Abstract
Hepatocellular carcinoma (HCC) is the third leading cause of cancer death worldwide. The mechanisms leading to development and progression of HCC are complicated and implicated in both genetic and epigenetic regulation. Thanks to the advance of high-throughput technologies, it becomes possible to use such technologies to investigate the genetic and epigenetic alteration in HCC with large amount of data. Some genes, proteins, miRNAs, and DNA methylation have been found important in the development of HCC, but these results are lack of a systematic view to elucidate the progression mechanisms of HCC. Therefore, with abundant data of mRNA profiles, miRNA profiles and methylation profiles from HCC patients, we used a systems biology approach and big database mining to construct genetic and epigenetic networks (GENs) that combine gene regulatory networks (GRNs), protein-protein networks (PPINs) and epigenetic network at different stages of hepatocarcinogenesis. For figuring out the most important core networks in each stage, principal network projection (PNP) is further facilitated to extract the core network markers. By comparing the pathways involved in core network biomarks between neighboring stages of HCC, we investigate how the changes of DNA methylation and aberrant regulations of miRNAs have impacts on the perturbation of ErbB, MAPK, TGF-beta, and JAK-STAT signaling pathways in the hepatocarcinogenesis, through which extracellular signals are further transduced to TFs to aberrantly regulate their target genes, resulting in favorable cellular responses for progression of HCC i.e. cell proliferation, anti-apoptosis, aberrant cell cycle, cell survival, and metastasis. Moreover, based on epigenetic network markers of each progression stage, we select several potential multiple drug targets for multiple drug design to prevent the progression for HCC patients; NTK2, MYC and AKT1 are selected as potential multiple drug targets and lestaurtinib, dinaciclib and perifosine are suggested as corresponding multiple drug molecules in HCC from stage I to stage II. DDIT3, PDGFB and JUN are selected as potential multiple drug targets and celecoxib, axitinib and vinblastine are suggested as corresponding multiple drug molecules in HCC from stage II to stage III. STAT3, IL1B and NFKB1 are selected as potential multiple drug targets and atiprimod, celastrol and bortezomib are suggested as corresponding multiple drug molecules in HCC from stage III to stage IV. The proposed method could not only investigate the progression molecular mechanisms of HCC but also provide potential multiple targets for multiple drug design of HCC.