mice were randomly divided into two groups of six mice per group. Huh7 cells infected with lentivirus containing hnRNPA2B1-sgRNA or NT sgRNA were injected right-back of each mouse. And 1 × 106 cells were injected into each mouse. 28 days after injection cells, the mice were sacrificed, and the tumors were isolated and photographed. All animals were approved by the Institutional Animal Care and Ethics Committee of Renji Hospital, with a maximum allowable tumor volume of 2000 mm3.

Hydrodynamic injection HCC models

We first constructed sgRNA plasmids targeting mouse hnRNPA2B1 and PCK1 genes based on the CRISPR-Cas9 method. The primers used to construct sgRNA plasmids are listed in Additional file 1: Table S2. The sgRNA plasmids were confirmed by sequencing. Next, six-week-old female C57BL/6 mice were randomly grouped into six mice per group. 2 ml plasmid mixture in 0.9% sodium chloride solution was prepared for each mouse. The plasmid mixture contained 13 μg pT3-EF1A- MYC-IRES-luc (Addgene 129775) (22), 6.5 μg pX330-p53 (Addgene 59910) (23), and 6.5 μg CMV-SB13 transposase, along with either hnRNPA2B1-sgRNA/PCK1-sgRNA or NT sgRNA plasmid. Then, 2 ml plasmid mixture was injected into the tail vein within 5–7 s. Mice were monitored by abdominal palpation and euthanized when they had a high burden of liver tumors.

RNA sequencing

First, we harvested hnRNPA2B1 sgRNA Huh7 cells and NT sgRNA Huh7 cells and total RNA was extracted using an RNA Isolation Kit (OMEGA, Norcross, USA). The purity and quality of total RNA were assessed by Nanodrop and Agilent 2100 Bioanalyzer. dsDNA library was constructed using VAHTS Universal V6 RNA-seq Library Prep Kit (Vazyme, China) and was purified using VIHTS DNA Clean Beads (Vazyme, China). Then, the dsDNA library was qualified and quantified by Qsep-400 and Qubit TM dsDNA Assay kit. The qualified dsDNA samples were subsequently sequenced using the Illumina Nova Seq 6000 platform (San Diego, USA). Raw sequencing reads were mapped to GRCh38 assembly of the human genome using Tophat2, version 2.0.10 [21]. Fragments Per Kilobase of transcript per Million mapped (FPKM) were computed using Stringtie and normalized with trimmed mean of M values (TMM) [22, 23]. Differentially expressed genes (DEGs) were analyzed using DESeq 2. Genes with log-fold change > 1.5 and false discovery rate (FDR) < 0.05 were considered significantly transcriptomic changes. This sequencing dataset is available at GEO: GSE226544.

Metabolite analysis

When cells were cultured up to 90% confluency, the cells were collected with a cell scraper after washed with PBS, and the cells were pelleted by centrifuging and stored in liquid nitrogen. A homogenate of 50 mg of sample mixed with 1 mL of cold methanol/acetonitrile/H2O (2:2:1, v/v/v) was sonicated twice (30 min/ once) at a low temperature and then incubated at − 20 ℃ for 1 h, followed by centrifugation for 20 min (140,00g, 4 °C). The supernatant was dried in a vacuum centrifuge. Sextuplicate samples were collected and sent to analyze metabolites by Metabolon-associated energy metabolism (Applied Protein Technology, Shanghai, China). For LC–MS analysis, the dried samples were dissolved in 100 μL acetonitrile/H2O (1:1, v/v). The samples were separated by Agilent 1290 Infinity LC ULTRA performance liquid chromatography system. Then A 5500 QTRAP mass spectrometer (MS, AB SCIEX) was used to analyze the chromatographic peak area and retention time in anion mode.

RNA binding protein immunoprecipitation (RIP) assay

For the RIP experiment, 1 × 107 cells were lysed in RIPA buffer (50 mM Tris–HCl, pH 8.0, 150 mM NaCl, 5 mM EDTA, 1% NP-40) for 10 min. After centrifugation, the cell lysates were pre-cleared with Protein A/G beads (Santa Cruz, Shanghai, China) and then incubated for 2 h with the antibody against hnRNPA2B1 or IgG. After washing, proteins and DNA were digested with proteinase K and DNase I. Total RNAs were extracted using TRIzol reagent and RT-PCR analysis was then performed on the purified RNA.

Methylated RNA immunoprecipitation (MeRIP) assay

MeRIP assay was performed with a riboMeRIP m6A transcriptome Profiling Kit (RIBOBIO, China) according to the guidelines of the manufacturer. Briefly, First, RNA fragmentation. The total RNA of HCC cells was first extracted using TRIzol reagent (Thermo Fisher Scientific, USA), and then 18 μγ RNA at 1 μγ/ml was fragmented in the RNA fragmentation Buffer according to the procedures of the manufacturer. Next, fragmented RNAs were precipitated with ethanol. Second, Preparation of the m6A magnetic beads. Protein A/G magnetic beads were incubated with m6A antibody for 30 min, then washing for 3 times. Third, Immunoprecipitation. Magnetic beads with m6A antibody were incubated with the above fragmented RNAs in IP buffer for 2 h at 4 ℃. Then, the magnetic beads with m6A Ab and RNAs were added into the pipes adsorbed on the magnetic frames and washed. After that, m6A methylated RNAs were eluted and purified. Fourth. RT-PCR according to described above. The primer of PCK1 was listed in Additional file 1: Table S3.

Immune infiltration analysis

ImmuCellAI (http://bioinfo.life.hust.edu.cn/web/ImmuCellAI) was used to predict the infiltration abundance of 36 immune cell types in cancer tissue using RNA-Seq data or gene-expression profiles derived from microarray data [24]. The normalized gene expression matrix was uploaded to ImmuCellAI for immune infiltration analysis. The Wilcoxon rank sum test to compare difference between groups. Spearman correlation analysis was performed to investigate the correlation of the expression of hnRNPA2B1 and immune cells.

Statistical analysis

Statistical analysis of bioinformatic analysis was performed by R software version 4.1.3. At least three independent tests were performed for each cell culture experiment and the data are presented as mean ± SEM and the statistical analysis was evaluated by SPSS software 22.0 and statistical significance was assessed using the Student t-test to compare two groups, and the Chi-square test was used for the comparisons of categorical variables. Pearson correlation analysis was employed to determine the association between hnRNPA2B1 and other genes. Kaplan–Meier survival analysis was conducted to investigate the impact of hnRNPA2B1 on overall survival (OS) and relapse-free survival (RFS) in patients with liver cancer. The log-rank test was used to determine the statistical significance. p-values less than 0.05 indicates statistical significance. *p <0.05, **p <0.01, ***p <0.001, ****p <0.0001.