Causes of molecular damage

Ageing is the progressive loss of physiological function that leads to decreasing fertility and increasing mortality with time.Reference Partridge and Gems⁴ It is driven by the accumulation of molecular damage to the organism; without this, intrinsic lifespans would not have evolved in any species. Lipids, proteins and nucleic acids (the three main biological macromolecules) accumulate damageReference Beckman and Ames⁵ and this can cause organelles and then cells to malfunction,Reference Partridge and Gems⁴ potentially leading to apoptosis.Reference Beckman and Ames⁵ This damage accumulates with time and is thought to manifest itself at the level of the organism as ageing. The exact nature of this damage (i.e. which molecules are important, the cause of the damage etc.) is unclear. Whether there is one unified mechanism of ageing or various separate ones,Reference Partridge⁶ or even only processes which one might label as ‘degenerative disease’Reference Peto and Doll⁷ is not certain. Furthermore, the distinction can be unclear.Reference Krause⁸ Nevertheless, some process which amounts to damage to the organism at some level must exist for animal species to have evolved to have an intrinsic lifespan, otherwise there would be no intrinsic limit to lifespan at all.

A potential mechanism for this damage is described by the free radical theory of ageing,Reference Harman⁹ which many propose as a universal theory of the mechanism of ageing.Reference Dowling and Simmons¹⁰ Telomere shortening is another potential mechanism.Reference Sahin and DePinho¹¹ More recently, DNA damage has been proposed as a final common pathway linking mechanisms of damageReference St Laurent, Hammell and McCaffrey¹² (DNA damage is a clear link, for example, between free radical damage and telomere shortening). However, individual genes can affect the ageing process (and therefore lifespan),Reference Partridge and Gems⁴ and so ageing is not just the result of ‘inevitable damage accumulation’. The mechanisms of free radical damage and/or telomere shortening, and their relationship to lifespan, is undoubtedly complex, and may well be linked to the numerous genetic pathways which have now been identified as influencing ageing.Reference Kenyon¹³ Suffice it to say that by some or all of these mechanisms, cellular damage occurs which provides a constraint within which a trade-off must occur for optimum survival of the gene line.

Evolutionary theories of ageing

The evolutionary theories of ageing are about examining the survival value of a particular lifespan. Though molecular damage is inevitably a problem that organisms must confront, it was not inevitable that no mechanism to completely control molecular damage evolved in any species. The evolutionary theories look at why no mechanism has evolved to completely control molecular damage, and hence why the vast majority of species have intrinsic lifespans.

The earliest evolutionary theory of ageing was that of Weismann,Reference Weismann¹⁴ who took the now discredited group selection (‘for the good of the species’) approach to suggest that ageing and death had evolved as a programmed mechanism to purge older individuals in order to increase the resources available for the better adapted individuals of subsequent generations.Reference Rose, Burke, Shahrestani and Mueller¹⁵ It has subsequently been shown that evolution rarely acts at the level of the group, because an altruistic trait for group benefit is open to exploitation by individuals, and so is not ‘evolutionarily stable’. These exploitative individuals produce the most offspring and so they are naturally selected, not those whose behaviour is for group benefit.

As well as this theoretical objection, the argument for ageing being programmed (implying a programmed lifespan) is thin. Partridge and Gems summarize the argument for programmed ageing as being that both the massive interspecies variation in lifespan, and the fact that single-gene mutations can cause marked changes in lifespan, show that ageing and lifespan must be a genetically determined – and therefore programmed – process.Reference Partridge and Gems⁴ Genes certainly do influence lifespan – though it is wrong to think of lifespan as being ‘genetically programmed’. No genes are known which specifically cause damage that leads to ageing.Reference Partridge and Gems⁴ Instead, genes influence life history strategies (such as reproduction schedules), or molecular mechanisms (such as those for cellular repair) that may indirectly influence the rate of ageing. Lifespan has not evolved ‘for the good of the species’, and is not genetically ‘programmed’, though it is of course influenced by many genes.

The two classic evolutionary theories of ageing are Medawar's mutation accumulation theoryReference Medawar¹⁶ and Williams’ theory of antagonistic pleiotropy.Reference Williams¹⁷ Kirkwood's disposable soma theoryReference Kirkwood¹⁸ is similar to the antagonistic pleiotropy theory, which may explain why it is rarely discussed as a separate entity. All three theories have in common the notion that the force of natural selection is greater on events happening early in an organism's life. This idea comes from the observation by Haldane that the invariably fatal dominant mutation that causes Huntington's disease is only passed on because the effects strike in people's thirties or forties, after most people have already had children.Reference Haldane¹⁹

Medawar developed this into the mutation accumulation theory of senescence, which says that the force of natural selection to purge deleterious genes weakens as the age at which the gene is expressed increases, and so across the generations late-acting deleterious mutations will accumulate in the genome, causing ageing and death. Hamilton mathematically justified the mutation accumulation theory,Reference Hamilton²⁰ though there are some unresolved issues.Reference Bonsall²¹

The antagonistic pleiotropy theory suggests that some genes that produce effects that are beneficial to survival value early in life (i.e. early fecundity or decreased early mortality) have late-acting deleterious effects that contribute to ageing. Pleiotropy refers to the fact that one gene can have multiple phenotypic effects. In the case of antagonistic pleiotropy these are temporally spaced and have opposing effects on survival value such that there is a trade-off. As with the mutation accumulation theory, natural selection acts more strongly on early effects: therefore the trade-off between the early beneficial effects and the late deleterious effects is resolved in favour of maintaining the genes. An example might be genes for testosterone in men. These genes are obviously vital for reproductive success, but the negative side of the pleiotropic trade-off is the risk of prostate cancer in later life, which has been linked to high levels of testosterone.Reference Bribiescas, Ellison, Stearns and Koella²²

Despite the manner in which some researchers (such as Partridge and GemsReference Partridge and Gems⁴) discuss these two theories, it is not the case that they are mutually exclusive, and indeed evidence does exist to support both theories to some extent. It is, however, difficult to assess the magnitude of the contribution of each explanation for ageing to the overall determination of lifespan.Reference Partridge and Barton²³

Evidence for mutation accumulation is mixed. Much work has concentrated on the prediction that the genetic contribution to mortality and declining fertility should increase with age (as the effects of deleterious genes are more likely to be expressed at later ages). Researchers have therefore looked for an increase in additive genetic variance for mortality and fertility with increased age. Rose and Charlesworth found no effect in Drosophila Reference Rose and Charlesworth²⁴ (though BonsallReference Bonsall²¹ identifies ‘methodological issues’) and some other studies showed the opposite effect.Reference Promislow, Tatar, Khazaeli and Curtsinger^25, Reference Shaw, Promislow, Tatar, Hughes and Geyer²⁶ However, Hughes and CharlesworthReference Hughes and Charlesworth²⁷ and then Hughes et al.Reference Hughes, Alipaz, Drnevich and Reynolds²⁸ showed an increase in genetic variance in mortality rate with age, and an increase in variance in age-specific reproductive success. Strangely, Partridge and Gems ignore these studies when they conclude that the evidence for mutation accumulation is poor.Reference Partridge and Gems⁴ Bonsall, by contrast, concludes there is ‘strong support for the mutation accumulation theory’.Reference Bonsall²¹ Mutation accumulation has a sound theoretical basis, and there is some evidence to suggest that it does play a role in the evolution of ageing and lifespan, though the extent and nature of this role is unclear. It seems strange that no more direct way of investigating mutation accumulation than examining additive genetic variance has been pursued.

Evidence for the role of antagonistic pleiotropy in the evolution of ageing and lifespan comes from artificial selection experiments in Drosophila. Comparisons have been made between lines bred only from the offspring of old Drosophila (so the line has genes for longevity) and lines from offspring only from young Drosophila (where longevity will have been less strongly selected). Antagonistic pleiotropy predicts that those who live longer (i.e. the ‘old’ line) will have to ‘pay’ for their increased longevity with reduced early lifetime fecundity, and this has been shown to be the case with late fertility and longer lifespans in the ‘old’ line, and earlier reproduction and shorter lifespans in the ‘young’ line.Reference Partridge and Gems^4, Reference Bonsall²¹

Confusion, however, emerges over what counts as pleiotropy due to observations from subsequent experiments. For example, sterilization of females in the ‘young’ Drosophila line abolishes the more rapid ageing in that line.Reference Partridge and Gems⁴ This indicates that rather than ageing being a late expression of the genes which increase early fecundity, it is in fact a latent, but direct effect of the damage caused by reproduction. Further experiments have confirmed this direct but latent effect of reproduction: restriction of egg-laying by limitation of oviposition sites, and the act of mating itself can shorten female lifespan.Reference Partridge and Barton²³ Partridge and Gems still consider this to be antagonistic pleiotropy,Reference Partridge and Gems⁴ presumably because these are distinct phenotypic effects resulting from a behaviour (mating when young) that is genetically influenced. However, Westendorp and Kirkwood use the increased longevity of flies that are prevented from mating as an example of a non-genetic correlation between longevity and fecundity.Reference Westendorp and Kirkwood²⁹

The confusion is because these researchers are looking at different levels. Antagonistic pleiotropy is a useful way of thinking about the evolution of these behaviours and lifespan, as Partridge and Gems discuss; however, in an individual, behavioural interventions can overrule the evolved behaviour, and this elucidates the non-genetic mechanism of the pleiotropic effect, to which Westendorp and Kirkwood seem to be referring. Therefore these experiments do still support the role of antagonistic pleiotropy in the evolution of lifespan, and further reveal the trade-offs that occur as a result of behaviour in an individual.

Unusually, given that most studies have been in Drosophila, there is evidence for this trade-off between reproduction and lifespan from human data. Westendorp and Kirkwood used a historical data set from aristocratic British families to show that amongst women who lived over 60 years (to exclude the effect of maternal death during childbirth), female longevity was negatively correlated with number of offspring.Reference Westendorp and Kirkwood²⁹

The experiments reported above also support the disposable soma theory of ageing, because they only reveal the trade-off principles, which are common to both antagonistic pleiotropy and the disposable soma theory. The disposable soma theory of ageing is that the demands of reproduction mean that not enough can be invested in somatic maintenance to prevent ageingReference Kirkwood and Rose³⁰ – i.e. genes which support reproduction (a beneficial early effect) are selected at the expense of repair (a detrimental late effect). This is achingly similar to antagonistic pleiotropy, and sometimes they are now lumped together as one ‘optimality theory’Reference Partridge and Barton²³ or ‘trade-off theory’.Reference Partridge and Gems⁴ The important conclusion of the experiments described is that they confirm there is a trade-off between reproduction and lifespan – manifested evolutionarily through antagonistic pleiotropy and mechanistically through the fact that increased reproduction reduces lifespan. The mutation accumulation and trade-off theories are not mutually exclusive, and given their sound theoretical bases and the supporting evidence for each theory, it seems likely that each contributes to the reasons why ageing occurs and lifespan is limited. It is difficult to determine the proportions of the contribution of each mechanism to intrinsic lifespan, and no-one has suggested a full description of this. However, trade-off theories may describe the primary reason why a particular lifespan was naturally selected, but mutation accumulation is a potentially important effect that is determined by the trade-off. Mutation accumulation only exists because individuals have offspring before the deleterious genes express their effects. The timing of reproduction is determined by the survival–reproduction trade-off, and therefore the trade-off theories determine the timing at which mutation accumulation occurs. Therefore the trade-off theories are the primary explanation for the life history strategy that determines lifespan, but the timing of mutation accumulation is an effect of that strategy, and mechanistically may be important in limiting lifespan. Indeed, given that many of the trade-off effects studied can be abolished behaviourally (such as by sterilization), it in fact does not seem that the pleiotropic effects are very intrinsic – mutation accumulation could be a more significant intrinsic cause of lifespan limitation.

Effects of extrinsic mortality

The evolutionary theories show why there is a lifespan, but it is extrinsic causes of mortality that determine the terms of the survival–reproduction trade-offs in a particular species such as humans, and so it is extrinsic mortality risks which has largely determined the interspecies variation in lifespans.

Extrinsic causes of mortality, such as accidents, disease or predation, shift the balance of the survival–reproduction trade-off away from prolonged survival and towards reproduction. The greater the extrinsic risk of mortality, the greater the shift away from adult survival towards reproduction. This effect is ‘predicted’ by the evolutionary theories of ageing,Reference Partridge and Gems⁴ and so evidence showing this effect of extrinsic risk can be said to support the evolutionary theories. According to the trade-off theories, a high extrinsic risk of mortality means that there is little investment in molecular repair mechanisms as this is likely to be a waste of resources since the individual will probably die young anyway. By contrast, low extrinsic mortality means that there is greater selection for repair mechanisms to retard senescence, and less advantage in fast, early reproduction. Mutation accumulation theory is also generally regarded as predicting this effect of extrinsic mortality – early death means that there would be no selection pressure to purge late-acting mutations.Reference Partridge and Gems⁴

Studies in fruitfliesReference Stearns, Ackermann, Doebeli and Kaiser³¹ and antsReference Keller and Genoud³² have found support for this theory. Reznick et al. found the opposite effect in guppies,Reference Reznick, Bryant, Roff, Ghalambor and Ghalambor³³ but their study does not disprove the predicted effects of extrinsic mortality, but rather highlights that the situation is complex. They showed that guppies taken from high predation areas live longer than others – but high predation could reduce the guppy population and so increase resources such as food for remaining guppies – and greater food resources would reduce the extrinsic risk of starvation.Reference Bronikowski and Promislow³⁴ High predation may also select for faster, fitter guppies, which then live longer.Reference Bronikowski and Promislow³⁴ Bronikowski and Promislow point out that the original prediction by Williams only applies to ‘random, condition-independent events’.Reference Bronikowski and Promislow³⁴ It seems possible that the intraspecific effects (like those demonstrated by Reznick et al.) may be different to the interspecific effects which we are interested in when comparing interspecies lifespans (for example, condition-dependent events may be more relevant to intraspecific competition).

This notion of extrinsic mortality risk ties in with the idea of a fast–slow continuum of species.Reference Read and Harvey³⁵ At the fast extreme there are small animals such as mice. These are often at a higher extrinsic risk due to factors such as predation. They therefore tend to show rapid development, have large numbers of small offspring, and have short intrinsic lifespans. At the slow extreme are animals such as elephants. They tend to be large and have long gestation and developmental periods, so they tend to produce few offspring. Humans are a ‘slow’ species, and so have long intrinsic lifespans.

However, though humans are clearly a ‘slow’ species, intrinsic human lifespan is actually a lot longer than one would predict from size alone. This comparative longevity is shared by other species, including birds, bats, bivalve molluscs, tortoises, ants and naked mole rats. The explanation lies in the ecology of each species: each has evolved a phenotype that markedly reduces extrinsic mortality. These species in themselves provide good evidence for the fact that extrinsic mortality influences lifespan. For example, birdsReference Partridge and Barton²³ and batsReference Brunet-Rossinni and Austad^36, Reference Partridge and Barton²³ tend to live much longer than their size would predict. It seems probable that the ability to fly is important, as this improves the chance of escaping from predators. Flightless birds show poor survival both in nature and under optimal conditions, lending further support to this theory.Reference Partridge and Barton²³ Molluscs and tortoises have hard shells to protect them.Reference Partridge and Barton²³ Ants and naked mole rats both live in large colonies where the physical structure helps defence;Reference Bourke and Franks^37, Reference Sherman and Jarvis³⁸ they also show co-operative behaviour in terms of defence and brood care.Reference Bourke and Franks^37, Reference Sherman and Jarvis³⁸ Their breeding is through a long-lived ‘queen’ who lives deep in the colony and who suffers a very low level of extrinsic risk.Reference Crozier³⁹ Naked mole rats in captivity have regularly lived over 20 years,Reference Sherman and Jarvis³⁸ which is remarkable given that comparably sized mice live for about 3 years.Reference Vijg and Campisi³ All this provides very good observational evidence for how a low risk of extrinsic mortality can lead to the evolution of a long intrinsic lifespan.

Distinctive features of human ecology mean that humans suffer from a very low risk of extrinsic mortality, and so our long intrinsic lifespan has been selected in accordance with the evolutionary theories of ageing. The unique feature of the human species is its unusually large brain size compared with its body size, and it is interesting that a correlation between brain size and longevity has long been known about.Reference Rose and Mueller⁴⁰ With this enlarged brain come associated cognitive capacities, and it seems that these capacities allow humans to reduce extrinsic mortality rates.Reference Rose and Mueller^40, Reference Charlesworth⁴¹ The ability to outwit predators would reduce predation hazard, and the ability to seek out and extract alternative food sources would reduce the hazard due to famine and disease.

The human brain is also vital for language and culture, which again are likely to reduce extrinsic mortality risks through the communication of information about environmental hazards. Communication of information has also allowed the accumulation of knowledge (by ‘standing on the shoulders of giants’) that has contributed to the rise of science and technology. Even at its most primitive level, science and technology will have reduced extrinsic mortality. This theory directly links a key feature of humans with the broader evolutionary theories of ageing, and so is a strong contender for explaining the long lifespans of humans.

Intergenerational resource transfers

Despite general support for this theory, the evolutionary theories did not seem to explain the uniquely human phenomenon of the menopause – why should our lifespan extend beyond the point at which reproduction is possible? Attempts to explain this have brought to prominence the role of intergenerational resource transfers in shaping lifespan.

Hawkes et al. have suggested that the effect of grandmothering selects for extended lifespan beyond childbearing age,Reference Hawkes, O'Connell, Blurton Jones, Alvarez and Charnov⁴² and this can still be described in relation to the classic evolutionary theories of ageing. Supporting grandchildren is of benefit to the survival of the grandmother's gene line – it contributes to her ‘inclusive fitness’ (the total ‘fitness’ or ability to pass on its genes, which includes the reproductive success of genetically related individuals), but initial mathematical analysis showed that the benefits of gene survival from grandmothering did not outweigh the potential benefits of continuing to reproduce.Reference Cant and Johnstone⁴³ However, Cant and Johnstone updated the model taking into account the changes in intergenerational competition brought about by the unusual fact that in human societies (as typified in many hunter-gatherer communities) it is generally the females that disperse to join other social groups rather than males, and this model survives mathematical analysis.Reference Cant and Johnstone⁴³

LeeReference Lee⁴⁴ and Kaplan and RobsonReference Kaplan and Robson⁴⁵ consider transfers from parents to offspring as well as from grandparents, and also examine why transfers have evolved. Both papers argue that human ecology means that heavy investment in a few offspring is optimal, and that this investment is only paid off if lifespan is long.

Lee suggests that ‘only the transfer effect shapes mortality’ in humans and in other species where there is considerable resource transfer. However, his model relies on an optimal allocation of resources between levels of fertility and investment per offspring – and it seems possible that this optimal allocation will have evolved because of the survival–reproduction trade-off. Equally, the theory depends on a situation where heavy investment in offspring is optimal, and such a situation is brought about under the notion of a fast–slow continuum of species, which is linked to the classic evolutionary theories through the effects of extrinsic mortality. It therefore seems that the transfer effect is one that occurs on top of the classic trade-off to determine human lifespan.

Kaplan and Robson look specifically at human ecology. They argue that humans evolved in a challenging niche that favoured large brains. They show that hunter-gatherer humans obtain 92% of their food from ‘extracted’ or hunted resources. This requires a high skill level and so requires a large brain. By contrast, 95% of chimpanzee calories come from sources readily gathered by hand. Lifetime calorie production profiles show that human cumulative lifetime net production is greater than that of a chimpanzee (demonstrating the comparative success of the human strategy), but that this net benefit comes comparatively late in life after a prolonged development period of 20 years (four times that of a chimpanzee) where net investment from adults is required. So, the human niche required large brains, but these only have a survival value if lifespan is long. Therefore they argue that the long human lifespan co-evolved with brain size.

The emphasis placed on brain size has clear echoes of the importance of brain size in reducing extrinsic mortality costs, but the key fact for Kaplan and Robson is not the benefit of having a large brain, but the fact that the benefit takes a long time to pay off its investment. However, it seems clear that the main benefit of having a large brain is in reduction of extrinsic mortality risks – and this benefit is one that will only pay off its investment in a long life. As such, the investment and transfer theory may describe an evolutionary force that is in synergy with the classic evolutionary theories in describing the evolution of the extraordinarily long human lifespan.

Conflicts of interest

The author declared no conflicts of interest.