Starving a pregnant mouse can cause changes in the sperm of her sons that apparently warp the health of her grandchildren, according to a new study. The finding offers some of the strongest evidence yet that a mother’s environment during pregnancy can alter the expression of DNA in ways that are passed on to future generations.
A number of studies have suggested that environmental stresses in a parent may harm the health of subsequent generations. For example, women who were pregnant during a 1944 famine in the Netherlands known as the Dutch Hunger Winter had children and grandchildren who were unusually small or prone to diabetes and obesity. Animal studies have also found that a stress to a parent, such as exposing a pregnant mouse to toxic chemicals or mildly shocking a mouse father to make it fear an odor, can result in effects such as infertility or changes in behavior that persist for two generations or more yet can’t be explained by genetic mutations.
Some scientists suspect that the effects are passed down via so-called epigenetic changes, chemical modifications of DNA that can turn genes on or off. A team led by geneticist Anne Ferguson-Smith of the University of Cambridge in the United Kingdom and diabetes researcher Mary-Elizabeth Patti of Harvard Medical School has now explored this idea by studying the DNA of two generations of mice descended from an undernourished mother.
The researchers gave pregnant mice chow containing only half the calories they needed during the last week of gestation—a time when the epigenetic patterns in a male embryo’s primordial sperm are erased, then reset. As Patti’s group had previously shown, this treatment resulted in offspring and grandchildren that were underweight and prone to diabetes.
The group next examined DNA from the sperm of the males born to the starved moms. Compared with sons of control mice, their sperm had fewer chemical tags known as methyl groups on about 110 stretches of DNA. Often, the methyl groups were missing near genes involved in metabolism that may play a role in obesity and diabetes. The expression of these genes was also altered in some body tissues.
However, although fetal tissues from the mother mouse’s grandchildren also had similar changes in gene expression, surprisingly, the DNA in these tissues did not carry these methylation differences. That suggests that the changes eventually disappear, the team reports online today in Science. Ferguson-Smith thinks methylation changes in the son’s sperm reflect the legacy of his undernutrition in the womb but, because they do not persist, would not directly explain the grandchildren’s disease. The methylation marks “are not the long-term memory that links one generation’s disease to the next,” she says.
“This is a very nice study” linking ancestral exposures to epigenetic changes, “but I wouldn’t say the book is closed on how these things work,” says epigenetics researcher Oliver Rando of the University of Massachusetts Medical School in Worcester. To show that such methylation patterns cause the health effects seen in the male sons and their offspring, one needs to artificially turn off or on the suspect genes and show that this leads to the same result, he says. “Perturbing the epigenome is the big challenge for the field.”
What’s more, the study does not rule out that the DNA methylation patterns are inherited for multiple generations because the researchers didn’t look for them in the sperm of the grandsons, Rando adds.
Some are more skeptical. Columbia University geneticist Timothy Bestor has “a number of issues” with the study. Among them is that instead of studying inbred mice that were genetically identical, the researchers used a strain in which individual mice vary genetically. Although that may have made the mice more similar to the human population, it raises the possibility that in the womb, only fetal mice with a particular genetic makeup may have survived starvation. Because genetics also shapes methylation patterns, these genetic differences could be why their sperm DNA methylation patterns differed from those of control mice, Bestor says—not because malnutrition directly altered the patterns.