Thursday, November 8, 2012

Fetal Origins and Epigenetics: Interview with Janet Currie

Douglas Clement has an insightful interview with Janet Currie in the September 2012 issue of The Region magazine published by the Federal Reserve Bank of Minneapolis.  As Currie says near the start of the interview: “Labor economists think a lot about human capital and investments in it. Traditionally, that’s something to do with education, ... [b]ut I’m interested in health as human capital as well, and understanding how health and education intersect. ... “It is a broad concept, human capital ... Not all these different boxes, but an integrated whole." The interview makes a number of interesting points about an array of subjects, including how financial incentives affect the practice of medicine ( no matter what doctors say!), early childhood education, and women in the economics profession.

Here, I'll focus on one topic: the fetal origins of inequality. Here's the question from Clement, and part of Currie's answer:


Region: Could you briefly review the fetal origins hypothesis and how economists have expanded its reach—to test scores, education and income as well as health?

Currie: I think the phrase itself was coined by David Barker, a physician who was interested in whether there was a biological mechanism such that if the fetus was starved in utero it would be more likely to be obese or more likely to have heart disease or diabetes, things related to that in later life.
... An infant programmed in this way would then be more likely to gain a lot of weight later on and to have diseases related to obesity.... I believe Thalidomide was the first thing that really shocked people and showed that if you give drugs to the woman, that it could have an effect on the fetus. People were also working on the Dutch “Hunger Winter” prior to Barker, looking into whether being literally starved in utero had long-term effects.

So economists have taken that idea and run with it. Economic studies are examining a wide range of things that might affect fetal health and asking whether they have long-term consequences. I think there’s pretty broad acceptance now of the idea that all kinds of things that happen when people are in utero seem to have a long-term effect.

One of the things I talked about in my Ely lecture was what mechanism might underlie the long term effects, and I raised the idea of “epigenetic” changes as one possibility. The way I like to think about that is you have the gene, which only changes very slowly when you have mutations. But then kind of on top of the gene you have the epigenome, which determines which parts of the gene are expressed. And that can change within one generation. There are animal experiments that do things like change the diet of guinea pigs and all the baby guinea pigs come out a different color. It can be pretty dramatic. ...   The idea is that the fetal period might be particularly important because these epigenetic switches are being set one way or another. And then once they’re set, it’s more difficult to change them later on.

I think we haven’t really been able to look at all of the implications of that given the limitations of the data. We don’t have very much data where we can follow people from, say, in utero to some later period. But, that’s where the frontier is, trying to do that kind of research and make those linkages....

I think a really interesting thing about the fetal origins hypothesis for public policy is that if it’s really important what happens to the fetus, and some people think that maybe the first trimester is the most important or the most vulnerable period, then you’re talking about women who might not even know that they’re pregnant. It really means you should be targeting a whole different population than, say, 15 years ago, when we thought, oh, we need to be targeting preschool kids instead of kids once they reach school age. Now we’re kind of pushing it back. Then it was, “We need to be playing Mozart to infants.” Now the implication is that we’ve got to reach these mothers before they even get pregnant if we really want to improve conditions.

Epigenetics implies that it does not make sense to talk about nature versus nurture. If nature is the gene and nurture is the thing that sets the switches, then the outcome depends on both of those things. So you can’t really talk about nature or nurture in most situations. It has to be some combination of both. ...

One thing that is interesting—and I’m starting to do a little bit of work like this myself—is thinking about children in developing countries. Things we’re looking at here in the United States, like the effects of in utero exposure to pollution on child health and economic outcomes, involve problems that are much worse in developing countries. So if we can find an effect here … for instance, my E-ZPass paper suggested that the incidence of low birth weight was 8 percent higher for pregnant women who are subjected to large amounts of auto exhaust because they live near highway toll plazas. If that is true here, then what must be the effect in Beijing? It must be even bigger than that.
Currie, along with Douglas Almond wrote on this topic in the Summer 2011 issue of my own Journal of Economic Perspectives in "Killing Me Softly: The Fetal Origins Hypothesis" (25(3): 153-72). That article, like all articles in JEP back to the first issue in 1987, is freely available to all courtesy of the American Economic Association. Another useful starting point to this literature is Currie's Ely lecture on this topic (mentioned in passing in her answer above): "Inequality at Birth: Some Causes and Consequences." American Economic Review, 101(3): 1-22. The AER is not freely available on-line, but many in academia will have on-line access through AEA memberships or library subscriptions.

 For yet another recent angle, there's an article in a recent Economist magazine on "Epigenetics and Health: Grandma's curse." The big question is whether an acquired characteristic--like, say, asthma that is caused by heavy smoking--can be inherited. A plain-vanilla theory of inheritance says this isn't possible, because smoking is bad for you, but it doesn't alter your genes. However, studies on pregnant rats apparently show that if a first generation of rats is dosed with nicotine, which leads to asthma, then the first generation of offspring also has a propensity to asthma. That result is unsurprising, because it's just the fetal origins hypothesis in action. But apparently, the next generation of rats after that also has a greater propensity to asthma, although that generation was never expose to nicotine directly. The epigenetic explanation is that nicotine doesn't change genes, but it can alter the "switches" that determine what characteristics are expressed by those genes, and those different "switches" can be to some extent inherited. As the article writes: [T]hose epigenetic changes that are inherited seem to be subsequently reversible. But the idea that acquired characteristics can be inherited at all is still an important and novel one ..."