Epigenetic inheritance and colorectal cancer

Lynch syndrome is the common name for hereditary non-polyposis colorectal cancer (HNPCC), in which mutations in two DNA mismatch repair genes, MLH1 and MSH2, strongly predispose carriers to develop colorectal cancer at an early age.

Carriers have up to 80 per cent risk of developing colorectal cancer and a high risk of developing endometrial and ovarian cancers.

Lynch syndrome generally has an autosomal dominant inheritance pattern - everyday garden variety Mendelian inheritance - but some researchers suspect there might be something else going on.

The suspicion is that a germline transmission of an epimutation - a mutation passed through the gametes without a change to the DNA sequence - of the MLH1 gene may be causing some of the cases of cancer susceptibility.

There has been a lot of contention over this suspicion, however. The phenomenon has been observed often in plants and in mice - Professor Emma Whitelaw's work on the agouti viable yellow (Avy) allele, for instance - but never before proven in humans.

A growing number of papers in recent years have reported germline MLH1 epimutations, but whether they could be inherited was still open to debate. Now, a Sydney research team has found evidence of one case of inter-generational epigenetic inheritance in humans and they are getting set to look for more.

Over the last few years, Professor Robyn Ward and colleagues from St Vincent's Hospital in Sydney have studied patients with Lynch syndrome in which there is no conventional mutation in the DNA sequence of MLH1.

Ward and colleagues have demonstrated that hypermethylation of one allele of MLH1 within the germline also caused HNPCC. (Hypermethylation is the aberrant methylation of genes - in this case very dense methylation of the CpG island spanning the promoter of MLH1.)

In these early papers, the researchers postulated that the epimutation could be inherited. That was open to question, however, with some researchers in the field arguing that while inter-generational inheritance was possible, real evidence wasn't actually there.

In a paper published earlier this year in the New England Journal of Medicine, however, Ward's team, with lead author University of NSW geneticist Dr Megan Hitchins, demonstrated a novel pattern of inheritance consistent with inter-generational epigenetic inheritance - in other words, an epimutation passed via the gametes.

However, whether the aberrant methylation itself was transmitted through the gamete, or merely the propensity for this to occur in the next generation, is still in question.

New pattern of inheritance

"We've found an entirely new pattern of inheritance of a familial form of cancer," Hitchins says. "It was always thought that the only way by which familial forms of cancer could be transmitted from a parent to a child was by a spelling mistake in the genetic code of a particular group of genes.

"But what we've found is that the methyl groups that attach to the DNA to form a coat around that particular gene have in fact been passed from a mother to a child, giving him a high risk of developing cancer at some stage during his life."

Ward says the really extraordinary part of the team's findings was that the tagged DNA was passed on from a mother who had cancer to one of her children, a young man who had no evidence of cancer at that time.

"The reason why that is extraordinary is because until now we thought only naked DNA passes through generations, not coated DNA. This describes a new way in which disease can be inherited."

Hitchins says that the methyl tags attached to MLH1 that silence it are normally released during gametogenesis - the development of sperm and eggs. In the case of the family her team has studied, however, while in two children in the family the alleles had returned to normal, in this son they had not.

"Previously we thought that the methyl tags were stripped away from the gene in the sperm and in the egg, so that by the time the sperm and the egg join at conception and the new embryo has formed, the gene has returned back to its normal state of activity," Hitchins says.

"In this particular family, one of the children inherited that piece of DNA with the tags still attached to it. So in one out of three the hypermethylation has been inherited even though his gene sequence appears to be absolutely normal, and he is a high risk of developing cancer sometime in his life. This is an entirely new pattern of inheritance in cancer."

Gametogenesis

Importantly, the researchers found that the epimutation was erased in the son's sperm. They believe that might point to eggs being less able to strip the methyl tags from the DNA than sperm.

"We believe that eggs may be less efficient at removing the coats from around genes like this than sperm are," Hitchins says. "This is because the way in which the DNA is packaged to fit in the nucleus of eggs and sperm differs. We may find out in the future that these patterns of inheritance may go down the maternal line, but we don't have sufficient evidence yet."

She also points out that it is possible the methyl groups weren't transmitted directly through the mother's egg to the child. "It's also possible that a susceptibility to developing the coat around the DNA was passed on from the mother to the child, through ordinary transmission of a genetic defect.

"In this scenario, the methyl tags may have been stripped away from the affected gene during egg production, as one would anticipate, but were then re-imposed after conception in the new embryo due to its genetic make-up.

"But we can't find anything wrong with the individual's DNA at all and the fact that it is reversed in two out of the three children suggests that there isn't an obvious genetic underpinning to it, as has been shown for other types of inheritance that involve a significant epigenetic component."

For Robyn Ward, the implications of the study are huge. "This is a fundamental change in the way we understand inheritance," she says. "It is a paradigm shift and to be honest it's going to take a lot more study and work to understand how this information can be applied in a day to day medical setting.

"How common or rare these things are we don't know. What we do know is that people will have to do work on other genes and other diseases to work out how important this phenomenon is in terms of determining the risk of other diseases other than cancer.

"For familial cancer this is very important, because with cancer the best thing is early detection and then getting in early with the treatments that are most likely to work. This offers us the opportunity to do that. In the case of bowel cancer, we can intervene by looking to see whether they've got any of the polyps that lead to cancer, so we can prevent them getting cancer altogether.

"We don't know (if this process applies to other diseases) because this is the first instance of this particular phenomenon, but like most things in medicine where it happens once, once you start looking hard enough you can find the same process in other diseases.

The researchers believe that germline epimutations must now be taken into account when looking at disease, along with DNA sequence mutations and environmental factors. They also believe that the discovery could quite quickly lead to re-screening of patients formerly cleared of any DNA mutation.

"These are two of the interesting points from the study," Hitchins says. "There is the possibility of alternative screening of individuals and families with a history of cancer, but no detectable sequence mutation within the gene associated with their disease. And this was just one gene that we looked at so far - there are dozens of others that could be similarly affected."

The work was funded by the Australian National Health and Medical Research Council.