Fat, Fate, and Disease : Why we are losing the war against obesity and chronic disease (23 page)

But we need to go down another layer in explaining the mechanisms of how a less than optimal start to life has long-term consequences. To do this excavation, we must use the techniques of molecular biology. We inherit the genes in our DNA from our parents but, as we explained earlier, next to the genes are segments of DNA that control gene function itself. These control regions can be envisaged as panels of switches—switches that tell genes when to be turned on and when to be turned off. For example, the insulin gene is turned on in pancreatic cells but as the settings are specific to cell types, the insulin gene will be turned off in liver cells.

Some switches are on or off at different times in life. So while the DNA sequence we inherit does not change during life, the switches controlling genes can be biochemically altered—as we go through puberty or old age, for example. But most importantly, these switches can be set to on or off in response to environmental signals received in early development. These control processes are complex because there can be thousands of these switches on every gene. They respond to molecules called transcription factors, which are the signal for turning the genes on and off and, because they are made by yet other genes, we end up with a very complex network of fine controls in every cell which use all sorts of feedback loops. But while changes in gene control can be very subtle, they can nonetheless have major effects.

These gene switch states are most modifiable in early development. This can be achieved by some chemical modifications such as methylating specific parts of the DNA—a surprisingly simple change which involves merely adding one carbon and three hydrogen atoms to a receptive site on the DNA. Other changes affect the histones—protein cores around which the DNA is wrapped and packaged and which are necessary to allow the approximately 2 m of DNA in every cell to be packaged into a small space. These are
very carefully regulated steps in cell biology and we call them epigenetic changes. Epigenetic changes are passed from one cell to its daughter cells when it divides. As we develop from a one-cell embryo into a complex being, we start differentiating into various types of cells and these have different types of epigenetic profiles in them. Epigenetic mechanisms play a critical role in many biological processes—for example, they stop certain types of viruses that got incorporated into our ancestors’ DNA from being active in us. Another function is to silence one of the two X chromosomes in every female cell so that the female does not get twice the dose of gene expression compared to males, who have only one X chromosome in each cell.

One of the most exciting discoveries in developmental biology has been that developmental plasticity depends on epigenetic processes. For example, the queen bee and the worker bee are both female bees; the only difference is that they have different epigenetic states, and we now know that these different states are induced early in life by what the larva is fed by the drones as it develops.

In recent years it has become possible to measure each individual gene switch by sophisticated molecular techniques, often using very expensive equipment. So, if our ideas are right we should find epigenetic changes associated with altered fetal experiences, and these might be related to what later happened to the offspring—especially in relation to disease risk. With this in mind, we decided to look for DNA methylation changes in very precise bits of the gene regulatory regions. We started searching in rats, while other groups around the world embarked on similar studies. Irrespective of how a poor start to life was induced, whether by changing maternal nutrition, or by giving the mother high doses of a stress hormone, changes in these gene switches were found in many important places in the DNA of the offspring. They were quite specific, however, occurring not randomly throughout the DNA, but on genes where the setting of the switch could have quite major consequences.

One set of experiments was particularly revealing. Michael Meaney, a neuroscientist in Montreal, used a well-established experimental model, in which he compared the offspring of rats that were well cared for with those poorly cared for by their mothers. How ‘good’ a rat mother is in looking after her young pups is indicated by how much licking and grooming she gives them. Those offspring that were poorly looked after grew up to be anxious animals, with high stress hormone levels in their bloodstream and, if they were female, became poor mothers themselves. By experiments in which he fostered pups onto different mothers he was able to show that this was not a genetic effect, but was set up by the early life experience. Meaney found that the effect was due to the triggering of an epigenetic change in a hormone receptor in the brain—indeed in the very receptor and the very pathway which would lead to high stress hormone levels. Lastly he used a drug which undid the epigenetic change and showed that it reversed the gene switch change and, even though the offspring were now adults, their behaviour returned to normal. Here was the proof that early, developmentally induced epigenetic changes could have lifelong effects on behaviour.

At the same time, Karen Lillycrop and Graham Burdge in Southampton were studying the effects of the diet fed to pregnant rats on epigenetic processes in their offspring. Using a diet low in protein and high in carbohydrates, they found that the pups showed epigenetic changes not only in stress hormone receptors but in the gene pathways controlling metabolism of fats in their livers. Just as in Meaney’s experiments, the effects seemed to be permanent in the adult offspring. But, once again, it was possible to prevent them by giving an additional nutritional stimulus during development—here it was the micronutrient folic acid, and the cancelling effect occurred even though the pregnant animals were still on the low-protein, high-carbohydrate diet.

Inspired by this study, we studied Mark Vickers’ rats that had been given leptin—these are the rats that had been tricked in their prediction and did not get fat, pre-diabetic, or eat to excess when given a high-fat diet. Leptin not only stopped the animals developing obesity; it stopped them having the epigenetic changes induced by prenatal under-nutrition. And Rebecca Simmons, a neonatologist in Philadelphia, used a different rat model of a poor start to life in which she found that giving newborn rats a drug related to a hormone made by the stomach reversed the abnormalities in the insulin-secreting cells of pancreas as well as the epigenetic changes in a gene controlling pancreas development.

These experiments make a compelling case that epigenetic processes are involved in the predictions which fetuses and infants make about their future lives, based on cues from the mother. This is a rapidly moving field—there is now a flood of studies showing epigenetic changes in animals which support these ideas.

As so often in experimental medicine, studies in animals reveal the important questions to ask in humans. As yet the human epigenetic data are more limited. There are some data suggesting that survivors of the Dutch Hunger Winter have changes in gene switches which can be seen in their white blood cells as adults, but the effects are small and the data difficult to interpret. Because white blood cells are made all the time, it is hard to relate epigenetic changes in them (in an adult) to a person’s early development. What was needed was a source of DNA collected from the person as a fetus which could be related to that person’s characteristics later.

It turned out that by good fortune or foresight, in the Southampton cohort studies umbilical cord tissue had been collected and
deep-frozen at birth to preserve it. This provided a source of DNA to study. Combining the forces and resources of our laboratories in Auckland and Southampton, we decided to measure the gene switches in this DNA. Our molecular biologists, led by Karen Lilly-crop and Allan Sheppard, had to develop new techniques to do so, and given the implications of anything we might find, we had to be very careful and pedantic with our methodology.

We did indeed find some gene switch changes that were associated with birth weight but these were not what interested us most. We wanted to know whether we could prove that developmental influence had important lifelong echoes even for children within the normal birth weight range because, as we discussed earlier, the risk of diabetes and cardiovascular disease is graded across the normal range. So if epigenetic changes are associated with this risk we should pick them up in normal weight babies.

The Southampton children have been followed for some years and have been extensively studied. They have had scans to measure the amount of fat and muscle and the density of bone in their bodies; they have had intensive cardiovascular measurements made; they have been examined for allergic outcomes; and they have even had cognitive and behavioural tests performed. So we set out to look for relationships between epigenetic changes at birth and measures of disease risk later in childhood. What we found astonished us.

First we looked at body composition. In our first study we found that the degree of epigenetic change measured at birth in one particular gene, associated with the control of fat metabolism, explained about 25 per cent of the differences in body fat between children nine years later. These results astounded us but we were nervous. The nature of science is such that extraordinary results can be an accident or a fluke. Therefore the more outlandish the result, the more important it is to replicate it. And for us this was an extraordinary result, maybe the most important of our careers—if it was right. It suggested that just measuring one gene switch at birth could tell us far more
about which person might develop obesity than all the billions of dollars spent on looking at structural mutations in genes had found. If this result was true it had two enormous implications: firstly, that the developmental period before birth was far more important in determining risk of obesity than most researchers and doctors had expected; secondly, it shifted the focus for future research from genomic variation studies following the Human Genome Project to epigenetic processes. We had to find out whether this first result was right or not.

So we repeated the study—not once, but twice, using a second birth cohort—the famous Southampton Women’s Survey we described before. The results were identical to those in the initial study. Methylation changes at birth in one particular gene at one particular site tell us much more about whether or not children will get obese than genetic variation.

And the work led to another important conclusion. We were able to show that the epigenetic switches measurable in the child’s umbilical cord after birth were related to the mother’s diet in early pregnancy. We showed that the switch was in the healthier direction in babies of women with higher carbohydrate and lower dairy protein intakes. We do not yet know how this comes about, let alone the answer to the
why
question that we flagged up earlier in the book, because association does not prove causation. It may be that it is some other aspect of the woman’s diet or lifestyle associated with carbohydrate intake that is important.

When the work was published in the journal
Diabetes
, journalists wanted to know what this meant for mothers’ diets. Unfortunately we cannot yet answer this fully, as it would be unsafe to go beyond the current data to be more precise. But with this knowledge we can now start evaluating which diets before and during pregnancy produce the best epigenetic profiles in the babies and thus the most healthy lives as they grow up. This will be the next phase of our work.

And our work has thrown up other tantalizing associations. We have found epigenetic signatures that are reflected in altered bone density, which is important because lower bone density starts one on the path to osteoporosis, to altered blood vessel function, to increased risks of allergy, and to altered attention and behaviour. This fits very well with the animal data, but as yet we have not replicated and confirmed these findings. Once we have we will publish them.

But it is clear that life before birth has an enormous influence on one’s destiny. Far more so than nearly everyone would have imagined. And life before birth can be influenced by what the mother does, working through epigenetic mechanisms to change the destiny of the child.

The focus has shifted: we can now see that human development has a critical influence on the risk of developing obesity, diabetes, and cardiovascular disease and probably on many other aspects of what makes us what we are. Now we can begin to see how our parents affect that development, although it is not through the genetic mechanism which the Human Genome Project had hoped to reveal (but did not). Rather, our developmental environment affects epigenetic gene switches which are set in expectation of the world in which we predict that we will live later. These processes appear to have evolved so that the prediction can change aspects of our bodies to maximize our survival until reproductive age, and reproductive fitness itself. The mismatches of our modern world, and our longer lives, have now made the epigenetic switches inappropriate, resulting in a greater risk of disease. With this information we can predict with greater certainty than ever before the likelihood that a child will become obese, and so be on the risky road to diabetes or cardiovascular disease.

The implications are broader than just having new knowledge. As epigenetic measurements become more established we can use that
information in many ways. We can use it in research to find out what mothers should do to get the best outcomes for their pregnancies. We can use it to predict the probable destiny of tomorrow’s children, and we can then use it to see whether various ways of bringing up our children—what and how we feed them, for example—might give them the healthiest possible destiny.

We can imagine (and we know from experience!) that about half of those reading this will be somewhat annoyed, maybe even a little angry, about this. Are we saying that what happens to us is all our mother’s fault? That the responsibility for chronic disease, and the solution to it—and the blame for it too—rest entirely with women?