by Greg Gibson
There is one major susceptibility factor, and that is a gene called ApoE. It is the only gene known beyond a shadow of a doubt to affect late onset Alzheimer’s disease. Your personal risk changes more than twentyfold, depending on which alleles you have. But even this is not sufficient to predict the course of disease, so physicians are loathe to even try to do so. In fact, ApoE probably has more of an influence on age of onset than whether you will get Alzheimer’s disease at all—namely, whether it is likely to come on in your sixties, seventies, or eighties.
ApoE is short for apolipoprotein E. That’s about where I turned off in college biochemistry as well; suffice it to say that it has an important role in what happens to your cholesterol. That it is a major player in the etiology of Alzheimer’s disease came as a major surprise, since until this discovery it was being investigated for its role in clogging up the arteries. Some investigators suggest that it is also involved in protection from viruses and possibly impacts fertility as well. All these functions are likely to have shaped the distribution of variation considerably more than the gene’s role in dementia late in life.
There are three common alleles, which we will call *E2, *E3, and *E4. If you had your druthers, you’d ask to have two copies of the *E2 allele, please, as it seems to be associated with lower incidence of heart disease as well as AD. However, few people are so lucky. Worldwide, between one half and two-thirds of us instead have two copies of the *E3 allele, which is fine. It seems to be neutral. Depending on where you live, with no easily described global patterns, between 10 and 20 percent of us have at least one copy of *E4, and just a few percent have two copies of it. For each additional copy of *E4, your chance of getting Alzheimer’s increases enough to make you worry if you start forgetting your keys, but not enough to be sure that it’s going to be a problem.
Now, if you’re wondering why anyone would have *E4 if it is so closely associated with disease, think again. It turns out that *E4 is the ancestral allele: Chimpanzees have it, and *E2 and *E3 are on their way to becoming the standard human condition. So, if anything, we are evolving resistance to Alzheimer’s disease, though with some setbacks in places such as Scandinavia, South Africa, and New Guinea where *E4 remains common. Across central Africa, more than a quarter of the alleles are *E4, but the good news there is that early reports suggest it is not associated with Alzheimer’s. If substantiated and combined with the fact that *E4 does increase LOAD susceptibility in African Americans, these results imply that our modern diet and lifestyle are interacting with ApoE to elevate the risk.
If you’re wondering what a protein that binds fats for a living is doing in the brain, the answer seems to be that it is involved in the cleavage or transport of APP into or out of neurons. In other words, it is not exactly clear, but gamma secretase makes Aβ at the surface of cells, right in the midst of the lipids that form the walls of the neurons, in a place where ApoE likes to spend some of its time. So it seems to have a direct role in pathogenesis, in addition to any indirect linkage between heart disease and aging.
Just Growing Old
Why do we grow old in general? For much the same reason that blue jeans and bicycles are disposable: It is a corrosive world out there. In time, chromosomes fray like downtrodden cuffs, and proteins oxidize like the metal of handlebars and spokes. But this facile answer belies the existence of extraordinary variability in the rate at which organisms age, much of this demonstrably genetic in origin.
As a species, humans are remarkably well off as far as aging is concerned. Our 80 years double the life span of our closest relative, the chimpanzee, and among mammals only elephants come close. That alligators are granted almost 70 years on Earth seems unjust, but swans make up for this by exceeding 100 years, and if you want a companion for life, then getting a parrot the day you are born is not a bad idea. Among dogs, small breeds such as terriers and miniatures typically surpass 14 years, but larger ones and bulldogs are lucky to make seven. In all these species, aging appears to follow a similar course, even in mice that live a scant four years.
We know that genes affect how long we live, because of mutant strains of worms, flies, and yeast that live much longer than usual. Remarkably, a common set of genes keeps emerging from studies of these model organisms, and it is widely believed that the biochemistry of aging is similar in mammals. One pathway is related to insulin signaling, which is best kept as low as possible throughout life. Similarly, and probably not coincidentally, dietary restriction (in essence, reducing how much you eat) is the most sure-fire way to live a long and healthy life. Another reduces the damage caused by “free radicals” generated by cellular energy factories over a lifetime: superoxide dismutase, or SOD, is basically rust proofing for the inside of cells. Then there are the Sirtuins, whose function remains obscure, except that they are the main target of resveratrol, the ingredient in red wine (and lately dietary supplements) that is supposed to make a glass of Shiraz each evening good for your health.
Step back far enough, and each of these pathways converges on a key gene called FOXO, which we can think of as the foreman of stress. When FOXO is active, it ensures that all kinds of stress responses are active. Heat shock, oxidation, xenobiotics, microbial immunity, DNA damage, metabolic dysfunction—protection from each of these keeps us from growing old prematurely. Some members of something called the Hormesis Society even go so far as to advocate taking a little bit of low-dose poison every now and then to keep FOXO on its toes, as it were.
Well, people who live long lives aren’t mutant in the same way that long-lived methuselah flies, age-1 worms, or Sir2 yeast are. It is likely, though, that variation in the hundreds if not thousands of genes that engage in these types of activity is contributing to variation in human longevity. Further, they each interact in a small way with ApoE to influence the onset of Alzheimer’s disease too. Crucially, different organ systems—muscle, liver, kidney, brain—all age at slightly different rates, in part because of the way they are wired genetically. Alzheimer’s is premature aging of the brain relative to the rest of the body, in the same way that muscle wasting and heart disease are premature aging of those tissues. People get old like the houses they live in, bit by bit.
We may never identify precisely which sequences of the DNA code are generically involved in aging, because their effects are too small, particularly since chance also plays such a huge role in aging. Some will be common variants, some rare; some will be prevalent in Sardinia and others in Alaska; some will be worse for men than women. Most will be present in the gene pool for reasons that have little to do with determining how long we live.
Evolutionary biologists like to theorize about aging as a necessary counterpoint to having babies. They do so because the existence of genetic variation for longevity as well as profound differences in life span among species implies that the rate at which organisms age is affected by evolution, and by the strategies adopted to ensure survival in different environments. Some species are so extreme about this that they basically all die immediately after spawning or otherwise seeding: Bamboo plants live for 100 years, and then commit suicide as soon as they flower. This trade-off between reproduction and existence also enters into the human equation: Having many children early is not generally conducive to a long life for their mother.
These days, two types of evolutionary theory predominate. The “don’t worry about it” theory, technically known as mutation accumulation, hypothesizes that mutations that act late in life have less of an effect on reproduction, so are more likely to build up in the gene pool. The “pay later” theory, technically known as antagonistic pleiotropy, hypothesizes that there are trade-offs between the good that mutations do early and the bad that mutations do late in life. Some aging-promoting mutations are actually thought to be beneficial to the species. The theories are not mutually exclusive, and both mechanisms are probably in play.
Both are predicated on the idea that natural selection favors alleles transmitted from generation to generati
on. Since the likelihood of death by predation and accident increases with age, a premium is placed on having children early, at least so long as those children can be raised successfully. In the don’t-worry-about-it scenario, mutations that don’t have an effect until after reproduction aren’t going to have as big an effect on genetic fitness as ones that affect a younger organism. We have already seen that more mutations happen every generation than evolution can possibly cope with, so it follows that deleterious alleles that act late in life are not weeded out as efficiently as ones that affect children. Natural selection literally doesn’t worry about them as much as it does about variation that acts early in life. Consequently, Alzheimer’s is more common than progeria, that nasty disease where 12-year-olds are physiologically in their seventies.
The pay-later model differs in that it postulates that there will sometimes be an advantage to mutations that are harmful late in life—not because there is anything to be gained from killing off the elderly, but because these mutations do other things that increase the survival and fertility of the young. For example, accelerating the onset of menstruation allows for earlier childbirth, but increases lifetime estrogen levels and hence the risk of breast cancer.
To be honest, I am not just on the fence about which of these theories is more plausible, but actually somewhat skeptical of their relevance to the human condition. Modern humans live on average five times longer than they have throughout most of the history of the species. Paleontologists estimate that fewer than half of all children in the Old Stone Age lived past age 10, and life expectancy even in the Middle Ages was less than 30. Just in the past few generations substantial decreases in child and adult mortality have occurred, and immigrants to developed countries can generally expect to live much longer than their parents.
So the conditions that pertain to each of the other common diseases that we have considered also exist here. Homo sapiens don’t reproduce until they are twice as old as other higher primates, and they are capable of living twice as long. Clearly there has been dramatic evolution of our genetic capacity for long life, too rapid for us to expect that we are at any sort of equilibrium. Correspondingly, cultural practices have dramatically extended human life span very recently. Together these factors ensure that a large fraction of those who reach their mid-sixties are no longer sheltered from the ravages of dementia.
8. Genetic normality
height and weight How tall and heavy we are is a function of thousands of genes, each with polymorphisms that have very small effects.
pigmentation A handful of genes suggest red hair and blue eyes, but on the whole human pigmentation is as genetically complex as any other trait.
the God gene Be wary of suggestions that God is in your genes, but don’t be surprised if spirituality is.
a few words about IQ Why it is wrong to assume that genetic variation within groups implies that genes guide differences between them.
on being human Selection and drift have shaped the evolution of the entire human genome, and it is an ongoing process.
the adolescent genome revisited Fact and fiction in understanding the origins of disease.
Height and Weight
Imagine that your child is asked to bring pictures of a half dozen athletes to school for a show-and-tell on the theme of human diversity. Whom would she choose? Shaquille O’Neal and Nadia Comaneci perhaps: a behemoth of a basketball center at 7 feet, 325 pounds of pure athleticism, and the diminutive gymnast who achieved perfection on the uneven bars at the Montreal Olympics. An equally odd couple would be Takanohana Koji, one of the great sumo wrestlers of our time, and Maria Sharapova, Wimbledon champion with the looks of a supermodel. Throw in El Guerrouj, the great Moroccan middle-distance runner, and the Aboriginal quarter-miler Kathy Freeman, and you have a pretty full coverage of human variation. We’re not quite as diverse as Chihuahuas and Bernese Mountain Dogs, or Brussels sprouts and broccoli for that matter, but our genetic potential is unarguably diverse.
Another perspective on this is provided by a marvelous YouTube video clip, “Women in Film.” As photographs of 77 Hollywood divas morph into one another—Bette Davis into Greta Garbo, Nicole Kidman into Catherine Zeta-Jones—you simultaneously feel how different and how similar people’s faces are. It really is amazing how sometimes two completely different people can look alike. The guy who bought your old bicycle could have been the kid across the street when you were growing up, even though one is Asian and the other of American Indian extraction. Apparently Charlie Chaplin once entered a Charlie Chaplin look-alike contest and came in third!
What do our genes have to do with normal human variation, and do they tell us anything about the origins of disease?
Let’s start with height, because this is possibly the most heritable of all human attributes and one of the most extensively studied. After factoring in generational effects, the correlation between children and their parents for height is around 80 percent (though whenever I try to demonstrate this in class the exercise invariably fails for some reason, presumably because we’re bad judges of our own nature). By now, 40,000 Englishmen have had their DNA subjected to whole genome genotyping for one reason or another. That’s plenty enough data to expect that scanning 500,000 genetic polymorphisms will lead geneticists to all the major genes that influence how tall a person is—except that it has only finger pointed 25 genes, and astonishingly they collectively explain a measly four percent of the variation.
Any one genetic variant is responsible for adding no more than a millimeter or two to height. Two people with the opposite alleles at most of these genes are certainly likely to differ by several inches in height, but they probably won’t be in the 5-foot and 7-foot categories. If all these genes were identical in all people, we’d barely notice a difference in the spectrum from short to tall. These data also tell us that if we could measure everyone, we’d find that literally thousands of genes affect how tall we are. Some will affect growth of your leg bones, some the muscular support of the spine, and some how big your head is, but all the common variants have very small effects.
On the other hand, when small and large breeds of dogs are compared, it turns out that a few genes seem to account for many of the differences. Unbeknownst to them, those responsible for breeding most of the small dogs were selecting bitches and studs that have in common a single mutation in the Insulin like growth factor 1 (IGF1) gene when they chose which smaller than average dogs to mate. When it comes to explaining just the short legs of Dachshunds and Corgis and the like, a similar mutation in another gene seems to be responsible. There is even speculation that complex behaviors such as loving to play in water might have relatively simple genetic underpinnings.
How can we reconcile these human and canine findings? It is probably a reflection of the histories of the two species: Artificial selection by dog breeders acts primarily on mutations that have a large effect, while natural selection and genetic drift during human evolution has acted on background variation. Within single families, it is possible that there are alleles that ensure that siblings, even nonidentical twins, can be several inches different in height despite being raised in the same home. But these are rare enough in the general population that they don’t make much of a contribution overall. Reciprocally, within breeds of dogs, there are thousands of polymorphisms that we will never know about that contribute to size differences.
The story for body weight is similar. Those tens of thousands of genome scans, many in the context of diabetes susceptibility, have pulled out just a couple of genes that have a relatively large effect. The major body mass gene in Caucasians is called FTO, which rumor has it is a contraction of fatso (but only fly geneticists can get away with such disrespect). We don’t know what FTO does yet, but if you have two copies of the heavy allele, you may be as much as 2 or 3 pounds heavier than if you had the other alleles. Most of the remaining variation is due to literally hundreds of genes with alleles that add only a few grams here and there.
On the whole, humans are not a particularly colorful species, at least not naturally. Until a few decades ago, pretty much anywhere you went in the world, most people you encountered would have been the same local hue, somewhere along the bland axis from black to white with a touch of rouge or ochre thrown in. Ninety percent of the world’s population has pitch-black hair, with only northern European derivatives experimenting with brown or blonde. Eye color is a little more variable, for reasons that are completely obscure.
Given this, Scandinavia would seem to be a good place to begin the search for the genes behind differences in skin color, and indeed the folks at deCODE Genetics in Iceland have obliged. Their scan of around 7,000 Icelanders and Dutchmen turned up five genes that convincingly influence eye and hair color and another region of a chromosome that is associated with freckles. Several of these were previously known to influence the transition from dark to pale skin in Europeans and East Asians as well, which makes sense because all these color attributes involve the deposition of a pigment in particular cells.
The biggest factor they found is a site near a gene called OCA2 that increases your odds of having blue instead of brown eyes about twentyfold, and blue instead of green eyes about fivefold. More than 80 percent of northern Europeans have it. Very few sites elsewhere in the genome are more different in frequency between Africans and Europeans. No other genes come close with respect to their effect on brown eyes, but if you also have a particular variant of SLC24A4 you can be reasonably confident that your eyes will be blue not green. It is not clear where hazel and gray eyes fit into the picture. Note that these are still by no means diagnostic indicators of color, which I find comforting, since my grandson has very blue eyes—even though his mother is without doubt an Italian woman. Were the dark brown eyes my stepson’s instead, I suspect we’d be wondering a little harder about his role in the conception.
