Integrative Analysis Of "OMICS" Data Of Hepatocellular Carcinoma And De Novo Germline Copy Number Variants In Hepatocellular Carcinoma

Part I: Integrative analysis of “OMICS” data of HepatocellularCarcinomaHepatocellular carcinoma (HCC) is one of the most common cancers and the thirdleading cause of cancer death worldwide. As the study of the molecular mechanismsand pathogenesis of HCC is still not comprehensive enough, there is no effectivetreatment. Identification of HCC related genes and revealing their mechanismsprovide an important theoretical foundation of prevention and treatment of HCC.microRNA (miRNA), a kind of endogenous small non-coding RNA with the length of22~24nt, plays a major role in regulating the expression of its target mRNAs, and itis estimated that approximately30%of human protein-coding genes are regulated bymiRNAs. miRNAs play a particularly important role in HCC and other cancers. Toidentify new HCC-related miRNAs, we profiled miRNA expression usinghigh-throughput sequencing in40HCC cases with tumor tissues and adjacentnon-tumor tissues, and totally found1030miNRAs.119differentially expressedmiRNA were identified, of which,64miRNAs were up-regulated while55weredown-regulated. The gene expression profiles of the40paired tissues were performedby Affymetrix microarrays. Then, the correlation-based Gene set enrichment analysis(GSEA) was performed to integrate the miRNA expression profile and mRNAexpression profile. Biclustering was carried out on the GSEA enrichment matrix and10biclusters were identified which contains73enriched miRNA and1809gene sets.After functional annotation, we found that the10biclusters enriched in metabolism,cell cycle, immune response, ECM, proteasome, tRNA aminoacylation, mitochondrialgenes, blood coagulation, E2F targets and SRF targets. Finally, to assess the clusterfunction inferred by bioinformatics method, we experimentally validated twomiRNAs previously reported irrelevant to HCC (miR-1306and miR-454*).Over-expression of miR-1306, derived from cell cycle bicluster, significantly increase the S phase of cell cycle in HepG2. Inhibition of miR-454*, derived from ECMbicluster, significantly inhibit the migration capability of HepG2. Thus, it’s expectedthat systematic functional studies of the miRNAs enriched in each cluster will providea series of new HCC-related miRNAs.In addition to single nucleotide variation (SNV), the copy number alteration (CNA) isanother type of common genomic variation. CNA can influence the expression dosageor the structure of protein-coding gene by changing the gene copy number, therebyaffecting the function of the gene. Lots of studies have been performed on CNA usingcomparative genomic hybridization (CGH) chips or SNP genotyping arrays, but mostof these studies lack of the gene expression profile as a evidence. In order to findHCC related CNA, we used analysis of CNAs by expression data (ACE) methodinferring CNA in the gene expression profiles of40HCC cases with tumor tissues andadjacent non-tumor tissues. Meanwhile, the ACE method was also applied to mRNAexpression profiles of other five independent HCC cohorts previously published.Finally, we found a total of15common significant CNAs in these six expressionprofiles, including eight amplifications:1q21.2-q21.3,1q21.3-q22,1q32.3-q41,1q41-q42.12,1q42.13-q42.2,8q24.13,8q24.3and20q11.21-q11.22etc., and sevendeletions:4q13.2-q13.3,4q21.3-q23,4q25,4q31.3-q32.3,8p21.3-p21.2,12p13.31-p13.2,16q12.2-q13and so on. We found one (8q24.13amplification) out ofthe15CNA was significantly associated with overall survival time (P<0.05) anddisease-free survival time (P<0.05) of HCC patients. Then, we performed highcontent screening experiments of the12genes within8q24.13. Among these genes,ATAD2and WDR67have a significant ability to promote the proliferation andmigration of HCC cell lines, suggesting that these two genes may be the keyHCC-related genes within8q24.13amplification.In conclusion, by using integrated “omics” analysis strategy based on the mRNAexpression profiles and miRNA expression profiles, we found a series of HCC-relatedmiRNA and CNA. Further systematic investigation of these miRNA and CNA willprovide candidate biomarkers of clinical diagnosis and prognosis of HCC, andcandidate drug targets of clinical treatment. Part II: de novo Germline Copy Number Variants in HepatocellularCarcinomaHCC is a complex disease, which accounts for various risk factors resulting fromgenetic and environmental factors interaction. Genome-wide association studies(GWAS) based on single nucleotide polymorphisms (SNP) and copy number variation(CNV) have revealed a series of candidate genes susceptible to HCC, but still acertain heritability could be explained. A large part of the individual suffering fromHCC can not be explained by these known genetic risk factors. Therefore, in additionto these known genetic factors, new genetic variation in gametes period (de novovariation) arouse the attention of people.In order to investigate the effect of de novo CNV on the risk of HCC, Genotyping wasperformed using Illumina Human Omni ZhongHua-8BeadChipin205HCC trios.Then PennCNV method was used for CNV identification and de novo CNV detection.We found that approximately5.4%(11/205) of HCC probands carrying de novoCNVs, which was significantly higher than the rate1.72%in control trios datapreviously reported (OR=3.19, P=0.002). Of the de novo CNVs identified in HCCtrios, only one recurrent (16p13.3duplication) CNV was found in two trios, whichaffects known HCC-related gene RBFOX1. Furthermore, we validated the prevalenceof de novo CNVs in295unrelated HCC cases and1,479non-HCC controls, andfound two additional recurrent CNVs at8p22and11q25, respectively.In conclusion, comparative study of de novo CNVs in HCC trios and control triossuggests that de novo CNVs has a significant effect on the risk of HCC,corresponding to a population attributable risk (PAR) of~4%for HCC in our sampleset, and is a new type of genetic causes of HCC.