Of mice, men, and genes: the best-laid plans o' DNA gang aft agley

Natural History, May, 2004 by Robert M. Sapolsky

Don't you love urban legends, those outrageous stories everyone believes? There are academicians who study urban legends for a living; they catalogue them, track their origins in Norse mythology, get into arguments at conferences about them. But amid all that intellectualizing, it's just plain fascinating to hear some of the made-up stuff that lots of people fall for. There's the endlessly repeated one about the poodle stuffed into a microwave to get dried off, or the classic about the scuba diver who gets scooped up along with a lot of water into the giant bucket of a firefighting plane, then is dropped onto a forest fire. And there's the one about the woman who leaves groceries in her car on a sweltering day: a tube of cookie dough explodes from the heat just as she gets back in, splattering the back of her head, and she's convinced she's been shot and the dough is her splattered brains.

And then there's the one about a bunch of scientists who sequenced the human genome: they can explain everything about you; all they have to do is look it up in the sequence of your genes. But it just ain't so: we're back in the domain of urban legend.

Why are people such suckers for the idea that genes are the be-all and end-all? The tendency is particularly bad right now. Not only has the human genome recently been (mostly) sequenced, but we've also just come off the golden anniversary of the discovery of the structure of DNA. The celebrations have been replete with religious imagery about the genetic code as holy grail, the Code of Codes. And this imagery even gets trotted out by biologists, people who get paid to know better.

And biologists really should know better, because they've had the sobering concept of "gene-environment interaction" hammered into their heads for much of their lives. (In fact, "gene-environment interaction" is probably one of the first utterances most biologists made as infants, along with "doggie, doggie come.") The trouble is, it is a phrase so often repeated that it has become as reflexive and ingrained--and, ultimately, ignored--as the words to "Elmo's Song."

The idea that genes and environment interact can mean a number of things. At the least, it means that people who get into black-and-white arguments about nature versus nurture are a century out of date. Of more relevance, it means that though genes can (indirectly) instruct cells, organs, and organisms about how to function in the environment, the environment can also regulate which genes become active at particular times. Of greatest relevance here, though, is that the thing a particular gene most proximally produces--a particular protein--can function quite differently in different environments. So, in theory, you might have a gene that in one environment causes you to grow antlers and, in another, causes you to fly south for the winter.

For folks who still want to fight the nature-nurture wars, the question up for debate becomes: OK, just how powerful are these gene-environment interactions? At one extreme are those who scoff at contrasts as gaping as the one between growing antlers and flying south. In their view, a gene does something or other, and the environment perhaps alters how fast or how strong or how long the gene does that something or other. But none of those environmental influences lead to dramatically different effects. Framed in the context of genes and disease, it's like saying, Yeah, how windy it is may alter the precise speed with which the anvil drops from a ten-story building and lands on your toe, but who cares about that environmental interaction with the anvil? And at the other extreme are those who assert that interactions can be of huge consequence--say, if what is dropping is a feather and not an anvil.

And so the scientists happily argue and experiment away, squandering tax dollars that could otherwise go for Halliburton contracts. Amid these debates, it's useful to be reminded just how powerful gene-environment interactions can be. Two recent studies provide some terrific examples.

The first study investigated the effects of one of the subtlest, least appreciated environments: the prenatal one. For many years, strains of laboratory rodents have been bred for various traits--one strain gets a type of diabetes, another strain gets hypertension, and so on. Each strain is developed by inbreeding generation after generation of animals with some trait, until all the members of the strain are as close as possible to being genetically identical--like clones of one another. If all the members of that strain show the trait, regardless of which laboratory they're raised in, there's some reason to think that the animals are subject to a strong genetic influence.

All the inbreeding is then followed by an experiment known as a "cross-fostering study," regarded as critical for detecting a genetic influence. Suppose all the mice of the Coke strain grow up preferring Coke to Pepsi, and all the mice of the Pepsi strain grow up displaying the opposite persuasion. Take some Coke-strain mice at birth and let Pepsi-strain moms raise them in a Pepsi-strain colony. If the Coke-strain mice still grow up craving Coke, the typical interpretation is that you've found a behavior that strongly resists environmental influence; score one for nature over nurture. But are cross-fostering experiments the last word?