Volume 34, Issue 1 p. 38-54
AN INTEGRATIVE VIEW OF SENESCENCE IN NATURE
Free Access

The untapped potential of reptile biodiversity for understanding how and why animals age

Luke A. Hoekstra

Luke A. Hoekstra

Department of Ecology, Evolution, and Organismal Biology, Iowa State University, Ames, IA, USA

Search for more papers by this author
Tonia S. Schwartz

Tonia S. Schwartz

Department of Biological Sciences, Auburn University, Auburn, AL, USA

Search for more papers by this author
Amanda M. Sparkman

Amanda M. Sparkman

Department of Biology, Westmont College, Santa Barbara, CA, USA

Search for more papers by this author
David A. W. Miller

David A. W. Miller

Department of Ecosystem Science and Management, Pennsylvania State University, University Park, PA, USA

Search for more papers by this author
Anne M. Bronikowski

Corresponding Author

Anne M. Bronikowski

Department of Ecology, Evolution, and Organismal Biology, Iowa State University, Ames, IA, USA

Correspondence

Anne M. Bronikowski

Email: [email protected]

Search for more papers by this author
First published: 09 September 2019
Citations: 36

Funding information

USA NIH RO1-AG049416 to AMB. USA NSF IOS-1558071 to AMB, DAWM, and AMS.

Abstract

  1. The field of comparative ageing biology has greatly expanded in the past 20 years. Longitudinal studies of populations of reptiles with a range of maximum lifespans have accumulated and been analysed for evidence of mortality senescence and reproductive decline. While not as well represented in studies of amniote senescence, reptiles have been the subjects of many recent demographic and mechanistic studies of the biology of ageing.
  2. We review recent literature on reptile demographic senescence, mechanisms of senescence, and identify unanswered questions. Given the ecophysiological and demographic diversity of reptiles, what is the expected range of reptile senescence rates? Are known mechanisms of ageing in reptiles consistent with canonical hallmarks of ageing in model systems? What are the knowledge gaps in our understanding of reptile ageing?
  3. We find ample evidence of increasing mortality with advancing age in many reptiles. Testudines stand out as slower ageing than other orders, but data on crocodilians and tuatara are sparse. Sex-specific analyses are generally not available. Studies of female reproduction suggest that reptiles are less likely to have reproductive decline with advancing age than mammals.
  4. Reptiles share many physiological and molecular pathways of ageing with mammals, birds and laboratory model organisms. Adaptations related to stress physiology coupled with reptilian ectothermy suggest novel comparisons and contrasts that can be made with canonical ageing phenotypes in mammals. These include stem cell and regeneration biology, homeostatic mechanisms, IIS/TOR signalling, and DNA repair.
  5. To overcome challenges to the study of reptile ageing, we recommend extending and expanding long-term monitoring of reptile populations, developing reptile cell lines to aid cellular biology, conducting more comparative studies of reptile morphology and physiology sampled along relevant life-history axes and sequencing more reptile genomes for comparative genomics. Given the diversity of reptile life histories and adaptations, achieving these directives will likely greatly benefit all ageing biology.

A free Plain Language Summary can be found within the Supporting Information of this article.

1 INTRODUCTION

Senescence is the complex biological process of disintegrating physiological and biochemical function with increasing age, leading to age-related frailty and functional decline. At the cellular level, senescence involves gene regulatory networks, cell signalling pathways, and cellular functional changes (Kennedy et al., 2014; Lopez-Otin, Blasco, Partridge, Serrano, & Kroemer, 2013). At the organismal level, senescence affects individual fitness by increasing the risk of mortality and reducing reproductive contributions with increasing age (Charlesworth, 2000). At the population level, senescence leaves a signature in the vital rates of accelerating age-specific mortality with advancing age, often accompanied by declining age-specific fecundity rates (Promislow & Bronikowski, 2006). Thus, the comprehensive study of senescence requires an integrative approach across many scales of biological organization, from the level of individual genes and single molecules upwards through cellular, organismal and population-level approaches (e.g. Massot et al., 2011). In addition, a complete understanding of senescence requires a comparative approach. Comparisons among different species can yield insights into those mechanisms of ageing that are fixed or constrained versus those that have evolved independently in different lineages, providing insights into the raw materials on which natural selection and evolution operated (Austad, 2010). Sauropsids (including ectothermic and endothermic lineages—hereafter ‘reptiles’ and ‘birds’) as the sister clade to mammals are well-positioned phylogenetically for comparative studies of ageing, altogether forming the monophyletic amniotes. Within amniotes, studies of ageing in ectothermic reptiles are far fewer in number than those of mammals and birds. This is due in part to the fact that the ecology and life history of many reptiles makes them difficult to monitor (McDiarmid, Foster, Guyer, Chernoff, & Gibbons, 2012) and, in concert with the increasingly imperiled conservation status of many reptile species (e.g. Rhodin et al., 2018), poses significant logistical and ethical constraints on research. Regardless, the actual contribution of existing reptile research to our understanding of the evolution of ageing in its demographic and mechanistic dimensions may also be underestimated due to a lack of substantive reviews of key, up-and-coming research on this topic—an issue this contribution seeks to address.

Evolutionary theories of senescence predict that the rate of senescence is determined by the strength of selection against mutations with age-specific effects; that is, selection varies with age class giving rise to age-specific mutation–selection balance (Hamilton, 1966; Medawar, 1952; Williams, 1957). The strength of selection on age-specific mutations is shaped by demography; indeed, skewed age structure (with few older individuals in populations than younger) is sufficient for senescence to evolve (Hamilton, 1966). Importantly, evolution moulds mortality and fertility trajectories within specific ecological contexts (e.g. environmentally driven mortality, resource availability and pathogens; Williams & Day, 2003) subject to physiological and genetic constraints. At the same time, life histories, realized within an ecological context, determine the relative selection for early- versus late-life fitness, creating a feedback loop that shapes the timing and rate of senescent processes. Selection for reduced rates of senescence (and increased reproductive lifespan) should occur when environmentally driven mortality is lowered and reproduction increases beyond the age of first reproduction (Hamilton, 1966; Medawar, 1952). Indeed, such an early- versus late-life fitness balance should shape senescence rates; and senescence rates should evolve if this balance changes and there is genetic variation underlying the balance. We discuss variation in mortality and reproductive senescence of reptiles in Section 2 below. Of note, while we might expect longevity and senescence to coevolve, similar rates of senescence do not necessarily equal similar longevity (e.g. within species: Bronikowski et al., 2002; Miller, Janzen, Fellers, Kleeman, & Bronikowski, 2014; across species: e.g. Bronikowski et al., 2011; Ruby, Smith, & Buffenstein, 2018). Nevertheless, in understudied taxa such as reptiles, longevity is often used as a proxy for senescence rate. We take care throughout this review to acknowledge and distinguish the difference between longevity and senescence.

The preceding mutation–selection balance theory of the evolution of senescence tends to make inference from age-specific demographic patterns. In addition, several theories have emerged that attempt to explain the evolution of longevity and senescence rates as a response to correlated selection on functional traits (e.g. metabolism). These include the oxidative stress theory (formerly free-radical theory) (reviewed in Speakman et al., 2015), the pace-of-life theory (formerly the rate-of-living theory) (reviewed in Dammhahn, Dingemanse, Niemela, & Reale, 2018) and related thermal theories (e.g. cold-living, brumation) (Garbarino, Orr, Rodriguez, & Buffenstein, 2015; Stark, Tamar, Itescu, Feldman, & Meiri, 2018) for the evolution of variable rates of senescence and concomitant longevity. Although these can be thought of as evolutionary hypotheses in the sense that they involve correlated evolution to adaptations in physiological and molecular traits, their relationship to the evolution of senescence would still be viz. age-specific mutation–selection balance. Thus, we discuss these further in Section 3 below on mechanisms of senescence.

Our review of senescence in reptiles is structured into two parts generally aligned with biological levels of organization (see Figure 1). In the first section, we review population-level, demographic senescence—where selection and evolution occur (Figure 1a). And, in the second section, we review mechanisms of individual senescence. In reviewing the demographic senescence of reptiles, we start by considering how indeterminate growth and the evolution of ‘protective phenotypes’ that lead to low environmentally driven mortality rates are expected to shape the evolution of senescence (Figure 1b). We then review patterns of senescence within reptiles, focusing on longevity and rates of demographic decline across reptilian lineages. In reviewing the mechanisms of individual senescence, we organize the evidence around seven pillars of (human) ageing put forth by the Geroscience Initiative at the National Institutes of Health (USA; Kennedy et al., 2014; Figure 1c), and highlight areas where reptiles age similarly and contrastingly with mammals. We end our review with suggestions for needed research on reptile senescence and their potential for bringing new insights to the field of ageing biology.

Details are in the caption following the image
Unique features of reptile life history may alter selection on the rate of senescence and affect the evolution of underlying ageing mechanisms. (a) Indeterminant growth in reptiles, contrasted with mostly determinant growth in mammals and birds, in combination with age-dependent fecundity, increases selection against senescence. (b) Basic phylogenetic relationships among reptiles and mammals with examples of reptiles with protective phenotypes. (c) Senescence rates are influenced by variation in seven underlying pillars of ageing as described in Kennedy et al. (2014). The seven pillars are colour-coded by the theory of ageing they relate to: pink, oxidative stress theory; purple, pace-of-life theory; green, thermal theories. Photograph Credits: Eirini Pajak (western diamondback rattlesnake and desert tortoise), John Finger (alligator)

2 POPULATION SENESCENCE: DEMOGRAPHY, LIFE HISTORY AND SELECTION

2.1 Indeterminate growth

Indeterminate growth—defined as extended growth past the age of first reproduction that may continue throughout life—has important implications for the evolution of senescence. Increasing fecundity with size reduces the relative contribution of early age classes to overall fitness, which at least in part counters the lower probability that animals survive and contribute at later ages. Consequently, indeterminate growth coupled with size-dependent fecundity can increase reproductive output and fitness contributions of older individuals, thus increasing selection against senescence. The extent to which this occurs depends on the proportion of overall growth that remains after reproduction begins and the degree to which size determines the quantity and quality of offspring (Vaupel, Baudisch, Dolling, Roach, & Gampe, 2004).

Debate continues about the extent to which reptile growth is indeterminate (e.g. Congdon, Gibbons, Brooks, Rollinson, & Tsaliagos, 2013). For example, recent reports on American alligators (Alligator mississippiensis) suggest that they are determinate growers with growth cessation sometime after maturation (e.g. Wilkinson, Rainwater, Woodward, Leone, & Carter, 2016). Similarly, a recent study of green sea turtles and loggerhead turtles argues that marine turtles exhibit determinate growth, adding to the growing evidence for determinate growth in turtles (Omeyer, Fuller, Godley, Snape, & Broderick, 2018). However, even if growth is not truly indeterminate, continued growth past the age of first reproduction is the norm for reptiles and distinguishes them from mammals and birds (Hariharan, Wake, & Wake, 2016). Birds generally reach asymptotic size by the time they gain flight (Gill, 2006). Similarly, mammals are generally near their full size by the time they reach the age of first reproduction. For instance, the ratio between length-at-maturity versus length-at-maximum adult body size is 0.95 for elephants (Elaphus maximus), 0.9 for New Zealand fur seals (Arctocephalus forsteri) and 0.97 for polar bears (Ursus maritimus) (Derocher & Wiig, 2002; McKenzie, Page, Goldsworthy, & Hindell, 2007; Mumby et al., 2015). By contrast, the proportion of remaining growth after reproduction among snakes, turtles and lizards averages 0.68, 0.70 and 0.74, respectively (Shine & Charnov, 1992; Shine & Iverson, 1995). This suggests that there is ample opportunity for increased strength of selection against senescence at later ages in reptiles relative to birds or mammals, at least in so far as indeterminate growth is coupled with indeterminate fecundity.

The relationship between size and fecundity is complex. Although in many species, fecundity increases as individuals grow, the pattern is by no means ubiquitous within reptiles (reviewed in Sauer & Slade, 1987; Thorbjarnarson, 1996). For example, Congdon and van Loben Sels (1993) show that although clutch size is positively related to maternal body size for Blanding's turtles (Emydoidea blandingi), clutch size is not related to age. In cases such as this, the size–fecundity relationship is likely attributable to individuals who grow faster and larger having larger clutches rather than individuals increasing reproductive output as they age. Even so, numerous studies have shown a positive relationship between reproduction and size or age. For instance, Sparkman, Arnold, and Bronikowski (2007) found increasing reproduction with size among populations of the western garter snake with individuals in populations of fast-growing snakes increasing litter size with age at a faster rate than individuals in populations of slow-growing snakes. Similarly, Congdon et al. (2013) found clutch size increased by an average of 1 egg for every 8.6 years of age for 13 populations of 9 species of North American freshwater turtles, with a range of 2.3–18.5 years/egg.

2.2 Protective phenotypes

Traits that are associated with low annual mortality have the potential to also be associated with reduced rates of senescence. If reduced senescence rates increase fitness, then a reduction in environmentally driven mortality risk should increase the efficacy of selection on senescence rate, particularly in populations with density-dependent regulation (Williams, 1957; Abrams, 1993, but see Moorad, Promislow, & Silvertown, 2019). Protective phenotypes (reviewed in Schwartz & Bronikowski, 2011) are putative adaptations that directly impact environmentally driven mortality—possibly in an ontogenetic manner if the protections develop or wane over the course of life. Such protective phenotypes in reptiles include skin armour (i.e. scutes) in crocodilians, external ribcages (i.e. shells) in turtles and venom or toxicity in some lizards and snakes. For example, adult crocodilians are apex predators with few natural predators and high annual survival probabilities of young and old adults (ca. 0.80–0.89, American alligators, Garner, 2016). Even if a trait did not evolve for protective purposes, protective phenotypes can still have demographic consequences. For instance, while conventional wisdom acknowledges that turtle shells are protective and substantially reduce adult predation (e.g. Spencer & Thompson, 2005), the original fossorial turtle shell is thought to be an adaptation for digging (Lyson et al., 2016). Regardless, among all reptiles, turtles and tortoises are associated with low mortality and long lifespan (Gibbons, 1987; Shine & Iverson, 1995). Still, a mortality comparison of hard-shelled and soft-shelled turtles, controlling for body size, has not been undertaken and would provide a welcome test of the importance of protective phenotypes for the evolution of senescence.

The protective role of venom or poison is perhaps even more equivocal and deserves further study. Chemical protection (venom or poison) is positively correlated with maximum lifespan after controlling for body size in fishes, snakes, caudatans and anurans (Blanco & Sherman, 2005). And relative to non-venomous colubrids, timber rattlesnakes (Crotalus horridus) and cottonmouths (Agkistrodon piscivorus) have some of the highest annual survival probabilities of all snakes (0.80–0.90) (Brown & Simon, 2018; Koons, Birkhead, Boback, Williams, & Greene, 2009). But these high annual survivals are also not unheard of in non-venomous snakes living in populations with low predation (western garter snakes: Miller, Clark, Arnold, & Bronikowski, 2011). Furthermore, a recent phylogenetic analysis of amphibian and snake species that are highly toxic or venomous suggests that venomous snakes, which use their venom primarily to acquire prey, are not more longevous than non-venomous snakes (Hossie, Hassall, Knee, & Sherratt, 2013).

2.3 Longevity and demographic senescence in reptiles

Longevity in reptiles ranges from one year (Labord's chameleon, Furcifer labordi) to ~150 years (Galapagos tortoise, Chelonoidis nigra) (Eckhardt, Kappeler, & Kraus, 2017). Although not a measure of change or decline with age, which is the definition of senescence, longevity provides important insights into ageing processes. Longevity requires that survival remain high into late ages and may be associated with increased fitness contributions from later-age phenotypes. The evolution of extreme longevity likely requires low environmentally driven mortality and strong selective pressure to reduce rates of senescence (Finch, 2009), resulting in negligible senescence or even negative senescence (i.e. a decline in mortality with age, Vaupel et al., 2004). Practically speaking, records of longevity are more easily obtained than measures of actuarial senescence. While estimates of longevity are still dependent on the number of observations for a species and ease of determining age, observations of maximum lifespan do provide insights into general patterns. Among the chordates, current records of animals living longer than 100 years are restricted to a handful of (mostly) marine fishes, a few large marine mammals, humans and five turtle species (de Magalhaes & Costa, 2009; Table 1). Within reptiles, turtles comprise the majority of species with the longest lifespan records (Gibbons, 1987), with 23 species documented to be capable of reaching at least 40 years (Table 1). Crocodilians are also relatively long-lived with 18 of 21 species in the AnAge database having longevity records >40 years, albeit mostly based on records of captive animals (Weigl, 2014). Similarly, tuatara, the only extant species in the order Rhynchocephalia, are reported to live as long as 90 or more years in captivity (Castanet, 1994). In contrast, only seven of the 401 squamate (snakes and lizards) species in the AnAge database have maximum recorded lifespans of >40 years. These patterns are generally consistent with adult survival rates, with turtles having consistently higher annual survival than squamates (Shine & Charnov, 1992; Shine & Iverson, 1995).

Table 1. A selection of maximum observed lifespan estimates for long-lived reptile species, defined here as >40 years, (AnAge database; de Magalhaes & Costa, 2009; Accessed March 13, 2019)
Species Max lifespan (years)
Galapagos tortoise (Chelonoidis nigra) 177a
Aldabra tortoise (Aldabrachelys gigantea) 152a
Eastern box turtle (Terrapene carolina carolina) 138
Greek tortoise (Testudo graeca) 127a
European pond turtle (Emys orbicularis) 120a
Other turtles (7 species) 40–77
Other turtles and tortoises (11 species) 40–70.3a
Tuatara (Sphenodon punctatus) 90a
West African dwarf crocodile (Osteolaemus tetraspis) 70a
Other crocodiles (10 species) 41−56a
Komodo dragon (Varanus komodoensis) 62
Other squamates (6 species) 40.4−54a
  • a Estimated from a captive individual.

Terrestrial and fresh-water turtles have long been models of longevity (e.g. Congdon et al., 2003; Gibbons, 1987), but recent studies suggest that turtles may exhibit significantly more variation in demography than previously appreciated (Warner, Miller, Bronikowski, & Janzen, 2016, see also Spencer & Janzen, 2010). While the representative turtle (Gopherus agassizii) used in a recent meta-analysis of ageing is notable for declining mortality with age (Jones et al., 2014), comparative analyses of ageing rates in painted turtles (Chrysemys picta) and western garter snakes found that mortality accelerated across the adult lifespan (Miller et al., 2014).

Unfortunately, despite considerable research on marine turtles, our understanding of sea turtle ageing is primarily limited to demographic studies and skeletochronological estimates of age (reviewed in Chaloupka & Musick, 1997). Estimates of annual survival in adult sea turtles, including age-class specific estimates for populations of loggerhead sea turtles (Caretta caretta, Chaloupka & Limpus, 2002) and green sea turtles (Chelonia mydas, Chaloupka & Limpus, 2005) do not yet allow estimation of actuarial senescence. Demographic studies are rare in general for reptiles, resulting in a lack of knowledge regarding demographic senescence relative to endothermic vertebrates (e.g. see review Ricklefs, 2010; also Jones et al., 2014). Studies that do exist indicate a wide range of patterns in the demographic senescence of reptile mortality and fecundity—a range which largely overlaps estimates from other animal lineages (e.g. Colchero et al., 2019; Warner et al., 2016). Ricklefs (2010) suggested that patterns of actuarial senescence in reptiles are similar to birds. However, all reptile studies included in the comparative analysis were of captive populations. Captivity can impact animal physiology and behaviour in complex ways with significant implications for animal health and longevity (e.g. Lahdenperä, Mar, Courtiol, & Lummaa, 2018). In contrast, studies by Congdon, Nagle, Kinney, and van Loben Sels (2001), Congdon et al. (2003) in two species of wild turtles showed no increases in mortality with age. More recently, Warner et al. (2016) reported evidence for increased mortality with age in a wild population of painted turtles, but the rate of increase was slow relative to western garter snakes, yellow-legged frogs (Rana sierrae) and other well-studied taxa (Miller et al., 2014). In fact, some of the best candidates for negligible senescence and extreme longevity are reptiles (Gibbons, 1987; de Magalhaes & Costa, 2009). A recent comparative study by Jones et al. (2014) demonstrated that three reptile species (freshwater crocodile [Crocodylus johnstoni], common lizard [Zootoca vivipara] and desert tortoise [Gopherus agassizi]) showed slow or negative rates of senescence with respect to mortality. In addition, two of the three showed no decline in reproduction with advancing age (crocodile and desert tortoise). Thus, while limited, the majority of studies in the wild suggest that slower ageing may in fact be widespread among reptiles.

While the limited availability of long-term demographic studies in reptiles has often precluded robust tests of life-history theories for ageing, several single-species estimates of longevity and somatic growth rate give some indication that sea turtle longevity may be governed by life-history tradeoffs and possibly an allometric relationship with body size. The smaller Kemp's ridley sea turtle (Lepidochelys kempii) has a maximum estimated age of 30.25 year with reproductive longevity estimated to be less than or equal to 10 year post-maturation (mean = 4.5 year) (Avens et al., 2017). The larger North Atlantic loggerhead sea turtles has a maximum estimated age of 70–77 year with reproductive longevity less than 46 year (but mean = 19 year) (Avens et al., 2015). It is possible that the relatively small size and shorter interbirth interval of the Kemp's ridley turtle contributes to its relatively shorter lifespan.

Studies of reproductive senescence in reptiles are even more rare than studies of actuarial senescence. We know that female three-toed box turtles can live >60 year with no apparent decline in reproductive performance (Miller, 2001). Furthermore, considerable increases in fecundity with age can occur in reptiles even in the oldest documented wild individuals (e.g. Sparkman et al., 2007; Warner et al., 2016, but see Brown, 2016)—although this does not preclude reproductive senescence at the end of life (Warner et al., 2016, see also Massot et al., 2011). Similarly, long-term studies of sea turtles generally find increasing reproductive output with age (Garner, 2012). Empirical data for this pattern come primarily from females, due to the inherent bias of sampling different sexes of some species (e.g. semi-aquatic turtles; Reinke, Hoekstra, Bronikowski, Janzen, & Miller, 2019). It is possible that sex-specific growth promotes sex-specific selective environments (e.g. Hoekstra, Weber, Bronikowski, & Janzen, 2018) and results in sex-specific ageing and longevity (Austad & Fischer, 2016), but this question remains elusive at this time for reptiles.

Overall, great strides have been made in documenting variation in longevity and rates of ageing among reptiles. The intriguing phenomenon of extreme longevity coupled with evidence for slower ageing relative to birds and mammals suggests that the field of comparative ageing would be greatly enriched by more in-depth, long-term demographic studies on a wider range of reptiles.

3 INDIVIDUAL SENESCENCE: MECHANISMS

We present this section in the framework of the seven pillars of ageing (Kennedy et al., 2014; see also Lopez-Otin et al., 2013), which represent common drivers of the ageing process contributing to ageing phenotypes and age-related disease across animals, but have typically focused on humans and other mammals. We choose this framework to (a) emphasize evidence for these ageing processes in reptiles, (b) contrast reptiles with mammals and (c) highlight fruitful future research directions using reptiles as models to understand ageing biology. Direct evidence for the pillars of ageing in reptiles is summarized in Table 2 and detailed below.

Table 2. Summary of evidence for mechanistic pillars of ageing in reptiles (sensu Kennedy et al., 2014)
Pillar of ageing Specific mechanism Effect on reptiles (# of citations)
Macromolecular damage DNA damage + Correlation with age/lifespan (3a)
No association with age (3b)
Lipid and protein damage Accumulates with age (2c)
Epigenetics Methylation Decreases with age (1d)
Inflammation Adaptive immune system No change with age (2e)
+ Association with ageing (1f)
Innate immune system No change with age (3g)
Expansion of gene families (1h)
Adaptation to stress Reduced oxidative stress + Association with longevity (4i)
Mixed results (2j)
Proteostasis Homology Sequence conservation (3k)
Expansion of gene families (2l)
Stem cells and regeneration Limits on cell division Correlated with lifespan (1m)
Telomere dynamics Declines with age (3n)
Sex-specific pattern (3o)
No change with age (1p)
Regeneration Unknown
p53 pathway Molecular evolution (1q)
Metabolism Nutrient sensing Molecular evolution (1r)

Note

  • Within each ‘pillar of ageing’ category, we summarize published evidence supporting the effect of ‘specific mechanism(s)’ on the ageing process of reptiles. In the absence of functional studies on mechanisms of ageing in reptiles, we highlight work on molecular sequence conservation and evolution within canonical ageing pathways.
  • a Bronikowski (2008); Robert and Bronikowski (2010); Schwartz and Bronikowski (2013).
  • b Woodhead et al. (1980); Regan et al. (1982); Schwanz et al. (2011).
  • c Manibabu and Patnaik (1991); Jena et al. (1995).
  • d Nilsen et al. (2016).
  • e Zimmerman et al. (2017); Palacios and Bronikowski (2017).
  • f Richard et al. (2012).
  • g Sparkman and Palacios (2009); Palacios et al. (2011); Schwanz et al. (2011).
  • h Quesada et al. (2019).
  • i Robert et al. (2007); Olsson et al. (2008); Olsson et al. (2009); Robert and Bronikowski (2010); Schwartz, Arendsee, and Bronikowski (2015).
  • j Schwartz and Bronikowski (2013); Isaksson, While, Olsson, Komdeur, and Wapstra (2011).
  • k Shaffer et al. (2013); Quesada et al. (2019), Schield et al. (2019).
  • l Harrison et al. (2010); Shaffer et al. (2013).
  • m Christiansen et al. (2001).
  • n Scott et al. (2006); Xu et al. (2009); Ujvari et al. (2017).
  • o Bronikowski (2008); Olsson et al. (2011); Rollings et al. (2017).
  • p Ujvari and Madsen (2009).
  • q Passow et al. (2019).
  • r Sparkman et al. (2012); McGaugh et al. (2015).

3.1 Macromolecular damage

Despite the existence of cellular protection and repair mechanisms, accumulating damage to proteins, lipids and DNA has long been a hallmark of ageing (Maynard, Fang, Scheibye-Knudsen, Croteau, & Bohr, 2015). Whether due to imperfect, age-independent repair mechanisms per se, or whether due to decreasing efficiency of repair mechanisms with age, studies across broad comparative landscapes show that accumulating damage has myriad negative downstream effects (e.g. MacRae et al., 2015). In reptiles, understanding of DNA damage repair mechanisms is limited, but does include some studies on lipid membrane damage (Jena, Das, & Patnaik, 1995; Manibabu & Patnaik, 1991). Less research has focused on damaged proteins. In this section, we highlight what is known about damage and repair in reptiles; in other sections, we focus on the degradation of protective mechanisms (e.g. telomeres, epigenetics, protein chaperones) with age.

3.1.1 DNA damage

The structure and function of DNA is 100% conserved among all living organisms. As a consequence, impairment resulting from DNA-damaging agents is also conserved, as are the mechanisms to repair damage (Uphoff, Reyes-Lamothe, Leon, Sherratt, & Kapanidis, 2013). Because deficiencies in DNA repair capacity have been linked to the induction of replicative senescence and progeroid syndromes, it is reasonable to suppose that functional DNA repair pathways are an essential component of longevity, both within and among species. Early studies on mammals supported this hypothesis, but have been criticized for failing to control for phylogeny and body size (Promislow, 1994). More recently, Page and Stuart (2012) demonstrated that body mass accounted for most of the variation seen in DNA repair capacity in an array of endothermic vertebrates, particularly in the context of oxidative DNA damage. Furthermore, early studies on DNA repair and longevity in some ectothermic vertebrates (turtles and fish) found no association; individual species had approximately equal amounts of repair after insult (Regan, Carrier, Samet, & Olla, 1982; Woodhead, Setlow, & Grist, 1980). Unfortunately, interpretation of these data is confounded by the small number of species and individuals tested, as well as technical issues—most notably heavy use of domesticated species, differences in euthanasia and sampling techniques, as well as differences in the specific methods used to quantify DNA repair. Clearly, more research is needed to test whether age-related DNA repair deteriorates similarly between mammals and reptiles, as well as to clarify the significance of these processes as they relate to organismal ageing.

Central to this mechanistic damage accumulation hypothesis for senescence is the assumption that declining ability to repair DNA damage with age causes an accumulation of damaged cells as well as somatic mutations (Gladyshev, 2013). In reptiles, in the absence of age-specific repair capability, we can look to studies of associations between DNA repair capacity and longevity for a first approximation. In painted turtles, juveniles and adults did not differ in their DNA repair capacity (Schwanz, Warner, McGaugh, Terlizzi, & Bronikowski, 2011). In snakes, long-lived species had more efficient DNA repair than short-lived species (Bronikowski, 2008), and slow-ageing ecotypes of garter snakes had higher capacity to repair DNA relative to fast-ageing ecotypes (Robert & Bronikowski, 2010; Schwartz & Bronikowski, 2013). In the same slow- and fast-ageing ecotypes of garter snakes, cells grown from larger/older females were more resistant to the lethal effects of the DNA-alkylating reagent methyl methanesulfonate (MMS) than were cell lines grown from smaller/younger females irrespective of ecotype (Alper, Bronikowski, & Harper, 2015). This result showed the opposite pattern than predicted if DNA repair capacity declines with age. Moreover, using dermal fibroblast cells, reptile cells were more resistant to a battery of cytotoxins, including MMS, than birds and mammals (Alper et al., 2015).

3.1.2 Lipid & protein damage

Membrane fatty acid composition changes over the lifespan, and profiles of damaged membrane fatty acids change with age in mammals (Hulbert, 2010; Hulbert, Pamplona, Buffenstein, & Buttemer, 2007). Changing fatty acids give rise to variation with age in lipid peroxidation, which is both damaging to lipids and other molecules. While a vast literature exists on mammal and bird age-specific membrane lipid profiles and their peroxidation index, data are sparse at best in reptiles. In garden lizards (Calotes versicolor), lipid composition changes with age in agreement with mammals, and lipid peroxidation increases with age in brain, kidney and liver (Jena et al., 1995; Manibabu & Patnaik, 1991). Accumulating protein carbonyls, which occurs when carbonyl groups are attached to proteins through interactions with free radicals or damaged lipids, is a stable hallmark of ageing. One study in asp vipers (Vipera aspis) linked protein carbonylization to the stress response (Stier et al., 2017) and another study in garden lizards correlated carbonyl accumulation with age (Jena, Das, & Patnaik, 1996), but no studies in reptiles have directly linked stress to ageing via increased carbonylization.

3.2 Epigenetics

Because the epigenome integrates environmental cues and changes over the lifespan, epigenetics has been a fast-growing area of research in ageing biology (Booth & Brunet, 2016; Pal & Tyler, 2016). Not only may epigenetic changes correlate with advancing age (i.e. the ‘epigenetic clock’, Horvath & Raj, 2018), but these age-related changes in the epigenome have been found to vary across species (de Paoli-Iseppi et al., 2017). Thus, they may be predictive of stress and ageing rates in other species as well (see Parrott & Bertucci, 2019).

Although few researchers have queried age-related changes in the epigenome of reptiles, some progress has been made in quantifying methylation. For example, alligators, like other vertebrates, show decreasing global methylation with advancing age (Nilsen et al., 2016). In general, reptiles (including crocodilians, lizards, snakes and turtles) have global methylation levels that span the range observed in mammals and fishes (i.e. relatively lower to higher levels). This greater range than expected global methylation occurs despite similar GC content and CpG motifs as other vertebrates (Head, Mittal, & Basu, 2014; Varriale & Bernardi, 2006). How methylation and other epigenetic modifications change with advancing age in reptiles and whether this affects transcriptional networks and age-related phenotypic decline in reptiles are areas that should be a future avenue for research.

3.3 Inflammation

The ubiquitous nature of increasing inflammation with age has been a long-appreciated aspect of ageing. Under this category, reports on age-related innate immune function are relatively plentiful in reptiles, while fewer studies have tested for changes in adaptive immunity with age (e.g. lymphocyte proliferation). No studies, to our knowledge, have sought to test whether senescent cells of reptiles secrete pro-inflammatory signals to neighbouring cells as is seen in model systems (Campisi, 2018).

Central to the immunosenescence hypothesis for organismal ageing is the assumption that declining immune function, or immunocompetence, causes organismal senescence (Brand & Nicholls, 2011; Maue & Haynes, 2009). For innate immunity, although reptile studies have demonstrated significant plasticity in innate immune function in relation to biotic and abiotic factors (Palacios, Cunnick, & Bronikowski, 2013; Refsnider, Palacios, Reding, & Bronikowski, 2015; Zimmerman, Carter, Bowden, & Vogel, 2017), there has been no consistent finding of an effect of age on innate immunity (Palacios, Sparkman, & Bronikowski, 2011; Sparkman & Palacios, 2009; Schwanz et al., 2011) in reptiles. Regarding acquired immune proliferative ability in reptiles, Zimmerman et al. (2017) found no decline with age in B-cell proliferative ability in red-eared slider turtles, a species that also exhibits plasticity in B-cell function with temperature (Palackdharry, Sadd, Vogel, & Bowden, 2017). Similarly, while we have found differences in T-cell proliferative ability between pregnant and non-pregnant females in garter snakes, we have not found a decline in immunocompetence with advancing age (Palacios & Bronikowski, 2017), despite differences between fast- and slow-ageing ecotypes (Palacios et al., 2013).

The immunosenescence mechanism of ageing has been examined in the common lizard (Z. vivipara), where polyandrous and monandrous females coexist and mating strategy is condition- and context-dependent. Polyandrous females have higher reaction to inoculation with phytohaemagglutinin, a reaction that involves both T-cell responses and innate immunity. Polyandrous females also have declining reproductive success (litter size and offspring quality) with age. Thus, there is a positive correlation between increased immunocompetence in an ecotype that has reproductive senescence, contrary to what one might predict under the immunosenescence hypothesis (Richard, Massot, Clobert, & Meylan, 2012).

Interestingly, Quesada et al. (2019) reported lineage-specific duplication of immunity-related genes in the long-lived giant tortoise and other turtles. These included gene duplications thought to be involved with the pro-apoptotic activity of granzymes and cytoxic T lymphocytes (e.g. PRF1, chymase), as well as the expansion of some gene families related to viral, microbial, fungal and parasitic innate immune defence. They argue that the expansion/diversification of innate immunity genes in turtles suggests that innate immunity may be relatively more important for turtles than mammals, but functional studies are required.

Important to both innate and adaptive immunity are the various cytokines. Cytokines are small proteins secreted by immune system cells, typically implicated in cell signalling and the inflammation response. While cytokine research in reptiles has lagged behind other aspects of immunity (reviewed in Zimmerman, Bowden, & Vogel, 2014), recent advances in reptile cytokines suggest that understanding their influence over senescence in reptile immunocompetence may be a reasonable near-term goal (e.g. Rayl, Wellehan, Bunick, & Allender, 2019; Zhang, Chen, Yuan, & Niu, 2015; Zhou, Guo, & Dai, 2009).

3.4 Adaptation to stress

Reptiles possess many morphological and physiological adaptations that may lend protection to individuals (Schwartz & Bronikowski, 2011). These may fundamentally alter the trajectory of age-specific mutation–selection balance in populations—skewing populations towards older individuals with more right-skewed age structures relative to mammals and birds. Examples include extended metabolic shutdown, starvation resistance, freeze and heat tolerance and hypoxia resistance. Clade-specific adaptations also include shells, super-cooling and anoxia tolerance in some turtles and venom and intermittent feeding and gut remodelling in some snakes (reviewed in Schwartz & Bronikowski, 2011). Many of these adaptations require protection and repair of cellular components that typically accumulate damage with age. One of the requirements for an effective response to environmental stressors is a set of robust mechanisms for proteostasis, including heat shock proteins (HSPs) for chaperoning unfolded proteins (e.g. Krivoruchko & Storey, 2010) and targeting denatured proteins for autophagy and proteosomal degradation. It is worth considering how the adaptive mechanisms for extreme-stress response may alter senescent parameters and perhaps even alter how different species age. For example, reptile species that have evolved a highly adapted HSP response may have enhanced ageing resistance via effective triage of damaged proteins, but still perhaps senesce through alternative mechanisms. We engage the state of the field with regard to the relationships between ageing and both thermal and oxidative stress in more detail below.

3.4.1 Thermal stress

Resilience to extreme temperature fluctuations, with ability to adjust metabolic rate from low to high with temperature (e.g. 4–40°C, e.g. Gangloff, Holden, Telemeco, Baumgard, & Bronikowski, 2016), raises the interesting prospect of correlated evolutionary responses of stress resistance. If present, these evolved stress responses should utilize molecular networks and cell-signalling pathways common to those found in animal models of ageing and could subsequently reveal nodes in these pathways that have been free to evolve (see the seventh pillar below on Metabolism for empirical data supporting this contention).

At the cold end of the thermal spectrum, brumation, like hibernation, is a metabolic adaptation that allows animals to persist in seasonally frigid environments. Both brumation and hibernation are associated with increased survival and the evolution of slower senescence in reptiles (brumation) and mammals (hibernation) (Ruf, Bieber, & Turbill, 2012). However, brumation is a very different metabolic state than hibernation and is characterized by dormancy (very low metabolism and heart rate) with periods of activity and is controlled by temperature. Both hibernating mammals and brumating reptiles would be expected to face the same oxidative stress upon increasing metabolism and reperfusion of oxygenated blood. Other theories of senescence arise from the observation that cold thermal environments are often associated with longevous species or representative populations (e.g. Stark et al., 2018). Lower body temperatures could simply delay the attritive effects of metabolic by-products, but the mechanisms behind the preservative effects of cold temperature are still elusive (Flouris & Piantoni, 2015).

At the warm end of the thermal spectrum, protective mechanisms consistent with mammals have been observed, such as upregulation of HSPs, increased activity of the hypothalamic–pituitary–adrenal axis and increased oxidative stress resistance (Schwartz & Bronikowski, 2013). This last factor, oxidative stress resistance, has been a long-studied protective mechanism for damage accumulation and has direct relation to the proposed mechanism of mitochondrial dysfunction with age.

3.4.2 Oxidative stress

As noted above (3.1 Damage), one of the most cited causes of ageing is damage to macromolecules. One specific type of damage is oxidative damage that is a by-product of aerobic respiration (Sena & Chandel, 2012) and oxidative phosphorylation by mitochondria, especially in the context of chronic stress conditions faced by free-living organisms (Salmon, Richardson, & Perez, 2010). More specifically, damage to macromolecules, such as mitochondrial DNA (Barja & Herrero, 2000) and membrane lipids (Hulbert et al., 2007), as a consequence of mitochondrial-derived reactive oxygen species (ROS), is thought to lead to age-related pathology and functional decline (Wallace, 2005). Both within and among mammals and birds, many studies have supported the notion that longer-lived species tend to have more efficient mitochondria, less ROS production and/or increased antioxidant capacities (e.g. Barja & Herrero, 2000; Ku, Brunk, & Sohal, 1993; Lambert et al., 2007). At the same time, many exceptions to this pattern exist, with perhaps the most famous being naked mole rats that demonstrate high levels of ROS and oxidative damage (Perez et al., 2009), with little impact on longevity (Ruby et al., 2018). Notwithstanding, life-extending dietary interventions, such as caloric restriction, methionine restriction and every-other-day feeding, are also associated with a decline in mitochondrial ROS (e.g. Caro et al., 2008; Gomez et al., 2007; Sanchez-Roman et al., 2012; Sohal, Ku, Agarwal, Forster, & Lal, 1994).

There is much interest in the role of oxidative stress, primarily through ROS, in shaping variation in life histories, including ageing rates and longevity in reptiles. For example, in painted lizards, Olsson et al. (2008), Olsson, Wilson, Uller, and Isaksson (2009) found that longevity-associated life-history traits are predicted by the degree of ROS production. In colubrid snakes, Robert, Brunet-Rossinni, and Bronikowski (2007) found a positive correlation between ROS production and longevity. Studies in painted and slider turtles have shown that they are well-suited to resist the effects of oxidative stress during cooling, with increased neuronal antioxidants during cooling (reviewed in Rice et al., 2002). Slow-ageing neonatal garter snakes have more efficient mitochondria and cellular antioxidant defences than fast-ageing neonates, despite equivalent mass-independent metabolic rates (Robert & Bronikowski, 2010). Not surprisingly, however, results that contest oxidative stress as a mechanism of ageing have been observed in reptiles as they have in mammals. For example, in a subsequent study on these same garter snakes populations, slow-ageing garter snakes produced higher baseline levels of ROS in the blood that decreased in response to acute heat stress, relative to fast-ageing populations (Schwartz & Bronikowski, 2013). Interestingly, in contrast to mammals, reptiles and birds have nucleated red blood cells with mitochondria and gene expression in these cells includes both antioxidants and HSPs (Waits, Simpson, Sparkman, Bronikowski, & Schwartz, 2019). Despite a lack of consensus with respect to the role of oxidative stress in reptile ageing, the combination of these intriguing findings and potential novel protective mechanisms in non-mammalian amniotes warrants more research to elucidate the similarities and differences in mitochondrial energetics between reptiles and endothermic vertebrates.

3.5 Proteostasis

Maintaining cellular function requires accurate protein synthesis, folding, maintenance and removal of degraded proteins. The cellular pathways responsible for maintaining protein homeostasis or proteostasis, in humans are themselves comprised of at least 2,000 components (Klaips, Jayaraj, & Hartl, 2018). Failures at various steps along the proteostasis network cause disease states, including neurodegenerative disorders Alzheimer's and Parkinson's disease (Ross & Poirier, 2004). More generally, decline in the function of the proteostasis network and increased protein aggregation have been associated with ageing in model organisms (e.g. Labbadia & Morimoto, 2015). For non-avian reptiles, although we lack a comprehensive comparison of reptile proteome evolution, genome sequencing has generally revealed conservation of core components of the proteostasis network in reptiles (e.g. Quesada et al., 2019; Schield et al., 2019; Shaffer et al., 2013). There is some evidence for expansion of gene families associated with protein homeostasis in reptile lineages (e.g. prion proteins, immunoglobulins; Harrison, Khachane, & Kumar, 2010; Shaffer et al., 2013), but the functional or adaptive role of these gene expansions is still unknown. Given the documented role of pseudogenization and gene loss across multiple reptile genomes (e.g. Liu et al., 2015; McGaugh & Schwartz, 2017; Shaffer et al., 2013), we anticipate that loss or change in function of proteostatic genes will also be a key feature of diversification of proteostasis in non-avian reptiles. Most functional work on proteostasis in reptiles has focused on the environmental stress response and molecular chaperone function (Chang, Knowlton, & Wasser, 2000; Heikkila, 2017; Takii et al., 2017). These studies provide basic knowledge of the composition and patterns of expression of the molecular stress response in reptiles. Notwithstanding, nothing is specifically known about age-related declines in molecular chaperone function and about other components of the proteostasis network, such as protein removal, in any context.

3.6 Stem cells and regeneration

3.6.1 Germline stem cell continuity

One aspect of reproductive biology that distinguishes reptiles among amniotes is the presence of oogenesis from stem cells throughout the reproductive lifespan of females (reviewed in Guraya, 1989). The germinal bed of reptilian ovaries is a unique morphological feature supplying the abundant supply of oogonia—the germ stem cells. Our understanding of reproductive senescence in reptiles is data deficient, particularly for longer-lived species whose longevity requires repeated captures of females to document reproduction. As mentioned earlier, the striking reproductive senescence seen in mammals is absent or nearly so in many reptiles (e.g. Sparkman et al., 2007; Warner et al., 2016) is perhaps in part due to the ‘unlimited’ supply of gametes without an expiration date, though more work needs to be done to investigate this possibility further,

3.6.2 Cellular senescence

The relationship between cellular replicative senescence and whole-organism senescence has been a topic of debate for several decades. Recent clarity has been brought to this question by the discovery that senescent cells are not simply accumulated non-functioning cells, but are actually cells that dramatically alter their near-environment by the release of pro-inflammatory signals (Campisi, 2018). In the field of human ageing, this has spurred research into chemicals (i.e. senolytics) that can selectively destroy senescent cells. Notwithstanding, normal cells cannot divide indefinitely due to the shortening of telomere ends of chromosomes with each cell division. Although whether the cell division upper limit of reptile cells is equal to that of mammals (i.e. the ‘Hayflick limit’, Hayflick & Moorhead, 1961) has not been well-studied, several turtle species indicate that the maximum number of cell divisions may be correlated with species' longevity (Christiansen et al., 2001).

3.6.3 Telomere dynamics

Telomeres are repetitive nucleotides at the ends of eukaryotic chromosomes that potentially protect DNA from replication errors. Telomere length is commonly associated with longevity in humans and telomerase-mutant mice have reduced longevity (reviewed in Gomes, Shay, & Wright, 2010). Telomere shortening with age is thought to be due to (a) telomeres shortening during replication (Blackburn, 2005) and (b) telomerase shutting off in adult somatic tissues as an anti-tumour mechanism (Tanaka, Horikawa, Barrett, & Oshimura, 2005). In contrast to mammals, some studies in reptiles indicate that telomerase may not be turned off in adult somatic tissues. A meta-analysis of wild-animal telomere dynamics, including 14 birds and 3 non-avian reptiles, found general support for the correlation between telomere maintenance and longevity (Dantzer & Fletcher, 2015). However, in the three non-avian reptiles used in the meta-analysis, the relationship between telomere length and age is complex (see Table 2). One study reported an association between telomere shortening and longevity in males (western terrestrial garter snake; Bronikowski, 2008). A second study also found evidence of age-related decline in telomere length of males, but not females, even though females with longer telomeres lived longer (sand lizard; Olsson et al., 2011). The third study found no evidence for telomere-related cellular ageing in either sex (water python (Liasis fuscus); Ujvari & Madsen, 2009). A study on the loggerhead sea turtle, which was not included in the meta-analysis, failed to detect a significant age-related decline in telomere length (Hatase et al., 2008), yet age-related decline was observed in crocodilians (Scott, Haussmann, Elsey, Trosclair, & Vleck, 2006; Xu, Wu, Yan, & Zhu, 2009). Even more complicating, a comprehensive study of telomere dynamics with age in frillneck lizards (Chlamydosaurus kingii) found that telomere length initially increases with age in the first few years of life followed by a decline, but that telomere length was unrelated to survival (Ujvari et al., 2017). Recent work in red-sided garter snakes found age-related telomere shortening in males, but not females, confirming that selection on telomere length is, at best, context-dependent (Rollings et al., 2017). An earlier review across Metazoans (Gomes et al., 2010, and references therein) found that many fishes, amphibians and reptiles have active telomerase activity in somatic tissue, whereas telomerase is repressed in most birds and mammals. Active telomerase has thus been hypothesized as a mechanism that allows for the regeneration capabilities seen in some species of reptiles (reviewed by Olsson, Wapstra, & Friesen, 2018). Comparability among studies is hampered by varying techniques used to measure telomere lengths (e.g. qPCR, telomere restriction fragments). Nonetheless, reptiles lend themselves to straightforward assessment of telomeres due to their nucleated red blood cells and thus the feasibility of sampling adequate blood volumes from even the smallest species.

3.6.4 Regeneration

Stem cells and progenitor cells may generally act as a somatic repair service capable of countering cellular senescence, but stem cells are also susceptible to the accumulation of deleterious changes that cause cellular senescence and may contribute to stem cell exhaustion. Senescence or exhaustion of various stem cell populations has been correlated with ageing in humans and other mammals (reviewed in Lopez-Otin et al., 2013). There is considerable interest in stem cell revitalization, perhaps through epigenetic mechanisms, to reverse ageing phenotypes (Rando & Chang, 2012). Within the animal kingdom, regenerative capacities correlate with ectothermy (Alvarado, 2000). Reptiles such as snakes and lizards and amphibians such as salamanders have tissue-regenerative capacities that far surpass that of mammals. Many salamanders and lizards can atomize their tails and regenerate the tail tissue, the extent of which is dependent upon the types of stem cells that remain active (Sun, Londono, Hudnall, Tuan, & Lozito, 2018). The gut tissue and many organs in pythons and boas that have infrequent meals degenerate in the weeks/months between feedings and then can fully regenerate within days of taking a meal (Secor & Diamond, 1998) utilizing molecular pathways known to regulate ageing (mTOR, NRF2; Andrew et al., 2017). These regenerative capacities of reptiles have brought considerable interest in the wound healing field (e.g. Jacyniak, McDonald, & Vickaryous, 2017; Lozito & Tuan, 2017), but there has been little consideration of how these regenerative capacities affect the rate of ageing in these species or whether these capacities themselves senesce. This is partly because we lack tools, such as reptilian cell lines, necessary to study the cellular biology of senescence in reptiles (see Alper et al., 2015). While evidence for cellular senescence and age-related decline of tail regeneration are still lacking, there are tantalizing examples of age-related patterns in the costs of tail regeneration (Vitt & Cooper, 1986) as well as condition- or environment-dependent effects on regenerative capacity such as decreased tail regeneration ability with increased parasite load (Oppliger & Clobert, 1997). Furthermore, there is also evidence for positive selection in geckos and anoles for genes potentially underlying tail regeneration, including genes related to cell proliferation and prostaglandin biosynthetic processing, which are known to be expressed during tail regeneration (Liu et al., 2015). All of these aspects could represent novel targets for ageing research.

3.6.5 Molecular network for cellular senescence: p53/p16

Cellular senescence can occur through mechanisms other than the replicative lifespan of cells. One of the primary drivers of cellular senesce is the p53 molecular network. Upon experiencing environmental stress that causes DNA damage, the p53 molecular network determines a cell's fate (i.e. death, senesce, repair), balancing anticancer and anti-ageing mechanisms. Recent work on the comparative genomics of the p53 molecular network provided evidence of positive selection on genes within p53 networks of reptile (both ecto- and endothermic) relative to mammals, with reptiles having more positive selection on upstream genes (p53 and its regulators), especially in lizards and snakes (Passow et al., 2019). This is in contrast to the purifying selection on p53 across mammals (Gaur et al., 2017; Passow et al., 2019), with the exception of the branch leading to elephants (Passow et al., 2019). We also found that species with longer lifespans experienced less variation in selection across the p53 network, whereas those with shorter lifespans were equally likely to have both positive and purifying selection (Passow et al., 2019). A recent publication on Galapagos tortoise genomes (Quesada et al., 2019) identified variants in p53 that are associated with hypoxia tolerance in other species (Zhao et al., 2013), providing an interesting link between this network and stress resistance in an exceptionally long-live species. Future work is needed to test functional consequences of the observed amino acid changes on rates of senescence and cancer.

3.7 Metabolism and deregulated nutrient sensing

In the last 25 years, molecular networks have been identified in laboratory model organisms that govern stress responses and senescence at the cellular level and ultimately govern stress responses and senescence at the organismal level (Greer & Brunet, 2008; Riera, Merkwirth, Magalhaes, & Dillin, 2016). These networks are antagonistically pleiotropic in the sense that increasing or decreasing signalling through their pathways shifts resources between growth and reproduction on the one hand or maintenance and survival on the other (Barzilai, Huffman, Muzumdar, & Bartke, 2012). This has led to a general hypothesis that the fundamental life-history trade-off between growth and reproduction versus maintenance and survival may be controlled by variation in signalling through these networks—both within and among species—and may underlie intra- and interspecific variation in longevity (Bartke, 2017; Bartke, Sun, & Longo, 2013; Schwartz & Bronikowski, 2011). Specifically, these homologous signalling networks include nutrient sensing pathways, namely the insulin/insulin-like signalling pathway (IIS), TOR, AMPK and Sirtuins. Across model species, dietary restriction extends health and/or longevity through these nutrient signalling networks (reviewed in Allison et al., 2014). Given the unique characteristics and plasticity of many reptiles in terms of nutrient signalling and metabolic plasticity, infrequent feeding in many species of snakes and downregulation of metabolism with low temperatures and hibernation, reptiles provide unique avenues to understand these networks and their regulation of ageing.

Across animals, reduced signalling through the IIS and TOR pathways across the lifespan, involving low circulating levels of IGF1 or mutations in the IGF1 receptor, has been associated with increased longevity (e.g. Fontana, Partridge, & Longo, 2010). Paradoxically, in adults, a decline in the level of circulating IGF1 is also a biomarker of ageing (e.g. Sanders et al., 2017). While there is currently no information on how dietary restriction or other manipulation of the IIS and TOR pathways alters ageing in reptiles, these networks have been examined in an evolutionarily comparative framework within amniotes including reptile, bird and mammalian clades (McGaugh et al., 2015; Sparkman et al., 2012). While the overall structure of these molecular signalling networks is highly conserved, there is considerable evidence of positive selection on many components of this network between the reptile and mammal clades and specifically within the squamate (snakes and lizards) lineage. Included in this positive selection are top extracellular regulators (IGF1, insulin receptor and IGF1 receptor), key intracellular nodes, insulin receptor substrate, PI3K and mTOR (McGaugh et al., 2015). Genetic variation found in humans and mice models in the same nodes is associated with shifts in longevity. This suggests that natural selection is working on these networks to alter their function across these taxonomic groups. Future work is needed to test functional consequences of this diversifying selection on life history and rates of senescence.

In the context of the process of senescence within a lifetime, we have essentially no knowledge of how these networks are functioning in reptiles and whether their function varies across reptile and mammal species. As with most aspects of ageing biology, the majority of study has been on laboratory rodent models. Mammals have relatively high levels of IGF1 (protein and gene expression) compared to reptiles. In addition, laboratory rodents, in contrast to humans, do not express IGF2 as adults (Brown et al., 1986). Reptiles are more similar to humans in that they express IGF2 throughout their lifespan (Reding, Addis, Palacios, Schwartz, & Bronikowski, 2016) and could be used to study IGF2 in the context of natural ageing. Biomarkers of senescence related to these pathways from mammalian studies, such as the decrease in IGF1 with age, have yet to be investigated in reptiles in longitudinal or even cross-sectional studies.

In addition to the IIS/TOR pathway, increased signalling in AMPK and sirtuins is associated with longevity in laboratory models. Genes in the NAD+/sirtuin pathway, of interest due to connections to FOXO and mitochondrial signalling, are conserved in reptiles and appear relatively evolutionarily constrained, particularly with respect to fishes (Gaur et al., 2017). Of note, SIRT2, a member of the sirtuin family of protein deacetylases and involved in cellular stress responses and tumour suppression, demonstrates some positively selected sites in reptiles (Gaur et al., 2017). The functional significance of positive selection in SIRT2, the positive selection across p53 and IIS/TOR networks for reptiles remains unknown at this time.

4 CONCLUSIONS

Ectothermic reptiles possess several features ideally suited for increasing our understanding of the evolution of ageing—including many with the life-history patterns of indeterminate growth and increasing fecundity with age, and many with plastic physiologies adapted to coping with stressful environments throughout the course of life. We believe that the ultimate cause of variation in longevity is its dependency on the evolution of correlated life-history traits. Reptiles have three times more variation in life-history traits than mammals (e.g. Babich Morrow, Kerkhoff, & Ernest, 2019), representing vast evolutionary potential for innovative solutions to senescence.

This review has uncovered key aspects of biology where an understanding in reptiles would facilitate a broad understanding of the biology of ageing. Foremost is further characterization of senescence in reptiles and whether biomarkers of ageing in mammals are robust across reptile species. For example, we have yet to reach a conclusion on telomere dynamics with age in reptiles. And, while increasing levels of p16 (CDKN2A) and decreasing IGF1 are the best-known biomarkers of senescence in organismal and cell culture studies (Krishnamurthty et al., 2004; Ressler et al., 2006; Sanders et al., 2017), we have no information on how or whether this translates to reptiles. Even so, from rapidly evolving molecular networks that regulate ageing, to regenerative capacities and to adaptations to extreme environmental stress, innovations in the reptile lineage have already brought new insights into the biology of ageing and are promising avenues of future research.

Comparative study of reptile ageing requires knowledge of the demography, ecology, physiology and genomics of more reptile species. We recommend four actions to remediate these deficiencies: (a) extend and expand long-term population monitoring of reptiles, (b) develop reptile cell lines to aid cellular biology, (c) conduct more comparative studies of reptile morphology and physiology sampled along relevant life-history axes and (d) sequence the genomes of more reptile species to help solve the sequencing disparity (David, Wilson, & Halanych, 2019). Given the untapped potential of reptile biodiversity, and its particular alignment with relevant dimensions of ageing biology, acting on these research initiatives promises fruitful rewards for our understanding of how and why animals age.

AUTHORS' CONTRIBUTIONS

All authors conducted literature review, drafted sections and contributed to the final manuscript.

DATA AVAILABILITY STATEMENT

This manuscript does not use data.