Fat, Fate, and Disease : Why we are losing the war against obesity and chronic disease (20 page)

The other line of evidence comes from studies on animals, mostly rats and mice. In such laboratory animals, which breed rapidly and have a lifespan that usually does not exceed two years, it is much easier to study effects of early development on adult health than it is in human populations. It has been shown repeatedly that reducing the level of food consumed by pregnant rats or mice, either in terms
of the total amount of food or just in its protein content, produces long-term effects on their offspring. These include high blood pressure, insulin resistance, a tendency to deposit fat, altered behavioural responses to stress, reduced activity, and even reduced ability to learn new tasks. This collection of characteristics is highly reminiscent of some of the problems experienced by humans who have conditions such as metabolic syndrome. There are even effects on reproduction and ageing. The offspring can develop osteoporosis, kidney failure, and die at an earlier age.

Many of the fundamental experiments in this field were conducted in our laboratories in Southampton and in Auckland, and our colleagues there have gone on to study in detail the mechanisms underlying these processes. One very striking observation comes from experiments in which the diets of the pregnant rat or mouse and of her subsequent offspring were manipulated experimentally to be very different—mismatched—so that the levels of nutrition the same animal was exposed to as a fetus and then after weaning were discordant. For example, if the pregnant animals were fed an inadequate diet, in terms of either its amount or its protein content, while their growing offspring were fed a diet rich in calories or fat, these offspring became obese and developed the equivalent of metabolic syndrome rapidly. But the key point—which was quickly realized in these labs by scientists such as Mark Vickers and Graham Burdge—was that these animals developed the unhealthy conditions more quickly and more severely than animals whose mothers had been well-nourished during pregnancy but which were also fed the high-calorie or high-fat diet as they grew up. It was as if the relative undernourishment during development had primed them to get fat, develop insulin resistance, and so forth more quickly when they were confronted with a high-fat or high-calorie diet. The speed and magnitude of the problem depended on the degree of the mismatch between their mother’s diet and the diet that they consumed themselves.

These observations have been confirmed many times by other laboratories. But when we see an important biological phenomenon it is important to shift from the observation itself to ask what it means and
why
it happens. Only then can the observation be truly put in context. As we saw, the ultimate answer to
why
questions generally comes from evolutionary biology, and so we should use evolutionary theory to try to explain what is going on.

Our predictive adaptive response theory proposes that aspects of the mother’s environment, including her nutritional status, stress levels, and the number of predators and other threatening events which occur during her pregnancy, are transformed into signals which she sends to her developing fetus. These signals in turn affect the development of her fetus in many ways. The effects do not disrupt development, but they tune it in terms of the relative sizes of the growing organs, the numbers of muscle cells in the heart or of filtering units in the kidney, the extent of fat deposition in the body, and even the levels at which some of its physiological control systems will be set, such as appetite and stress responses. All these developmental processes are made in expectation of the fetus living in an environment similar to that of its mother after it is born and as it grows up. She has educated her baby about the world in which she lives, and this information will be of great importance to her child as he or she grows up. This forecasting (or predictive) strategy is present in humans as well as in many other mammals, and even in some amphibians, plants, and insects.

A good example of this strategy at work can be observed in the little Pennsylvania meadow vole. Because it can be born in either spring or autumn, the baby vole needs to know whether it should be born with a thick or a thin coat of fur, to be prepared for either the cold winter or the hot summer that is coming. The vole has to set the type and the density of its hair follicles before it is born. But it cannot use temperature clues directly, because the temperature in the womb and in the nest is very similar throughout the year. It is only after it
leaves the nest that its thick or thin coat becomes important for survival. If it does not survive the winter or summer it will not reproduce, so there is an evolutionary advantage to getting this right. So the vole has evolved the ability to predict whether summer or winter is coming while it is still a fetus. It does so by sensing its mother’s melatonin levels, which show a different pattern if the days are shortening as winter comes on, or lengthening as summer approaches. This is a fairly safe prediction—hard to get wrong—but it illustrates the simple point that animals make biological decisions early in development which have long-term consequences for their survival, health, and reproductive success.

There are many examples of these predictive processes in insects. Some butterflies develop very different wing colourations depending on the season in which they hatch. The temperature the larva is exposed to changes the biochemical signals that determine wing colouration. The larva picks up the signal about the season and the changes occur during metamorphosis and development of the wing, but the advantage of altered wing colouration only comes later when the butterfly is mature. Leaf colouration is different in different seasons and it is essential to have the right wing colours for camouflage and to hide from predators.

So we proposed that the human fetus was responding to signals for, say, a nutritionally rich or nutritionally poor future environment and set its physiology accordingly. But when a fetus which predicted a nutritionally poor environment was actually faced with a nutritionally rich environment, its physiology would not be correctly tuned and obesity and all the other problems ensued.

This is all very well, but these ideas are based purely on observations—things which happen in nature and for which it is possible to provide a plausible explanation. But how can we be sure that this is the correct or even the best explanation? We need to go back to the lab and conduct further experiments to test the theory. This is precisely what Mark Vickers and his colleagues did in
Auckland. They reasoned that if the development of the newborn rat pup had been influenced by its mother’s diet, and that this would affect the likelihood of its becoming obese when it was given a high-fat diet, then it ought to be possible to prevent or reverse this process by sending a counter-signal at an appropriate time. Suppose, for example, that it was possible to trick the newborn pups into believing that they were already obese very soon after they were born. Would this somehow counteract the signals that they had received before birth and prevent them becoming obese later?

Vickers and colleagues tested this idea directly by giving newborn pups injections of the hormone leptin. We saw earlier in the book that this hormone is produced by fat and acts on the areas of the brain which control appetite. So high levels of leptin should indicate to the growing brain of the newborn rat that it is not hungry and in fact is already relatively obese. When Vickers conducted these experiments the effects were dramatic. The leptin injections completely prevented pups whose mothers had been fed a poor diet during pregnancy from becoming obese—even if they were fed high-fat diets. It also prevented them from developing insulin resistance and high blood pressure, and made them more active and less anxious. Leptin had effectively corrected the message which the developing animals had received from their mothers, based on the poor diet which the mothers were fed. They no longer over-ate the high-fat diet and so did not become obese or develop the characteristics equivalent to human metabolic syndrome. It certainly looked as if the predictive adaptive response theory applied well to rats.

So far, so good, for the theory. Now we had to go back to look for evidence to support this concept of the mismatch pathway in humans. The confirmation came from some work we undertook with Terrence Forrester, who directs one of the world’s most famous nutritional
research centres, the Tropical Metabolism Research Unit in Kingston, Jamaica. This Unit was set up after the Second World War, when Jamaica was still a British colony. There was an extraordinarily high level of under-nutrition there, as emphasized by the Unit’s first director, John Waterlow, one of the founders of the modern science of nutrition, who died only recently.

Jamaica still has a high incidence of severe infant malnutrition, and children are admitted to Forrester’s ward in the hospital in a terrible nutritional state every week. Many families live on the edge of poverty and only one thing needs to happen—the father loses his job or leaves the home, the mother has another child, a hurricane strikes, or just that the child gets an infection—and the child’s nutritional state collapses. Severely malnourished children may develop one of two syndromes: marasmus or kwashiorkor. Infants with marasmus look terribly emaciated, with their muscles wasted away. Infants with kwashiorkor have swollen bodies with pot bellies distended by fluid accumulation. We have all seen such children in the heart-rending pictures of famines in Africa. Infants with kwashiorkor are more likely to die. But tragically, even though they die of malnutrition, they still have fuel stores left in the muscle and fat in their bodies. It is as if they cannot mobilize their fuel supplies, unlike the marasmic children who seem to cling onto life, getting thinner and thinner, until there is no more fuel left to burn.

But no one really understood why some infants develop marasmus and others get kwashiorkor—there had been many theories but none satisfactorily explained the difference. They seem to come from similar families and to be exposed to similar levels of under-nutrition. Then Forrester thought of looking back at their birth weights. And there an explanation lay, for marasmic children had lower birth weights than those who developed kwashiorkor. It seemed that fetuses who were less well nourished before birth, and therefore grew less and had a lower birth weight, had predicted a poor
nutritional environment and had adjusted their physiology to be good at mobilizing their fuel supplies to withstand famine. When famine struck they developed marasmus—becoming terribly thin, but often surviving. In contrast, fetuses who were better nourished and had higher birth weights had not predicted such famine. When it struck they were not so well prepared, developed kwashiorkor, and had a higher chance of dying. For those children, making the right prediction was a matter of life and death.

Would it be possible to get even more direct proof of the predictive adaptive response theory in Jamaica? Vickers’ rats had shown different regulation of appetite according to whether they predicted good or bad times, based on their mother’s diet. If our prediction idea was right we would expect to see that survivors of kwashiorkor and marasmus in Jamaica had different appetite control. We decided to test this idea.

The ecologist David Raubenheimer is South African by birth. The media have dubbed him the Indiana Jones of nutrition. After some years in Oxford he emigrated to New Zealand. He now spends his life in exotic places studying the control of food intake in animals ranging from locusts to snow leopards, from lizards to lemurs. He worked closely with Steve Simpson—who is now at the University of Sydney, but was for some time curator at Oxford University’s Museum of Natural History, where Thomas Huxley had defended Darwin’s ideas against the disbelieving Bishop of Oxford. Raubenheimer and Simpson had developed a very precise method of studying appetite control. It worked in creatures ranging from plague locusts to Oxford medical students. So we enlisted Raubenheimer’s help to see if the appetite control of Jamaican children differed according to their developmental history.

In Jamaica, Raubenheimer and Forrester carefully measured appetite control in a group of survivors of famine, who had suffered either kwashiorkor or marasmus as young children. These people are now between 25 and 40 years of age. Sure enough, those people who had
predicted a bad environment based on their prenatal cues had very different appetite control mechanisms from those of higher birth size, just as with Mark Vickers’ rats.

Clear though this seems, our predictive adaptive response idea has met with some criticisms. The main criticism has come from those researchers who said it was unlikely that the fetus would shift its development for such a long-term advantage. They argued that preserving the mother’s health would be much more important. The argument disappeared when it became clear from mathematical modelling that the major determinant of reproductive success in humans is survival to puberty, so the adaptive advantages of predicting and surviving the childhood environment are indeed large. Another argument, that any conflict for resources between the mother and the fetus must be resolved in favour of the mother, also disappeared—mothers do not sacrifice fetal growth, or indeed breast milk production, even during times of extreme famine. Even under horrific wartime conditions, in refugee camps and famine conditions, babies can be born of normal size, and breast milk production is protected. And, importantly, fetuses do not have to be small at birth to have made predictions about their future. It seems that the forecasting or prediction process forms part of normal human biology.

Other data had already pointed in this direction. Many studies have been done on the Dutch Hunger Winter, as we saw earlier. The famine lasted for seven months from late 1944 until the Allies liberated the Netherlands in 1945. Throughout this period Dutch nurses and doctors continued to keep careful records on births. Thanks to the dedication of researchers such as Tessa Roseboom, many of these children have now been followed into middle age. It turns out that those who were born to mothers who were pregnant during the famine are much more likely to become obese and diabetic in middle age—the mismatch effect again. But those who were only exposed to famine for the first third of their gestation, because they were
conceived late in the war, did not have reduced birth weights. The argument that the fetus is sacrificed for the mother is thus untenable.