Telomere length mirrors age structure along a 2200-m altitudinal gradient in a Mediterranean lizard.

The timing of organisms' senescence is developmentally programmed but also shaped by the interaction between environmental inputs and life-history traits. In ectotherms, ageing dynamics are still poorly understood even though their body temperature, metabolism, or growth trajectory are very sensitive to environmental changes. Here, we investigated the role of life-history traits such as age, sex, body size, body condition, and tail autotomy (i.e self-amputation) in shaping telomere length in six populations of the Algerian sand lizard (Psammodromus algirus) distributed along an elevational gradient from 300 to 2500 m above the sea level. Additionally, we compile the available information on reptiles' telomere length in a review table. Our cross-sectional study shows that older lizards have longer telomeres, which might be mostly linked to the selective disappearance of individuals with shorter telomeres or, alternatively, mediated by a higher expression of telomerase across their life. In fact, variation in telomere length across elevation was explained by age structure of lizards; thus, in contrast to our predictions, altitude had no effect on the telomere length in this study system. Telomere length was unaffected by tail regeneration and was sex-independent, but positively correlated with body condition, which might be linked to high somatic investment. Hence, our results suggest that life-history traits such as age or body condition can be major drivers of telomere dynamics for this species, whereas environmental conditions apparently had scarce or no effects on lizard telomeres. Our findings emphasize the relevance of understanding species' life histories for fully disentangling the causes and consequences of differences in ageing in ectotherms.


Introduction
Environmental conditions can modulate the physiology of individuals and therefore alter their ageing rate (Marasco et al., 2017;Ratikainen and Kokko, 2019). The study of the evolutionary underpinnings of ageing has been a long-standing topic both in ecological and medical research. Although several studies have shown the relation between mitochondrial metabolism and the variation in lifespan across taxa (Selman et al., 2012;Ziegler et al., 2015), the machinery governing ageing remains unclear. Most studies on vertebrate ageing have been conducted in endotherms. In contrast, ectothermic vertebrates have received less attention although their body temperature, metabolism, or growth trajectory are very sensitive to environmental changes, which might alter their ageing rate (Bronikowski, 2008;Olsson et al., 2018;Monaghan et al., 2018). Understanding the link between environmental conditions, life-histories, and senescence in wild ectotherms will increase the current knowledge about their evolutionary and ecological dynamics.
Telomeres are non-coding repeated sequences (TTAGGG n in vertebrates) located at the termini of chromosomes, essential for maintaining genomic stability and for protecting cells from chromosome degradation and fusion (O'sullivan and Karlseder, 2010). Telomeres often shorten after each cell division, which likely explains that these regions tend to be shorter with age in endotherms, as observed in several mammals (e.g. Whittemore et al., 2019) and birds (e.g. Hall et al., 2004; Table 1 Summary of the studies describing the relationship between telomere length (TL) with age and/or other traits in reptiles. Western terrestrial garter snake

Thamnophis elegans
Hatase et al.  (continued on next page) P. Burraco, et al. Comparative Biochemistry and Physiology, Part A 247 (2020) 110741 telomerase can alleviate telomere erosion. Telomerase is known to be active across development in some ectothermic vertebrates (Klapper et al., 1998;Bousman et al., 2003). In this line, signs of telomere elongation have been found throughout larval development of the Atlantic salmon (Salmo salar, McLennan et al., 2016) and of the common water frog (Rana temporaria; Burraco et al., 2019), as well as during the first years of life in some reptiles (e.g. Olsson et al., 2010;Ujvari et al., 2017). Telomere length, at a given ontogenetic point, is not only a function of cell replication but also of the organisms' ability to cope with stress across their life (Dugdale and Richardson, 2018). In vertebrates, harmful conditions often enhance glucocorticoids secretion, which can induce an oxidative state and damage essential biomolecules like lipids, proteins or DNA (Luceri et al., 2018;Florencio et al., 2020), including telomeres (Reichert and Stier, 2017; but see Chatelain et al., 2020). As a consequence of the apparent sensitiveness of telomeric sequences to environmental inputs, telomere length is often used as an indicator of the amount of stress accumulated by an organism across time (Young, 2018). Several studies across taxa have found positive relationships between telomere length and organisms' life expectancy (e.g. Heidinger et al., 2012;Barrett et al., 2013;Wilbourn et al., 2018), reproductive outcome (Eastwood et al., 2019), or immunocompetence (Alder et al., 2018). In ectothermic vertebrates, telomere shortening has been associated with increased growth rate, bold personality, or predator exposure (reviewed in Olsson et al., 2018). In reptiles, a paraphyletic group, several studies have investigated the variation in telomere length with age (Table 1). Five studies found that telomeres shorten with age, although in some cases this relationship was sex-dependent (Table 1). A quadratic sex-dependent relationship between telomere length and age was observed in three studies, i.e. telomeres increase their length until a certain age, and then shorten (Table 1). Meanwhile, three studies found no effect of age on telomere length (Table 1). The high inter-species variation in the relation between telomere length and age highlights the need of further research. Furthermore, several studies have addressed the relationship between telomere length and some life-history traits. For instances, some studies show a positive relationship between telomere length and lifetime reproductive success or body condition. In contrast, no association or unexpected relationships were observed for other traits, or it was species-dependent (Table 1). For instance, one might expect telomeres to shorten as body size increases across lifetime, since it implies a higher number or rate of cellular replications. However, only a few studies on reptiles have observed a significant negative effect of body size or growth rate on telomere length (Table 1), unlike in fish (McLennan et al., 2016) or amphibians (Burraco et al., 2017a). Therefore, so far, we should not generalise when discussing telomere dynamics in reptiles, further studies being needed.
Here, we also aim to understand the role of elevation on telomere length of the Algerian sand lizard (Psammodromus algirus). To this end, we studied a substantial altitudinal gradient from 300 to 2500 above the sea level (m.a.s.l. thereafter) in Sierra Nevada mountain system (Spain). Mountains cover circa a quarter of the Earth's surface and show deep variations across the elevational gradient in biotic and physical conditions, such as predator abundance, temperature, or ultraviolet radiation (Körner et al., 2011). Consequently, physiological adaptations to divergent habitats across altitude have been reported in different taxa (Keller et al., 2013). Such physiological adjustments often imply elevational variation in the relative allocation of energy expenditure to reproduction and somatic maintenance (Bronikowski and Arnold, 1999), which can affect telomere dynamics (Stier et al., 2016). In this altitudinal gradient, as a consequence of environmental temperature decrease with altitude, these lizards reduce their activity while hibernation time increases with ascending elevation (Zamora-Camacho et al., 2013). Hibernation is known to slow down telomere attrition (Hoelzl et al., 2016;Kirby et al., 2019), hence telomere dynamics of lizards at higher altitudes might benefit from longer hibernation. On the other hand, the Algerian sand lizard is heliothermic, meaning that they References are included in the Supplementary Material. P. Burraco, et al. Comparative Biochemistry and Physiology, Part A 247 (2020) 110741 spend a notable amount of time exposed to solar radiation. This lizard devotes more time to basking with increasing elevation (Díaz, 1997), where UV radiation is higher (Reguera et al., 2014a), thereby compensating the dwindling environmental temperature (Zamora-Camacho et al., 2013;Zamora-Camacho et al., 2016). The exposure to UV radiation is known to damage DNA (Cadet et al., 2015), which might negatively affect telomeres in high-elevation lizards. However, lizards at high elevations are darker, which may protect them from UV radiation (Reguera et al., 2014a). Probably as a consequence of reduced activity time, elongated hibernation, and darker colorations, oxidative stress is lower in high-elevation lizards (Reguera et al., 2014b;Reguera et al., 2015). Overall, according to the information gathered on this species along the elevational gradient, we predicted longer telomeres at high elevations.
The main goal of our study was to investigate telomere length across altitude in a lizard species. However, other factors may potentially affect telomere dynamics. Psammodromus algirus lizards do not show a noticeable sexual dimorphism, as both sexes often have similar body sizes, although males can show orange or blue colorations (Carretero, 2002). Therefore, we did not expect differences in telomere length between sexes beyond the putative costs linked to reproductive investment linked to each sex. Importantly, and given that telomere length is often affected by age, we estimated the age of lizards. Regarding the available information on the relationship between telomere length and age in reptiles, one might expect either a positive, negative, or quadratic relationship between both factors (Table 1). Overall, we aimed to increase the available information on telomere length in reptiles, thus helping to gradually fill the gap of knowledge on telomeres in ectothermic vertebrates.

General procedures
The lizard P. algirus is a medium-large lacertid (53-80 mm snoutvent length -SVL-in our study area) that inhabits shrubby habitats in the Mediterranean region from south-western Europe and north-western Africa (Salvador, 2015). In the Sierra Nevada mountain system (SE Spain), we sampled individuals from six populations, at 300 (N = 18), 700 (N = 16), 1200 (N = 18), 1700 (N = 19), 2200 (N = 15), and 2500 (N = 20) m.a.s.l. (Fig. 1). In total, we assessed 106 individuals (50 males and 56 females): 7 in 2010, 28 in 2011, 65 in 2012 and 6 in 2013. Additionally, we estimated telomere length in 37 neonates (see below). Because lizards were part of a long-term study, we marked individuals by toe clipping, a marking method frequently used in lizards, with limited impact on their welfare (Perry et al., 2011). We conserved toe samples in ethanol and used them for age class determination using phalanx skeletochronology (more details below). We collected a portion of the terminal region of lizards' tail (~1 cm) in the field and immersed it in an Eppendorf tube filled with 1.5 mL of absolute ethanol for genetic analyses. Many species of lizards regenerate lost tails, a trait that has evolved with the ecology and with the evolutionary history of lizard lineages (Higham et al., 2013). We took special care to disinfect the wounds caused by both toe clipping and tail sampling with chlorohexidine, closing the wounds with a tissue adhesive glue (Dermabond®).
We measured lizard body mass with a digital balance (Model Radwag WTB200; to the nearest 0.01 g) and SVL with a metal ruler (to the nearest 1 mm). We estimated body condition index (BCI) as the residuals of the regressing log mass on log SVL. This widely used index represents the relative energy reserves of an animal (Schulte-Hostedde et al., 2005). We also recorded whether the tail was intact or regenerated. Males were distinguished from females mainly because they have more femoral pores in their hind limbs and an orange spot in the corners of their mouths (Carretero, 2002). A subset of gravid females (different to the females used for collecting telomere data), recognized by palpation of developing eggs inside the trunk, were transported to a lab and placed in individual terrariums (100 × 20 × 40 cm) with a heat cable at one end of the cage to allow thermoregulation, indirect access to sun light, and water (in form of aqueous nutritious gel) and food (Tenebrio molitor larvae) ad libitum. Substrate was bare soil from the study area. We maintained eggs laid in terrariums until hatching. Then, we took a portion of tail of hatchlings for genetic analyses (see below). In order to avoid pseudoreplication, only one neonate per litter (N = 37) was randomly used for telomere analyses. Females and their neonates were released at the point where the female was caught. No lizard died or suffered permanent pain during the study.

Telomere length measurement
Once in the laboratory, we stored tail samples at −20°C until assayed. We extracted DNA from epidermis using a high-salt DNA extraction protocol (Lahiri and Nurnberger Jr., 1991). This method eliminates the use of toxic reagents such as phenol or chloroform, and yields high amount of good-quality DNA. We used a Nanodrop (Thermo Scientific) to quantify DNA concentration and quality. Since storage conditions, extraction method, or tissue type can affect telomere length measures (Nussey et al., 2014) we used the same conditions for all samples to avoid confounding factors.
We quantified relative telomere length through quantitative polymerase chain reactions (qPCRs), which is one of the most widely used methods for estimating telomere length (Nussey et al., 2014). We compared the cycle threshold (C t ) of telomeric sequences with the C t of a control sequence that is autosomal and non-variable in copy number (Cawthon, 2002;Nussey et al., 2014). We used previously published primer sequences for GAPDH and telomere fragments (Criscuolo et al., 2009). As a reference sequence, we amplified GAPDH sequences using 5´-AACCAGCCAAGTACGATGACAT-3′ (GAPDH-F) and 5′-CCATCAGCA GCAGCCTTCA-3′ (GAPDH-R) as forward and reverse primers, respectively. The use of GAPDH as a single copy gene is widely spread in telomere studies in vertebrates and has been previously used in sand lizards (Pauliny et al., 2018). We confirmed that the among-individual variation was low for this gene (the average Cq-value was 25.37 with a standard error (S.E.) of ± 0.32). For telomere sequences, we used 5′ CGGTTTGTTTGGGTTTGGGTTTGGGTTTGGGTTTGGGTT-3′ (Tel1b) and 5′-GGCTTGCCTTACCCTTACCCTTACCCTTACCCTTACCCT-3′ (Tel2b) as forward and reverse primers, respectively. Conditions of qPCR for GAPDH fragment consisted of 10 min at 95°C and 40 cycles of 10 s at 95°C, 20 s at 58°C, and 1 min at 72°C, and for telomere fragment of 10 min at 95°C, and 10 s at 95°C, 20 s at 58°C, and 1 min at 72°C. We conducted qPCR assays for each gene in separate plates on a LightCycler 480 (Roche) and ran a melting curve from 65 to 95°C, as a final step in each qPCR to check for specific amplicons. Melting curve showed a normal shape, indicating the high specificity of GAPDH and telomere primers (Supplementary Material S1). For each sample, we added 20 ng of genomic DNA and used both sets of primers at a final concentration of 100 nM in a 20 μL master mix containing 10 μL of Brilliant SYBR Green (QPCR Master Mix, Roche). All samples were run in duplicate. Samples with coefficient of variation higher than 5% were measured again. We calculated qPCR-plates efficiency by including five serial diluted standards in triplicate (120, 40, 10, 2.5 and 0.66 ng/uL for GAPDH and telomere sequences), obtained from a golden standard sample containing a pool of three samples from each elevation. We calculated relative telomere length by applying the following formula (Pfaffl, 2001) ]; where E telomere and E GAPDH are the qPCR efficiency of telomere and GAPDH fragment, respectively; ΔCt telomere (controlsample) and ΔCt GAPDH (control-sample) are the deviation of standard -telomere or GAPDH sequences for each sample, respectively. Efficiencies of qPCR were 1.99 ± 0.02 S.E. and 1.93 ± 0.02 S.E. for GAPDH and telomere fragments, respectively. The intra-assay CV% was 4.07% for GAPDH gene and 1.38% for telomere gene. The inter-assay P. Burraco, et al. Comparative Biochemistry and Physiology, Part A 247 (2020) 110741 CV% was 11.26% for relative telomere length. All the R 2 of the standard curves were higher than 0.985.

Estimation of age class with skeletochronology
We estimated individual age class by phalanx skeletochronology (Comas et al., 2019), one of the most accurate techniques to estimate age in many vertebrates, including reptiles (Zhao et al., 2019). Vertebrate ectotherms show indeterminate growth, and consequently present a cyclic growth pattern in hard body structures such as bones, corresponding to alternate periods of growth and resting. This pattern is particularly marked in temperate climates, where age can be fairly estimated by counting annual growth rings in the phalanges (Comas et al., 2016). Growth rings are called lines of arrested growth (LAGs). Toes sampled were decalcified in 3% nitric acid for 3 h and 30 min. Crosssections (10 μm) were prepared using a freezing microtome (CM1850 Leica), stained with Harris hematoxylin for 20 min and dehydrated through an alcohol chain (more details in Comas et al., 2016). Next, cross-sections were fixed with DPX (mounting medium for histology), mounted on slides, and examined for the presence of LAGs using a light microscope (Leitz Dialux20) at magnifications from 50 to 125×. We took several photographs (with a ProgresC3 camera, at the University of Barcelona UB) of various representative cross-sections, discarding those photographs in which cuts were unsuitable for observing LAGs.
The number of LAGs detected in the periosteal bone was independently and blindly counted three times by a single observer (MC) on three independent occasions.

Statistical analysis
In order to meet parametric assumptions, we log-transformed relative telomere length, body mass, and body condition data. We examined the presence of outliers through a Cleveland plot, which revealed that an individual had an extremely abnormal low value (almost zero) of relative telomere length, so we decided to omit this datum from all the analyses. Cleveland plots also showed a possible outlier within the body condition data, thus we followed the recommendations of Quinn and Keough (2002)and, for analyses implying body condition data, we performed the test with and without the datum and reported both statistical results.
Given that not all individuals had associated data for all variables (e.g., neonates always had complete tails and were not sexed) and some variables presented collinearity (e.g. SVL and age), we first ran some analyses in order to test whether variables potentially affecting telomere dynamics covaried with either relative telomere length or altitude. We performed linear models to check for sexual differences in body mass, age or relative telomere length. A chi-squared test was used to test whether the frequency of males and females differed with Fig. 1. Sampling locations in this study across Sierra Nevada mountain altitudinal gradient (SE Spain). Numbers from one to six correspond with each location, i.e. 300, 700, 1200, 1700, 2200, and 2500 m.a.s.l. respectively. P. Burraco, et al. Comparative Biochemistry and Physiology, Part A 247 (2020) 110741 elevation. A linear model was also used to test relative telomere length according to the year of capture in order to evaluate possible cohort effects. Since we sampled lizards with intact tails (n = 44) and regenerated tails (n = 58), and tail regeneration could affect telomere dynamics in tail tissue (Alibardi, 2016), we tested whether there were differences in relative telomere length between lizards with intact or regenerated tail through linear models. A chi-squared test was used to test whether the frequency of individuals with intact or regenerated tails differed with elevation. We also performed linear models to test how relative telomere length varied with age class. In these models, given that there were only five lizards 4 years old and two lizards 5 years old, age was reclassified as neonate, 1, 2, and ≥ 3 years. Moreover, we repeated the analysis without including lizards with 4 and 5 years old. We used Pearson correlations to test for the covariation between relative telomere length and variables such as age, body mass, SVL, and body condition, and for the relationships between age and both body mass or SVL. Using a linear model, we tested for the variation in body condition with altitude.
Finally, we tested the altitudinal variation in relative telomere length in lizards. Given that both altitude and age class were significantly related to relative telomere length, we tested the effect of the two variables as predictors on relative telomere length, as a dependent variable. We also tested for the altitudinal variation in telomere length for neonates, in order to check for the variation in telomere length with altitude at birth. For all the linear models, we confirmed that data met parametric assumptions. All statistical analyses were conducted in Statistica software (version 8.0).

Discussion
Life-history trade-offs and environmental conditions can shape ageing across taxa (Wilbourn et al., 2018;Eastwood et al., 2019). In our study system, we expected to find Algerian sand lizards with longer telomeres at high elevation. This hypothesis was based on the grounds that ectotherms typically live longer at high altitude (Cabezas-Cartes et al., 2018) and that the populations studied herein show reduced time of activity and oxidative stress with elevation (see Introduction, Zamora-Camacho et al., 2013;Reguera et al., 2014b). However, our hypothesis was not supported by the data, as lizards showed an altitudinal pattern of telomere length that simply mirrored the altitudinal distribution in average age. Our cross-sectional study also suggests that, in these lizards, telomeres are longer with age, although we acknowledge the putative role of selective disappearance in explaining this pattern, as discussed below. Likewise, older (and thus larger) lizards  had longer telomeres. Moreover, our findings suggest that differences in telomere length were sex-independent, unlike adult sand lizards of other species (Lacerta agilis, Olsson et al., 2011). Sex differences in telomere length may result from sex differences in growth rate, body size, and/or age . However, in our study system, lizards did not show sexual dimorphism in size or age structure. Tail regeneration did not affect telomere length despite the fact that differences in the regulation of telomere length may be driven by evolutionary pressures such as predation (Olsson et al., 2010), and also by metabolic demands during tissue regeneration. Moreover, no cohort effect was detected, as telomere length did not differ with year of sampling.
Longer telomeres were associated with older age in lizards. This result agrees with previous studies in snakes and lizards Madsen, 2009 andUjvari et al., 2017, respectively). We also found a positive relationship between telomere length and body size. Although telomere length and survival had no association in other lizards such as in the frillneck lizard (Chlamydosaurus kingii, Ujvari et al., 2017), larger body size can include lower mortality risk in ectotherms with indeterminate growth (Angilletta Jr et al., 2004). If extrinsic conditions selectively remove individuals in poor condition -with expected shorter telomeres and likely smaller body size-, then the fact that older lizards have longer telomeres might indicate a prolonged survival of individuals in better condition (Van de Pol and Wright, 2009;Salmón et al., 2017). Despite the putative role of selective disappearance in explaining differences in telomere length, the positive relationship between telomere length and body size suggests that increase in body size -likely involving higher cell replication-does not imply shorter telomeres by itself. Previous studies have showed that ectotherms, unlike endotherms, can show longer telomeres along their lifetime . Such contrasting patterns of telomere dynamics may be related to a higher telomerase expression after birth in somatic cells in ectotherms than in endotherms (Gomes et al., 2010). Hence, telomerase may be relevant for buffering downstream effects of cellular damage in organisms with indeterminate growth such as lizards (Jones et al., 2014). However, telomerase expression may not be enough to protect from telomere shortening in ectothermic vertebrates. For instance, telomerase is expressed in tissues of adult medaka fish (Klapper et al., 1998) but telomeres shorten with age in this species (Hatakeyama et al., 2008). Furthermore, the maintenance of telomerase expression in species with indeterminate growth can imply a trade-off suggested by a higher cancer occurrence in ectotherms (Gomes et al., 2010;Olsson et al., 2018). However, the knowledge about cancer in wildlife is still meagre. In our study system, the use of a longitudinal approach would allow to disentangle the possible role of selective disappearance or telomere elongation (and telomerase activity) in explaining differences in telomere length in older lizards.
We expected longer telomeres in populations at higher elevation. However, elevation did not shape telomeres in these lizard populations since the variation in telomere length was explained by the age structure at each altitude. Contrary to our results, Dupoué et al. (2017) found shorter telomeres and higher extinction risk in low-elevation populations of the common lizard (Zootoca vivipara). In our study system, lowland populations suffer poor habitat quality since they face low thermal quality (risk of overheating, Zamora-Camacho et al., 2016), high ectoparasitism (Álvarez-Ruiz et al., 2018), low food availability (Moreno-Rueda et al., 2017), and high oxidative damage (Reguera et al., 2014b(Reguera et al., , 2015. Additionally, low-elevations lizards increase their activity time while decreasing hibernation time (Zamora-Camacho et al., 2013). Nonetheless, low-elevation populations did not have shorter telomeres than populations at high elevations.
Lizard body condition, environmental temperature, oxidative stress, and telomerase expression might explain the lack of variation in telomere length in lizards inhabiting at different elevations. In this study, body condition of lizards was greater in populations at higher elevation and correlated positively with telomere length. It is known that telomere length can show a positive correlation with body condition in other reptiles (Thamnophis sirtalis; Rollings et al., 2017), suggesting that body condition is an indirect measure of somatic investment. In addition, a temperature-mediated regulation of telomerase expression is likely, thus telomerase might show a higher expression at low elevation, then compensating for telomere erosion Fitzpatrick et al., 2019). At high elevations, the reduction in metabolic rate due to cold conditions may have favoured a reduction in the rate of telomere erosion due to a reduced production of ROS (Reguera et al., 2014b(Reguera et al., , 2015. Indeed, increases in lifespan are often orchestrated by reductions in metabolic rate (Speakman, 2005), as for example show in the snake Thamnophis elegans across an elevational gradient (Bronikowski and Vleck, 2010). Furthermore, the variation in the pace-of-life as a consequence of particular environmental conditions is also known to alter telomeres, thus resulting in complex or unexpected patterns (Giraudeau et al., 2019). In ectotherms, shorter telomeres are associated with higher survival in migratory Atlantic salmon (McLennan et al., 2017), which may indicate a trade-off between investment in migration and investment in telomere maintenance. Likewise, amphibian larvae inhabiting ponds under permanent desiccation risk, showed shorter telomeres (Burraco et al., 2017b). In our system, other factors like diseases or intraspecific interactions might have also modulated ageing in lizards at each elevation. A cross-fostering approach would help to fully clarify the evolutionary impact of both environment and life-history traits on telomeres of this lizard metapopulation.

Conclusions
In contrast to our expectations, altitude had no effect on lizard telomere length. Our results suggest that telomeres are longer with age, and telomere length variation with elevation reflects variation in age along the mountain gradient. Nevertheless, in our cross-sectional study, we cannot disentangle whether this age-dependent variation in telomere length is due to telomere elongation with age or to selective disappearance of low-quality individuals with shorter telomeres. P. Burraco, et al. Comparative Biochemistry and Physiology, Part A 247 (2020) 110741 Likewise, larger lizards (and those with higher body condition) had longer telomeres. This study highlights the relevance of understanding species' life histories and habitat characteristics for disentangling the causes and consequences of differences in ageing in ectotherms.

Declaration of competing interest
No conflict of interest declared.