Abstract
Background Hepatocellular carcinoma (HCC) is the main form of malignant liver cancer with poor survival. Although some critical driver aberrations were identified in HCC, neither the transcriptomic nor the clinical associations of many of these driver genes have been well-characterized.
Methods We used MutSigCV and OncodriveFM to identify the consensus driver genes across six HCC cohorts with significant functional impacts. We performed mutual exclusivity analysis among these drivers, as well as survival and association analyses between these drivers and physiological/clinical factors. We used generalized linear regression model to assess the impact of driver genes on transcriptome.
Results Using six international HCC cohorts comprising 1,494 samples in total, we obtained a consensus of 11 driver genes including TP53, CTNNB1, ALB, AXIN1, RB1, ARID1A, RPS6KA3, ACVR2A, NFE2L2, CDKN2A and HNF1A. We found that some of these driver genes are significantly associated with gender (TP53, CTNNB1, ALB), grade (TP53, ALB, RB1) and age (CTNNB1, AXIN1, RB1) in multiple cohorts. Moreover, these driver genes are significantly associated with patients’ overall survival. Integrative analysis of driver mutations, copy number variations and transcriptomic data reveals that these driver genes are associated with the majority (63%) of the mRNA transcriptome, but only a small fraction (9%) of miRNAs. Genes associated with TP53, CTNNB1, ARID1A and HNF1A mutations contribute to four most densely connected clusters of biological pathways.
Conclusions In summary, this study reveals the vast impacts of driver gene mutations (genotype) in HCC phenotypes. The consensus drivers may be valuable therapeutic targets of HCC.
Footnotes
Authors’ email addresses: LL: liangqunlu{at}gmail.com KC: kchaudhary{at}cc.hawaii.edu OBP: OPoirion{at}cc.hawaii.edu TC: TChing{at}cc.hawaii.edu