ON A VISIT TO THAILAND IN 2001, I SPENT THE FIRST DAY, while suffering the miasma of jet lag, at the National Museum complex in Bangkok. Though I am a museum lover, I predictably found it hard to pay much attention to the displays and especially to the captions. After a half-conscious hour or so, my stomach churning, I found a welcome and therapeutic diversion in people watching. Particularly engrossing were the multitude of local school children visiting the museum on field trips. Most were dressed in particularly drab school uniforms—khaki shirts and short pants, dark brown socks worn high, and lighter brown shoes. The color of their attire seemed somewhat incongruous in this tropical city, though the shorts were quite practical.
But more noteworthy was their behavior. No shouting or screaming or running wildly up and down the galleries. They were quite attentive to their instructors and could be efficiently mustered when it was time to move as a group. Even upon exiting the museum to the crowded and chaotic streets, they retained their single-file formation and poised demeanor.
My own son was ten at the time, so I could not help but notice the differences regarding his own school outings, which were always attended by multiple chaperones, who, as I can attest from my own experience, were invariably exhausted by the end of the day from their only partially successful attempts at maintaining order.
On subsequent trips to Thailand, I would spend a couple of days in Bangkok at the beginning and end of the trip continuing my informal study of the school kids while visiting various temples and palaces. My focus gradually shifted, though, from their behavior to their physical attributes. Most obvious was their height, which was much closer to American standards than their behavior. As is true throughout much of Asia, each generation since the last world war has experienced a size increase. Many twelve-and thirteen-year-old Thai kids are already taller than their grandparents and will soon exceed the height of their parents. There are obvious reasons for this, most notably enhanced nutrition, especially protein.
Weight-wise, however, the Thai youth were still lagging far behind their American counterparts. On my first visit, one boy had stood out. I’ll call him Paradorn. He was the only Thai kid in the museum who was clearly overweight. On subsequent visits I noticed more and more overweight boys (but not girls). By 2005 there seemed to be about one per class of about twenty-five kids. This is about the ratio I remember from my own grade school days in the 1960s: about one clearly overweight kid in each of my class pictures. That was much less than the obesity rate in my son’s classes (for both boys and girls).
Weight-wise, the Thai kids more closely resembled American kids of the 1960s than those of my son’s age. It is only recently that any preadolescent Thai kids have become overweight, and obesity seems to be restricted to the more urban areas; you will still have to search long and hard to find a fat Thai child in the Issan region, for example.
There is an obvious cause for this urban-rural weight difference in Thailand. Urban Thais are markedly more affluent that rural Thais; this discrepancy is the main cause of the recent political unrest. With affluence comes increased calories and often less physical activity—the standard formula for weight gain. But the source of the increased calories also matters, and here, too, urban Thais differ from their rural counterparts.
The diet of my friend Aniwat, though he lives in Bangkok, is typical of the rural village where he grew up, near Kaeng Krachan National Park in Petchburi Province: mostly a diverse range of fruits and vegetables, some of which he acquires at open-air markets, some he forages himself. On every walk in the forest, Wat would gather morning glory and other greens, as well as the fruits of countless species of eggplant, none of which looked much like the large, elongate, purple-skinned stuff that I once grew.
He eats much less animal-derived protein than a typical American, and virtually no beef. Chicken and fish are the primary animal proteins consumed by rural Thais, augmented by a diversity of creatures most Americans would not consider edible, notably a variety of insects, from grubs to cicadas and cockroaches. (For Wat, a green papaya salad without a cockroach is not a proper green papaya salad.) Wat consumes virtually no dairy products or processed foods, though he did acquire a taste for peanut butter during an extended visit to the United States. His desserts largely consist of processed fruit confections that are not particularly sweet. At age sixty plus, he retains the physique of the Muay Thai boxer he once was.
The diet of urban Thais, especially those recently arrived from the hinterlands, encompasses many of these elements—including the insects—but also a much greater amount of prepared foods, both Thai and increasingly European and American. The latter comes primarily in the form of fast-food chains—McDonald’s, KFC, and such—which provide the most efficient path to obesity known. There are some interesting peculiarities in Thai fast-food preferences. Thailand is one of the few places I have visited where there seemed to be more Kentucky Fried Chicken franchises than McDonald’s. In part this can be explained by the Thai preference for chicken over beef. But according to Wat, there is also the “fried factor.” Much traditional Thai food is fried, albeit in a wok and with much less oil. That predisposition toward fried foods would also explain the recent success of doughnut purveyors in urban Thailand.
Originally, I attributed the accelerating weight gain in affluent urban Thais to their Americanized eating habits, a trend that began when the United States used Thai air bases as staging areas for attacks on Vietnam, Laos, and Cambodia from the mid-1960s through the early 1970s. Thailand is also the place where physically and psychically exhausted soldiers went for R&R. Naturally, they were looking for American food and they got it. An increasing number of Thais acquired a taste for American food in some of its most unwholesome incarnations—and they gained weight. When you eat like an American, you get fat like an American. That all seemed pretty obvious. Subsequently though, I have come to understand that the obesity of neither Americans nor Thais, like Paradorn, is quite that simple.
Though it is certainly true that McDonald’s, KFC, and the like are major contributors to obesity in the United States, and more recently in places like Thailand, these fast-food outlets should be viewed more as a precipitating cause, the cause that tips a predisposed body over the edge. What is it that makes a body predisposed? The conventional answer is genes. Some individuals or ethnic groups are predisposed toward obesity because of their biological inheritance. In this chapter we will explore a different sort of predisposition: epigenetic. These epigenetic predispositions generally develop in the womb or in infancy.
Thrifty Genes?
Obesity per se is not a public health problem; it’s the bad things that obesity does to your physiology, notably the so-called metabolic syndrome, which is basically a problem in the way your body processes food and can lead to cardiovascular disease and diabetes. It was in an attempt to explain diabetes in particular that the “thrifty genes” hypothesis was proposed.1 James Neel noted in the early 1960s that when non-European populations were exposed to a Western diet, the incidence of diabetes (and obesity) skyrocketed to levels that were much higher even than those in the United States. He proposed, by way of explanation, that these populations had evolved in an environment of periodic famine in which individuals whose bodies were particularly efficient at turning calories into fat stores were at a selective advantage. These individuals could thrive during lean times because of their “thrifty genes,” but they become fat and diabetic in an environment where food is plentiful.
The thrifty-genes hypothesis was criticized on a number of grounds. Most damningly, there is no evidence that humans experienced periodic famine until the agricultural revolution 9,000 years ago.2 Neel himself soon abandoned the idea, but it survives to this day in mutated forms.3 The thrifty-genes hypothesis reflects a genocentric view of the obesity epidemic, a view that is also manifest in the search for obesity genes.4 There is no lack of candidate obesity genes, but none of the candidate genes, alone or combined, go very far toward explaining who gets fat and why.5
While the gene hunters were busy at their labors, other researchers approached the obesity issue from a different angle, based on the fact that Americans and many Europeans were getting fatter at an accelerating rate. It was hard to pin this obesity epidemic on obesity genes. Instead, it is widely recognized that the American diet was too much of a muchness, and that the physical activity of the average American was insufficient to burn all of those calories. This all seems obvious enough, conventional wisdom even.
The Western lifestyle certainly provided a better explanation for the fact that many non-Western peoples, like Inuits, Pacific Islanders, and urban Thais like Paradorn, experienced dramatic weight gain when exposed to McDonald’s and Kentucky Fried Chicken. They were just newly experiencing essentially what native-born Americans had been experiencing for decades. Whatever genetic differences may have existed among these groups, the signal was quite faint compared with that of diet and physical activity.
But what about the individual variation within these groups? Paradorn is still the exception among urban Thais. Moreover, not all Americans of European descent are overweight; in fact, there is tremendous variation in weight within this group. While most of this variation can be attributed to lifestyle factors, individuals with essentially the same eating and exercise habits can differ substantially in their weight. It is this variation that currently motivates most of those searching for obesity genes. The gene hunters’ reasoning goes something like this: Even if we account for diet, exercise, and so on, there still remains the fact that people vary in weight and from an early age; ergo, they are born with different predispositions to obesity; ergo, they are genetically different in their propensity to gain weight. On this view, Paradorn got dealt a bad genetic hand.
But this seemingly obvious chain of logical inferences is only as strong as its weakest link. And its weakest link is the one that connects “ergo, we are born with different predispositions to obesity” to “ergo, we have genetic differences.” That is simply a non sequitur, one you often find in discussions of gay genes, intelligence genes, and genes for a host of other human attributes. We cannot infer that there is an obesity gene or a gay gene or any other sort of gene from the fact that we are born with a predisposition toward obesity, homosexuality, or whatever. As we saw in the previous chapter, many of us are born with a predisposition toward a hyperactive stress response that is not there at conception. This predisposition develops in the womb as a result of epigenetic processes. The same may be true of obesity and its associated disorders in people like Paradorn.
From Thrifty Genes to Thrifty Phenotypes
Of course, our genes affect our weight. At issue is whether, in coming to grips with the obesity epidemic, our research dollars are best spent searching for particular mutant alleles at particular genetic loci. Obesity is not a simple trait like Huntington’s disease, which can be traced to a mutation in a single gene. Rather, obesity is influenced by any gene that affects the way we process food, of which there could be hundreds, each having a fairly small effect.6 So the gene hunters’ task is to identify variant or mutant alleles among this large number of genes that may contribute slightly to obesity. This is a huge task under any circumstances, the payoff of which is uncertain at best.
Meanwhile, a quite different research program has proven productive. The objective in this alternative approach is to identify events in the womb that might result in obesity. It has long been common knowledge that events in the womb affect the health of infants, hence the emphasis on prenatal care. But this research greatly extended the range of conditions that we know are more or less directly affected by the in utero environment and their duration, not least obesity and other elements of the metabolic syndrome.
As was evident from the Dutch famine, one indicator of the quality of the womb environment is birth weight. Low birth weight generally indicates poor conditions in the womb. Not surprisingly, neonates born with a below-average weight are subject to a host of health disorders during early infancy. What is surprising is that these individuals continue to be less healthy throughout their lives and have a lower life expectancy as a result.7 As we saw in Chapter 1, one of the adverse consequences of low birth weight is obesity in adulthood. Why would a small neonate become an overweight adult? The current consensus is that this association results from a process called fetal programming,8 much of which occurs in the womb.
James Barker proposed that when the fetus receives insufficient nutrition through the placenta, it becomes programmed in the womb for a thrifty phenotype.9 As was proposed for the thrifty-genes hypothesis, those with a thrifty phenotype have a more efficient metabolism than babies born at a normal birth weight. But the thrifty phenotype can result from diverse genetic backgrounds and without the aid of specific obesity genes. It is, rather, simply a function of the intrauterine environment.
The thrifty phenotype works out well in traditional non-Western cultures where the postnatal environment is often one of scarcity. In those cases, the prenatal environment predicts the postnatal environment in an adaptive way. Problems arise, however, if the postnatal environment is enriched food-wise relative to the prenatal environment. When this mismatch occurs, thrifty phenotypes result in obesity and its consequences. The Barker hypothesis nicely accounts for the correlation between low birth weight and adult obesity and has been bolstered by much, though certainly not all, subsequent research.10
But what is the nature of this so-called programming?
Barker himself has little interest in the mechanism through which this effect of the uterine environment occurs. Others, though, have pursued mechanistic investigations of thrifty phenotypes, and, as is usually the case, the initial studies have been conducted on nonhuman mammals—especially mice, rats, and sheep. For studies of this sort, biologists often look for changes in the expression of particular genes, as evidenced either by the abundance of their associated proteins (translation products) or messenger RNA (transcription products). In this case, they are looking for long-term differences in gene expression between those individuals born underweight and those born at a normal birth weight. And indeed, scientists have found a number of differences in gene expression associated with birth weight.11
Many of these gene expression differences are tissue specific. For example, one gene may be more (or less) active in the liver of a small neonate, while another gene may be more (or less) active in the adipose (fat cells) tissue. Other genes, notably the glucocorticoid receptor gene (GR; see previous chapter), show different expression patterns in many tissues, including several parts of the brain, the liver, the adrenal gland, the heart, and the kidney.12 These different gene expression patterns often persist into adulthood and old age.
The products of many of the genes that vary in expression according to the uterine environment are transcription factors, each of which influences the expression of many other genes. The net result is a host of long-term differences in gene expression in many tissues that relate to the intrauterine environment. The challenge is to sort out the cause-and-effect relationships in these gene expression patterns and to causally connect them to events in the womb. Because these gene expression differences are long-term, researchers have begun to search for epigenetic signals.
Methylation Patterns Vary with Diet
The expression of one family of genes, in particular, seems to be directly connected to nutrient availability in the womb: the genes that code for DNA methyltransferases (Dnmt).13 Dnmt promotes and maintains methylation in genes subject to epigenetic regulation. Hence, when Dnmt levels are high, these genes tend to be turned off, or silenced. Conversely, low Dnmt levels and the consequent reduced methylation result in increased expression of those genes.
In rats fed a protein-restricted diet during pregnancy, Dnmt expression is low.14 Low Dnmt levels mean that some genes that should be methylated are not. Because these genes are not methylated, we would expect them to be more active than they would normally be, producing more of whatever it is they produce. One of the genes methylated by Dnmt is GR.15 This gene is another example of tissue-specific (and hence context-sensitive) gene regulation. Recall that in the hippocampus, NGF binds to GR. In the liver, on the other hand, it is Dnmt that binds to GR, thereby deactivating it. In rats fed a low-protein diet, Dnmt levels drop, resulting in lower methylation of GR and hence increased GR expression. Just as lower-than-normal levels of GR expression in the hippocampus cause problems such as a hypersensitive stress response, higher-than-normal levels of GR expression in the liver also cause problems. The higher-than-normal levels of GR in the liver and other tissues cause these tissues to be too sensitive to stress hormones. The long-term effect is an increased risk of diabetes, obesity, and other elements of the metabolic syndrome.16
The link between GR expression and the metabolic syndrome has led some to speculate that ultimately, the low nutrient levels in the womb are just another sort of stressor, the effects of which are mediated by the stress response.17 If so, other forms of stress in utero that result in high cortisol levels should mimic the effects of bad nutrition. In fact, there is evidence that social stress experienced by the mother increases the chances that her offspring will develop the metabolic syndrome. A fetus subject to both kinds of stress, as occurred during the Dutch famine, would be especially vulnerable to the metabolic syndrome.
Maternal stress adds another potential dimension to the obesity epidemic. Some researchers have proposed that increased obesity levels are in part attributable to the highly stressful Western way of life, especially in urban settings.18 This stress is transmitted to the fetus through the placenta, resulting in obesity, diabetes, and so on. Thus it is possible that Paradorn is overweight because of the stress his mother was under while pregnant. That stress could have a number of sources, of course—poverty, for example. Or it may have been her social environment. Perhaps she moved away from her family in the rural Issan region for a better life in bustling Bangkok. That not only would amount to severe culture shock; it would have left her isolated, without the social support of traditional rural Thai families. Whatever the source of her stress, it may have affected both Paradorn’s weight and stress response.
It is worth noting as well that too much of a good thing in the womb can also be a stressor. A fetus that gets too many calories also has an elevated stress response and is more prone to obesity.19 Perhaps it is for this reason that overweight neonates, as well as their underweight counterparts, tend to become overweight adults. For Paradorn, that would mean a quite different explanation vis-à-vis his mother as it relates to his plight. Having moved to the city, she abandoned her traditional diet for McDonald’s and KFC, and her cravings for these foods only increased during pregnancy. The effect of these excessive calories on Paradorn—whether directly through its effect on his metabolic rheostat, or indirectly through his stress response—was to predispose him toward obesity. This purely speculative scenario is intended solely to provide some sense of the potential diversity of environmentally induced epigenetic changes relevant to obesity.
From DNA to Histones
So far, we have only considered one avenue by which methylation exerts its epigenetic effect on gene activity: through the binding of methyl groups to or near a particular gene. But many of the effects of methylation on gene expression are more indirect; these indirect effects come by way of a class of proteins called histones.20 There is evidence that diet-induced histone modifications in the fetus are a factor in the metabolic syndrome.21
When I first learned about DNA in high school biology, my mental picture of things at the molecular level had the naked double helices sort of floating around the nucleus, always at the ready for protein synthesis. It was only later, and with some mental effort, that I came to understand that DNA is far from naked but, rather, intimately entangled with proteins. It is this DNA-protein complex that constitutes chromosomes. DNA and protein are so entangled that, as mentioned in Chapter 2, after the discovery of chromosomes, it was quite unclear whether it was the DNA or the proteins that were the genetic material. Naturally, once DNA was proven to be the genetic material, the protein component of chromosomes was largely ignored.
The proteins in chromosomes were thought to function primarily in efficiently packaging inactive DNA in a condensed state, one that occupied far less space than that of the expanded, active form of DNA—sort of like archiving a computer file. It is only recently, primarily as a result of epigenetic research, that a quite different view of these chromosomal proteins has emerged. On this new view, histones are much more dynamic than previously supposed and they play an important role in regulating gene expression.
In general, histones are less tightly bound to the DNA where the genes are actively engaged in protein synthesis, and more tightly bound to the DNA where genes are inactive. The degree to which a histone is bound to the DNA is a function of epigenetic processes. These histone-related epigenetic processes involve various types of biochemical alterations to the histone, one of which is methylation.22 As in DNA methylation, histone methylation usually (but not always) blocks gene expression. And as with DNA methylation, histone methylation is passed, intact, from a cell to its descendents.
Rats that experience low-protein diets during development evidence histone modifications near the GR gene that cause it to be expressed at higher-than-normal levels.23 Whether these histone-based alterations precede or follow alterations in the expression of this gene caused by DNA methylation is unclear. DNA methylation and histone methylation are often coordinated. For the purpose of fine-tuned therapeutic treatments, it is important to know, in a precise way, just how these two forms of methylation are coordinated. Currently, it is known that folic acid and other key nutrients (for example, zinc, vitamin B12, and choline) can remedy somewhat the effects of poor nutrition in the womb through their epigenetic effects.24
Folic acid was first used as a prophylactic against spina bifida and other neural tube defects. It has proven quite effective in that regard when consumed by a mother-to-be in the first trimester. This effect of folic acid occurs via epigenetic modifications of certain key genes in neural development. Subsequently, it was discovered that folic acid has other epigenetic effects during development, some of which can meliorate the metabolic syndrome.25 These epigenetic effects of folic acid extend well beyond birth, perhaps to adulthood. For this reason, most food manufacturers fortify all grain products from cereal to flour with extra folic acid (which is usually obtained from fruits and vegetables). This is perhaps the first application of nutritional epigenetics.
But there are reasons for caution with regard to this uncontrolled experiment. Given its potential epigenetic potency, too much folic acid could be a bad thing. Some suspect a link between high levels of folic acid and autism, for example, based on epigenetic considerations.26 The putative increase in autism roughly corresponds to the time that folic acid became widely added to our foodstuffs and consumed in high doses by pregnant women. Moreover, epigenetic differences have been identified in some diagnosed with autism.27
At this point, the folic acid–autism link is almost purely conjecture. But there can be no doubt that nutritional epigenetics has a bright future, both as prophylaxis and therapy. As prophylaxis the big payoff will come when scientists can influence the “fetal programming” of obesity, diabetes, and other conditions, in people like Paradorn, through nutritional silver bullets, carefully timed and targeted. Therapeutically, the payoff will come from diets formulated specifically for those either susceptible to or experiencing these ailments—from childhood on. Both the prophylactic and therapeutic potential of nutritional epigenetics extends to many other ailments as well, such as cancer, which I will discuss later in the book.
What Predisposed Paradorn?
We have explored several possible explanations for Paradorn’s weight; these can be divided into two broad categories: genetic and epigenetic. They are not mutually exclusive. It is possible that Paradorn represents a rare combination of thrifty and obesity genes, which caused problems given his food-enriched environment. Alternatively, Paradorn’s predisposition may have originated in the womb or his early postnatal environment. If so, his predisposition would be largely epigenetic. I have discussed one possible epigenetic mechanism involving Dnmt and the GR gene, as they relate to nutritional factors and stress. The epigenetic and genetic explanations both involve genes but in fundamentally different ways. Genetic explanations for Paradorn’s predisposition require sequence variation, that is, variation in the alleles at a particular genetic locus. Such variation is immune to environmental influence except via mutation. Epigenetic explanations for Paradorn’s plight, on the other hand, involve variation in chemical attachments, either to key genes or to adjacent histones, which can be quite sensitive to the external environment.
One reason genetic explanations for obesity continue to garner so much attention is the observation that obesity runs in families. Epigenetic processes begin and end in a single lifetime. Or so it was thought. Recently, it has become apparent that epigenetic processes, including those involved in obesity, can be transgenerational as well. That is the subject of the next chapter.