Comparative study of active and inactive hepatocarcinogens using a QSAR-based prediction model

Abstract

In this study, we investigated global gene expression in primary rat hepatocytes treated with three active hepatocarcinogens (prednisolone, dehydroepiandrosterone and monocrotaline), three inactive hepatocarcinogens (hydrocortisone, glycyrrhetinic acid and lithocholic acid) and two unclassified chemicals (griseofulvin (possibly active) and prednisone (possibly inactive)). At 48 h after treatment with these eight chemicals, cells were harvested for RNA extraction. Global gene expression analyses were conducted using oligonucleotide microarrays to detect genes whose expression was altered. Differentially expressed genes (DEG) analysis, principal component analysis (PCA), gene set enrichment analysis (GSEA) and KEGG pathway analysis were also conducted. Seven-hundred and sixty genes whose expression was altered >1.2-fold (P<0.05; unpaired Welch’s t-test) were identified as DEGs for the active and inactive carcinogens. In PCA, prednisolone and dehydroepiandrosterone were located away from the inactive carcinogens. In contrast, monocrotaline was close to the inactive carcinogens. Hydrocortisone, glycyrrhetinic acid and lithocholic acid were separate from the active carcinogens. PCA identified griseofulvin as an active carcinogen and prednisone as an inactive carcinogen. GSEA detected several genes associated with hepatocarcinogenesis, such as glutathione S-transferase A2 and NADPH oxidase. KEGG pathway analysis showed that several pathways might be associated with hepatocarcinogenesis. Our results suggest that it may be feasible to differentiate active and inactive hepatocarcinogens by PCA, GSEA and KEGG pathway analysis.