Data collection
The RNA sequencing data of 374 HCC samples and 50 adjacent samples and their associated clinical data were downloaded from the TGCA database (http://portal.gdc.cancer.gov/repository). The autophagy-associated gene set was obtained via Human Autophagy Database (http://www.autophagy.lu/index.html), and the ferroptosis and pytoptosis-associated gene sets were summarized from previously reported studies (46, 47). The three types of gene sets are presented in Supplementary Table S1.
Identification of ferroptosis as the target analysis gene sets
The “limma” package was used to analyze the differentially expressed gene sets of the ferroptosis, autophagy, and pytoptosis-associated gene sets. The cut-off value for differential genes was as follows: the absolute value of log fold change (│logFC│) ≥1.5, p-value < 0.05. And then used ssGSEA to get the enrichment scores of the ferroptosis-, autophagy-, and pytoptosis-associated gene sets using the “GSVA” R package. Finally, the “survival” R package was applied to compare the survival differences of the ssGSEA scores of the ferroptosis-, autophagy-, and pytoptosis-associated gene sets based on the median cutoff value. The ferroptosis-associated gene set was determined to be aimed for further exploration through survival, univariate, multivariate, and differential analysis.
Analysis of co-expression module construction of HCC
According to the mRNA, lncRNA, and miRNA expression data in LIHC, the co-expression analysis was carried out with the “WGCNA” package. The gradient method was used to test the independence and the average connectivity degree of different modules with different power values (ranging from 1 to 30). The appropriate power value was determined when the degree of independence was 0.8. The correlation between each co-expression module and ferroptosis-related ssGSEA scores was calculated, and the module genes with the strongest correlation were screened for subsequent ceRNA network construction.
Functional enrichment analysis of ferroptosis-related modules
Based on the clusterProfiler R package (48) and GOplot R package (49), candidate genes were analyzed for functional enrichment. Using a statistical algorithm (Fisher's exact test) to find the specific functional items with which the module gene is most associated. The candidate genes here refer to the mRNA modules most associated with ferroptosis. Each item in the results corresponded to a statistical value P-value to indicate significance.
LncRNA-miRNA-mRNA network construction
The lncRNA-miRNA-mRNA network was constructed according to the classical ceRNA network as follows: (a) Interaction prediction of miRNA-mRNAs in the miRmap (50), miRanda (51), miRDB (52), TargetScan (53) and miRTarBase (54), and miRNA-lncRNAs in the miRcode (55), STRBase (56) were identified; (b) Both the lncRNAs and mRNAs were identified by co-expressing negatively with one same miRNA; (c) After intersecting the above-identified lncRNAs, mRNAs, miRNAs, we got the candidates for network construction. The ceRNA regulatory networks were visualized using Cytoscape 3.8.1 software (57).
Cell Lines and Cell Transfection
HepG2 cells, BEL7402 cells, Hep2b, and HuH7 cells were all purchased from ATCC and were cultured in DMEM (Gibco) supplements with 10% fetal bovine serum (Hyclone), 100 U/ml of penicillin (Gibco), and 0.1 mg/ml of streptomycin (Gibco) at 37°C in a humidified atmosphere of 95% O2, 5% CO2.
The overexpression plasmids and siRNAs of DNAJC27-AS1, PPIF, mimic and inhibitor of miR508-3p were provided by GenePharma Co., Ltd. (Shanghai, China). Overexpression plasmid vector is a pEX-3 vector, siRNA sequence information of DNAJC27-AS1 and PPIF are presented in Supplementary Table S2. Briefly, cells were transfected with various siRNAs in MEM medium with 90 nM of each siRNA duplex or were transfected with 0.8 ug different plasmids using Lipofectamine 2000 transfection reagent (Thermo Fisher Scientific, Inc.) following the manufacturer's protocol. After transfection for 48 h, cells were harvested for other analyses.
Main reagents and antibodies
Cell Counting Kit-8 (CCK-8) (cat. no. CK04) was purchased from Dojindo Laboratories (Kumamoto, Japan). GSH ELISA kit (cat. no. CEA294Ge) and TP53 antibody (cat. no. PAA928Ra01) were purchased from Cloud-Clone Corp. (Wuhan, China). Cellular ROS Assay Kit (cat. no. ab113851), SLC7A11 antibody (cat. no. ab175186), GPX4 antibody (cat. no. ab40993), and GAPDH antibody (cat. no. ab8245) were purchased from Abcam (Cambridge, MA, USA). Dual Luciferase Reporter Gene Assay Kit (cat. no. KGAF040) was purchased from KeyGEN BioTECH Co., Ltd (Jiangsu, China). The JC-1 assay kit (cat. no. M8650) was obtained from Solarbio Life Science Co., Ltd (Beijing, China).
Fluorescence Quantitative Real-Time PCR
Quantitative RT-PCR was conducted as per the manufacturer’s instructions. In brief, total RNA was isolated using the TRIpure Reagent (RP1001, Bioteke Co., Ltd., Beijing, China). Reverse transcription was performed on 1 µg of RNA at 60°C for 35 min using a BeyoRT II M-MLV (D7160L, Beyotime, Shanghai, China). After reverse transcription, the cDNAs were used for semi-quantitative PCR using 2×Taq PCR MasterMix (PC1150, Solarbio) and SYBR Green (SY1020, Solarbio). Amplification was carried out in the qPCR instrument (LightCycler®480 II, Roche). 25 μl of reaction mixture contained 12.5 μl of SYBRGreen mastermix (SY1020, Solarbio), the appropriate primer concentration, and 1μl of cDNA. The amplification program included the initial denaturation step at 95°C for 10 min, 40 cycles of denaturation at 95°C for 10 s, and annealing and extension at 60°C for 1min. Fluorescence was measured at the end of each extension step. After amplification, melting curves were acquired and used to determine the specificity of the PCR products. 2-△△CT method was used in the data process. The sets of specific primers are presented in Supplementary Table S3.
Luciferase reporter assay
The bioinformatics tool (Targetscan) was used to predict the miR-23b-3p binding sites of PPIF. Bibiserv was used to predict the miR-23b-3p binding sites of DNAJC27−AS1. According to the above prediction results, the wild-type (wt) binding site sequence was amplified by PCR and then ligated into a dual-luciferase reporter vector, pmirGLO, and site-directed mutation (mut) was made at the possible target site of the target gene. The miR-23b-3p mimics and the above recombinant plasmids were co-transfected into 293T cells in the logarithmic growth phase. The first experiment group is as follows: pmirGLO-wtDNAJC27-AS1+NC mimics, pmirGLO-mutDNAJC27-AS1+NC mimics, pmirGLO-wtDNAJC27-AS1+miR-23b-3p mimics, pmirGLO-mut DNAJC27-AS1+ miR-23b-3p mimics. The second group of experiments is grouped as follows: pmirGLO-PPIF-wtUTR+NC mimics, pmirGLO-PPIF-mutUTR+NC mimics, pmirGLO-PPIF-wtUTR+miR-23b-3p mimics, pmirGLO-PPIF-mutUTR+miR-23b-3p mimics. Forty-eight hours after transfection, the luciferase activity was measured with a multifunction microplate reader (M200Pro, Tecan, Switzerland). The first group of each experiment was set as the control.
Western blotting
HepG2 and BEL7402 cells were harvested and washed with PBS after drug treatment or genetic manipulation. Total cellular protein was isolated using the protein extraction buffer (containing 150 mM NaCl, 10 mM Tris (pH 7.2), 5 mM EDTA, 0.1% Triton X-100, 5% glycerol, and 2% SDS). Equal amounts of proteins (50 μg/lane) were fractionated using 10% SDS–PAGE and transferred to PVDF membranes. The membranes were incubated with primary antibodies. After washing with PBS, the membranes were incubated with corresponding peroxidase-conjugated goat anti-mouse or anti-rabbit secondary antibody, followed by enhanced chemiluminescence staining through the enhanced chemiluminescence system. GAPDH was used to normalize protein loading.
ROS and mitochondrial membrane potential detection by flow cytometry
2'-7'-dichlorodihydrofluorescein-diacetate (DCFHDA) penetrates cells and becomes hydrolyzed to non-fluorescent dichlorodihydrofluorescein (DCFH). DCFH reacts with ROS to form the highly fluorescent dichlorofluorescein (DCF), which flow cytometry can measure. Briefly, HepG2 and BEL7402 cells were treated and collected at least 1.5×104 per assayed condition. Staining cells in culture media with 20 µM DCFHDA and incubating for 30 minutes at 37°C. Washing cells with 1× Buffer after incubation. Then DCF fluorescence distribution of at least 10000 cells was detected by a Novocyte Flow Cytometer (ACEA, Bioscience, USA). Establish forward and side scatter gates to exclude debris and cellular aggregates from the analysis. DCF is excited by the 488 nm laser and detected at 535 nm.
The JC-1 kit determined the mitochondrial membrane potential. The working solution was prepared as instruction. Firstly, 1ml JC-1 working solution was used to suspend the cells and incubated at 5% CO2 and 37 ℃ for 20 min. After washing twice with incubation buffer, the cells were resuspended with 1 ml incubation buffer; then, flow cytometry analysis was conducted. Excitation was set at 488 nm.
Cell viability and GSH assay
The cells were seeded into 96-well microtiter plates at 3,000 cells/well density and allowed to adhere for 12 h. Cells were then treated differently according to the grouping, then exposed to CCK-8 (10 µl/well) for 2 h at 37°C. Absorbance was measured at 450 nm on a Tecan Sunrise microplate reader (Tecan Group AG, Männedorf, Switzerland). For GSH detection, cells (1×106 cells/well) were seeded in 6-well plates and then treated differently. Following treatment, cells were harvested by centrifugation and lysed on ice for 30 min using lysis buffer (50 mmol Tris-HCl, 1.0 mmol/l EDTA, 150 mmol/l NaCl, and 0.1% SDS) and centrifuged once more at 5,000×g for 10 min at 4°C. The supernatant was used for the GSH assays as per the manufacturer's instructions. Briefly, add 50µL samples to each well. And then add 50µL prepared Detection Reagent A immediately. Incubate for 1 hour at 37℃. Add 100µL prepared Detection Reagent B after aspiration and wash. Incubate for 30 minutes at 37℃. Add 90µL Substrate Solution after aspiration and wash. Incubate for 15 minutes at 37℃. Add 50µL Stop Solution and read at 450 nm immediately.
Statistical analysis
Data are presented as the mean±standard deviation (SD) of at least three independent experiments. The Student's t-test for comparison of two groups or one-way analysis of variance (ANOVA) for comparison of multiple groups followed by Tukey's multiple comparison test. For multiple testing, a Bonferroni post hoc test of P-values was made. Statistical analysis was performed using GraphPad Prism 8 (GraphPad, Inc., San Diego, CA, USA). P value less than 0.05 means statistically significant.