Dysregulated genes in sorafenib-resistant hepatocellular carcinoma
Resistance to sorafenib in patients with hepatocellular carcinoma reduces their prognostic survival time. Therefore, exploring the molecular mechanism of drug resistance is of great significance for the treatment of hepatocellular carcinoma. First of all, Principal Component Analysis (PCA) was performed on samples to detect abnormal samples (Fig. 1a). Then, differential expression analysis was performed to identified differentially expressed genes (DEGs) in sorafenib-resistant samples compared with sorafenib-sensitive samples. A total of 827 significantly DEGs were identified (p<0.05, |logFC|>0.5) (Fig. 1b, Table S1), including 475 up-regulated and 352 down-regulated. It was considered that these DEGs are partly related to sorafenib resistance in HCC. And the top 5 DEGs were further filter by PI value, namely FXYD3, FXYD1, TBX4, NPNT, and ADH1C (Fig. 1c).
Gene function annotation
Gene-set enrichment analysis (GESA) for all genes and GO function and KEGG pathway enrichment analysis for DEGs were performed, respectively. It has been showed that genes in sorafenib-resistant HCC were significantly participated in various signaling pathways associated with HCC and drug metabolism pathways (Fig. 2a), such as PPAR signaling pathway, TGF-β signaling pathway, JAK-STAT signaling pathway, MAPK signaling pathway and drug metabolism cytochrome p450, etc.
The differentially expressed genes were significantly enriched in 1137 biological process entries, 173 cell component entries, 189 molecular function entries, and 28 KEGG pathway entries (Table S2). It was found that these genes were significantly involved in neutrophil associated biological processes, dephosphorylation, regulation of vesicle-mediated transport, etc. (Fig. 2b). The results of KEGG pathway enrichment analysis showed that DEGs mainly participated in MAPK signaling pathway, sphingolipid signaling pathway, phagosome, etc. (Fig. 2c). Further, pathways which enriched by DEGs with PI value > 2 were focused (Fig. 2d), and for subsequent analysis.
Transcription factors significantly regulating dysregulated genes
It is well known that the expression, transcription, and post-transcription of genes in vivo are strictly regulated. The transcription and post-transcriptional regulation of dysregulated genes may be closely related to the occurrence and development of various diseases. Here, TFs significantly regulating the drug resistance-related genes were predicted to help us to further explore the transcriptional regulation mechanism of sorafenib resistance in hepatocellular carcinoma. There were 18 TFs were identified (Fig. 3a) and differentially expressed in sorafenib-resistant HCC. It was observed that NFKB1 regulated the most DEGs, suggesting NFKB1 may affect the resistance of hepatocellular carcinoma to sorafenib and play an essential regulatory role in the process of drug resistance in hepatocellular carcinoma. Further, the strong correlation between TFs and DEGs indicated that the prediction results are more authentic (Fig. 3b). Moreover, these TFs (correlation coefficient >0.85) were identified as critical regulatory factors associated with drug resistance in hepatocellular carcinoma and played an essential role in the whole regulatory network (Fig. 3c).
Effect of crucial disorders on drug resistant hepatocellular carcinoma
To confirm potential prognostic efficacy of critical molecules for drug resistant HCC, ROC analysis and survive analysis were performed on two GEO datasets (GSE73571 and GSE62813) and TCGA dataset of HCC, respectively. ROC curve showed that DEG CDK1, CDKN1A, TAPBP and MYC with high AUC (AUC>=0.8, Fig. 4a), suggesting the correlation between these molecules and disease. Survive analysis also showed the four DEGs tightly associated with HCC survival and prognosis (p<0.05, Fig. 4b).