TP53 R72P polymorphism modulates DNA methylation in hepatocellular carcinoma

Khadija Rebbani, Agnès Marchio, Sayeh Ezzikouri, Rajaa Afifi, Mostafa Kandil, Olfa Bahri, Henda Triki, Abdellah Essaid El Feydi, Anne Dejean, Soumaya Benjelloun, Pascal Pineau

Hepatocellular carcinoma (HCC) is characterized by widespread epidemiological and molecular heterogeneity. Previous work showed that in the western part of North Africa, a region of low incidence of HCC, mutations are scarce for this tumor type. As epigenetic changes are considered possible surrogates to mutations in human cancers, we decided, thus, to characterize DNA methylation in HCC from North-African patients. A set of 11 loci was investigated in a series of 45 tumor specimens using methylation-specific and combined-bisulfite restriction assay PCR. Results obtained on clinical samples were subsequently validated in liver cancer cell lines. DNA methylation at tumor suppressor loci is significantly higher in samples displaying chromosome instability. More importantly, DNA methylation was significantly higher in Arg/Arg when compared to Pro/Pro genotype carriers at codon 72 rs1042522 of TP53 (65% vs 20% methylated loci, p = 0.0006), a polymorphism already known to affect somatic mutation rate in human carcinomas. In vitro experiments in cell lines indicated that enzymes controlling DNA methylation were differentially regulated by codon 72 Arg or Pro isoforms of p53. Furthermore, the Arg72-carrying version of p53 was shown to re-methylate DNA more rapidly than the pro-harboring isoform. Finally, Pro-carrying cell lines were shown to be significantly more resistant to decitabine treatment (two-fold, p = 0.005). Our data suggest that Arg72Pro polymorphism in a WT p53 context may act as a primary driver of epigenetic changes in HCC. It suggests, in addition, that rs1042522 genotype may predict sensitivity to epigenetic-targeted therapy. This model of liver tumorigenesis that associates low penetrance genetic predisposition to epigenetic changes emerges from a region of low HCC incidence and it may, therefore, apply essentially to population living in similar areas. Surveys on populations submitted to highly mutagenic conditions as perinatally-acquired chronic hepatitis B or aflatoxin B1 exposure remained to be conducted to validate our observations as a general model.