“…There have been several recent reviews of mammalian DNA methylation (Holliday, 1979;Riggs & Jones, 1983;Nyce et al,1983;Doefler, 1983;Razin et al, 1984;Jones, 1986), and considering all data, it appears that methylation of DNA cytosine is one among several mechanisms significant in the control of mammalian gene expression (Riggs & Jones, 1983;Doefler, 1983;Razin et al, 1984), perhaps functioning primarily as a locking mechanism for the stable maintenance of the transcriptionally silent state (Razin & Riggs, 1980). There has been speculation that DNA methylation changes may play some role in tumour initiation or progression (Holliday, 1979;Riggs & Jones, 1983;Nyce et al, 1983;Jones, 1986), and some experimental data are now available. Evidence that DNA methylation changes may be associated with tumorigenesis is as follows: (a) The DNA of tumour cells has often been found to be undermethylated compared to the DNA of normal cells (although exceptions have been found) (Lapeyre et al, 1981;Gama-Sosa et al, 1983;Goelz et al, 1985); the methylation patterns of some genes that are expressed in tumour cells have been found to be altered in association with increased gene expression (Feinberg & Vogelstein, 1983a, b;Cheah et al, 1984); (b) the DNA of tumour metastases is less-methylated than the DNA of primary tumours (Gama-Sosa et al, 1983); (c) the DNA of primary malignant tumours is less-methylated than the DNA of benign tumours (Gama-Sosa et al, 1983); transformed cells have been shown to have their metastatic properties altered by the potent inhibitor of DNA methylation, 5-azacytidine (Olsson & Forchammer, 1984;Trainer et al, 1985;Ormerod et al, 1986); (e) several carcinogens have been shown to cause demethylation (Boehm & Drahovsky., 1979Salas et al, 1979;Wilson & Jones, 1983); (f) 5-azacytidine, an inhibitor of DNA methylase, has been shown to be a carcinogen in some experimental animal studies …”