Tuesday, October 18, 2011

Genes that ensure a long life? Not likely

The genome of a woman who was 115 years old when she died has been sequenced, and according to a report in the BBC (the analysis hasn't yet been published, but the results were reported at the recent Human Genetics meeting in Montreal), the woman was the oldest in the world at the time of her death, and had no signs of dementia, and in fact when tested for cognitive ability, was found to have the thinking facility of a woman in her 60's or 70's.

According to one of the geneticists involved, the woman had some "rare genetic changes" in her DNA.
Dr Holstege told the BBC: "We know that she's special, we know that her brain had absolutely no signs of Alzheimer's.
"There must be something in her body that is protective against dementia.
"We think that there are genes that may ensure a long life and be protective against Alzheimer's."
Ok, let's pick these few statements apart.

First, we all have "rare genetic changes" in our DNA, so in fact she should get no points for this. 

Second, is dementia really the default state for the elderly?  That is, is it right to conclude that  someone over 100 without dementia must have some innate protection against it?  A quick search suggests that the data vary wildly, and aren't strictly comparable -- a small study in Oregon found that almost 40% of the study group of people over age 97 were dementia-free, while a compilation of data from around the world, published in The Lancet in 2005 suggests that dementia comes nowhere close to being the default state in people 85 or older.  So let's say for the sake of argument that the prevalence is somewhere in between both numbers -- still, dementia is pretty clearly not the default state.

Table 1
Group mean consensus estimates (SD) for prevalence of dementia (%) for each region and age-group

Age-group (years)

0·9 (0·1)  
1·5 (0·2)
  3·6 (0·2)
  6·0 (0·2)
12·2 (0·8)
   24·8 (1·0)
0·9 (0·1)
1·3 (0·1)
  3·2 (0·3)
  5·8 (0·3)
12·2 (0·3)
24·7 (2·3)
0·9 (0·1)
1·3 (0·1)
  3·2 (0·2)
  5·8 (0·2)
11·8 (0·5)
24·5 (1·8)
0·8 (0·1)
1·7 (0·1)
  3·3 (0·3)
  6·5 (0·5)
12·8 (0·5)
30·1 (1·1)
0·8 (0·1)
1·7 (0·1)
  3·4 (0·2)
  7·6 (0·4)
14·8 (0·6)
33·2 (3·5)
0·7 (0·1)
1·5 (0·3)
  2·8 (0·4)
  6·2 (1·1)
11·1 (2·0)
28·1 (5·2)
0·9 (0·3)
1·8 (0·1)
  3·5 (0·3)
  6·6 (0·2)
13·6 (0·8)
25·5 (2·3)
1·2 (0·3)
1·9 (0·2)
  3·9 (0·3)
  6·6 (0·4)
13·9 (1·3)
23·5 (2·3)
0·6 (0·1)
1·4 (0·1)
  2·6 (0·3)
  4·7 (0·6)
10·4 (1·2)
22·1 (3·5)
0·6 (0·1)
1·8 (0·2)
  3·7 (0·4)
  7·0 (0·9)
14·4 (1·9)
26·2 (3·9)
1·0 (0·1)
1·7 (0·2)
  3·4 (0·2)
  5·7 (0·5)
10·8 (1·2)
17·6 (2·7)
0·4 (0·1)
0·9 (0·1)
  1·8 (0·2)
  3·7 (0·4)
7·2 (1·2)
14·4 (2·7)
0·3 (0·1)
0·6 (0·1)
  1·3 (0·2)
  2·3 (0·5)
4·3 (1·0)
9·7 (1·9)
0·5 (0·3)
1·0 (0·4)
  1·9 (0·9)
  3·8 (1·7)
7·0 (3·6)
14·9 (7·2)

And the fact that prevalence varies by region, being highest in developed countries, suggests either that diagnosis or record-keeping are less reliable in developing countries, or that risk increases with level of development.  If the latter is true, the idea that something innate protects against dementia -- genes, as the researcher suggests -- is less convincing.

Finally, as reported in the BBC story, "Further work [on this woman's genome] could give clues to why some people are born with genes for a long life, says a UK scientist."

Now, we noted earlier that whole genome sequences, of which there are now quite a few, have shown that we all have 'rare' variants.  At known mutation rates, we all have about 155 brand-spanking-new mutations.  Most, of course, will be in non-functional places (so far as 'function' is known).  But a few will be in known functional sites.  But the Methuselady paper suggests that she had rare longevity variants, and raised the hope that we'll find it by looking, blearily, at all 6.2 billion nucleotides in her genome.

More important by far is the flip side of this story.  With few if any exceptions, every person from whom we have whole genome DNA sequence has many variants that inactivate a gene, usually some known 'disease' variants, and large numbers of totally knocked-out (inactivated) genes...in both copies.  Yet the samples are from middle-aged or older people who don't have any particular disease.  So no one should be saying with a straight face that this lady's genome is going to reveal the Gene for Immortality.  Single variants rarely cause major traits in this sense.

Given what we already, clearly know, can there be genes "to ensure a long life"?  How would they have evolved?  Natural selection can only act on traits that affect fitness, reproductive success.  That is, traits that are apparent during reproductive years, because it's only by having more offspring than people without a given trait that a trait can become more common.  Longevity is not one of those traits. Especially decades beyond menopause.

But, ok, let's say that this woman did in fact have genes 'for a long life' -- even then it's unlikely they will be alleles that are found in significant frequency in the population.  Thus, looking for, or even finding them, in this woman's DNA will not tell us much at all that can be generalized to the rest of us.

Aging is an interesting phenomenon and there are some really important questions about it, and even some good work being done.  For example why do mice, with a largely similar genome as humans, live 2 or 3 years, while we can live over 100?   With orders of magnitude fewer cells at risk, why do mice get cancer at an age-accelerating rate that seems similar in shape to that of human cancers?  And if there is a random component in life -- and certainly at least there is the luck involved in not being affected by lethal infections, accidents, and so on -- then how do we calibrate our view of how likely is a single long-lived person is to have been so for some 'causal' reason, genetic or otherwise?

These are the scientific questions.  Instead of serious science, what this kind of warm human interest story tells us most is about the assumptions first, that everything has to be genetic, and second, even more obvious, is that everything has to be milked for its publicity and grant-justifying value.


Holly Dunsworth said...

This is a terrific post.

Makes me wonder if there's any difference between explaining a one-off like this with "DNA did it" as opposed to with "God did it."

Also, rearing offspring contributes to fitness so traits can be selected for post-reproduction for sure. How far can selection reach down the road past parenthood and towards death? That's the question to my mind. (Not really relevant at all for Methusalady, just thinking about longevity.)

Ken Weiss said...

'Science', the western-culture-derived method of understanding Nature from a materialistic perspective, is not good at one-off events. It grew out of induction--generalizing from repeated observations of similar cause-effect pairings.

In the 19th and 20th centuries, statistical reasoning was introduced as the formal basis of much of 'the scientific method'. Statistical reasoning is based on probabilities, another kind of repeatability or induction-based concept.

At least we can hope to construct hypotheses based on one-off observations---if we can set up situations where we can observe the 'same' event somehow (like the effect of some particular variant we may see in the Aged Lady.

Rearing offspring contributes to reproductive success in many ways, but that isn't the same as fitness if it is not based on genes. The effect of parenting is often taken into account indirectly by estimating fitness as the number of offspring who reach adulthood.

The older you are the less your net or average effect of your genotype on the future fates of the variants you carry, which is why the 'grandmother hypothesis', that we live a long time because we can care for our g'children and hence enhance their fitness, is problematic.

There were tons of research done to try to explain how selection could favor people living well past fertility (if that's what your question referred to). The upshot is related to my previous point. Selection favoring the survival of the hoary is very weak, if it works at all.

Fortunately, there is little problem here because humans are not very different, for their body size, from other mammals in terms of their longevity. For what we were doing in the past, maybe in terms of how rapidly or not we grow to adulthood (or how immature at birth), our life-clock evolved appropriately. Likewise, there was no 'need' for mice to live more than they do (about 2 years), and they are on the same body-size/longevity line as we are.

This has been a point very difficult for people on the NIH gravy train or anthropologists wanting some exceptionalism for humans, to digest.

Anne Buchanan said...

Thanks, Holly. Agreed, "DNA did it" can be as faith-based as "God did it".

Erick Howenstine said...

Problem almost boils down to a sample size of 1. However, in the past, before written record, the elderly would have provided useful information with survival benefits. And birds are also small like mice, but can live 100 years. It has to do with metabolism required for flight, I've read.

Ken Weiss said...

The survival information you refer to arguably would accrue to the group, not to the individual, and it has to have sufficient impact on fitness to overcome genetic drift, to be of any import. Also, in the past, if you mean the 'indigenous' long pre-agricultural past, which is the vast majority of our evolutionary history as a species distinct from others, the fraction of people who lived to really old age would have been very very small. That raises the relative impact of genetic drift vis-a-vis selection. And opposed to this is the burden that caring for the elderly imposes. There are many reasons why that could be far greater than the benefit of elders' wisdom. If anything, I would guess that baby-sitting services would be the greatest factor. But elders seem often to have baby-sat for multiple children whose mothers were out gathering. So personally, at least, I find the longevity evolution arguments to be forced.

As to birds and some other examples,they have long raised questions about simple, or simplistic, models of the biology and evolution of aging and lifespan. Those questions aren't really answered, as far as I know, but I haven't worked directly in aging and its evolution for a number of years, and am undoubtedly out of date (if not too old).

Erick Howenstine said...

No not the group, the elders would be helping kin specifically. I rarely see group selection, though I try. But in addition to the baby sitting, gathering, care and knowledge they may provide, care of elders may result from the general growth of empathy, which would have other more direct bemeficial effects.

But care won't make them live longer, it's the benefit of the elderly, not benefit to the elderly which would matter. So it is a mystery, still. However I don't see the countervailing force of drift. Would you explain?

BTW my ref on the birds, I believe is Nick Lane's Power Sex. Suicide

Ken Weiss said...

I was not referring to group selection, but to the idea that in small demes these kinds of things like grandparent babysitting are done not just for close relatives. The point about drift is that when the selective advantage is very small in a small group, chance can be as important as that advantage in terms of the actual future fate of the responsible genes. This is just the mathematical consequence of genetic sampling.

Anyway, these are long-standing differences of view

EllenQ said...

I happened to be in that talk and there were a couple things brought up that weren't covered much in the published accounts. The dementia-free/plaque-free claim was bolstered by some specific brain function tests where she performed 30-70 years better than expected for age.

The most problematic aspect, to me, was that she maintained a lifestyle that would be most likely to increase lifespan for anyone. She didn't drink, smoke, or have children. She kept her brain active. She had been planning to devote her body to science upon her death since the 1950s and kept up on popular science literature. While she may have some genetic variants that kept her alive, the lifespan tradeoffs that may or may not have to do with having kids or a heavy disease burden were not influencing her longevity.

Personally, I think looking at people like Ozzy Osbourne (also sequenced) is more informative as he has done many many things that should have killed him by now.

Anne Buchanan said...

The poor woman, she forwent all the fun in life, and then had to live to 115! As for Ozzy Osbourne, probably somatic mutations would be of more interest!

Erick Howenstine said...

Why wouldn't a selective effect still work on the face of drift, I wonder just more slowly, maybe much more slowly. But as you say, we can agree to disagree on this point, there is so much more we obviously agree on anyway.

Ken Weiss said...

Just to finish up. The effect of selection is measured by a selection coefficient, call it 's', relative to a given genetic variant, compared to other variants in the same population. Relevant facts are how systematic or steady that 's' value is, how small the population is, how much genetic identity (kinship) there is among the members, what the mating pattern is in this regard, the average number of offspring per parent, the social dynamics of the effect (e.g., whether directly or indirectly elders help other grandchildren as well as their own), and so on.

A parent is related by half to each of its children, but only by a quarter to each grandchild. That reduces the relative impact of any gene related to longevity leading to grandparenting.

Reproduction is highly probabilistic, and it's just the mathematics of probability by which in a small group, small selective effects have only about the same chance of resulting in fixation (or loss, if the effect is negative) relative to what would happen by drift alone.

If in one generation an allele helps advance the reproductive success of a grandchild because of its systematic effect on the longevity of the grandparent, it could (for example) have the opposite effect in other times, such as of resource shortage, and actually hamper the grandchild's prospects.

The situations are complex, but since humans are not really all that longevous relative to other species in protective environments like ours, there is not that much (if any) effect to explain.

Mathematical simulations (that I've been aware of, at least) have had to specify rather specific situations for this to work as an effective selective force. That doesn't make it impossible, but the effect of grandparenting (or analogous contributions of elders) could work in purely cultural ways, or some have argued in group-selection ways.

And the potential effectiveness for staying healthy up to, roughly, menopausal ages to rear slow-growing human offspring are mathematically much larger and thus, if any selective explanation is needed, much more likely to account for calibrating our life histories.

At least, these are roughly my reasons for holding the view on this that I do.