Hepatoblastoma (HB) is a rare embryonal liver tumor of children. Although intrinsic biological differences between tumors can affect prognosis, few groups have studied these differences. Given the recent increased attention to epigenetic mechanisms in the genesis and progression of these tumors, we aimed to classify HB samples according to the stages of liver development and DNA methylation machinery.

We evaluated the expression of 24 genes associated with DNA methylation and stages of hepatocyte differentiation and global DNA methylation. Using bioinformatics tools and expression data, we propose a stratification model for HB.

Tumors clustered into three groups that presented specific gene expression profiles of the panel of DNA methylation enzymes and hepatocyte differentiation markers. In addition to reinforcing these embryonal tumors’ molecular heterogeneity, we propose that a panel of 13 genes can stratify HBs (TET1, TET2, TET3, DNMT1, DNMT3A, UHRF1, ALB, CYP3A4, TDO2, UGT1A1, AFP, HNF4A, and FOXA2). DNA methylation machinery participates in the characterization of HBs, directly reflected in diverse DNA methylation content. The data suggested that a subset of HBs were similar to differentiated livers, with upregulation of mature hepatocyte markers, decreased expression of DNA methylation enzymes, and higher global methylation levels; these findings might predict worse outcomes.

HBs are heterogeneous tumors. Despite using a small cohort of 21 HB samples, our findings reinforce that DNA methylation is a robust biomarker for this tumor type.