Tumour-suppressor genes act as safeguards that prevent normal cells from forming tumours. But how are they taken out of action during cancer development? Initial findings focused on genetic mutations or deletions. More recent research, however, has unveiled a role for epigenetic events in the inactivation of tumour-suppressor genes. A paper in Nature now sheds further light on the epigenetic mechanisms that operate in human cancer.

An important factor in the control of gene expression is DNA methylation. Several genes that normally suppress tumour development are switched off in human tumours and this is associated with methylation of their promoter regions, but relatively little is known about the mechanisms by which these genes become methylated. The enzymes that catalyse DNA methylation are DNA methyltransferases (DNMTs). The prototype of this enzyme family is Dnmt1, which accounts for most DNA methylation in mouse cells. Surprisingly, however, human cells that lack DNMT1 maintain a significant level of overall methylation and gene silencing.

Ina Rhee and colleagues therefore set out to investigate whether other DNMTs might have a more prominent role, and inactivated DNMT3b in a human colorectal carcinoma cell line. As was the case for disruption of DNMT1, the overall levels of genomic DNA methylation dropped only by a small percentage. By contrast, the simultaneous deletion of DNMT1 and DNMT3b resulted in an almost complete elimination of cellular DNA methyltransferase activity and genomic methylation. Every individual gene locus analysed showed a substantial loss of methylation and increased levels of expression, including several that have been implicated in tumour progression. Furthermore, the proliferation of cells that lack DNMT1 and DNMT3b was markedly reduced. These results clearly establish that methylation is required to shut off tumour-suppressor genes in human cancer cells, and show that the activity of more than one DNA methyltransferase is required to keep tumour suppressors in check if a cell is to escape normal growth control.

Future research will undoubtedly investigate how specific methylation events are controlled in normal cells and during tumour development. Understanding these mechanisms will have important therapeutic implications: changes in DNA methylation are more readily reversible than mutations or deletions, so a deeper knowledge of how DNA methylation goes awry in cancer might pave the way to restoring the function of tumour-suppressor genes in human cancer.