Epigenetic inheritance is a wild goose chase

This week's award of the Nobel Prize for medicine to John Gurdon and Shinya Yamanaka effectively recognizes the science of epigenetics. Dr. Gurdon showed that almost any cell (in a frog) contains all the genetic information to become an adult. What makes the cell develop a certain way is a pattern of "epigenetic" modifications to the DNA specific to each tissue-turning genes on and off. Dr. Yamanaka showed that if you can remove that epigenetic modification (in a mouse) you can reprogram a cell to be an embryo.

Yet to most people the word "epigenetics" has come to mean something quite different: the inheritance of nongenetic features acquired by a parent. Most scientists now think the latter effect is rare, unimportant and hugely overhyped.

There are several mechanisms of modifying DNA without altering the genetic code itself. The key point is that these modifications survive the division of cells.

This is crucial to the development of the body: It means brain cells express different genes than kidney cells. Among the implications: a faulty modification in the womb, caused perhaps by maternal dietary deficiency, may condemn the baby to future disease. Babies gestated during the "hunger winter" of 1944-45 in the Netherlands were more likely to be obese and diabetic as adults. Likewise, rat pups insufficiently licked and groomed by their mothers are more likely to be stressed when they grow up.

The far bolder claim is that these modifications can be transmitted between generations, surviving not only the normal division of cells during growth (by "mitosis") but the special processing of cells that prepares them for sexual reproduction as sperm or eggs ("meiosis").

This theory is controversial for three reasons. First, the entire epigenetic mechanism is normally stripped away when an egg or sperm is made. Second, this version of epigenetics rehabilitates the theories of the French scientist Jean-Baptiste Lamarck, who two centuries ago postulated that traits acquired during a lifetime can be genetically passed on to the next generation-a theory long since buried by experiments. Third, the evidence for such a process is sketchy-while the evidence that it has negligible impact, even if it can occasionally happen, is strong.

Caroline Relton of Britain's Newcastle University and George Davey Smith of Bristol University, the editors of a recent special issue of the International Journal of Epidemiology, conclude that epigenetic inheritance may be a distracting wild-goose chase. Yet headlines proclaim "a turning point in our understanding of heredity" and "why your DNA isn't destiny."

The evidence to back up such claims is threadbare. Frequently mentioned is a study of a remote Swedish province called Overkalix, which suffered famines whose effects are felt in the health of the second generation of descendants. But the sample size is small, the effects marginal and no specific epigenetic reprogramming has been established as the cause. Evidence from rats is slightly better: One study found energy metabolism in pregnant rats affected by what happened to the parent. But the most famous animal case, involving a mouse's coat color tied to diabetes and obesity, is somewhat unpredictable and untypical.

Moreover, beginning a century ago with Wilhelm Johannsen, who coined the word "gene," many experiments have ruled out all but the most trivial Lamarckian effects. These now- forgotten tests involved "pure lines" of genetically identical plants or animals. The variation in pure lines, in weight of beans in bean plants for example, again and again shows no heritability, ruling out heritable nongenetic effects.

As Dr. Davey Smith puts it: "The conclusion from over 100 years of research must be that epigenetic inheritance is not a major contributor" to physical resemblance across generations.