November 3, 1998
Johns Hopkins scientists have found a genetic alteration associated with common forms of colon cancer in patients' normal cells. The same abnormality, called loss of imprinting (LOI), also appears, the researchers say, in a significant number of healthy people, offering a possibility of predicting as many as 40 percent of new colon cancer cases before they start.
"It's the first cancer-related genetic alteration found at high frequency in healthy people," says Andrew P. Feinberg, M.D., who with Hengmi Cui, Ph.D., led the study appearing in this month's Nature Medicine. By contrast, APC and HNPCC, the known colon cancer genes that predict high risk of the disease, together account for less than 3 percent of new cases.
In the study, the Hopkins team analyzed both tumors and normal parts of the colon in 30 patients with colon cancer, looking for loss of imprinting. LOI -- a recently discovered phenomenon with a strong tie to cancer and other diseases -- refers to an abnormal switching-on or off of genes.
People inherit copies of the same gene from both parents, and both genes or sets of genes work in tandem in a person's cells. In normal development, in a process called imprinting, the copy from one parent, in certain genes, is turned off. This inherited preferential silencing usually holds throughout all cells in a person's body.
In their investigations of cancer, however, researchers have learned that such genes can lose their imprinting. "We know there is a link between LOI and cancer," Feinberg says, "because we've repeatedly found LOI in cancers, notably in genes that either encourage cell growth or that suppress cancer." Earlier work by Feinberg and by others has showed, for example, that LOI appears frequently in Wilms' tumor, the most common non-blood cancer in children.
In this new study, researchers focused on a specific growth-promoting gene called IGF2 (for insulin-like growth factor), one of several cell growth genes normally imprinted so that only the father's gene gets expressed. But in 44 percent of the cancer patients, that imprinting was lost in their tumor cells.
"Even more surprising to us," says Cui, "was the appearance of LOI in these patients' healthy colon tissue and blood cell samples." Finding such flaws in healthy cells is rare, he adds.
Thirteen percent of the study's 31 control patients, with no sign of cancer, also tested positive for LOI in colon tissues or blood. Before doctors can use LOI to screen patients, however, Feinberg says the studies must extend to a large number of healthy people. "Further research will be needed to establish LOI's predictive value," says Feinberg.
What strengthens the case for LOI's cancer-predicting potential is a link, in this study, between LOI in healthy cells and a second defect found in the actual cancers. This defect, called microsatellite instability (MSI), involves an unusual duplication of certain stretches of DNA -- a sign of a glitch in the way DNA reproduces or repairs itself. Oncologists associate MSI with cancers that come at a relatively early age and run in families. Some use MSI to predict a patient's outcome.
When the Hopkins researchers examined the colon cancer patients' normal cells, they found that those with abnormal imprinting were "for the most part, the same people with MSI," says Feinberg. And those same patients got colon cancer 14 years earlier than the others, on average.
"We don't yet know if LOI is causal for cancer," says Feinberg. Knowing that, he says, would be especially useful because, unlike conventional mutations, LOI is potentially reversible.
Imprinting was first linked to cancer about a decade ago, with the childhood tumor, rhabdomyosarcoma. In addition to Wilms' tumor, LOI is also definitely implicated in Prader-Willi syndrome, which causes retardation, and in some adult lung cancers.
The most striking case of large-scale genome imprinting involves crosses between horses and donkeys. Cross a female horse and a male donkey and you get a mule. Cross a male horse and a female donkey and you get a hinny, altogether a different creature. Clearly, the same genes act out different roles, depending on whether they come from "mom" or "dad."
Other members of the research team were Stanley R. Hamilton, M.D., and Rolf Ohlsson, of the University of Uppsala, Sweden.
A patent is pending on the techniques in this research.
Funding for this study was from NIH grants.
The research described in this press release was supported by the National Institutes of Health. Under a license agreement between the Johns Hopkins University and Epigene, Inc., Feinberg is entitled to a share of sales royalty received by the University from sales of the technology described in this press release. The license will become effective once the company has achieved certain financial milestones. Dr. Feinberg and the University own Epigene stock, which is subject to certain restrictions under University policy. Feinberg also is a consultant to Epigene and serves on its Board of Directors and Scientific Advisory Board. The terms of this arrangement are being managed by the University in accordance with its conflict of interest policies.