by Greg Gibson
Unfortunately, as yet we have almost no idea what FTO does, where it does it, or why the different flavors do it differently. Surely it will not take long for scientists to figure this out, but it is one of the frustrating things about contemporary genetics. Like rising gasoline prices, often we can see what the problem is but can’t do anything about it.
Another example of this is the gene called INSIG2. My first thought on reading the paper describing the discovery of this particular obesity gene was that the researchers must have a wry sense of humor and a fearless attitude toward funding agencies. So many findings of highly significant associations between genes and diseases have turned out to be false leads that it is almost asking for trouble to give your gene a name that conjures up “insignificance.” But it turns out that INSIG2 actually stands for Insulin-induced gene 2. The protein encoded by the gene seems to be involved in the synthesis of fatty acids and cholesterol, which obviously makes sense if you are looking for something to do with obesity.
The actual DNA changes that alter the function of the gene have yet to be identified, but three different studies of several thousand Caucasians on both sides of the Atlantic, and of African Americans, suggest that the ten percent of individuals who have two versions of the less common allele are more likely to be obese. The risky variant is at a relatively constant frequency of around one-third of the alleles in all populations examined, but oddly it does not seem to promote weight gain in all populations. The obesity-associated allele seems to be the ancient one that was present in humans before they began moving around the globe and practicing agriculture. The protective type is the more modern one, implying that we are evolving a genetic constitution that is less predisposed to putting on weight.
Another gene, ENPP1, also known as PC1, popped out of a search for genes that might be involved in diabetes. It encodes a protein that binds to and turns down the function of the insulin receptor. If you think of the insulin receptor as the main control on your dashboard that allows you to work the air conditioning in your car, then ENPP1 is like the knob on the air vent that provides a little more control. Cell biology is full of such devices that you can do without, but that make life easier.
A common form of ENPP1 in humans has one different amino acid, affecting how the protein binds to the insulin receptor. Many other alleles also affect how much of the protein is made in particular tissues such as the pancreas, the liver, and fat cells. It is easy to imagine that such different forms affect the development of resistance to insulin, and hence susceptibility to diabetes, and it seems that they do. But it turns out that they also predispose carriers toward obesity. Not a lot, but enough to make a difference, at least in Caucasians: One study of Japanese found no effect at all. This is emerging as a common occurrence in the genetics of disease: Whether a variant flavor of a gene matters is peculiar to each population.
Just like cancer, a few percent of morbid obesity is actually explained by heritable cases that can be attributed to single severe mutations. One of the main culprits is a gene that mediates the feeling of satiety, called the fourth melanocortin receptor, MCR4. Melanocortin is one of those fascinating genes that seem to be involved in everything. Variation in one receptor gene, MCR1, is responsible for the white patches on many of our furry friends, while another one is implicated in erectile dysfunction, and seasonal affect disorder ties in there somehow as well. Numerous studies now encompassing tens of thousands of cases establish a weak link between common variation in MCR4 and obesity. I can’t wait for the advertising campaigns that, after drugs are developed that overcome the genetic legacy here, solemnly warn that if you experience an erection lasting for more than four hours while trying to lose weight, consult your doctor.
Type 2 Diabetes
Diabetes is to obesity as crime is to unemployment. You can have one without the other, but the latter certainly leads to the former. Consequently, the obesity gene map heavily overlaps with the diabetes gene map, though we now know that it is not sufficient. Many common polymorphisms contribute to diabetes without playing much of a role in normal weight control.
The body has evolved exquisite mechanisms based around insulin to keep glucose concentrations in balance. Normally it copes with excess by storing it as fat. But if the system is overloaded for too long, it becomes stressed, and eventually throws in the towel. Diabetes then arises from a double dose of disequilibrium. First is the disequilibrium between genes and the environment; second is the loss of equilibrium in the balance of hormones that function to ensure that energy reserves are kept within a healthy range.
Why would the body build up resistance to the very hormone that makes life possible? Why would evolution tolerate the accumulation of genetic variants that cause insulin signaling to go so terribly wrong? This is really not a case of bad genes, but rather of good genes forced to do bad things under abnormal circumstances.
Insulin levels go up after a meal because the hormone tells the body systemically how to deal with all the new glucose, but the levels cannot stay up indefinitely. While there is a tap on insulin production at the source in the production, it is more efficient to control usage of the hormone at each tissue. This provides a type of buffering.
It might help to think of insulin resistance as caller ID. Telephones are great things when they help us remember to pick up some tomatoes on the way home or let us talk to loved-ones thousands of miles away. But when telemarketers worked out that they are also an effective device for hawking unwanted banking services just as we sit down to dinner, telephones became a threat. So we evolved caller ID so that we can locally screen the incoming messages, thereby making a relatively simple communication device somewhat complicated, but much better regulated and controlled.
One of the wondrous features of the insulin buffering mechanism is that it is flexible enough to adjust its sensitivity as life unfolds. During puberty or pregnancy, for example, we have different physiological responses to eating and adjust glucose metabolism accordingly, in part by elevating insulin release. Similarly, what is ordinarily a good level of insulin to get the job done, may not be so good as we go through phases of weight gain and healthy exercise, or illness, or just getting older. So the body is constantly adjusting levels of insulin resistance in proportion to the production of insulin in the first place. As insulin production goes up, so too does resistance. This sort of feedback loop serves to keep the whole system in equilibrium.
In obese people with constantly high levels of something called non-esterified fatty acids and with large rolls of abdominal fat deposits, long-term increase in resistance occurs, and the body enters something of a cry-wolf situation. Used to receiving the insulin signal all the time, eventually the person’s cells become insensitive to it. At that point, resistance eventually leads to impaired insulin release in the pancreas as the beta cells shut down. What normally functions as a negative feedback loop to keep glucose in the normal range becomes the source of disease after a lifetime of stress. The genes aren’t bad; they’re just trying to do the right thing, but are getting it wrong because they are out of equilibrium with their environment.
Among the 50 or so genes thought to contribute to diabetes in this way, three are well enough characterized that even the most hardened skeptics would have difficulty refuting their involvement.
Calpain 10 is the textbook example both for how to find a gene for a complex disease and for differences among populations in the effect it has. Initially identified in Finland and quickly confirmed in Mexican Americans, for whom it is responsible for as much as one-tenth of disease susceptibility, it is nevertheless not involved in diabetes in Japanese, Samoans, or Africans.
PPARG has such a long formal name that it even puts biochemists to sleep: peroxisome proliferator-activated receptor-gamma. It encodes a growth factor receptor involved in regulation of adipocyte (that is, fat cell) development, whereas Calpain10 encodes a protein that digests other proteins and affects pancreatic function. PPARG’s oddity is that the
susceptibility allele, thought to be a change of the twelfth amino acid from one type to another, is the common type in humans. More than three-quarters of us have what simplistically might be called the bad gene, but it has such a small effect that it does not lead to diabetes for most of us. On the other hand, most diabetics have it, so in the end it explains quite a bit of the disease susceptibility.
A third gene only just emerged from a series of whole genome scans late in 2006. It is evidently such a bad gene that it gets a criminal name: TCF7L2. Call it the Lex Luthor of genes, because it is also a mastermind, encoding a transcription factor whose role it is to control other genes. Like the other two genes, variation in TCF7L2 accounts for as much as 20 percent of the incidence of diabetes in Africans and Europeans. Even more telling, the old, risky, allele has almost disappeared from east Asia, where correspondingly diabetes is relatively rare.
Debunking the Thrifty Genes Hypothesis
We now need to address the issue of why risk alleles are found at a common frequency in the human genome. There are logically three possibilities. First, they are just drifting around, too inconsequential for natural selection to pay much attention to. Second, a balance between the advantages and the disadvantages they confer may actively keep them around. Third, some of the alleles may be truly beneficial, but they are too young and there has not yet been enough time for them to displace the old ones. Evidence is accumulating that all these mechanisms are involved.
Let’s start our discussion, though, with an explanation that you might have read about in the Sunday papers, the so-called “thrifty genes hypothesis.” To be thrifty is to be wise with your resources, to be economical and not wasteful, to plan for the future. It follows that thrifty genes would be ones that allow you to conserve your food reserves for a rainy day, or another cache of berries or kill of wild meat as the case may have been for an early human.
It is precisely this concept that the University of Michigan geneticist Jim Neel had in mind back in 1962 when he coined “thrifty genes” as an explanation for the recent upsurge in obesity. Individuals who are better able to convert a surplus of calories into fat reserves would be more likely to see through times of famine that were presumably common as the species was first evolving. Yet confronted with the supermarket diet rich in sugars and fats and all things in plenty, those genetically same individuals are now predisposed to develop unwanted fat reserves.
Professor Neel was one of the world’s leading human geneticists in the latter half of the twentieth century, but he had a case of foot-in-mouth disease at the end of his seminal paper. In a section titled “Some Eugenic Considerations” he made the precious remark that:
If...the mounting pressures of population numbers means an eventual decline in the standard of living with, in many parts of the world, a persistence or return to seasonal fluctuations in the availability of food, then efforts to preserve the diabetic genotype through this transient period of plenty are in the interests of mankind.
He concluded: “Here is a striking illustration of the need for caution in approaching what at first glance seem to be ‘obvious’ eugenic considerations!” We might conclude that here is a striking illustration of the need for prominent geneticists to shut the heck up when they start speculating about and proposing genetic remedies for the future of mankind.
He should have known better. Food scarcity is going to cause a lot more pressing problems for humankind than preserving genetic variants, as the situations in places such as the Sudan and Rwanda indicate. Just how he envisaged the West implementing this engineering of thrifty gene prevalence in developing countries is disturbing to contemplate. We should give him the benefit of the doubt, though: Perhaps he was writing in late May as gray winter in Ann Arbor was finishing its eighth straight month, a stress mere mortals cannot be reasonably expected to handle with grace.
Lamentable eugenics does not however invalidate the thrifty genes argument, which seems to be thriving. Humans have a magnetic attraction to simple ideas that seem to have great explanatory power without a whole lot of tangible evidence in their favor. All the bits and pieces to the argument seem to make sense. All three premises of the theory of natural selection are met: There is variation among individuals, this is somewhat heritable, and there is differential survival related to the trait. Put them together, and you arrive at the conclusion that genetic variants predisposing to obesity must have been positively selected in modern human history.
Alas, close reading uncovers three features that a lot of evolutionary medicine arguments have in common: hyperadaptationism, the hereditarian fallacy, and sloppy quantitative reasoning. Hyperadaptationism is the notion that if an organism exhibits a trait, the trait must have evolved for that purpose. We now know that often traits evolve as a side effect of something else. Rather than assuming adaptation, evolutionary geneticists now accept a high burden of proof in establishing it.
The hereditarian fallacy is the idea that if there is genetic variation for a trait, and the frequency of the trait varies among populations, then genetics must account for the differences between the populations. It is that old racist curveball that keeps coming up in discussions of American IQ, raised here as evidence that selection accounts for the high prevalence of obesity among Pacific Islanders. None other than Jared Diamond, author of two brilliant books about the rise and fall of civilizations (Guns, Germs and Steel and Collapse), has even gone so far as to suggest that Polynesian founders were likely to have been especially enriched for thrifty genes; otherwise they would not have survived the voyages across the Pacific. But Captain Bligh and his Bounty men were lean exemplars of the British Navy and survived their 49-day postmutiny ordeal, and there are numerous modern examples of survivors of calamities twice this long, so such inference of strong selection must be taken with a grain of salt.
Quantitative reasoning needs to be carefully applied to any proposal for adaptation, and on close examination the thrifty genes hypothesis falls apart. We can calculate what sort of benefit a genetic difference must afford for it to arise in an individual and be found in one-fifth of people a few thousand years later. It is in the vicinity of five percent, meaning that people with the advantageous allele have a 1 in 20 better chance of having children than those without it.
However, a thrifty genes critic from Aberdeen, John Speakman, has demonstrated that famine survival is highly unlikely to be anywhere near this strong. Basically, famines only rarely kill more than a few percent of the population in any generation, when they do they preferentially affect the very young and very old and so have little effect on genetic transmission. In any case, usually people die of infectious disease, not starvation. So we should be skeptical about claims that diabetes is the result of genes for weight gain being advantageous to nomads and pastoralists but bad for modern urbanites.
One gene actually does fit the thrifty criteria, but it is responsible for lactose tolerance, not weight gain. Prior to the domestication of goats and cattle, humans would not have drunk milk after weaning, or eaten cheese, yogurt, or any of those other great dairy products, as adults. The reason why many of us can digest lactose as adults is because we’ve converted a baby gene into an adult one. The enzyme lactase-phlorizin hydrolase (LPH or lactase for short) is used in the small intestine to allow babies to turn lactose into glucose and other sugars. On at least two different occasions in the past 10,000 years, once in Europe and once in East Africa, novel mutations have arisen that allow the gene to keep on working in adults. Population geneticists can estimate that possessing the mutation would have conferred something like a five percent fitness advantage to individuals in early pastoral societies that came quickly to depend on milk as a staple.
By contrast, the most compelling argument that diabetes susceptibility didn’t get into the gene pool a few tens of thousand years ago actually comes from the genes themselves. In just about every case, it turns out that the ancestral allele, the one we share with chimpanzees and other primates, is the
one that is more risky. In other words, the protective types are the ones that have been increasing in frequency for the last few millennia. Without these, obesity and diabetes would be even more prevalent than they are. Frankly, we have no idea yet what forces, if any, are favoring them, but we can thank our lucky stars that they are around.
Disequilibrium and Metabolic Syndrome
A possible explanation for the increase in resistance factors is that there has been selection against the disease of diabetes. This seems unlikely since until recently it was not prevalent enough to affect reproduction with the necessary selection intensity. Perhaps there is selection involving a more subtle aspect of diet, and the impact of diabetes incidence is a side effect of this. Alternatively, the regulation of metabolism may be so complicated, involving hundreds of genes each with a variety of alleles, that it is inevitable that the network malfunctions in some people after a few key components of the system change.
One of the central ideas of this book is that humans are now outside their normal buffering zone. Instead of thinking of some optimal body mass or blood glucose level, we ought to be thinking of well-buffered systems of interactions that keep weight and energy within a critical range in the face of a constantly shifting environment. On this view it is absolutely critical that there be genetic variation to absorb the environmental insults that any organism faces: seasonal food sources, droughts, famines, illness, pregnancy, and growing old. They all put pressure on our metabolism, and it pays to have a flexible response.