Sydney researchers uncover cancer's silent culprit

Two Sydney cancer patients have made medical history this week as the first individuals to be diagnosed with cancer induced by spontaneous silencing of an otherwise normal tumour-suppressor gene.

The discovery, described in the April issue of Nature Genetics, points to the likelihood that a small percentage of sporadic and inherited cancers are triggered by germline epimutations -- the silencing of normal genes by hypermethylation.

It confirms a theory advanced 17 years ago by eminent CSIRO molecular geneticist Dr Robin Holliday, that methylation-induced silencing of key genes that regulate cell growth and division regulation could trigger cancer.

The study's lead investigator, Dr Catherine Suter, of the Garvan Medical Research Institute, says the team's discovery of a new, generic mechanism in cancer has much broader implications.

Because all 30,000-odd genes in the human genome are potentially susceptible to misprogramming by the same mechanism, hypermethylation may explain enigmatic cases in which individuals exhibit the classical symptoms of inherited, single-gene diseases, in the absence of any point mutation or deletion in the relevant gene.

Suter and her colleagues -- Prof Robyn Ward, also at the Garvan, and Dr David King of the Victor Chang Cardiac Research Institute -- now suspect that methylation-induced gene silencing underlies some metabolic disorders previously thought to involve multiple genes, and may even be involved mental diseases like schizophrenia and bipolar disorder.

In the past decade, research has identified methylation as a major player in the elaborate regulatory machinery of gene expression in higher organisms, including humans.

In animals, it is involved in the sex-specific regulation of certain genes in particular organs and tissues, depending on whether the genes are inherited via the male or female parent -- in defiance of Mendel's laws of inheritance.

Plant discovery

Plant geneticists have discovered examples of epimutation-induced gene silencing in Arabidopsis, toadflax and maize. The changes are stable for the life of the plant, and potentially heritable, but can be reversed by chemical agents that strip methyl molecules from DNA, restoring the silenced genes to normal function.

Suter and her colleagues suspected the same mechanism might be involved in pathological changes in normal gene expression in mammals, including humans.

They intensively screened 94 patients with colorectal cancers, 65 of them with a family history of cancer, to winnow out those with mutations classically associated with their particular forms of cancer.

They ended up with two individuals -- a man (TT) and a woman (VT) -- who had developed non-polyposis colorectal cancer (NPCC), in the absence of any mutations in three tumour-suppressor genes -- MSH2, MLH1 and APC -- known to be associated with NPCC. Their medical histories were uncannily similar: both had been treated successfully for six tumours over 20 years, and developed their first colorectal cancers at age 43 (TT) and 45 (VT).

Both had developed tumours of the colorectum and small bowel, while the woman had also developed a breast tumour and an endometrial cancer in the lining of the uterus -- the same tissues affected in cancers with classical, mutation-induced non-polyposis colorectal cancer. The woman had also been treated for a melanoma skin cancer.

Using a special DNA-sequencing technique called COBRA, which can detect methylated genes, the Sydney researchers showed that both patients had a mutation in one allele of the MLH1 gene. The gene encodes a DNA-repair enzyme that normally detects and corrects mutations in newly replicated DNA after cells divide.

In both patients, the epimutation was present in all tissues tested: blood cells, hair-follicle cells and cells from the cheek lining. These tissues represent the three primordial cell lineages that form all the body's tissues and organs during embryogenesis.

Suter says this finding indicates that the epimutation had been inherited from a parent, or had occurred spontaneously before the fertilised egg began to divide.

Any epimutation - or mutation - that occurs after a newly fertilised egg begins to divide will be expressed in a mosaic pattern in the body, with only those organs or tissues arising from the original, mutant cell being affected.

The result is a 'phenocopy' of the disease - the silenced gene produces symptoms indistinguishable from those associated with a typical point mutation or deletion in the same gene.

Just as with classical inherited cancers, the individual is born with one of their two copies of the gene inactivated in every cell - putting them at greatly increased risk of developing cancer in childhood, or before middle age - any cell that loses its 'backup' allele to mutation can turn cancerous.

Time question

Suter and her colleagues were unable to determine whether, in the two Sydney patients, the epimutation was inherited, or occurred as a germline event before the newly fertilised egg began to divide.

In both cases, the patients' mothers both died of cancer -- colorectal cancer, in TT's case, and ovarian cancer, in VT's case. This raised the possibility that epimutation was inherited, and both individuals could have transmitted it their own children.

Because the mothers of both patients had both died overseas, and VT's mother had had her cancerous ovaries removed, the Sydney team could not obtain tissue samples to explore this possibility.

However, tests on some of the patients' own children failed to detect any methylation abnormality in the MLH1 gene. And when they tested TT's sperm, they found that only 1 per cent of the sperm cells carried the MLH1 epimutation.

"It's possible the epimutation is cleared during sperm formation, or that, very early in embryogenesis, normal MLH1 gene function was restored in the cell lineage that gave rise to his germ cells," Suter says.

Suter's team now plans to look for further cases that will allow them to do pedigree studies, to determine if epimutations can be inherited.

How do genes become accidentally methylated? Suter says one possibility is that genes like MLH1 may lie in chromosomal regions populated by 'junk' DNA, including transposons -- mobile genetic elements that are suppressed by methylation to prevent them activating and disrupting gene function.

She says their discovery is likely to lead to researchers in other fields investigating the possible role of epimutation in inherited disorders -- or susceptibilities -- where there is no sign of causal mutations in the relevant genes.

Suter says the mosaic, tissue-specific pattern of epimutations that occur after the one-cell stage of embryogenesis could also throw new light on disorders that have been assumed to be genetically complex.

Rather than being caused by multiple, interacting genes, some polygenic disorders could actually result from epimutations in single genes, that result in abnormal gene expression in some tissues or organs, but leave others functioning normally.