The Truth About The Genetics Of Obesity.

Evolutionary selection favored those who became fat easily. That's the essence of the "thrifty gene hypothesis". It's like Madonna. On the wrong side of 50, and ripe to be dethroned by something with greater sex appeal. In this case the contender's name is the "drifty gene hypothesis". Here is why you shouldn't be too dazzled about it. [tweet this].    

Exactly 50 years ago, Neel suggested that the high rate of diabetes in our society is the result of  evolutionary selection which favored those of our ancestors whose genes made them store fat more efficiently during periods of food abundance [1]. It's such a marvelously simple explanation that it doesn't take the brains of an Einstein to chatter about it at any dinner party where one wants to be remembered as quite the hobby geneticist. But to every party there is a party pooper. In this case two of them. John R. Speakman and Klaas R. Westerterp are telling us that the high prevalence rate of obesity and diabetes actually disproves the thrifty gene hypothesis [2].

In a nutshell their argument goes like this: our human and hominin ancestors have gone through so many feast and famine cycles over the past 2 million years, that, if it was for genetic selection, we should by now all be carriers of the genes that made caveman survive and modern man fat and diabetic. Since this is clearly not the case, the TGH can't be correct. 

I'm a sucker for theories which challenge common wisdoms, so I enthusiastically read the authors' arguments. Now, let's see how this enthusiasm evaporated.

To a considerable extent, obesity is determined by genes. If you want to put a number on it, genetic factors explain about 60% of the variance in obesity metrics, such as the body mass index (BMI). That's the numbers we are getting from studies, which compare such metrics between identical twins and other sibling types [3]. Just as an aside: When you consider genes as the one condition which you can't change, 60% heritability still leaves a lot of wiggling room for you to fashion your own fate. That's good because obesity comes with a host of nasty diseases, none of which makes your life longer or more pleasant. Think diabetes. Of course, you know all that, and it is not really our subject here. We want to know why there is such a high prevalence of obesity prone people.

To answer this question Speakman and Westerterp compiled some insights from genetics and put them through a mathematical blender. That sounds far simpler than it really was. For that blender to give you an intelligent answer you need to feed it with intelligent data. Otherwise it's the old nonsense-in-nonsense-out" paradigm. In our case at hand there are three data segments which need to be considered. 

First, there is biology: what happens to a human organism when it is exposed to fasting? How long will it survive?

Second, there is genetics: what do we know about those 60% of genetic causes? Are they concentrated in a handful of genes, or are they spread over hundreds? And what do we know about the mutation rates of genes?  Obviously, the more causative genes, and the smaller the mutation rates the longer it will take for any genetic mutation (or allele) to become fixed in the genetic pool. "Fixed" being geneticist speak for "(almost) everybody has it".

Third, there is evolution & environment: how often did famines happen, and how many of our ancestors were affected by them at any one event?

Get the figures slightly wrong in any of those three segments and your result will be off track. And so will be your conclusions.  
To get intelligent data, the two authors first went through an exemplary exercise of modeling what happens to a human organism when it is exposed to a zero-intake famine. That's not as straight forward as you might think, because our metabolism goes through at least three distinct phases when fasting in the extreme. These three phases are determined by our organism's way of storing energy reserves. 
First, there is glucose, the building block of virtually all carbohydrates in our food. While our brain thrives almost exclusively on glucose, the body's glucose stores are remarkably small. Glucose is predominantly stored in the form of glycogen in muscle and liver tissue. It is these reserves which are tapped first, and they are typically depleted within 24 hours. If you are a marathon runner you do this depletion business a lot faster, say after 20 miles or so. 

Since your brain still needs glucose, your body then starts to produce its own. Largely from lactate and glycerol, a component of fat. Which brings us to the second phase, where the body metabolizes its fat reserves. But even fat reserves don't last forever. Once they are depleted, the body begins to cannibalize its protein. Actually, weight loss in phase 2 is never a pure loss fat only. Proteins are being burnt at the same time but a at a lesser rate, until fat reserves have been depleted. And that's where fasting gets critical, because to your body, burning proteins for energy is like burning banknotes for warming your house: you go broke in no time. And "broke" means "dead" to your body. 

Since time to death is a critical element in the mathematical model, the authors went through an exemplary effort of mapping the course from fully fed to fully dead. Interestingly, everybody reacts differently to this fasting business. Some people survive longer than others, even when they have the same BMI to start with. That's why Speakman and Westerterp applied three different models to predict survival time, all models representing those known different ways of adapting to starvation. For a severely obese 1.64 m tall female weighing 100 kg, the models predicted a survival time of 249-289 days. Imagine, that's about 8-9 months with no food at all. 

Onto the genetics assumptions. The one thing we know for sure is that obesity is a multi-gene condition. Very multi-gene in fact, because genome-wide association studies (GWAS) have thrown up about 30 odd genes with a combined effect of explaining only 7% of those 60% of weight variance. So, we are assuming that the unexplained difference resides within another 200 or so genes, which we haven't even identified yet. Speakman translated this knowledge into an assumption of each individual gene having a net effect on fat storage of about 80g. That is, a carrier of a gene's "thrifty mutation" (or allele) would store 80g more fat than his peer with the "lean" version of the gene, with those 80g, translating into a 0.25% better chance of surviving a famine. With these assumptions the authors could then calculate how many famines it would take to weed out the unlucky ones whose "lean" genes didn't give them the 80g advantage. That calculation in itself is no rocket science. The authors took a given population size of 5 million people, exposed them to a virtual famine, after which the population had been appropriately decimated, and the percentage of "thrifty gene" carriers among the survivors had increased. They all mated happily after that until the population again reached 5 million. Then the next virtual famine struck, and so on. 

How many famines would it take to eliminate the lean gene from the gene pool? Under the authors' assumptions about 6000 famine events.  
They then made their final assumption: one famine happening every 150 years. That's 900000 years altogether for those 6000 famine events. Their conclusion: if the thrifty gene hypothesis and its assumption of selection pressure from catastrophic events was correct, we all should be obese today. Since we are not, the TGH is false. 

The alternative explanation, which the authors offer is a "drifty gene hypothesis" as opposed to the thrifty version. "Drifty" referring to genetic drift, meaning that mutations of the genes, which regulate fat storage were never really subject to selection pressure, and what we see today is simply the result of a natural drift of genetic mutations over the eons of human existence. 

The authors argue further that excessive fat storage was a distinct disadvantage for our earliest hominin ancestors, for reasons of predation. Think of it like that: while neither a fat man nor a lean man can outrun a saber toothed tiger, it's enough for the lean guy to run just a little faster than his fat bro'. Call it a stone-age version of the "first come, first serve" principle, at least from the tiger's perspective: the first man I get is the first man to serve me as breakfast. 

The authors then suggest that once our ancestors discovered fire and spears and other things which placed them on top of the food chain, the selective pressure for the lean gene had vanished. Its thrifty sibling started to flourish, not because it was favored by famine-based selection pressure, but simply because man had taken tiger and co. out of the equation, and with it the selective pressure to NOT get fat. During those zillions of generations which separate the man-known-for-throwing-spears from the man-known-for-throwing-tantrum-when-the-iphone-doesn't-work, those 200 odd genes accumulated just enough mutations for many, but not all, of us to become obese and diabetic. 

Up to this point one might buy into Speakman's and Westerterp's story. But here is the twist:  

Speakman has written about the subject before. With a different tagline. In his 2006 paper he suggested that the selection pressure of famines in human history was too small to have caused the effects attributed to it by the thrifty gene hypothesis [4]. According to that paper, famines with severe mortality rates were rare and, most tellingly, a phenomenon of agricultural societies. 

Indeed, the consensus view on famines in pre-agricultural vs. agricultural societies is that our hunter/gatherer ancestors were better fed and better protected against famines than their agricultural descendents. The hunter simply doesn't depend on a crop. Whereas when a crop fails, food shortage is inevitable for the agriculturalist. But even then, a true famine, where there is no food at all, typically requires a back-to-back failure of crops in consecutive years. And even then, as Speakman pointed out in his 2006 paper, mortality rates rarely exceeded 10% of the population, with those 10% coming almost exclusively from those who are either too young or too old to reproduce and thereby contribute to the gene pool after the famine is over. The author's message in 2006:  Genetic mutations towards thrifty genes didn't have sufficient advantage or time to spread. 

This little twist shows us that somebody is taking potshots at TGH: 

Shot 1 (2006): Famines haven't been with us for long enough nor with sufficient severity to have exerted the selective pressure on which the thrifty gene hypothesis rests. Ergo, TGH is wrong.
Shot 2: Famines were so numerous and severe during human history that their combined selective pressure on the thrifty genes was sufficient to have made them a fixture in EVERYBODY'S genetic make-up. Since this is not the case, the TGH is wrong.

Science shouldn't be about taking potshots. Science is about the testing of falsifiable hypotheses in reproducible experiments. A mathematical model, such as the one presented in Speakman's most recent paper does not qualify as such.  

Here is why: Given that mutations happen at the rate of 1.1 per 30-100 million base pairs, we all carry about 100 to 200 mutations in our DNA [5].  Not necessarily do those mutations affect actual genes coding for proteins. And if they do, most mutations confer a slight disadvantage, many have no effect on an organism's fitness, and only a few are favorable. Natural selection will weed out the deleterious ones, quickly fix the favorable ones and let the neutral ones accumulate at the given mutation rate. To complicate matters, all those processes happen at vastly different rates depending on the location on the DNA. That much we do know. What we don't know is how much these rates differ. We certainly can't know it for those genes, which we haven't even identified yet, as is the case for most of the hypothesized fat storage genes. That's why the mathematical model with which Speakman supports his argument against the validity of the thrifty gene hypothesis is in all likelihood not reflective of what has happened throughout evolution. Which means, it doesn't add any quantitative or objective evidence against the TGH. 

In my next post I will tell you why I believe that the entire discussion misses the point. What we really want to know now is how to help people avoid becoming fat and diabetic in the first place. Decoding the genome and its evolutionary history doesn't do that trick. Because genes do not make us fat and diabetic, genes make proteins, nothing else. One part of those proteins are the hormones. They drive our moods and emotions, our likes and our dislikes and, believe it or not, all our behaviors, from feeding to physical activity. For those latter two I have suggested an explanatory model in my dissertation thesis. 
This model tries not only to explain why we eat too much and move too little, despite having the best intentions to do otherwise, and while being aware of all the life threatening consequences. But, more importantly, without having to have a complete understanding of all those hormonal happenings, the model suggests a practical and testable solution to oppose those genetically encoded mechanisms for a longer and healthier life. Think of your car: You don't need to understand the mechanism of its gearbox to operate it for an optimal ride. 
Achieving the same thing with your life could turn out to be a gratifying pastime while my geneticist colleagues work on unraveling the enigma of the genetics of obesity. Whatever newer and sexier model they develop to explain the genetic origins of obesity, we might look at it like we look at Madonna and her variants: offering lots of entertainment value, but little of practical use. [tweet this].    

1. Neel JV: Diabetes mellitus: a "thrifty" genotype rendered detrimental by "progress"? Am J Hum Genet 1962, 14:353-362.
2. Speakman JR, Westerterp KR: A mathematical model of weight loss under total starvation and implications of the genetic architecture of the modern obesity epidemic for the thrifty-gene hypothesis. Disease models & mechanisms 2012.
3. Segal NL, Allison DB: Twins and virtual twins: bases of relative body weight revisited. Int J Obes Relat Metab Disord 2002, 26(4):437-441.
4. Speakman JR: Thrifty genes for obesity and the metabolic syndrome--time to call off the search? Diabetes & vascular disease research : official journal of the International Society of Diabetes and Vascular Disease 2006, 3(1):7-11.
5. Xue Y, Wang Q, Long Q, Ng BL, Swerdlow H, Burton J, Skuce C, Taylor R, Abdellah Z, Zhao Y et al: Human Y chromosome base-substitution mutation rate measured by direct sequencing in a deep-rooting pedigree. Curr Biol 2009, 19(17):1453-1457.

NEEL JV (1962). Diabetes mellitus: a "thrifty" genotype rendered detrimental by "progress"? American journal of human genetics, 14, 353-62 PMID: 13937884

Speakman JR, & Westerterp KR (2012). A mathematical model of weight loss under total starvation and implications of the genetic architecture of the modern obesity epidemic for the thrifty-gene hypothesis. Disease models & mechanisms PMID: 22864023

Segal NL, & Allison DB (2002). Twins and virtual twins: bases of relative body weight revisited. International journal of obesity and related metabolic disorders : journal of the International Association for the Study of Obesity, 26 (4), 437-41 PMID: 12075568

Speakman JR (2006). Thrifty genes for obesity and the metabolic syndrome--time to call off the search? Diabetes & vascular disease research : official journal of the International Society of Diabetes and Vascular Disease, 3 (1), 7-11 PMID: 16784175

Xue Y, Wang Q, Long Q, Ng BL, Swerdlow H, Burton J, Skuce C, Taylor R, Abdellah Z, Zhao Y, Asan, MacArthur DG, Quail MA, Carter NP, Yang H, & Tyler-Smith C (2009). Human Y chromosome base-substitution mutation rate measured by direct sequencing in a deep-rooting pedigree. Current biology : CB, 19 (17), 1453-7 PMID: 19716302