Gene expression profiling of hepatitis B- and hepatitis C-related hepatocellular carcinoma using graphical Gaussian modeling

Background & aims: Gene expression profiling of hepatocellular carcinoma (HCC) and background liver has been studied extensively; however, the relationship between the gene expression profiles of different lesions has not been assessed. Methods: We examined the expression profiles of 34 HCC specimens (17 hepatitis B virus [HBV]-related and 17 hepatitis C virus [HCV]-related) and 71 non-tumor liver specimens (36 chronic hepatitis B [CH-B] and 35 chronic hepatitis C [CH-C]) using an in-house cDNA microarray consisting of liver-predominant genes. Graphical Gaussian modeling (GGM) was applied to elucidate the interactions of gene clusters among the HCC and non-tumor lesions. Results: In CH-B-related HCC, the expression of vascular endothelial growth factor-family signaling and regulation of T cell differentiation, apoptosis, and survival, as well as development-related genes was up-regulated. In CH-C-related HCC, the expression of ectodermal development and cell proliferation, wnt receptor signaling, cell adhesion, and defense response genes was also up-regulated. Many of the metabolism-related genes were down-regulated in both CH-B- and CH-C-related HCC. GGM analysis of the HCC and non-tumor lesions revealed that DNA damage response genes were associated with AP1 signaling in non-tumor lesions, which mediates the expression of many genes in CH-B-related HCC. In contrast, signal transducer and activator of transcription 1 and phosphatase and tensin homolog were associated with early growth response protein 1 signaling in non-tumor lesions, which potentially promotes angiogenesis, fibrogenesis, and tumorigenesis in CH-C-related HCC. Conclusions: Gene expression profiling of HCC and non-tumor lesions revealed the predisposing changes of gene expression in HCC. This approach has potential for the early diagnosis and possible prevention of HCC. (C) 2013 Elsevier Inc. All rights reserved.