Mothers modify the cost of reproduction by dynamic changes in antioxidant function and filial cannibalism

Abstract

Investment in current reproduction may negatively influence subsequent fitness. Oxidative stress has been proposed as a potential mediator of this trade-off between current and future reproductive success. However, evidence of reproduction causing oxidative stress is limited, possibly owing to compensatory mechanisms that counteract oxidative insults. Here we test the idea that organisms protect against oxidative challenges through a dynamic interaction between behavioural and physiological adjustments at different stages of reproduction. To test this idea, we manipulated maternal care in the mouthbrooding cichlid fish Astatotilapia burtoni by allowing females to continue care (brooders) or by preventing care (non-brooders). We found that brooders depleted the pool of antioxidants as brood care progressed; however, we only observed increased oxidative DNA damage at the early stage of care relative to non-brooders, possibly owing to upregulated antioxidant protection during later stages of care. Most brooders adjusted parental investment by consuming some of their offspring during mouthbrooding. Intriguingly, the level of filial cannibalism was positively related to liver antioxidant function. These changes in antioxidant function and filial cannibalism allow parents to manage the cost of reproduction and are important for our understanding of how oxidative stress mediates life-history trade-offs.

1. Introduction

In many animal species, parents are faced with a fundamental decision: how to allocate resources optimally when investment in current reproduction can come at the expense of subsequent fitness [1–3]. Parental care entails substantial costs that may negatively impact the health and future reproductive success of the parents. Consequently, the degree of parental care in a particular breeding attempt may be adjusted to both the parent's physiological state or the prevailing environmental conditions [4,5]. For example, parents may reduce care or even abandon their brood under unfavourable conditions in favour of of future breeding opportunities when environmental conditions improve [6]. The physiological costs of reproduction have received considerable attention in a range of animal taxa; however, how these costs are managed is often unclear [7–9].

Oxidative stress, which occurs when the production of reactive oxygen species (ROS) overwhelms the antioxidant systems, has been suggested as a potential mediator of life-history trade-offs [10–13]. The high metabolic cost of parental care increases the production of ROS, potentially giving rise to oxidative stress [10,14]. Since oxidative stress is linked to disease risk, accelerated ageing and reduced reproductive potential [15], it may mediate the trade-off between reproduction and somatic maintenance, resulting in a negative association between reproductive effort and lifespan [10]. There is evidence that parental care increases oxidative stress [16,17], but many studies have found no effect or even reduced oxidative damage in parents that were forced to increase care [7,8,18]. This is due to animals managing or minimizing potential oxidative challenges of reproduction [12,19]. Therefore, identifying potential strategies that animals use to cope with oxidative challenges may provide important insights into the role of oxidative stress in mediating life-history trade-offs.

In the cichlid fish Astatotilapia burtoni, females exhibit maternal care after spawning in the form of mouthbrooding [20–22]. During mouthbrooding, females hold their fertilized eggs and developing embryos in the buccal cavity for approximately two weeks. During this time, the female is unable to consume food and consequently experiences obligatory starvation [21,23]. The lack of feeding may increase ROS production [24] and prevent the female from replacing energy reserves and dietary antioxidants, which may compromise the mother's ability to maintain redox homeostasis. However, in a previous study, we found that the number of embryos varied significantly across stages of mouthbrooding (electronic supplementary material, figure S1), suggesting that mothers may ingest some of their offspring during mouthbrooding, as has been reported for other cichlid fish [25–27]. Consumption of nutrient-rich offspring may boost antioxidant function in the parent and indicates reduced parental care in favour of somatic maintenance.

Consistent with the high demands of reproduction in this species, we have previously found that egg laying and mouthbrooding elevated oxidative stress [28]. To isolate the physiological cost of maternal care (mouthbrooding) from the cost of egg production, we randomly assigned females to either a brooding (care) or a non-brooding (no-care) treatment after spawning and collected tissue at three stages of mouthbrooding (early, middle and late) to assess changes in oxidative damage and antioxidant function. In the brooding females, we quantified the number of fry left in the mouth to assess how oxidative balance is linked to the rate of filial cannibalism. We predicted that brooding females would have higher levels of oxidative stress than their non-brooding counterparts, and we predicted that this effect would become more pronounced with increasing brood care duration. Since fry may supplement antioxidant resources, we also predicted that the rate of filial cannibalism would be positively linked to antioxidant capacity.

2. Material and methods

A detailed Material and methods section can be found in the electronic supplementary material.

(a) . Animals and experimental design

Animals used in this study were Astatotilapia burtoni bred from a laboratory stock population originating from Lake Tanganyika, Africa [29].

Females were housed in six mixed-sex groups. We observed fish daily in the morning to identify females that had just spawned and initiated mouthbrooding. Once a brooding female was identified, she was randomly assigned to either the brooding treatment or the non-brooding treatment (figure 1). In the brooding treatment, we removed eggs from the mother's buccal cavity (number of eggs: mean ± s.e. = 63.7 ± 2.8, n = 31) and returned 25 of her eggs with a plastic pipette before returning her to the tank [30]. The same procedure was followed for females assigned to the non-brooding treatment (number of eggs: mean ± s.e. = 64.0 ± 4.3, n = 32) to mimic handling but no eggs were returned to the buccal cavity of these females. Brooding and non-brooding females were sampled for tissue collection 2 days (early-stage), 6 days (middle-stage) or 12–15 days (late-stage) after egg manipulation. We returned eggs into the mouth of 43 females (all brood stages combined) and 31 of these held their brood until the assigned sampling date. The remaining 12 females that did not hold their brood were not sampled, and we note that this may have influenced the main findings of this paper. We also sampled nine control females from the same groups from which we sampled the manipulated females. Control females were not handled and were not brooding at the time of sampling.

Figure 1.

Mouthbrooding impacts oxidative stress and body mass/condition in a time-dependent manner. (a) Oxidative stress and (b) body mass and body condition in brooding (yellow) and non-brooding (blue) females at different stages of brood care. Shown are the means ± s.e. For statistics, see table 1. ⁽*⁾p < 0.1, *p < 0.05, **p < 0.01, *** p < 0.001. ROMs: reactive oxygen metabolites; 8-OHdG: 8-hydroxy-2′-deoxyguanosine; TAC: total antioxidant capacity; TE: Trolox equivalent; SOD: superoxide dismutase.

At tissue collection, we removed the eggs or fry from the female's buccal cavity and took body weight and standard length measurements. Blood and liver tissues samples were collected, processed and stored at −80°C as described previously [31]. The ovaries were dissected, and we visually determined gonadal maturation on a scale from 1 (eggs are less than 0.5 in diameter) to 5 (eggs are ready to be laid). We calculated body condition as weight/(length)³ × 100 [32].

(b) . Filial cannibalism

Eggs and fry left in the mouth of the mother at tissue sampling were counted in mouthbrooding females. To calculate the number of fry that disappeared as an estimate of filial cannibalism, we subtracted this number from 25 since 25 eggs were returned to the buccal cavity of each female in the brooding treatment.

(c) . Oxidative stress

We following established procedures [33,34] to evaluate oxidative stress in blood plasma and liver samples. We measured plasma reactive oxygen metabolites (ROMs) as a measure of overall oxidative damage. Plasma total antioxidant capacity (TAC [35]) was measured as an indicator of overall antioxidant protection. We measured superoxide dismutase (SOD) activity since this enzymatic antioxidant can be upregulated in response to oxidative insults [15]. As a more specific marker of oxidative damage, we assessed oxidative DNA damage by measuring levels of 8-hydroxy-2′-deoxyguanosine (8-OHdG). TAC, SOD and oxidative DNA damage were measured in the liver because metabolic processes in the liver often vary with life-history stage.

(d) . Statistical analysis

The statistical package R v3.6.1 [36] was used for all statistical analyses of the data. We used linear mixed models (LMMs) to test whether brooding status of the female and stages of brood care influenced markers of oxidative stress. ‘Tank’ was used as a random effect. We used Tukey's multiple comparison post hoc tests for pairwise comparisons between groups. We analysed the effect of brood care duration on the proportion of young cannibalized by the mother using a generalized linear mixed model assuming a beta distribution. We then examined how markers of oxidative stress were linked to the rate of filial cannibalism using LMMs within each brooding stage.

3. Results

We observed dramatic changes in markers of oxidative stress across stages of brood care in both brooding and non-brooding females (figure 1a; electronic supplementary material, table S1). Markers of oxidative stress differed in both brooding and non-brooding females at various stages of brood care relative to control females (figure 1a; electronic supplementary material, table S2).

To isolate the oxidative cost of mouthbrooding, we tested for differences between brooding females and non-brooding females in each stage of brood care. We found that oxidative DNA damage in the liver was significantly higher in brooding females compared with non-brooding females, but in the early stage only (figure 1a; table 1). Plasma ROMs were not significantly higher in brooding females compared with non-brooding females at any stage of brood care (figure 1a; table 1). Plasma TAC but not liver TAC was lower in brooding females compared with their non-brooding counterparts, although this effect was only significant in middle-stage females and borderline non-significant in late-stage females (figure 1a; table 1). The enzymatic antioxidant SOD activity was significantly higher in the liver of middle- and late-stage brooding females compared with their non-brooding counterparts (figure 1a; table 1).

Table 1.

Statistical comparison between brooding and non-brooding females for each stage of brood care. Shown are the output results of Tukey's multiple comparison post hoc tests after implanting linear mixed models for each variable using brooding status (brooding (B) and non-brooding (NB) females) and brood care stage (early-, middle-, late-stage) as predictors. Samples sizes are shown for B and NB females. The estimated marginal means for NB females are subtracted from those for B females. Significance indicated same as in figure 1. ROMs: reactive oxygen metabolites; TAC: total antioxidant capacity; SOD: superoxide dismutase.

	estimate	d.f.	t-ratio	p-value	B	NB
plasma ROMs
early-stage	−1.17 ± 1.13	58.3	−1.038	0.3035	7	7
middle-stage	1.15 ± 0.93	60.9	1.238	0.2204	10	11
late-stage	−1.67 ± 0.93	61.0	−1.792	0.0780	11	10
liver DNA damage
early-stage	0.0253 ± 0.0088	62.5	2.865	0.0057**	10	10
middle-stage	0.0058 ± 0.0090	62.0	0.646	0.5204	9	10
late-stage	0.0075 ± 0.0088	62.3	0.860	0.3929	10	10
plasma TAC
early-stage	−588 ± 1555	58.1	−0.378	0.7067	8	8
middle-stage	−3450 ± 1374	59.1	−2.510	0.0148*	10	11
late-stage	−2731 ± 1377	59.3	−1.984	0.0519	11	10
liver TAC
early-stage	0.0222 ± 0.0727	62.3	0.306	0.7607	9	10
middle-stage	0.0850 ± 0.0686	62.2	1.239	0.2200	10	11
late-stage	0.0714 ± 0.0679	62.7	1.052	0.2969	11	11
liver SOD
early-stage	−0.0100 ± 0.0217	63.0	−0.462	0.6460	9	10
middle-stage	0.0617 ± 0.0204	62.8	3.020	0.0037**	10	11
late-stage	0.0790 ± 0.0202	63.6	3.909	0.0002***	11	11

We found that 29 out of 31 females (93.5%) lost at least some fry during the mouthbrooding phase (figure 2a). Lost fry were most likely cannibalized by the mother. The rate of filial cannibalism was higher at early stages of brooding (linear model, days of brood care: −0.00448 ± 0.00213, F_1,29 = 4.439, p = 0.044; electronic supplementary material, figure S2).

Figure 2.

The degree of cannibalism covaries with markers of antioxidant function. (a) The proportion of fry that disappeared (out of the original 25 eggs) and the rate of fry loss estimated in females that were sampled for tissue collection at various brood care stages. Shown are the means ± s.e. (b) Liver TAC and SOD as a function of the proportion of fry that disappeared at the time of tissue sampling. For statistics, see table 2. Solid lines are significant effects. The shaded area shows the 95% confidence interval of the regression line. For abbreviations see figure 1.

We found that the proportion of fry cannibalized was a positive predictor of two antioxidant measurements (liver TAC and liver SOD, figure 2b) in middle-stage and/or late-stage brooding females only (table 2). There was no such effect for the other measures of oxidative stress (data not shown, all p > 0.2).

Table 2.

Statistical results examining the effect of proportion of fry disappeared (cannibalized) on liver TAC and liver SOD for each stage of brood care using LMM (mouthbrooding females only). *p < 0.05, **p < 0.01. TAC: total antioxidant capacity; SOD: superoxide dismutase.

	estimate	d.f.	t-value	p=value
liver TAC
early-stage	0.0225 ± 0.1860	9.00	0.121	0.906
middle-stage	0.3547 ± 0.1272	5.91	2.789	0.032*
late-stage	0.520 ± 0.152	11	3.43	0.006**
liver SOD
early-stage	0.0473 ± 0.0311	9.00	1.522	0.162
middle-stage	0.0658 ± 0.0803	9.66	0.82	0.432
late-stage	0.1360 ± 0.0570	6.85	2.387	0.049*

4. Discussion

We predicted that maintaining oxidative balance becomes more challenging with increasing brood care duration since the female will have expended a considerable amount of energy caring for the developing embryos while under forced starvation [37,38]. Although the reduced plasma TAC in middle- and late-stage brooding females is consistent with this prediction, we only found evidence for increased oxidative damage in liver oxidative DNA damage in early-stage females only. It is possible that this early cost of mouthbrooding is due to metabolic adjustments to cope with an extended period of mouthbrooding and starvation. Our data also suggest that females were able to manage oxidative balance quite well as brood care progressed by upregulating SOD activity, which may have helped limit oxidative damage.

We found evidence for filial cannibalism in mouthbrooding A. burtoni females. Filial cannibalism enables a parent to recoup some of the lost energy and perhaps also gain dietary antioxidants from nutrient-rich offspring, both of which could promote protective or repair mechanisms against oxidative insults. In a separate study, we found some evidence that filial cannibalism is high at the beginning of mouthbrooding and that body condition may even increase at the early stages of mouthbrooding in unmanipulated females (electronic supplementary material, figure S1), consistent with the idea that cannibalism may benefit the mother. In addition, filial cannibalism may boost antioxidant function because dietary antioxidants promote the expression of endogenous antioxidants [39], allowing for an enhanced and more efficient antioxidant response. This potential for dietary antioxidants having a priming effect on the antioxidant response is supported by the observation that the proportion of fry cannibalized positively predicted SOD. Body condition was positively linked to the degree of cannibalism in middle-stage females only (LMM, final body condition: 1458.81 ± 414.42, t₁₀ = 3.520, p = 0.0055), which suggests that females that were in better condition also had a higher tendency to cannibalize offspring. These data support the idea that mothers adaptively sacrifice offspring to boost antioxidant function and their own somatic maintenance. However, it is also possible that oxidative status influences the tendency to engage in filial cannibalism. More experiments are needed to clarify the causal relationship.

In conclusion, we demonstrated time-dependent effects of parental care on oxidative balance, with only a transient elevation of oxidative damage at the beginning of mouthbrooding, presumably due to physiological adjustments to counteract the oxidative challenges of reproduction. Our data also suggest that mothers may modulate oxidative balance by adjusting the degree of filial cannibalism. Both antioxidant responses and reductions in parental investment have obvious and less obvious costs (see supplementary electronic material for extended discussion). Thus, future studies should focus on behavioural and physiological strategies to manage oxidative challenges associated with reproduction and their potential consequences for life-history evolution [40–42] rather than simply assessing whether or not reproduction results in oxidative stress.

Ethics

Cichlids were studied under protocols approved by the Animal Care and Use Committees of Central Michigan University (IACUC protocol 15–22). We attempted to reduce stress of the animals during the experiment by ensuring sufficient enrichment in the form of flowerpots and gravel. All fish were carefully monitored after the manipulation and no fish were injured or died after the tagging procedure or egg manipulation.

Data accessibility

Data and code are available in the Figshare repository: https://doi.org/10.6084/m9.figshare.20301729 [43].

Supplementary material is available online [44].

Authors' contributions

J.S.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, visualization, and writing—review and editing; P.D.D.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, supervision, visualization, and writing—original draft.

Both authors gave final approval for publication and agreed to be held accountable for the work performed herein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

This work was supported by a graduate student grant from the Office of Research and Graduate Studies at Central Michigan University to J.S.