aging

I'm usually pretty measured when I respond to dumb ideas about evolution reported in the press. After all, scientists are often misquoted, or misunderstood by reporters. So, I didn't really think it was worth writing about this story covering a lecture by UCL geneticist Steve Jones. After all, I'm hardly going to attend a faculty talk in London, and there's really no news here -- Jones has been arguing for more than ten years that human evolution has slowed or stopped.

For example, this 1995 article in the NY Times describes his book, The Language of Genes:

"Natural selection has to some extent been repealed" in the case of humans, says Dr. Steve Jones, a geneticist at University College London. Most social changes "seem to be conspiring to slow down human evolution," he argues[.]

His ideas have been publicized for years outside of his books; for example, a 2002 public debate.

But this latest Steve Jones kerfuffle seems to have impressive reach. It hit Slashdot, for goodness' sake. The Guardian has pubished an exchange of opinion pieces about it. Bloggers of note have picked it up, almost universally to criticize it as a wrong idea.

Why it's wrong

What I haven't yet seen, in all the commentary, is a short and simple refutation for each element of his argument. Let me lay out the components of Jones' argument, as explained in the current article and previous works:

1. Evolution includes natural selection, mutation, and random change.

Jones excludes gene flow, one of the usual four mechanisms of evolution -- this allows him later to argue that population mixing is a sign of evolution stopping, when in fact it is evolution.

2. Older fathers have a higher mutation rate than younger fathers or mothers, and the proportion of older fathers is now much less than in the past.

This is true, but minor compared to the main factor affecting the introduction of mutations into human populations: the population size. The rate of new mutations in the population is 2Nu, where u is the rate per individual, and N the number of people. The population of the world has increased tenfold since 1700. All other things equal, this means ten times as many mutations -- and twice as many mutations per generation today as in 1960. There are a smaller proportion of older fathers now than in 1700, but the absolute number of older fathers is much, much greater.

Besides all this, the story of paternal age at birth is not so simple. Over the past twenty years in the U.S., the birth rate to fathers over 35 has been increasing, while the birth rate to fathers under 30 years has been decreasing. Reproduction in men aged 20-35 grew markedly after World War II, but the fraction of births to older fathers has been climbing since 1970. To be sure, the current rate of births to older fathers remains substantially less than before 1940, but this is part of an overall reduction in fertility across all age classes. Over the past 200 years, a reduction in average male fertility has been made up by an increase in infant and juvenile survival, so even though the birth rate to older (and younger) fathers declined, the population size continued to grow.

3. Mortality of young people has reduced to near-zero.

Jones acknowledges that this is only true in industrialized economies, so I'll set aside the obvious point (also made by Chris Stringer in this article: Mortality is still high among young people in a global context.

But Jones entirely neglects fertility. Fertility selection depends on the variance in lifetime reproduction (some people have more children than others), as well as the variance in age at reproduction (some people have children earlier than others). Selection does not stop, even if mortality does. He also neglects the high human rate of spontaneous abortion, a continuing source of mortality selection.

Also, the decrease in mortality means that some mutations that once were deleterious are now neutral. These mutations now will be retained in the population rather than rapidly eliminated, and some of them will increase under genetic drift. In terms of the rate of change in frequency of these previously rare deleterious alleles, this means that the population will henceforth evolve more, not less.

4. Small isolated populations allow rapid evolution by drift. But today's population is large and highly interconnected.

There's no denying this, at least if we're talking about the rate of change of allele frequencies within each small population. However, the rate of fixation of neutral alleles is independent of population size. And the global rate of evolution is far slower in a network of partially isolated subpopulations than in a single large population. So this argument depends on what we mean by "evolution." Here's what we have from Jones in this context:

“Small populations which are isolated can evolve at random as genes are accidentally lost. Worldwide, all populations are becoming connected and the opportunity for random change is dwindling. History is made in bed, but nowadays the beds are getting closer together. We are mixing into a global mass, and the future is brown.”

Jones' definition of evolution (in argument 1, above) leads inevitably to confusion here. Clearly, this "mixing into a global mass" is actually rapid evolution of human populations, measured in terms of changes in allele frequencies. If Europe becomes "brown" in 500 years, that's a whole lot faster than the 20,000 years it took Europe to become "non-brown." Jones apparently means that sometime in the future, after this current, rapid period of evolutionary change, human evolution will be slow.

But he's wrong. Mutations will be entering this large population faster than in the smaller global population of the past. This future population will be vastly more variable than any of the small human populations of today. Alleles under selection will be able to move and mix much faster than through the disjointed network of population contacts that existed 200 years ago. Only by one measure will evolution be slower: the rate of change in frequency of neutral mutations. But even that will be faster in the mixed future than the semi-isolated past, if we consider it globally instead of locally.

Longevity

Several respondents to Jones' arguments have taken an approach that I think is misguided. Pointing out that human longevity was much lower in the past, they argue that a much lower proportion of children were being born to older fathers.

Actually, longevity has a very limited application to this argument. The high juvenile and infant mortality rates of the past have no influence on the average age of fathers at reproduction, since fathers are a subset of people who survive juvenile and infant mortality. For example, in the U.S. around 1850, only around 60 percent of all people born would survive to age 20. These deaths greatly reduced the average longevity, but had no effect at all on the proportion of births to older fathers.

What matters is the fraction of young men (20-30) who survive to be older men (35-50). That fraction was high: roughly 85% of twenty-year-old men in the U.S. in 1850 survived to age 35, and two thirds survived to age 50. Older men were fathering a large fraction of the infants in early America and in pre-industrial populations. This is because once they started, they didn't stop, as long as they had a wife who could have children (and widowers often remarried). This was probably true in all agricultural societies, and likely in modern human hunter-gatherers back to 30,000 years ago or earlier. If we go back further in time, we find a much higher mortality rate among young men, so that fathers over 35 made much less of a contribution. But in historic times, older fatherhood has been very important to rate of new mutations per individual.

What's important is that (a) the proportional reduction of older fatherhood is a small effect compared to the increase in mutations due to the growth of the population, (b) part of the decline of birth rates to older men is compensated by a reduction of infant and juvenile mortality, and (c) older fatherhood is now rising, not falling.

I think Jones ought to pursue a far more interesting interpretation of these facts. European and American men today are increasingly pursuing part of a reproductive strategy that was common in the past, but less common in postwar Europe and America. Today, with lower infant and childhood mortality, the consequences of that strategy are potentially more powerful than in the past.

Bottom line

As always, claims about the rate of evolution in the future depend only slightly on empirical observations, and mainly on assumptions. In this case, Steve Jones has defined the "rate" of evolution in a very particular way, to come to the story that he prefers.

I generally don't mind when prominent people say silly things about evolution. It gives the rest of us a chance to explain why they're wrong, and teach about the mathematical basis of evolution as we do it. In this case, it's sort of sad: Jones is out there making arguments and selling books, but he's clearly trapped in the pre-genome era. The exciting thing about genetics today is the extent to which we can observe human evolution happening!

There's also an antiquated version of ethnocentrism here: how can we talk about the future of human evolution without considering the intense dynamics in today's developing nations? Relative to Africa and Asia, Europe is now a population sink.

Kim Hill and colleagues (2007) report in the current Journal of Human Evolution on the mortality profile of recent Hiwi hunter-gatherers. Here is their abstract:

Extant apes experience early sexual maturity and short life spans relative to modern humans. Both of these traits and others are linked by life-history theory to mortality rates experienced at different ages by our hominin ancestors. However, currently there is a great deal of debate concerning hominin mortality profiles at different periods of evolutionary history. Observed rates and causes of mortality in modern hunter-gatherers may provide information about Upper Paleolithic mortality that can be compared to indirect evidence from the fossil record, yet little is published about causes and rates of mortality in foraging societies around the world. To our knowledge, interview-based life tables for recent hunter-gatherers are published for only four societies (Ache, Agta, Hadza, and Ju/'hoansi). Here, we present mortality data for a fifth group, the Hiwi hunter-gatherers of Venezuela. The results show comparatively high death rates among the Hiwi and highlight differences in mortality rates among hunter-gatherer societies. The high levels of conspecific violence and adult mortality in the Hiwi may better represent Paleolithic human demographics than do the lower, disease-based death rates reported in the most frequently cited forager studies.

The mortality rates reported for the Hiwi are higher than those for other hunter-gatherers -- especially the African groups (Hadza and !Kung), but not stunningly so. Among pre-1960 Hiwi males, 57 percent could expect to survive to age 15, and 43 percent to age 30, with an average young adult mortality rate of around 2 percent annually. So it is not anything like as high as has been suggested for Neandertals and earlier humans (with annual mortality rates as high as 6 percent).

The most interesting aspects of the paper are the comparisons between the Hiwi and other ethnographically-known hunter-gatherers. Many of the differences in mortality profiles are attributable to strong cultural differences:

Cause of death among the groups differs considerably. Disease is an important cause of death in all groups, but represents only 20% of deaths in the precontact Ache, 45% among the precontact Hiwi, and about 75–85% of all Hadza, !Kung, and Agta deaths. Respiratory disease is the main killer of the Ache, whereas gastrointestinal pathogens are most important among the Hiwi and probably Hadza. Among the !Kung, respiratory and gut infections are about equally important. Violence is the major cause of death among the precontact Ache (55% of all deaths) and very important among the Hiwi (30% of all deaths), but notably less important in the two African societies and the Agta (3–7% of all deaths). Indeed, the crude homicide/warfare death rates per year lived are more than ten times higher among the Hiwi and Ache than among the Hadza or !Kung (1/100 and 1/200 per year for precontact Hiwi and Ache, respectively, vs. 1/2500 and 1/3000 for the Hadza and !Kung, respectively). Blurton Jones et al. (2002) suggested that this may be due to the more pervasive effects of colonial governments in Africa and the reduction of intertribal warfare. Even so, within-group homicide and infanticide rates are also much lower among African foragers, suggesting real cultural differences in violence rates.

The most notable contrast among hunter-gatherer life tables is the overall similarity of child mortality followed by subsequent high mortality of the Hiwi and Agta in adulthood compared to the Ache, !Kung, and Hadza (Fig. 3). The number of individuals at risk in each yearly category and the number of deaths observed have been only published for the Ache and Hadza. Thus, statistical analyses of differences in mortality rates between these groups and the Hiwi can be performed using logistic regression. The results suggest that all foragers are not characterized by a single “typical” mortality schedule. Analyses of the differences for infants, children, adults, and elderly using logistic regression (Table 6) shows significantly lower Ache infant mortality and early-adult mortality relative to the Hiwi, and lower Hadza adult mortality (both young and old) relative to the Hiwi. Particularly striking is the fact that Hiwi early-adult mortality rates are about double those of the Ache and Hadza (Hill et al. 2007:449-450).

Violence is as important a cause of death as disease for young Hiwi adults, and for infants as well. On page 451, the paper points out that violence and accident cause as many deaths in the Hiwi young adults as occur in most other hunter-gatherers from all causes combined. Hill and colleagues discuss this issue in relation to the possible life history pressures on Paleolithic hunter-gatherers:

If high mortality, warfare, homicide, and accidental trauma are typical of our Paleolithic ancestors, the Hiwi mortality patterns may be more representative of the past than those derived from other modern hunter-gatherers. If so, several observations about the Hiwi are important. First, conspecific violence was a prominent part of the demographic profile, accounting for many deaths in all age and sex categories. Most of the adult killings were due to either competition over women, reprisals by jealous husbands (on both their wives and their wives' lovers), or reprisals for past killings. The criollo-caused killings were motivated by territorial conquest. Moreover, infanticide (especially on females) constituted the highest mortality rate component of all Hiwi conspecific violence. Second, no predation deaths were reported despite attacks by anacondas, Orinoco caimans, and piranhas, and the presence of jaguars in the area. Accidents associated with the active-forager lifestyle were common, but disease was a more important killer, accounting for nearly half of all deaths. This suggests an adaptive landscape in which success in social relations, competitive violence, and disease resistance are paramount. This may partially explain why many of the genes that appear to have been under strong selection in the past 50,000 years affect either disease resistance or cognitive function (Wang et al., 2006), presumably related to success in an atmosphere of frequent violent social competition (Hill et al. 2007:451).

The paper also includes a substantial discussion of the implications of high young adult mortality for intergenerational investments, such as grandmothering. This is an important issue, and Hill and colleagues end their discussion with a suggestion that neither the "grandmothering" nor the "embodied capital" models for the evolution of long life spans is sufficient to explain the human pattern. In their view, the key difference between humans and other primates (notably, chimpanzees) is not life span itself, but the markedly lower mortality rate among young adults. This low mortality rate directly causes the long life span (if you don't die young, you'll live long!). Hill and colleagues favor extrinsic factors such as greater protection of children, nursing the sick, and food sharing as possible causes of reduced mortality rates in humans.

This is a thought-provoking paper, beyond the valuable data, because of its discussion. I will have more to say about the early hominids later on.

References:

Hill K, Hurtado AM, Walker RS. 2007. High adult mortality among Hiwi hunter-gatherers: implications for human evolution. J Hum Evol 52:443-454. doi:10.1016/j.jhevol.2006.11.003

FuturePundit points me to a study of telomerase expression in mice. Here's the abstract:

In multicellular organisms, telomerase is required to maintain telomere length in the germline but is dispensable in the soma. Mice, for example, express telomerase in somatic and germline tissues, while humans express telomerase almost exclusively in the germline. As a result, when telomeres of human somatic cells reach a critical length the cells enter irreversible growth arrest called replicative senescence. Replicative senescence is believed to be an anticancer mechanism that limits cell proliferation. The difference between mice and humans led to the hypothesis that repression of telomerase in somatic cells has evolved as a tumor-suppressor adaptation in large, long-lived organisms. We tested whether regulation of telomerase activity coevolves with lifespan and body mass using comparative analysis of 15 rodent species with highly diverse lifespans and body masses. Here we show that telomerase activity does not coevolve with lifespan but instead coevolves with body mass: larger rodents repress telomerase activity in somatic cells. These results suggest that large body mass presents a greater risk of cancer than long lifespan, and large animals evolve repression of telomerase activity to mitigate that risk.

The study is discussed pretty well in this University of Rochester press release:

"Mice express telomerase in all their cells, which helps them heal dramatically fast," says Gorbunova. "Skin lesions heal much faster in mice, and after surgery a mouse's recovery time is far shorter than a human's. It would be nice to have that healing power, but the flip side of it is runaway cell reproduction -- cancer."

Up until now, scientists assumed that mice could afford to express telomerase, and thereby benefit from its curative powers, because their natural risk of developing cancer is low -- they simply die before there's much likelihood of one of their cells becoming cancerous.

"Most people don't know that if you put mice in a cage so the cat can't eat them, 90 percent of them will die of cancer," says Gorbunova.

I for one didn't know that. Of course, there is no paradox here -- early reproduction in mice has a much higher impact on their fitness than late reproduction, and they really shouldn't live long enough to compete reproductively with their daughters -- if they can devote that late reproductive effort to earlier reproduction, they certainly should do so.

But isn't it interesting that cancer in particular should be a high risk for late-living mice, and that it might be linked to the facility for healing early injuries?

Telomerase has long been recognized as one of the big baddies behind most cancers. Here's a paragraph from a 2001 review by Shay and colleagues:

Telomerase, a eukaryotic ribonucleoprotein (RNP) complex (26-33), helps to stabilize telomere length in human stem cells, reproductive cells (34) and cancer cells (35,36) by adding TTAGGG repeats onto the telomeres using its intrinsic RNA as a template for reverse transcription (37). Telomerase activity has been found in almost all human tumors but not in adjacent normal cells (35,36). The most prominent hypothesis is that maintenance of telomere stability is required for the long-term proliferation of tumors (38-42). Thus, escape from cellular senescence and becoming immortal by activating telomerase, or an alternative mechanism to maintain telomeres (43), constitutes an additional step in oncogenesis that most tumors require for their ongoing proliferation. This makes telomerase a target not only for cancer diagnosis but also for the development of novel anti-cancer therapeutic agents.

The study of telomerase knockout mice has shown that they get messed up in some ways characteristic of aging (e.g., grey hair, hair loss) and that they start having wounds in places with chronic mechanical stresses, like the distal limbs, snout, and throat (Rudolph et al. 1999). Additionally, old telomerase deficient mice had slow wound healing. In contrast, they did not suffer generalized effects of aging to other organ systems -- the effects seem to be concentrated in the skin. This makes some sense because the skin functions by attrition -- constantly wearing off into the environment and fixing itself based on environmental insults. The other system that depends on constant loss, the blood, was also affected by the lack of telomerase, with the rate of blood cell replenishment significantly lower in older telomerase-knockout mice.

Kim and colleagues (2002) presented a good review of the relation of telomeres to cancer and aging, and mention the relation to wound healing from several different studies. It's a good review of the literature to that point, and they also discuss a number of other proteins besides telomerase that are associated with maintaining telomere structure or function.

The evolution of telomere stabilization has involved multiple pathways, and that does generate an apparent paradox: a deficiency in telomerase helps to trigger certain cancers. These manage to spread by a telomerase-independent pathway, which enables cell replication to continue without telomerase. In other words, there is no single cancer switch involving telomere length, and the removal of the ordinary regulator protein may cause other pathways to spiral out of control.

References:

Kim S-H, Kaminker P, Campisi J. 2002. Telomeres, aging and cancer: in search of a happy ending. Oncogene 21:503-511. DOI link

Rudolph KL, Chang S, Lee H-W, Blasco M, Gottlieb GJ, Grieder C, DePinho RA. 1999. Longevity, stress response, and cancer in aging telomerase-deficient mice. Cell 96:701-712. DOI link

Seluanov A and 7 others. 2006. Telomerase activity coevolves with body mass not lifespan. Aging Cell (online early) DOI link

Shay JW, Zou Y, Hiyama E, Wright WE. 2001. Telomerase and cancer. Hum Mol Genet 10:677-685. Free full text

The New York Times is carrying an article by Gina Kolata that discusses research on genetics and longevity. She has quotes from several big figures in the field -- Vaupel, Finch, Christensen -- and it's a good article.

The main idea is that lifespan is weakly heritable:

Now, Dr. Christensen and his colleagues have analyzed the data. They restricted themselves to twins of the same sex, which obviated the problem that women tend to live longer than men. That left them with 10,251 pairs of same-sex twins, identical or fraternal. And that was enough for meaningful analyses even at the highest ages. "We were able to disentangle the genetic component," Dr. Christensen said.

But the genetic influence was much smaller than most people, even most scientists, had assumed. The researchers reported their findings in a recent paper published in Human Genetics. Identical twins were slightly closer in age when they died than were fraternal twins.

But, Dr. Christensen said, even with identical twins, "the vast majority die years apart."

Basically, the argument is that death is complicated -- lots of "random" events can cause it:

The likely reason is that life span is determined by such a complex mix of events that there is no accurate predicting for individuals. The factors include genetic predispositions, disease, nutrition, a woman's health during pregnancy, subtle injuries and accidents and simply chance events, like a randomly occurring mutation in a gene of a cell that ultimately leads to cancer.

Of course, it can't be only that. After all, stature (one of the strongly heritable traits used here as a contrast) is likewise determined by disease, nutrition, mother's health, subtle injuries, and so on. It is not only the complexity, but also the size of the random component. Essentially, what is left to kill people are all the things that genes can't avoid very easily. This includes stuff that is fundamentally rare (so strength of selection is slight) as well as stuff that feeds on complexity (with multiple subsystems that can break).

There is also an interesting bit about the heritability of deaths in given age intervals -- apparently there is no significant heritability for deaths under 60. Of course, that won't be true for societies where disease causes a significant number of deaths under 60. Sure in Western Europe where you have basically sporadic cancers, heart disease and accidents, the heritability will be zero. But throw in malaria and a few disease resistance alleles, like sickle-cell, and the story will be quite a bit different. So part of the story is that we have done so well eliminating some of the main causes of mortality that interact significantly with a person's genetics.

But as the article points out, the kinds of quantification that work well for classes of people don't work so well as applied to individuals -- even identical twins. And there is some confusion there, since "heritable" doesn't mean the same thing as "identical in monozygotic twins." If you're looking for articles on genetics to give out to students, this is a good one for raising points for discussion!

This is one of the most beautiful openings to a paper, ever:

Consider four hypothetical genes in man. Suppose all are limited in their expression to the female sex and also age-limited in the following way: each gives complete immunity against some lethal disease but only for one particular year of life. Suppose the first gives immunity for the first year, the second for the fifteenth, the third for the thirtieth, and the fourth for the forty-fifth. What are the relative selective advantages of these genes?

If for further simplicity parental care is ignored and it is assumed that the menopause always comes before age 45, it is at once obvious that the fourth gene is null, whereas all the others do confer some advantage. It is also fairly obvious that the third gives less than the second. But how much less? Does the second give a maximum becasue it occurs at the age of puberty? Does the first give less than the second?

The importance of questions of this kind for an evolutionary theory of senescence has been realized for some time. Most of the answers that will be given in this paper agree with the theory of Williams (1957). Although perhaps not obvious, they are so simple that it is surprising to find almost no indication that they had been realized earlier. Several writers have in effect answered the last two questions in the affirmative, which is for the one inexact and the other wrong.

If anybody in biology ever wrote like Nero Wolfe talks, it was Bill Hamilton. That intro is from "The Moulding of Senescence by Natural Selection," (Hamilton 1966:12-13), which ought to be required reading, if it isn't.

Later in the paper Hamilton presents almost fully-formed his idea of "sibling replacement" to explain why subadult mortality should be concentrated in infants:

Suppose that the catching of a disease in the immature period is ineveitable and that hte first infection has only two possible outcomes: death or survival with perfect immunity against second infection; and suppose that the probabilities do not change according to the age of the first infection. Then if there is any degree of sibling replacement at all, a gene bringing forward the expected age of first infection will be selected, for it can easily be seen that the more commonly the gene is appearing in a progeny the larger its expected completed size will be, while at the same time the expected frequency of the gene within the progeny is unchanged.

If the bringing forward of the susceptibility involves a disadvantage, for instance by a slight increase in the chance that death is the outcome of infection, the situation is more complex and will require mathematical analysis to delimit the possibilities. This is because it is then no longer true that the proportion of the gene is unchanged by the amount of replacement that goes on; positive selection could fail even if it was guaranteed that the amount of replacement could more than compensate for the extra mortality. This is in effect a problem of a more "altruistic" versus a more "selfish" trait (ibid:39).

And so the theoretical power of inclusive fitness is made plain: it encompasses even death itself in the service of first-order relatives.

I found this quote from Vaupel et al. (2003) very interesting:

Death of the frail alters the composition of a cohort, lowering subsequent mortality and possibly offsetting increases in mortality resulting from cumulative damage.

Their paper is a commentary on a study of dietary restriction and longevity in Drosophila (Mair et al. 2003). Dietary restriction is known to promote longer lifespan in many model animals, including Drosophila. Mair et al. (2003) found that the effect of dietary restriction was not necessarily a long-term phenomenon -- instead, they found that the death rates of the flies were strongly determined by what they had been eating in the 2 previous days (!):

These investigators show that when flies fed a restricted diet are switched to a full diet, mortality soars to the level suffered by fully fed flies. Conversely, when the diet of fully fed Drosophila is restricted, mortality plunges within 2 days to the level enjoyed by flies that have experienced a lifelong restricted diet.

Vaupel et al. (2003) press the issue to consider the effects of lifestyle change on human mortality rates:

Following unification of East and West Germany (1989-1990), mortality in the East declined toward prevailing levels in the West, especially among the elderly (8, 9) (see the figure). Although conditions early in life do significantly influence human health and survival late in life (10), the German example--and other demographic data--provide strong evidence that such effects are of less importance (at least in more recent decades) than changes in current conditions (2, 11, 12). The second half of the 20th century saw a radical (and continuing) decline in old-age mortality in most developed countries: in Western Europe, for instance, from 1950 to 2000 the probability of surviving from age 80 to 100 increased 20-fold (13). Most of this increase is due to improvements in economic and social conditions and to ongoing medical advances (14).

Epidemiological and clinical research provide further evidence of the malleability of old age. For example, the risk of death for elderly smokers who quit falls, within 1 or 2 years, to a lower level than that suffered by recalcitrant smokers (15). There is a growing appreciation that even octogenarians and nonagenarians can substantially benefit from medical interventions such as cataract surgery (16) and hip replacement (17). Low-tech interventions at advanced ages can have an important impact. For instance, Fiatrone et al. found that physical training leads to significant gains in muscle strength, size, and functional mobility among frail residents of nursing homes: The oldest person studied (and helped) was 96 years old (18).

The bottom line is that mortality rates are malleable. They are changed not only by improvements in health, nutrition, and environment across the lifespan; they are also improved by short-term changes in older adults. That shouldn't really be a surprise (yes, I mean beyond the fact that this is a 2003 paper!), since most of us know somebody who, thanks to modern medicine, is living with some condition that would have killed them 40 years ago. But somehow medical changes seem to be exceptions, not "natural" in some way. The point of the experimental work and Vaupel et al.'s (2003) commentary is that medicine is not the only late-acting life-extender; broader changes in behavior, stress, diet, or social relationships.

Aristotle contrasted premature death with natural death due to old age--he asserted that nothing could be done about old age (19). More than 23 centuries later, many still believe that death rates at older ages are intractable (20). This view is reinforced by evolutionary theories of aging, which emphasize that senescence is inevitable because the force of selection against deleterious, late-acting mutations declines with age (21). Research over the past decade strongly supports an encouraging alternative--that aging is plastic and survival can be substantially extended by various genetic changes and nongenetic interventions (1-2, 5, 7-18, 20). For most species, damage to cells and tissues accumulates with age, and mortality rises. Nonetheless, aging is so remarkably pliable that interventions do not have to be lifelong. As illustrated by the Mair et al. report and other studies, interventions even late in life can switch death rates to a lower, healthier trajectory.

I'm thinking about this in the context of the new study by Rachel Caspari and Sang-Hee Lee (2006) in AJPA. Here's the abstract:

Increased longevity, expressed as the number of individuals surviving to older adulthood, represents a key way that Upper Paleolithic Europeans differ from earlier European (Neandertal) populations. Here, we address whether longevity increased as a result of cultural/adaptive change in Upper Paleolithic Europe, or whether it was introduced to Europe as a part of modern human biology. We compare the ratio of older to younger adults (OY ratio) in an early modern human sample associated with the Middle Paleolithic from Western Asia with OY ratios of European Upper Paleolithic moderns and penecontemporary Neandertals from the same region. We also compare these Neandertals to European Neandertals. The difference between the OY ratios of modern humans of the Middle and Upper Paleolithic is large and significant, but there is no significant difference between the Neandertals and early modern humans of Western Asia. Longevity for the West Asian Neandertals is significantly more common than for the European Neandertals. We conclude that the increase in adult survivorship associated with the Upper Paleolithic is not a biological attribute of modern humans, but reflects important cultural adaptations promoting the demographic and material representations of modernity.

This would seem to indicate that recent human cultural innovations, as encompassed within the Upper Paleolithic of Europe but presumably shared by all living and recent people, were responsible for a vast increase in the survivorship of younger adults. More people were reaching older ages. This doesn't necessarily mean that these early modern people matched the longevity of historic populations, but it does mean that they must have escaped or delayed many of the major sources of mortality affecting earlier modern and archaic humans. And rather than being an aspect of human phylogeny, specifically the cultural changes indicated by 25,000 years ago appear to have been key.

The work on experimental animals and recent human demography comes together to indicate that nongenetic, environmental changes in mortality can be substantial. Not only can they be great, but they may be highly sensitive -- small changes in parameters may have a large impact on the chance of death. It doesn't take much figuring to realize that a small decline in age-specific mortality may have a large effect on average lifespan.

The death of the strong alters the composition of a cohort also. When a key person dies, it has ripple effects on small human groups. For people living on the edge, when a key person takes a risk, it impacts the entire group's chances of survival. When risks are less necessary, the entire group may benefit.

The secret to understanding the evolution of longevity is to show the channels through which changes in human culture allowed reductions in mortality. These changes are technological, organizational, and -- very possibly -- spiritual. All ways that humans deal with risk.

References:

Vaupel JW, Carey JR, Christensen K. 2003. Aging: It's never too late. Science 301:1679--1681. DOI link

Caspari R, Lee S-H. 2006. Is human longevity a consequence of cultural change or modern biology? Am J Phys Anthropol (in press) DOI link

Here's some timeless advice for anyone:

If you wish to be a prophet, first you must dress the part. No more silk ties or tasseled loafers. Instead, throw on a wrinkled T-shirt, frayed jeans, and dirty sneakers. You should appear somewhat unkempt, as if combs and showers were only for the unenlightened. When you encounter critics, as all prophets do, dismiss them as idiots. Make sure to pepper your conversation with grandiose predictions and remind others of your genius often, lest they forget. Oh, and if possible, grow a very long beard.

In this particular instance, it's from an article about biogerontological prophet Aubrey de Grey in the Chronicle of Higher Education. The article has a short review of the recent Cambridge conference on Strategies for Engineered Negligible Senescence. But mostly it's a profile.

And it's quite a profile, if you can suffer through some "Oh, please" moments describing his personal life.

This is my favorite part, although not the most entertaining:

One will not find Mr. de Grey in the laboratory hovering over petri dishes or test tubes. He readily acknowledges that he lacks the qualifications to perform experiments. What some might view as a handicap, he sees as a strength: Rather than spending his time behind a microscope, he reads the literature and searches for connections that a specialist may have missed.

Imagine that! And what does it get him?

Mr. [David] Finkelstein has little respect for Mr. de Grey's own research contributions. "I am very underwhelmed," he says. The fact that Mr. de Grey does not set foot inside a laboratory also bothers him: "Look, you either work at the bench, or you don't work at the bench," he says.

Wow.

The gene's name is Klotho, the nicest of the three Fates. Science is reporting on a not-yet-available-to-me-at-least article about this gene, which affects longevity in mice.

Whereas lab mice can live about 2 years, mice engineered to overproduce this protein, called Klotho, have celebrated third birthdays, Makoto Kuro-o of the University of Texas Southwestern Medical Center in Dallas and his colleagues report online in this week's Science Express (www.sciencemag.org/cgi/content/abstract/1112766). The mutant rodents represent a rare case of a single gene substantially influencing life span in mammals.

Klotho also occurs in humans, so the news part of the story is the hope of developing new drugs to increase longevity in humans. That is also where the controversy is, since there are some reasons -- notably insulin resistance -- to think that it may not work easily in humans.

But I'm more interested in the evolution of the gene, and here there are some interesting hints. First, there is this:

While Klotho is produced only in the kidney and brain, a fragment of it slips into the blood and may act like a hormone.

And this:

As with lab animals coaxed to have lengthy life spans, the altered rodents had fertility problems. They produced about half the expected number of offspring.

That is pretty interesting. Could it be that an evolutionary increase in brain size had the side effect of higher expression of this hormonal fragment? Such a mechanism for this (or other) genes might explain human longevity as a side effect of brain evolution. (Also interesting to me because Gould (2002) suggested just the opposite: that brain size might be a side effect of increasing longevity).

And could it be that increased longevity comes with the inevitable cost of lower fertility? In that context, it would seem to take a very special genetic change to allow longevity to increase without incurring this cost to fertility. One strategy, of course, would be to increase survivorship of the offspring that are born, thereby reducing the risk of low fertility.

A question that the Science news piece doesn't answer: Is there variation in the human version of Klotho?

Fortunately, we have OMIM, which has a long entry on Klotho, including this:

Arking et al. (2002) identified a functional variant of klotho, a haplotype termed KL-VS, which is defined by the presence of 6 single-nucleotide polymorphisms (SNPs) in an 800-bp region spanning exon 2 and flanking sequence. Allele-specific oligonucleotide hybridization analysis showed complete linkage disequilibrium for the coding region mutations. Of the 3 mutations in exon 2, 1 is silent and 2 encode amino acid changes, phe352 to val (F352V) and cys370 to ser (C370S). The variant was associated with reduced human longevity when in homozygosity. The prevalence of the variant in the general population was estimated to be 0.157.

The study by Arking et al. (2002) also indicates another allele associated with reduced longevity, recognized by its linkage to a marker (allele 17) but without any recognized coding sequence variants. I suspect its effect is due to a difference in expression, but there is no evidence one way or another. Additionally, the KL-VS allele not only influenced survivorship to age 75, but also increased longevity slightly in heterozygotes older than 80 years of age.

Also of interest, the frequency of the KL-VS allele appears to be very similar (18.7 percent, 23.8 percent, and 25.7 percent) in the three study populations, including Bohemian, Baltimore Caucasian, and Baltimore African-American groups. That similarity may suggest that the variant is in selective balance broadly in human populations, and the authors do observe a slight heterozygote advantage in the Bohemian sample, although they did not look for an effect on net fertility.

I'd like to see a gene tree.

Carl Zimmer (The Loom) is back from appendicitis, and has a Times article on ... appendicitis!

When I got back home, I began doing some research and found that each year, more than 250,000 Americans had appendectomies. The more I thought about that figure, the more absurd it seemed. Why should an organ fail in so many healthy people? What makes it more puzzling is that no one ever needs an appendix transplant. Appendix-free, I can expect a normal life.

It's an oldie, but a goodie, in evolutionary terms. Zimmer interviews Rebecca Fisher (Midwestern University) about her research on appendices in primates (apes have them, some monkeys do, others don't):

Still, I wondered how such a dangerous and disposable organ could survive over evolutionary time. "We consider it maladaptive because we want to live to a very old age," Dr. Fisher said. "But from a strictly Darwinian view, it might not be."

Fisher describes a stabilizing selection hypothesis: the appendix kills some people, but it saves others because of a role in immune defense. This would especially be powerful if the defensive role was greater in childhood (and thereby highly effective in getting people to survive to reproduce) and the killer role greater in elderly people.

Zimmer's Loom post also describes a hypothesis by Nesse and Williams, who proposed that smaller appendices might be even more prone to infection, thus leading to an evolutionary blind alley --- they can't get smaller, because smaller is worse.

Either of these could be true. Or there are other explanations. Indeed, one reason the appendix is such a difficult problem is that there are so many unknowns.

For example, 250,000 appendectomies a year sounds severe --- it's around 1 in 1000 people. This means that your odds of going 60 adult years without an appendectomy would be around 94 percent. Lifetimes in the past were shorter, but if the incidence was the same per year, then the odds of a Pleistocene human going 30 adult years without needing an appendectomy would be a bit greater than 97 percent. All added up, that would seem to amount to pretty strong selection.

Moreover, epidemiological data show that appendicitis is most frequent among people nearing or at childrearing age. The highest incidence overall is in women ages 35 to 44, with all people ages 10 to 19 also having a high risk. So it's not just a disease of the old and infirm. If it had a similar incidence in other groups and in the past, it would have been a strong target of selection.

But that's a problem: we just don't know what its incidence may have been in the past. One recent study found that children with appendectomies had consumed significantly less dietary fiber; that would imply that today's diets have increased appendicitis.

And we don't know another key fact necessary to infer selection: how many of those people with appendectomies would have died without them? Of course, surgeons want you to think the number is close to 100 percent. I know that I sure wouldn't risk a perforation: the surgery is relatively easy, routine, and very low-risk compared to the high risk of peritoneal infection.

However, the fact remains that a substantial proportion of people with inflamed appendices will not have the burst. And a substantial proportion of burst appendices will not kill. We don't know what proportion that might be. And so, we can't say how much selection there may have been on the appendix.

And adding insult to surgery, around 5 percent of appendectomies are entirely negative.

Does it matter? Isn't any selection enough to take the appendix away, given enough time? That depends.

In particular, it depends what other structural effects may be caused by the genetic pathways that make the appendix. For a genetic change to result in no appendix, it must alter the regulation of developmental genes in the lower intestine. These genes probably don't only influence the appendix, but also other structures, a relation called "pleiotropy". Fixing one thing can break something else.

Another barrier to selection is genetic drift. If selection was small enough -- say less than one in a thousand over a lifetime, then it may have been very weak relative to genetic drift in the ancient human population.

Still, if I were taking odds, I would bet on Fisher's adaptive explanation. The conservation of an appendix in apes, along with its fairly distinctive form, argue that it it probably has a function that in the old days we couldn't have done easily without. It probably isn't just hanging around as a hanger-on.

References:

Fisher RE. 2000. The primate appendix: a reassessment. Anat Rec 261:228-236. PubMed

A paper by Hunter Fraser and colleagues (2005) in PLoS Biology describes a survey of gene expression in the cortex of humans and chimpanzees and the cerebellum of humans. The study investigates the relationship of gene expression with age in a series of human individuals of different ages and chimpanzees of different ages (via Gene Expression.

The paper briefly discusses the theories describing the evolution of aging, but they have little relevance to the procedures or results. It's not clear to what extent the setup of the paper was influenced by the ultimate results, as indicated by the following:

One promising approach to answering this question of evolutionary conservation lies at the level of gene expression: Do orthologous genes tend to undergo the same patterns of expression changes with age in diverse species, or can a common factor such as ROS lead to different gene expression patterns in different organisms? Using DNA microarrays, this question can now be addressed in a systematic, genome-wide manner. One such study found that a small but significant portion of aging-related gene expression changes are shared by the very distantly related nematode and fruit fly [14]; another study comparing aging patterns in muscle cells of two more closely related species, mouse and human, also found a great deal of divergence in aging patterns [15]. Although both of these studies are informative, neither addresses the questions of how quickly age-related gene expression patterns can evolve over short periods of time, and if humans in particular show unique patterns of aging not shared by closely related primates.

I'm not sure that "evolutionary conservation" would be a relevant prediction of either antagonistic pleiotropy or power of selection hypotheses for aging, since there is no reason to think that the same alleles that are adaptive in one species are necessarily adaptive in its relatives. In any event, the place to look for evolutionary conservativism is not between human and chimpanzee brains! Look among macaque species if you must. There is much more interesting stuff to consider in the evolution of humans.

That's why I say this may have followed the results rather than framed the research. Because the results show what one might expect: human and chimpanzee brains differ in their patterns of gene expression change with age.

That's a long phrase to gulp down. The study looked at 841 genes that show changes in expression with age in the frontal pole of the brain. They found that these genes tend to change in the same way in other parts of the cortex.

In contrast, they find that relatively fewer genes change expression with age in the cerebellum. Their results give good reasons to think that the cerebellum actually exhibits less aging-related change than the cerebral cortex. This is very interesting, especially if the majority of recent evolutionary changes have affected the cortex and not the cerebellum. If so, then we might expect many new advantageous mutations to have had suboptimal pleiotropic effects late in life. The development of a larger brain might well have led to one in which age-related changes were more detrimental. Or, conversely, the evolution of greater longevity in humans may have led to strong selection on regulator genes that had widespread effects on many cortical genes, but few effects on cerebellar genes.

The other major result is that genes in the chimpanzee cortex do not show the same pattern of change with age as the same genes in the human cortex. The chimpanzee cortex instead has its own, different pattern of age related changes in gene expression.

How does this difference arise?

Given this difference in aging patterns between humans and chimpanzees, we examined the expression levels of the chimpanzee orthologs of the 841 human genes that change expression with age in frontal pole in order to see if the expression levels of the chimpanzee orthologs of these genes resemble young humans, old humans, or neither. To test this, we first reanalyzed the expression data by masking all microarray probes with sequence differences between humans and chimpanzees [19]. We then calculated, for both human and chimpanzee prefrontal cortex, the average expression level for the set of genes that increase expression with age in frontal pole, as well as the average for the genes that decrease expression with age. The result is that in chimpanzee cortex, the orthologs of both sets of human genes (up-regulated and down-regulated) are expressed at the levels of their young human counterparts (Figure 5). In other words, chimpanzee cortex expression levels strongly resemble expression levels in young but not old humans, at least among the set of genes tested here. Humans then diverge from these average expression levels as they age, whereas chimpanzee gene expression levels change in an almost entirely different set of genes.

In their discussion, Fraser et al. focus on the possibility that radical oxygen species (ROS) related damage may differ between humans and chimpanzees. Perhaps so, but it may be pointless to look for a functional explanation for the difference. A non-functional explanation is that human genes active in the brain have undergone many rounds of natural selection, each leaving the possibility of slightly deleterious or neutral age-related effects. Added up over six million years, that leaves a lot of age-related changes in expression that just haven't been selected out. A few may be important causes of neurodegenerative disease, others may be barely noticeable.

It's a wide-open frontier for evolutionary explanation.

References:

Fraser HB, Khaitovich P, Plotkin JB, Paabo S, Eisen MB. 2005. Aging and gene expression in the primate brain. PLoS Biol 3:e374. Full text online

MIT Technology Review is offering a prize to any researcher in biogerontology who can write an essay demonstrating why Aubrey de Grey's SENS project is unworthy of serious scientific consideration.

Aubrey de Grey is a well-known futurist, with a twist. A few months ago, Technology Review had a profile of de Grey, that starts like this:

Aubrey David Nicholas Jasper de Grey is convinced that he has formulated the theoretical means by which human beings might live thousands of years -- indefinitely, in fact.

Perhaps theoretical is too small a word. De Grey has mapped out his proposed course in such detail that he believes it may be possible for his objective to be achieved within as short a period as 25 years, in time for many readers of Technology Review to avail themselves of its formulations -- and, not incidentally, in time for his 41-year-old self as well. Like Bacon, de Grey has never stationed himself at a laboratory bench to attempt a single hands-on experiment, at least not in human biology. He is without qualifications for that, and makes no pretensions to being anything other than what he is, a computer scientist who has taught himself natural science. Aubrey de Grey is a man of ideas, and he has set himself toward the goal of transforming the basis of what it means to be human.

De Grey has set himself up as just the gadfly to prod biologists to make it happen. Consider this essay, where he describes researchers on aging as "catatonic" for standing mute before his agenda.

And an agenda is what SENS is. The central idea is that the causes of aging are finite. De Grey lists seven basic causes, and argues that there likely aren't any more to be found, since no new ones have emerged in over twenty years. The seven range from an accumulation of mitochondrial mutations to programmed cell death and the breakdown of extracellular support structures. SENS seeks to find ways to nullify these seven causes of aging.

Not bad for a start. De Grey and others have identified in principle some ways that each of the seven causes could be stopped. It's when you read into these details that you begin to comprehend the breathtaking ambition behind the agenda. How to solve the problem of mtDNA mutations? No problem --- just replace the body's stem cells with others that have the 13 mtDNA genes present in the nucleus. And find ways to make them active in the right amounts. And work out a way to get them across the mitochondrial membrane. The thing is, they have ideas about how all these things might be done. The only barrier is, of course, is finding out whether all these steps will work.

The main focus is cellular aging, but to the extent that stem cells might be able to replace tissues with younger, more functional versions, the focus is probably what is needed. But there are plenty of stumbling blocks. Cells are Rube Goldberg devices, with complex sequences of protein interactions necessary to make them work. Aging happens because these sequences have only been tuned by evolution over a short lifespan. It just doesn't matter that much to most organisms' reproduction if their cells break down and stop working after a certain period of time. It can even be a good thing to have them die on purpose -- when cells just refuse to follow their molecular controls, we call it cancer.

Many of de Grey's proposed solutions would have us sweep aside these complex mechanisms, starting from scratch with new genetically customized replacement stem cells. That idea would limit to some extent what you would need to know about gene interactions -- it's like those custom kits that let you make a Ford Pinto look like a '26 Packard.

Perhaps the way to go, but is there some critical flaw? There are lots of interesting evolutionary reasons why aging should be an intractable problem, but it looks to me like many problems are swept away by the assumption that stem cells can be engineered to possess or lack any feature de Grey would require. The challenge may flush out somebody with a better idea of why it wouldn't work.

It will be interesting to see if anyone takes up the challenge. Personally, I wish somebody would offer a prize to show some anthropological research project is not worth serious consideration. Of course, with my luck it would be one of mine....