Methylation tests herald new cancer diagnostics and treatments

Eighteen years ago CSIRO biochemist Dr Robin Holliday first advanced the radical idea that that perfectly normal genes inactivated by hypermethylation played a role in the onset of cancer. Today, specific, sensitive tests for methylation are being developed which could play a role in early detection of cancer.

Methylation is known to mimic the effects of mutation; methylated and mutant genes conspire to derange cell growth and replication in many cancers.

At the recent Lorne Cancer Conference on Phillip Island, American breast-cancer researcher Professor Nancy Davidson, of the Johns Hopkins University Oncology Centre, in Baltimore, Maryland, said her team's studies of breast ductal cells had shown abnormal patterns of methylation in apparently healthy women with no physical sign of breast cancer.

Testing for abnormal patterns of methylation-silenced genes in tissue samples could save lives by providing early warning of cancer.

Davidson said the discovery had important clinical implications for early detection and risk assessment. "It might tell you who is going to get breast cancer," she said.

While there was little that oncologists could do to correct mutations in mutant genes like BRCA1 or BRCA2, de-methylating agents like azocytidine could prevent or delay cancer by restoring function to otherwise normal genes silenced by hypermethylation.

Davidson's team uses a very fine catheter, inserted into the nipple, to sample lavage fluid containing cells shed from the mammary epithelium.

In a comparative study, they found that 96 per cent of ductal cells from women with invasive breast cancer had hypermethylated genes - the main hypermethylated genes were Cyclin D2, TWIST and RAR-B. In contrast, these genes were unmethylated in ductal cells from normal breasts.

"We extended our search to other genes, and looked at a wider range of different breast cancers," she said. "One or more of these genes appear to be hypermethylated in a large majority of tumours."

In one case, said Davidson, researchers had detected an occult breast cancer in an apparently healthy woman on the basis of a particular, abnormally methylated gene.

"We're developing a quantitative, multiplex, methylation-specfic PCR test to detect breast cancer early," she said.

"It is highly specific, very sensitive, and we can detect cells with hypermethylated genes, in the presence of a large excess of normal cells - around 1 in 100,000 cells."

Although the estrogen receptor (ER) gene is over-expressed in many breast cancers, Davidson's team has shown that the gene is silenced by hypermethylation in 100 per cent of breast cancer cell lines that do not express estrogen receptors. The ER gene is not methylated in ER-expressing breast cancer cell lines.

"If you take ER-negative cell lines and treat them for several days with demethylating agents, you see nice re-expression of the ER alpha protein and messenger RNA, but there is no impact on ER-positive cell lines."

Histone deacetylase inhibitors (HDACs) might also be useful in treating breast cancer. When Davidson's team treated the same ER-negative cell lines with two prototype HDACs developed by Novartis, they saw a progressive increase in expression of estrogen-receptor messenger RNAs, but no accompanying change in the methylation status of the ER gene.

"The two mechanisms are highly interrelated, so a synergistic therapy may be more effective," she said.

One of the Novartis HDACs is already being trialled in women undergoing primary breast cancer therapy.

Asked whether such therapies could activate genes that were normally methylated or acetylated, Davidson said her team was checking for adverse side-effects that might be due to inappropriate activation of silenced genes.