In the Feb. 24 online version of Science, the researchers report that mice engineered to have a double dose of insulin-like growth factor 2 (IGF2) develop more so-called precursor cells within the lining of the colon than normal mice. When these mice also carried a colon-cancer-causing genetic mutation, they developed twice as many tumors as those with normal IGF2 levels, the researchers report.
"Both clinically and scientifically, this discovery should expand attention in colon cancer research to earlier events, situations present well before tumors appear," says the study's leader, Andrew Feinberg, M.D., M.P.H., professor of medicine and director of the Center for Epigenetics in Common Human Disease at Johns Hopkins.
"In the mice with a double dose of IGF2, everything is pretty normal except for the extra precursor cells," says Christine Iacobuzio-Donahue, M.D., assistant professor of pathology and oncology. "But when the genetic mutation is present, too, we found a clear cost for what otherwise appears to be a benign effect of extra IGF2."
The team's analysis of colon tissue samples from a dozen or so Johns Hopkins patients with suspected colon cancer suggests that IGF2's effect in people may be similar, the researchers report. A larger study of samples from patients with and without suspected colon cancer is underway, Feinberg notes.
In the mice -- as well as in about 30 percent of colon cancer patients and 10 percent of the general population -- the extra IGF2 stems not from a genetic problem, or mutation, but an "epigenetic" problem that im properly turns on the copy of the IGF2 gene that should remain off.
Unlike most genes, the copy of IGF2 that should be silent depends only on which parent it came from, a situation called genomic imprinting. For IGF2, the copy inherited from the mother is always supposed to be turned off.
In the mice and in some people, however, cells lack the epigenetic "marks" that sit on the DNA and keep the maternally inherited copy turned off. As a result, cells make a double dose of the IGF2 protein and are said to have "loss of imprinting" of IGF2.
Although Feinberg and others have already noted an association between loss of imprinting of IGF2 and colon cancer in people, the current experiments were designed to find out whether the loss of imprinting is involved in cancer's development or just in its progression.
"Most researchers, including me, expect epigenetic differences to influence progression -- whether a tumor would grow slowly or quickly, or whether it would spread," says Feinberg. "But, in this case, our results show that loss of imprinting of IGF2 contributes to colon cancer's development in the mice. It doesn't cause tumors directly, but it creates an environment which is ripe for cancer to start."
Because precursor cells in the colon's lining had been identified as a likely starting point for tumors, Feinberg and his team tossed a cancer-causing genetic mutation into the mix. The IGF2 mice were crossed with mice carrying a mutation in a gene called APC, which had been tied to colon cancer by researchers studying families with excessive growths, or polyps, in the colon.
Mice with extra IGF2 and the APC mutation developed twice the number of tumors as mice with the mutation but whose IGF2 levels were normal. The tumors grew at the same rate in both sets of mice, suggesting that more tumors get started in the mice with extra IGF2, notes Feinberg.
"In the mice, loss of imprinting of IGF2 roughly doubles the ris k that the genetic mutation will cause a tumor," says postdoctoral fellow Atsushi Kaneda, Ph.D. "Double the risk may not seem like much, but this loss of imprinting is common."
The researchers' mice mirror two situations in people because the double dose of IGF2 was accomplished in two ways. One set of mice, obtained from Shirley Tilghman at Princeton, have a double dose of IGF2 because they are missing another gene, H19, whose sequence overlaps the region that usually shuts off one copy of IGF2. As a result, these mice lack H19 and have double IGF2.
To isolate the effect of the extra IGF2, Rolf Ohlsson at the Uppsala University, Sweden, developed a set of mice missing only the control region for IGF2; their H19 gene was intact. These mice likely mimic the 30 percent of colon cancer patients and 10 percent or so of the human population who have loss of imprinting of IGF2.
Only the Princeton mice have been crossed with the APC mice. Both the Princeton and the Swedish mice have the extra precursor cells in the lining of the colon, suggesting the effect on cancer development would be similar.
Co-author Dan Longo, M.D., of the National Institute on Aging notes that the mice with both double IGF2 and the APC mutation should be a useful animal model to evaluate the impact of colon cancer prevention strategies, including dietary interventions and targeted drugs.