We studied the pattern of aberrant methylation during the multistage pathog
enesis of cervical cancers. We analyzed a total of 73 patient samples and 1
0 cervical cancer cell lines. In addition, tissue samples [peripheral blood
lymphocytes (n = 10) and buccal epithelial cells (n = 12)] were obtained f
rom 22 healthy volunteers, On the basis of the results of preliminary analy
sis, the cervical samples were grouped into three categories: (a) nondyspla
sia/low-grade cervical intra-epithelial neoplasia (CIN; n = 37); (b) high-g
rade CIN (n = 17); and (c) invasive cancer (n = 19), The methylation status
of six genes was determined (pld, RAR beta, FHIT, GSTP1, MGMT, and hMLH1),
Our main findings are as follows: (a) methylation was completely absent in
control tissues; (b) the frequencies of methylation for all of the genes e
xcept hMLH1 were >20% in cervical cancers; (c) aberrant methylation commenc
ed early during multistage pathogenesis and methylation of at least one gen
e was noted in 30% of the nondysplasia/low-grade CIN group; (di an increasi
ng trend for methylation was seen with increasing pathological change; (e)
methylation of RAR beta and GSTP1 were early events, p16 and MGMT methylati
on were intermediate events, and FHIT methylation was a late, tumor-associa
ted event; and V) methylation occurred independently of other risk factors
including papillomavirus infection, smoking history, or hormone use. Althou
gh our findings need to be extended to a larger series, they suggest that t
he pattern of aberrant methylation in women with or without dysplasia may h
elp identify subgroups at increased risk far histological progression or ca
ncer development.