DNA methyltransferase expression and DNA methylation in human hepatocellular carcinoma and their clinicopathological correlation.

Aberrant DNA methylation on CpG islands is one of the most consistent epigenetic changes in human cancers, and the methylation process is catalyzed by DNA methyltransferase (DNMT). We evaluated i) the mRNA levels of three DNMTs; DNMT1, DNMT3a and DNMT3b, in 25 hepatocellular carcinomas (HCCs), in their corresponding non-cancerous liver tissues and in 7 normal livers by using real-time reverse transcriptase-polymerase chain reaction; ii) nuclear expression of DNMT1 and DNMT3a proteins in the HCCs by immunohistochemistry, iii) the methylation status of 5 genes; p16, p15, E-cadherin, HIC-1 and RASSF1A in the same tissues, and iv) the relationships between the above results and the clinicopathological characteristics, including prognosis. The differences in mRNA expression levels for DNMT1, DNMT3a and DNMT3b were statistically significant between HCC and normal livers (p<0.001), HCC and chronic hepatitis (p<0.001) and HCC and cirrhosis (p<0.001). An increase in mRNA expression levels of >4-fold for DNMT3b in HCCs was significantly associated with a poorer overall survival (p=0.027) and shorter metastasis-free survival (p=0.0299). A poorer recurrence-free survival was noted in HCCs with a >4-fold increase in DNMT3a mRNA (p=0.0120). The average numbers of methylated genes were 0, 1.27, 1.38 and 2.72 for normal livers, chronic hepatitis, cirrhosis and HCCs, respectively, and this progressive increase from normal livers to chronic hepatitis/cirrhosis through HCC may suggest that tumor suppressor gene methylation is an early event in hepatocarcinogenesis. These results first suggest that hepatocarcinogenesis involves an increased expression of DNMT1, DNMT3a and DNMT3b mRNA and a progressive increase in the number of methylated genes from normal liver, chronic hepatitis/cirrhosis to HCC and secondly that an increase in the DNMT3a and DNMT3b mRNA levels in HCCs relative to their non-cancerous tissues may be a predictor of poor survival.