Paternal predatory risk alters parental behavior and offspring phenotypes in biparental Brandt’s voles

Abstract

Paternal predation risk can program offspring phenotypes via maternal responses and epigenetic marks of spermatozoa. However, the processes and consequences of this experience in biparental species are unknown. Here, we examined how preconception and postconception paternal cat odor (CO) exposure affects anxiety-like behavior and antipredator response in Brandt’s voles (Lasiopodomys brandtii). We found that preconception paternal CO exposure inhibited maternal investment when offspring were raised by mothers alone, while postconception exposure increased paternal investment towards the offspring raised by both parents. The increased paternal behavior may be associated with an increasing grooming behavior received from their mates, which alleviated the anxiety-like behavior in CO-exposed males. Both paternal experiences increased the levels of anxiety-like behavior in adolescent offspring but differentially altered adult phenotypes. Specifically, adult females from preconception CO-exposed fathers spent less time in defensive concealing, whereas the offspring of postconception CO-exposed fathers showed more in response to acute cat urine exposure. Correspondingly, baseline corticosterone levels were decreased and increased in these offspring, respectively. Our results indicate that in biparental species, paternal predation risk exposure affects offspring phenotypes in pathway-dependent and age-specific manners and that only the presence of both parents can elicit adaptive responses to a high predation-risk environment.

Graphical Abstract

Graphical Abstract.

Phenotypic similarity between parents and offspring is caused by shared genes, whereas epigenetic variations arise from heritable nongenetic sources such as parental effects (Badyaev and Uller 2009; Merila and Hendry 2014). When the offspring phenotype is influenced by the phenotype or experienced environment of 1 or both parents (i.e., predation risk, low food availability, social instability, or weather) and not by the effect of direct genetic transmission, parental effects occur (Badyaev and Uller 2009; Wolf and Wade 2009). This nongenetic inheritance has been predominantly studied in mothers. For example, 1 classic study of water fleas (Daphnia cucullata) showed that offspring from mothers prenatally exposed to predator chemical cues grow larger helmets, promoting their survival (Agrawal et al. 1999). In wild cavies (Cavia porcellus), an unstable social environment experienced by mothers during pregnancy and lactation induced camouflage behavior in male offspring, whereas a stable social environment caused earlier reproductive behavior (Siegeler et al. 2017).

Predation risk is a strong selective pressure that can drive the behavioral and physiological adaptation of prey through nongenetic maternal effects (Agrawal et al. 1999; Storm and Lima 2010; Coslovsky and Richner 2011; Giesing et al. 2011; Bestion et al. 2014; St-Cyr and McGowan 2015; Cattelan et al. 2020); however, its transgenerational response induced by fathers has received little attention until recently. In three-spine sticklebacks (Gasterosteus aculeatus), Stein and Bell (2014) demonstrated that males exposed to predation risk (sculpin, Cottus spp.) produce offspring with smaller size, reduced body condition, and behavioral activity. Hellmann et al. (2020, 2021, 2023) further proved that prefertilization exposure of fathers to predation risk can be transmitted to offspring via sperm and decrease offspring survival against predators, increasing the activity of sons under high-perceived predation pressure, making them less active, causing higher cortisol in them following a simulated predator attack, and making them exhibit conserved paternal care response. In rodents, Korgan et al. (2016) exposed male rats to cat odor (CO) before conception, which influenced maternal investment and caused the development of anxiety-like behavior in offspring. Azizi et al. (2019) found increased plasma corticosterone (CORT) levels and anxiety in the juvenile offspring of male rats repeatedly exposed to a live cat. In C57BL/6J mice, male offspring from a father exposed to 2,4,5-trimethylthiazoline (TMT) before mating were more active, exhibited less anxiety-like behavior, and had decreased baseline plasma CORT levels compared with those in controls (Brass et al. 2020). Overall, paternal exposure to predation risk can affect the offspring phenotype; however, studies are rather limited and mostly focused on the model animals and the preconception period.

Increasing evidence suggests that preconception paternal stress impacts offspring phenotypes primarily through epigenetic signals transferred through sperm (e.g., DNA methylation, small noncoding RNAs, and histone posttranslational modifications) (Chan et al. 2018; Champagne 2020; Cunningham et al. 2021). However, many pathways of father-offspring influence are reportedly mediated by maternal responses, termed “paternally driven maternal effects.” Male phenotypes can influence the physiology and behaviors of females, which can affect offspring development. For example, female mice mated with males that have experienced social enrichment or food restriction showed increased licking/grooming towards their offspring (Mashoodh et al. 2012, 2018), whereas females mated with stressed males (subjected to 1-h restraint stressors or a 6-min forced swim daily for 6 weeks) gained less weight during pregnancy and provided less care towards their offspring (Mashoodh et al. 2023). Regarding the predation risk, only 1 study showed that female rats in a seminaturalistic environment rearing offspring from predator-odor-exposed males engaged in elevated licking/grooming and arched-back nursing (Korgan et al. 2016). Extensive research suggests that females can dynamically adjust the provision of prenatal and postnatal care towards offspring based on the quality of male mates; however, how paternal predation risk affects the parental investment and subsequent offspring phenotype, particularly in biparental species, is only partially understood.

In species where both parents provide care for their offspring, paternal experience could impact offspring development directly through father-offspring interaction. For example, a male California mouse (Peromyscus californicus) that experienced social defeat over 3 days before pairing exhibited increased paternal behavior towards their young (Kowalczyk et al. 2018). Furthermore, in these species, the presence of a pair-bonded partner may provide a buffer to alleviate paternal stress; even strong social contact could help individuals recover from a stressful experience. For example, the partner’s presence during the 1-h immobilization (IMO) stress reduced the levels of anxiety-like behavior in male and female Prairie voles (Microtus ochrogaster) elicited by this stress (Donovan et al. 2018), while the stress response of females caused by the IMO experience can be alleviated by social contact from their male partners (Smith and Wang 2014). Thus, whether paternal stress in these cases can affect the offspring phenotype remains unknown. A recent study revealed that when female mice were allowed to mate naturally with food-restricted males, they demonstrated maternal masking via increased prenatal and postnatal care for their offspring, which caused no differences in depression- or anxiety-like phenotypes compared with those of control fathers (Mashoodh et al. 2018). Whether maternal interaction with a father who has experienced predation risk in biparental species may lead to masking via “social buffering” or “reproductive compensation” is not understood.

The Brandt’s vole (Lasiopodomys brandtii) is a typical herbivorous rodent inhabiting the open plains of northern China, Mongolia, and the Trans-Baikal region in Russia. Field studies showed that they live in extended family groups, breeding seasonally and displaying a variable mating system ranging from promiscuity and polygyny to monogamy (Yu et al. 2004; Liu et al. 2013). In a family group, both parents provide care for their young, and juveniles may also assist in rearing younger siblings. In the wild, their predators include foxes, polecats (Mustela eversmanii), Pallas’ cats (Otocolobus manul), saker falcons (Falco cherrug), upland buzzards (Buteo hemilasius), and steppe eagles (Aquila nipalensis) (Samjaa et al. 2000). Similarly, our laboratory study showed that they exhibited the strongest fear and defensive responses to cat (Felis catus) and weasel (Mustela sibirica) urine and feces (Hegab et al. 2014b); thus, we used cat urine and feces as a source of predator odor in the following study. Recent studies have demonstrated that maternal CO exposure during adolescence, pregnancy, and the postpartum period has a transgenerational effect on the antipredator responses of their offspring (Gu et al. 2018; Pang et al. 2022; Wu et al. 2022). Pregnant voles repeatedly exposed to CO produce more offspring and have a higher female offspring ratio in adulthood (Gu et al. 2020). These offspring also exhibit changes in social behavior, increased anxiety-like behavior, and higher serum adrenocorticotropic hormone and CORT (Wu et al. 2023). Thus, this species provides an ideal model for investigating the transgenerational effects of paternal predatory risk.

In this study, we aimed to investigate how paternal predation risk alters offspring phenotypes in biparental Brandt’s voles. To this end, we exposed male voles to cat urine, rabbit urine (as the nonpredator odor sources), and distilled water (DW) for 18 consecutive days in preconception and postconception stages. We predicted that preconception and postconception paternal predation risk exposure would impact offspring phenotype in a pathway-dependent and behavior-specific manner in biparental Brandt’s voles.

Materials and Methods

Animal housing

All subjects were laboratory-reared offspring from a wild population of Brandt’s vole in Inner Mongolia, China. The nonbreeding animals were housed individually in 22 cm × 18 cm × 15 cm cages, and the breeding animals were housed in pairs in 46 cm × 35 cm × 19 cm polycarbonate cages. They were kept at 22 ± 3 °C under a 12-h L:D cycle (lights on at 07: 00) and provided with standard rodent chow (Yizheng Animal Biotechnology Co., Ltd., Yangzhou, China) and water ad libitum. All experimental procedures adhered to the Guide for the Care and Use of Laboratory Animals of China and were approved by the Institutional Animal Care and Use Committee of Yangzhou University (No. 202302051).

Odor source

The basic procedure for urine collection was performed in our previous studies (Wu et al. 2021, 2022). The cat urine was taken from an adult male domestic cat which was caught from the Wenhui campus of Yangzhou University. Rabbit urine was obtained from an adult male rabbit (Oryctolagus cuniculus) which was purchased from the Laboratory Animal Center of Nantong University. They were housed in 120 cm × 40 cm × 30 cm wire cages in different rooms and provided with water and food ad libitum. In order to eliminate variations among collecting sessions, urine was collected only once from a clean tray placed under the cage for 48 h. Feces or fur were subsequently filtered out, and the urine samples were stored at −20 °C for further use.

Experimental design

For the preconception exposure, 41 male naïve voles (50–60 g, at approximately 90 days of age) were used and randomly exposed to DW (n = 14), rabbit odor (RO, n = 12), and CO (n = 15) for 60 min daily for 18 consecutive days. The exposure procedure was identical to that in our previous studies (Gu et al. 2018; Wu et al. 2020, 2022, 2023). Briefly, the subjects were transported to the exposure room and individually habituated to the exposure apparatus (75 cm × 37 cm × 40 cm) for 15 min. The urine was thawed and diluted with 4 volumes of DW to ensure that the experiment would have sufficient odor resources. A cotton ball soaked with 1 mL of diluted urine, or DW, and 10 g of their diet were placed into 2 dishes and introduced into the apparatus. The subjects were individually exposed to these conditions for 60 min before returning them to their home cages. The apparatus was cleaned with a 70% alcohol solution and dried after each test. These procedures were repeated for 18 consecutive days. Subsequently, the males were mated with unfamiliar, sexual, and experimental naïve females. The female partners were checked daily and considered likely to mate and become pregnant only if the vaginal openings and plugs were present. The males were removed from the breeding cage, and pregnant females were housed individually. The number of females that produced pups was 9 for each treatment, and the averaged little size were 7.1 ± 0.6 (DW), 8.7 ± 0.4 (RO), 8.1 ± 0.4 (CO), respectively. On postpartum days (PP) 1, 3, 6, 9, 12, 15, and 18, the maternal behavior in the home cage was recorded (n = 8 for each treatment). On postnatal day 21 (PND 21), offspring were separated from their parents and housed individually. Experimental naïve offspring (no more than 2 males and 2 females; n = 16 for each group) were randomly selected from each litter to evaluate anxiety-like behavior at adolescence (PND 30 and 32) and adulthood (PND 90 and 92). An antipredator behavioral test was conducted between PND 94 and 100. Furthermore, 2 days after the last behavioral test, the offspring were sacrificed to measure the baseline CORT concentrations to analyze the underlying mechanism. The experimental procedure is illustrated in Figure 1A. All behavioral observations were conducted during the light cycle.

Figure 1.

Effects of preconception paternal predation risk on maternal investment. (A) Schematics of experimental procedures used in this study; (B) maternal investment (n = 8 for each treatment). DW, distilled water; RO, rabbit odor; CO, cat odor; PND, postnatal day; PP, postpartum day. Main effect of paternal treatment (*P < 0.05). Turkey’s post hoc test (*P < 0.05). The error bars represent ± standard error.

In the experiment of postconception exposure, 60 male naïve voles (90 days old) were paired with unfamiliar naïve females. When a possible pregnancy was detected, the males were randomly exposed to DW, RO, or CO (n = 20 for each treatment). The exposure procedure was identical to that mentioned above. On exposure days 1, 3, 6, 9, 12, 15, and 18, social interactions between the subjects and their mates were assessed immediately following odor exposure in their home cage. On the 2nd day after the continuous odor exposure, the anxiety-like behaviors of the subjects and their partners were observed in an open-field box. Subsequently, the voles were bred as normal until they gave birth. The numbers of females mated under the different treatments that produced pups were 10, 11, and 15, and the averaged little size were 7.5 ± 0.7 (DW), 8.0 ± 0.3 (RO), 8.7 ± 0.6 (CO), respectively. On PP 1, 3, 6, 9, 12, 15, and 18, paternal and maternal behavior was recorded in their home cage. After weaning, experimental naïve offspring were randomly selected from each litter and subjected to behavioral and neuroendocrinological tests (Figure 3A).

Figure 3.

Effects of postconception paternal predation risk on social interaction and parental investment. (A) Schematics of experimental procedures used in this study; (B) retreating of male subject following odor exposure; (C) licking/grooming of their female mates; (D) % time of male subject and (E) their female mates spent in the center of open field; n = 10 for each treatment; (F) representative trajectory diagram of the open-field test; (G) paternal investment; (H) maternal investment; n = 8 for each group. Main effect of paternal treatment (*P < 0.05). Turkey’s post hoc test (*P < 0.05, **P < 0.01). DW, distilled water; RO, rabbit odor; CO, cat odor; PND, postnatal day; PP, postpartum day. The error bars represent ± standard error.

Social interaction test

This test was observed in the home cage as described previously (Wu et al. 2013, 2021, 2023), with slight modifications. The subjects and their partners acted as stimuli for each other. To identify individual animals, male voles were marked by cutting a bundle of hair on their back. Following the odor exposure, male voles were returned to the home cage and placed away from their partners. Subsequently, the behaviors were video-recorded for 15 min and scored by a trained observer blinded to the treatment groups using the event-recording program BORIS (www.boris.unito.it) as follows: approaching, sniffing, attacking, chasing, staring, retreating, submission, licking/grooming, body contact (or huddling), sexual behaviors (mounts and lordosis for the males and females, respectively), and other nonsocial behaviors (self-grooming, digging, exploring, and inactivity).

Parental behavior test

A parental behavior test was conducted in the home cage. Maternal behaviors were recorded in the preconception exposure experiment, while paternal and maternal behaviors were recorded in the postconception exposure experiment. At the beginning of each test, all pups were removed from the home cage, and the nest was destroyed. After 10 s, the pups were returned and randomly placed away from the original nest, and the behavioral performances were video-recorded for 15 min. The following behaviors were scored using BORIS software (Wu et al. 2016; Cai et al. 2022): pup sniffing, pup retrieval, pup licking/grooming, hovering over pups, nest-building, and other behaviors. The parental investment was calculated as the time spent on pup-directed behavior.

Open-field and light/dark box tests

Locomotor activity and anxiety-like behavior were assessed in open-field (50 cm × 50 cm × 30 cm) and light/dark (45 cm × 30 cm × 30 cm) boxes, as conducted in our previous study (Wu et al. 2023). For the open-field test, the animals were placed individually in the center of the box, and their exploration was recorded for 5 min. The open-field box was divided into 16 quadrants (4 central and 12 peripheral), and the time spent in the central and peripheral areas, total distance traveled, and duration of inactivity were scored using the SuperMaze software (version 3.3, Shanghai Xinruan Information Technology Co., Ltd., China).

The light/dark box is a linear box divided into light (30 cm × 30 cm × 30 cm) and dark (15 cm × 30 cm × 30 cm) chambers using a black sheet with an opening at the bottom (7.5 cm × 7.5 cm). The animals were placed in the center of the light chamber with their backs towards the opening. The test lasted for 5 min. The time spent in the light and dark chambers and the number of transmissions between the 2 chambers were scored. The apparatus was cleaned using a 70% alcohol solution between tests.

Antipredator behavior test

The offspring were sequentially exposed to DW, RO, or CO to analyze the antipredator behavior. To eliminate interference of the former test to the later test, there was a 3-day interval between each test. The basic procedure was identical to that of the paternal odor exposure. The offspring were individually placed into the apparatus, and their behavioral responses were recorded for 15 min. The following behaviors were scored by BORIS as described in our previous study (Wu et al. 2022): sniffing odor sources, freezing, jumping, avoidance, concealing, head-out, vigilant rearing, self-grooming, and others.

Serum CORT level assays

The offspring were decapitated 48 h after the last behavioral test. Blood samples were collected immediately to measure baseline CORT concentrations. The serum was stored at − 20 ^oC following 30 min of placement and centrifugation. Serum CORT concentration was determined using the vole-specific enzyme-linked immunosorbent assay kits (JL21595, Jianglai Biological Science and Technology, Shanghai, China) according to the manufacturer’s instructions.

Statistical analyses

Statistical analyses were conducted using SPSS software (version 19.0; SPSS Inc., Chicago, IL, USA), and data visualization was performed using GraphPad Prism. For all statistical tests, we used α = 0.05. Repeated measures analysis of variance (ANOVA) was used to determine whether there was a main effect of odor exposure or time on social interaction and parental care and to identify potential interaction effects between 2 factors. One-way ANOVA was used to evaluate the group differences on different testing days and behavioral parameters in open-field and light/dark box tests, followed by Tukey’s Honestly Significant Difference (HSD) post hoc test. Two-way ANOVA was used to evaluate whether there was a main effect of odor exposure or sex on offspring behavioral and neuroendocrinological data and to identify potential interaction effects between 2 factors. Tukey’s HSD post hoc tests were conducted only if the main effects of treatment or interaction were found. Independent sample t-tests were used to clarify the main effects of sex. All data are presented as mean ± standard error (SEM).

Results

Effects of preconception paternal predatory risk exposure

When the subject’s mate gave birth, the maternal investment was measured to determine whether there was a reproductive compensation based on the male quality. Repeated measures ANOVA showed a main effect of odor exposure (F_{2, 21} = 5.412, P = 0.013), with preconception male exposure to RO (P = 0.013) and CO (P = 0.063) significantly reducing and marginally reducing maternal investment, respectively (Figure 1B). This reduction may be owing to decreased nest-building behavior (ANOVA: treatment: F_{2, 21} = 7.976, P = 0.003; treatment × day: F_{12, 126} = 2.189, P = 0.016; Turkey’s test: DW > CO, P = 0.032, 0.035, 0.004, and 0 for PP 6, 12, 15, and 18, respectively) (Supplementary Figure S1). For pup licking/grooming, odor exposure and PP interacted (F_{12, 126} = 1.872, P = 0.044), with paternal CO exposure slightly inhibiting the female’s pup licking/grooming on PP3 (P = 0.072) and increasing it on PP 15 (P = 0.012) and 18 (P = 0.019), causing no differences in averaged licking/grooming (Supplementary Figure S1). In addition, no effects of paternal treatment were observed on the litter size, survival rate, female offspring ratio, and pup weight during the lactation period (all P > 0.05, Supplementary Figure S2).

Offspring’s locomotor activity and anxiety-like behavior were evaluated in the open-field and light/dark boxes in adolescence (Figure 2 and Supplementary Figure S3) and adulthood (Supplementary Figure S4). ANOVA showed a main effect of paternal treatment only in adolescence, including total distance (F_{2, 90} = 5.611, P = 0.005), total transmission between the light and dark boxes (F_{2, 90} = 9.381, P < 0.001), and % time in the light box (F_{2, 90} = 14.690, P < 0.001). Similarly, there was a sex effect (F_{1, 90} = 19.806, P < 0.001) and an interaction of paternal treatment and sex (F_{2, 90} = 7.264, P < 0.001) on the percentage of time in the light box. In the open field, CO offspring showed increased travel distance (P = 0.004) compared with those of DW offspring (Figure 2A–C). In the light/dark box, these offspring entered the dark box more frequently (P = 0.002, Figure 2D), and females even spent less time in the light box (P < 0.001, Figure 2E and F), showing an increased anxiety-like behavior.

Figure 2.

Effects of preconception paternal predation risk on adolescent and adult offspring phenotypes. (A) Total distance and (B) % time in the center of the open field in adolescent offspring; (C) representative trajectory diagram of the open-field test; (D) frequency of transmissions and (E) % time in the light chamber of light/dark box in adolescent offspring; (F) representative trajectory diagram of the light/dark box test; concealing time towards acute distilled water (G), rabbit urine (H) and cat urine (I) exposure; (J) baseline serum CORT levels of adult offspring; n = 16 for each group. DW, distilled water; RO, rabbit odor; CO, cat odor. Main effect of paternal treatment (*P < 0.05). Turkey’s post hoc test (*P < 0.05, **P < 0.01 indicates differences between different paternal treatment groups; #P < 0.05 indicates differences between males and females). The error bars represent ± standard error.

To investigate the antipredator behavioral response (Figure 2 and Supplementary Figure S5–S7), the adult offspring were exposed to the 3 odor cues in the following sequence: DW, rabbit urine, and cat urine, with a 3-day interval between each exposure. When exposed to DW, adult voles from the RO- and CO-exposed fathers spent significantly and marginally less concealing time than that of the DW offspring (ANOVA, F_{2, 90} = 5.144, P = 0.008; RO, P = 0.007; CO, P = 0.056, Figure 2G), respectively. CO fathers’ offspring also displayed more head-out behavior (ANOVA, F_{2, 90} = 6.973, P = 0.002; DW vs. CO, P = 0.001, Supplementary Figure S5A). Upon exposure to RO, we found that paternal CO exposure significantly increased self-grooming, but not other behaviors (ANOVA: F_{2, 90} = 5.206, P = 0.007; CO vs. DW, P = 0.008; CO vs. RO, P = 0.045, Supplementary Figure S6). Exposure to CO caused a main effect of paternal treatment (F_{2, 90} = 3.764, P = 0.027) and sex (F_{1, 90} = 14.798, P < 0.001) on concealing. Female RO (P = 0.066) and CO (P = 0.093) offspring displayed slightly decreased concealing behavior in response to acute cat urine exposure (Figure 2I).

After the behavioral test, baseline serum CORT concentrations were measured. Two-way ANOVA revealed a main effect of paternal treatment (F_{2, 90} = 23.734, P < 0.001), with the CO offspring showing lower levels of serum CORT than the DW (P < 0.001) and RO offspring (P = 0.019) did, and the RO offspring exhibited lower serum CORT levels than the DW offspring did (P < 0.001) (Figure 2J). These indicate that preconception paternal CO and RO exposure can decrease the activity of the basal hypothalamic-pituitary-adrenal (HPA) axis in adult offspring.

Effects of postconception paternal predatory risk exposure

To examine the effect of social buffering between the male subjects and their mates, we conducted the social interaction test immediately following the 1-h odor exposure (n = 10 for each treatment, Figure 3 and Supplementary Figures S8 and S9). Repeated measures ANOVA showed a main effect of odor exposure on male retreating (F_{2, 27} = 6.064, P = 0.007) and female grooming (F_{2, 27} = 8.780, P = 0.001) behavior. CO males spent significantly more time retreating from their mates (P = 0.005, Figure 3B), while their mates exhibited more grooming (P = 0.001, Figure 3C) towards them, possibly as a buffering effect to alleviate the stress response in the subjects. Subsequently, we compared the levels of anxiety-like behavior following repeated odor exposure (Figure 3 and Supplementary Figure S10). CO males spent significantly more time in the central open-field area than DW males did (ANOVA, F_{2, 27} = 4.659, P = 0.018; CO vs. DW, P = 0.016, Figure 3D and F), suggesting that partner presence after CO exposure could alleviate the subject’s anxiety through increased grooming.

When the pup was born, we compared the parental-offspring interactions among the 3 treatments (n = 8 for each treatment, Figure 3 and Supplementary Figures S11 and S12). Although there were no statistically significant differences in paternal (F_{2, 21} = 2.944, P = 0.075) and maternal (F_{2, 21} = 0.447, P = 0.646) investments, planned comparisons revealed a trend of higher paternal care in CO fathers (P = 0.075, Figure 3G). Similarly, we analyzed the litter size, the survival rate of weaning pups, female offspring ratio, and averaged pup weight, which showed no effects of paternal treatment (all P > 0.05, Supplementary Figure S13).

For the offspring’s locomotor activity and anxiety-like behavior (Figure 4 and Supplementary Figures S14 and S15), the main effect of paternal treatment was observed for the % time in the center of the open field (F_{2, 90} = 4.018, P = 0.021) and the light chamber of light/dark (F_{2, 90} = 4.150, P = 0.019) boxes in adolescence, as well as total distance traveled (F_{2, 90} = 3.534, P = 0.033) and inactivity time (F_{2, 90} = 3.635, P = 0.030) of open-field test at adulthood. Adolescent CO offspring spent significantly less time in the center of the open field (P = 0.029) and the light (P = 0.014) boxes (Figure 4A–D), while adult offspring showed an increased total distance (P = 0.032) and a decreased inactivity time (P = 0.022) compared with those of DW offspring (Supplementary Figure S15).

Figure 4.

Effects of postconception paternal predation risk on offspring phenotypes. (A) % time in central open field in adolescent offspring; (B) representative trajectory diagram of the open-field test; (C) % time in the light chamber of light/dark box in adolescent offspring; (D) representative trajectory diagram of the light/dark box test; concealing time of adult offspring towards acute distilled water (E), rabbit urine (F) and cat urine (G) exposure; (H) baseline serum CORT of adult offspring; n = 16 for each group. DW, distilled water; RO, rabbit odor; CO, cat odor. Main effect of paternal treatment (* P < 0.05). Turkey’s post hoc test (*P < 0.05, **P < 0.01). The error bars represent ± standard error.

When the adult offspring were exposed to DW, no effects of paternal treatment were observed (all P > 0.05, Figure 4E and Supplementary Figure S16). Upon exposure to rabbit urine, RO and CO offspring showed an increased tendency towards self-grooming (F_{2, 90} = 3.341, P = 0.040; vs. RO, P = 0.067; vs CO, P = 0.077, Figure 4F and Supplementary Figure S17), whereas these offspring spent significantly more time self-grooming (F_{2, 90} = 4.161, P = 0.019; vs. RO, P = 0.042; vs. CO, P = 0.038, Supplementary Figure S18) even after cat urine exposure. Notably, adult offspring from the postconception CO-exposed fathers showed an increased concealing behavior (ANOVA, F_{2, 90} = 3.152, P = 0.048; CO vs. DW, P = 0.049) in response to CO exposure (Figure 4G).

For the baseline CORT levels, there was a main effect of paternal treatment (F_{2, 90} = 99.243, P < 0.001) and sex (F_{1, 90} = 4.763, P = 0.032), with the CO and RO offspring having higher levels of baseline serum CORT than the DW offspring did (all P < 0.001), and male CO offspring showing higher CORT levels than male RO offspring did (P = 0.002) (Figure 4H). These indicate that postconception paternal CO and RO exposure increased basal HPA activity in adult offspring.

Discussion

Although paternal stress can induce transgenerational effects (Mashoodh et al. 2012; Korgan et al. 2016, 2018; Chan et al. 2018; Azizi et al. 2019; Brass et al. 2020; Champagne 2020; Cunningham et al. 2021; Hellmann et al. 2021, 2023), the processes and consequences of paternal predation risk effect in biparental species are only partially understood. Using Brandt’s voles, we examined the impact of paternal CO exposure on offspring phenotypes in 2 experiments. The first was a replication experiment, similar to that of Korgan et al. (2016) and Azizi et al. (2019), where female voles mated with males exposed to CO before mating and raised their offspring alone. The second experiment was postconception exposure, where males and females were housed together and jointly raised their young. Our results indicate that paternal predation risk exposure can also induce a transgenerational response in biparental species, and preconception and postconception paternal exposure differentially affects offspring phenotypes in age- and behavior-specific fashions.

Predation risk on social interaction between the male subjects and their mates

Predator odor is an ecologically relevant stressor for rodents. Some studies presented predator odor sources (e.g., cat urine and feces, fox feces and fur, and TMT) to prey species to elicit aversive, avoidance, and defensive responses and induce fear, anxiety, and stress (McGregor et al. 2004; Takahashi et al. 2005; Staples 2010; Rosen et al. 2015). In this study, male Brandt’s voles, in the company of their pregnant mates after each CO exposure, exhibited lower levels of anxiety-like behavior than the controls did. This result is partly consistent with those of the studies described in the introduction (Smith and Wang 2014; Donovan et al. 2018), which demonstrated that partner presence during and after stressful events can attenuate stress response (Burkett et al. 2016; Chun et al. 2022). Repeated exposure to cat urine in the absence of a partner produced behavioral habituation in male Brandt’s voles (data not shown), as indicated in our previous studies (Hegab et al. 2014a; Peng et al. 2021). Thus, our results suggest that partner presence following stress exposure acts as social buffering to ease the stress response and promotes subjects’ good emotions, possibly through increased partner-directed grooming (otherwise called “allogrooming,” an indicator of consolation behavior).

Social buffering effects and consolation behavior have been reported in various rodent species, including rats (Kiyokawa et al. 2018; Lu et al. 2018), mice (Sterley et al. 2018; Miao et al. 2019), prairie voles (Smith and Wang 2014; Gobrogge and Wang 2015; Burkett et al. 2016), and mandarin voles (Li et al. 2019, 2020). For example, in CD1 and C57BL/6J mice, social avoidance and anxiety-related behavior induced by chronic social defeat stress (CSDS) are alleviated by the presence of their pregnant partner without body contact during the stress process (Miao et al. 2019). In prairie voles, females spend more time interacting with an IMO restraint tube containing their male partner, and the limited physical contact decreases typical anxiety-like stress responses in males (Chun et al. 2022). In Sprague–Dawley rats, the cagemate observer increased allogrooming towards conspecifics that had received an intraperitoneal injection of acetic acid during a dyadic social interaction (Kiyokawa et al. 2018). Prairie and mandarin voles displayed enhanced allogrooming towards their stressed (including footshock stress, IMO stress, and CSDS) partners, buffering the partners’ stress response (Smith and Wang 2014; Burkett et al. 2016; Li et al. 2020). In this study, we observed consistent results, as the female mate spent significantly more time grooming CO-exposed males that displayed social avoidance after each CO exposure, which possibly alleviated the anxiety-like behavior. These results emphasize that positive social interaction can propagate good mental health and normal behavioral routines. A study has reported that wild rats usually live in gregarious colonies where social interaction may benefit predator avoidance and other stressful conditions (Macdonald et al. 1999). In wild chimpanzees, daily grooming from bond partners reportedly reduces stress hormone levels when encountering potentially life-threatening stressors (Wittig et al. 2016).

Stressed individuals may affect the emotional status of their partner while the partner provides stress buffering or consolation behavior to them. For example, observing social defeat or exposure to a stressed demonstrator can induce depressive-like behavior, elevate circulating CORT, and alter the cardiovascular tone in a state-matching pattern (Carnevali et al. 2017; Finnell et al. 2017). However, female Brandt’s voles that interacted with their mate exposed to CO showed no altered anxiety-like behavior. No stress transmission or stress contagion occurred in this study. A possible explanation for this is that although repeated predatory risk exposure had led to behavioral habituation in males (Hegab et al. 2014a), the presence of the female partner may have attenuated the male’s stress response (Chun et al. 2022), reducing the transmitted stress that the female partner received. Another possibility could be that the females did not receive increased grooming behavior during the social interaction. Based on these speculations, female voles showed no alterations in anxiety-like behavior.

Paternal predation risk on parental investment

Most mammalians, particularly the altricial mammalian species, require significant parental care during the early stages of postnatal life. In rodents, parental care includes thermoregulatory huddling, nursing, licking/grooming, and predator protection (Numan and Insel 2003; Bauer et al. 2016), which are essential for offspring survival and development. Females reportedly regulate the prenatal and postnatal parental care towards offspring in response to mate quality in the form of “differential allocation” or “reproductive compensation” (Cunningham and Russell 2000; Gowaty et al. 2007). For example, female C57BL/6 mice that mated with food-restricted or socially enriching males showed increased postnatal care (Mashoodh et al. 2012, 2018). Female rats rearing offspring from predator odor-exposed males engaged in elevated licking/grooming and arched-back nursing in a semi-naturalistic environment (Korgan et al. 2016). However, in our study, female Brandt’s voles that mated with CO-exposed males (before mating) showed decreased maternal investment in their offspring. This discrepancy could be explained by the differences in maternal behavior. Parental care in these studies included huddling, nursing, and licking/grooming (Mashoodh et al. 2012, 2018; Korgan et al. 2016), whereas nest-building is considered a parental behavior in our study, which was decreased by paternal CO exposure and caused the decreased maternal investment. Thus, our result was consistent with those of Korgan (Korgan et al. 2016), where females showed no alterations in maternal care to the offspring from predator-odor-exposed males under the standard housing. As nest-building is a behavior directly related to reproduction (Perez et al. 2023), our results indicated that paternal CO exposure inhibited postnatal maternal investment. This is possibly because exposed males are less attractive as potential mates. The changes in maternal and offspring outcomes are associated with female preference for a male during a free choice preference test rather than specific male qualities (Champagne 2020). Laboratory conditions do not enable free mate choice, that is, only forced mating with non-preferred mates occurs. However, these are our speculations, which should be examined in the future.

Regarding postconception exposure, we exposed male voles to odor sources after confirming the pregnancy of their mates. They have established close bonds with no forced mating. In this case, our results showed that paternal CO exposure facilitated subsequent paternal behavior. To the best of our knowledge, this is the first study to investigate the effects of male stress during a mate’s pregnancy on postnatal paternal care in a biparental species. Previous studies examined this effect after males had pups or before mating (Bales et al. 2006; Harris et al. 2013). For example, in Prairie voles, acute stress exposure through forced swimming increased parental care in males but not in females (Bales et al. 2006). In California mice, acute elevation of plasma CORT (30 and 60 mg/kg CORT injection) in first-time fathers did not alter direct paternal behavior, body mass, or reproductive outcomes (Harris et al. 2011), while chronic variable-stress fathers spent less time with their mate and pups and more time auto-grooming, and separated fathers spent more time behaving paternally and grooming the female mate (Harris et al. 2013). CSDS male mice engaged in more paternal behavior than the controls did and had reduced anxiety-like behavioral responses in the open-field test (Kowalczyk et al. 2018). These studies suggest that the effect of male stress on paternal behavior may depend on the stress type and duration. In our study, repeated exposure to cat urine after pairing facilitated the male’s parental behavior but did not affect that of their mate. We speculated that male voles could make a “reproductive compensation” based on their previous adverse experience. In addition, some studies have reported that positive social interaction promotes subsequent parental behavior (Champagne and Meaney 2007; Curley et al. 2009). Male Brandt’s voles exposed to predator odor received more grooming from their partner and thus exhibited more consistent parental care to their offspring.

In this study, repeated CO exposure in fathers preconception or postconception and the resulting alterations in maternal and paternal behaviors caused no detectable changes in litter size, female offspring ratio, pup survival, and offspring development, suggesting that slight alterations in the current reproductive investment might not affect parental fitness significantly. The effects could be subtle, not translating to detectable changes in pup outcomes, similar to the findings of Harris et al. (2013). In this study, preconception paternal CO exposure only inhibited nest-building behavior and not primary maternal care behavior, and postconception paternal exposure only showed an increasing tendency for subsequent paternal behavior. Thus, the effects of paternal CO exposure were relatively subtle on offspring development. Another possibility is that the effects on offspring were stage-specific. This hypothesis was supported by our current study, as paternal CO exposure increased the anxiety-like behavior of the adolescent but not of the adult offspring.

Paternal predation risk effects on offspring phenotypes

We found that the offspring of CO-exposed fathers before mating were more hyperactive (owing to less time spent inactive) and entered the dark box more frequently, and the females spent less time in the light box, indicating increased anxiety-like behavior, particularly in females. The sex-specific effect may result from mothers providing different levels of care to sons and daughters, with sons often receiving more care than daughters (Moore and Morelli 1979), and thus the decreased maternal investment observed in the present study may pertain to female offspring. Our results are broadly consistent with those of a previous study, where adolescent female offspring (but not male offspring) sired by a male exposed to a cat collar before mating displayed increased anxiety-like behavior in the open field and elevated plus maze (EPM) (Korgan et al. 2016). Similarly, adolescent rats with fathers chronically exposed to a live cat before mating exhibited more anxiety-like behavior in the EPM test (Azizi et al. 2019). In contrast, paternal TMT exposure in C57BL/6J mice increased activity and reduced anxiety-like behavior in the open field, and these behavioral patterns were stable from weaning to adulthood (Brass et al. 2020). This difference may be related to the severity of the stressor used, as they chose a 10% TMT dilution (Brass et al. 2020). Previous studies showed that small TMT concentrations (approximately 1%) are close to the natural conditions of animals (Buron et al. 2007), and a high concentration could be regarded as a harmful stimulus (Hacquemand et al. 2010a, 2010b, 2013). However, most of the results indicated that paternal predation stress increases anxiety in adolescent offspring; this phenotype could be adaptive when predators are common. Fear and anxiety are confirmed to be adaptive responses in animals against a potential or ambiguous threat (Ennaceur 2014). When life is threatened by these conditions, the physiological and behavioral responses need to be rapid and precise.

Moreover, this paternal effect on anxiety is age-dependent, as it is not shown in adulthood. This is possibly owing to different behavioral outcomes caused in adult offspring. In this study, adult CO offspring spent less time concealing and more time exhibiting a head-out behavior when exposed to DW, and females showed decreased concealing even in response to cat urine exposure. These results suggest that paternal predation risk contributes to higher levels of impulsivity or risk-taking behavior in adult offspring (particularly in females), which was supported by the findings of lower baseline CORT in this study. Human studies corroborate this hypothesis because disadvantageous patterns of decision-making were observed in the participants with the lowest cortisol levels (van Honk et al. 2003). A recent study, which indicated that paternal history of maternal separation mice displayed increased risk-taking behavior by failing to moderate their approach to the reward in the presence of a predator odor TMT (Thivisol et al. 2023), also supported our results. In three-spined sticklebacks, predator-exposed fathers produced sons that were more risk-prone, causing reduced survival when confronted with a predator (Hellmann et al. 2020). Thus, we speculated that the high risk-taking caused by preconception paternal predation risk in Brandt’s voles may be maladaptive; however, high risk implies high reward access to resources under high predation pressure (Bell et al. 2010). In addition, preconception paternal RO exposure also caused higher levels of risk-taking behavior in adult offspring. Actually, the olfactory environment was altered when males were exposed to RO. Previous studies showed that the artificial lemon odor altered the normal olfactory environment and subsequently disrupted the full and normal expression of maternal behavior (Shah et al. 2002). It was also found that odor acetophenone-induced fear conditioning in F0 male animals caused the same odor-induced startles in both F1 and F2 offspring generations (Dias and Ressler 2014). These studies may support our result, but the transgenerational effect was not as strong as paternal CO exposure.

We found consistent results in both experiments for the anxiety-like behavior in offspring, with adolescent CO offspring from postconception CO-exposed fathers also displaying increased anxiety-like behavior in the open-field and light/dark boxes. They did not display increased locomotor activity, indicating that the change in anxiety-like behavior was not owing to altered locomotor activity levels, which can confound emotional measures in some cases (Stanford 2007). Although the results of anxiety-like behavior were similar, the mechanism by which paternal predation risk affects offspring differs between preconception and postconception. In the experiment of preconception exposure, the effect can be mediated via epigenetic changes to sperm and/or seminal fluid, as well as maternal investment in responses to perceived paternal quality. Previous studies, through the microinjection of fertilized oocytes and zygotes, demonstrated that paternal preconception stress-associated changes to sperm small noncoding RNAs in mice influenced anxiety-like behavior and the stress-induced CORT response of adult F1 offspring (Rodgers et al. 2013, 2015; Gapp et al. 2014, 2020). However, another study in paternal social defeat stress showed that the transgenerational effects on depression and anxiety-like behaviors in offspring were mediated by changes in the maternal investment in her offspring, as this effect was absent when offspring were generated via in vitro fertilization (Dietz et al. 2011). Owing to the possibility of behavioral interaction between the mother and father at the time of mating, the transgenerational effect of preconception paternal predation risk on anxiety-like behavior in this study may be through these 2 pathways. In the experiment of postconception exposure, only the changes in the father’s behavior were demonstrated, including decreased anxiety-like behavior and increased paternal investment. Here, we did not allow the epigenetic change in sperm to influence offspring phenotypes, as the paternal CO exposure was conducted after mating. Furthermore, no change in the mother’s behavior was observed. Therefore, the transgenerational effect of postconception paternal predation risk in this study may be mediated via father-offspring interaction. Several studies in biparental species have demonstrated the role of father-offspring interaction on offspring phenotypes, including anxiety, aggression, social behavior, and response to reward (Bales and Saltzman 2016; Feldman 2016; Gromov 2022). In this study, we emphasized that increased father-offspring interaction owing to postconception paternal predation risk could cause increased locomotor activity and defensive or cautious behavior (indicated by increased defensive concealing and baseline serum CORT in adult offspring) in response to threatening situations. Being motionless and concealing are usually considered defensive responses to risks (Blanchard and Blanchard 1989; McGregor et al. 2004). Based on this and the description above, the phenotypic changes in offspring induced by postconception paternal predation risk were found to be adaptive, as they prepared offspring for a consistently stressful environment. Interestingly, these changes were consistent with those observed in our previous study of maternal predation risk exposure during pregnancy (Gu et al. 2023). Despite the similarity in outcomes, the transgenerational mechanisms differed, potentially involving maternal stress hormone-induced fetal programming, postnatal maternal care, and paternal and social interactions (Sheriff et al. 2017; Weinstock 2017; Chan et al. 2018). Notably, in biparental species, the presence of both parents rather than just the mother, contributes to these adaptive phenotypic changes in offspring.

Taken together, the results showed that preconception paternal exposure inhibited postnatal maternal investment and induced increased anxiety in adolescents and more risk-taking in adult offspring, while postconception paternal exposure alleviated paternal anxiety, promoted postnatal paternal investment, and caused high anxiety in adolescence and increased defense in adulthood. Paternal predation risk can affect offspring in pathway-dependent and behavior-specific manners. These findings support existing reports that preconception paternal stress can program offspring phenotypes via epigenetic marks of spermatozoa (e.g., DNA methylation, histone modification, and noncoding RNAs) and maternal responses. Notably, our study demonstrated that the transgenerational effect induced by paternal stress in biparental species could affect the offspring through father-offspring interaction. In the presence of a father, these observed effects are anticipatory and may prepare offspring well for a threatening environment.

Supplementary Material

Supplementary material can be found at https://academic.oup.com/cz.

Authors’ Contributions

R. Wu designed and supervised the study and drafted and revised the manuscript; J. Zhu, P. Wang, Z. Xu, and L. Chen participated in experiment performing and data analysis; Y. Chen, J. Xu, and Q. Wang assisted with experiments performing and behavioral analysis; S. Yang and W. Wei helped modify the manuscript. All authors gave final approval for publication and agreed to be held accountable for the work performed herein.

Conflict of Interest

We declare we have no competing interests.

Ethical Statement

All animal procedures were conformed to the Guide for the Care and Use of Laboratory Animals of China, and approved by the Institutional Animal Care and Use Committee of Yangzhou University (No 202302051).

Funding

This work was supported by grants from the National Natural Science Foundation of China (31770422 and 31900334), Natural Science Foundation of Jiangsu Province (BK20190910), Postdoctoral Science Foundation of China (2018M630610 and 2019T120468), Students’ Project for Innovation and Entrepreneurship Training Program of Yangzhou University (XCX20230796), and Priority Academic Program Development of Jiangsu Higher Education Institutions.

Data Accessibility

All data generated or analyzed during this study are included in the figures and in electronic supplementary material. Further inquiries can be directed to the corresponding authors.