Data collection. This study intended to preliminarily explore the regulatory mechanisms of ceRNA during liver fibrosis progression by analyzing the gene expression differences and co-expression analysis between the healthy group and a liver fibrosis group. GEO database (https://www.ncbi.nlm.nih.gov/geo/), a public international and functional genomics data repository, is used for high-throughput microarray and next-generation sequences38. Therefore, GSE12392 and GSE84044 were mined for bioinformatics analysis by searching the keywords "liver fibrosis" and "Homo sapiens" in the GEO database. In the dataset, lncRNA and mRNA expression in six healthy individuals and six patients were measured using the Agilent MicroArray V4 platform.
DEGs analysis. The DEGs were located using edgeR39 in the R Bioconductor package. The expression data in GSE12392 were analyzed using the LIMMA package of the R language, and the probe ID was converted into GeneSymbol according to the CORRESPONDING GPL file of the probe. The mRNA or lncRNA were selected from the probe for differential expression and gene co-expression network analysis. The criteria for selecting DEGs were P < 0.05 and log2 Fold-Change less than -1 or greater than 1.
WGCNA analysis. Gene co-expression network analysis was used to analyze gene expression in the two groups at the mRNA and lncRNA levels in combination with basic clinical information of the samples (i.e., age, gender, and whether liver fibrosis occurred) aiming to explore gene co-expression modules related to disease occurrence.
LncRNAs target gene prediction. The results of co-expression analysis and differential expression analysis were combined to obtain common lncRNAs, that is, to identify DEGs and co-expressed genes associated with clinicopathological features. StarBase software (http://starbase.sysu.edu.cn/) and bioinformatics were used to predict the target genes of lncRNAs.
MiRNA target gene prediction. StarBase software (http://starbase.sysu.edu.cn/) and bioinformatics methods were used to predict the target genes of human miRNAs, and the default parameters of the official database were adopted.
Establishment of the ceRNA regulatory network. An lncRNA-miRNA-mRNA network was constructed according to ceRNA theory40. Based on the prediction results of these target genes, a ceRNA regulatory network of lncRNA-miRNA-mRNA was constructed by combining mRNA co-expression network analysis (i.e., genes co-expressed between mRNA level and liver fibrosis) and mRNA differential expression analysis (i.e., genes with differential expression). We used Cytoscape 3.7.241 to visualize the lncRNA-miRNA-mRNA network, and the CytoHubba plugin in Cytoscape 3.7.242 was used to calculate all node degrees.
GO and pathway enrichment analysis. GO analysis43 included biological process, cellular component, and molecular function, generated using the bioinformatics tool DAVID44 (P < 0.05). This analysis was used to examine the unique biological significance of the high-throughput transcriptome. KEGG enrichment analysis was used to predict the biological pathways45. The results were visualized using the R language.
Analysis of the PPI network. The PPI information for the common DEG network was evaluated using the search tool STRING (https://string-db.org/)46 . Any potential correlations between these DEGs were analyzed using Cytoscape47. The PPI network modules were recognized using the Cytoscape plugin MCODE, and only those were presented based on a node degree ≥3.
Hub gene selection. We analyzed the PPI network using the Cyto-Hubba plugin of Cytoscape 3.7.2 and selected candidate hub genes with top node degrees. Subsequently, hub genes associated with liver fibrosis were identified using GSE84044.
Human samples. From the Xiangya Hospital of Central South University, ten healthy subjects and 16 patients with liver fibrosis were recruited. Informed consent was obtained, and ethical approval was obtained from the Ethics Committee of Xiangya Hospital. Human liver tissues were immediately frozen in liquid nitrogen and stored at − 80 °C.
Cell culture experiments. LX-2 cells were cultured in DMEM containing 1% fetal bovine serum and 1% penicillin/streptomycin. LX-2 cells were passaged using trypsin. LX-2 cells were stimulated by TGF-β1 (0, 5, 10, 20, 40 ng/ml; Sigma, Cat No. SAB4502958) for 48 h. All cells were stored in an incubator at 37 °C in 5% CO2.
QRT-PCR. Total RNA was isolated from liver fibrosis human samples and LX-2 cell lines using TRIzol reagent (Invitrogen, CA, USA). The cDNAs were synthesized using a commercial kit (Bio-Rad, Hercules, CA). Gene expression was measured using the ABI 7900HT Fast Real-Time PCR System. Relative mRNA expression levels were calculated using the 2−ΔΔCT method. The following primer sequences were used: lncRNA PCBP1-AS1, forward, 5'-ACTACTCAGTCAATTGCTCCA-3', reverse, 5'-ATTTCCTTACTGACCTGCAT-3'; lncRNA AC100861, forward, 5'-GCACATGAC
ACGGGATGAGA-3', reverse, 5'-GGCTTTCGG GAGGCT GATTA -3'; lncRNA TEX41,forward, 5'-TGGCCAAGAGACAACACCAA-3', reverse,5'-GGCAGAGTGA
GTCCAAAGG-3'; GAPDH, forward, 5'-TGGAAATCCCATCACCATCT-3', reverse, 5'-TGGACTCCACGAC GTA C TC A -3'.
Statistical analysis. Data obtained from three independent experiments were expressed as mean ± standard deviation. The t-test was performed to compare differences between two groups, and one-way analysis of variance was used for multi-group comparisons. Differences with p < 0.05 were considered statistically significant.
The statement. All methods are carried out in accordance with relevant guidelines and regulations.