Hormones and Behavior 71 (2015) 10–15
Contents lists available at ScienceDirect
Hormones and Behavior journal homepage: www.elsevier.com/locate/yhbeh
Mild maternal stress disrupts associative learning and increases aggression in offspring L. Eaton a,⁎, E.J. Edmonds a, T.B. Henry b,c,d, D.L. Snellgrove e, K.A. Sloman a a
Institute of Biomedical and Environmental Health Research, School of Science and Sport, University of the West of Scotland, Paisley, Scotland PA1 2BE, UK School of Life Sciences, Heriot-Watt University, Edinburgh, Scotland EH14 4AS, UK Center for Environmental Biotechnology, University of Tennessee, Knoxville, TN 37996, USA d Department of Forestry, Wildlife and Fisheries, University of Tennessee, Knoxville, TN 37996 USA e WALTHAM Centre for Pet Nutrition, Freeby Lane, Waltham-on-the-Wolds, LE14 4RT, Leicestershire, UK b c
a r t i c l e
i n f o
Article history: Received 26 November 2014 Revised 17 March 2015 Accepted 24 March 2015 Available online 1 April 2015 Keywords: Guppies Maternal effects Mirror-image test Plus-maze learning Viviparity
a b s t r a c t Maternal stress has been shown to affect behaviour of offspring in a wide range of animals, but this evidence has come from studies that exposed gestating mothers to acute or severe stressors, such as restraint or exposure to synthetic stress hormones. Here we show that exposure of mothers to even a mild stressor reduces associative learning and increases aggression in offspring. Female guppies were exposed to routine husbandry procedures that produced only a minimal, non-significant, elevation of the stress hormone cortisol. In contrast to controls, offspring from mothers that experienced this mild stress failed to learn to associate a colour cue and food reward, and showed a greater amount of inter-individual variation in behaviour compared with control offspring. This mild stress also resulted in offspring that were more aggressive towards their own mirror image than controls. While it is possible that these results could represent the transmission of beneficial maternal characteristics to offspring born into unpredictable environments, the potential for mild maternal stress to affect offspring performance also has important implications for research into the trans-generational effects of stress. © 2015 Elsevier Inc. All rights reserved.
Introduction There are many ways that parents can influence offspring phenotype through non-genomic pathways both directly (e.g. nutrient and hormone transfer) and indirectly (e.g. parental care behaviours) (Bernardo, 1996; Donaldson et al., 2008a,b; Mousseau and Fox, 1998). While in some species these parental effects may be due to both parents, maternal effects on offspring have received the most attention and have been demonstrated in both oviparous and viviparous animals. A key element of non-genomic maternal influence is maternal stress and there is considerable evidence in mammals that exposure to pre-natal stress affects the hormonal and behavioural development of offspring (Huizink et al., 2004; de Kloet et al., 2005; Kofman, 2002). In particular, gestational stress can increase the incidence of anxiogenic and depressive-like behaviours in offspring and impair cognitive performance (Vallée et al., 1999; Weinstock, 2005). The majority of studies investigating maternal stress and offspring performance have subjected mothers to significant levels of stress, either through direct imposition of an acute stressor or by supplementation of endogenous hormone levels with synthetic hormones (Eriksen et al., 2006; Janczak et al., 2006; Sloman, 2010; Takahashi et al., 1998; Vallée ⁎ Corresponding author at: School of Science and Sport, University of the West of Scotland, Paisley, Renfrewshire, Scotland PA1 2BE, UK. E-mail address: [email protected] (L. Eaton).
http://dx.doi.org/10.1016/j.yhbeh.2015.03.005 0018-506X/© 2015 Elsevier Inc. All rights reserved.
et al., 1999; Weinstock, 2005). However, the effect of milder stressors, which are likely to be a more common experience for a mother, has received little attention. Transmission of a variety of information about a mother's environment to her offspring has been previously demonstrated (Dadda and Bisazza, 2012; Giesing et al., 2011), possibly through the involvement of glucocorticoid hormones, although the exact mechanisms remain elusive. Thus, it seems feasible that mild environmental disturbance could alter offspring physiology and behaviour with implications for our interpretation of non-genomic effects. The aim of this study was to examine the effects of mild maternal stress induced on offspring in the viviparous guppy. To our knowledge this is the first study to investigate the effects of a mild maternal husbandry stressor in any animal and the first to consider the effects of maternal stress in a viviparous fish. Maternal stress was applied both prior to fertilisation and during gestation. Offspring behaviour was tested using measures of associative learning and competitive ability. Methods Adult virgin female guppies (total length 4.51 ± 0.9 cm), Poecilia reticulata, of various colour morphs were obtained from three different home aquarium fish suppliers and were kept in high density 10 l (21.5 × 19.5 × 40 cm) stock tanks (~ 1.5 g l− 1) on 200 l recirculating systems (flow rate: 0.6 l min− 1 ; 26.9 ± 0.1 °C; pH 6.5 ± 0.1,
L. Eaton et al. / Hormones and Behavior 71 (2015) 10–15
dissolved oxygen 7.52 ± 0.02 mg l − 1 , 12:12 h light:dark) for 4 weeks prior to the start of the study. Fish were fed to satiation daily with a diet supplied by WALTHAM Centre for Pet Nutrition (http://www.waltham.com/). We used fish from three different suppliers to ensure that any significant effects were repeatable in different varieties of guppy and not restricted to one particular batch or strain of fish. Different sources of fish were included as ‘tank replicates’ in data analysis. Groups of six virgin females were then assigned to either a control or stress treatment (n = 3 groups of six females per treatment, i.e. 36 females in total) and held in 19 l tanks on the recirculating system with a plastic plant as environmental enrichment. Each replicate consisted of fish from a different fish supplier. Typical home aquaria husbandry procedures were used to induce mild maternal stress and were carried out twice weekly for 4 weeks. Females assigned to the control treatment experienced 2 min of tank siphoning, where a siphon was used to remove waste food and faeces from the tank, along with a small amount of water. During siphoning, water was continuously replaced via the recirculating system and care was taken to disturb the fish as little as possible. In the stress treatments, tanks were siphoned for 2 min, fish were then carefully netted and transferred to a plastic tank containing 2 l of water from their home tank system for a further 2 min. During this time, the plastic plant in their home tank was moved to a new location. Fish were then re-netted and returned to their original tank. This stress treatment was based on typical home husbandry procedures for pet fish tanks determined via a survey of fish owners (Sloman, K. unpublished data) where fish are removed from their tank while the home tank is thoroughly cleaned. After 4 weeks, three randomly chosen males of various colour morphs (orange delta, yellow delta, blue delta, black delta and/or snakeskin delta morphs) were introduced into each tank. These males were allowed 3 days to interact and inseminate the females, at which time they were removed and another set of three males were introduced for a further 3 days. Females were allowed to multiple mate as this reduces gestation time and produces larger broods in comparison to singly-mated females (Evans and Magurran, 2000). Husbandry treatments continued post-mating as previously detailed, but their frequency was reduced to once weekly to prevent gestating females from selectively aborting broods (Evans and Magurran, 2000). After three weeks of gestation, females were easily identifiable due to their size as being in the latter stages of pregnancy. Treatments were then stopped and females were held in nylon breeding nets (3 mm mesh size) in their groups of six. Fry were born 21–27 days after the males had been removed. Mothers and breeding nets were then removed from the tanks and offspring were raised in situ (n = 3 tanks of offspring per treatment). Fry were fed a mixture of live Artemia, a commercial flake food (Aquarian) and high protein pelleted food (ZM200) daily. Six juveniles were randomly selected from each replicate for plus-maze trials at 14 weeks of age (i.e. n = 18). At 18 weeks of age, when fish were reproductively mature, six individuals not previously used in the plus-maze trials were selected for mirror-image tests (i.e. n = 18). Water-borne cortisol sampling To determine the effect of the husbandry stress treatments on levels of cortisol, an additional experiment was carried out. A new group of female guppies were acclimated to tank conditions in the same way as before and water-borne cortisol was sampled after one exposure to the stress treatments. Water cortisol concentrations following the control and experimental husbandry treatments were compared with water cortisol released from guppies following an acute 1 h transport stress which represented a transport stress from pet shop to home. For the measurements of water cortisol following husbandry procedures, females were removed from their tanks and placed in groups of three (n = 16 groups of three per husbandry treatment) into a
11
beaker containing 150 ml of clean water for 30 min based on the methods of Sebire et al. (2007). For the acute transport stress, groups of guppies (n = 9 groups of three) were held in a plastic bag containing 150 ml water, placed inside a dark box, and moved around the laboratory for 1 h prior to cortisol measurements. After this time, water cortisol concentration was extracted according to the methods of Ellis et al. (2004) and measured using a commercial ELISA (DRG Instruments). The commercial ELISA used a cortisol antibody sensitive to 2.5 ng ml−1 (6.9 nmol l−1), the average intra and inter assay variation was 5.6% and 5.9% respectively. Blank samples of tank system water were used to account for background concentrations of cortisol. All stressors and cortisol measurements were conducted at the same time of day to control for circadian fluctuations in endocrine status. Offspring behavioural tests Plus-maze test The four-armed radial plus-maze (10 l) supplied with system water (26.9 ± 0.1 °C), was based on the design of Sison and Gerlai (2010) and made of white, opaque polypropylene. Each arm was 30 cm long, 10 cm wide and 10 cm high and arms were connected by a central 10 × 10 cm central square. At the end of each individual arm there was a section of tubing connected to an external syringe loaded with bloodworm. Fish were held for 10 s within the central square in a ‘start-box’, an opentopped transparent container attached to a pulley system to acclimate to the maze environment before the start-box was remotely removed. Fish were subjected to five habituation trials in the plus maze, one per day on consecutive days, in decreasing group sizes and times. In the first of these trials, six fish were placed into the start-box, in single treatment groups, and then allowed to explore the plus-maze for 1 h, with food released from all arms after 30 min. The second trial allowed groups of three fish to explore the maze for 30 min with food administered after 15 min in all arms. The third allowed pairs of fish to explore the maze for 20 min with food administered after 10 min and the fourth and fifth, single fish to explore the maze alone for 10 min with food administered in all arms after 5 min. Fish were fed at no other time other than within the plus-maze. After habituation trials, each fish spent 10 min individually within the maze twice daily for 10 days, once in the morning and once in the afternoon. Each period within the plus-maze was recorded from above using a mounted camcorder (Panasonic SDR-S50). A large piece of red card with a hole in the centre was placed at the end of one arm with the food-dispensing tubing pushed through. Food was administered after 5 min of fish being within the plus-maze and only from this colour-coded arm (target-arm), never from non colour-coded arms (non-target arms). All arms, target and non-target, were loaded with bloodworm to ensure that any scent of food leaked equally from each. The plus-maze was curtained with black cloth to prevent any external features acting as guides to the target arm. The order in which fish were subjected to plus-maze trials and the orientation of the target-arm was randomised each day. Maze performance was analysed using JWatcher (http://www. jwatcher.ucla.edu/). The percentage of time spent within the target arm and the percentage of entries made to the target arm were recorded for 10 min after the fish was released from the start-box. Arm entries were recorded as soon as a fish entered that specific arm, if a fish exited an arm and remained within the central square, it was considered to still be within the last arm visited until it entered a different arm. Each fish completed the plus-maze 20 times over a period of 10 days. Mirror image test Mirror image tests have previously been used as a measure of competitive ability since fish respond to their mirror image as if it were another individual (Brick and Jakobsson, 2002; Sloman and Baron, 2010). At approximately 18 weeks of age, mature single males (n = 18) from each maternal treatment were randomly selected for
12
L. Eaton et al. / Hormones and Behavior 71 (2015) 10–15
mirror image tests. Mature males were used in this experiment, as opposed to mixed sex juveniles for plus-maze tests, since mature males are aggressive towards other males and would therefore interact with a mirror image to a greater extent than immature or female fish. Fish were placed individually in static aerated 2 l tanks with a covered mirror, and were left to acclimate for 20 h. The mirror was mounted vertically on the side facing into the tank. The mirror was then uncovered for 1 min and re-covered for a further hour to acclimate the fish to the action of mirror uncovering. Following this, the mirror was uncovered and fish were recorded with a camcorder (Panasonic SDR-S50) for a period of 30 min (Sloman, 2010). The number of behavioural interactions with their mirror image (defined as a bite, a charge or touching the mirror) as well as the latency to touch the mirror was determined. Video analysis of behaviour was randomised to prevent observer bias, all observations were carried out by LE. Statistical analyses Data were analysed using PASW Statistics version 18.0 (http:// www.spss.com.hk/statistics/). Data were tested for normality (Kolmogorov–Smirnov: K–S) and homogeneity of variance, parametric or non-parametric statistics were then used where appropriate. Cortisol data were transformed prior to analysis (√Cortisol + 0.1), were normally distributed (K–S: P = 0.09); and thus analysed using a one-way ANOVA with LSD post-hoc testing. Percentage of time spent within and the percentage of entries made to the target-arm of the plus-maze were arcsine square root transformed, were normally distributed (K–S: P N 0.05) and analysed using repeated-measures (RM) ANOVAs with maternal stress treatment, time and replicate as factors. The interindividual coefficient of variation (CV) was calculated for both the percentage of time spent within and the percentage of entries made to the target-arm. Inter-individual coefficient of variation data were not normally distributed (K–S: percentage of time: P b 0.05; percentage of entries: P b 0.005), thus, a natural log transformation was applied and data were normalised (K–S: percentage of time: P = 0.31; percentage of entries: P = 0.82). Measures of inter-individual CV of the percentage of time spent within, and the percentage of entries made to the target arm were analysed using un-paired t-tests. Number of interactions with a mirror image were normally distributed (K–S: P = 0.06), data were therefore analysed using a two-way ANOVA. Latency to interact with the mirror image was not normally distributed (K–S: P b 0.001) and was analysed using a two-way Scheirer–Ray–Hare (S–R–H) test. For both the number of interactions with, and the latency to interact with the mirror image, maternal treatment and tank replicate were included as factors. Effect sizes were estimated using eta squared (η2) and Cohen's d for RM ANOVAs, Scheirer–Ray–Hare and t-test analyses. Effect sizes equal to or less than 0.2, 0.5 and 0.8 have been categorised as small, medium and large effect sizes according to Cohen (1988). Ethical note The study was approved by the WALTHAM ethical committee and carried out in accordance with University of Plymouth and University of the West of Scotland ethical guidelines under a U.K. Home Office project licence to KAS. Fish were not subjected to any invasive techniques. Stress exposures comprised typical home-husbandry procedures and all fish had access to environmental enrichment. All fish were held under appropriate water quality conditions. Results Maternal cortisol A significantly higher concentration of water-borne cortisol was measured when fish were subjected to the simulated transport stress compared to the control or mild husbandry stress (Fig. 1: K–S: 0.09;
Fig. 1. Cortisol concentrations. Mean (±SEM) total water-borne concentration (ng/fish/h) from gestating female guppies exposed to a control treatment, a mild husbandry stressor and a simulated transport stress. Asterisk denotes significant between treatments (ANOVA; LSD).
one-way ANOVA: F1,38 = 14.378, P b 0.001, η2 = 0.39, post-hoc LSD). When cortisol excretion rates were analysed independently of the transport stress data, there was still no statistical significance between treatments (K–S: P = 0.30; one-way ANOVA: F1,28 = 0.146, P = 0.767, η2 = 0.001), Offspring behaviour Offspring whose mothers were subjected to mild stress treatments prior to and during gestation spent a lower percentage of time in the target arm than control offspring (Fig. 2A: RM ANOVA: F1,30 = 70.79, P b 0.001, η2 = 0.86). There was a significant interaction between treatment and repeated trials i.e. time (F19,30 = 435.00, P b 0.001, η2 = 0.12); further analysis showed that offspring from control treatments increased the percentage of time spent within the target arm with repeated trials (Fig. 2A: RM ANOVA: F19,323 = 4.21, P b 0.001, η2 = 0.15) whereas offspring from stress treatment did not (RM ANOVA: F19,323 = 0.87, P = 0.618, η2 = 0.03). Offspring from the stress treatment also made a lower percentage of entries to the target arm of a plus maze than controls (Fig. 2B: RM ANOVA: F1,30 = 76.87, P b 0.001, η2 = 0.31); again when considered separately, offspring from the control treatment increased the percentage of entries made to the target arm with repeated trials (Fig. 2B: RM ANOVA: F19,323 = 5.25, P b 0.001, η2 = 0.16) whereas offspring from the stress treatment did not (F19,340 = 0.73, P = 0.79, η2 = 0.02). For both percentage of time spent in the target arm and percentage of entries to the target arm, there was a significant effect of replicate i.e. origin of the fish and a significant interaction with treatment (RM ANOVA: percentage time: F2,30 = 57.04, P b 0.001; stress* replicate: ANOVA: F2,30 = 5.16, P b 0.05; number of entries: F2,30 = 73.15, P b 0.001; stress* replicate: ANOVA: F2,30 = 11.98, P b 0.001). Further analysis of each replicate separately showed that the significant effect of maternal stress treatment was consistent across all replicates individually (RM: ANOVA P b 0.01), but that the magnitude of response varied between these different batches of fish. Mild maternal stress increased the inter-individual CV of time spent within the target-arm (Fig. 3: un-paired t-test: df = 118, t = − 4.75, P b 0.01, Cohen's d = 0.86), but there was no difference in interindividual CV for entries made to the target-arm (un-paired t-test: df = 118, t = − 1.08, P = 0.281, Cohen's d = 0.69). Offspring from the stress treatment interacted more with their mirror image than those from the control treatment (Fig. 4: two-way ANOVA: F1,52 = 19.03, P b 0.001, η2 = 0.22) with a significant effect of tank replicate (F2,52 = 4.36, P b 0.05 η2 = 0.10) and interaction (F2,52 = 3.67,
L. Eaton et al. / Hormones and Behavior 71 (2015) 10–15
13
different batches of fish. The latency for fish to interact with a mirror image was not affected by maternal stress (S–R–H: df = 1, SS = 9.41, P = 0.93) and there was no significant effect of tank replicate (S–R–H: df = 1, SS = 1577.51, P = 0.5) or maternal stress*tank replicate interaction (S–R–H: df = 2, SS = 37.7, P = 0.98).
Discussion
Fig. 2. Maze performance. Mean (±SEM) percentage of (A) time spent in and (B) entries made to the target arm of a plus maze by offspring from control and stress treatments over 20 trials. n = 18 fish per treatment.
P b 0.05, η2 = 0.08). Further analysis of each replicate showed the effects of maternal stress on offspring aggression to be consistent across all replicates (P b 0.05), but the magnitude of difference varied between
Fig. 3. Maze performance variability. Mean (±SEM) inter-individual coefficient of variation of time spent within and entries made to the target arm of the plus maze. Asterisk denotes significance between treatments (un-paired t-tests). n = 18 fish per treatment.
Offspring from mothers which experienced prenatal mild stress (both pre-conception and during gestation) did not exhibit improved performance over time in an associative learning task. In contrast, offspring from control mothers learnt to associate a colour cue with a food reward. Previous studies investigating the consequences of maternal stress have focussed on substantial stress events to the mother and have not addressed the possibility that mild stressors could have long-term effects on offspring behaviour. To our knowledge, this is also the first study to demonstrate the effects of maternal stress on offspring behaviour in a viviparous fish which presents an interesting intermediate between studies in mammals and oviparous animals. The lack of associative learning by offspring from mothers experiencing mild stress may not necessarily be disadvantageous. While there is a tendency for phenotypic responses in offspring attributed to maternally-derived stress to be interpreted as negative, there is increasing evidence that maternal stress can have adaptive benefits (Love et al., 2013; Sheriff and Love, 2013). Strong association between a visual cue and food reward may be disadvantageous if either the cue is false or the environment highly variable. Therefore, lack of associative learning, or capacity to forget, may be an adaptive mechanism (Warburton, 2003). In the present study, mild maternal stress involved the movement of environmental enrichment each time the tank was cleaned, thus mothers experienced a variable environment and this information may have been transferred to their offspring. Giesing et al. (2011) demonstrated that female sticklebacks can transfer information regarding predation levels in their environment to embryos and hypothesised that this might occur through maternal cortisol transfer to eggs. Similarly, in gold belly minnows, light intensity of maternal environment altered expression of lateralisation in offspring in both visual and motor tasks (Dadda and Bisazza, 2012), and again the authors hypothesised that this might be mediated through maternal cortisol. Even though we did not detect a significant elevation in cortisol between maternal stress treatments in the present study, the involvement of cortisol in mediating these effects remains likely but at much lower levels than previously documented. It seems unlikely that the water cortisol concentrations measured in the present study represent true basal levels; a drawback to this non-
Fig. 4. Interactions with a mirror image. Mean (±SEM) number of interactions with a mirror image. Asterisk denotes significance between treatments (two-way ANOVA). n = 18 fish per treatment.
14
L. Eaton et al. / Hormones and Behavior 71 (2015) 10–15
invasive methodology for measuring stress hormones is that to detect such low levels fish are confined in a relatively small volume of water which can represent a confinement stress. However, fish from all treatments experienced the same methodology making comparisons of water cortisol between treatments possible as has been done in other studies (Ellis et al., 2007; Scott et al., 2008). It is also possible that fish experiencing husbandry stressors repeated over seven weeks could acclimate to the stress (Wong et al., 2008). Our measurements of cortisol were taken immediately after the first husbandry stressor experience, it was therefore assumed at this point the rise in cortisol concentrations would be the greatest due to exposure to an unfamiliar stressor with no acclimation period. Inter-individual variation in the percentage of time spent within the target arm was significantly higher in offspring from the stress treatment, which may also be beneficial in an unpredictable environment. ‘Bet-hedging’, where mothers increase phenotypic variability among offspring to ensure survival of at least some individuals in a changeable environment has been suggested (Gagliano and McCormick, 2007; Philippi and Seger, 1989). In birds, variable maternal investment to eggs within the same clutch, including concentrations of maternal hormones, may create degrees of competitive asymmetry between individuals (Muller and Groothuis, 2013). Whether such intentional intrauterine manipulation of offspring can occur in viviparous fishes is unknown. Offspring from mothers exposed to a mild stressor performed more bites towards their own mirror image than offspring from control mothers. Being born into an unpredictable environment may necessitate greater competition for food and mates which may be aided by higher innate levels of aggression. Higher levels of aggression have also been demonstrated in juvenile brown trout exposed pre-fertilisation to synthetic cortisol (Sloman, 2010). Interestingly, however, Burton et al. (2011) showed that exposure of brown trout embryos to synthetic cortisol immediately after fertilisation produced offspring with reduced competitive abilities suggesting that timing of exposure to maternal glucocorticoids is critical in shaping offspring behaviour. The mammalian literature traditionally divides the effects of maternal stress into pre-natal and post-natal (Love et al., 2013), since immediate post-natal motheroffspring interactions can have profound effects on offspring behaviour (Sheriff and Love, 2013). No parental care occurs in the guppy, so maternal stress can only act prenatally, which can be further subdivided into pre-conception and gestation periods. In the present study mothers were stressed both pre- and post-fertilisation. If maternal cortisol effects on offspring behaviour in viviparous fish mirror those seen in oviparous fishes where pre- and post-fertilisation cortisol exposure increases and decreases offspring aggression respectively, it could be hypothesised that when both pre- and post-fertilisation stressors occur that effects of pre-fertilisation exposure dominate. However, the importance of exposure timing in viviparous fishes requires further investigation. In mammals, exposure of adults to elevated glucocorticoids reduces neurogenesis (Gould et al., 1997; Kozorovitskiy and Gould, 2004) and similar effects have been seen in fish (Sørensen et al., 2011), which could underlie behavioural changes in cognition. In oviparous fish, the HPI axis does not become functional until after hatching (Alsop and Vijayan, 2008; Auperin and Geslin, 2008) and in mammals the HPA axis becomes functional around mid-gestation (Tegethoff et al., 2009). Exactly when the HPI axis becomes functional in viviparous fishes is currently unknown. Given that exposure to maternal stress hormones can exert effects pre-fertilisation (Sloman, 2010), it seems unlikely that these behavioural changes are a direct result of interference with the HPI axis (Li et al., 2010). Within the mammalian literature there is increasing evidence for epigenomic programming via early life stress exposure (Li and Leatherland, 2012a). In bonnet macaques (Macaca radiata), enhanced behavioural stress reactivity following early life stress was associated with epigenetic changes (Kinnally et al., 2011) and unpredictable maternal separation during postnatal development in mice induced aggressive-like behaviours and altered DNA methylation
of several candidate genes (Franklin et al., 2010). In fish, there is evidence that cortisol exposure of oocytes causes changes in expression patterns of genes associated with the innate immune system and growth (Li et al., 2010; Li and Leatherland, 2012b). Thus, it seems increasingly likely that maternal effects of cortisol very early in life act via epigenetic mechanisms, most likely before the HPI/A axis is functional. The pronounced behavioural effects seen in offspring as result of this routine husbandry stressor highlight that even relatively mild changes in maternal environment can have concomitant effects on offspring phenotype. Whether or not these effects are advantageous will depend on the environmental conditions they encounter (McCormick, 1998, 1999, 2006, 2009; McCormick and Nechaev, 2002; Sheriff and Love, 2013). The mild maternal stress used in the present study represented a husbandry procedure used in home aquaria. Many animals are bred in captivity for a variety of reasons including as companion animals, for stocking programmes and for research. Therefore, the present study also demonstrates the importance of carefully considering the exposure of captive mothers to routine mild stressors as they have the potential to alter offspring phenotype. Further work is needed to address the specific mechanisms through which maternal stress acts on offspring behaviour and to combine work with oviparous and viviparous animals to investigate the effects of timing of maternal stress. Acknowledgments We thank Stanley McMahon, Deirdre Galbraith and Edith Burns for fish husbandry and Rick Preston for maze construction. The project was generously supported by a WALTHAM Foundation grant to KAS. References Alsop, D., Vijayan, M.M., 2008. Development of the corticosteroid stress axis and receptor expression in zebrafish. Am. J. Physiol. 294, R711–R719. Auperin, B., Geslin, M., 2008. Plasma cortisol response to stress in juvenile rainbow trout is influenced by their life history during early development and by egg cortisol content. Gen. Comp. Endocrinol. 158, 234–239. Bernardo, J., 1996. Maternal effects in animal ecology. Am. Zool. 36, 83–105. Brick, O., Jakobsson, S., 2002. Individual variation in risk taking: the effect of a predatory threat on fighting behavior in Nannacara anomala. Behav. Ecol. 13, 439–442. Burton, T., Hoogenboom, M.O., Armstrong, J.D., Groothuis, T.G.G., Metcalfe, N.B., 2011. Egg hormones in a highly fecund vertebrate: do they influence offspring social structure in competitive conditions? Funct. Ecol. 25, 1379–1388. Cohen, J., 1988. Statistical Power Analysis for the Behavioural Sciences. Routledge Academic, New York, NY. Dadda, M., Bisazza, A., 2012. Prenatal light exposure affects development of behavioural lateralization in a livebearing fish. Behav. Process. 91, 115–118. de Kloet, E.R., Sibug, R.M., Helmerhorst, F.M., Schmidt, M., 2005. Stress, genes and the mechanism of programming the brain for later life. Neurosci. Biobehav. Rev. 29, 271–281. Donaldson, J.M., Munday, P.L., McCormick, M.I., 2008a. Parental effects on offspring life histories: when are they important? Biol. Lett. 5, 262–265. Donaldson, J.M., McCormick, M.I., Munday, P.L., 2008b. Parental condition affects early life-history of a coral reef fish. J. Exp. Mar. Biol. Ecol. 360, 109–116. Ellis, T., James, J.D., Stewart, C., Scott, A.P., 2004. A non-invasive stress assay based upon measurement of free cortisol released into the water by rainbow trout. J. Fish Biol. 1233–1252. Ellis, T., James, J.D., Sundh, H., Fridell, F., Sundell, K., Scott, A.P., 2007. Non-invasive measurement of cortisol and melatonin in tanks stocked with seawater Atlantic salmon. Aquaculture 698–706. Eriksen, M.S., Bakken, M., Espmark, A., Braastad, B.O., Salte, R., 2006. Prespawning stress in farmed Atlantic salmon Salmo salar: maternal cortisol exposure and hyperthermia during embryonic development affect offspring survival, growth and incidence of malformations. J. Fish Biol. 69, 114–129. Evans, J.P., Magurran, A.E., 2000. Multiple benefits of multiple mating in guppies. Proc. Natl. Acad. Sci. U. S. A. 97, 10074–10076. Franklin, T.B., Russig, H., Weiss, I.C., Gräff, J., Linder, N., Michalon, A., Vizi, S., Mansuy, I.M., 2010. Epigenetic transmission of the impact of early stress across generations. Biol. Psychiatry 68, 408–415. Gagliano, M., McCormick, M.I., 2007. Maternal condition influences phenotypic selection on offspring. J. Anim. Ecol. 76, 174–182. Giesing, E.R., Suski, C.D., Warner, R.E., Bell, A.M., 2011. Female sticklebacks transfer information via eggs: effects of maternal experience with predators on offspring. Proc. R. Soc. B Biol. Sci. 278, 1753–1759. Gould, E., McEwen, B.S., Tanapat, P., Galea, L.A.M., Fuchs, E., 1997. Neurogenesis in the dentate gyrus of the adult tree shrew is regulated by psychosocial stress and NMDA receptor activation. J. Neurosci. 17, 2492–2498.
L. Eaton et al. / Hormones and Behavior 71 (2015) 10–15 Huizink, A.C., Mulder, E.J.H., Buitelaar, J.K., 2004. Prenatal stress and risk for psychopathology: Specific effects or induction of general susceptibility? Psychol. Bull. 130, 115–142. Janczak, A.M., Braastad, B.O., Bakken, M., 2006. Behavioural effects of embryonic exposure to corticosterone in chickens. Appl. Anim. Behav. Sci. 96, 69–82. Kinnally, E.L., Feinberg, C., Kim, D., Ferguson, K., Leibel, R., Coplan, J.D., Mann, J.J., 2011. DNA methylation as a risk factor in the effects of early life stress. Brain Behav. Immun. 25, 1548–1553. Kofman, O., 2002. The role of prenatal stress in the etiology of developmental behavioural disorders. Neurosci. Biobehav. Rev. 26, 457–470. Kozorovitskiy, Y., Gould, E., 2004. Dominance hierarchy influences adult neurogenesis in the dentate gyrus. J. Neurosci. 24, 6755–6759. Li, M., Leatherland, J.F., 2012a. The implications for aquaculture practice of epigenomic programming of components of the endocrine system of teleostean embryos: lessons learned from mammalian studies. Fish Fish. 14 (4), 528–553. Li, M., Leatherland, J.F., 2012b. The interaction between maternal stress and the ontogeny of the innate immunity during fish embryogenesis: implications for aquaculture practice. J. Fish Biol. 81, 1793–1814. Li, M., Bureau, D.P., King, W.A., Leatherland, J.F., 2010. The actions of in ovo cortisol on egg fertility, embryo development and the expression of growth-related genes in rainbow trout embryos, and the growth performance of juveniles. Mol. Reprod. Dev. 77, 922–931. Love, O.P., McGowan, P., Sheriff, M.J., 2013. Maternal adversity and ecological stressors in natural populations: the role of stress axis programming in individuals, with implications for populations and communities. Funct. Ecol. 27, 81–92. McCormick, M.I., 1998. Behaviorally induced maternal stress in a fish influences progeny quality by a hormonal mechanism. Ecology 79, 1873–1883. McCormick, M.I., 1999. Experimental test of the effect of maternal hormones on larval quality of a coral reef fish. Oecologia 118, 412–422. McCormick, M.I., 2006. Mothers matter: crowding leads to stressed mothers and smaller offspring in marine fish. Ecology 87, 1104–1109. McCormick, M.I., 2009. Indirect effects of heterospecific interactions on progeny size through maternal stress. Oikos 118, 744–752. McCormick, M.I., Nechaev, I.V., 2002. Influence of cortisol on developmental rhythms during embryogenesis in a tropical damselfish. J. Exp. Zool. 293, 456–466. Mousseau, T.A., Fox, C.W., 1998. The adaptive significance of maternal effects. Trends Ecol. Evol. 13, 403–407. Muller, M., Groothuis, T.G.G., 2013. Within-clutch variation in yolk testosterone as an adaptive maternal effect to modulate avian sibling competition: evidence from a comparative study. Am. Nat. 181, 125–136.
15
Philippi, T., Seger, J., 1989. Hedging one's evolutionary bets, revisited. Trends Ecol. Evol. 4, 41–44. Scott, A.P., Hirschenhauser, K., Bender, N., Oliveira, R., Earley, R.L., Sebire, M., Ellis, T., Pvalidis, M., Hubbard, P.C., Huertas, M., Canario, A., 2008. Non-invasive measurements of steroids in fish-holding water: important considerations when applying the procedure to behaviour studies. Behaviour 145, 1307–1328. Sebire, M., Katsiadaki, I., Scott, A.P., 2007. Non-invasive measurement of 11-ketotestosterone, cortisol and androstenedione in male three-spined stickleback (Gasterosteus aculeatus). Gen. Comp. Endocrinol. 152, 30–38. Sheriff, M.J., Love, O.P., 2013. Determining the adaptive potential of maternal stress. Ecol. Lett. 16, 271–280. Sison, M., Gerlai, R., 2010. Associative learning in zebrafish (Danio rerio) in the plus maze. Behav. Brain Res. 207, 99–104. Sloman, K.A., 2010. Exposure of ova to cortisol pre-fertilisation affects subsequent behaviour and physiology of brown trout. Horm. Behav. 58, 433–439. Sloman, K.A., Baron, M., 2010. Conspecific presence affects the physiology and behaviour of developing trout. Physiol. Behav. 99, 599–604. Sørensen, C., Bohlin, L.C., Øverli, O., Nilsson, G.E., 2011. Cortisol reduces cell proliferation in the telencephalon of rainbow trout (Oncorhynchus mykiss). Physiol. Behav. 102, 518–523. Takahashi, L.K., Turner, J.G., Kalin, N.H., 1998. Prolonged stress-induced elevation in plasma corticosterone during pregnancy in the rat: implications for prenatal stress studies. Psychoneuroendocrinology 23, 571–581. Tegethoff, M., Pryce, C., Meinlschmidt, G., 2009. Effects of intrauterine exposure to synthetic glucocorticoids on fetal, newborn, and infant hypothalamic–pituitary–adrenal axis function in humans: a systematic review. Endocr. Rev. 30, 753–789. Vallée, M., Maccari, S., Dellu, F., Simon, H., Le Moal, M., Mayo, W., 1999. Long-term effects of prenatal stress and postnatal handling on age-related glucocorticoid secretion and cognitive performance: a longitudinal study in the rat. Eur. J. Neurosci. 11, 2906–2916. Warburton, K., 2003. Learning of foraging skills by fish. Fish Fish. 4, 203–215. Weinstock, M., 2005. The potential influence of maternal stress hormones on development and mental health of the offspring. Brain Behav. Immun. 19, 296–308. Wong, S.C., Dykstra, M., Campbell, J.M., Earley, R.L., 2008. Measuring water-borne cortisol in convict cichlids (Amatitlania nigrofasciata): is the procedure a stressor? Behaviour 145, 1283–1305.
Contents lists available at ScienceDirect
Hormones and Behavior journal homepage: www.elsevier.com/locate/yhbeh
Mild maternal stress disrupts associative learning and increases aggression in offspring L. Eaton a,⁎, E.J. Edmonds a, T.B. Henry b,c,d, D.L. Snellgrove e, K.A. Sloman a a
Institute of Biomedical and Environmental Health Research, School of Science and Sport, University of the West of Scotland, Paisley, Scotland PA1 2BE, UK School of Life Sciences, Heriot-Watt University, Edinburgh, Scotland EH14 4AS, UK Center for Environmental Biotechnology, University of Tennessee, Knoxville, TN 37996, USA d Department of Forestry, Wildlife and Fisheries, University of Tennessee, Knoxville, TN 37996 USA e WALTHAM Centre for Pet Nutrition, Freeby Lane, Waltham-on-the-Wolds, LE14 4RT, Leicestershire, UK b c
a r t i c l e
i n f o
Article history: Received 26 November 2014 Revised 17 March 2015 Accepted 24 March 2015 Available online 1 April 2015 Keywords: Guppies Maternal effects Mirror-image test Plus-maze learning Viviparity
a b s t r a c t Maternal stress has been shown to affect behaviour of offspring in a wide range of animals, but this evidence has come from studies that exposed gestating mothers to acute or severe stressors, such as restraint or exposure to synthetic stress hormones. Here we show that exposure of mothers to even a mild stressor reduces associative learning and increases aggression in offspring. Female guppies were exposed to routine husbandry procedures that produced only a minimal, non-significant, elevation of the stress hormone cortisol. In contrast to controls, offspring from mothers that experienced this mild stress failed to learn to associate a colour cue and food reward, and showed a greater amount of inter-individual variation in behaviour compared with control offspring. This mild stress also resulted in offspring that were more aggressive towards their own mirror image than controls. While it is possible that these results could represent the transmission of beneficial maternal characteristics to offspring born into unpredictable environments, the potential for mild maternal stress to affect offspring performance also has important implications for research into the trans-generational effects of stress. © 2015 Elsevier Inc. All rights reserved.
Introduction There are many ways that parents can influence offspring phenotype through non-genomic pathways both directly (e.g. nutrient and hormone transfer) and indirectly (e.g. parental care behaviours) (Bernardo, 1996; Donaldson et al., 2008a,b; Mousseau and Fox, 1998). While in some species these parental effects may be due to both parents, maternal effects on offspring have received the most attention and have been demonstrated in both oviparous and viviparous animals. A key element of non-genomic maternal influence is maternal stress and there is considerable evidence in mammals that exposure to pre-natal stress affects the hormonal and behavioural development of offspring (Huizink et al., 2004; de Kloet et al., 2005; Kofman, 2002). In particular, gestational stress can increase the incidence of anxiogenic and depressive-like behaviours in offspring and impair cognitive performance (Vallée et al., 1999; Weinstock, 2005). The majority of studies investigating maternal stress and offspring performance have subjected mothers to significant levels of stress, either through direct imposition of an acute stressor or by supplementation of endogenous hormone levels with synthetic hormones (Eriksen et al., 2006; Janczak et al., 2006; Sloman, 2010; Takahashi et al., 1998; Vallée ⁎ Corresponding author at: School of Science and Sport, University of the West of Scotland, Paisley, Renfrewshire, Scotland PA1 2BE, UK. E-mail address: [email protected] (L. Eaton).
http://dx.doi.org/10.1016/j.yhbeh.2015.03.005 0018-506X/© 2015 Elsevier Inc. All rights reserved.
et al., 1999; Weinstock, 2005). However, the effect of milder stressors, which are likely to be a more common experience for a mother, has received little attention. Transmission of a variety of information about a mother's environment to her offspring has been previously demonstrated (Dadda and Bisazza, 2012; Giesing et al., 2011), possibly through the involvement of glucocorticoid hormones, although the exact mechanisms remain elusive. Thus, it seems feasible that mild environmental disturbance could alter offspring physiology and behaviour with implications for our interpretation of non-genomic effects. The aim of this study was to examine the effects of mild maternal stress induced on offspring in the viviparous guppy. To our knowledge this is the first study to investigate the effects of a mild maternal husbandry stressor in any animal and the first to consider the effects of maternal stress in a viviparous fish. Maternal stress was applied both prior to fertilisation and during gestation. Offspring behaviour was tested using measures of associative learning and competitive ability. Methods Adult virgin female guppies (total length 4.51 ± 0.9 cm), Poecilia reticulata, of various colour morphs were obtained from three different home aquarium fish suppliers and were kept in high density 10 l (21.5 × 19.5 × 40 cm) stock tanks (~ 1.5 g l− 1) on 200 l recirculating systems (flow rate: 0.6 l min− 1 ; 26.9 ± 0.1 °C; pH 6.5 ± 0.1,
L. Eaton et al. / Hormones and Behavior 71 (2015) 10–15
dissolved oxygen 7.52 ± 0.02 mg l − 1 , 12:12 h light:dark) for 4 weeks prior to the start of the study. Fish were fed to satiation daily with a diet supplied by WALTHAM Centre for Pet Nutrition (http://www.waltham.com/). We used fish from three different suppliers to ensure that any significant effects were repeatable in different varieties of guppy and not restricted to one particular batch or strain of fish. Different sources of fish were included as ‘tank replicates’ in data analysis. Groups of six virgin females were then assigned to either a control or stress treatment (n = 3 groups of six females per treatment, i.e. 36 females in total) and held in 19 l tanks on the recirculating system with a plastic plant as environmental enrichment. Each replicate consisted of fish from a different fish supplier. Typical home aquaria husbandry procedures were used to induce mild maternal stress and were carried out twice weekly for 4 weeks. Females assigned to the control treatment experienced 2 min of tank siphoning, where a siphon was used to remove waste food and faeces from the tank, along with a small amount of water. During siphoning, water was continuously replaced via the recirculating system and care was taken to disturb the fish as little as possible. In the stress treatments, tanks were siphoned for 2 min, fish were then carefully netted and transferred to a plastic tank containing 2 l of water from their home tank system for a further 2 min. During this time, the plastic plant in their home tank was moved to a new location. Fish were then re-netted and returned to their original tank. This stress treatment was based on typical home husbandry procedures for pet fish tanks determined via a survey of fish owners (Sloman, K. unpublished data) where fish are removed from their tank while the home tank is thoroughly cleaned. After 4 weeks, three randomly chosen males of various colour morphs (orange delta, yellow delta, blue delta, black delta and/or snakeskin delta morphs) were introduced into each tank. These males were allowed 3 days to interact and inseminate the females, at which time they were removed and another set of three males were introduced for a further 3 days. Females were allowed to multiple mate as this reduces gestation time and produces larger broods in comparison to singly-mated females (Evans and Magurran, 2000). Husbandry treatments continued post-mating as previously detailed, but their frequency was reduced to once weekly to prevent gestating females from selectively aborting broods (Evans and Magurran, 2000). After three weeks of gestation, females were easily identifiable due to their size as being in the latter stages of pregnancy. Treatments were then stopped and females were held in nylon breeding nets (3 mm mesh size) in their groups of six. Fry were born 21–27 days after the males had been removed. Mothers and breeding nets were then removed from the tanks and offspring were raised in situ (n = 3 tanks of offspring per treatment). Fry were fed a mixture of live Artemia, a commercial flake food (Aquarian) and high protein pelleted food (ZM200) daily. Six juveniles were randomly selected from each replicate for plus-maze trials at 14 weeks of age (i.e. n = 18). At 18 weeks of age, when fish were reproductively mature, six individuals not previously used in the plus-maze trials were selected for mirror-image tests (i.e. n = 18). Water-borne cortisol sampling To determine the effect of the husbandry stress treatments on levels of cortisol, an additional experiment was carried out. A new group of female guppies were acclimated to tank conditions in the same way as before and water-borne cortisol was sampled after one exposure to the stress treatments. Water cortisol concentrations following the control and experimental husbandry treatments were compared with water cortisol released from guppies following an acute 1 h transport stress which represented a transport stress from pet shop to home. For the measurements of water cortisol following husbandry procedures, females were removed from their tanks and placed in groups of three (n = 16 groups of three per husbandry treatment) into a
11
beaker containing 150 ml of clean water for 30 min based on the methods of Sebire et al. (2007). For the acute transport stress, groups of guppies (n = 9 groups of three) were held in a plastic bag containing 150 ml water, placed inside a dark box, and moved around the laboratory for 1 h prior to cortisol measurements. After this time, water cortisol concentration was extracted according to the methods of Ellis et al. (2004) and measured using a commercial ELISA (DRG Instruments). The commercial ELISA used a cortisol antibody sensitive to 2.5 ng ml−1 (6.9 nmol l−1), the average intra and inter assay variation was 5.6% and 5.9% respectively. Blank samples of tank system water were used to account for background concentrations of cortisol. All stressors and cortisol measurements were conducted at the same time of day to control for circadian fluctuations in endocrine status. Offspring behavioural tests Plus-maze test The four-armed radial plus-maze (10 l) supplied with system water (26.9 ± 0.1 °C), was based on the design of Sison and Gerlai (2010) and made of white, opaque polypropylene. Each arm was 30 cm long, 10 cm wide and 10 cm high and arms were connected by a central 10 × 10 cm central square. At the end of each individual arm there was a section of tubing connected to an external syringe loaded with bloodworm. Fish were held for 10 s within the central square in a ‘start-box’, an opentopped transparent container attached to a pulley system to acclimate to the maze environment before the start-box was remotely removed. Fish were subjected to five habituation trials in the plus maze, one per day on consecutive days, in decreasing group sizes and times. In the first of these trials, six fish were placed into the start-box, in single treatment groups, and then allowed to explore the plus-maze for 1 h, with food released from all arms after 30 min. The second trial allowed groups of three fish to explore the maze for 30 min with food administered after 15 min in all arms. The third allowed pairs of fish to explore the maze for 20 min with food administered after 10 min and the fourth and fifth, single fish to explore the maze alone for 10 min with food administered in all arms after 5 min. Fish were fed at no other time other than within the plus-maze. After habituation trials, each fish spent 10 min individually within the maze twice daily for 10 days, once in the morning and once in the afternoon. Each period within the plus-maze was recorded from above using a mounted camcorder (Panasonic SDR-S50). A large piece of red card with a hole in the centre was placed at the end of one arm with the food-dispensing tubing pushed through. Food was administered after 5 min of fish being within the plus-maze and only from this colour-coded arm (target-arm), never from non colour-coded arms (non-target arms). All arms, target and non-target, were loaded with bloodworm to ensure that any scent of food leaked equally from each. The plus-maze was curtained with black cloth to prevent any external features acting as guides to the target arm. The order in which fish were subjected to plus-maze trials and the orientation of the target-arm was randomised each day. Maze performance was analysed using JWatcher (http://www. jwatcher.ucla.edu/). The percentage of time spent within the target arm and the percentage of entries made to the target arm were recorded for 10 min after the fish was released from the start-box. Arm entries were recorded as soon as a fish entered that specific arm, if a fish exited an arm and remained within the central square, it was considered to still be within the last arm visited until it entered a different arm. Each fish completed the plus-maze 20 times over a period of 10 days. Mirror image test Mirror image tests have previously been used as a measure of competitive ability since fish respond to their mirror image as if it were another individual (Brick and Jakobsson, 2002; Sloman and Baron, 2010). At approximately 18 weeks of age, mature single males (n = 18) from each maternal treatment were randomly selected for
12
L. Eaton et al. / Hormones and Behavior 71 (2015) 10–15
mirror image tests. Mature males were used in this experiment, as opposed to mixed sex juveniles for plus-maze tests, since mature males are aggressive towards other males and would therefore interact with a mirror image to a greater extent than immature or female fish. Fish were placed individually in static aerated 2 l tanks with a covered mirror, and were left to acclimate for 20 h. The mirror was mounted vertically on the side facing into the tank. The mirror was then uncovered for 1 min and re-covered for a further hour to acclimate the fish to the action of mirror uncovering. Following this, the mirror was uncovered and fish were recorded with a camcorder (Panasonic SDR-S50) for a period of 30 min (Sloman, 2010). The number of behavioural interactions with their mirror image (defined as a bite, a charge or touching the mirror) as well as the latency to touch the mirror was determined. Video analysis of behaviour was randomised to prevent observer bias, all observations were carried out by LE. Statistical analyses Data were analysed using PASW Statistics version 18.0 (http:// www.spss.com.hk/statistics/). Data were tested for normality (Kolmogorov–Smirnov: K–S) and homogeneity of variance, parametric or non-parametric statistics were then used where appropriate. Cortisol data were transformed prior to analysis (√Cortisol + 0.1), were normally distributed (K–S: P = 0.09); and thus analysed using a one-way ANOVA with LSD post-hoc testing. Percentage of time spent within and the percentage of entries made to the target-arm of the plus-maze were arcsine square root transformed, were normally distributed (K–S: P N 0.05) and analysed using repeated-measures (RM) ANOVAs with maternal stress treatment, time and replicate as factors. The interindividual coefficient of variation (CV) was calculated for both the percentage of time spent within and the percentage of entries made to the target-arm. Inter-individual coefficient of variation data were not normally distributed (K–S: percentage of time: P b 0.05; percentage of entries: P b 0.005), thus, a natural log transformation was applied and data were normalised (K–S: percentage of time: P = 0.31; percentage of entries: P = 0.82). Measures of inter-individual CV of the percentage of time spent within, and the percentage of entries made to the target arm were analysed using un-paired t-tests. Number of interactions with a mirror image were normally distributed (K–S: P = 0.06), data were therefore analysed using a two-way ANOVA. Latency to interact with the mirror image was not normally distributed (K–S: P b 0.001) and was analysed using a two-way Scheirer–Ray–Hare (S–R–H) test. For both the number of interactions with, and the latency to interact with the mirror image, maternal treatment and tank replicate were included as factors. Effect sizes were estimated using eta squared (η2) and Cohen's d for RM ANOVAs, Scheirer–Ray–Hare and t-test analyses. Effect sizes equal to or less than 0.2, 0.5 and 0.8 have been categorised as small, medium and large effect sizes according to Cohen (1988). Ethical note The study was approved by the WALTHAM ethical committee and carried out in accordance with University of Plymouth and University of the West of Scotland ethical guidelines under a U.K. Home Office project licence to KAS. Fish were not subjected to any invasive techniques. Stress exposures comprised typical home-husbandry procedures and all fish had access to environmental enrichment. All fish were held under appropriate water quality conditions. Results Maternal cortisol A significantly higher concentration of water-borne cortisol was measured when fish were subjected to the simulated transport stress compared to the control or mild husbandry stress (Fig. 1: K–S: 0.09;
Fig. 1. Cortisol concentrations. Mean (±SEM) total water-borne concentration (ng/fish/h) from gestating female guppies exposed to a control treatment, a mild husbandry stressor and a simulated transport stress. Asterisk denotes significant between treatments (ANOVA; LSD).
one-way ANOVA: F1,38 = 14.378, P b 0.001, η2 = 0.39, post-hoc LSD). When cortisol excretion rates were analysed independently of the transport stress data, there was still no statistical significance between treatments (K–S: P = 0.30; one-way ANOVA: F1,28 = 0.146, P = 0.767, η2 = 0.001), Offspring behaviour Offspring whose mothers were subjected to mild stress treatments prior to and during gestation spent a lower percentage of time in the target arm than control offspring (Fig. 2A: RM ANOVA: F1,30 = 70.79, P b 0.001, η2 = 0.86). There was a significant interaction between treatment and repeated trials i.e. time (F19,30 = 435.00, P b 0.001, η2 = 0.12); further analysis showed that offspring from control treatments increased the percentage of time spent within the target arm with repeated trials (Fig. 2A: RM ANOVA: F19,323 = 4.21, P b 0.001, η2 = 0.15) whereas offspring from stress treatment did not (RM ANOVA: F19,323 = 0.87, P = 0.618, η2 = 0.03). Offspring from the stress treatment also made a lower percentage of entries to the target arm of a plus maze than controls (Fig. 2B: RM ANOVA: F1,30 = 76.87, P b 0.001, η2 = 0.31); again when considered separately, offspring from the control treatment increased the percentage of entries made to the target arm with repeated trials (Fig. 2B: RM ANOVA: F19,323 = 5.25, P b 0.001, η2 = 0.16) whereas offspring from the stress treatment did not (F19,340 = 0.73, P = 0.79, η2 = 0.02). For both percentage of time spent in the target arm and percentage of entries to the target arm, there was a significant effect of replicate i.e. origin of the fish and a significant interaction with treatment (RM ANOVA: percentage time: F2,30 = 57.04, P b 0.001; stress* replicate: ANOVA: F2,30 = 5.16, P b 0.05; number of entries: F2,30 = 73.15, P b 0.001; stress* replicate: ANOVA: F2,30 = 11.98, P b 0.001). Further analysis of each replicate separately showed that the significant effect of maternal stress treatment was consistent across all replicates individually (RM: ANOVA P b 0.01), but that the magnitude of response varied between these different batches of fish. Mild maternal stress increased the inter-individual CV of time spent within the target-arm (Fig. 3: un-paired t-test: df = 118, t = − 4.75, P b 0.01, Cohen's d = 0.86), but there was no difference in interindividual CV for entries made to the target-arm (un-paired t-test: df = 118, t = − 1.08, P = 0.281, Cohen's d = 0.69). Offspring from the stress treatment interacted more with their mirror image than those from the control treatment (Fig. 4: two-way ANOVA: F1,52 = 19.03, P b 0.001, η2 = 0.22) with a significant effect of tank replicate (F2,52 = 4.36, P b 0.05 η2 = 0.10) and interaction (F2,52 = 3.67,
L. Eaton et al. / Hormones and Behavior 71 (2015) 10–15
13
different batches of fish. The latency for fish to interact with a mirror image was not affected by maternal stress (S–R–H: df = 1, SS = 9.41, P = 0.93) and there was no significant effect of tank replicate (S–R–H: df = 1, SS = 1577.51, P = 0.5) or maternal stress*tank replicate interaction (S–R–H: df = 2, SS = 37.7, P = 0.98).
Discussion
Fig. 2. Maze performance. Mean (±SEM) percentage of (A) time spent in and (B) entries made to the target arm of a plus maze by offspring from control and stress treatments over 20 trials. n = 18 fish per treatment.
P b 0.05, η2 = 0.08). Further analysis of each replicate showed the effects of maternal stress on offspring aggression to be consistent across all replicates (P b 0.05), but the magnitude of difference varied between
Fig. 3. Maze performance variability. Mean (±SEM) inter-individual coefficient of variation of time spent within and entries made to the target arm of the plus maze. Asterisk denotes significance between treatments (un-paired t-tests). n = 18 fish per treatment.
Offspring from mothers which experienced prenatal mild stress (both pre-conception and during gestation) did not exhibit improved performance over time in an associative learning task. In contrast, offspring from control mothers learnt to associate a colour cue with a food reward. Previous studies investigating the consequences of maternal stress have focussed on substantial stress events to the mother and have not addressed the possibility that mild stressors could have long-term effects on offspring behaviour. To our knowledge, this is also the first study to demonstrate the effects of maternal stress on offspring behaviour in a viviparous fish which presents an interesting intermediate between studies in mammals and oviparous animals. The lack of associative learning by offspring from mothers experiencing mild stress may not necessarily be disadvantageous. While there is a tendency for phenotypic responses in offspring attributed to maternally-derived stress to be interpreted as negative, there is increasing evidence that maternal stress can have adaptive benefits (Love et al., 2013; Sheriff and Love, 2013). Strong association between a visual cue and food reward may be disadvantageous if either the cue is false or the environment highly variable. Therefore, lack of associative learning, or capacity to forget, may be an adaptive mechanism (Warburton, 2003). In the present study, mild maternal stress involved the movement of environmental enrichment each time the tank was cleaned, thus mothers experienced a variable environment and this information may have been transferred to their offspring. Giesing et al. (2011) demonstrated that female sticklebacks can transfer information regarding predation levels in their environment to embryos and hypothesised that this might occur through maternal cortisol transfer to eggs. Similarly, in gold belly minnows, light intensity of maternal environment altered expression of lateralisation in offspring in both visual and motor tasks (Dadda and Bisazza, 2012), and again the authors hypothesised that this might be mediated through maternal cortisol. Even though we did not detect a significant elevation in cortisol between maternal stress treatments in the present study, the involvement of cortisol in mediating these effects remains likely but at much lower levels than previously documented. It seems unlikely that the water cortisol concentrations measured in the present study represent true basal levels; a drawback to this non-
Fig. 4. Interactions with a mirror image. Mean (±SEM) number of interactions with a mirror image. Asterisk denotes significance between treatments (two-way ANOVA). n = 18 fish per treatment.
14
L. Eaton et al. / Hormones and Behavior 71 (2015) 10–15
invasive methodology for measuring stress hormones is that to detect such low levels fish are confined in a relatively small volume of water which can represent a confinement stress. However, fish from all treatments experienced the same methodology making comparisons of water cortisol between treatments possible as has been done in other studies (Ellis et al., 2007; Scott et al., 2008). It is also possible that fish experiencing husbandry stressors repeated over seven weeks could acclimate to the stress (Wong et al., 2008). Our measurements of cortisol were taken immediately after the first husbandry stressor experience, it was therefore assumed at this point the rise in cortisol concentrations would be the greatest due to exposure to an unfamiliar stressor with no acclimation period. Inter-individual variation in the percentage of time spent within the target arm was significantly higher in offspring from the stress treatment, which may also be beneficial in an unpredictable environment. ‘Bet-hedging’, where mothers increase phenotypic variability among offspring to ensure survival of at least some individuals in a changeable environment has been suggested (Gagliano and McCormick, 2007; Philippi and Seger, 1989). In birds, variable maternal investment to eggs within the same clutch, including concentrations of maternal hormones, may create degrees of competitive asymmetry between individuals (Muller and Groothuis, 2013). Whether such intentional intrauterine manipulation of offspring can occur in viviparous fishes is unknown. Offspring from mothers exposed to a mild stressor performed more bites towards their own mirror image than offspring from control mothers. Being born into an unpredictable environment may necessitate greater competition for food and mates which may be aided by higher innate levels of aggression. Higher levels of aggression have also been demonstrated in juvenile brown trout exposed pre-fertilisation to synthetic cortisol (Sloman, 2010). Interestingly, however, Burton et al. (2011) showed that exposure of brown trout embryos to synthetic cortisol immediately after fertilisation produced offspring with reduced competitive abilities suggesting that timing of exposure to maternal glucocorticoids is critical in shaping offspring behaviour. The mammalian literature traditionally divides the effects of maternal stress into pre-natal and post-natal (Love et al., 2013), since immediate post-natal motheroffspring interactions can have profound effects on offspring behaviour (Sheriff and Love, 2013). No parental care occurs in the guppy, so maternal stress can only act prenatally, which can be further subdivided into pre-conception and gestation periods. In the present study mothers were stressed both pre- and post-fertilisation. If maternal cortisol effects on offspring behaviour in viviparous fish mirror those seen in oviparous fishes where pre- and post-fertilisation cortisol exposure increases and decreases offspring aggression respectively, it could be hypothesised that when both pre- and post-fertilisation stressors occur that effects of pre-fertilisation exposure dominate. However, the importance of exposure timing in viviparous fishes requires further investigation. In mammals, exposure of adults to elevated glucocorticoids reduces neurogenesis (Gould et al., 1997; Kozorovitskiy and Gould, 2004) and similar effects have been seen in fish (Sørensen et al., 2011), which could underlie behavioural changes in cognition. In oviparous fish, the HPI axis does not become functional until after hatching (Alsop and Vijayan, 2008; Auperin and Geslin, 2008) and in mammals the HPA axis becomes functional around mid-gestation (Tegethoff et al., 2009). Exactly when the HPI axis becomes functional in viviparous fishes is currently unknown. Given that exposure to maternal stress hormones can exert effects pre-fertilisation (Sloman, 2010), it seems unlikely that these behavioural changes are a direct result of interference with the HPI axis (Li et al., 2010). Within the mammalian literature there is increasing evidence for epigenomic programming via early life stress exposure (Li and Leatherland, 2012a). In bonnet macaques (Macaca radiata), enhanced behavioural stress reactivity following early life stress was associated with epigenetic changes (Kinnally et al., 2011) and unpredictable maternal separation during postnatal development in mice induced aggressive-like behaviours and altered DNA methylation
of several candidate genes (Franklin et al., 2010). In fish, there is evidence that cortisol exposure of oocytes causes changes in expression patterns of genes associated with the innate immune system and growth (Li et al., 2010; Li and Leatherland, 2012b). Thus, it seems increasingly likely that maternal effects of cortisol very early in life act via epigenetic mechanisms, most likely before the HPI/A axis is functional. The pronounced behavioural effects seen in offspring as result of this routine husbandry stressor highlight that even relatively mild changes in maternal environment can have concomitant effects on offspring phenotype. Whether or not these effects are advantageous will depend on the environmental conditions they encounter (McCormick, 1998, 1999, 2006, 2009; McCormick and Nechaev, 2002; Sheriff and Love, 2013). The mild maternal stress used in the present study represented a husbandry procedure used in home aquaria. Many animals are bred in captivity for a variety of reasons including as companion animals, for stocking programmes and for research. Therefore, the present study also demonstrates the importance of carefully considering the exposure of captive mothers to routine mild stressors as they have the potential to alter offspring phenotype. Further work is needed to address the specific mechanisms through which maternal stress acts on offspring behaviour and to combine work with oviparous and viviparous animals to investigate the effects of timing of maternal stress. Acknowledgments We thank Stanley McMahon, Deirdre Galbraith and Edith Burns for fish husbandry and Rick Preston for maze construction. The project was generously supported by a WALTHAM Foundation grant to KAS. References Alsop, D., Vijayan, M.M., 2008. Development of the corticosteroid stress axis and receptor expression in zebrafish. Am. J. Physiol. 294, R711–R719. Auperin, B., Geslin, M., 2008. Plasma cortisol response to stress in juvenile rainbow trout is influenced by their life history during early development and by egg cortisol content. Gen. Comp. Endocrinol. 158, 234–239. Bernardo, J., 1996. Maternal effects in animal ecology. Am. Zool. 36, 83–105. Brick, O., Jakobsson, S., 2002. Individual variation in risk taking: the effect of a predatory threat on fighting behavior in Nannacara anomala. Behav. Ecol. 13, 439–442. Burton, T., Hoogenboom, M.O., Armstrong, J.D., Groothuis, T.G.G., Metcalfe, N.B., 2011. Egg hormones in a highly fecund vertebrate: do they influence offspring social structure in competitive conditions? Funct. Ecol. 25, 1379–1388. Cohen, J., 1988. Statistical Power Analysis for the Behavioural Sciences. Routledge Academic, New York, NY. Dadda, M., Bisazza, A., 2012. Prenatal light exposure affects development of behavioural lateralization in a livebearing fish. Behav. Process. 91, 115–118. de Kloet, E.R., Sibug, R.M., Helmerhorst, F.M., Schmidt, M., 2005. Stress, genes and the mechanism of programming the brain for later life. Neurosci. Biobehav. Rev. 29, 271–281. Donaldson, J.M., Munday, P.L., McCormick, M.I., 2008a. Parental effects on offspring life histories: when are they important? Biol. Lett. 5, 262–265. Donaldson, J.M., McCormick, M.I., Munday, P.L., 2008b. Parental condition affects early life-history of a coral reef fish. J. Exp. Mar. Biol. Ecol. 360, 109–116. Ellis, T., James, J.D., Stewart, C., Scott, A.P., 2004. A non-invasive stress assay based upon measurement of free cortisol released into the water by rainbow trout. J. Fish Biol. 1233–1252. Ellis, T., James, J.D., Sundh, H., Fridell, F., Sundell, K., Scott, A.P., 2007. Non-invasive measurement of cortisol and melatonin in tanks stocked with seawater Atlantic salmon. Aquaculture 698–706. Eriksen, M.S., Bakken, M., Espmark, A., Braastad, B.O., Salte, R., 2006. Prespawning stress in farmed Atlantic salmon Salmo salar: maternal cortisol exposure and hyperthermia during embryonic development affect offspring survival, growth and incidence of malformations. J. Fish Biol. 69, 114–129. Evans, J.P., Magurran, A.E., 2000. Multiple benefits of multiple mating in guppies. Proc. Natl. Acad. Sci. U. S. A. 97, 10074–10076. Franklin, T.B., Russig, H., Weiss, I.C., Gräff, J., Linder, N., Michalon, A., Vizi, S., Mansuy, I.M., 2010. Epigenetic transmission of the impact of early stress across generations. Biol. Psychiatry 68, 408–415. Gagliano, M., McCormick, M.I., 2007. Maternal condition influences phenotypic selection on offspring. J. Anim. Ecol. 76, 174–182. Giesing, E.R., Suski, C.D., Warner, R.E., Bell, A.M., 2011. Female sticklebacks transfer information via eggs: effects of maternal experience with predators on offspring. Proc. R. Soc. B Biol. Sci. 278, 1753–1759. Gould, E., McEwen, B.S., Tanapat, P., Galea, L.A.M., Fuchs, E., 1997. Neurogenesis in the dentate gyrus of the adult tree shrew is regulated by psychosocial stress and NMDA receptor activation. J. Neurosci. 17, 2492–2498.
L. Eaton et al. / Hormones and Behavior 71 (2015) 10–15 Huizink, A.C., Mulder, E.J.H., Buitelaar, J.K., 2004. Prenatal stress and risk for psychopathology: Specific effects or induction of general susceptibility? Psychol. Bull. 130, 115–142. Janczak, A.M., Braastad, B.O., Bakken, M., 2006. Behavioural effects of embryonic exposure to corticosterone in chickens. Appl. Anim. Behav. Sci. 96, 69–82. Kinnally, E.L., Feinberg, C., Kim, D., Ferguson, K., Leibel, R., Coplan, J.D., Mann, J.J., 2011. DNA methylation as a risk factor in the effects of early life stress. Brain Behav. Immun. 25, 1548–1553. Kofman, O., 2002. The role of prenatal stress in the etiology of developmental behavioural disorders. Neurosci. Biobehav. Rev. 26, 457–470. Kozorovitskiy, Y., Gould, E., 2004. Dominance hierarchy influences adult neurogenesis in the dentate gyrus. J. Neurosci. 24, 6755–6759. Li, M., Leatherland, J.F., 2012a. The implications for aquaculture practice of epigenomic programming of components of the endocrine system of teleostean embryos: lessons learned from mammalian studies. Fish Fish. 14 (4), 528–553. Li, M., Leatherland, J.F., 2012b. The interaction between maternal stress and the ontogeny of the innate immunity during fish embryogenesis: implications for aquaculture practice. J. Fish Biol. 81, 1793–1814. Li, M., Bureau, D.P., King, W.A., Leatherland, J.F., 2010. The actions of in ovo cortisol on egg fertility, embryo development and the expression of growth-related genes in rainbow trout embryos, and the growth performance of juveniles. Mol. Reprod. Dev. 77, 922–931. Love, O.P., McGowan, P., Sheriff, M.J., 2013. Maternal adversity and ecological stressors in natural populations: the role of stress axis programming in individuals, with implications for populations and communities. Funct. Ecol. 27, 81–92. McCormick, M.I., 1998. Behaviorally induced maternal stress in a fish influences progeny quality by a hormonal mechanism. Ecology 79, 1873–1883. McCormick, M.I., 1999. Experimental test of the effect of maternal hormones on larval quality of a coral reef fish. Oecologia 118, 412–422. McCormick, M.I., 2006. Mothers matter: crowding leads to stressed mothers and smaller offspring in marine fish. Ecology 87, 1104–1109. McCormick, M.I., 2009. Indirect effects of heterospecific interactions on progeny size through maternal stress. Oikos 118, 744–752. McCormick, M.I., Nechaev, I.V., 2002. Influence of cortisol on developmental rhythms during embryogenesis in a tropical damselfish. J. Exp. Zool. 293, 456–466. Mousseau, T.A., Fox, C.W., 1998. The adaptive significance of maternal effects. Trends Ecol. Evol. 13, 403–407. Muller, M., Groothuis, T.G.G., 2013. Within-clutch variation in yolk testosterone as an adaptive maternal effect to modulate avian sibling competition: evidence from a comparative study. Am. Nat. 181, 125–136.
15
Philippi, T., Seger, J., 1989. Hedging one's evolutionary bets, revisited. Trends Ecol. Evol. 4, 41–44. Scott, A.P., Hirschenhauser, K., Bender, N., Oliveira, R., Earley, R.L., Sebire, M., Ellis, T., Pvalidis, M., Hubbard, P.C., Huertas, M., Canario, A., 2008. Non-invasive measurements of steroids in fish-holding water: important considerations when applying the procedure to behaviour studies. Behaviour 145, 1307–1328. Sebire, M., Katsiadaki, I., Scott, A.P., 2007. Non-invasive measurement of 11-ketotestosterone, cortisol and androstenedione in male three-spined stickleback (Gasterosteus aculeatus). Gen. Comp. Endocrinol. 152, 30–38. Sheriff, M.J., Love, O.P., 2013. Determining the adaptive potential of maternal stress. Ecol. Lett. 16, 271–280. Sison, M., Gerlai, R., 2010. Associative learning in zebrafish (Danio rerio) in the plus maze. Behav. Brain Res. 207, 99–104. Sloman, K.A., 2010. Exposure of ova to cortisol pre-fertilisation affects subsequent behaviour and physiology of brown trout. Horm. Behav. 58, 433–439. Sloman, K.A., Baron, M., 2010. Conspecific presence affects the physiology and behaviour of developing trout. Physiol. Behav. 99, 599–604. Sørensen, C., Bohlin, L.C., Øverli, O., Nilsson, G.E., 2011. Cortisol reduces cell proliferation in the telencephalon of rainbow trout (Oncorhynchus mykiss). Physiol. Behav. 102, 518–523. Takahashi, L.K., Turner, J.G., Kalin, N.H., 1998. Prolonged stress-induced elevation in plasma corticosterone during pregnancy in the rat: implications for prenatal stress studies. Psychoneuroendocrinology 23, 571–581. Tegethoff, M., Pryce, C., Meinlschmidt, G., 2009. Effects of intrauterine exposure to synthetic glucocorticoids on fetal, newborn, and infant hypothalamic–pituitary–adrenal axis function in humans: a systematic review. Endocr. Rev. 30, 753–789. Vallée, M., Maccari, S., Dellu, F., Simon, H., Le Moal, M., Mayo, W., 1999. Long-term effects of prenatal stress and postnatal handling on age-related glucocorticoid secretion and cognitive performance: a longitudinal study in the rat. Eur. J. Neurosci. 11, 2906–2916. Warburton, K., 2003. Learning of foraging skills by fish. Fish Fish. 4, 203–215. Weinstock, M., 2005. The potential influence of maternal stress hormones on development and mental health of the offspring. Brain Behav. Immun. 19, 296–308. Wong, S.C., Dykstra, M., Campbell, J.M., Earley, R.L., 2008. Measuring water-borne cortisol in convict cichlids (Amatitlania nigrofasciata): is the procedure a stressor? Behaviour 145, 1283–1305.
