Evolution of ageing

Old man at a nursing home in Norway.

Enquiry into the evolution of aging aims to explain why almost all living things weaken and die with age. There is not yet agreement in the scientific community on a single answer. The evolutionary origin of senescence remains a fundamental unsolved problem in biology.

Historically, ageing was first likened to "wear and tear": living bodies get weaker, shoes get wrecked with use or, with exposure to air and moisture, iron objects rust. But this idea was discredited in the 19th century when the second law of thermodynamics was formalized. Entropy (disorder) must increase inevitably within a closed system, but living beings are not closed systems. It is a defining feature of life that it takes in free energy from the environment and unloads its entropy as waste. Living systems can even build themselves up from seed, and routinely repair themselves. There is no thermodynamic necessity for senescence. In addition, generic damage or "wear and tear" theories could not explain why biologically similar organisms (e.g. mammals) exhibited such dramatically different lifespans. Furthermore, this initial theory failed to explain why most organisms maintain themselves so efficiently until adulthood and then, after reproductive maturity, begin to succumb to age-related damage.

History

August Weismann was responsible for interpreting and formalizing the mechanisms of Darwinian evolution in a modern theoretical framework. In 1889, he theorized that aging was part of life's program because the old need to remove themselves from the theatre to make room for the next generation, sustaining the turnover that is necessary for evolution. This theory again has much intuitive appeal, but it suffers from having a teleological or goal-driven explanation. In other words, a purpose for aging has been identified, but not a mechanism by which that purpose could be achieved. Aging may have this advantage for the long-term health of the community; but that doesn't explain how individuals would acquire the genes that make them get old and die, or why individuals that had aging genes would be more successful than other individuals lacking such genes. (In fact, there is every reason to think that the opposite is true: aging decreases individual fitness.) Weismann later abandoned his theory.

Theories suggesting that deterioration and death due to aging are a purposeful result of an organism's evolved design (such as Weismann's "programmed death" theory) are referred to as theories of programmed aging or adaptive aging. The idea that the aging characteristic was selected (an adaptation) because of its deleterious effect was largely discounted for much of the 20th century, but a theoretical model suggests that altruistic aging could evolve if there is little migration among populations.

Mutation accumulation

The first modern theory of mammal aging was formulated by Peter Medawar in 1952. It formed from discussions in the previous decade with J. B. S. Haldane and the selection shadow concept. Their idea was that aging was a matter of neglect. Nature is a highly competitive place, and almost all animals in nature die before they attain old age. Therefore, there is not much reason why the body should remain fit for the long haul – not much selection pressure for traits that would maintain viability past the time when most animals would be dead anyway, killed by predators, disease, or accident.

Medawar's theory is referred to as Mutation Accumulation. The mechanism of action involves random, detrimental germline mutations of a kind that happen to show their effect only late in life. Unlike most detrimental mutations, these would not be efficiently weeded out by natural selection. Hence they would 'accumulate' and, perhaps, cause all the decline and damage that we associate with aging.

Modern genetics science has disclosed a possible problem with the mutation accumulation concept in that it is now known that genes are typically expressed in specific tissues at specific times (see regulation of gene expression). Expression is controlled by some genetic "program" that activates different genes at different times in the normal growth, development, and day-to-day life of the organism. Defects in genes cause problems (genetic diseases) when they are not properly expressed when required. A problem late in life suggests that the genetic program called for expression of a gene only in late life and the mutational defect prevented proper expression. This implies existence of a program that called for different gene expression at that point in life. Why, given Medawar's concept, would there exist genes only needed in late life or a program that called for different expression only in late life? The maintenance mechanism theory (discussed below) avoids this problem.

Medawar's concept suggested that the evolution process was affected by the age at which an organism was capable of reproducing. Characteristics that adversely affected an organism prior to that age would severely limit the organism's ability to propagate its characteristics and thus would be highly "selected against" by natural selection. Characteristics that caused the same adverse effects that only appeared well after that age would have relatively little effect on the organism's ability to propagate and therefore might be allowed by natural selection. This concept fit well with the observed multiplicity of mammal life spans (and differing ages of sexual maturity) and is important to all of the subsequent theories of aging discussed below.

Medawar did not suggest that there were fundamental limitations on lifespan. Organisms exhibiting negligible senescence suggest that aging is not a fundamental limitation, at least not in the scale of mammal lifespan.

Antagonistic pleiotropy

Main article: Antagonistic pleiotropy hypothesis

Medawar's theory was further developed by George C. Williams in 1957, who noted that senescence may be causing many deaths, even if animals are not 'dying of old age.' In the earliest stages of senescence, an animal may lose a bit of its speed, and then predators will seize it first, while younger animals flee successfully. Or its immune system may decline, and it becomes the first to die of a new infection. Nature is such a competitive place, said Williams, (turning Medawar's argument back at him), that even a little bit of senescence can be fatal; hence natural selection does indeed care; ageing isn't cost-free.

Williams's objection has turned out to be valid: Modern studies of demography in natural environments demonstrate that senescence does indeed make a substantial contribution to the death rate in nature. These observations cast doubt on Medawar's theory. Another problem with Medawar's theory became apparent in the late 1990s, when genomic analysis became widely available. It turns out that the genes that cause ageing are not random mutations; rather, these genes form tight-knit families that have been around as long as eukaryotic life. Baker's yeast, worms, fruit flies, and mice all share some of the same ageing genes.

Williams (1957) proposed his own theory, called antagonistic pleiotropy. Pleiotropy means one gene that has two or more effects on the phenotype. In antagonistic pleiotropy, one of these effects is beneficial and another is detrimental. In essence, this refers to genes that offer benefits early in life, but exact a cost later on. If evolution is a race to have the most offspring the fastest, then enhanced early fertility could be selected even if it came with a price tag that included decline and death later on. Because ageing was a side effect of necessary functions, Williams considered any alteration of the ageing process to be "impossible."

Antagonistic pleiotropy is a prevailing theory today, but this is largely by default, and not because the theory has been well verified. In fact, experimental biologists have looked for the genes that cause ageing, and since about 1990 the technology has been available to find them efficiently. Of the many ageing genes that have been reported, some seem to enhance fertility early in life, or to carry other benefits. But there are other ageing genes for which no such corresponding benefit has been identified. This is not what Williams predicted. This may be thought of as partial validation of the theory, but logically it cuts to the core premise: that genetic trade-offs are the root cause of ageing.

Another difficulty with antagonistic pleiotropy and other theories that suppose that ageing is an adverse side effect of some beneficial function is that the linkage between adverse and beneficial effects would need to be rigid in the sense that the evolution process would not be able to evolve a way to accomplish the benefit without incurring the adverse effect even over a very long time span. Such a rigid relationship has not been experimentally demonstrated and, in general, evolution is able to independently and individually adjust myriad organism characteristics.

In breeding experiments, Michael R. Rose selected fruit flies for long lifespan. Based on antagonistic pleiotropy, Rose expected that this would surely reduce their fertility. His team found that they were able to breed flies that lived more than twice as long as the flies they started with, but to their surprise, the long-lived, inbred flies actually laid more eggs than the short-lived flies. This was another setback for pleiotropy theory, though Rose maintains it may be an experimental artifact.

Disposable soma theory

A third mainstream theory of ageing, the Disposable soma theory, proposed in 1977 by Thomas Kirkwood, presumes that the body must budget the amount of energy available to it. The body uses food energy for metabolism, for reproduction, and for repair and maintenance. With a finite supply of food, the body must compromise, and do none of these things quite as well as it would like. It is the compromise in allocating energy to the repair function that causes the body gradually to deteriorate with age. A caveat to the disposable soma theory suggests that time, rather than energy, is a limiting resource that may be critical to an organism. The concept is that each organism must reproduce in an optimal period in order to ensure the greatest chance of success for the offspring. This optimal period is dictated by the ecological niche of the organism but in essence, it limits the time that any given organism can devote to growth and development prior to bearing offspring. Thus, developmental rate and gestational rate are subject to evolutionary pressure. The need to accelerate gestation limits the time allocated to damage repair at the cellular level, resulting in an accumulation of damage and a decreased lifespan relative to organisms with longer gestation. This concept stems from a comparative analysis of genomic stability in mammalian cells.

The term disposable soma came from the analogy with disposable products—why spend money making something durable if it will only be used for a limited amount of time?

There are arguments against the disposable soma theory. The theory clearly predicts that a shortage of food should make the compromise more severe all around; but in many experiments, ongoing since 1930, it has been demonstrated that animals live longer when fed substantially less than controls. This is the caloric restriction (CR) effect, and it cannot be easily reconciled with the Disposable Soma theory. Though by decreasing energy expenditure the damage generated (by free radicals, for instance) is expected to be reduced and the total energy budget might indeed be reduced, but the investment in repair function might still be relatively the same. But dietary restriction has not been shown to increase lifetime reproductive success (fitness), because when food availability is lower, reproductive output is also lower. So CR does thus not completely dismiss disposable soma theory.

Experimentally, some animals lose fertility when their life spans are extended by CR and some suffer no appreciable loss. Males, for example, typically remain fertile when underfed, while females do not. And, even females present an enigma because their fertility decline is not tightly coupled to their longevity gain. For example, in female mice that are restricted to 60% of a free-feeding diet, reproduction is shut down altogether. But female life span continues to increase linearly right up to the threshold of starvation – around 30% of free-feeding levels.

A difficulty with the disposable soma theory is that the energy required for maintenance and repair would appear to be relatively minor when compared to the energy required for gestation (repair should take less energy than producing an entire new organism). Yet gestating animals seem able to perform the maintenance while post-reproductive animals do not. A similar difficulty is that female animals seem to have longer life spans than males, despite the apparently higher energy requirement for gestation and other reproductive activities.

With respect to such limitations Kriete proposed consideration of systems-level properties like robustness to characterize aging as a robustness tradeoff. According to this concept living systems evolve into a state of highly optimized tolerance promoting traits beneficial for survival and fitness at the cost of fragilities driving the aging phenotype. The view is compatible with aspects of the antagonistic pleiotropy and the disposable soma theory, but offers additional mechanisms rooted in complex systems theory.

Other problems with the classical ageing theories

A raised criticism for all three mainstream theories based on classical evolutionary process concepts is the potential existence of 'deliberate' metabolic mechanisms that work to promote death.

One is apoptosis, or programmed cell death. Apoptosis is responsible for killing infected cells, cancerous cells, and cells that are simply in the wrong place during development. There are clear benefits to apoptosis, so the existence of apoptosis isn't a problem for evolutionary theory. The problem is that apoptosis seems to ramp up late in life and kill healthy cells, causing weakness and degeneration. And, paradoxically, apoptosis has been observed as a kind of 'altruistic suicide' in colonies of yeast under stress. This seems to be a direct hint that senescence arose because it conferred a direct evolutionary advantage, rather than some kind of side effect of genes that have other evolutionary advantages (pleiotropy).

A second 'deliberate' mechanism is called replicative senescence or cellular senescence. Metaphorically, a cell may be said to 'count' (with its telomeres) the number of times that it has divided, and after a set number of replications, it languishes and dies. It has been proposed that this mechanism evolved to suppress cancer. Many invertebrates experience replicative senescence, though they never die of cancer. Even one-celled organisms count replications, and will die if they don't replenish their telomeres with conjugation (sex).

More strictly, cells cannot 'count' the number of times they have divided. Telomeres are not a counting mechanism, though they may be used to indicate the number of times a particular chromosome has been replicated. Cellular processes for genetic material replication occur in both directions along DNA, 5' to 3' and on the other strand, 3' to 5'. As the 3' or 5' end is impossible for DNA polymerase to grab at the 1 base pair mark, a handful of basepairs (10-15) are cut off each replication. Over time, this cutting short of the DNA results in no telomeres, and the cell is unable to replicate that chromosome without cutting into genes.

The dilemma is that classical evolutionary theory says that what is maintained in a lineage is that which ensures the viability of an organism and its offspring. Ageing can only cut off an individual's capacity to reproduce. So, according to classical theory, ageing could only evolve as a side effect, or epiphenomenon of selection. The disposable soma theory and antagonistic pleiotropy theory are examples in which a compensating individual benefit, compatible with classical evolution theory (See neo-Darwinism and modern evolutionary synthesis) is proposed. Nevertheless, there is accumulated evidence that ageing looks like an adaptation in its own right, selected for its own sake.

Semelparous organisms and others that die suddenly following reproduction (e.g. salmon, octopus, marsupial mouse (brown antechinus), etc.) also represent instances of organisms who incorporate a lifespan-limiting feature. Sudden death is more obviously an instance of programmed death or a purposeful adaptation than gradual ageing. Biological elements clearly associated with evolved mechanisms such as hormone signalling have been identified in the death mechanisms of organisms such as the octopus.

Impact of new evolution concepts on ageing theories

At the time most of the non-programmed ageing theories were developed, there was very little scientific disagreement with classical theories (i.e. Neo-Darwinism or modern evolutionary synthesis) regarding the process of evolution. However, in addition to suicidal behaviour of semelparous species (not handled by the classical ageing theories) other apparently individually adverse organism characteristics such as altruism and sexual reproduction were observed. In response to these other conflicts, adjustments to classical theory were proposed:

• Various group selection theories (beginning in 1962) propose that benefit to a group could offset the individually adverse nature of a characteristic such as altruism. The same principle could be applied to characteristics that limited life span and theories proposing group benefits for limited lifespans appeared.
• Evolvability theories (beginning in 1995) suggest that a characteristic that increased an organism's ability to evolve could also offset an individual disadvantage and thus be evolved and retained. Multiple evolvability benefits of a limited lifespan were subsequently proposed in addition to those originally proposed by Weismann.

Ageing theories based on group selection

Group selection is often criticized to be too slow to happen in real biology. However, Jiang-Nan Yang recently showed with an individual-based model that the evolution of altruistic aging occurs under fairly general conditions by kin/group selection. Group selection can be based on population viscosity (limited offspring dispersal, first proposed by Hamilton (1964) for kin selection) that is widely present in natural populations. This population structure builds a continuum between individual selection, kin selection, kin group selection and group selection without a clear boundary for each level. Although early theoretical models by D.S. Wilson et al. (1992) and Taylor (1992) showed that pure population viscosity cannot lead to cooperation/altruism because of the exact cancelling out of the benefit of kin cooperation and the cost of kin competition, this exact cancelling out also suggests that any additional benefit of local cooperation would be sufficient for the evolution of cooperation. Mitteldorf and D.S. Wilson (2000) later showed that if the population is allowed to fluctuate, then local populations can temporarily store the benefit of local cooperation and promote the evolution of altruism. By assuming individual differences in adaptations, Yang (2013) further showed that the benefit of local altruism can be stored in the form of offspring quality and thus promote the evolution of altruistic aging even if the population does not fluctuate, this is because local competition among the young will result in an increased average local inherited fitness of survived progenies after the elimination of the less adapted by natural selection, since the young do not have strong age-associated abilities and have to depend more on inherited abilities to compete. In Yang (2013)'s model, altruistic aging is stabilized by higher-level selection instead of just kin selection.

Mitteldorf proposed a group benefit of a limited lifespan involving regulation of population dynamics. Populations in nature are subject to boom and bust cycles. Often overpopulation can be punished by famine or by epidemic. Either one could wipe out an entire population. Senescence is a means by which a species can 'take control' of its own death rate, and level out the boom-bust cycles. This story may be more plausible than the Weismann hypothesis as a mechanistic explanation, because it addresses the question of how group selection can be rapid enough to compete with individual selection.

Libertini also suggests benefits for adaptive ageing.

Inversely, within a Negative Senescence Theory R.D. Lee (similarly J.W. Vaupel) considered positive group effects performing a selection force directed to survival beyond the age of fertility. Often also postreproductive individuals make intergenerational transfers: bottlenose dolphins and pilot whales guard their grandchildren; there is cooperative breeding in some mammals, many insects and about 200 species of birds; sex differences in the survival of anthropoid primates tend to correlate with the care to offspring; or an Efe infant is often attended by more than 10 people. Lee developed a formal theory integrating selection due to transfers (at all ages) with selection due to fertility.

Ageing theories based on evolvability

Goldsmith proposed that in addition to increasing the generation rate, and thereby evolution rate, a limited lifespan improves the evolution process by limiting the ability of older individuals to dominate the gene pool. Further, the evolution of characteristics such as intelligence and immunity may specially require a limited lifespan because otherwise acquired characteristics such as experience or exposure to pathogens would tend to override the selection of the beneficial inheritable characteristic. An older and more experienced, but less intelligent animal would have a fitness advantage over a younger, more intelligent animal except for the effects of ageing.

Skulachev has suggested that programmed ageing assists the evolution process by providing a gradually increasing challenge or obstacle to survival and reproduction, and therefore enhancing the selection of beneficial characteristics. In this sense, ageing would act in a manner similar to that of mating rituals that take the form of contests or trials that must be overcome in order to mate (another individually adverse observation). This suggests an advantage of gradual ageing over sudden death as a means of lifespan regulation.

Weissmann's 1889 ageing theory was essentially an evolvability theory. Ageing or otherwise purposely limited lifespan helps evolution by freeing resources for younger, and therefore, presumably better-adapted individuals.

Yang (2013)'s model is also based on mechanisms of evolvability. Aging accelerates the accumulation of novel adaptive genes in local populations. However, Yang changed the terminology of "evolvability" into "genetic creativity" throughout his paper to facilitate the understanding of how aging can have a shorter-term benefit than the word "evolvability" would imply.

Ageing mechanism concepts

If organisms purposely limit their lifespans via ageing or semelparous behaviour, the associated evolved mechanisms could be very complex, just as mechanisms that provide for mentation, vision, digestion, or other biological function are typically very complex. Such a mechanism could involve hormones, signalling, sensing of external conditions, and other complex functions typical of evolved mechanisms. Such complex mechanisms could explain all of the observations of ageing and semelparous behaviours as described below.

It is typical for a given biological function to be controlled by a single mechanism that is capable of sensing the germane conditions and then executing the necessary function. The mechanism signals all the systems and tissues that need to respond to that function by means of organism-wide signals (hormones). If ageing is indeed a biological function, we would expect all or most manifestations of ageing to be similarly controlled by a common mechanism. Various observations (listed below) indeed suggest the existence of a common control mechanism.

It is also typical for biological functions to be modulated by or synchronized to external events or conditions. The circadian rhythm and synchronization of mating behaviour to planetary cues are examples. In the case of ageing seen as a biological function, the caloric restriction effect may well be an example of the ageing function being modulated in order to optimize organism lifespan in response to external conditions. Temporary extension of lifespan under famine conditions would aid in group survival because extending lifespan, combined with less-frequent reproduction, would reduce the resources required to maintain a given population.

Theories to the effect that ageing results by default (mutation accumulation) or is an adverse side effect of some other function are logically much more limited and suffer when compared to empirical evidence of complex mechanisms. The choice of ageing theory therefore is logically essentially determined by one's position regarding evolutionary processes, and some theorists reject programmed ageing based entirely on evolutionary process considerations.

Maintenance theories of ageing

It is generally accepted that deteriorative processes (wear, other molecular damage) exist and that living organisms have mechanisms to counter deterioration. Wounds heal; dead cells are replaced; claws regrow.

A non-programmed theory of mammal ageing that fits with classical evolution theory and Medawar's concept is that different mammal species possess different capabilities for maintenance and repair. Longer-lived species possess many mechanisms for offsetting damage due to causes such as oxidation, telomere shortening, and other deteriorative processes that are each more effective than those of shorter-lived species. Shorter-lived species, having earlier ages of sexual maturity, had less need for longevity and thus did not evolve or retain the more-effective repair mechanisms. Damage therefore accumulates more rapidly, resulting in earlier manifestations and shorter lifespan. Since there are a wide variety of ageing manifestations that appear to have very different causes, it is likely that there are many different maintenance and repair functions.

A corresponding programmed maintenance theory based on evolvability suggests that the repair mechanisms are in turn controlled by a common control mechanism capable of sensing conditions, such as caloric restriction, and also capable of producing the specific lifespan needed by the particular species. In this view, the differences between short- and long-lived species are in the control mechanisms, as opposed to each individual maintenance mechanism.

Summary of empirical evidence favouring programmed ageing

• Examples of complex programmed death mechanisms exist in semelparous species (e.g. octopus), including hormone signalling, nervous system involvement, etc. If a limited lifespan is generally useful as predicted by the programmed ageing theories, it would be unusual for an octopus to possess a more complex mechanism for accomplishing that function than a mammal.
• Discovery of "ageing genes" with no other apparent function.
• Caloric restriction effect: reduction of available resources increases lifespan. This behavior has a plausible group benefit in enhancing the survival of a group under famine conditions and also suggests common control.
• Progeria and Werner syndrome are both single-gene genetic diseases that cause acceleration of many or most symptoms of ageing. The fact that a single gene malfunction can cause similar effects on many different manifestations of ageing suggests a common mechanism.
• Although mammal lifespans vary over an approximately 100:1 range, manifestations of ageing (cancer, arthritis, weakness, sensory deficit, etc.) are similar in different species. This suggests that the deterioration mechanisms and corresponding maintenance mechanisms operate over a short period (less than the lifespan of a short-lived mammal). All the mammals therefore need all the maintenance mechanisms. This suggests that the difference between mammals is in a control mechanism or repair efficiency.
• Lifespan varies greatly among otherwise very similar species (e.g. different varieties of salmon 3:1, different fish 600:1) suggesting that relatively few genes control lifespan and that relatively minor changes to genotype could cause major differences in lifespan—suggests common control mechanism.

Problems with programmed aging theories

Contrary to the theory of programmed death by aging, individuals from a single species usually live much longer in a protected (laboratory, domestic, civilized) environment than in their wild (natural) environment, reaching ages that would be otherwise practically impossible. Also, in majority of species, there doesn't exist any critical age after which death rates change dramatically, as intended by the programmed death by aging, but the age-dependence of death rates is very smooth and monotonic. However, as mentioned above, V.P. Skulachev explained that a process of gradual aging has the advantage of facilitating selection for useful traits by allowing old individuals with a useful trait to live longer. It is also easy to imagine that animals with gradual aging will live longer in a protected environment.

The death rates at extreme old ages start to slow down, which is the opposite of what would be expected if death by aging was programmed. From an individual-selection point of view, having genes that would not result in a programmed death by aging would displace genes that cause programmed death by aging, as individuals would produce more offspring in their longer lifespan and they could increase the survival of their offspring by providing longer parental support.

Biogerontology considerations

Theories of aging affect efforts to understand and find treatments for age-related conditions (see biogerontology):

• Those who believe in the idea that aging is an unavoidable side effect of some necessary function (antagonistic pleiotropy or disposable soma theories) logically tend to believe that attempts to delay aging would result in unacceptable side effects to the necessary functions. Altering aging is therefore "impossible", and study of aging mechanisms is of only academic interest.
• Those believing in default theories of multiple maintenance mechanisms tend to believe that ways might be found to enhance the operation of some of those mechanisms. Perhaps they can be assisted by anti-oxidants or other agents.
• Those who believe in programmed aging suppose that ways might be found to interfere with the operation of the part of the aging mechanism that appears to be common to multiple symptoms, essentially "slowing down the clock" and delaying multiple manifestations. Such effect might be obtained by fooling a sense function. One such effort is an attempt to find a "mimetic" that would "mime" the anti-aging effect of calorie restriction without having to actually radically restrict diet.

See also

• List of life extension-related topics
• Negligible senescence (Describes observation of animals without measurable symptoms of ageing)
• Senescence
• Gerontology
• Phenoptosis

Further reading

• Fabian, D. & Flatt, T. (2011) The Evolution of Aging. Nature Education Knowledge 3(10):9
• Gavrilova, N.S., Gavrilov, L.A. Human longevity and reproduction: An evolutionary perspective. In: Voland, E., Chasiotis, A. & Schiefenhoevel, W. (eds.): Grandmotherhood - The Evolutionary Significance of the Second Half of Female Life. Rutgers University Press. New Brunswick, NJ, USA, 2005, 59-80.
• Gavrilova, N.S., Gavrilov, L.A. Evolution of Aging. In: David J. Ekerdt (ed.) Encyclopedia of Aging, New York, Macmillan Reference USA, 2002, vol.2, 458-467.

External links

• Evolutionary Theories of Aging and Longevity
• The Evolutionary Theory of Aging by João Pedro de Magalhães.
• Programmed-Aging.Org Site provides comprehensive information on programmed ageing, the programmed/non-programmed controversy, and underlying evolution controversies.
• How Evolutionary Thinking Affects People's Ideas About Aging Interventions
• AnAge Animal Ageing and Longevity Database Provides maximum observed ages and sexual maturity ages for many animals.
• The Case for Programmed Mammal Aging Describes empirical data, evolutionary rationale, and historical perspective supporting programmed ageing in mammals.
• Life Table for USA 2005 Probability of death as a function of age