The role of telomeres in the mechanisms and evolution of life-history trade-offs and ageing

(a). Life-history evolution and the search for underlying mechanisms

Discussions of the role that telomeres may play in the mechanisms and evolution of life-history trade-offs and ageing can be usefully grounded in life-history theory [1,2], whose logic is paralleled by that of the disposable soma theory of ageing ([4–6]; itself compatible with the antagonistic pleiotropy theory of ageing [5–7]). Life-history theory recognizes that the range of attainable evolutionary outcomes will be bounded by so-called ‘absolute’ evolutionary constraints (inescapable constraints born of the laws of physics), such as the need to allocate limited resources across multiple competing traits, including growth, reproduction and ‘somatic maintenance’ (the suite of molecular proof-reading and damage mitigation mechanisms that slow the progressive accumulation of errors and damage in the body's tissues [4–6]) [1,2,9–11,20]. Selection for optimal resource utilization strategies, for example, might then be expected to give rise to negative phenotypic and genetic correlations between some pairs of life-history traits (i.e. trade-offs; trade-offs need not be apparent between all pairs of traits as the optimal strategy may be to co-express a given pair while trading them off against others). This approach can intuitively account for the existence of trade-offs between traits that are simultaneously expressed, such as clutch size and egg size, as resource invested in one trait cannot be invested in others. And it can also account for ‘current–future’ life-history trade-offs, such as those between current reproduction or growth and future reproduction or survival. For example, current–future trade-offs are often envisaged to arise because current investment in reproduction can entail shortfalls in investment in somatic maintenance that will accelerate age-related declines in somatic integrity (the extent to which the body's tissues are free from errors and damage) that, unless recovered, are carried forward to future time steps, with detrimental effects on future reproduction and survival. In this way, current–future trade-offs have the potential to deflect an organism's senescence trajectory (characterized by the onset and rate of late-life declines in components of fitness), due to impacts of current actions on the rate of age-related decline in somatic integrity [8,17,18].

This resource allocation model has proved a useful heuristic for studying the evolution of life histories, but the extent to which the relevant evolutionary constraints in reality conform to the model's assumptions is currently far from clear [1–3,9–12,20–23]. While ‘absolute’ constraints of various kinds clearly will shape life-history evolution, the nature of the absolute constraints at play and their relative importance is less clear [1,2]. Moreover, the importance of what I shall collectively term ‘mechanistic constraints' (i.e. evolutionary constraints born of aspects of an organism's existing genetic, developmental and physiological mechanisms, themselves a product in part of phylogenetic history) in life-history evolution remains a matter for debate [9,12,20–23]. For these reasons, evolutionary biologists have become increasingly concerned with elucidating the mechanisms that underpin life-history trade-offs, with a view to shedding light on the constraints at play [9–16,21,22]. One major focus is the role that oxidative stress may play, given the potential for energetically demanding current activities to cause oxidative damage to diverse biomolecules, thereby compromising somatic integrity and future performance [13–15]. However, uncertainty regarding the central importance of oxidative stress [13–15], coupled with the expectation that multiple mechanisms of somatic deterioration are likely to act in concert [24,25], motivates continued attention to the roles of other pathways. Intuitively, in seeking a proximate mechanism that underpins current–future trade-offs one might seek a broadly conserved biological structure that can be damaged by ‘current’ actions, and whose deficits pass forward to ‘future’ time steps with the potential for causal detrimental effects on future performance. A wealth of evidence from biomedical and ecological research now highlights that telomeres may constitute just such a structure [16].

(b). What role might telomeres play?

Telomeres are nucleoprotein complexes that cap the ends of the linear chromosomes of eukaryotes [26]. They comprise a repetitive non-coding DNA sequence (TTAGGG repeats in vertebrates) bound up in a multi-protein complex [26]. Their structural conservation across eukaryotes suggests an ancient origin and reflects their importance in overcoming two challenges posed by the evolution of linear chromosomes [26,27]. First, as conventional DNA polymerases do not replicate the ends of linear chromosomes, the presence of terminal non-coding telomeric DNA averts the loss of coding nucleotides during cell replication [28,29]. Second, the telomeric-binding proteins wrap up the chromosomal end in such a way as to shroud it from detection by the cell's DNA repair machinery, to avoid triggering a DNA damage response and/or pathological chromosomal fusions [26,30]. Telomeres are, however, dynamic structures that can change in length, and it is these telomere length dynamics, born of the balance of telomere attrition (the shortening of telomeres over time via the loss of terminal repeats) and elongation (e.g. the enzyme telomerase can extend telomeres via the addition of terminal telomeric repeats [26,31]), that are of particular interest when considering their role in the mechanisms and evolution of life histories.

Interest in a potentially causal role for telomere length dynamics in current–future trade-offs and ageing stems from a series of findings that highlight their potential to leave current actions having negative downstream effects on future performance [16]. First, while telomerase expression in unicellular eukaryotes is thought to maintain telomere lengths within a species-specific range, telomerase insufficiency or complete repression is commonplace in the somatic tissues of metazoans, frequently leading to progressive declines in telomere length with advancing organismal age [26,27,31–34]. Second, in vitro studies suggest that such shortening arises principally from cellular replication and oxidative stress [28,29], providing candidate pathways by which organismal growth and reproduction could accelerate telomere shortening [16]. Third, the accrual of short telomeres within cells is implicated in the triggering of apoptosis (programmed cell death) and a state of irreversible cell-cycle arrest termed cellular senescence; cell fates that are strongly implicated in age-related declines in tissue and organismal performance [25,35]. Accordingly, shorter mean telomere lengths in vivo frequently predict poorer organismal health and survival outcomes both in the laboratory and the wild [36]. As such, telomere attrition is now widely recognized as one of the hallmarks of ageing [25], and telomeric assessments are being widely adopted in evolutionary biology as a biomarker of somatic integrity. Despite this, limited attention has been paid to addressing the fundamental and increasingly pressing question posed by these relationships: what role, if any, do telomeres play in the mechanisms and evolution of life-history trade-offs and ageing?

Here, I seek to address this question by evaluating the mechanistic and evolutionary plausibility of telomeres playing a proximate causal role in current–future life-history trade-offs (and their impacts on ageing trajectories), before considering the wider question of whether telomere biology is likely to shed new light on the evolution of life-history trade-offs and ageing. First, I consider the case for telomeres being one of the proximate causal mechanisms that gives rise to current–future trade-offs (§2). I conclude that while it is mechanistically plausible that telomeres are one proximate cause of current–future trade-offs in a subset of species, whether telomere attrition per se plays a significant causal role in any species, relative to alternative mechanisms, is currently far from clear.

Second, I consider the evolutionary plausibility of telomeres being one proximate cause of current–future trade-offs (§3), given that selection would be expected to decouple such mechanistic links between current and future performance in the absence of evolutionary constraint (see above). In evaluating adaptive explanations for a strategy of progressive telomere attrition in somatic cells, I consider two broad forms of explanation that are not mutually exclusive: the costly maintenance hypothesis (telomere attrition is adaptive because telomere maintenance entails costs, such as the utilization of resources that could be invested in other traits) and the functional attrition hypothesis (telomere attrition is adaptive because it serves a particular function; such as its widely invoked function in countering cancer). I also highlight a potential and hitherto unexplored signalling function of telomere attrition: as a biological clock for regulating the pace of an individual's life history according to somatic integrity (the life-history regulation hypothesis).

Finally, I consider the extent to which telomere biology is likely to shed new light on the evolution of life-history trade-offs and ageing (§4). Importantly, evidence of a proximate causal role for telomeres in current–future trade-offs and ageing would not implicate a role for telomeres in shaping their evolution. Whether telomere biology informs our understanding of the evolution of life histories and ageing trajectories hinges upon (i) the extent to which telomere biology sheds new light on the constraints that explain why selection has not decoupled current–future trade-offs (see above) and (ii) the extent to which the evolution of telomeric mechanisms has influenced the ‘form’ of current–future trade-offs (i.e. the probability, magnitude and timing of the impacts of current actions on future performance).

(a). Do current reproduction, growth and adversity accelerate telomere attrition?

Telomere attrition can arise through a range of processes [28,29], but two principal mechanisms provide candidate pathways by which current reproduction and growth could hasten telomere attrition [16,38]. First, in the absence of telomerase, telomeres shorten during each round of cellular replication, due in part to the inability of conventional DNA polymerases to completely replicate the ends of linear chromosomes [28,29]. Second, larger tracts of telomeric repeats are occasionally lost in seemingly ‘sporadic’ events [39], which appear to be attributable in part to oxidative damage to telomeric DNA ([28,29], but see [40]). Both cellular replication and oxidative damage provide viable mechanisms by which organismal growth and reproduction could impact telomere attrition. Growth and reproduction typically both entail significant cellular replication, and investments in either are generally expected to increase exposure to oxidative damage via associated increases in the rates of reactive oxygen species (ROS) production and/or compromised investment in antioxidant defences [13,14,38]. Accordingly, telomere attrition rates in vivo are often faster during the growth phase than later in life (e.g. [34,38]), and evidence suggests that growth acceleration may indeed entail oxidative stress and accelerate telomere shortening ([41,42], see [38] for a review). Whether reproduction entails an oxidative stress cost remains a matter of debate, as a recent meta-analysis has revealed that, while breeding females exerting greater reproductive effort show higher levels of oxidative damage, breeders typically show lower levels of oxidative damage than non-breeders [14]. These patterns have led to the suggestion that selection has favoured physiological strategies that mitigate oxidative stress during reproductive episodes (the oxidative shielding hypothesis [14]). Nevertheless, despite limited empirical attention, a number of studies have now provided evidence consistent with the expectation that reproductive effort accelerates telomere attrition (e.g. [43–45]). The proximate mechanisms that generate current–future trade-offs may also underpin the widely documented deleterious effects of current-life adversity on future performance [46,47]. It is notable then that exposure to diverse stressors as well as elevated glucocorticoid (stress hormone) concentrations has been demonstrated to accelerate telomere attrition, in ovo, in utero, during post-natal development and in adulthood (see [46,47] for recent reviews).

(b). Does telomere attrition causally impact future organismal performance and ageing?

Telomere attrition could have causal negative effects on later-life performance because the accumulation of critically short telomeres in cells is understood to trigger apoptosis or cellular senescence; cell fates strongly implicated in organismal ageing [25,35,48–50]. The progressive loss of telomeric repeats is thought to gradually undermine the shielding of the chromosomal end from the cell's DNA damage recognition machinery [26,30], ultimately triggering a DNA damage response that in turn induces apoptosis or cellular senescence [26,30,35,50,51]. Accordingly, experimental evidence in vitro suggests that telomere attrition limits the proliferative potential of telomerase-negative cells, and the induction of telomerase expression restores proliferative potential and staves-off cellular senescence [51]. Caution is needed when interpreting the findings of telomerase manipulations, however, given the apparent potential for telomerase to rescue proliferative potential in the absence of telomere lengthening [26,52]. Moreover, recent findings have highlighted that while telomeric mechanisms may indeed be the major trigger of cellular senescence in vivo [53,54], such triggering may actually occur independent of telomere length [54,55] (see §2c(ii) below). Either way, the age-related accumulation of such senescent cells [53,54] is thought to play a key role in age-related declines in organismal performance, due to the associated passive loss of tissue function and their unusual pro-inflammatory secretory profiles [25,35]. Indeed, remarkable recent experiments have revealed that the clearance of senescent cells by two different methods counteracts age-related declines in performance in house mice, Mus musculus [48,49]. Telomere-attrition-induced apoptosis (as opposed to cellular senescence) is expected to compound these effects, as cell replacement via somatic stem cell division is generally expected to hasten stem cell telomere attrition and consequent stem cell exhaustion [25,35].

While it is widely accepted that telomere attrition can have causal effects on cell fates, the extent to which it contributes to age-related declines in organismal performance is less clear. Key areas of uncertainty include the relative importance of telomere-mediated and telomere-independent mechanisms (such as oxidative damage to other structures [13,35,50]), and the role of telomere-mediated mechanisms that may act independent of telomere length [54,55] (see next section). Nevertheless, given the panoply of mechanisms that could act in concert to precipitate organismal ageing [24,25], it is notable that three lines of evidence are at least consistent with a causal role for telomeres. First, humans and wild vertebrates with shorter mean blood cell telomere lengths and/or higher telomere attrition rates frequently show poorer survival prospects [36]. Moreover, recent comparative studies have revealed that the telomere attrition rates of wild animals are also higher in species with shorter lifespans and faster life histories [32,56]. While fewer studies have investigated whether individuals with shorter telomeres suffer poorer reproductive prospects, such relationships do exist (e.g. [57]). Second, short telomeres are also prognostic of age-related disease and poorer survival in humans [58], and telomeric pathologies are strongly implicated in diverse premature ageing disorders [25]. Third, studies revealing exacerbated tissue degeneration and shortened lifespans in telomerase-deficient house mice, and the restorative effects of telomerase restoration, illustrate the potential for telomere-related pathologies to causally impact organismal health (e.g. [59]). With regard to the causal role of telomeres in the natural ageing process (i.e. in the absence of telomerase deficiency), recent evidence that telomerase overexpression in adult and aged house mice extends natural lifespans is notable ([60], but see [61]). Indeed, there is considerable wider interest in the potential for telomerase manipulations to offer insights into the causal effects of changes in telomere length (see [62] for a review), despite the challenges posed by the diverse actions of telomerase aside from telomere lengthening [26,52].

(c). Reasons for caution: the non-causal biomarker hypothesis, alternative mechanisms and questionable generality

Given cellular- and organismal-level evidence outlined above, it is certainly mechanistically plausible that telomere dynamics are one proximate cause of current–future trade-offs and ageing. However, multiple reasons exist for caution regarding causality. It is frequently highlighted, for example, that telomere length's utility as a predictor of health and fitness could instead reflect it acting as a non-causal biomarker of accumulated damage to other biological structures that themselves have causal deleterious effects on future performance (e.g. [61,63]; the non-causal biomarker hypothesis). There are also notable reasons for caution regarding the relative causal importance of telomere attrition and alternative mechanisms, as well as the taxonomic generality of any causal role for telomeres. I discuss these issues below.

(a). Telomere attrition as an adaptive strategy: the costly maintenance hypothesis

The costly maintenance hypothesis proposes that telomere attrition in the somatic cells of adults is adaptive because telomere length maintenance entails costs which collectively render tolerating attrition adaptive. Telomere maintenance must entail a resource cost of some kind, but it would seem implausible that this cost alone is so high as to account for selection favouring telomere attrition if attrition was a major causal agent in current–future trade-offs and ageing. Human telomeres, for example, have been estimated to shorten at a rate of just 20–30 bp per year, the restoration of which (including telomeric elongation and any associated changes in the shelterin complex, etc.) would presumably generate negligible resource demands relative to the maintenance and regulation of the 6.47 billion bp human genome [80]. The pervasive nature and quantity of so-called ‘junk’ DNA (see [81] for a recent review) and the success and persistence of polyploid genomes [82] also give cause to question whether the resource costs entailed in the maintenance of such a small genomic region (i.e. the telomeres) is likely to impact life-history evolution. That said, the potential importance of even small maintenance costs cannot currently be dismissed, as if telomere attrition is only a minor causal player in current–future trade-offs and ageing (see above) even small maintenance costs could conceivably be sufficient to favour the toleration of attrition.

The costs entailed in telomere maintenance need not arise solely from resource expense, however. It is conceivable that telomere maintenance entails other types of costs arising from interactions between biological mechanisms and the laws of physics. For example, aspects of telomere maintenance could conceivably entail risks to chromosomal integrity, and hence organismal fitness, if they require physical conformational rearrangements of the telomere (including its shelterin complex) or other chromosomal regions with which telomeres interact. While selection could have the potential to mitigate such costs via evolutionary changes to the mechanisms involved (e.g. the evolution of additional chromosomal stabilizing mechanisms), the costs associated with such solutions could conceivably be even higher than those that arise from telomere attrition. Alternatively, some form of absolute constraint might render the complete mitigation of such risks impossible. Closer attention to the plausibility of maintenance costs of this kind could be instructive, given their potential importance and the possibility that they cannot be readily reconciled with classical resource allocation models of life-history evolution.

As telomere maintenance must entail costs (born at least in part from resource expense), such costs alone could conceivably favour (i) telomere maintenance only in those cell lineages in which attrition would entail substantial fitness costs (e.g. germ cells and some immune cells [26,31]), while (ii) tolerating attrition in other somatic lineages, even if the latter had causal negative effects on later-life performance. The greater these causal negative effects are, however, the more acute the need becomes to invoke adaptive explanations for attrition born in part of a function for telomere attrition (i.e. a role for the functional attrition hypothesis).

(b). Telomere attrition as an adaptive strategy: the functional attrition hypothesis

Telomere attrition in the somatic cells of adults could be adaptive (i.e. of net fitness benefit over the life-course), despite having causal negative effects of later-life performance, because progressive telomere attrition yields fitness benefits by serving one or more functions (the functional attrition hypothesis). While diverse functions are conceivable (e.g. [83]), I first consider the leading adaptive explanation for telomere attrition in the biomedical literature: that it serves to mitigate the risk and fitness consequences of cancer [29] (the cancer surveillance hypothesis). I then highlight the possibility that telomere attrition has an additional and hitherto unexplored function: allowing organisms to adaptively regulate their physiology, behaviour and life history according to residual somatic integrity (which I term the life-history regulation hypothesis).