Paternal deprivation induces vigilance-avoidant behavior and accompanies sex-specific alterations in stress reactivity and central proinflammatory cytokine response in California mice (Peromyscus californicus)

Abstract

Rationale

Parental neglect upregulates neuroinflammatory signaling, increases anxiety, and reduces prosocial behaviors throughout development. This knowledge greatly stems from studies investigating relationships between mothers and offspring using uniparental rodent species, like Rattus and Mus. Biparental care is rare among mammals; thus, our understanding of the effects of paternal care on offspring development is limited, and the role paternal care plays in neuroimmune functioning remains unexplored.

Objectives

Using the biparental California mouse (Peromyscus californicus), we examined to what extent paternal deprivation induces sex-dependent social impairments, anxiety-like behavior, and proinflammatory cytokine production following an acute stressor in adulthood.

Methods

Biparentally-reared and paternally-deprived (permanent removal of the father 24h post-birth) adult California mice were assessed for sociability, preference for social novelty, social vigilance, and innate social avoidance, followed by the novelty-suppressed feeding test for general anxiety. Following an acute stressor, circulating corticosterone concentrations and region-specific proinflammatory cytokine concentrations were determined via radioimmunoassay and Luminex multianalyte analysis, respectively.

Results

Regardless of paternal deprivation and sex, adult California mice displayed impaired sociability and innate social avoidance behavior. In response to a novel same-sex conspecific, social vigilance was associated with reduced sociability and increased avoidance in paternally-deprived offspring. This effect was absent in biparental mice. Moreover, paternal deprivation increased the latency to consume (greater anxiety-like behavior) during novelty-suppressed feeding testing. These behavioral deficits were accompanied by lower circulating corticosterone concentration in paternally-deprived females and more hypothalamic interleukin-1β protein concentration in paternally-deprived males following an acute stressor in adulthood.

Conclusion

These data suggest that impaired sociability in biparental and paternally-deprived adult California mice may be mediated by different behavioral mechanisms induced by rearing experience. Moreover, paternal deprivation may disrupt adult emotional regulation regardless of sex. Sex-specific neuroinflammatory and HPA-axis may underlie these behavioral outcomes following paternal deprivation.

1. Introduction

Early-life stress (ELS) is commonly known as exposure to severe trauma/abuse or parental, emotional, or physical neglect occurring from the perinatal period through adolescence. (Lupien et al. 2009). ELS-exposed adults are 2.7 times more likely to develop an anxiety disorder than adults without a history of ELS (Li et al. 2016). This increases the risk of developing many behavioral disorders, like oppositional–defiant disorder and conduct disorder (Kessler et al., 2010). While ELS is equally experienced by both men and women (Pawlby et al. 2011), women with a history of ELS are twice as likely to develop an anxiety disorder compared to men (McHenry et al. 2014). Additionally, ELS-induced anxiety and behavioral disorders are related to alterations in brain regions associated with sociability, social cognition, and anxiety (e.g., prefrontal cortex, hippocampus, and amygdala). For example, ELS reduces gray matter volumes in the hippocampus and medial prefrontal cortex and decreases general anxiety in adults (Gorka et al. 2014). Moreover, the findings of Zhong and colleagues (2020) suggest that childhood trauma increases sensitization to psychosocial stress and dysregulates neural activity in the adult prefrontal cortex. Therefore, ELS may induce long-term structural and functional changes in the brain that may mediate behavioral impairments later in life.

Our understanding of the effects of ELS on behavioral and neurobiological development stems largely from studies assessing disrupted mother-child relationships (Cabrera et al. 2018). As a result, research investigating the enduring effects of father-child relationships on social and emotional development in children is scarce (Mitchell et al. 2007; Cabrera et al. 2018; Schoppe-Sullivan and Fagan 2020). This should be addressed since human studies suggest that fathers may uniquely contribute to a child’s behavioral development and mental health (Ruiz-Ortiz et al. 2017; Fitzsimons and Villadsen 2019). For instance, independent of the maternal parent, father absence during childhood may increase internalizing behaviors (e.g., unhappiness, anxiety, clingy behaviors) in boys and girls as well as increase the likelihood of boys developing externalizing behaviors (e.g., aggression and hyperactivity) later in life (Fitzsimons and Villadsen 2019). To our knowledge, no studies have assessed the effects of paternal absence on brain development in children; consequently, it is unclear to what extent alterations in the brain may be associated with enduring sex-specific behavioral impairments in children with absent fathers.

While mechanistic studies in humans are limited, our understanding of potential neurobiological mechanisms underlying adult psychopathology induced by paternal absence stems from research in biparental rodent species (i.e., where the paternal rodent engages in the same offspring care as the dam, except for nursing). In biparental rodent species, like the California mouse (Peromyscus californicus), the maternal parent will spend most of her time nursing and engaging in anogenital pup-licking, while the paternal rodent engages in more non-anogenital pup-licking (Gubernick and Alberts 1987). In California mice, maternal care declines throughout the neonatal period; accompanied by an increase in paternal care (Gubernick and Alberts 1987). Interestingly, maternal California mice do not increase offspring care (e.g., pup-grooming, pup-retrieval) to compensate for the absence of the father (Dudley 1974; Madison et al. 2022). Similar findings have been demonstrated in other biparental rodent species like degu (Octodon degus), prairie voles (Microtus ochrogaster), and mandarin voles (Microtus mandarinus). In fact, permanently removing the paternal male (paternal deprivation) results in pups receiving less parental care (i.e., licking and grooming, pup-retrieval, huddling, and play fighting) than offspring reared biparentally (Ahern 2009; Helmeke et al. 2009; Jia et al. 2009; Rogers and Bales 2020).

This reduction in early-life social stimulation can be highly stressful for biparental rodent offspring. Given that the stress response system plays a vital role in mediating emotional regulation and social behaviors (Packard et al. 2016), the role of paternal deprivation on stress axis (hypothalamic-pituitary-adrenal [HPA] axis) activity is important to note. In fact, neonatal paternal deprivation upregulates circulating corticosterone (CORT) production in juvenile mandarin voles and sensitizes CORT reactivity to mild social stress (i.e., 30 minutes of social isolation; Wang et al. 2014). Furthermore, sexually-naïve, paternally-deprived adult female mandarin voles exhibit higher circulating CORT and adrenocorticotropic hormone (ACTH) concentrations than control-reared female mandarin voles (Wu et al. 2014). These studies suggest that paternal deprivation may be a form of ELS, resulting in disruptions in HPA axis activity that could potentially induce lasting effects on adult offspring behavior. In fact, sex-specific HPA axis regulation and social behavior outcomes are observed in paternally-deprived rodents. Paternal deprivation increases serum CORT concentrations and reduces social investigation in adult male mandarin voles, whereas paternal deprivation blunts CORT production and increases social anxiety-like behaviors in adult female mandarin voles (Yu et al. 2012). These sex-specific behavioral consequences of paternal deprivation have also been reported in other biparental rodent species. Following paternal deprivation, female California mice display more aggressive behaviors and more passive social behaviors than males (Bambico et al. 2015). Therefore, paternal care may mediate offspring behavioral development in a sex-specific way and also may be required for the expression of certain behaviors that are important in biparental species that are highly social.

The long-term effects of paternal deprivation are observed in measures of structural brain plasticity as well. In juvenile mandarin voles, paternal deprivation increases arginine vasopressin (AVP) immunoreactivity in the paraventricular nucleus of the hypothalamus (PVN), while decreasing oxytocin (OXT) immunoreactivity (Wang et al. 2014). Similar disruptions in neuropeptide systems are observed in prairie voles. Paternally-deprived adult female prairie voles exhibit reduced AVP1a receptor density in the medial amygdala, compared to biparental females (Rogers et al. 2021). Paternal deprivation did not alter AVP1a receptor density in males; however, paternal male absence decreased OXT receptor density in the central amygdala in adult males (Rogers et al. 2021). The hippocampus and medial prefrontal cortex also present sex-specific vulnerabilities following paternal deprivation. Paternal absence diminishes adult hippocampal neurogenesis in female, but not male, California mice (Glasper et al. 2018) and more profoundly increases glutamate transmission in the medial prefrontal cortex of adult females, compared to males, of this species (Bambico et al. 2015). Taken together, these data suggest that the paternal male contributes to the neurobiological development of offspring.

Accumulating evidence in humans and rodents suggests that inflammatory mediators may play a role in the development of mood disorders and social deficits displayed in ELS-exposed individuals (Coelho et al. 2014; Brydges and Reddaway 2020; Carlton et al. 2021; Lumertz et al. 2022). ELS may induce systemic inflammation by increasing proinflammatory cytokines, such as interleukin (IL)-1β, IL-6, and tumor necrosis factor-α (TNF-α) (Coelho et al. 2014; Brydges and Reddaway 2020; Carlton et al. 2021; Lumertz et al. 2022). This is alarming, given that chronic low-grade systemic inflammation is related to elevated stress sensitivity (Carpenter et al. 2010) and a variety of anxiety disorders (e.g., social phobia, generalized anxiety disorder, post-traumatic stress disorder; Michopoulos et al. 2017). In addition, a relationship among ELS, neuroimmune functioning, and structural neuroplasticity have been observed in rodent models examining mother and offspring interaction (Brydges and Reddaway 2020; Lumertz et al. 2022); however, the role paternal contributions may play in neuroimmune functioning remain unexplored. Using the California mouse, we examined to what extent paternal deprivation induces sex-dependent behavioral change (e.g., social and affective behaviors) and modulates region-specific central proinflammatory cytokine concentration in adult offspring following adult stress.

2. Materials and Methods

2.1. Animals

Virgin California mice (90-120 days of age) were obtained, or were descendants of mice obtained, from the Peromyscus Genetic Stock Center (University of South Carolina, SC). These mice were used to form breeding pairs and offspring from these breeding pairs were used in this study. Breeding pairs were housed in standard mouse cages (11.5 inches x 7.5 inches x 5 inches) in a climate-controlled vivarium, under a 16:8 reversed light: dark cycle (lights off at 11:00h) and provided ad libitum access to food and water. Breeding pairs were left undisturbed (except for routine husbandry care) until the birth of their first litter (postnatal day [PND] 0). On PND 1 (i.e., 24h post-birth), litters were randomly assigned to one of two rearing conditions: biparental care (i.e., reared by the maternal and paternal rodent) or paternal deprivation (i.e., paternal male permanently removed; Figure 1A). All litters were left undisturbed (except for routine husbandry care) until weaning ~PND 31, when they were housed in mixed rearing, but same-sex dyads or triads. All experiments were approved by the University of Maryland Institutional Animal Care and Use Committee and conformed to the National Institutes of Health guidelines for the care and use of animals.

Fig. 1.

Study design and behavioral testing timeline. a) On postnatal day (PND) 1 (i.e., 24 hours after the birth of offspring), litters were randomly assigned to either the biparental care group (i.e., reared by both the maternal and paternal rodent) or the paternal deprivation group (i.e., reared solely by the maternal parent). All litters were left undisturbed (except for routine husbandry care) until weaning. b) Biparentally-reared and paternally-deprived California mice were weaned between PNDs 30 and 35 and placed in same-sex housing dyads or triads. For Experiment 2, one week after weaning, mice were weighed weekly until the commencement of behavioral testing. Between PNDs 90 and 120, adult offspring underwent behavioral testing for three-chamber social interaction and novelty-suppressed feeding behavior to assess social behaviors and general anxiety-like behavior, respectively. One day after the conclusion of the novelty-suppressed feeding test, mice received an intraperitoneal (IP) saline injection (i.e., an acute physical stressor). The prefrontal cortex, hippocampus, hypothalamus, and blood samples were collected 4-6 hours (h) after the saline injection.

2.2. Experimental Design

2.2.1. Experiment 1

At weaning, biparentally-reared mice (N sizes; female = 11 and male = 8) and paternally-deprived mice (N sizes; female = 5 and male = 12) remained undisturbed (except for routine husbandry) until blood and tissue samples were collected, on ~ 93 ± 3.20 days of age, for circulating corticosterone concentrations and proinflammatory cytokine expression, respectively. Mice were rapidly decapitated, and trunk blood was collected; blood samples were stored on ice before being centrifuged (14,000 RPM at 4°C), blood serum was aspirated and stored at −80°C for later corticosterone assay (see Corticosterone Radioimmunoassay below). Brains were quickly dissected from the skull and hemisected. The hippocampus, frontal cortex, and hypothalamus were microdissected, flash-frozen in liquid nitrogen, and stored at −80°C until assessment for pro-inflammatory cytokine concentration analysis (see Tissue Homogenization and Pro-inflammatory Cytokine Analysis below).

2.2.2. Experiment 2

At weaning, biparentally-reared (N sizes; female=10 and male=10) and paternally-deprived California mice (N sizes; female=11 and male=11) were weighed weekly until they were ~102 ± 0.95 days of age (Figure 1B). Mice were then assessed for sociability, preference for social novelty, social anxiety-like behavior, and novelty-suppressed feeding (NSF) behavior (see Behavioral Assessments below). All social behaviors were recorded from a top-down positioned camera and tracked using behavioral tracking software (EthoVision^®XT, Noldus Information Technology, Leesburg, Virginia). NSF behavior was hand-scored during testing. Mice were intraperitoneally (IP) injected with isotonic saline the next day (~22h following NSF testing), rapidly decapitated, and tissue was collected, as described in Experiment 1.

2.3. Behavioral Assessments

2.3.1. Three-chamber Social Interaction Test

Habituation.

With minor modifications, the three-chamber social interaction test procedures were adapted from Kaidanovich-Beilin et al. (2011). In a covered cart, Mice were transferred in their home cages to a red-light illuminated behavior room adjacent to the vivarium. Immediately after the final habituation to the behavior room, mice were individually placed in the center chamber of an acrylic, three-chambered testing arena (15.5 inches x 23.5 inches x 9 inches; Noldus Information Technology, Leesburg, Virginia) and allowed to explore it for five minutes. An acrylic holding cage (diameter: 3.9 inches, height: 7.9 inches) was placed in the two outer chambers. After five minutes, doors separating the center chamber from the outer chambers were removed, mice were allowed to explore all three chambers for 10 minutes, and then returned to their home cages. These procedures were repeated on three consecutive days to ensure acclimation to the testing environment. The three-chamber apparatus and holding cages were cleaned with 70% ethanol and 1:5:1 Clidox-S (Pharmacal) between each mouse.

Sociability.

After three days of habituation, mice were transferred to the behavior room, acclimated (as described above), and allowed to explore the center chamber of the three-chambered apparatus for five minutes. Next, an empty holding cage was placed in each outermost chamber; an unfamiliar sex- and age-matched, biparentally-reared conspecific (Stranger 1) was placed in one of the holding cages. The sliding doors were removed, and the subject mouse could freely explore all chambers for 10 minutes (Fig. 1a). Percent time spent exploring all chambers was recorded, and sociability was deemed more percent time spent exploring the chamber housing Stranger 1, compared to the chamber housing the empty holding cage (i.e., non-social stimulus). If a mouse spends more time with Stranger 1 versus a non-social stimulus, the mouse prefers sociability and displays typical social behaviors (e.g., social affiliation; Kaidanovich-Beilin et al., 2011). Social vigilance behavior, a measure of risk assessment was assessed by adapting previously described procedures (Williams et al. 2020; Duque-Wilckens et al. 2020). In the absence of approaching the social stimulus, a subject mouse was displayed social vigilance if it oriented its head towards the social stimulus while outside the interaction zone (within 1.15 centimeters of the holding cage).

Preference for Social Novelty.

Immediately after the sociability test, a second sex- and age-matched unfamiliar conspecific (Stranger 2) was placed in the previously empty holding cage (Fig. 2a); the test mouse was allowed to explore the apparatus for an additional 10 minutes. Percent time spent exploring the center chamber, the chamber with Stranger 1 (now the familiar rodent), and the chamber housing Stranger 2 (a novel rodent) was determined. The apparatus, and holding cages, were cleaned with 70% ethanol and 1:5:1 Clidox-S (Pharmacal) between each mouse. If a mouse spends more time with Stranger 2 than with Stranger 1, the mouse prefers social novelty and displays typical social behaviors (i.e., social recognition; Kaidanovich-Beilin et al., 2011). Immediately following testing, all mice were returned to their home cages. The arena, and holding cages, were cleaned with 70% ethanol and 1:5:1 Clidox-S (Pharmacal) between each subject mouse.

Fig. 2.

Regardless of sex and rearing, adult California mice display impaired sociability. a) Using a three-chamber social interaction chamber, mice explored an empty chamber, a center chamber, or a chamber that contained a novel same-sex conspecific (Stranger 1) for 10 minutes. Percent time spent within each chamber was measured. Mice demonstrate sociability by exploring the chamber holding Stranger 1 more than the empty chamber. b) Regardless of rearing, adult female California mice spent more time with Stranger 1 and in the empty chamber than in the center chamber. Bars represent mean ± SEM. *, p≤0.05. ***, p≤0.001. c) Regardless of rearing, adult male California mice spent more time with Stranger 1 and in the empty chamber than in the center chamber. Bars represent mean ± SEM. ****, p≤0.0001.

Social Anxiety.

Twenty-four hours later, mice were returned and acclimated to the behavior room (as described above). Social anxiety-like behavior was assessed for 10 minutes by placing the test mouse in the center chamber and again placing Stranger 1 in one of the outer chambers within a holding cage; the other holding cage remained empty (Fig. 3a). The three-chambered apparatus and holding cages were cleaned with 70% ethanol and 1:5:1 Clidox-S (Pharmacal) between each mouse. Percent time spent exploring all chambers was recorded. If a mouse spends more time with the non-social stimulus (i.e., the empty holding cage) compared to Stranger 1, then the mouse displays social avoidance behavior (Toth and Neumann, 2013). Social vigilance behavior was also assessed during this phase; the same procedure described in the “Sociability” section was used. Immediately after testing, all mice were returned to their home cages.

Fig. 3.

Social vigilance reduces sociability and increases avoidance in paternally-deprived adult California mice. a) Regardless of sex and rearing, the duration of social vigilance during the sociability test was not affected. b, right) In biparental females, no significant correlations were observed between social vigilance and time spent with Stranger 1 and time spent in the empty chamber. b, left) In paternally-deprived females, significant correlations were observed between social vigilance and time spent with Stranger 1 and time spent in the empty chamber during the sociability test. c, right) In biparental males, no significant correlations were observed between social vigilance and time spent with Stranger 1 and time spent in the empty chamber. c, left) In paternally-deprived males, significant correlations were observed between social vigilance and time spent with Stranger 1 and time spent in the empty chamber during the sociability test.

2.3.2. Novelty-Suppressed Feeding Test

The protocol for the NSF test, which measures general anxiety-like behavior (Samuels and Hen 2011), was adapted for use in California mice. Specifically, 48h following social anxiety testing in the three-chambered apparatus, mice were weighed, placed in a clean cage, and food-restricted for 24h. At the conclusion of the 24h food restriction, mice were re-weighed and transported to an adjacent behavioral room and allowed to habituate under general white lighting for 10 minutes. NSF testing occurred in a Plexiglas arena (17.5 inches x 17.5 inches x 12 inches) with the floor evenly covered with clean mouse bedding (aspen shavings). Testing occurred under ambient lighting, with an overhead lamp illuminating the center of the arena (~1000 lux). After the 10-minute habituation period, the mouse was placed in the corner of the arena, where a food pellet was positioned in the center of the arena on a piece of white filter paper. Mice freely explored the arena for 5 minutes, and latency to consume the food pellet (i.e., initial bite) was recorded. If the mouse failed to consume the food pellet, a latency of 5 minutes was recorded. A longer latency to consume the food pellet suggests increased anxiety-like behavior (Samuels and Hen 2011). Immediately after the test, the mouse was returned to their home cage and given ad libitum food. The arena was cleaned with 70% ethanol, and clean mouse bedding was placed in the arena between each mouse.

2.4. Acute Stress

The next morning (~11:30 am), mice were re-weighed, and then was IP injected with ~0.034 mL of sterile saline (0.9% NaCl), which is an acute stressor in mice (Drude et al. 2011). Studies have shown that acute stress can upregulate central and peripheral proinflammatory cytokines (Yamakawa et al. 2009; Grippo and Scotti 2013). ~Five hours later, mice were re-weighed, rapidly decapitated, and trunk blood was collected. Similar procedures for neural tissue collection and serum storage were followed as described in Experiment 1.

2.5. Corticosterone Radioimmunoassay

Corticosterone concentrations were assessed using the Corticosterone Double Antibody Radioimmunoassay (RIA) Kit (MP Biomedicals, Orangeburg, NY). Since corticosterone concentrations in Peromyscus are higher relative to Mus musculus and Rattus, the blood serum was diluted 5.2-folds more than required for other rodents (Agarwal et al. 2020). Two additional standard dilutions were added to the low end of the standard curve. All samples were run in duplicate or triplicate within the same assay. The intra-assay coefficient of variation (CV) was <15% and the inter-assay CV was <10%. For Experiment 2, two samples were excluded due to experimental error resulting in the following N sizes female: biparentally-reared=9, paternally-deprived=11; male: biparentally-reared=9, paternally-deprived female=11.

2.6. Tissue Homogenization and Proinflammatory Cytokine Analysis

Brain regions were weighed and added to tubes containing lysis buffer (pH 7.4 containing 50 mM Tris-HCL with 2mM EDTA and the following protease inhibitors: aprotinin, antipain, leupeptin, and pepstatin A (all at 1ug/mL) and 2mM phenylmethylsulfonyl fluoride at 9X tissue weight, and homogenized (Tissueruptor, Qiagen) for 30 seconds while submerged in an ice bath. Samples were then centrifuged (13,000xg, at 4°C), followed by aspiration of the supernatant, which was stored at −80°C. Lysates were assayed via Luminex Multianalyte System analysis at the Cytokine Core Laboratory (University of Maryland School of Medicine, Baltimore, MD) for IL-1β, IL-6, and TNF-α concentrations. Proinflammatory cytokine concentrations were not detected in tissue samples from Experiment 1.

2.3. Statistical Analysis

Data were analyzed using GraphPad Prism version 9.3.1 for Windows (GraphPad Software, La Jolla, California). All statistical outliers were removed using the ROUT method, where Q=1%. Two-way repeated-measures analysis of variance (ANOVA) was used to determine the effects of time, chamber, or brain region and rearing on the following outcome measures: weight, three-chamber apparatus behavior, and proinflammatory cytokine concentrations in male and female mice. For the sociability and social anxiety tests, a two-way ANOVA was conducted to assess the effects of rearing experience on the duration of social vigilance in male and female mice. Simple linear regressions were performed to determine the relationships between social vigilance duration and percent time spent in either the same-sex conspecific chamber or the empty chamber. When appropriate, Holm-Šídák’s multiple comparisons testing was performed. An unpaired t-test was performed to determine the effects of rearing on serum corticosterone concentration in male and female mice. When variances were statistically different (i.e., NSF test), non-parametric tests (i.e., Mann-Whitney test) were performed. When p ≤ 0.05, mean differences were reported as statistically significant.

3. Results

3.1. Experiment 1

3.1.1. Paternal deprivation does not alter basal corticosterone concentration

Serum was assayed for CORT concentration, via RIA, to determine the effects of rearing on a measure of HPA axis activity in male and female California mice. Unpaired t-test revealed no significant effect of paternal deprivation on basal CORT concentration in adult female or male California mice (female: t=1.10, df=12, p=0.29; male: t=1.48, df=17, p=0.16; Table 1).

Table 1.

Effects of Rearing Experience on Circulating Corticosterone Concentrations (ng/ml) in California mice

	Biparental Care		Paternal Deprivation
	Female	Male	Female	Male
Basal	6954 ± 1681	4072 ± 764.8	10191 ± 2541	2748 ± 497.8
Stress Response	10523 ± 1876	8288 ± 1726	3584 ± 943.4**	8057 ± 1822

Basal): For Experiment 1, in the absence of stress later in life (i.e., behavioral testing and intraperitoneal [IP] saline injection), paternal deprivation does not alter basal corticosterone (CORT) concentrations in adult male and female California mice.

Stress response): For Experiment 2, in biparentally-reared and paternally-deprived adult California mice, sex-dependent modulations in circulating CORT were assessed following behavioral testing and an IP saline injection (administered ~22h following novelty-suppressed feeding testing).

^~Five hours after the IP saline injection, paternally-deprived females displayed lower circulating CORT concentrations than acutely-stressed biparentally-reared females. No difference in circulating CORT was observed among males.

^**,p<0.01

3.2. Experiment 2

3.2.1. Paternal deprivation does not alter growth rate

Female.

Body weight was assessed over 9 weeks. A significant main effect of week was observed (F_2,35=92.14, p<1x10⁻⁴). Post-hoc analysis revealed that females weighed less during week 1 than weeks 2-9 (Online Resource 1a. All comparisons were statistically significant (p<1x10⁻⁴). No significant interaction between week and rearing experience (F_8,143=1.32, p=0.24) and no main effect of rearing (F_1,18=1.32, p=0.26) was observed.

Male.

Body weight was assessed over 9 weeks. A significant main effect of week was observed (F_2,39=86.72, p<1x10⁻⁴). Post-hoc analysis revealed that males weighed less during week 1 than weeks 2-9 (Online Resource 1b). All comparisons were statistically significant (p<1x10⁻⁴). No significant interaction between week and rearing experience (F_8,144=1.07, p=0.38) and no main effect of rearing (F_1,18=1.37, p=0.26) was observed.

3.2.2. Regardless of sex and rearing, adult California mice display impaired sociability

Female.

Percent time spent in each chamber (Stranger 1 versus center chamber versus empty chamber) was determined (Fig. 2a). Sociability was displayed if the mouse spent more of its time in the chamber with Stranger 1 compared to the empty chamber. Two-way repeated-measures ANOVA revealed a main effect of chamber (F_1,22=8.93, p<0.01) and a main effect of rearing (F_1,17=4.55, p=0.05). Post-hoc analysis revealed that, overall, female mice spent more time with Stranger 1 than in the center chamber (p=0.01) and also spent more time in the empty chamber than in the center chamber (p=1x10⁻⁴) (Fig. 2b). Percent time spent with Stranger 1 and in the empty chamber did not differ (p=0.19). While a main effect of rearing was found, post-hoc analysis did not indicate any statistical differences in the percent time spent in any of the chambers between biparentally-reared and paternally-deprived female mice (p>0.05, for all comparisons). No interaction between chamber and rearing (F_2,34=0.43, p=0.65) was observed.

Male.

Percent time spent in each chamber was also determined in males, and sociability was displayed if the male mouse spent more percent time in the chamber with Stranger 1 compared to the empty chamber (Fig. 2a). Two-way repeated-measures ANOVA revealed a main effect of chamber (F_1,19=15.34, p=5x10⁻⁴). Post-hoc analysis revealed that, overall, male mice spent more time with Stranger 1 than in the center chamber (p=1x10⁻⁴) and spent more time in the empty chamber than in the center chamber (p=1x10⁻⁴) (Fig. 2c). Percent time spent with Stranger 1 and in the empty chamber did not differ (p=0.19). No main effect of rearing (F_1,16=0.23, p=0.64) and no interaction between chamber and rearing (F_2,32=1.59, p=0.22) were observed.

3.2.3. Social vigilance reduces sociability and increases avoidance in paternally-deprived adult California mice

In response to a novel same-sex conspecific, social vigilance duration was assessed during sociability testing within the 3-chamber apparatus. An increase in social vigilance behavior may indicate greater social fear (Williams et al. 2020; Duque-Wilckens et al. 2020). A two-way ANOVA revealed no interaction between rearing and sex (F_{1, 34}=1.88, p=0.18), and no main effects of rearing (F_{1, 34}=1.67, p=0.21) or sex (F_{1, 34}=3.046, p=0.09), on social vigilance behavior (Fig. 3a).

Female.

A simple linear regression was conducted to determine the relationship between social vigilance duration and percent time within the chamber of Stranger 1 as well as the empty chamber during the sociability test. Paternal deprivation significantly influenced the relationship between social vigilance and percent time spent with Stranger 1 and percent time spent in with the empty chamber in female mice. Specifically, as social vigilance increased, percent time spent with Stranger 1 decreased (R²=0.57, p<0.01) and percent time spent in the empty chamber increased (R²=0.69, p<=0.01) (Fig. 3b, right). No significant relationships were observed between social vigilance and percent time spent with Stranger 1 (R²= 0.15, p=0.30) or percent time spent with the empty chamber (R²=0.15, p=0.30) among biparentally-reared female mice (Fig. 3b, left).

Male.

A simple linear regression was conducted to determine the relationship between social vigilance duration and percent time within the chamber of Stranger 1 as well as the empty chamber during the sociability test. Paternal deprivation significantly influences the relationship between social vigilance and percent time spent with Stranger 1 and percent time spent in the empty chamber in male mice. Specifically, as social vigilance increased, percent time spent with Stranger 1 decreased (R²=0.61, p=0.02) and percent time spent in the empty chamber increased (R²=0.64, p=0.02) (Fig. 3c, right). No significant relationships were observed between social vigilance and percent time spent with Stranger 1 (R²=0.01, p=0.78) or percent time spent in the empty chamber (R²=0.00, p=0.85) in biparentally-reared male mice (Fig. 3c, left).

3.2.4. Paternal deprivation does not alter preference for social novelty

Female.

Preference for social novelty was determined by comparing percent time spent in each chamber (Stranger 1 versus center chamber versus Stranger 2) (Fig. 4a). A preference for social novelty was exhibited if more percent time was spent with Stranger 2 (a novel rodent) than with Stranger 1 (a now-familiar rodent). Two-way repeated-measures ANOVA revealed a significant interaction between chamber and rearing (F_2,34=3.67, p=0.04); however, the post-hoc analysis revealed no significant differences within either rearing condition (p>0.05). Percent time spent with stranger 2 and in the center did not reach statistical significance in biparentally-reared females (p=0.06, Fig. 4b). No significant main effects of chamber (F_2,34=0.38, p=0.69) or rearing (F_1,17=3.41, p=0.08) were observed.

Fig. 4.

Paternal deprivation does not alter preference for social novelty. a) Immediately after sociability testing, preference for social novelty was assessed within a three-chamber social interaction chamber for 10 minutes. Mice explored an empty chamber, or a chamber housing Stranger 1 (a now-familiar rodent) and a novel same-sex conspecific (Stranger 2). Percent time spent within each chamber was measured. Mice exhibiting a preference for social novelty should spend more time exploring the chamber holding Stranger 2 than Stranger 1. b) Percent time spent exploring each chamber did not differ because of rearing experience in female mice. c) Regardless of rearing, males spent more time exploring the novel mouse (Stranger 2) than the center chamber. Bars represent mean ± SEM. **, p≤0.01.

Male.

Preference for social novelty was also determined in males; a preference for social novelty was exhibited if more percent time was spent with Stranger 2 (a novel rodent) than with Stranger 1 (a now-familiar rodent) (Fig. 4a). Two-way repeated-measures ANOVA revealed a significant main effect of chamber (F_2,31=6.7, p<0.01). Post-hoc analysis indicates that males spent more percent time in the empty chamber than in the center chamber (p<0.01, Fig. 4c). All other comparisons did not statistically differ (p>0.05). No interaction between chamber and rearing (F_2,32=0.33, p=0.72) and no main effect of rearing (F_1,16=6.19e-013, p>0.05) was observed.

3.2.5. Adult California mice exhibit innate social avoidance, irrespective of sex and rearing experience

Female.

Innate social anxiety-like behavior (i.e., social avoidance) was determined by comparing percent time spent in each chamber (Familiar rodent versus center chamber versus empty chamber). A mouse displayed social anxiety-like behavior if it spent more percent time in the empty chamber than it spent with the Familiar rodent (Fig. 5a). Two-way repeated-measures ANOVA revealed a significant main effect of chamber (F_2,25=51.44, p<1x10⁻⁴). Post-hoc analysis revealed that, regardless of rearing, that females spent more percent time in the empty chamber than with the Familiar rodent (p<1x10⁻⁴) and in the center chamber (p<1x10⁻⁴) (Fig. 5b). All other comparisons were not statistically significant (p>0.05). Moreover, no interaction between chamber and rearing (F_2,34=2.05, p=0.14) and no main effect of rearing (F_1,17=0.14, p=0.71) was shown.

Fig. 5.

Adult California mice exhibit innate social avoidance, irrespective of sex and rearing experience. a) Twenty-four hours following preference for social novelty testing within the three-chamber social interaction chamber, mice explored an empty chamber, a center chamber, or a chamber that contained a familiar same-sex conspecific (Familiar rodent; Stranger 1 from Day 1 of testing) for 10 minutes. Percent time spent within each chamber was measured. Mice demonstrate social anxiety by exploring the empty chamber more than chamber holding the Familiar rodent. b) Female mice spent more percent time exploring the empty chamber than the center chamber and the chamber holding the familiar rodent. Bars represent mean ± SEM. ****, p≤0.0001. c) Regardless of rearing, males spent more percent time in the empty chamber than with the Familiar rodent and the center chamber. Also, males spent more percent time with the Familiar rodent than in the center chamber. Bars represent mean ± SEM. *, p≤0.05. **, p≤0.01. ****, p≤0.0001.

Male.

Innate social anxiety-like behavior was also determined in males; males displayed innate social anxiety-like behavior (i.e., social avoidance) if they spent more percent time in the empty chamber than with the Familiar rodent (Fig. 5a). Two-way repeated-measures ANOVA revealed a significant main effect of chamber (F_1,21=18.57, p=1x10⁻⁴). Post-hoc analysis revealed that, regardless of rearing, males spent more percent time in the empty chamber than with the Familiar rodent (p=0.04) and the center chamber (p<1x10⁻⁴) (Fig. 5c). Also, regardless of rearing, males spent more percent time with the Familiar rodent than in the center chamber (p<0.01, Fig. 5c). No main effect of rearing (F_1,16=0.39, p=0.54) or interaction between chamber and rearing (F_2,32=0.90, p=0.42) was observed.

3.2.6. Regardless of sex and rearing, social vigilance is not associated with innate social anxiety-like behavior in California mice

In response to a familiar same-sex conspecific, the duration of social vigilance was determined in males and females during the social anxiety test. Increased social vigilance indicated a social fear-like response (Williams et al. 2020; Duque-Wilckens et al. 2020). A two-way ANOVA revealed no main effect of rearing (F_{1, 34}=1.09, p=0.30) or sex (F_{1, 34}=1.32, p=0.26) and no interaction between rearing and sex (F_{1, 34}=3.38, p=0.08) (Fig. 6a).

Fig. 6.

Regardless of sex and rearing, social vigilance is not associated with innate social anxiety-like behavior in California mice. a) Regardless of sex and rearing, the duration of social vigilance during the social anxiety test was not affected. b, right) In biparental females, no significant correlations were observed between social vigilance and percent time spent with the Familiar rodent and percent time spent in the empty chamber. b, left) In paternally-deprived females, no significant correlations were observed between social vigilance and percent time spent with the Familiar rodent and percent time spent in the empty chamber. c, right) No significant correlations were observed between social vigilance and percent time spent with the Familiar rodent and percent time spent in the empty chamber in biparentally-reared males. c, left) No significant correlations were observed between social vigilance and percent time spent with the Familiar rodent and percent time spent in the empty chamber in paternally-deprived males.

Female.

A simple linear regression was conducted to determine the correlations between social vigilance and percent time spent with the Familiar rodent and percent time spent in the empty chamber during the social anxiety test. In biparental females, no significant correlations were observed between social vigilance and percent time spent with the Familiar rodent (R²=0.30, p=0.12) and percent time spent in the empty chamber (R²=0.31, p=0.12) (Fig. 6b, right). In paternally-deprived females, no significant correlations were observed between social vigilance and percent time spent with the Familiar rodent (R²=0.23, p=0.13) and percent time spent in the empty chamber (R²=0.22, p=0.15) (Fig. 6b, left).

Male.

A simple linear regression was conducted to determine the correlations between social vigilance and percent time spent with the Familiar rodent and percent time spent in the empty chamber during the social anxiety test. In biparental males, no significant correlations were observed between social vigilance and percent time spent with the Familiar rodent (R²=0.31, p=0.097) and percent time spent in the empty chamber (R²=0.30, p=0.10) (Fig. 6c, right). In paternally-deprived males, no significant correlations were observed between social vigilance and percent time spent with the Familiar rodent (R²=0.12, p=0.41) and percent time spent in the empty chamber (R²=0.43, p=0.08) (Fig. 6c, left).

3.2.7. Paternal deprivation increases general anxiety-like behavior

General anxiety-like behavior was assessed following 24h of food deprivation in the NSF test. A longer latency to consume indicated anxiety-like behavior (Fig. 7a). Two-way ANOVA revealed a significant main effect of rearing was (F_1,33=4.72, p=0.04). After collapsing across sex, Mann-Whitney test revealed a longer latency to consume in paternally-deprived mice, compared to biparentally-reared mice (U=74.50, p=0.02, Fig. 7b). No interaction between sex and rearing experience (F_1,33=0.70, p=0.41) and no main effect of sex (F_1,33=0.03, p=0.87) (data not shown) was demonstrated.

Fig. 7.

Paternal deprivation increases general anxiety-like behavior. a) Forty-eight hours after assessing social anxiety-like behavior, mice were food-deprived (24 hours) and tested for novelty-suppressed feeding behavior. The latency to consume a food pellet in the center of a brightly lit was recorded, with a longer latency characterized as general anxiety-like behavior. b) Mice exposed to paternal deprivation exhibited a longer latency to consume than biparentally-reared mice. *, p≤0.05. Bars represent the median ± 95% confidence interval.

3.2.8. Paternal deprivation induces sex-dependent modulations in circulating corticosterone and proinflammatory cytokine concentration

Corticosterone.

Serum was assayed for CORT concentration in mice exposed to behavioral testing and an acute saline injection to determine the effects of paternal deprivation on stress reactivity in male and female California mice. Unpaired t-test revealed that paternally-deprived females exhibited significantly lower concentrations of serum CORT than biparentally-reared mice (t=3.71, df=15, p<0.01; Table 1); however, no effect of rearing was observed in male CORT concentration (t=0.09, df =16, p=0.93).

IL-1β.

Female.

IL-1β protein concentration was measured in the frontal cortex, hippocampus, and hypothalamus. Two-way ANOVA revealed a main effect of region (F_2,32=8.041, p<0.01); post-hoc analysis revealed a higher concentration of IL-1β in the hypothalamus than in the frontal cortex (p=0.0064) (Fig. 8a). Other comparisons did not significantly differ (p>0.05, for all comparisons). Also, no interaction between brain region and rearing (F_2,36=0.42, p=0.66) or a main effect of rearing (F_1,18 =0.50, p=0.49) was shown.

Fig. 8.

Sex-dependent modulations in circulating corticosterone and proinflammatory cytokine concentration following paternal deprivation. a) Regardless of rearing, females had a higher concentration of IL-1β in the hypothalamus than in the frontal cortex. Bars represent mean ± SEM. **, p≤0.01. b) Hypothalamic IL-1β concentration was significantly higher in paternally-deprived male mice than in biparentally-reared males. Bars represent mean ± SEM. *, p≤0.05. c) In females, there was a significant effect of brain region on IL-6 concentrations. Regardless of rearing, females had a higher concentration of IL-6 in the hypothalamus than in the frontal cortex. Bars represent mean ± SEM. *, p≤0.05. d) In males, there were no significant differences between biparentally-reared and paternally-deprived mice across brain region-specific comparisons of IL-6. e) In females, regardless of rearing, there were higher concentrations in the frontal cortex. Bars represent mean ± SEM. ***, p≤0.001. f) Regardless of rearing, males had higher concentrations of TNF-α in the hypothalamus than in the frontal cortex and hippocampus. Bars represent mean ± SEM. ***, p≤0.001. ****, p≤0.0001.

Male.

IL-1β protein concentration was quantified in the frontal cortex, hippocampus, and hypothalamus. Two-way ANOVA revealed a significant interaction between region and rearing (F_2,38=3.69, p=0.03). Post-hoc analysis revealed significantly higher concentrations of IL-1β in the hypothalamus of paternally-deprived males than biparentally-reared males (p=0.04) (Fig. 8b). Rearing did not alter frontal cortex or hippocampus IL-1β concentrations (p>0.05, for all comparisons). Since the interaction effect was significant, the main effects of rearing (F_1,19=4.89, p=0.04) and region (F_2,36=13.91, p<1x10⁻⁴) were difficult to interpret.

IL-6

Female.

IL-6 protein concentration was quantified in the frontal cortex, hippocampus, and hypothalamus. Two-way ANOVA revealed a significant main effect of brain region was (F_2,30=5.90, p<0.01); post-hoc analysis revealed a higher concentration of IL-6 in the hypothalamus than in the frontal cortex (p=0.04, Fig. 8c). Other region-specific comparisons were not statistically significant (p>0.05, for all comparisons). Also, no significant interaction between region and rearing (F_2,34=2.87, p=0.07), and no main effect of rearing (F_1,17=0.05, p=0.83) was shown.

Male.

IL-6 protein concentration was quantified in the frontal cortex, hippocampus, and hypothalamus. Two-way ANOVA revealed a significant interaction between region and rearing (F_2,24=3.74, p=0.04). However, post-hoc analysis did not reveal significant differences between biparentally-reared and paternally-deprived male mice across region-specific comparisons (p>0.05, for all comparisons). While not statistically significant, paternally-deprived males may demonstrate different concentrations of IL-6 in the frontal cortex (p=0.05) and hippocampus (p=0.07) compared to biparentally-reared mice (Fig. 8d). A main effect of region was observed (F_1,13=4.96, p=0.04); however, the main effect of rearing is difficult to interpret given the significant interaction. No main effect of rearing (F_1,12=0.10, p=0.76) was observed.

TNF-α.

Female.

TNF-α protein concentration was measured in the frontal cortex, hippocampus, and hypothalamus. Two-way ANOVA revealed a significant main effect of brain region (F_2,35=8.30, p<=0.01), and post-hoc analysis revealed higher concentrations of TNF-α in the hypothalamus than in the frontal cortex (p=5×10⁻⁴; Fig. 8e). Other region-specific comparisons were not statistically different (p>0.05, for all comparisons). Moreover, there was no main effect of rearing (F_1,18=0.62, p=0.44) and no significant interaction between region and rearing (F_2,36=0.67, p=0.52).

Male.

TNF-α protein concentration was quantified in the frontal cortex, hippocampus, and hypothalamus. Two-way ANOVA revealed significant main effect of brain region was observed (F_2,28=25.89, p<1x10⁻⁴). Post-hoc analysis revealed a higher concentration of TNF-α in the hypothalamus than in the frontal cortex (p<1x10⁻⁴) and the hippocampus (p=1x10⁻⁴) (Fig. 8f). There was no significant difference between the frontal cortex and the hippocampus (p>0.05). There was no main effect of rearing (F_1,17=3.56, p=0.08) and no significant interaction between region and rearing (F_2,34=0.19, p=0.83).

4. Discussion

We investigated the degree to which paternal deprivation, in California mice, results in sex-specific alterations in adult social (i.e., sociability, preference for social novelty, and social anxiety-like behaviors) and general anxiety-like behaviors. Furthermore, we investigated the extent to which paternal deprivation influences central proinflammatory cytokine production and stress reactivity. Our data suggest that regardless of rearing experience and sex, adult California mice exhibit impaired sociability and innate social anxiety-like behavior (i.e., social avoidance). Interestingly, in response to a novel same-sex conspecific, increased social vigilance was associated with reduced sociability and greater social avoidance in paternally-deprived male and female California mice; this effect was not shown in biparental offspring. Rearing experience did not alter social novelty preference or growth rate post-weaning. Paternal deprivation also resulted in general anxiety-like behavior in a highly anxiogenic novel environment independent of sex. While basal CORT was not significantly altered by rearing experience, acute stress CORT concentration was lower in paternally-deprived females compared to control-reared females. Male CORT concentrations did not differ. Lastly, in males with a history of acute stress, hypothalamic IL-1β protein concentration was higher in paternally-deprived males compared to control-reared mice. This effect was absent in females.

Using the three-chamber social interaction test, sociability is characterized by an increase in time spent with a novel same-sex and age-matched conspecific versus being in an empty chamber (Kaidanovich-Beilin et al. 2011). Adult male and female California mice showed impaired sociability (i.e., no difference in percent time spent with a novel same-sex conspecific versus being alone) independent of rearing experience. Our findings are incongruent with those of Bambico and colleagues (2015), who demonstrated reduced sociability in adult paternally-deprived male and female California mice. The dissimilar findings may be related to age and methodological approaches. Bambico and colleagues (2015) assessed sociability at PND 70, whereas the present study examined sociability when the mice were over 90 days of age. This suggests that paternal deprivation may induce transient social deficits in California mice. Notably, paternal deprivation increases passive social behaviors (i.e., side-by-side contact without olfactory investigation) in adult California mice, regardless of sex (Bambico et al., 2015). In this study, side-by-side contact could not be assessed, as the stimulus mouse was housed in enclosures to prevent its movement throughout the 3-chamber apparatus and to allow the subject mouse the choice to initiate social investigation. However, by using the 3-chamber apparatus, we were able to assess social vigilance (i.e., the avoidance of a social stimulus while orienting the head towards the social stimulus). Social vigilance is a risk assessment behavior that may indicate a passive stress-coping strategy induced by novel or threatening social experiences (Wright et al. 2020). In our study, increased social vigilance in paternally-deprived California mice was associated with reduced sociability and greater social avoidance when interacting with a novel same-sex conspecific; this effect was present regardless of sex. In biparental offspring, social vigilance was not associated with sociability or social avoidance. According to human studies, the “vigilance-avoidance” theory suggests that patients with anxiety disorders will display an attentional bias towards a stressful stimulus followed by active avoidance of that stressful stimulus to mitigate emotional distress (Chen and Clarke 2017). This suggests that, in response to a novel same-sex conspecific, impaired sociability in paternally-deprived male and female California mice may be mediated by a “vigilance-avoidance” stress coping mechanism. Moreover, the impaired sociability displayed by biparentally-reared and paternally-deprived California mice may be mediated by different behavioral mechanisms induced by rearing experience. Specifically, in response to a novel non-threatening mouse, impaired sociability in biparental offspring may be mediated by low social motivation. In contrast, impaired sociability in paternally-deprived offspring may be mediated by a “vigilance-avoidance” stress coping mechanism. Given that social vigilance is a risk assessment behavior in rodents that can be indicative of social fear and/or a passive stress coping mechanism (Wright et al. 2020), it is possible that novel social experiences may be more stress-inducing for paternally-deprived mice than biparentally-reared mice.

Additionally, we were able to assess innate social anxiety-like behavior (i.e., social avoidance). Innate social avoidance behavior may influence performance during the sociability test; therefore, innate social avoidance is often assessed by re-introducing the previously novel mouse 24h later. In most rodents, social avoidance is an indicator for social anxiety (Toth and Neumann, 2013). In this study, California mice preferred to explore the empty chamber rather than explore the chamber with the now-familiar mouse. This suggests that regardless of sex and rearing experience, California mice may display innate social avoidance behavior. This finding contradicts the findings of Bambico et al (2015). As previously stated, for the three-chambered social interaction test, social stimuli were individually housed in enclosures to prevent their movement throughout the 3-chamber apparatus and to allow the subject mouse the choice to initiate social investigation. On the other hand, in the Bambico et al (2015) study, social stimuli were allowed to freely roam the testing apparatus. These differences in methodical procedures may have contributed to the conflicting findings. Furthermore, regardless of sex and rearing, social avoidance of a familiar same-sex conspecific was not associated with social vigilance in California mice. Given the findings of the sociability and social anxiety tests, paternally-deprived California mice may perceive novel, but not familiar, social experiences as stress-inducing. Moreover, the social deficits displayed during the sociability test may be due to innate social avoidance behavior in California mice. This innate social avoidance behavior may be mediated by different behavioral mechanisms depending on the social environment and rearing experience.

In uniparental rodents, a preference for social novelty suggests intact social recognition capabilities (Moy et al. 2004). Here, assessing the preference for same-sex familiar and novel conspecifics, California mice did not exhibit a preference for social novelty (i.e., no difference in percent time spent with a novel same-sex conspecific versus a familiar same-sex conspecific). It is important to note that the preference for social novelty test was originally standardized using uniparental rodents (Moy et al. 2004); however, comparative studies suggest affiliative social behaviors and social novelty preferences can vary among rodent species/strains (Moy et al. 2004; Beery and Shambaugh 2021). While a preference for social novelty is typical in uniparental rodents (Moy et al., 2004), it may not always be observed in biparental rodents. Specifically, when given a choice, prairie voles display a preference for a familiar same-sex conspecific compared to a novel same-sex conspecific (Beery and Shambaugh 2021). This preference for social familiarity does not appear until an extended cohabitation period with the familiar mouse (Beery et al. 2018). While 10 minutes of social interaction was insufficient for prairie voles to develop a social novelty preference (Beery et al. 2018), 60 minutes of social interaction was sufficient to observe a preference for social familiarity (Beery and Shambaugh 2021). This lack of social novelty preference in prairie voles aligns with our lack of a preference for social novelty in California mice following a 10-minute familiarization period with a familiar and novel same-sex conspecific. A longer familiarization and testing period may be necessary to observe stable social preferences in California mice.

As assessed during the NSF test, paternally-deprived mice demonstrated greater general anxiety-like behavior (i.e., increased latency to consume) than biparentally-reared mice. Both groups of mice lost the same amount of weight during the 24-hour food deprivation period (data not shown). Unfortunately, our data cannot rule out the possibility that the paternally-deprived mice were contextually less motivated to consume during the novelty-suppressed feeding test. Other tests, like effort-based discounting, would be necessary to determine how motivated the paternally-deprived mice were compared to the biparentally-reared mice. Our paternal deprivation effects on NSF test behavior have been observed in other models of ELS in uniparental rodents (Qin et al. 2021). Other measures of general anxiety-like behavior, like the elevated plus-maze test, haveClick or tap here to enter text.Click or tap here to enter text. failed to demonstrate increased anxiety-like behavior following paternal deprivation in California mice and prairie voles (Tabbaa et al. 2017; Glasper et al. 2018). Additionally, the open field test has also failed to show anxiety-like behavior in paternally-deprived California mice (Bambico et al. 2015). These dissimilar findings may reflect differences in the stress sensitivity of these behavioral anxiety tests. While the elevated plus maze and open field test assess anxiety-like behavior, they typically do so under stress-neutral conditions. The NSF test is performed under highly anxiogenic conditions; animals undergo a 24h food deprivation period and are exposed to an open arena with bright light illuminating a food source. Paternal deprivation, along with highly stress-intensive behavioral tests, may result in more significant disruptions to affective behaviors.

While paternal deprivation is an adverse early-life experience in biparental species (Bales and Saltzman, 2016), paternal deprivation alone may not be enough to alter HPA axis responsivity. In fact, California mice and prairie voles have failed to exhibit greater CORT responses in adulthood following paternal deprivation alone (Ahern 2009; Agarwal et al. 2020; Rogers et al. 2021). In rodents, a hyperactive HPA axis can increase susceptibility to stressors later in life, increase anxiety-like behaviors, and induce social impairments in adulthood (Packard et al. 2016). Circulating CORT and ACTH are elevated in neonatal and juvenile mandarin voles following paternal deprivation (Wang et al. 2014). Furthermore, paternal deprivation primes the HPA axis to respond to mild social stress (i.e., 30 minutes of social isolation) in adult mandarin voles (Wang et al., 2014). This suggests that paternal deprivation may influence the HPA-axis early in life and may increase stress sensitivity later in life. Our data show that in adult California mice, basal CORT concentrations were unaffected by paternal deprivation. However, an effect of paternal deprivation on circulating CORT was revealed in mice that experienced behavioral testing and an acute stressor (i.e., IP saline injection). Specifically, paternal deprivation and acute stress induced a hyporeactive CORT response in females, compared to biparentally-reared females, while males did not differ in their post-stress CORT response. Several HPA axis-related mechanisms may contribute to the lower post-stress CORT observed in this study: 1) decreased production of corticotrophin-releasing hormone (CRH)/AVP and/or ACTH from the PVN and/or pituitary gland, respectively, 2) downregulation of target receptors of CRH/AVP and/or ACTH, or 3) increased glucocorticoid receptor (GR) expression at different levels of the HPA axis (Fries et al., 2005). Recently, Rogers et al. (2021) showed that paternal deprivation reduced AVP1a receptors in the medial amygdala in adult female prairie voles; this effect was absent in males (Rogers et al. 2021). Interestingly, this paternal deprivation-induced neural deficit was not mitigated by introducing a second female caregiver to the litter (Rogers et al. 2021). This suggests that the paternal male might be necessary for the proper neural development of AVP systems in female, but not male, prairie voles. Therefore, our findings showing that paternal deprivation reduced CORT reactivity in adult female California mice may be related to disruptions in AVP systems in the HPA axis. More studies will have to be conducted to determine to what extent paternal deprivation may disrupt the HPA axis in a sex-specific manner.

A secondary stressor may be necessary to reveal the effects of paternal deprivation on HPA axis activity and may be involved in emotional dysregulation in adult female California mice following paternal deprivation. Comparable findings are observed in high-anxiety adult CD1 mice following acute psychological stress (i.e., forced swim and elevated plus maze testing; Sotnikov et al. 2014). High-anxiety mice produced less circulating CORT than moderate- and low-anxiety mice (Sotnikov et al. 2014). This blunted CORT response in high-anxiety mice was associated with greater GR expression in the PVN and the pituitary gland, as well as increased hippocampal GR, compared to mice with a moderate anxiety phenotype (Sotnikov et al. 2014). Taken together, paternal deprivation and acute stress may enhance HPA axis negative feedback, resulting in a hyporeactive CORT response in adult female California mice. To what extent this potentially enhanced HPA axis negative feedback underlies the observed increase in anxiety-like behavior in adult female California mice following paternal deprivation is still unknown.

Proinflammatory cytokines can modulate the stress response system and induce a hyperactive HPA axis (Silverman et al. 2005). In mice exposed to behavioral testing and acute stress, hypothalamic IL-1β protein concentrations were higher in paternally-deprived males than in biparentally-reared males; paternally-deprived females were unaffected. IL-1β is a potent stimulator of the HPA-axis (Dunn 2006) and may underlie general anxiety-like behavior and impaired sociability following paternal deprivation in males. The upregulation of central IL-1β can reduce social investigation in rodents (Goshen and Yirmiya 2009). However, the upregulation in hypothalamic IL-1β may not be associated with the observed innate social anxiety-like behavior in paternally-deprived males. A recent study indicated that neuroinflammatory signaling might not play a role in developing social anxiety-like behavior (i.e., social avoidance) in stressed male mice (McKim et al. 2018). On the other hand, increased IL-6 in the periphery promotes social anxiety-like behavior in stressed male mice (Hodes et al. 2014); therefore, peripheral proinflammatory signaling may have induced the social anxiety behavior observed in this study. To what extent central and peripheral proinflammatory signaling may mediate social avoidance and social vigilance in paternally-deprived male and female California mice is unknown. Central and peripheral proinflammatory signaling should be investigated when assessing the sex-specific effects of paternal deprivation on social anxiety-like behaviors.

Moreover, infusion of IL-1β into the PVN increases anxiety-like behavior in adult Sprague Dawley rats (Shim et al. 2019). In the current study, region-specific effects of rearing on IL-1β concentration in the hypothalamus were not examined. The extent to which the PVN or other hypothalamic regions drove our observed increase in IL-1β should be further explored. More so, since TNF-α and IL-6 concentrations did not differ due to rearing in male or female mice, region-specific analyses may also be warranted within the frontal cortex, hippocampus, and hypothalamus for these cytokines. It is worth noting that without behavioral testing or acute stress, proinflammatory cytokines were too low to detect and add to several studies that suggest a challenge (i.e., immune, behavioral) is necessary to upregulate proinflammatory cytokines (Ishii and Yoshida 2000).

Additionally, region-specific analyses within the frontal cortex and hippocampus may be warranted when assessing proinflammatory cytokine concentrations. Studies have shown that paternal deprivation induces more significant alterations in functional and structural plasticity in the medial prefrontal cortex of adult females than in males (Bambico et al. 2015; de Schultz et al. 2020). It should be noted that de Schultz and colleagues (2020) showed that replacing the paternal rodent with a second female caregiver does not protect the medial prefrontal cortex from paternal deprivation-induced deficits in dendritic spine density and dendritic complexity in the adult female degus. This suggests that the paternal rodent may be necessary for the proper development of the medial prefrontal cortex in adult female offspring. Moreover, findings from our lab demonstrate that paternal deprivation increases microglia density in the dentate gyrus in females but not males (Madison et al. 2022). Given that proinflammatory cytokines can influence structural and functional neuroplasticity (Khairova et al. 2009; Boulanger 2009) and microglia are the primary immune cells of the brain, paternal deprivation may induce sex- and region-specific modifications in neuroinflammatory systems in the frontal cortex and hippocampus. Specifically, neuroinflammatory systems in the medial prefrontal cortex and the dentate gyrus may be more vulnerable to the effects of paternal deprivation than other regions in the frontal cortex and hippocampus, respectively; also, these disruptions may be greater in females than males.

Many varieties of ELS can significantly impact developmental and neurobiological outcomes in offspring, including impaired socio-cognitive function, increased emotional deficits, hypo- or hyper-responsive HPA axis activity, increased neuroinflammatory signaling, and reduced growth rate (Marco et al. 2015; Brydges and Reddaway 2020; Kentrop et al. 2020). ELS can also increase vulnerability to metabolic disruptions (e.g., weight gain) in both humans and rodents (Malik and Spencer 2019). We, however, did not observe an effect of rearing on growth rate in this study; previous studies from our lab also failed to observe weight differences during adolescence and into young adulthood (i.e., PNDs 35, 60, 68) in biparentally-reared or paternally-deprived California mice (Glasper et al. 2018). However, following seven days of chronic variable stress in adulthood, paternal deprivation increases terminal body weight in adult California mice regardless of sex (Agarwal et al. 2020). It is possible that the early-life adversity of paternal deprivation coupled with chronic stress disrupted metabolic processes in the California mouse. Paternal deprivation alone may not be sufficient to alter growth rate and related metabolic processes.

5. Conclusion and Future Directions

Our study examined the sex-specific effects of paternal deprivation on sociability, emotional regulation, stress responsivity, and neuroinflammatory response. We indicated that regardless of rearing experience and sex, adult California mice displayed impaired sociability towards a novel same-sex conspecific and innate social avoidance behavior towards a familiar same-sex conspecific. Interestingly, paternally-deprived adult offspring displayed vigilance-avoidant behavior in response to a novel, but not familiar, same-sex conspecific. Therefore, impaired sociability in adult California mice may be mediated by different behavioral mechanisms induced by rearing experience. Paternal deprivation also increased general anxiety-like behavior in adult offspring regardless of sex. Paternal deprivation-induced behavioral impairments were accompanied by increased hypothalamic IL-1β in male offspring. However, females exposed to paternal deprivation exhibited decreased stress reactivity – an effect not observed in males. These findings contribute to our understanding of neuroinflammatory- and stress reactivity-related processes following paternal deprivation and may highlight sex-specific mechanisms that may underlie behavioral dysfunction. Future studies should investigate the extent to which sex-specific impairments in neuroinflammatory and HPA axis mechanisms may mediate emotional regulation and social behaviors in adult paternally-deprived California mice.