The post-reproductive lifespan: evolutionary perspectives.

Aging (or senescence) is a well studied topic for several decades. Hamilton (1966) [1], Williams (1957) [2] and Medawar (1946) [3], and Kirkwood (1979) [4] have addressed this issue and have builded the first theories about this phenomenon. Aging is defined as a «decline in performance and fitness with advancing age»  [5]. Three theories have been proposed to explain why we die: the deleterious mutation accumulation theory [3,6], the antagonistic pleitropy theory [2] and the disposable soma theory [4,7,8]. The first two theories explain that deleterious mutations are accumulated because they have deleterious effects only in the late life, so that there is few selection on it. In the second theory, those mutations are selected because they have positive effects in early life. The third theory explains that there is a trade-off in the energy investments allocated for reproduction and for repairing the deleterious effects of the metabolism (oxidative catabolism), so the body age because we invest in reproduction.

It has been observed that the loss of reproductive ability can appear before death in metazoans. In some species, females stop to reproduce a short or a long time before they die. This phenomenon is called «menopause». Female menopause has received considerable attention over the past thirty years. Understanding the evolution of menopause is particular because of the common conception of the theory of evolution. According to the theory of evolution, surviving after the cessation of reproduction should not be selected, because it is completely useless regarding the fitness. Indeed, living species are « survival machines... for replicators (genes) » [9], which have to reproduce as much as possible to spread their genome, so we should reproduce and die when we couldn't reproduce anymore. Why non-reproductive individuals survive? Another formulation can be proposed: why selection didn't favor female who can reproduce longer? [1,2] Nowadays, studies have clearly shown that post-reproductive lifespan (PRLS) is common in animals, especially in mammals. But menopause seems to be only present in humans and some whales and most of studies published focuses on those species [10‑15]. Indeed, human females have much more to gain by stopping to reproduce and provide parental or grand parental care, than going on reproducing [10]. The prospects are huge, understanding the proximate causes and the evolutionary process of this dramatic physiological change could help us to manage with menopause, by finding biological models, elucidating the causes, or finding how comes that females live after their reproductive cessation.

It is important to distinguish the two terms «menopause» and «post-reproductive lifespan». Indeed, menopause is a biological term more related to humans, which is «the irreversible loss of the physiological capacity to produce offspring due to intrinsic biological factors»  [16]. However we can extend this term to other mammals as cetaceans, which clearly show the same kind of reproductive cessation [17]. The biological definition of menopause includes the loss of fertility around a mean of 50 years in humans. The proximate cause of this loss is the cessation of ovulation accompanied by hormonal changes in neuroendocrine system [18] and the primary origin is the progressive depletion of ovarian follicles [19]. The term PRLS is a term used to describe «the time between age at reproductive cessation and age at death» (for individuals living beyond reproductive cessation) [16]. No study on PRLS focuses on the different strategies of « stopping to reproduce because of the cessation of ovulation » and « stopping to reproduce because it is just too costly ». To sum up, menopause is a biological term, which relies on a physiological state, while PRLS is just a time scale, between the time of reproductive cessation and the time at death [16]. So, menopause is included in PRLS. The two terms will be used during the discussion.

A lot of theories try to explain the emergence of menopause or PRLS (Table 1). There are non-adaptative theories and theories based on natural selection. Non-adaptative theories assume that PRLS could be just an artifact of life histories, such as the current raise of civilization for humans or the domestication for animals [2,20,21]. As a consequence, menopause should be just the deterioration of the reproductive system. An other explanation is that PRLS appeared as an insurance against indeterminacy [22]. On the other hand, there is two main theories, which assume that menopause is an evolutive process. The first is the Grandmother hypothesis, which proposes that non-reproductive mothers help their daughters to enhance their reproductive fitness. The second is the Mother hypothesis, which explains that old females stop reproducing because it is to risky for them, and so for their children.

Here we discuss about the old and the recent theories of the evolution of PRLS and its evolutionary implications. First, we will discuss about who experiences menopause, which is still a debated topic. Second, we will successively analyze the evidence for and against the two main evolutive explanations of PRLS (the grandmother and the mother hypothesis), which have been the two main hypothesis supported for a long time. Then we will discuss about alternatives to the first two hypothesis and we will finally study the case of the emergence of menopause in humans.

Nowadays, menopause is considered for a lot of researchers as a selective process. Mayer proposed in 1982 [36] one of the two principal theories, explaining the existence of PRLS and menopause. The grandmother hypothesis predicts that non-reproductive (or old) mothers can help their daughters to reproduce and breed children in order to enhance their own fitness. It means that surviving old mothers increase their inclusive fitness [36], for instance, because the presence of living grandmother has a positive effect on the number of grandchildren and on the fitness of those grandchildren [15]. Grandmothers enhance the survival of their grandchildren through protection and/or provisioning [23,37]. The more the grandmother is genetically close from his grandchild, the more she favors his survival, as measured Fox et al. (2010) [38]. So family with grandmothers are selected and PRLS is increased [11].

The case of humans is well studied and a good example to understand the evolution of PRLS and is particularly interesting because humans have a longer PRLS than other species, even other primates [26]. In humans it has been shown that, from an economic point of view (regarding the fitness benefits and the cost of the strategy used), this strategy of kin selection outweigh the cost of reproduction [10,15]. Cant and Johnstone (2008) [10] drew a model to explain the grandmother hypothesis and bound it to the conflict between generations in human. This model includes the cost of reproductive competition and the benefits of grandmothering. It predicts that menopause appears in old females because it becomes too costly to compete their own daughter to reproduce, and it is more beneficial for their inclusive fitness to help them to breed their children. As a consequence, the overlap of reproduction between generations in human is close to zero (Figure 2). This model comes as a complement to the grandmother hypothesis. Lahdenpera et al. (2004) [15] showed with a large study in Canadian and Finns population that human grandmothers actually increase their inclusive fitness by breeding their grandchildren. They studied the fitness benefits of post-reproductive women during the eighteenth and nineteenth centuries by measuring the effect of the presence of a grandmother on: number of grandchildren, reproductive performances of the offspring (lifetime fecundity, lifetime reproductive success) and grandchildren survival probability. Indeed, the presence of a grandmother has a positive effect on the number of grandchildren, as grandmother enhance the lifetime and the reproductive success of their offspring (earlier breeding, higher frequency, more success). All of these results show that selection can increase PRLS. Interestingly, the mortality rate of the grandmother increase dramatically when their daughter stop to reproduce. The role of demography and kinship has also been demonstrated in whales with the support of the grandmother hypothesis [39]. The social structure of the group increases the local relatedness between individuals (and particularly between females), which favors the evolution of menopause, because relatedness favors the probability of helping behaviors.

Figure 2. Overlap from three generation in four primates.

Overlap from three generation in four primates.

Rectangle account for the total lifespan and black parts account for the reproductive lifespan. (From Cant and Johnstone, 2008 [10]) .

In addition, Peccei (1995) [24] proposed the «altriciality lifespan» hypothesis, which is an alternative to the grandmother hypothesis. She identified encephalization and infant altriciality as the critical selection pressure, thanks to a demographic model measuring the extra-reproductive value on females. Infant born in more and more helpless states because of the increasing encephalization in Pliocene, so that helping behaviors are selected.

Most of the work in this field supports the grandmother hypothesis [23,37,40,10], although there is some criticism [41,13,20]. Adaptative hypothesis are balanced by non-adaptative hypothesis. Different authors found cues to contest the grandmother hypothesis. Indeed, some data support that the cessation of reproduction is not a selective process [41], but a by product of life histories, which increased dramatically in humans [20]. Packer (1998) [41] proved in two social species, baboon and lion, that the presence of the mother does not enhance the fitness of the offspring (survival, but also the number of children). Other authors argue that the inclusive fitness benefits are to small to influence natural selection [13], but they admit that phenomenons as allocare could influence life history.

Other cases of PRLS-like behaviors have been reported and can be considered as particular examples for the grandmother hypothesis: post-reproductive aphids become transformed into glue bombs (transformation of their abdomen, previously dedicated to embryo production, into a manufacture of defensives wax secretion), which can explode to protect their colony [30]. The point is that aphids live in clonal colony. So there are old helpers in the colony, which enhance the survival of their younger relatives.

Is menopause a consequence of the increasing lifespan of humans, as a byproduct of the industrialization ? We may suppose that menopause appeared in humans because they lived longer thanks to modernization and medicine, what fit with the «artifact hypothesis», that is to say that menopause appeared because women lived longer and the reproductive system get exhausted [20,21,2]. But as some studies shown, menopause emerged when humans were already organized in hunter-gatherer populations [45]. They argue that the features of our species (large brain, long lives, parental investment in breeding, grandparental support, long child dependency of parents) appeared before the preagricultural history [46]. This a reason why today, scientists test their hypothesis on the evolution of menopause on current hunter-gatherer populations [47] or populations living in developing countries [48]. There is some evidence that menopause in humans did not appeared because of civilization and modern circumstances, or the effect of agricultural and industrial economies [48]: an emergence because of civilization could be too near from today to be the result of natural selection, as he says. So there is a lot of speculations on the date of the apparition of human menopause. Hawkes (1998) recorded three dates for life history changes mediated by increasing PRLS: the initial appearance of Homo erectus 1,8 millions years ago, in early Homo sapiens 600,000 years ago, or 50,000 years ago, which is when the Homo sapiens had a huge ecological and competitive success compared to the other Homo .

Here is still a debate about how appeared menopause, and more generally PRLS. If only humans and whales clearly show menopause, PRLS is not unusual in mammals. Many theories converge to conclude that PRLS evolved because it is adaptative and because it increases the fitness of animals who dispose of this feature. So non-reproductive mother increase their inclusive fitness, through the protection and the help of their offspring. On the other hand, there is also evidence for non-adaptative theories: the increase in life span gave the possibility to use all the ovocyte stock.

Other factors influencing PRLS have been shed in light in the past ten years (mortality risk for example), which can explain the variations of PRLS in populations, in animals as well as in humans. Indeed the duration of PRLS can change from a population to another in a species. Humans, for example, is a species, which can be divided in many different populations, in which life histories vary, because of different life styles, environment, or wealth.

The fact that PRLS in particularly commons in mammals should be more investigated. Indeed, mammals have a lots of different life-histories: longevity, size, body mass, weaning age, social structure vary between species. We can assume that PRLS could be bound to the social structure in mammals. Indeed, if we agree with selective based theories as the grandmother and the mother hypothesis, the emergence of this feature based on parental care, helping behavior require a minimal social context. Animals had to live together to provide care to weaned, adult or sub-adult offspring. So it could be very interesting to investigate if the sociality level could influence PRLS. One prediction that could be made is that PRLS should be more selected in highly social species. Indeed, the more social is a species, the more high is the probability to find helper in the group, which can enhance the new-born and young survival. Moreover, the helper may be or should be, if we refer to the theory of the «selfish gene» [9], related to the offspring helped. So, non-reproductive grandmothers should be more selected more in social species.

Finally, there is some evidence for each theory and also against (summarized in Figure 4). May be should we consider that more than one theory could account for the emergence of menopause. We can imagine a mixed scenario between non-adaptative and adaptative hypothesis: an increase in female lifespan should have lead to a longer presence of grandmother, and increased the fitness and the lifespan of those family thanks to helping behaviors of the grandmother, leading finally to the emergence and the increase of PRLS. Moreover, a favorable environment or less predators could have allowed humans and whales to increase their lifespan and PRLS appeared. Indeed, we can have such phenomenon in captive animals as bird, in which we observed PRLS [31]. According those assumptions, PRLS could have appeared several times independently.

Figure 4.  Summary of the different theories on PRLS.

Summary of the different theories on PRLS.

Five reported hypothesis are reported in the middle, while evidence for and against are on the right and on the left. Arrows go from evidence to an hypothesis. Vertical bars account for evidence against an hypothesis. Dotted lines group theories together: adaptative hypothesis in green, hypothesis related to the grandmother hypothesis in purple. Colors helps to indicate which hypothesis or which group of hypothesis arrows and vertical bars point.

I want to thank Aurélie Cohas and Jean-François Lemaitre for helping me to write this report and also Ben, Marion and Vérane, who helped me to deal with LateX and made me suggestions.