Fat, Fate, and Disease : Why we are losing the war against obesity and chronic disease (14 page)

The understanding that there were many small differences in the structure of genes between people led to the search for associations between these differences and a high risk of disease. The concept became simplified to the idea that there might be genes ‘for’ certain conditions such as diabetes, obesity, or high blood pressure, when what was really being meant was that there were particular structural changes in genes that led to greater disease risk. In the 1990s immense amounts of research dollars were ploughed into the Human Genome Project to build on this very idea. If only, the scientists argued, we knew the complete structure of the human genome, all the sequence of the nucleotides in the DNA strands that make up our chromosomes and which provide the code for developing a human being, we might be able to find the particular regions where small differences between individuals change disease risk.

These small differences in an individual’s genetic make-up are called polymorphisms. Because they are found in the genetic
material, they are likely to be passed from mother and father to son and daughter, along lines of inheritance. The search for genes ‘for’ disease depended on having a dictionary of all the genes in the human genome. Compiling that dictionary became a race between a publicly funded consortium in the UK and in Washington and the private enterprise of Craig Venter in the USA. After an enormous amount of pioneering work, and with a great fanfare, the human genome sequence was published almost simultaneously by these two independent endeavours.

Now the fun really started. Where were the gene variants that explained cardiovascular disease or diabetes? Some genes were found that were associated with obesity. An example was the
FTO
gene (the researchers involved apparently wondered whether to call it the
FATSO
gene but thought better of it). People who have one particular variant of the
FTO
gene have a higher chance of becoming fat. Great! The concept of the thrifty gene or the fat gene was proven. Or was it? Well, not quite, because while it is true that people with this particular polymorphism are at greater risk of becoming fat, the reverse is not true. The vast majority of people who become fat do not have this polymorphism in their
FTO
gene. All cats are animals, but not all animals are cats. In fact the FTO polymorphism did not explain much of the obesity pattern at all.

So the search continued. More gene candidates turned up where variations in the gene sequence were associated with some increase in the risk of obesity, but the degree of added risk they conferred and the amount of obesity they explained were generally rather small. By now there about two dozen genes clearly linked to obesity but the amount of fatness in the population that these variants explain still remains small. Indeed, even taking all these genes into account identifies less than 10 per cent of the people in a Western population who become fat. The conclusion must be that genes do not hold the answer—well, at least, inherited polymorphisms are not the answer. Some researchers have argued that the problem is an analytical one
and we need more sophisticated techniques, but only a relatively small fraction of obesity and disease risk—perhaps a quarter—can be explained by using even the most optimistic estimates.

And then other forms of genetic variation were found. One of the surprises of the Human Genome Project was the discovery that our genetic material varies more than expected because some of us carry multiple copies of some genes. This so-called copy number variation was then studied to see if it could provide the genetic link to disease risk that polymorphisms had not. Again small effects were seen, and again, in terms of explaining the variation in disease risk, the results were disappointing.

This was depressing for the genetic research community, although from a medical point of view we could see it as encouraging. After all, if it had turned out that our risk of being fat versus staying slim was something primarily inherited from our parents as fixed genetic variation, and thus determined inexorably at conception, we would have to take a very deterministic and fatalistic view of life. It would mean that while we could identify those children at birth who would most probably become fat, there would be little we could do about it. We can’t re-engineer the human genome in a particular individual. Forget the ideas of gene therapy which have largely turned out to be ineffective, even for diseases where a defect in only one single gene operates, such as cystic fibrosis. For a disorder such as obesity, where a large number of genes might be involved, gene therapy approaches are implausible.

So it was gradually acknowledged that the strong genetic determinism which had driven much of biomedical research for two decades was not particularly helpful in understanding the human condition. Indeed, with only 21,000 genes to explain all our body functions, layers of additional biological control remain to be discovered, and exploring them is a most exciting part of the next generation of scientific enquiries. The complexity of control of gene switches is extraordinary and every year more layers are being uncovered. It is an exciting story to which we will return in the latter part of the book.

While the Human Genome Project was a technical and scientific
tour de force
, it did not provide the answers that its most earnest protagonists had hoped for and in intellectual terms it brought with it many problems. Why was this? Part of the problem lay in the mindset of the scientists doing the work and the media which covered them. We still see it operating today, in claims that a gene ‘for schizophrenia’ or a gene ‘for breast cancer’ has at last been found. There are of course no such genes because these diseases are complex. Our genes have evolved over time to generate biological control over our development and function. They do very prosaic things in controlling the making of proteins and the function of other genes—the design of our bodies has been refined through the Darwinian selection of these very fundamental functions. Virtually all bodily functions involve many genes acting together, and many genes have multiple forms and functions. All of them operate through complex control processes about which we are still learning. So it is naive to imagine that there would be one gene that causes a particular disease.

Indeed even the concept of a gene is changing. We used to think that genes had only one role—to be the template for proteins to be made, so we identified genes by the proteins for which they coded—the
FTO
gene made the FTO protein, and so on. That template involves DNA being read, which in turn leads to the production of an intermediary called RNA. Like DNA, RNA is also made of nucleotides. It is the RNA that then gets read by the cell’s machinery, which leads to amino acids being joined together to make proteins. But now we know that most RNA made on instruction from DNA does not act to make proteins at all, but rather acts to control gene function itself. These non-coding RNAs (so called because they do not code for proteins) create new levels of molecular control—they represent an exciting new area of biology in which new discoveries emerge every week.

As so often in science, part of the problem lies in our use of words. Genes do not cause disease, but paradoxically we tend to name them after the diseases we identify when an abnormality of a gene leads to disease. So the cystic fibrosis gene is named after the disease that occurs when one particular gene is mutated, but in reality that gene’s function is to regulate the amount of chloride going in and out of cells by making a protein which acts as a chloride channel across the cell membrane. It is the malfunction of that channel, caused by a mutation in the gene, that leads to the disease cystic fibrosis. It has been this misuse of genetic deterministic language in talking about genes ‘for’ a disease that has limited the thinking of many scientists. And cystic fibrosis is a straightforward example in that it is a very common and well-documented purely genetic disease which occurs only if both the mother and father of the affected child carry a mutated copy of this gene.

But in the chronic disease story, there is no simple genetic pattern to justify a belief in a strong genetic basis. Certainly some genetic variation across a number of genes plays some role in generating greater disease risk, but not that much and we have not been able to find genetic variation of a type that can account for the common occurrence of obesity and the non-communicable diseases.

Even the ‘thrifty gene’ concept with which we started this chapter turns out to have a flaw. Neel thought that our Palaeolithic ancestors would have been exposed to cycles of feast and famine and that this would have favoured the survival of individuals who could better lay down fat because they had thrifty genes. In fact there is little evidence that our hunter-gatherer forebears did experience such cycles—they probably moved when food sources became low. Skeletal remains do not suggest that our earliest ancestors suffered greatly from malnutrition.

Furthermore, humans have lived in many different environments over many generations. We are a generalist species, good at responding to changes in environments, and it is probably our ability to use technology to stabilize our environment through, for example,
making clothes and shelter and fire which has given us a major advantage over other species on the planet. If we could change the environment, we did not need to change our biology. So trying to find the environment in which our genes are best ‘fitted’ is wrong, because neither that environment nor a particular set of genes suited to it exists.

Yet we cannot ignore genes completely, for they are important in another sense. Selective mechanisms can only operate across the range of environments our predecessors have been exposed to. For example, if none of our ancestors had been exposed to a particular toxin, whether from a plant or a parasite or a new pesticide, we could not possibly have evolved the genes needed for the detoxifying mechanisms to deal with it. Indeed there are some foods that some species eat with impunity, because they have evolved with the right detoxification mechanism, while they are deadly for other species. For example, there are many berries which are poisonous for humans but which some birds can eat. The differences between the species in this respect are usually genetic, and presumably the genetic basis for the detoxification process has evolved in one species and not another because it was necessary for the former to make use of that food source.

In the same way, it is fairly certain that our ancestors were never exposed to the kind of nutritional levels we now have—they would not have had access to highly refined foods and they did not consume energy-dense diets over prolonged periods of time. The processes of selection that operated on our ancestors could only ensure that their metabolic capacities matched the nutritional intakes of their time. So from an evolutionary perspective the nutritional environments we now confront are essentially entirely novel and we cannot expect to have the genetic repertoire to cope with them.

Until perhaps only 100 years ago, not many of our ancestors would have faced high nutritional loads consistently and continuously. Thus we have evolved with a limit to our capacity to cope with
modern energy-dense diets—a limit that we are now increasingly likely to exceed. These concepts of evolutionary novelty and the limits in our evolved capacity to adapt are important because they explain the inevitability of the non-communicable disease epidemic in a world where food supplies have changed so dramatically and so quickly. Because evolutionary processes on the genes are slow, there is no way that we can evolve now to cope with this new nutritional environment. We need to look beyond the gene.

Genetic variation is not the only way in which people differ. They differ in the way they grow up, how they learn, what they are taught, and what they observe. That is why New Zealanders speak English (although the authors disagree on this point) and Brazilians speak Portuguese. That is why orthodox Jews do not eat pork and devout Hindus do not eat beef. That is why some people can swim or fly-fish and some cannot.

What we are is not only a function of our genes. It is a function of the environment we live in, our families, our society, our culture, our physical world. A young woman in rural Mauritania is more likely to be obese than one in Japan for reasons that are embedded in her culture, rather than her genes. A yak herder in Tibet is likely to have to eat much more each day to meet his basic energetic needs than a desk jockey in Hong Kong. And, just as individuals vary in their propensity to get fat or to suffer from a chronic disease, so do populations—and for very much the same reasons.

Populations inhabit and create very different environments: the experiences of a child in northern California are likely to be very different from those of a child in Ethiopia. One will be more likely to have had a well-nourished and healthy mother, a safe birth, good nutrition and immunization in infancy, as well as an education, but at the same time to have been exposed to excess amounts of energy-dense foods
and to have spent much of their time in front of a TV screen. Populations clearly also vary in their economic status, in family structures, and in the types of exercise they habitually undertake. One child may go out with friends to play a ball game. The other may have the role in the family of bringing 50 litres of water from the river a kilometre away each day.

We can summarize the problem we have been discussing by looking at a famous study by Paul McKeigue and colleagues, published in the early 1990s. McKeigue’s group related the waist–hip ratio, which is a good measure of the level of abdominal obesity, to the occurrence of diabetes in different population groups living in the UK. For white Europeans of Caucasian extraction the risk of diabetes for those with a ratio of waist circumference to hip circumference of 1 was about 10 per cent. However, for people of South Asian origin, many of whom had come from the Indian subcontinent early in their lives or in the previous generation, the incidence of diabetes for those with a waist–hip ratio of 1 was about 25 per cent. And at any higher waist–hip ratio, South Asians had about three times the risk of developing diabetes. Of course there are well-known differences in dietary habits between these groups, as well as differences in levels of physical activity and recreational pursuits. If this were the main cause of the difference, then dieting and exercise should be more effective in reducing diabetes in South Asians than in Caucasians, but this is not the case. So is the difference genetic? While some genetic differences clearly exist between the populations, these do not appear to explain the differences in susceptibility to diabetes. Is it because South Asian babies start their lives in a very different way? There are hints that this might be so—these babies are on average smaller at birth than Caucasian babies. We will soon come back to this possibility.