Paternal absence and increased caregiving independently and interactively shape the development of male prairie voles at subadult and adult life stages

Abstract

The influence of maternal caregiving is a powerful force on offspring development. The absence of a father during early life in biparental species also has profound implications for offspring development, although it is far less studied than maternal influences. Moreover, we have limited understanding of the interactive forces that maternal and paternal caregiving impart on offspring. We investigated if behaviorally upregulating maternal care compensates for paternal absence on prairie vole (Microtus ochrogaster) pup development. We used an established handling manipulation to increase levels of caregiving in father-absent and biparental families, and later measured male offspring behavioral outcomes at sub-adulthood and adulthood. Male offspring raised without fathers were more prosocial (or possibly less socially anxious) than those raised biparentally. Defensive behavior and responses to contextual novelty were also influenced by the absence of fathers, but only in adulthood. Offensive aggression and movement in the open field test changed as a function of life-stage but not parental exposure. Notably, adult pair bonding was not impacted by our manipulations. Boosting parental care produced males that moved more in the open field test. Parental handling also increased oxytocin immunoreactive cells within the supraoptic nucleus of the hypothalamus (SON), and in the paraventricular nucleus (PVN) of biparentally-reared males. We found no differences in vasopressinergic cell groups. We conclude that male prairie voles are contextually sensitive to the absence of fathers and caregiving intensity. Our study highlights the importance of considering the ways early experiences synergistically shape offspring behavioral and neural phenotypes across the lifespan.

Introduction

Early life experiences serve a crucial role in shaping the development of physiological, neural, and behavioral outcomes in offspring. Most notably, naturally and experimentally induced variation in a mother’s behaviors and their offspring-directed care has been repeatedly shown to direct developmental processes in rodents (Curley and Champagne, 2016). The maternal mediation hypothesis originally suggested that maternal care induces the development of subsequent stress resistance in offspring and provided a powerful way of explaining how seemingly mundane differences in maternal care or maternal separation could have astonishing influences over offspring physiology and behavior (Levine, 2000). For example, rodent maternal behaviors like licking and grooming reflect the quality of the postnatal environment (Champagne and Curley, 2005). When dams exhibit high levels of licking and grooming, their adult offspring display fewer anxiety-like behaviors and attenuated physiological responses to stressors compared to adult offspring of dams who lick and groom comparatively less (Caldji et al., 1998; Liu et al., 1997). Although quite powerful, the maternal mediation hypothesis becomes even more compelling when additional factors like exposure to novelty (Tang et al., 2006) or environmental stress (Macrì and Würbel, 2006) are considered in tandem. Moreover, maternal care among rodents induces epigenetic modifications to several genes important for neuroendocrine signaling, serving as a mechanism by which social contact in the natal nest drives diversity in adult offspring social behavior (Champagne, 2011). In this way, mothers can behaviorally tune the gene regulatory profiles of their offspring in early life, guiding their pups down distinct neurodevelopmental trajectories.

If variation in maternal behavior confers such powerful effects, then the contributions of other neonatal caregivers should also impact offspring development. Most mammalian species are uniparental, with 10% or less exhibiting biparental care (Kleiman and Malcolm, 1981). Of this subset, which notably includes humans, the socially monogamous prairie vole (Microtus ochrogaster) has enabled investigations into the impacts that fathers exert to shape offspring phenotypes. Prairie vole mothers and fathers display all the same care behaviors towards their pups with the exception of nursing (Thomas and Birney, 1979). Prior studies have revealed that when fathers were removed from the natal nest, offspring were slower to begin consuming solid foods and egressing from the nest (Wang and Novak, 1992), spent less time engaged in alloparental care as juveniles (Wang and Novak, 1994), and had longer latencies to form adult pair bonds (Ahern and Young, 2009). Additionally, pups that were reared by only mothers received significantly less parental care than pups that were reared biparentally (Ahern and Young, 2009; Rogers and Bales, 2020; Tabbaa et al., 2017). Taken together, these studies suggest that offspring phenotypic development may be mediated by the reduced levels of care that stem from the absence of a father, profoundly expanding the scope of what the maternal mediation hypothesis originally proposed.

Under standard laboratory conditions, prairie vole mothers do not increase their caregiving to compensate for the absence of a father (Ahern and Young, 2009; Ahern et al., 2011; Bosch et al., 2018; Hiura et al. 2023; McGuire et al., 2007; Rogers and Bales, 2020; Tabbaa et al., 2017; but see Kelly et al., 2020). However, both maternal and paternal effects resulting from variation in caregiving and the postnatal environment have been reported (Bales and Saltzman, 2016) indicating that each parent can serve as a major source of variation in caregiving with life-long consequences. Here, we investigate the extent to which the maternal mediation hypothesis can be extended to include fathers, and to what extent variation in mothers might rescue the impacts of absent paternal care due to loss of fathers. We hypothesized that if mothers were induced to upregulate their caregiving behavior, they may be capable of offsetting the loss in care that results from a missing father. Experimental handling manipulations increase the levels of care that prairie vole offspring receive from their parents (Bales et al., 2007; Carter et al., 2008; Perkeybile et al., 2019). Therefore, the present study leveraged handling manipulations to ask if behaviorally upregulated caregiving among mothers could compensate for the developmental effects resulting from absent fathers in the postnatal nest, and/or if handled mothers and fathers enhanced offspring phenotypes synergistically. We examined behavioral outcomes in male prairie vole offspring that were reared in families with and without their fathers, and with or without experimental handling. We also assessed behavioral profiles at two important points in prairie vole life-history to determine if the effects of rearing conditions were age-specific.

Development of social behavioral neural substrates are driven by the specific social experiences that offspring encounter. Notably, oxytocin (OT) and arginine vasopressin (VP) are involved in the control of a variety of social behaviors (including parental care, pair bonding, social recognition, aggression, and sexual behaviors) and they are sensitive to variation in parental care in several rodent species (e.g., Bales and Saltzman, 2016). OT and VP are primarily produced in the paraventricular nucleus (PVN) and supraoptic nucleus (SON) of the hypothalamus (Gainer et al., 2002). These sites project to the posterior pituitary where they release hormones into the periphery, but they also have extensive extrahypothalamic efferents and can release peptides axonally to provide signals that act on oxytocin and vasopressin receptors throughout the forebrain (Landgraf and Neumann, 2004). Indeed, activation and inhibition of PVN OT neurons promote and inhibit paternal care, respectively, in monogamous and biparental Mandarin voles (He et al., 2021). Smaller OT and VP cell groups are also found in the bed nucleus of the stria terminalis (BST), a region closely tied to social behavior (Kelly and Goodson, 2014b). As such, variation in the numbers of OT and VP neurons in these cell groups may have significant implications for the volume of peptide available for release, and therefore, the neuroendocrine modulation of behavior. We investigated the impact of rearing experiences on the number of OT and VP cells within the PVN, SON, and BST to provide potential neural corollaries that map onto behavioral outcomes.

Methods

Experimental Animals

We created F1 breeders derived from wild-caught animals originally trapped in Champagne-Urbana, Illinois and reared at Cornell University. Only breeders that birthed a litter of three or more (of which at least two were male) were used in this study. We bred experimental families in cohorts to accommodate the number of animals to be used within the spatial constraints of the colony and testing rooms. From these breeders, 55 F2 offspring from 26 breeding cohorts served as subjects in this study. All animals were housed in standard polycarbonate rodent cages (29 × 18 × 13 cm) lined with Sani-chip bedding and provided nesting material. We provided animals ad libitum access to water and food (Rodent Chow 5001, LabDiet, St. Louis, MO, USA). Animals were kept under a 14:10 light-dark cycle with ambient temperature maintained at 20 ±2 °C. No animals in this experiment were raised in isolation. Sex was assigned based on external genitalia. Our study focused on male offspring, in part because male prairie voles appear to be exceptionally sensitive to social factors in the natal nest (e.g., Ahern and Young, 2009; Ahern et al., 2011; Bales et al., 2011; Bales and Perkeybile, 2012; Bosch et al., 2018; Carter et al., 2008; Danoff et al., 2023; Hiura and Ophir, 2018; Kelly et al., 2018a; Kelly et al., 2020; Perkeybile et al., 2013; Prounis et al., 2015; Prounis and Ophir, 2019; Prounis et al., 2018; Rogers and Bales, 2020; Rogers et al., 2021; Tabbaa et al., 2017; Wang et al., 2012; Wang and Novak, 1994; Wang and Novak, 1992). All procedures used in this study were approved by the Institutional Animal Care and Use Committee of Cornell University (ACUP 2013-0102).

Experimental conditions

We created four rearing conditions that manipulated father presence or absence and parental caregiving (via experimental handling) during preweaning development of subjects. For most behavioral tests, we also conducted a repeated measure assessment of subjects by age at two stages of maturation (sub-adulthood [PND 24-31] and adulthood [PND 60-74]). Therefore, most comparisons followed a 2 x 2 x 2 design (exposure to fathers, manipulation of parents to modify caregiving, and subject age, see below).

Exposure to Father

Prairie vole gestation is typically 21 days from fertilization to birth. To create a Father Absent treatment group, we removed fathers from home cages after 20 days of cohabitation with a dam to ensure that pups were never exposed to their fathers. For families in the Father Present condition, we removed fathers then immediately returned them to their home cages to account for the potential effect of nest disturbance. Nests were carefully monitored for the birth of pups, the number of pups was recorded at birth, and the litter was culled to three to standardize litter size across breeders. Males were preferentially spared for testing in the following experiments.

Modification of Parental Caregiving

By removing fathers, we reduced the total amount of care pups would receive (Hiura et al., 2023). Our second manipulation aimed to increase caregiving by physically handling parents, a method established in prior vole studies and confirmed in our lab (Carter et al., 2008; Hiura et al., 2023). For details on the impacts of these manipulations on caregiving that subject pups received, see Hiura et al. (2023). On postnatal days (PND) 2, 9, and 16, families were transferred to a clean cage during weekly cage changes. In the Handled condition, parents (either just mothers, or mothers and fathers; see above) were gently scruffed by a gloved hand, whereas parents in the Non-handled condition were scooped and moved with a plastic beaker (Bales et al., 2007). Nursing prairie voles engage in tenacious nipple attachment (McGuire et al., 2011), and therefore young pups (PND 2 and PND 9) were typically securely latched to their mother’s nipples. In such cases, pups were carefully moved with their mothers while nursing. When pups were unattached to mothers, they were transferred in the same manner as their parent(s), either by hand or by cup. One-hour video recordings of the home cages were taken immediately following the cage transfers for a separate study.

At PND 21, all pups were weaned from their parents and moved to a new home cage with their same sex sibling(s). One male from each cage was randomly assigned to be the focal subject for behavioral testing and histological analysis. Taken together we created four groups: Father Present / Non-handled (i.e., a control group, N = 15), Father Present / Handled (N = 16), Father Absent / Non-handled (N = 13), and Father Absent / Handled (N = 11).

Behavioral Testing

All subjects underwent the same order of behavioral testing (Figure 1). After weaning, subjects were run though a fixed sequence of behavioral testing once as subadults (PND 24-31) and then again as adults (PND 60-67). The testing battery consisted of an open field test to evaluate anxiety/exploration, a social interaction test to assess pro-and anti-social behavior, a resident-intruder test to assess territorial behavior, and a tube test to assess behavioral dominance. Animals always had one day between behavioral tests during the subadult and adult testing phases. Finally, all subjects were housed with a sexually receptive female and then tested once in a partner preference test to assess pair bonding (at PND 73-74). The testing apparatuses were thoroughly cleaned between tests using 70% ethanol or soapy water.

Figure 1.

Timeline of experimental treatments, behavioral testing schedules, and brain extractions for all subjects by postnatal day (PND). Grey boxes group experimental phases together.

Open Field Test

On PND 24/25 and PND 60/61, subjects underwent a 10-minute open field test. Subjects were transferred by cup to the center of a transparent Plexiglas arena (57 cm x 57 cm) on top of a 4x4 grid of squares (14.25 cm x 14.25 cm). The centermost four squares comprised the arena “center” (28.5 cm²). We measured the duration of time spent in the center, frequency of visits to the center, and the total distance moved during the test.

Social Approach Test

On PND 26/27 and PND 62/63, subjects were placed under a plastic beaker on one side of a transparent Plexiglas arena (57 cm x 57 cm). Another plastic beaker on the opposite side held an unfamiliar adult male conspecific. Both beakers were simultaneously lifted, and animals were allowed to explore and interact for 10 mins. We analyzed the total distance moved, the duration of body contact between animals, and the mean distance between animals.

Resident-Intruder Test

On PND 28/29 and PND 64/65, we removed the siblings of the focal males from the home cage and allowed subjects to reacclimate for five minutes. After the acclimation period, we placed a novel unfamiliar conspecific male intruder into the home cage and recorded the subjects’ behaviors for five minutes. At the end of the test, we removed the stimulus animal and returned the sibling(s) back to the home cage. During the test, we measured the latency to the first attack, and counted each time the subject chased or attacked the stranger, fled, submitted (laid supine), and reared up on their hindlegs. We combined the counts of attacks and chases into a single outcome we called ‘offensive aggression’. We collapsed submissions and flees into a measure we called ‘defense behavior’.

Dominance Tube Test

Testing took place on PND 30/31 and PND 66/67. A transparent Plexiglas tube (30 cm length x 3.15 cm diameter) was used in the testing process. In this test, animals were released into opposite ends of a tube, where they could interact at the point at which they meet. In this test, the more dominant animal forces its opponent out of the tube. Before initiating the test, the male subject and an unfamiliar age-matched, and weight-matched male were individually allowed to pass through the tube to acclimate to the procedure. The acclimation order was randomized. At the start of a test trial, both animals were simultaneously released by hand into opposite ends of the tube. An animal was declared the loser when it had all four paws out of the tube. A trial ended as a draw if two minutes elapsed without a loser. The starting side was randomly assigned at the beginning of the test and held constant across three consecutive test trials, with one minute between each trial. The proportion of wins were counted as the dependent variable, and animals that won at least two trials were labeled as the winner of the test.

Partner Preference Test

Two unrelated sexually mature and inexperienced females were sexually primed with dirty bedding from an unrelated male’s cage (Dluzen et al., 1981). The next day, one of these females was randomly assigned to be the cage-mate and moved into a clean cage with the focal subject on PND73/74. After 24 hours of cohabitation, the female was considered the ‘partner’. Prior work has demonstrated that 24 hours of cohabitation in the absence of mating is sufficient for voles to form a pair bond (Williams et al., 1992). After this cohabitation phase, subjects underwent a 3 h partner preference test in a 3-chamber apparatus (51 x 102 x 30.5 cm). The chambers were separated by opaque walls, which contained small access doors to the center chamber that were offset in opposite directions from the midline of the apparatus. The subject’s partner was tethered in one side chamber (51 x 28 cm) and the sexually primed unfamiliar stimulus female was tethered in the other side chamber. The chamber that would contain each animal was determined randomly. Tethering involved using a plastic zip-tie as a collar connected to a lightweight chain constructed of fishing barrel swivels attached to the apparatus, allowing for normal activity (i.e., moving and mating). Stimulus animals were allowed 30 min to acclimate to the tethering in the testing apparatus. The focal male was then gently placed into the center chamber (51 x 46 cm) and allowed to freely explore for three hours. We measured the duration of time spent in side-by-side contact with the partner and the stranger.

Behavioral quantification

All tests were video recorded using a Sony HDR-CX330 camcorder (Sony, New York City, NY, USA) or a GoPro HERO3 camera (GoPro Inc, San Mateo, California, USA). EthoVision XT v13 software (Noldus Information Technology, Leesburg, VA, USA) was used to score the open field test, social approach test, resident-intruder test, and partner preference test. The dominance tube test was evaluated without software by an experimenter blind to experimental conditions.

Histology and Immunocytochemistry

Immediately following the partner preference test, subjects were sacrificed using isoflurane overdose and underwent transcardiac perfusion (0.1M phosphate-buffered saline [PBS, pH = 7.4], 4% paraformaldehyde in PBS). Brains were post-fixed in 4% paraformaldehyde (24 h) and sunk in 30% sucrose (48 h) before storage at −80°C. Each brain was coronally sectioned into three series (40 μm slice) in cryoprotectant. One free-floating series from each subject was fluorescently double labeled for OT and VP immunoreactivity (-ir). Sections were rinsed in PBS (2 x 30 m), blocked (1 h, PBS + 10% normal donkey serum + 0.03% Triton-X-100), and incubated in primary antibodies (48 h, mouse anti-OT 3:1000, Millipore, Billerica, MA; guinea pig anti-VP 1:1000, Peninsula Laboratories, San Carlos, CA). Sections were then rinsed in PBS (2 x 30 m), incubated in biotinylated donkey anti-guinea pig (1h, 1:8000, Jackson Immunoresearch West Grove, PA), and rinsed again in PBS (2 x 15 m). Sections were then incubated in secondary antibodies (2 h at room temp, streptavidin conjugated to Alexa Fluor 488 3:1000; donkey anti-mouse Alexa Fluor 594 5:1000, ThermoFisher Scientific, Waltham, MA). Sections were washed in PBS (overnight at 4°C), mounted onto microscope slides, and cover-slipped with Prolong Gold antifade + DAPI nuclear stain (ThermoFisher Scientific, Waltham, MA) before imaging.

Microscopy and quantification

A Zeiss AxioImager II scope with an AxioCam MRm attachment, z-drive, and Apotome optical dissector (Carl Zeiss Inc., Gottingen, Germany) was used to capture images of the PVN, BST, and SON at two coronal sections (separated by 240 μm along the rostral-caudal axis). Observers blind to the experimental conditions manually labeled OT-ir cells and VP-ir cells in the GNU Image Manipulation Program (GIMP, 2.8.22). An ImageJ script (National Institutes of Health, Bethesda, MD) compiled counts of the labeled images, and these cell counts were combined within each brain region for statistical analysis.

Statistical Analysis

All the statistical analyses were calculated in R software v.4.2.2 (Team, 2013). Behavioral and neural data were analyzed using generalized linear mixed models (GLMM) in the glmmTMB package (Brooks et al., 2017). Models were constructed with Father Presence, Handling Condition, and Age as fixed factors (as well as their full interactions), and individual Subject ID as a random effect. The partner preference test data and the neural data both left out Age as a fixed factor because these measures were only taken once in adult animals. Stimulus ID was included as a fixed effect in the model when analyzing huddling duration in the partner preference tests. The family function and link function for each GLMM was selected based upon the type and distribution of each dependent variable. Model selection was conducted by comparing Akaike’s Information Criterion values among potential models using the stats package. The DHARMa package (Hartig, 2019) was used to assess model fits with simulated residuals and regression diagnostic plots. A dispersion parameter was included if model diagnostics indicated overdispersion. Type II Wald chi-squared tests were conducted via the Anova function of the car package to generate p-values and assess significant effects of independent variables. The emmeans package (Lenth et al., 2020) was used to conduct post-hoc tests with a Bonferroni correction for multiple comparisons. When interactions were significant, we report the main effects, and the highest order interaction and related post-hoc contrast(s) rather than post-hoc contracts on lower order effects that include the same factors. For all tests, an alpha cutoff of 0.05 was used to determine statistical significance.

Results

Exploratory behaviors are higher in subadult males and males raised by experimentally handled families

We assessed how Father Presence, Handling Condition, and Age impacted exploratory behavior using the open field test. The number of visits made to the center of the apparatus differed by Age (X² = 11, df = 1, p < 0.001), where subadults visited the center more frequently than adults (t₍₆₂₎ = 3.03, p < 0.01; Figure 2A). Center visit frequency was not impacted by Father Presence (X² = 0.03, df = 1, p = 0.8), or by Handling Condition of a subjects parent(s) (X² = 3.5, df = 1, p = 0.06).

Figure 2.

Offspring performance in the Open Field Test (OFT) as a function of experimental conditions. Data are presented as box plots with individual data overlaid. Thick vertical bars represent group median values, boxes span interquartile ranges (IQR) with whiskers corresponding to the min and max values within 1.5*IQR. A) The total number of visits to the center compared between offspring life stage (i.e., Age). B) The total amount of time (in minutes, min) subjects spent in the center of the OFT between offspring life stage. C) The total distance moved during the OFT test by offspring life stage. D) The total distance moved during the OFT test by whether family groups were manipulated to increase caregiving (Handling Condition). *p ≤ 0.05.

The total distance that offspring traveled in the open field test differed as a function of Age (X² = 14.5, df = 1, p = 0.0001), where subjects traveled more as subadults compared to when they were adults (t₍₆₂₎ = 3.5, p < 0.001; Figure 2C). Total distance traveled also varied by Handling Condition (X² = 8.1, df = 1, p < 0.01), where handled animals traveled more than non-handled animals (t₍₆₂₎ = 2.8, p < 0.01; Figure 2D). Still, Father Presence did not impact total distance traveled in the open field test (X² = 1.9, df = 1, p = 0.17) and no interactions were found among these three factors (all p values > 0.05).

Finally, the duration of time spent in the center of the apparatus did not differ across any factors (Father Presence: X² = 0.12, df = 1, p = 0.73; Handling Condition: X² = 0.06, df = 1, p = 0.8; Age: X² = 1.2, df = 1, p = 0.28, Figure 2B), and no interactions were found (all p values > 0.05).

Offspring age, father presence, and experimental handling during the rearing period all impact aggressive/defensive/investigatory behavioral phenotypes

We assessed how Father Presence, Handling Condition, and Age impacted social behavior in dyadic interactions using the social interaction test. Being raised by both fathers and mothers increased social distance, but not overall social interaction among unfamiliar conspecifics. Specifically, Father Presence impacted the mean distance between the subject and a stimulus male (X² = 8.1, df = 1, p < 0.01), where animals reared with fathers were further from the stimulus males on average compared to animals reared without their fathers (t₍₉₄₎ = 2.8, p < 0.01; Figure 3F). Average social distance was not impacted by Age (X² = 0.01, df = 1, p = 0.9; Figure 3D) or Handling Condition (X² = .25, df = 1, p = 0.6; Figure 3E). Notably, none of the experimental factors impacted the time that subjects and stimulus males spent in direct body contact (Father Presence: X² = 3.8, df = 1, p = 0.051; Handling Condition: X² = 0.03, df = 1, p = 0.87; Age: X² = 1.4, df = 1, p = 0.24; Figure 3A–C), however the effect of Father Presence showed a non-significant trend for pups reared without fathers to be in relatively more contact.

Figure 3.

Results from offspring Social Interaction Tests. Data are presented as box plots with individual data overlaid. Thick vertical bars represent group median values, boxes span the interquartile range (IQR) with whiskers corresponding to the min and max values within 1.5*IQR. Total duration of body contact between focal and stimulus males in minutes (min) by A) offspring life stage (i.e., age), B) Handling Condition, and C) Father Presence. The mean distance between focal and stimulus males during the interaction test contrasted by D) offspring life stage, E) Handling Condition, and F) Father Presence. *p ≤ 0.05.

We also assessed how Father Presence, Handling Condition, and Age impacted aggressive behavior in dyadic interactions using the resident-intruder test. Our results showed that different forms of aggression were influenced in a variety of ways by Age or by an interaction between Father Presence and Age. Offensive behavior was significantly affected by Age (X² = 13, df = 1, p < 0.001), where adult animals displayed more aggressive behaviors than subadults (t₍₈₂₎ = 3.7, p < 0.001; Figure 4A). Offensive behavior was not impacted by Father Presence (X² = 1.8, df = 1, p = 0.17; Figure 4C) or by Handling Condition (X² = 9.05x10⁻⁵, df = 1, p = 0.99; Figure 4B), and no higher order interactions were found (all p values > 0.05). The latency to initiate the first attack did not vary as a function of Father Presence (X² = 0.18, df = 1, p = 0.67), Handling Condition (X² = 0.05, df = 1, p = 0.82), or Age (X² = 2.5, df = 1, p = 0.1), and no higher order interactions were found (all p values > 0.05).

Figure 4.

Male offspring performance in the Resident-Intruder test. Data are presented as box plots with individual data overlaid. Thick vertical bars represent group median values, boxes span the interquartile range (IQR) with whiskers corresponding to the min and max values within 1.5*IQR. The total number of offensive behaviors as a function of A) offspring life stage (i.e., age), B) Handling Condition, and C) Father Presence. D) The variation in the total number of defensive behaviors between offspring in the Father Absent condition compared to the Father Present condition, split by offspring life stage. E) The total number of Rear Ups between offspring in the Father Absent condition compared to the Father Present condition, split by life stage. Data are presented as box plots with individual data overlaid. Thick vertical bars represent median values, and boxes span interquartile ranges. *p ≤ 0.05.

Defensive behavior was impacted by a combination of the rearing contexts we investigated. Although there were no main effects of Handling Condition (X² = 2.9 x 10⁻⁵, df = 1, p = 0.99), or Age (X² = 0.57, df = 1, p = 0.45) on the number of defensive behaviors the resident subjects displayed toward the intruders, Father Presence significantly interacted with Age (X² = 4.2, df = 1, p = 0.04). Specifically, Father Presence did not produce group differences when offspring were subadults (t₍₈₂₎ = 0.2, p = 0.8), but adult offspring reared without fathers displayed more defensive behaviors than offspring reared biparentally(t₍₈₂₎ = 2.9, p < 0.01; Figure 4D).

Lastly, rearing up on hind legs is often considered a response to environmental novelty (Lever et al., 2006). Rearing behavior differed by an interaction between Father Presence and Age (X² = 4.5, df = 1, p = 0.03). Specifically, the presence of the father had no impact on rearing behavior when subjects were subadults (t₍₈₂₎ = 0.1, p = 0.9), but as adults, the absence of the father led to more rear-ups (t₍₈₂₎ = 3.0, p < 0.01; Figure 4E). Neither Handling Condition (X² = 0.004, df = 1, p = 0.95), nor Age (X² = 1.2, df = 1, p = 0.28) significantly affected the frequency of rear ups, and no other higher order interactions were found (all p values > 0.05).

Social dominance was unaffected by early life social conditions

We next assessed how Father Presence, Handling Condition, and Age impacted dominance behavior in dyadic interactions using the dominance tube test. The proportion of winning animals in each experimental condition was not influenced by Age (X² = 3.0, df = 1, p = 0.08), Father Presence (X² = 1.6, df = 1, p = 0.2), or Handling Condition (X² = 1.6, df = 1, p = 0.2; Figure 5A–C). No higher order interactions were found (all p values > 0.05).

Figure 5.

Proportions of “Winners” in the social dominance tube test (i.e. best out of three trials). Comparisons between A) offspring life stage (i.e., age), B) Handling Condition, and C) Father Presence. Dark bar: proportion of winners. Grey bar: proportion of losers. *p ≤ 0.05.

Early-life experiences did not impact preferences for partners or total huddling behavior

In our final assessment of behavior, we asked how Father Presence and Handling Condition impacted the establishment of adult pair bonds using the partner preference test. Subjects across conditions demonstrated robust partner preferences in adulthood. Specifically, the identity of the stimulus animal drove a significant difference in the amount of time subjects spent huddling in the partner preference test (X² = 20.0, df = 1, p < 0.001). Offspring preferred to huddle with their female partners over the novel female stimuli (t₍₉₈₎ = 5.8, p < 0.001), regardless of Father Presence or Handling Condition of parents (Figure 6A–B).

Figure 6.

Offspring performance in the partner preference test. Data are presented as box plots with individual data overlaid. Thick vertical bars represent group median values, boxes span the interquartile range (IQR) with whiskers corresponding to the min and max values within 1.5*IQR. The duration in minutes (min) that offspring spent huddling with each stimulus animal across A) Handling Condition and B) Father Presence. The total duration (min) that subjects spent in huddling contact with any stimulus animal as a function of C) Handling Condition and D) father absence/presence. *p ≤ 0.05.

To assess if early social conditions impacted offspring affiliation in general, we combined the time spent huddling with both the partner and the stranger. This total time spent huddling was not affected by Father Presence (X² = .06, df = 1, p = 0.5), Handling Condition (X² = 0.23, df = 1, p = 0.6), or by an interaction between Father Presence and Handling Condition (X² = 0.1, df = 1, p = 0.8; Figure 6C–D).

Hypothalamic OT cell numbers vary across early-life experiences

We analyzed the number of OT-ir and VP-ir cells in the PVN, SON, and BST to ask how early-life social experiences shape neuroendocrine cell populations that are implicated in social behaviors. OT-ir cell counts, but not VP-ir cell counts, were differentially impacted by our experimental conditions in the PVN and SON, but not the BST.

We found no main effect of Father Presence on PVN OT-ir cell counts (X² = 3.2, df = 1, p = 0.08). However, we did find a main effect of Handling Condition (X² = 15.0, df = 1, p = 0.0001), and a significant interaction between Father Presence and Handling Condition for the density of OT-ir detected in the PVN (X² = 6.4, df = 1, p = 0.01). Specifically, we found no differences in PVN OT-ir cell counts between uniparentally-reared (mother-only) and biparentally-reared offspring when parents were not experimentally handled (t₍₅₀₎ = 0.5, p = 0.6). In contrast, when parents were handled, males raised with fathers and mothers showed a greater total number of PVN OT-ir cells than males raised without fathers (t₍₅₀₎ = 3.1, p = 0.004; Figure 7A). Unlike OT-ir cells in the PVN, the total number of PVN VP-ir cells did not differ between males raised with or without fathers (X² = 1.8, df = 1, p = 0.18) or between males raised in families that were handled or not-handled (X² = 2.3, df = 1, p = 0.13; Figure 7B), and no interaction between these factors was found (X² = .53, df = 1, p = 0.47).

Figure 7.

Plots of OT-ir and VP-ir cell counts across experimental conditions. Thick vertical bars represent group median values, boxes span the interquartile range (IQR) with whiskers corresponding to the min and max values within 1.5*IQR. Contrasts between Handling Condition and Father Presence on the totals of OT-ir and VP-ir within the A-B) PVN, C-D) BST, and E-F) SON. *p ≤ 0.05.

Experimentally handling parents had a significant effect on the number of OT-ir cells in the SON of adult offspring (X² = 5.8, df = 1, p = 0.02), where subjects from handled families had more OT-ir neurons than those from non-handled families (t₍₄₉₎ = 2.3, p = 0.03; Figure 7E). However, we found no effect of Father Presence (X² = 0.39, df = 1, p = 0.53) and no interaction effect between Handling Condition and Father Presence (X² = 0.55, df = 1, p = 0.46) for OT-ir cells in the SON. Moreover, the total number of SON VP-ir cells did not differ by Father Presence (X² = 1.4, df = 1,p = 0.24) or by Handling Condition (X² = 1.9, df = 1, p = 0.16; Figure 7F), and no interaction between these factors was found (X² = 0.62, df = 1, p = 0.43).

Finally, we found no group differences in either the total number of OT-ir cells or VP-ir cells within the BST for Father Presence (OT-ir: [X² = 0.0, df = 1, p = 0.99], VP-ir [X² = 0.01, df = 1, p = 0.91]), Handling Condition (OT-ir: [X² = 2.4, df = 1, p = 0.13], VP-ir [X² = 0.47, df = 1, p = 0.49]), or interactions between these factors (OT-ir: [X² = 0.4, df = 1, p = 0.53], VP-ir [X² = 0.1, df = 1, p =0.75]; Figure 7C–D).

Discussion

The maternal mediation hypothesis originally offered an explanation for how and why variation in maternal caregiving could alter offspring developmental trajectories, particularly (but not exclusively) in terms of stress reactivity (Levine, 2000). Yet, fathers of biparental species provide a unique experience and play a central role in caregiving, which can also determine a child’s emotional and cognitive development or impact health and wellbeing (Amato, 1994; Dubois et al., 1994; Forehand and Nousiainen, 1993; Phares and Compas, 1992; Williams and Radin, 1999). Similarly, the absence of a father in the natal nest significantly alters the developmental trajectory of prairie voles (Ahern and Young, 2009; Bales and Saltzman, 2016; Kelly et al., 2020; Rogers and Bales, 2020; Wang and Novak, 1994; Wang and Novak, 1992). Thus, offspring phenotypic development may be mediated by the reduced levels of care that stem from the absence of a father, the enhanced caregiving that a mother might provide in compensation for an absent father, or enhanced caregiving from two highly attentive parents.

The current study aimed to assess if offspring phenotypes that follow paternal absence could be rescued by behaviorally upregulating parental care, thereby expanding the scope of the maternal mediation hypothesis. To accomplish this task, we reared animals with and without their fathers, and under two different experimental handling conditions to measure behavioral outcomes at subadult and fully adult stages of life. We found that offensive aggression increased in adulthood when compared to sub-adulthood. On the other hand, adults were less exploratory (or possibly more anxious) than subadults in the open field test. We also found that males that had been raised by parents manipulated to increase caregiving moved more in the open field test than males raised by parents that were not experimentally handled. Furthermore, males reared without a father spent more time near strangers in the social interaction test than those raised by both parents. Significant higher-order interactions in the resident intruder test also indicated that males raised without fathers showed more defensive behaviors and responded more to contextual novelty, but this difference was only observable at adulthood. These results indicate male prairie voles demonstrate developmental sensitivity to the absence of fathers under a range of behavioral domains. Notably, despite other studies that have shown that pair bonding is sensitive to the absence of fathers in the postnatal nest (e.g., Ahern and Young, 2009; Rogers and Bales, 2020, but see Prounis and Ophir, 2019), we found that pair bonding was robust across different forms of access to caregiving during development.

We also examined whether variation in early social experiences impacts nonapeptide cell populations due to their direct involvement in modulating social behavior. We found no evidence that the number of BST nonapeptide neurons were affected by the early experiences we manipulated. The number of OT-ir cells, but not VP-ir cells, of the SON were greater among offspring that were raised by one or more parents that were experimentally induced to increase caregiving through handling. Similarly, experimentally boosting caregiving increased the number of OT-ir cells in the PVN of males raised by both parents compared to males raised by only their mothers. These results indicate that OT-producing cells of male offspring were particularly sensitive to early-life manipulations in which both mothers and fathers were induced to provide increased care. These results were only detectable by considering the impacts of multiple dimensions of early experiences on brain and behavior across the lifespan.

The impact of fathers on social behavior phenotype: Context matters.

The impact that the presence or absence of a father during the rearing period had on offspring development and adult phenotype was pronounced across several of the behaviors we tested. We found that the absence of fathers in the natal nest significantly decreased the distance that males maintained between themselves and a novel conspecific in the social interaction test. Consistent with this, we also found that males raised without fathers tended to spend more time in direct body contact with an unfamiliar conspecific; however, this fell just short of our significance threshold (i.e., p = 0.051). Notably, these outcomes did not interact with handling condition (or age), suggesting that behaviorally upregulating maternal care did not offset the effects of absent fathers for these behavioral outcomes.

On first glance, our data suggest that males raised without fathers are more likely to be prosocial. Indeed, several studies on prairie voles have found that early-life experiences impact subadult and adult phenotypic outcomes (Ahern and Young, 2009; Ahern et al., 2011; Bosch et al., 2018; McGuire et al., 2007; Prounis et al., 2015; Rogers and Bales, 2020; Tabbaa et al., 2017). For instance, consistent with our social approach data, offspring reared without fathers spent more time in side-by-side contact with a novel conspecific during a social interaction task compared to biparentally reared offspring (Tabbaa et al., 2017). Similarly, father-absent reared adult males spent more time in contact with a novel infant pup during an alloparental care test than did males reared by communal pairs of a mother and an alloparental older sister (Rogers and Bales, 2020). Finally, Prounis and Ophir (2019) raised offspring with both parents and then in adolescent social groups (i.e., ‘socially enriched’) or raised offspring with mothers only followed by social isolation (i.e., ‘socially limited’). They found that socially limited males exhibited increased social contact in juvenile affiliation tests as subadults (PND 35 and PND 42) and were more likely to form a partner preference as adults. Thus, it is possible to conclude that paternally deprived male prairie voles (or males that received relatively lower levels of socialization during the postnatal period) demonstrate higher levels of prosocial behavior compared to males that were reared by multiple caregivers.

However, other studies have indicated that being raised without a father decreases prosocial behavior among male prairie voles (Ahern and Young, 2009; Kelly et al., 2020; Wang and Novak, 1994). For example, Wang and Novak (1994) reported that time engaged in allogrooming and play was lower in males reared without fathers compared to those reared with both parents, and Ahern and Young (2009) demonstrated that father-absent reared males required a week of cohabitation before a partner preference (i.e., pair bond) was formed. Although it is difficult to resolve the differences between these studies, we speculate that the social context and other factors associated with development might offer an explanation. For instance, in Wang and Novak (1994) and Ahern and Young (2009), only the presence of the father was manipulated, whereas in the current study, presence of the father and the intensity of caregiving (manipulated by experimenter handling) were altered in combination. In the case of Kelly et al. (2020), subjects raised by either both parents or just mothers were also raised in contexts where food acquisition for the parent(s) required either little or considerable effort – forcing a trade-off for parents in terms of offspring caregiving or self-care. So perhaps secondary factors (such as access to food, or the specific experimental conditions in which animals are tested) alter, mask, or reverse the impact that the absence of a father might have on pro-sociality. Indeed, Prounis and Ophir (2019) provide additional evidence for the hypothesis that different contexts interact to have different impacts on social development (see above). Clearly, parental caregiving (quality, quantity, and how they interact with other factors in the postnatal environment) has the potential to alter the degree to which male prairie voles are prosocial or reclusive. What is less clear is what factors increase or decrease their prosocial tendencies in an adult bonding context.

Differences in how the presence of a father impacts prosocial behavior can also be found across biparental rodent species. For example, socially monogamous Mandarin voles (Lasiopodomys mandarinus) become more social when deprived of social interactions in early life. Specifically, male siblings that were isolated from each other for six hours were less aggressive to and displayed more contact with their siblings upon reunion if they were raised without fathers compared biparentally reared sibling pairs (Wang et al., 2012). On the other hand, paternally deprived male and female California mice (Peromyscus californicus) took longer to contact a conspecific and spent less time investigating conspecifics compared to biparentally reared mice when observed in a social interaction test (Bambico et al., 2015). Thus, although the impact of social experiences can have long lasting impacts on developing offspring, the consequences of early socialization on subsequent social behaviors can be species-specific and context-specific. This highlights the need to take a comparative approach (Kelly and Ophir, 2015; Stevenson et al., 2018) if we are to ever fully understand the evolution of developmental programming and what common factors underlie the interactions of social experiences and development.

The impacts of age and life stage on aggression and exploration

Immature offspring often affiliate extensively with caregiving parents, then progress to interactions with siblings and peers during the juvenile and subadult stages, and finally transition to the expression of courtship and sexual behaviors with mates (Nelson et al., 2016). Age-related changes in animal aggression are commonly observed with this progression through different social contexts and have been reported in a variety of species (Moore et al., 2014; Ricci et al., 2013; Sakakura and Tsukamoto, 1999; Singh, 1989; Takahashi and Lore, 1982). The overarching pattern depicts an increase in aggression from immature offspring to reproductively mature adults. Prior work in prairie voles recapitulates this pattern, showing that between the subadult and adult stages of life, same-sex conspecific-directed prosocial behavior decreases with age, whereas aggressive behavior increases with age (Kelly et al., 2018b). The results from our resident-intruder test replicate this developmental shift towards a more offensively aggressive adult phenotype. The subadult age we tested corresponds to when prairie voles begin to become reproductively active (around PND30; Solomon, 1991). This shift toward offensive aggression at the time of sexual maturation has been explained as a means to facilitate resource and mate acquisition and defense (Buss and Shackelford, 1997).

Interestingly, our open field test results depict a decrease in exploratory behavior (possibly motivated by an increase in anxiety-like behavior) from the subadult age to the adult age. We tested subadults just after weaning when they are capable of independent living and soon begin dispersing from the nest (Arias del Razo and Bales, 2016; McGuire et al., 1993). It is possible that a high level of exploratory behavior during sub-adulthood could encourage males to leave the nest and consequently seek their own territories. Alternatively, previous experience in the open field test as subadults might explain the reduction in exploration on their second exposure as adults. The effects of age can be challenging to dissociate from the confounds of experience in developmental research. Future work may benefit from using different measures to assess exploratory/anxiety-like behavior across the lifespan to account for effects related to prior testing experience.

We did not find an effect of father presence on total distance traveled in the open field test, in contrast to Ahern and Young (2009), who found that father-absent reared offspring travelled more than biparentally-reared offspring in the first 5 minutes of an open field test. Therefore, it is possible that father-absent reared males are more active or exploratory than biparentally reared males. However, we also note that Ahern and Young (2009) reported that this effect disappeared when the first 5 minutes of the test were excluded from the analysis, suggesting that exploratory behavior in the open field test is dynamic across the testing period. Because our analysis collapsed performance across the session, it is possible that we missed the nuances of temporal variation in exploratory patterns as a function of early-life experiences.

Latent impact of paternal care as a function of age and life stage

In addition to the direct impacts that paternal presence or absence had on prosocial behaviors, the presence or absence of the father in the natal nest also interacted with the age (or life-history stage) of males in some notable ways. Specifically, rearing up on hind legs has been characterized as a behavioral response to environmental novelty (Lever et al., 2006). We observed that father-absent raised male prairie voles reared up more than males raised by both parents in the resident-intruder test; however, this pattern was only observed among adults. The increased rearing behavior among adult males raised by only mothers may indicate that these animals were more prone to gather information in the presence of a social stimulus before engaging in other behaviors. This result could just as well be interpreted as increased social interest (see above) because it could be interpreted as greater social vigilance and hesitation within an aggressive context.

Notably, we also found that defensive behaviors in the resident-intruder test were also contingent on the presence or absence of fathers and age. Specifically, subadult offspring demonstrated comparable amounts of defensive behavior regardless of whether their fathers were present in the natal nest. However, upon reaching adulthood, males raised by only mothers demonstrated significantly more defensive behaviors than those also raised by fathers. Why this effect was most observable during adulthood and not earlier is unclear. It is plausible that adult males must be more responsive to intruders into home territories (Getz et al., 1993; Madrid et al., 2020). It is also plausible that defensive behavior, and not offensive aggression, was particularly affected in these animals if balancing a propensity for prosocial behavior (see above) with territorial or aggressive interactions are placed at odds. Furthermore, it is unclear to what extent these results would be exhibited in nature, or how they might influence reproductive success of parents and sons. Indeed, additional studies that consider the reproductive outcomes of parents and offspring under natural settings are sure to provide answers addressing why or if these patterns exist. Nevertheless, these results indicate that paternal care has great potential to impact development in ways that could affect reproductive success in real-world settings, and that some effects of paternal care might only be observable at different stages of life-history.

Pair bonding behavior was robust irrespective of early-life experience

The capacity to form a pair bond is a key feature in the socially monogamous life history strategy, and pair bonding marks a complex transition in the evolution of mammalian mating systems (Lukas and Clutton-Brock, 2013). Prairie voles are known for their pair bonding behavior (Young et al., 2011), but the strength of preference for a partner can dramatically vary across individuals (Forero et al., 2023; Vogel et al., 2018), and this diversity could be attributable to environmental variation or life history stage (Forero et al., 2023; Madrid et al., 2020).

We asked if variation in parental care could account for some of the environmental variation that leads to differences in pair bonding. Inducing increased parental care via experimental handling does not impact partner preferences in male prairie voles (Bales et al., 2007). However, the absence of the father in the postnatal nest has been shown to delay or entirely disrupt the formation of a bond (Ahern and Young, 2009; Rogers and Bales, 2020). Interestingly, Rogers and Bales (2020) showed that care from a second caregiver (an older sister) was sufficient to rescue the disruptions to bonding that both they and Ahern and Young (2009) found, but only in females; male offspring raised by their mother and an older sister were equally disrupted in their pairbonding as males raised by only mothers. It is unclear to us why we did not replicate these results, but our results provided no evidence that early-life manipulations (handling or absent fathers) altered pair bonding, which is consistent with other work investigating the influence of early-life manipulations on pair bonding (e.g., Prounis and Ophir, 2019). It is possible that genotype or genotype x experience interactions could account for the variation between studies (e.g., Ahern et al., 2021). Alternatively, it is possible that the expression of pair bonding may be relatively robust to complex interactions between multiple environmental perturbations. Indeed, the nature of early-life social experiences provided by fathers and mothers (and other caregivers) potentially differ in important qualitative ways that, in turn, differentially impact the development of certain aspects of social behavior in offspring. If true, this underscores the idea that fathers have a unique role to play in the social development of offspring (e.g., Amato, 1994; Dubois et al., 1994; Ember and Ember, 1994; Flinn and England, 1997; Forehand and Nousiainen, 1993; Williams and Radin, 1999). How that interacts with other social factors like maternal care, alloparental care, or any other number of variables remains to be seen and might help explain the inconsistent nature of how adult pair bonding can be affected by developmental experience.

Early postnatal experiences shape oxytocin, but not vasopressin, cell groups in adult brains

Developmental shifts in behavior must be accompanied by changes in the underlying neural mechanisms. Previous work has demonstrated that oxytocin and vasopressin systems in prairie vole brains are susceptible to distinct early-life experiences (Ahern and Young, 2009; Bales et al., 2011; Bales and Perkeybile, 2012; Carter et al., 2008; Danoff et al., 2023; Kelly et al., 2018a; Perkeybile and Bales, 2015; Perkeybile et al., 2019; Rogers et al., 2021). Some work has expanded upon this foundation by including the analysis of multiple dimensions of the early postnatal experience on nonapeptide system profiles (Hiura and Ophir, 2018; Kelly et al., 2020; Prounis et al., 2015; Prounis et al., 2018). Here, we report that experimental manipulation to increase parental care altered the number of oxytocin neurons in adult males in the SON and PVN. Notably, father presence or absence was also a key mediating factor in shaping OT expression in the PVN. To our surprise, we found no evidence that VP positive cells were affected by either manipulation.

Bales et al. (2011) demonstrated that prairie vole families that were induced to provide more parental care (by experimental manipulation) on PND1 had male offspring with significantly more OT-ir positive cells in the SON compared to control groups that were not handled. We replicated this pattern and report that the number of OT-ir positive cells in the SON was greater in animals whose parents were induced to provide more care (via three experimental handling bouts) compared to controls. These corresponding findings suggest that experimentally elevated caregiving (via handling) has a predictable effect on oxytocin cells in the SON. We acknowledge that it is not directly clear if the differences we and others have reported are a function of the experimental handing itself, or of the consequences of the modifications to parental behavior that handling appears to cause (Bales et al., 2011).

We also found that OT-ir cells in the PVN were sensitive to early-life manipulations. Unlike OT-ir cells in the SON, however, PVN cells only demonstrated an increase in OT-ir if males were raised by parents that were experimentally handed and raised by both mothers and fathers (i.e., males receiving the most parental care). Therefore, whereas neurons that produce OT appear to be particularly sensitive to variation in parental care in males, the PVN may be specifically responsive to enhanced paternal care.

Notably, the changes we found in OT-ir cells did not appear to correspond to the behavioral results we found, making it difficult to ascertain the functional implications of changes in this cell group. The PVN is the primary source of OT in the brain and is also highly involved in modulating social behavior (Kelly and Goodson, 2014a). Thus, this neural phenotype certainly has the potential to impact social behavior in several contexts beyond what was tested in the present study. Indeed, PVN OT is involved in the modulation of the Hypothalamic-Pituitary-Adrenal axis (Grippo et al., 2007; Neumann et al., 2000), social grouping (Kelly and Goodson, 2014b), parental care (Da Costa et al., 1996; Kelly et al., 2017), consolation behavior (Li et al., 2019), social reward (Hung et al., 2017), and of course bonding and partner loss (Hirota et al., 2020; Sun et al., 2014). In any case, our results provide direct evidence for the synergistic actions of two qualitatively distinct early experiences on the phenotypic profile of a cell group critical in the control of a variety of social behaviors.

We did not find any effects of early-life experiences on nonapeptide cell counts in the BST, or group differences in VP-ir cell neurons overall. Although we were somewhat surprised by this outcome, these data are consistent with previous work that found prairie vole VP-ir did not differ in the PVN or in the SON as a function of early-life handling (Bales et al., 2011). Our results are also similar to those published by Perkeybile and Bales (2015) that reported that VP-ir did not differ in the PVN or the SON between males born to high care and low care biparental families. Thus, although vasopressin receptors of male prairie voles appear to be highly sensitive to early-life perturbations (Hiura and Ophir, 2018; Prounis et al., 2018), the vasopressin peptide does not appear to show this sensitivity.

Furthermore, our results are consistent with the hypothesis that OT-ir neuron plasticity is highly sensitive to variation in early-life experiences, and more plastic than VP-ir hypothalamic cell groups. Interestingly, it is the VP receptors of males that appear to be most sensitive to environmental perturbations during postnatal development, even though both OT and VP receptors show substantial developmental plasticity in males and females (Hiura and Ophir, 2018; Prounis et al., 2018). Together, these results imply that the plasticity of OT and VP systems may differ by which level of their signaling systems are influenced by developmental experiences (peptide-producing neurons for OT vs. receptor expression for VP). Future work would greatly benefit from the study of simultaneous changes in both receptor densities and peptide containing cells to better understand the relationship between these systems over social development. Furthermore, few empirical studies have directly manipulated these cell groups or receptor profiles in early development, which will be necessary to functionally assess their roles in the ontogeny of social behavior (Kelly and Goodson, 2014b).

Expanding the maternal mediation hypothesis to include paternal mediation?

Finally, we have reported that the handling manipulation used in this study also increased pup directed care behavior in both parents (Hiura et al., 2023). Yet the only outcome we found in the current study that was directly attributable to having been raised by parents manipulated to increase caregiving was i) how much males moved in the open field test, and ii) the number of OT-ir cells found within the SON or in the PVN as an interaction with father presence. We hypothesized that the logic underlying the maternal mediation hypothesis could be extended to include paternal care (i.e., a paternal mediation hypothesis), and that increased maternal care might compensate (or mediate) impacts of an absent father according to the maternal mediation hypothesis. Based on our results and that from other studies discussed above, the overwhelming evidence indicates that, indeed, fathers clearly impact offspring development – and in parent-specific ways. However, the maternal mediation hypothesis was originally discussed in terms of how maternal care serves a protective function by facilitating stress resiliency whereas mother separation enhances stress reactivity (Levine, 2000). Here, we found that offspring of handled parents moved more in the open field test, reflecting enhanced exploration or possibly reduced anxiety. However, because this was a main effect irrespective of the presence or absence of fathers, it is difficult to determine if this effect is exclusively attributable to maternal mediation or a broader parental mediation of behavior (though certainly an argument could be made for the former). Moreover, we believe that expanding the focus of the maternal mediation hypothesis to domains beyond anxiety-like and stress-like behaviors will be a powerful framework for contextualizing how mothers — and fathers — tune multiple dimensions of offspring development.

Notably, we found very little evidence indicating that the impact of the mother can compensate for the absence of a father to produce the same phenotypic outcome. Similarly, Rogers and Bales (2020) demonstrated that replacing a father with an older sister as a secondary caregiver increases the total care that offspring receive, but did not match the kind of care that fathers provide. Indeed, male offspring raised without fathers (with or without a secondary caregiving sister) did not form pair bonds. Similarly, animals raised without fathers, whether or not a secondary alloparent was present, showed greater oxytocin receptor density in the central amygdala (Rogers et al., 2021). Results such as these support the notion that, in some instances, fathers provide a unique form of caregiving that has its own suite of outcomes on offspring development (Amato, 1994; Forehand and Nousiainen, 1993; Williams and Radin, 1999). Taken together, we suggest that, like the maternal mediation hypothesis, paternal behavior also mediates offspring behavior, particularly in domains that relate to offspring’s social interactions and that reflect aspects of social anxiety (see above). Thus, while both parents might have the capacity to mediate the development of offspring brain and behavior, each parent might assert a more pronounced effect on different domains of offspring phenotype. To this end, the interactive effects of maternal mediation and paternal mediation should be carefully considered, and not characterized as being directly interchangeable.

Conclusion

Distinct forms of early-life experiences are heterogenous in their effects on offspring development. We found evidence that multiple factors of the early environment synergistically shape both offspring behavioral phenotypes and OT cell group profiles. Specifically, we found that a father’s presence in the natal nest has lasting consequences on a suite of behaviors that determine a male prairie vole’s social permissiveness, but that this effect is often contingent upon enhanced parental care (due to experimental handling of parents) and life-stage. Collectively, our results suggest that males born to single mothers are generally more social and demonstrate more defensive behavior and possibly novelty seeking as adults. Access to different forms of caregiving opportunities though natural means (i.e., father presence) and synthetic means (i.e., enhanced caregiving by experimenter handling) also interacted to shape the number of oxytocin, but not vasopressin, immunoreactive cells in the PVN and SON, suggesting that early social experiences with parents are encoded in specific neuroendocrine profiles of adult animals. The experimental variation we implemented only captures a fraction of the complexity of offspring development in the natural world. However, incorporating the influence of variation in multiple dimensions of postnatal experiences will bring us closer to understanding and appreciating the mechanisms that promote the development of behavioral plasticity.

Highlights.

• multiple factors in early life synergistically shape offspring phenotype

• males born to single mothers were more social and more defensive

• offensive aggression increased and exploration decreased with onset of adulthood

• pair bonding was unaffected by access to different forms of caregiving

• boosted caregiving increased the number of oxytocin cells in the hypothalamus