Telomere attrition with age in a wild amphibian population

Abstract

Telomere shortening with age has been documented in many organisms, but few studies have reported telomere length measurements in amphibians, and no information is available for growth after metamorphosis, nor in wild populations. We provide both cross-sectional and longitudinal evidence of net telomere attrition with age in a wild amphibian population of natterjack toads (Epidalea calamita). Based on age-estimation by skeletochronology and qPCR telomere length measurements in the framework of an individual-based monitoring programme, we confirmed telomere attrition in recaptured males. Our results support that toads experience telomere attrition throughout their ontogeny, and that most attrition occurs during the first 1–2 years. We did not find associations between telomere length and inbreeding or body condition. Our results on telomere length dynamics under natural conditions confirm telomere shortening with age in amphibians and provide quantification of wide telomere length variation within and among age-classes in a wild breeding population.

1. Introduction

Telomere length (TL) variation throughout the ontogeny is driven by the balance between erosion of telomeres over successive cellular divisions and their elongation by telomerase [1–4]. Net telomere shortening with age has been widely documented across taxa [1,5,6], although some studies have failed to find significant correlations between age and TL [5,7]. This suggests the relationship between age and TL is complex, rendering TL a poor indicator of individual age [5]. Furthermore, intrinsic and extrinsic factors causing physiological stress, like enhanced growth rate [8–11], high reproductive investment [12–14], inbreeding depression [15], low resource availability [16] or pathogen infection [17,18] significantly reduce TL [19], with consequences spanning across generations [9,20–25]. The environmental drivers of telomeric dynamics and their fitness consequences are receiving increasing attention in ecological studies [26–28]. However, robust information on natural variation of TL in target organisms is still scarce [29].

Net telomere shortening with age has been demonstrated principally in mammals and birds [30–34], but also in reptiles [35,36], fishes [37,38], insects [39], crustaceans [40] and platyhelminthes [41]. In contrast, very few studies have reported TL measurements in amphibians, which constitute a singular group of terrestrial vertebrates owing to their biphasic life cycle. This particularity renders them an interesting model to test intrinsic and environmental drivers of TL dynamics, owing to the possibility of experimentally decoupling individual growth and development [10,42]. Information about amphibian TL dynamics is specially lacking for post-metamorphic (juvenile and adult) growth or in wild populations [10,42–45] owing to two main difficulties: assessing the age of adult amphibians in natural populations and tracking them in their habitat over their lifespan. For this reason, individual-based monitoring programmes like those built on capture–mark–recapture data provide excellent opportunities to fill these knowledge gaps [31,34,46,47].

Here we aim to characterize TL dynamics in a wild amphibian population by quantifying (i) TL variation across age-classes using a cross-sectional approach and (ii) telomere shortening by repeated TL measures (longitudinal approach). We also test the relationship of both TL and attrition with two individual traits potentially associated to fitness: body condition and inbreeding. Our dataset combines skeletochronological and DNA samples with morphometric measures obtained in the framework of an individual-based demographic monitoring programme. We expected net telomere shortening in our longitudinal data, and a concomitant negative relationship between TL and age in our cross-sectional assessment. We also hypothesized that individuals showing higher body condition and/or lower inbreeding might present longer telomeres and lower attrition rates.

2. Material and methods

(a). Study species and field methods

This work is part of a demographic monitoring programme conducted continuously since 2010 on a wild amphibian community in Central Spain [48–50]. We focused on the natterjack toad Epidalea calamita (Laurenti, 1768) [51]. Activity patterns in E. calamita show wide interannual fluctuations associated with temperature and rainfall, so we took advantage of the high attendance of individuals to the breeding area in 2015 to study TL patterns and their association with age, body condition and inbreeding in a representative sample from the population. We additionally obtained a second sample from 33 individuals recaptured again during the 2018 reproductive season, to characterize telomere attrition dynamics under natural conditions.

Field sampling consisted of nocturnal surveys, mostly during the breeding season of E. calamita, actively searching for adult individuals. We recorded sex, body length (snout-to-vent length) and body mass of all captured individuals. We also marked all toads with passive integrated transponder (PIT) tags, conferring an individual alphanumeric code. A tissue sample consisting of a piece of the 4th toe (containing three phalanges, muscle, skin and cartilage) of the right hind limb of all individuals was clipped at the time of first capture for its use in skeletochronological and TL analyses. In 2018, we also collected tissue samples of the 4th toe of the left hind limb of 33 individuals (32 males, one female) that had previously been recorded at the breeding area in 2015, for telomere attrition assessments.

(b). Age estimation by skeletochronology

We estimated the age of 211 individuals (143 males, 68 females) captured during the 2015 breeding season (although some of them had been marked in previous years, so tissue samples were collected between 2010 and 2015) by histological inspection of the number of hematoxylinophilic lines of arrested growth (LAGs) present in their toe phalanxes [52–54] (see detailed protocol in electronic supplementary material, appendix 1).

(c). DNA extraction and inbreeding estimation

We isolated DNA from the tip of 244 toe samples (211 individuals of the skeletochronological analysis plus 33 individuals recaptured in 2018) using a magnetic beads-based protocol in a robotic Freedom EVO platform (Tecan). All individuals were genotyped using 16 microsatellite loci. Multilocus genotypes were published in a related paper [48] and used to calculate the inbreeding coefficient F for each individual using the ‘inbreeding’ function in R package adegenet [55,56].

(d). Telomere length estimation

We assessed the relative length of telomeres (T/S ratio) for all individuals at the moment of tissue sampling (between 2010 and 2015 for first captures and 2018 for recaptures) using qPCR (electronic supplementary material, appendix 2). Since samples were obtained from clipped toes, these included mostly muscle and skin cells, but also cartilage and bone.

(e). Growth models and associations between telomere length, telomere attrition and individual traits

We used body length measures of individuals at the moment of first capture along with their ages estimated by skeletochronology to parametrize von Bertalanffy growth functions for males and females using R package FSA [57]. The relationships between TL, telomere attrition and individual variables were tested by fitting general linear models for males and females separately (electronic supplementary material, appendix 3).

3. Results

The ages of breeding E. calamita males in 2015 ranged between 1 and 9 years (mode = 3), and in females, between 2 and 7 years (mode = 4, electronic supplementary material, figure S2). There were no significant differences in age distribution across sexes (Wilcoxon test W = 4735.5, p = 0.515). TL ranged between 0.405 and 3.103 (median = 0.998) in males and between 0.364 and 2.636 (median = 0.896) in females. There were no significant differences in TL across sexes (W = 5634, p = 0.094).

(a). Growth models and association between telomere length, telomere attrition and individual traits

The von Bertalanffy growth function in males approached a mean maximum body length of 77.6 mm with a rate constant of 1.56 (p < 0.001). Females showed a higher mean maximum body length (84.8 mm) but a lower rate constant (0.76, p = 0.004, figure 1). TL significantly decreased with age in females ( β = −0.137, p = 0.010) but not in males (table 1). TL did not show significant associations with body condition or average inbreeding coefficient (table 1, figure 1).

Figure 1.

(a) Growth functions and simple linear regression models for TL versus (b) age, (c) body condition and (d) inbreeding in males (black dots, solid lines) and females (white dots, dashed lines). The p-value is indicated for the single significant TL model.

Table 1.

Linear regression models for TL and attrition for each sex. Intercept and coefficient, standard error (s.e.) and p-value of the slope of the corresponding explanatory variable (exp. var.) and the adjusted R² are shown for each model. Sqrt D: square-root transformation of D-adjusted telomere attrition. Significant p-values (less than 0.05) in italics.

response	exp. var.	sex	intercept	slope	s.e.	p-value	R2
telomere length	age	males	0.085	−0.020	0.036	0.581	<0.01
		females	0.367	−0.137	0.051	0.010	0.09
	body condition	males	0.331	−0.716	0.487	0.143	0.01
		females	0.231	−0.649	0.542	0.235	0.01
	inbreeding	males	−0.013	0.261	0.418	0.534	<0.01
		females	−0.329	1.352	0.762	0.080	0.03
Sqrt D	initial TL	males	0.529	−0.021	0.044	0.644	<0.01
telomere attrition	age	males	2.118	−0.464	0.151	0.006	0.28
	body condition	males	1.034	−0.654	1.643	0.695	<0.01
	inbreeding	males	0.649	0.530	0.961	0.587	<0.01

From the 33 E. calamita individuals with two time-separated TL measures (32 males and one female), 27 showed TL reductions, whereas the remaining six showed slight TL increases (figure 2). Telomere attrition between 2015 and 2018 for E. calamita male breeders depended on the age of individuals, with younger males showing higher telomere attrition (β = −0.464, p = 0.006, table 1), but did not show any significant relationship with TL, average inbreeding or body condition in 2015 (table 1, figure 2).

Figure 2.

(a) TL measured at the dates of first (black dots) and second capture (grey dots). (b) Observed TL attrition in the same individuals as in (a): 32 males (dark dots) and one female (white dot). The horizontal line delimits telomere shortening (positive attrition) versus elongation (negative attrition). Linear regression models for telomere attrition versus (c) initial TL, (d) age and (e) body condition in 2015 and (f) inbreeding. The p-value is indicated for the single significant TL attrition model.

4. Discussion

Our results provide robust evidence of net shortening of telomeres with age in E. calamita. Longitudinal results revealed telomere shortening in most E. calamita males, although shortening rates were variable, with individuals first-captured in 2015 showing higher rates than individuals first-captured in previous years, which could be indicative of cohort-dependent attrition rates. Furthermore, six males showed net TL increases. Telomere elongation has been documented in vertebrates, suggesting telomere repair could increase the fitness of some aged individuals [31,42,58] (although measurement errors have also been invoked to explain TL increases, [59,60]). The different telomere attrition rates observed in our data, together with wide TL variation within age-classes in our cross-sectional results, suggest that a substantial portion of TL variability in E. calamita cannot be solely explained by age-dependent causes.

Nevertheless, our longitudinal and cross-sectional results suggest that most toads experience a net reduction in TL over time, and that telomere attrition rate decreases with age. Both facts are supported by significant negative associations between TL and age (although only in females) and between telomere attrition and age, although the proportion of variance explained was relatively low in the first case (9% and 28%, respectively). Growth and metabolism early in an individual's life are well-documented determinants of telomere dynamics, which in turn condition individual fitness, affecting longevity and lifetime reproductive success [34,61–65]. Our results support the importance of the first years of post-metamorphic life in the TL balance of wild amphibians.

Two main endogenous processes prevail in the years following metamorphosis in amphibians: somatic growth and sexual maturation [66]. Growth rate and the energetic demand of both processes likely drive telomere attrition in amphibians, at least after metamorphosis. Our results show that females attain larger sizes than males across most age-classes. Enhanced growth, rather than enhanced metabolism, results in greater telomere attrition in amphibians [10], and it is therefore not surprising that the relationship between age and TL is stronger in females than in males. Our TL measurements suggest that the physiological demand causing telomere attrition also decreases after the juvenile growing period, and thus the first 1–2 years after metamorphosis probably set the baseline TL for adult life in temperate amphibians.

However, 1- and 2-year-old individuals in our sample did not show disproportionately longer telomeres than the remaining age-classes, nor did we observe a clear association between TL and age in our male cross-sectional data. Both results are probably caused by the biased representation of the first two age-classes in our samples. For ethical reasons, we only PIT-tagged individuals above a certain body size, mainly focusing on the adult, potentially breeding portion of the population. Although it is possible that our 1- and 2-year age-class samples could include some immatures encountered near the breeding area, most of them were reproductive individuals. They could be early maturing individuals, while many other toads of these age-classes delayed maturity to older ages, mostly 3 years as suggested by the modal age-class. The first two age-classes are thus not comprehensively represented with respect to their actual frequencies in the population. Moreover, telomere measures from our 1- and 2-year samples are probably biased representatives of their age-classes, because early maturing individuals might show shorter telomeres among their cohort [67]. We predict that representative sampling across all male age-classes will confirm the negative association between TL and age as observed in females and also suggested by longitudinal data, and that individuals of both sexes presenting slow growth and delayed maturity probably have significantly longer telomeres, especially during their first year.

Environmental effects, particularly during the first 1–2 years, may also cause significant telomere attrition during the post-metamorphic life of amphibians. Arguably, hibernation and aestivation periods are the main climatic challenges faced by E. calamita in Mediterranean regions [68,69], characterized by important interannual variability. Consequently, individuals of the same age-cohort should experience a similar effect in their TL, because of their shared environmental background. The mixture of data from different cohorts probably obscures the association between TL and age in cross-sectional datasets. Some studies have reported a negative effect of inbreeding on telomeres [15], but we did not find any association between TL or telomere attrition and inbreeding rate in E. calamita, nor with body condition. Nevertheless, individual traits involved in fitness and breeding performance of amphibians are expected to have indirect consequences on their telomeric dynamics, as age does.

Selective mortality can blur the association between TL and age in cross-sectional studies [70–72], and may explain the lack of association between telomere attrition and initial TL in our longitudinal data. Similarly, selective mortality of poor-conditioned or inbred individuals could explain the lack of association between TL and body condition and inbreeding in our dataset, although we did not find a significant relationship between the probability of return to the breeding site in 2018 and TL, age, body condition or inbreeding of individuals sampled in 2015 (electronic supplementary material, figure S3). Our assessment of heterozygosity with microsatellite markers is a partial descriptor of genome-wide heterozygosity; future assessments using large numbers of single nucleotide polymorphisms (SNPs) may illuminate this question.

We present evidence of telomere attrition with age in a long-lived, seasonal-breeding anuran, and a quantitative report of TL variation in a wild population. Further research exploring different individual traits potentially involved in maintaining long telomeres in amphibians, their interaction with environmental factors, and the usefulness of TL as a predictor of individual lifespan and growth trajectory will provide valuable insights on telomere dynamics in vertebrates.

Ethics

All procedures were approved by the Ethics Committee of Consejo Superior de Investigaciones Científicas, Spain (ref.: 710/2018) and Comunidad de Madrid (refs.: PROEX 040/19, 10/069513.9/18).

Data accessibility

Raw data analysed in this study are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.37pvmcvgs [73].

Authors' contributions

G.S.-M. participated in the design of the study, carried out the field sampling, skeletochronological and molecular laboratory work, performed data analysis and drafted the manuscript; I.M.-S. conceived, designed and coordinated the study and collected field data; C.D.-P. participated in skeletochronological assessments and interpretations; A.V. collected field data; A.H.A. coordinated skeletochronological laboratory work and revised statistical analyses; I.G.-M. conceived and designed the study and coordinated telomere length assessments. All authors critically revised the manuscript, gave final approval for publication and agree to be held accountable for the work performed therein.

Competing interests

We declare we have no competing interests.

Funding

This work was supported by FEDER/Ministerio de Ciencia, Innovación y Universidades, Spain (grant nos CGL2017-83131-P, CGL2017-83407-P) and Asociación de Amigos de la Universidad de Navarra (predoctoral grant to G.S.-M.).