Bioinformatics Analysis And Medicine Prediction Based On Active Enhancers Developed De Novo In Cirrhosis And Conserved In Hepatocellular Carcinoma

BackgroundHepatocellular carcinoma(HCC)is a serious threat to human health,most hepatocellular carcinomas arise in the background of cirrhosis,but the mechanism of how cirrhosis induces hepatocellular carcinoma is less clear yet.Aberrant epigenetic changes in cirrhosis provide a conducive environment for HCC tumorigenesis.Active enhancers are essential for epigenetic regulation and play an important role in cell development and the progression of many diseases by up-regulating their target genes.Active enhancers exist abundantly in cirrhosis and HCC,however,the role of active enhancers in the progression from cirrhosis to HCC remains unclear.In this study,we used bioinformatics analysis,combined with multi-omics and multiple public data to identified the landscape of active enhancers that developed de novo in cirrhosis and were conserved in HCC,referred to as CL-HCC active enhancers.We investigated the mechanism of how these active enhancers induced HCC,their persistent effects on HCC patients,and the potential worth in diagnosis,prognosis,and treatment of HCC.Methods and results1.Identification of CL-HCC active enhancers and exploration of their potential mechanism on hepatocarcinogenesis and the molecular subtype of HCCTo investigate the role of CL-HCC active enhancers,we first identified 620 active enhancers and their associated 483 target genes that were developed de novo in cirrhosis and were conserved in HCC based on Chip-seq data of histone modifications H3K27ac and H3K4me1 and gene expression profile data from GSE112221.Then we analyzed the variation of H3K27me3,H3K4me3,DNA methylation,5hmc methylation,and mutations in CL-HCC activation enhancers target genes from normal liver to cirrhosis and HCC.The mutation rate of these genes in HCC was low.By comparing the active enhancers of Hep G2 cell,the CL-HCC active enhancers were classified into hepatocyte-intrinsic CL-HCC active enhancers and tumor microenvironment-associated active enhancers.Through enrichment analysis(including GO,KEGG,disease,and hallmarks of cancer),MCPcounter(based on mark gene)and EPIC(based on the inverse convolution algorithm)analysis which used to analyze sample immune cell infiltration and correlation analysis,these two types of CL-HCC active enhancers were found to contribute to the development of cirrhosis to HCC by promoting hepatocellular carcinogenesis and generating a tumor microenvironment with inflammatory and CD8~+T cell function suppression,respectively.Based on the expression of CL-HCC active enhancers'target genes in TCGA-LIHC patients and NMF,TCGA-LIHC patients were classified into three different molecular subtypes.Survival analysis showed the prognostic differences between molecular subtypes,and analyzing of CNV and mutation profiles showed the different genomic landscapes across the molecular subtypes.GSVA analyzed the KEGG pathway,hypoxic state and T cell dysfunction and MCPcounter calculated the immune cell infiltration,the results revealed differences among molecular subtypes in metabolism and tumor microenvironment.TIDE was used to predict patients'immunotherapy response based on gene expression profiles and p RRophetic to calculate sensitivity to classic chemotherapy medicines and three classic molecular targeted medicines based on GDSC,the results revealed differences in immunotherapy responsiveness and medicine sensitivity among molecular subtypes.2.Establishment and verification of CL-HCC active enhancers target genes relate diagnostic model and prognostic signature for HCCTCGA-LIHC samples were divided into a training group and a testing group.Three machine learning algorithms,including logistic regression,random forest,and support vector machine,were combined to build a diagnostic model on the training group and the AUC value was used to evaluate their diagnostic ability.we found that the expression of CL-HCC active enhancers'target genes could accurately classify normal and HCC tissues(RF AUC=0.883,SVM AUC=0.930,LR AUC=0.951),and the diagnostic model including 5 genes(THBS4,OLFML2B,CDKN3,GABRE,and HDAC11)constructed according to the LR algorithm was the best.The model was utilized to examine the test set,external independent HCC datasets and TCGA pan-cancer dataset.The model could predict the occurrence of HCC.The accuracy of the model was above 0.9 in both the test group and external independent liver cancer data,and above 0.7 for the tumor data set of 15 TCGA datasets.One-way Cox regression of CL-HCC active enhancers'target genes revealed that most(47.6%)CL-HCC active enhancer target genes could be used as independent prognostic factors for HCC and most of them were risk factors for HCC patients'prognosis.Then LASSO regression with ten-fold cross-validation descending obtained 12 optimal prognostic genes.After dividing TCGA-LIHC into training and test groups,a 4-gene prognostic signature was established from the training group by using stepwise regression multifactorial COX analysis for 12 optimal prognostic genes.The risk score of each sample in the training set was calculated according to the expression of the five AR lnc RNAs by using the following formula:risk score=C5orf30 expression*(-0.12684)+KITLG expression*0.20732+SPP1expression*0.08163+UBE2S expression*0.52305,and the median risk score of the TCGA-LIHC training group was used as the cut-off point for discriminating high-risk and low-risk.Then the prognostic signature was validated in the TCGA-LIHC test group,all TCGA-LIHC patients and the ICGC LIRI-JP dataset,high and low-risk groups divided by this signature showed significant differences in survival(OS p=0.0009 in the TCGA-LIHC test group,OS p<0.0001 in all TCGA-LIHC patients and OS p<0.0001 in ICGC LIRI-JP).Time-dependent ROC curves and C-index were used to compare the sensitivity and specificity of 4-gene prognostic signature with other recently published signatures.The model had better predictive power and stability across different data sets when compared with other recently published signatures.GSEA was used to analyze the differences of CL-HCC active enhancers'target genes expression and KEGG pathway,and it found that the expression of CL-HCC active enhancers'target genes was higher in high-risk patients than in low-risk group,and tumor-related pathways and inflammatory pathways were activated in high-risk group patients.Comparing the genomic landscape of high and low-risk groups,it was found that there was no difference in gene mutation frequency between high and low-risk groups except TP53.The model was found to serve as an independent prognostic factor by univariate and multifactorial COX analyses,and the nomogram was constructed by combing signature and the TNM stage.Calibration curves showed that the 1-year and 3-year predictive values of the nomogram in both data were almost the same as those in the observation,and DCA analysis showed that the net benefit rate of the nomogram was higher than that of the TNM stage in both data.3.Medicine prediction using CL-HCC active enhancers and their target genes as targets and prediction of therapeutic agents for patients in high-risk groups identified from 4-gene prognosis model delineationUsing GSEA analysis GSE51143,the BET inhibitor JQ1 was found to significantly down-regulated the expression of hepatocyte-intrinsic CL-HCC active enhancers'target genes in Hep G2 cells.15 medicines with different mechanisms that may down-regulated the expression of CL-HCC active enhancers'target genes were predicted by the CMap,PPI network analysis showed metformin's target INS interacted with a variety of CL-HCC active enhancer target genes which involved in the tumor pathway.Analysis of GSE131175,we found the expression of CL-HCC active enhancers'target genes was significantly decreased after metformin intervention in CCl4 treated transgenic mouse with the background of cirrhosisUsing the CTRP and PRISM databases combined with p RRophetic to calculate medicine sensitivity in patients from TCGA-LIHC and ICGC LIRI-JP.five compounds from CTRP and four compounds from PRISM were found to be significantly more sensitive in the high-risk patients than the low-risk patients,and the sensitivity was correlated with risk scores.Under the CMap database and literature research,it was found that patients in the high-risk group were more sensitive to clofarabine and BI-2536,which may be effective medicines for high-risk patients.Conclusions and implications1.CL-HCC active enhancer has an important role in the development of HCC:(1)CL-HCC active enhancer was identified for the first time and found to be the main cause of abnormal expression of its target genes.CL-HCC active enhancer can lead to the development of cirrhosis to hepatocellular carcinoma by affecting hepatocytes and the liver microenvironment.(2)The differential expression of CL-HCC active enhancer target genes can be classified into different molecular subtypes,and there are great differences among these molecular subtypes,indicating that CL-HCC active enhancer target gene expression has an important role,which provides a new idea to explore heterogeneity and molecular typing of liver cancer.2.CL-HCC active enhancers have important significance and application prospects for the diagnosis and prognosis of liver cancer:(1)The expression of CL-HCC active enhancer target genes can be used as potential biological markers for liver cancer diagnosis,and the5-gene diagnostic model has the potential to diagnose liver cancer,which provides ideas for studying and finding potential biological markers for liver cancer diagnosis.(2)The prognostic model of liver cancer based on the expression levels of 4 CL-HCC active enhancer target genes can effectively predict the prognosis of liver cancer patients,and it is more accurate and stable than the three recent models,and the nomogram constructed in combination with TNM staging has important significance and application prospects for clinical decision-making.3.CL-HCC active enhancers can be used as potential medicine targets to discover anti-liver cancer or anti-cirrhosis medicines:(1)Both JQ1 and metformin can inhibit the expression of CL-HCC active enhancer target genes,which may be a potential mechanism of JQ1 and metformin anti-tumor,providing theoretical support for developing medicines with CL-HCC active enhancers and their target genes as therapeutic targets.(2)Targeting high-risk patients in risk models,our study provides potential therapeutic agents for them(clofarabine and BI-2536),which could potentially improve their prognosis.Overall,this study provides new insights into the development of HCC through CL-HCC active enhancers,which can help in the diagnosis,prognosis prediction,precision treatment,and medicine development of hepatocellular carcinoma.