Proceedings of the Royal Society B: Biological Sciences
You have accessResearch articles

Oxidative costs of cooperation in cooperatively breeding Damaraland mole-rats

Rute Mendonça

Rute Mendonça

Mammal Research Institute, Department of Zoology and Entomology, University of Pretoria, Pretoria, South Africa

Laboratory of Evolutionary Ecophysiology, Institute of Biology, University of Neuchâtel, Neuchâtel, Switzerland

Department of Biological and Environmental Sciences, University of Gothenburg, Gothenburg, Sweden

[email protected]

Google Scholar

Find this author on PubMed

, ,
Nathan Katlein

Nathan Katlein

Mammal Research Institute, Department of Zoology and Entomology, University of Pretoria, Pretoria, South Africa

Laboratory of Evolutionary Ecophysiology, Institute of Biology, University of Neuchâtel, Neuchâtel, Switzerland

Kalahari Meerkat Project, Kuruman River Reserve, Northern Cape, South Africa

Google Scholar

Find this author on PubMed

,
Armelle Vallat

Armelle Vallat

Neuchâtel Platform of Analytical Chemistry, Faculty of Sciences, University of Neuchâtel, Neuchâtel, Switzerland

Google Scholar

Find this author on PubMed

,
Gaétan Glauser

Gaétan Glauser

Neuchâtel Platform of Analytical Chemistry, Faculty of Sciences, University of Neuchâtel, Neuchâtel, Switzerland

Google Scholar

Find this author on PubMed

,
Nigel C. Bennett

Nigel C. Bennett

Mammal Research Institute, Department of Zoology and Entomology, University of Pretoria, Pretoria, South Africa

Google Scholar

Find this author on PubMed

and
Fabrice Helfenstein

Fabrice Helfenstein

Laboratory of Evolutionary Ecophysiology, Institute of Biology, University of Neuchâtel, Neuchâtel, Switzerland

Google Scholar

Find this author on PubMed

Published:https://doi.org/10.1098/rspb.2020.1023

Abstract

Within cooperatively breeding societies, individuals adjust cooperative contributions to maximize indirect fitness and minimize direct fitness costs. Yet, little is known about the physiological costs of cooperation, which may be detrimental to direct fitness. Oxidative stress, the imbalance between reactive oxygen species (by-products of energy production) and antioxidant protection, may represent such a cost when cooperative behaviours are energetically demanding. Oxidative stress can lead to the accumulation of cellular damage, compromising survival and reproduction, thus mediating the trade-off between these competing life-history traits. Here, we experimentally increased energetically demanding cooperative contributions in captive Damaraland mole-rats (Fukomys damarensis). We quantified oxidative stress-related effects of increased cooperation on somatic and germline tissues, and the trade-off between them. Increased cooperative contributions induced oxidative stress in females and males, without increasing somatic damage. Males accumulated oxidative damage in their germline despite an increase in antioxidant defences. Finally, oxidative damage accumulation became biased towards the germline, while antioxidant protection remained biased towards the soma, suggesting that males favour the maintenance of somatic tissues (i.e. survival over reproduction). Our results show that heightened cooperative contributions can ultimately affect direct fitness through oxidative stress costs, which may represent a key selective pressure for the evolution of cooperation.

1. Introduction

Individual variation in cooperative behaviour is inherently associated with variation in benefits and costs accrued. In many cooperatively breeding species, where helpers postpone reproduction and assist the breeders in raising their offspring and/or maintaining their territory [1,2], the benefits of cooperative behaviours have been intensely studied [310]. Additionally, several types of cooperative behaviours have been shown to be energetically demanding [1117] and are thus expected to generate costs. Indeed, higher cooperative contributions are linked to a higher metabolic rate [16,17] and a lower growth rate [3,15], weight gain [12,13,18] or foraging time [1922]. However, it remains largely unexplored whether cooperative behaviours carry important physiological costs capable of affecting Darwinian fitness.

Oxidative stress may represent a physiological cost of cooperative behaviours, which can ultimately decrease an individual's fitness. Oxidative stress is an imbalance between reactive oxygen species (ROS; inevitable by-products of energy production) and antioxidant protection, in favour of the former [23]. Under oxidative stress, ROS are more likely to oxidize important biomolecules and accumulation of oxidative damage may disrupt cell and tissue functioning, compromising survival and/or reproduction [2427]. Cooperative contributions are supported by appropriate tuning of metabolism and energy production [17], and may induce oxidative costs due to such energetic demands [24,2830] (but see [31]). In cooperatively breeding birds, recent correlative work suggests that social status is a good predictor of oxidative stress at the end of the breeding season, when cooperation is intense [3234]. However, whether this is the consequence of differences in social status or of their associated differences in energetically demanding cooperative contributions remains unclear.

Upon increased energetic demands, individuals may face a trade-off between protecting somatic and germline tissues against damaging effects of oxidative stress, reflecting the trade-off between survival and reproduction [3436]. Notably, the germline and particularly sperm cells are highly sensitive to ROS due to a high proportion of polyunsaturated fatty acids in their membranes and vulnerable DNA [37,38]. Hence, competing demands between somatic and germline functions in terms of protection against ROS may represent a physiological mechanism underlying the survival/reproduction trade-off [39]. Optimal investment in either function may be plastic and depend on reproductive prospects, such that individuals with low/uncertain opportunities may benefit from increased investment in somatic functions, preserving survival and thus future reproduction [40,41]. Strikingly, whether changes in cooperative contributions affect the trade-off between somatic and germline functions remains untested.

Here, we experimentally tested whether increases in energetically demanding cooperative contributions (i) affect oxidative balance, (ii) affect the trade-off between somatic and germline tissues, and (iii) are dependent on initial oxidative status in captive Damaraland mole-rats (Fukomys damarensis). Damaraland mole-rats are cooperative breeders with only one female and up to three males monopolizing reproduction [3537], whereas the breeders' offspring (so-called helpers) are unlikely to breed while staying in their natal colony [35,3739], but do produce functional sperm [40]. This subterranean rodent dwells in niches where burrowing behaviours—to maintain and expand the tunnel system in search for food, which is then shared by all colony members [41]—are the most pronounced and energetically demanding form of cooperation [17]. All members of the colony contribute to burrowing activities, and variation in cooperative contributions, within and between individuals, has been previously reported [37,42,43].

Oxidative balance was assessed in a somatic tissue in both males and females, and in the germline of the males. To account for the complexity of the oxidative balance system [44], we described oxidative status through: (i) the activity of the antioxidant superoxide dismutase (SOD), a key enzymatic antioxidant in the first line of defence against ROS [45]; (ii) the ratio between oxidized and reduced glutathione (GSSG/GSH) which provides a measure of cellular oxidation and ROS formation [46], highlighting oxidative threats that disturb the oxidative balance [47] (in harmony with SOD, GSH reduces ROS, producing GSSG [45]); and (iii) the concentration of malondialdehyde (MDA), a marker of lipid peroxidation (i.e. oxidative damage) [45].

Based on the hypothesis that the expression of energetically demanding cooperative behaviours stimulates ROS formation, we predicted that experimentally increasing cooperative contributions would lead to increased antioxidant protection (SOD) and cellular oxidation (GSSG/GSH). We expected stable or increased levels of oxidative damage (MDA) to be related to a good or poor antioxidant protection, respectively. Finally, we predicted that upon an oxidative threat, helpers would bias their investment towards somatic tissues and thus favour survival, while breeders would largely invest in their germline, favouring reproduction.

2. Methods

(a) Study animals and husbandry

We used six wild-caught colonies of Damaraland mole-rats (79 individuals: 36 females, 43 males; mean group size = 13.2 individuals), maintained in captivity for at least 2 years at the Kuruman River Reserve, Northern Cape, South Africa. Breeding status (breeder/helper) of males was determined through genetic paternity analysis of the offspring born in captivity as late as three months after the end of the experiment, following [48]. Paternity analysis revealed the presence of one breeding male in one colony, two breeding males in four colonies and three breeding males in one colony used in this study. Female breeders were identified by the presence of elongated nipples and the obvious pregnancies and parturitions observed in each colony. Only one female breeder was present in each of the experimental colonies. To allow individual identification, animals carried subcutaneous RFID microchips and unique dye marks.

Each colony was individually housed in standardized artificial tunnel systems made of modified PVC pipes covered with a transparent PET sheet allowing behavioural observations. Each tunnel system contained a nest-box, food storage areas and a large plastic waste box, for animals to discard unwanted material. Three vertical pipes, hereafter referred to as sand dispensers, were placed close to the extremities of the tunnel systems to provide sand to the colonies (fig. S2 in [43]).

Every morning, tunnel systems were cleaned, and nest material (small pieces of paper towel) was provided away from the nest. Twice daily, animals were fed ad libitum with large non-transportable and small transportable items of sweet potatoes and cucumbers. Food items were provided behind the sand dispensers, which were filled on the same occasion, to encourage burrowing activities.

(b) Experimental design

Sand provisioning was manipulated to experimentally increase burrowing contributions. Each colony was subjected to a control and a sand treatment, each lasting 8 days and separated by an 8-day resting period (electronic supplementary material, figure S1). The treatment order was balanced across colonies. Daily, sand dispensers were filled twice (at 7.00 and 19.00) during the control treatment and 12 times (i.e. every hour from 7.00 to 19.00) during the sand treatment, resulting in the daily provisioning of 16 and 92 kg (range: 70.4–112 kg) of sand, respectively. Although these amounts of sand do not necessarily reflect the amounts burrowed by wild colonies, they illustrate Damaraland mole-rats’ ability to adjust cooperative contributions according to the existing needs. To control for possible hourly disturbances during the sand treatment, the sand dispensers were touched every hour during the control treatment.

(c) Behavioural observations

Behavioural observations were carried out on days 2 and 7 of each treatment (electronic supplementary material, figure S1). A scan protocol [49] was used to record the behaviour of each colony member every 4 min from a list of pre-defined behaviours (electronic supplementary material, table S1). Scan observations lasted 12 h, resulting in the collection of 180 behavioural data points for each individual. Observations were shared equally between two observers with shifts of 2–4 h. Five behavioural measures were considered for our analyses: active, non-cooperation, burrow, food carry and nest build. Active referred to any behaviour displayed while outside of the nest; non-cooperation grouped all behaviours that are not of cooperative nature (see below); burrow included all behaviours related to tunnel excavation; food carry referred to transportation of food pieces; nest building included preparation and transportation of nest material (electronic supplementary material, table S1). The latter three are cooperative activities [43].

(d) Sampling procedures

Blood and ejaculates were collected from subjects anaesthetised with isoflurane, 2 days before the start and on the last day of each treatment (day 8) (electronic supplementary material, figure S1). Blood was collected from all individuals older than six months, while ejaculates were collected from males older than eight months. A vein from a pre-warmed foot was pierced using a sterile 16G needle, and blood was collected into lithium-heparinized tubes and centrifuged immediately for 5 min at 2000g. Plasma was separated from the cell pellet and both fractions were kept at −80°C until laboratory analysis. After blood sampling, ejaculates were collected by electro-ejaculation (electronic supplementary material, methods), following minor adjustments to the procedure described by Fasel et al. [50]. Ejaculate cell density was assessed when ejaculate volume was greater than that needed for oxidative marker quantification. Ejaculates were loaded into a counting-chamber under a microscope, and cell density was calculated using cell count adjusted for the area examined, the volume of the chamber used and the final dilution. Subjects were then removed from anaesthesia, weighed and placed back in their original tunnel system, once full mobility was regained.

(e) Quantification of oxidative stress markers

SOD was quantified in erythrocyte and ejaculate homogenates; MDA in erythrocytes, ejaculate and plasma; GSH and GSSG in erythrocytes. SOD activity was determined, using a commercial kit (Cayman Chemical), following the supplier's instructions and adjusting the samples' final dilution (electronic supplementary material, methods). Concentrations of GSH and GSSG were determined following minor adjustments to the description of Mora et al. [51] (electronic supplementary material, methods). Total MDA was quantified following Mendonça et al. [52], applying minor modifications (electronic supplementary material, methods).

(f) Statistical analysis

All analyses and graphical representations were done using the software R (v. 3.4.0) [53]. Whenever possible, (generalized) linear mixed-effects models ((G)LMMs) were used, with the package lme4 [54]. When measurements of colony and/or individual were repeated, these were specified as random effects. An observation level random effect (OLRE) was added in the presence of overdispersion of GLMM residuals [55]. Covariates specified in interaction terms were scaled to allow the interpretation of their main effects [56]. Full models were used for interpretation. p-values for fixed effects were extracted from ANOVA tables using Wald F tests with the Kenward–Roger approximation of degrees of freedom from the package car [57]. Statistical significance was set at p < 0.05.

(i) Effect of treatment on behaviour and body condition

To confirm that the sand treatment specifically increased energetically demanding cooperative contributions, the proportions of active, non-cooperation and burrow were used as response variables in distinct binomial GLMMs (n = 77 after exclusion of two pups younger than three months). Rare behaviours (nest build and food carry) were analysed using Wilcoxon signed-rank tests with continuity corrections. All GLMMs included treatment, sex, breeding status and all possible interactions as predictors to assess possible differences between breeders and helpers of both sexes. Body mass, day of treatment and their interaction with treatment were also included.

To determine the effect of treatment on body condition, the change in body mass throughout each treatment was used as a response variable in an LMM including treatment, sex, breeding status and all possible interactions as predictors. To control for regression to the mean effects [58], body mass at the beginning of the week was included as a covariate.

(ii) Effect of treatment on oxidative stress in blood

The final measure (day 8) of each somatic oxidative marker (SOD, GSSG/GSH, GSSG, GSH and MDA) was ln-transformed and set as a response variable in distinct LMMs including treatment, sex, breeding status and all possible interactions, and body mass and its interaction with treatment as well as the interaction between treatment and the initial (day −2) measure of the respective marker were specified as predictors.

We determined whether individuals going through the treatment sequence sand–control (electronic supplementary material, figure S1b) experienced carry-over effects (i.e. the 8-day resting period was not sufficient for oxidative markers to return to a baseline level similar to that encountered before the sand treatment). The initial (day −2) measure of each somatic oxidative marker for these individuals was ln-transformed and set as a response variable in distinct LMMs with treatment as a predictor.

(iii) Effect of treatment on oxidative stress in ejaculates

The final (day 8) measure of each germline oxidative marker (SOD and MDA) was ln-transformed and set as a response variable in distinct LMMs. Due to smaller sample sizes, only body mass and the interaction between treatment and breeding status were specified as predictors. Due to a reduced number of ejaculates collected before the start of each treatment, initial measures were not included. An additional LMM investigated whether final ejaculate cell density was affected by treatment, which could be the underlying mechanism for the observed effects on oxidative markers.

(iv) Effect of treatment on the trade-off between somatic and germline tissues

The ratio between ejaculate and erythrocyte final (day 8) levels for each marker was used as a response variable in two distinct LMMs. Positive ratios indicate biased antioxidant protection (SOD) or damage accumulation (MDA) towards the germline, while negative ratios indicate a bias towards the soma. The model structure followed the one described in the section ‘Effect of treatment on oxidative stress in ejaculates’.

(v) Relationship between initial oxidative stress and changes in cooperative contributions

We tested whether individuals better equipped to cope with an oxidative challenge, either due to higher initial antioxidant levels or lower initial oxidative damage, could raise cooperative contributions to a greater extent. The difference in the percentage of burrow between the two treatments (sand–control) was set as a response variable in six LMMs (one for each marker), including the initial measure of the respective marker (day −2 of the sand treatment), and its interaction with sex and breeding status as well as initial body mass as predictors. To control for regression to the mean effects, the percentage of burrow during the control treatment was included as a covariate [58].

3. Results

(a) Effect of treatment on behaviour and body condition

Animals were significantly more active when more sand was provided to the colony, compared with the control treatment (electronic supplementary material, table S2a). Such increase in individual activity was specifically driven by the significant increase in burrowing behaviours: the proportion of burrowing doubled during the sand treatment (electronic supplementary material, table S2b; figure 1), while other cooperative and non-cooperative activities remained unchanged (nest build females: V = 145, p = 0.6; males: V = 142.5, p = 0.6; food carry females: V = 117, p = 1; electronic supplementary material, figure S2a,b) or even decreased (food carry males: V = 382.5, p = 0.008, electronic supplementary material, figure S2a; non-cooperation: electronic supplementary material, table S2c and figure S2c). The increase in individual burrowing activities, caused by increased sand provisioning, did not drive individuals to exhaustion as the expression of these behaviours throughout the week did not differ between treatments (treatment × day: all models p > 0.1).

Figure 1.

Figure 1. Effect of treatment on the proportion of time spent burrowing in two sessions of 12 h scan observations. Boxplots show median, first and third quartiles and whiskers extend no further than 1.5× interquartile range (IQR). Each individual is represented twice in each treatment, once from observation on day 2 and once from observation on day 7. Lines between points illustrate the repeated measurements from same individuals. ***p < 0.001.

Weekly changes in body mass did not differ between treatments (electronic supplementary material, table S3).

(b) Effect of treatment on oxidative stress in blood

Cellular oxidation (GSSG/GSH) in erythrocytes was elevated upon increased burrowing activities (electronic supplementary material, table S4a; figure 2a). This effect was driven by a greater rise in oxidized (GSSG) than in reduced (GSH) glutathione (45.5% and 14.8%, respectively; t = −2.7, p = 0.008; electronic supplementary material, table S4b,c; figure 2b,c). Females, but not males, showed a decrease in erythrocyte lipid damage (MDA) during the sand treatment (electronic supplementary material, table S5a and figure S3; least-squares means comparisons with Tukey's adjustment; females: t = 2.8, p = 0.037; males: t = −0.7, p = 0.9). Increased burrowing activities did not affect erythrocyte antioxidant SOD activity (electronic supplementary material, table S5b) nor plasma MDA levels (electronic supplementary material, table S5c).

Figure 2.

Figure 2. Effect of treatment on final (day 8) erythrocyte (a) cellular oxidation (GSSG/GSH), (b) oxidized glutathione (GSSG) and (c) reduced glutathione (GSH). Boxplots show median, first and third quartiles, and whiskers extend no further than 1.5 × IQR. Points represent individual measurements, and lines between points illustrate the repeated measurements from same individuals. *p < 0.05; **p < 0.01; ***p < 0.001.

The initial levels of all oxidative stress markers were positively correlated with their final levels. Other variables of interest such as sex, breeding status and body mass appeared to have little effect on oxidative stress markers (electronic supplementary material, tables S4 and S5).

The initial (day −2) oxidative marker levels did not differ between treatments in individuals that started the experiment with the sand treatment (all p > 0.3), showing that the 8-day resting period was sufficient to avoid potential carry-over effects.

(c) Effect of treatment on oxidative stress in ejaculates

Independently of breeding status, SOD and MDA in ejaculates were elevated upon increased burrowing activities (electronic supplementary material, table S6; figure 3a,b). These effects were not driven by changes in sperm density as this did not differ between treatments (F1,14 = 2.3, p = 0.2).

Figure 3.

Figure 3. Effect of treatment on final (day 8) (a) ejaculate SOD antioxidant activity, (b) ejaculate oxidative damage to lipids (MDA) and on the trade-off between germline and somatic tissues, expressed by the ratio ejaculate/erythrocyte levels, of (c) SOD and (d) MDA. Boxplots show median, first and third quartiles, and whiskers extend no further than 1.5 × IQR. y-axes of plots (c) and (d) are logarithm transformations, where negative values indicate greater levels in the soma, and positive values indicate greater levels in the germline. Points represent individual measurements, and lines between points illustrate the repeated measurements from same individuals *p < 0.05; **p < 0.01.

(d) Effect of treatment on the trade-off between somatic and germline tissues

Upon increased burrowing activities and independently of breeding status, SOD and MDA were both elevated in the germline to a greater degree than in the soma, resulting in a higher ratio germline/soma during the sand treatment (electronic supplementary material, table S7; figure 3c,d). SOD activity was biased towards somatic tissues in both treatments (negative ratio), while MDA levels were relatively balanced during the control treatment and became biased towards the germline during the sand treatment (positive ratio).

(e) Relationship between initial oxidative stress and changes in cooperative contributions

The initial levels of oxidative stress markers did not significantly predict changes in the proportion of time spent burrowing between the two treatments (electronic supplementary material, tables S8 and S9). An initial model revealed a trend for the interaction between initial plasmatic MDA and sex (p = 0.053), where males showed a negative (and females a positive) relationship between initial plasmatic MDA and changes in burrowing activities. Visual inspection of this interaction suggested that one single male may be responsible for this result, and its removal from the model rendered the interaction non-significant (electronic supplementary material, table S9c).

4. Discussion

Our results offer the first experimental demonstration of an oxidative cost of energetically demanding cooperative behaviours. Increased energetically demanding cooperative contributions generated oxidative stress in somatic and germline tissues, as witnessed by elevated cellular oxidation and oxidative damage, respectively. Upon increased cooperative contributions, Damaraland mole-rat males favoured somatic over germline functions, revealing the greater cost of cooperation and illustrating a trade-off prioritizing survival over reproduction. The specific effect of our sand manipulation on burrowing behaviours makes us confident that the oxidative costs observed were driven by the increased energetically demanding cooperative contributions. The lack of an effect of increased cooperative contributions on body condition, in a captive setting, highlights the possibility that the oxidative costs of cooperation in the wild may be exacerbated due to limited resources and harsher environment conditions [5961]. Together, our results provide evidence on a physiological mechanism underlying the greater cost of cooperative behaviours—that induced to reproduction through oxidative stress.

Increased cooperative contributions, and their associated energetic demands, induce an oxidative threat to somatic tissues, as witnessed by the increased levels of cellular oxidation (GSSG/GSH) in erythrocytes [47,62]. This effect was driven by increased oxidized glutathione (GSSG), supporting the view of reduced glutathione's (GSH) important role in neutralizing ROS [24]. Despite the extensive oxidation of GSH (into GSSG), its concentration was also increased during the sand treatment, probably through de novo synthesis [45], suggesting an investment in somatic antioxidant protection upon heightened cooperative contributions.

Individuals were able to deal with the oxidative threat imposed to the soma by increased cooperative contributions, avoiding the accumulation of somatic oxidative damage. Firstly, increased cooperative contributions did not affect antioxidant SOD activity in erythrocytes, suggesting that this was within sufficient levels to neutralize the exceeding superoxide anions formed upon higher energetic demands [63,64]. Secondly, regardless of sex and breeding status, increased cooperative contributions did not increase blood oxidative damage to lipids (MDA). Such results could be suggestive of individual adjustments of cooperative contributions according to their oxidative status, yet initial oxidative markers did not predict the extent to which individuals increased cooperative contributions during the sand treatment. Studies in the wild often suggest that individual condition and oxidative status affect energetically demanding cooperative contributions [13,32,33,65], an effect that may have been shadowed in our study by captive conditions and ad libitum food provision. Direct manipulations of individuals' oxidative balance [66] are needed to experimentally assess the role of oxidative status on cooperative contributions.

Female Damaraland mole-rats are better equipped than males to prevent and even reduce somatic oxidative damage upon increased cooperative contributions. While males were able to prevent increased somatic lipid damage, females reduced erythrocyte damage levels at the end of the sand treatment. Sex differences in antioxidant defences not measured in our study [67] and/or efficiency of mechanisms repairing oxidative damage to lipids [27] could explain this reduction in MDA levels in females, and should be addressed in future studies. Decreased oxidative damage may represent an adaptive oxidative shielding response of females to the possibility of encountering foreign mates of neighbouring colonies in the wild [68].

The greater oxidative costs of energetically demanding cooperative behaviours may be those imposed to the germline, with potential detrimental effects on direct fitness. Although the increased SOD activity in ejaculates suggests a protection response against ROS, increased cooperative contributions also resulted in elevated oxidative damage in ejaculates, possibly compromising individuals' reproductive success. Indeed, high levels of lipid peroxidation in the sperm are associated with decreased sperm quality [69] due to reduced motility and ability to fuse with oocytes [70,71]. Although Damaraland mole-rat helpers do not reproduce within their natal colony, reproductive opportunities in the wild may arise when the underground tunnel systems of two colonies become connected, for example due to animals extending their tunnels in search of new food sources. In such situations, compromised sperm quality can negatively affect direct fitness of male helpers. The prediction that male breeders may invest in germline protection to a greater extent than helpers was not supported by our study. Instead, our results may highlight a lack of reproductive competition between the males, allowing them to withstand some level of damage to the germline.

Our study supports the hypothesis that oxidative stress mediates a trade-off between survival and reproduction through a trade-off between somatic and germline tissues [72]. Males accumulated oxidative damage in the germline to a greater extent than in the somatic tissue, suggesting that overall antioxidant protection has been traded-off. Oxidative damage integrates the negative effects of ROS and the protective effects of antioxidants and repair mechanisms [45] and thus provides reliable information about the outcome of a potential trade-off between somatic and germline tissues. Upon an oxidative challenge, Damaraland mole-rat males prioritize survival over reproduction by avoiding increased oxidative damage to the soma at the expense of damage to the germline. For male helpers, such a strategy may be adaptive because reproduction is unlikely while staying in their natal colony [35,3739], and investment in survival may result in increased indirect fitness and increased prospects of future reproduction [73,74]. For male breeders, the lack of reproductive competition may have favoured investment in survival as long as reproduction is not fully compromised (i.e. spermatozoa remain capable of fertilization). Aligned with these results, spermatozoa characteristics are similar in Damaraland mole-rat helpers and breeders [40], suggesting that breeding status may not play a significant role in germline investment in this species.

Although heightened cooperative contributions induce oxidative costs capable of leading to fitness reductions in the short term, whether they do so in the long term [4,75] remains to be confirmed. Individuals may avoid the long-term build-up of oxidative damage through two non-mutually exclusive mechanisms. Firstly, individuals can adjust cooperative contributions according to their past contributions and current condition [15,76,77]. Secondly, individuals can adjust their oxidative stress physiology [31], though upregulation of antioxidant production and oxidative damage repair mechanisms, and changes of lipid composition of cellular membranes [30,7880]. Longer-term experimental manipulations of cooperative contributions would be extremely valuable to address this question.

Our study provides the first experimental demonstration that oxidative stress represents a key physiological cost of energetically demanding cooperative behaviours, susceptible to ultimately compromise individuals’ fitness. Furthermore, assessing such costs in somatic and germline tissues is of great importance to gain insights into the physiological mechanisms underpinning one of the best-studied life-history trade-offs that between survival and reproduction. Future studies will gain from addressing this trade-off in females. Finally, these findings have implications on our appreciation of the selective pressures posed by energetically demanding behaviours on the evolution of cooperation in animal societies.

Ethics

All protocols were approved by the Animal Ethics Committee at the University of Pretoria (EC093-14).

Data accessibility

Data and R code used in this study are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.f7m0cfxt3 [81].

Authors' contributions

R.M., P.V., N.C.B. and F.H. designed experiments. R.M. and N.K. collected data. R.M., G.G. and A.V. conducted laboratory analysis. R.M., P.V. and F.H. conducted statistical analysis. R.M. authored the manuscript with contributions from all co-authors.

Competing interests

We declare we have no competing interests

Funding

R.M. was supported by a Postgraduate Research Support bursary awarded by the University of Pretoria. N.C.B. was funded by the Department of Science and Technology/National Research Foundation SARChI chair in Behavioural Ecology and Physiology (GUN 64756); laboratory work was supported by grant nos PP00P3_139011 and PP00P3_165840 awarded from the Swiss National Science Foundation to F.H. Animal husbandry and facilities at the Kalahari Research Centre were funded by European Research Council Advanced Grants (nos 294494 and 742808).

Acknowledgements

We thank the two anonymous reviewers and the associate editor, who provided relevant and positive feedback to improve the manuscript. We thank Ophélie Gning for her assistance during the laboratory analysis. We are grateful to Tim Clutton-Brock for providing access to facilities, captive animals and records of individual identities and life histories collected at the KRC, which has been supported by the Kalahari Research Trust. We thank Tim Vink and Ian Stevenson for their work on database maintained at the KRC. We are grateful to Martha Manser for her contribution to maintaining the KRC and to Nichola Raihani for her contributions at the inception of the Kalahari Mole-Rat Project. We thank Dave Gaynor, Katy Goddard and Markus Zöttl for helping to organize the research project and on-site employees for animal husbandry.

Footnotes

Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5104026.

Published by the Royal Society. All rights reserved.

References