Pharmacological manipulation of oxytocin receptor signaling during mouse embryonic development results in sex-specific behavioral effects in adulthood

Abstract

The oxytocin (Oxt) system is a known neuromodulator of social behaviors, but also appears to contribute to the development of sex-specific neural circuitry. In this latter role, the Oxt system helps to lay the foundation for sex-specific behaviors across the life span. In mice, the Oxt system emerges in early development, with sex differences in the expression of Oxt and a temporal offset in the expression of the Oxt receptor (Oxtr) relative to Oxt. In females, Oxt mRNA is detectable by embryonic day (E) 16.5, but in males, Oxt mRNA is not measurable until after birth. However, in both sexes, Oxtr mRNA is detectable by E12.5 and binding by E16.5. While the postnatal Oxt system has been studied, little is known about the embryonic Oxt system. Therefore, we hypothesize that it directly contributes to the developmental trajectory of the brain, ultimately affecting adult sex-specific behaviors. To test this hypothesis, Oxtr signaling was transiently disrupted at E16.5 using an Oxtr antagonist (OxtrA) and the effects on adult behavior evaluated. OxtrA-treated adult males displayed increased agonistic behavior, social investigation, and depressive-like behavior compared to vehicle-injected controls, while OxtrA-treated adult females had impaired social recognition memory compared to vehicle-injected controls. These data are the first to identify a functional link between the organizational activity of the embryonic Oxt system and adult behavior. Further, this work suggests that the Oxt system does more than serve as a neuromodulator in adulthood, but rather, may help shape the development of the neural circuitry regulating sex-specific behaviors.

Introduction

The oxytocin (Oxt) system, while well known for its modulation of behavior in adult and juvenile animals, appears to also be important to sex-specific brain formation during early development. In mice, the Oxt receptor (Oxtr) is expressed in both males and females as early as embryonic day (E) 16.5. However, only females produce Oxt mRNA during embryonic development, i.e. as early as E16.5, while males do not appear to express Oxt mRNA until postnatal day (P) 2 (Tamborski et al., 2016). Thus, it appears that females have both the ligand and the receptor during embryonic development and males only have the receptor. Males however, have significantly more arginine vasopressin (Avp) mRNA than females at E16.5 (Aulino and Caldwell, 2020), and there is documented receptor cross talk between the Oxt and Avp systems in adulthood (Song et al., 2014, 2016b). Thus, it is plausible that males do not have Oxt during embryonic development as Avp can signal through the Oxtr. Nevertheless, while there are known sex differences in the Oxt system later in life, the functional consequences for sex differences in the embryonic Oxt system remain largely unknown. Though, elegant work in neonatal rodents may provide some hints as to the broader impacts of the Oxt system in development.

In neonatal prairie voles, a peripheral injection of Oxt or an Oxtr antagonist (OxtrA) results in cFos expression patterns that differ between male and female littermates (Cushing et al., 2003). When their neurochemistry is examined later in life, female voles administered Oxt or an OxtrA on P1 have increases in Oxt immunoreactivity in the paraventricular nucleus (PVN), which were not observed in male littermates. Conversely, male voles administered an OxtrA on P1 have decreased Avp immunoreactivity in the PVN not found in female littermates (Yamamoto et al., 2004). Similarly, when females voles are administered Oxt on P1 they have region-specific increases in estrogen receptor alpha; those given an OxtrA have region-specific decreases, and male littermates have no significant changes to expression (Yamamoto et al., 2006). Moreover, when rats are given a peripheral injection of an OxtrA at P1, females have a decrease in pituitary Oxt at P60, while males have increases in Oxt (Withuhn et al., 2003; Young et al., 2005). These collective findings illustrate the various, regionally distinct ways in which the developing Oxt system differentially impacts neurochemistry in males and females.

Beyond these measurable changes in neurochemistry, and perhaps reflective of the sex- and region-dependent differences in Oxt system function in adulthood, manipulations of the Oxt system in early life affect a variety of sex-specific behaviors in adult rodents. Male, but not female, prairie voles, administered a peripheral injection of an OxtrA at P0 have increases in pup aggression and decreases in parental care (Bales et al., 2004b), while both sexes have changes in reproductive behaviors (Bales et al., 2004a; Cushing et al., 2005). Additionally, female prairie voles treated peripherally with Oxt at P0 have increases in intrasexual aggression and decreases in sociability following exposure to a male, but the same decrease is not observed in males following an exposure to a female (Bales and Carter, 2003). Similar increases in intrasexual behavior have been reported in mandarin voles given Oxt on P0. Specifically, partner preference formation in male and female mandarin voles is facilitated by peripheral Oxt treatment on P0, although its maintenance is impaired in females only (Jia et al., 2008a, 2008b). Lastly, in mice treated with an OxtrA at P0, females but not males have decreases in sociability, while males but not females have increases in their latencies to retrieve pups (Mogi et al., 2014). Taken together, these studies suggest that manipulation of the Oxt system in early postnatal life has sex- and species-specific effects on neurochemistry and behavior. Given that the Oxt system is present, and likely functional, during embryonic development, it is reasonable to hypothesize that it too has an important, and perhaps a unique, function in sex-specific brain development and behavior.

In order to interrogate the impact of disrupted embryonic Oxtr signaling on adult sex-specific behaviors, we injected an OxtrA into the brains of embryos at E16.5. This timepoint was selected as this is when Oxtr protein is detectable in both male and female mice (Tamborski et al., 2016). Based on earlier work, which has reported that early postnatal manipulations of the Oxt system impacts a range of adult behaviors (Bales et al., 2004b, 2004a, 2007; Bales and Carter, 2003; Jia et al., 2008a, 2008b), we chose to examine inter-male aggression, as well as anxiety-like behaviors, social recognition memory, and depressive-like behaviors in both males and females. We hypothesized that male mice administered an OxtrA at E16.5 would exhibit increased inter-male aggression and that females administered an OxtrA at E16.5 would have impaired social recognition memory compared to vehicle-injected controls. Additionally, we hypothesized that both males and females administered an OxtrA at E16.5 would have increased anxiety- and depressive-like behaviors compared to vehicle-injected controls. The findings of this work are important as they will link the organizational activity of the embryonic Oxt system with its function in adulthood, establishing the embryonic Oxt system as an important player in the sex-specific development of the social brain.

Methods

1.1. Subjects

C57BL/6J mice were bred and housed in the vivarium at Kent State University. All subjects were provided food and water ad libitum and kept on a 12:12 light-dark cycle. Additional mice were used in validation experiments (see supplemental). All experiments were conducted in accordance with protocols approved by the Kent State University Institutional Care and Use Committee.

1.2. Disruption of Oxtr signaling in utero and evaluation of adult behavior

1.2a. Embryonic transuterine microinjection surgery

Due to primiparous dams often eating their young, we only injected embryos from dams that were in their second pregnancy, as this increased the likelihood that they would care for their offspring. Briefly, sexually naïve, female C57BL/6J mice between the ages of 2 and 4 months were paired with an experienced C57BL/6J breeder male for two weeks. On P0, after the dam gave birth, the females were immediately re-paired with a breeder male overnight. As females enter postpartum estrus within hours of giving birth (Gilbert, 1984), the overnight pairing improves the likelihood of pregnancy and ensures that dating of the pregnancy is accurate. On E16.5, a transuterine microinjection was performed (Figure 1). Specifically, pregnant dams were anaesthetized with isoflurane and administered a subcutaneous injection of 5 mg/kg Ketofen^® [Zoetis, Parsippany, NJ, USA]. An Ioban^™ [3M^™, St. Paul, MN, USA] drape was applied before making a midline incision, exposing the uterine horns. The uterus was sprayed with warm 0.9% saline periodically to prevent drying. Each embryo was carefully isolated from the uterine horn, stabilized by hand, and injected with 2μL of either 0.9% saline (n=14 litters) or the Oxtr antagonist d(CH₂)₅[Tyr(Me)², Thr⁴, Tyr-NH₂⁹]OVT [0.375 ng/μL, courtesy of Maurice Manning](Elands et al., 1988; Manning et al., 1989; Neumann et al., 1994) in 0.9% saline (n=13 litters) into the lateral ventricle. This dose was selected because in Sprague-Dawley rats it was found to be effective at altering behavior when given intracerebroventricularly (Insel and Winslow, 1991; Pedersen and Boccia, 2006). The accuracy of the injection was confirmed visually by the presence of 0.025% fast green in the vehicle, since the brain and skull are translucent at E16.5. Thus, an injection was considered a “hit” when green dye flooded the ventricles. In the case of a surgical “miss”, the entire litter was excluded from study. The uterine horns were then gently returned to the body cavity, the muscle sutured with Unify^® PDO sutures [AD Surgical, Sunnyvale, CA, USA] and wound clips used to close the skin. Dams then continued with their pregnancies and underwent parturition. The experimental offspring were weaned at P21, group-housed by sex and allowed to develop until behavioral testing was initiated at 2–4 months of age.

Figure 1:

Schematic representation of trans-uterine fetal microinjection surgery. Created by E.A.A.

1.2b. Behavioral testing of offspring

Male (vehicle: n=10 from n=6 litters, OxtrA: n=11 from n=5 litters) and female (vehicle: n=9 from n=6 litters, OxtrA: n=7 from n=6 litters) offspring were used for each test in the order in which they are described below, with a few exceptions. The first is with the open field test, where three fewer female OxtrA mice (N=4) were included due to a tracking error during testing. The other exceptions are with the two-trial social discrimination and forced swim tests, where one fewer female vehicle mouse (N=8) was included after one of the cages flooded. Males underwent resident-intruder testing first and were subsequently given a one-week break prior to entering the next phase of testing. As a result, males and females did not undergo behavioral testing at the same time. Female offspring were freely cycling during testing, and the estrus cycles were not monitored. All testing occurred during the dark period of the light/dark cycle, one hour after lights out.

Resident-intruder test (males only)

Males were single-housed for two weeks without a cage change before undergoing testing. For three consecutive days, experimental and control males as well as gonadally- intact male Balb/C intruders [The Jackson Laboratory, Bar Harbor, ME, USA] were moved to the testing room and left to acclimate under red light for one hour. As described previously (Caldwell and Young, 2009; Dhakar et al., 2012; Wersinger et al., 2007a, 2007b), testing consisted of an intruder being placed into the resident’s home cage opposite of the resident. The resident and intruder were then permitted to interact for five minutes in the absence of an attack. If no attack occurred, the resident and intruder were permitted to interact for five minutes. Following an attack, testing was capped at two additional minutes. Thus, testing could last as little as two or as many as seven minutes per animal. All interactions were filmed and scored later for agonistic and aggressive behaviors using Observer^® XT software [Noldus Information Technology Inc., Leesburg, VA, USA] by a researcher with no knowledge of the experimental groups. Agonistic behaviors were scored up to the point of attack, and so, rate per minute is used instead of duration as total time spent being agonistic was variable. For aggressive behaviors, only the period of time after the initial attack was scored. For a list of all behaviors scored, see Table 1.

Table 1:

Table of behaviors scored during resident-intruder testing and their descriptions.

AGONISTIC BEHAVIORS
MEASURED IN RATE PER MINUTE
NON-SOCIAL BEHAVIOR	Facing away from intruder and/or across cage
SNIFFING	Investigation within 1cm of intruder
AFFILIATIVE GROOMING	Grooming of intruder in areas not easily reached and not impeding free movement
AGONISTIC GROOMING	Grooming of intruder such that movement is impeded and/or causing intruder to freeze
ANOGENITAL CONTACT	Contact and investigation of the intruder’s anogenital region
PURSUIT	Close, rapid following with contact outside the anogenital region
MEASURED IN COUNTS
MOUNTING	Close contact following of the intruder with pelvic thrusting
REARING	Resident stands on hind legs and extends body
SIDE THREATS	Approach from the intruder’s side followed by a bite or agonistic grooming
TAIL RATTLING	Rapid waving of the resident’s tail
BITING	As performed by the resident
AGGRESSIVE BEHAVIORS
MEASURED DURATION IN SECONDS
NON-SOCIAL BEHAVIOR	Facing away from intruder and/or across cage
SELF-GROOMING	Resident grooming himself away from intruder
CONTACT	Time spent physically interacting with intruder not during an attack, initiated by resident
LATENCY TO FIRST ATTACK	Time from introduction of intruder to first attack
AGGRESSIVE BEHAVIOR	Includes agonistic behaviors as well as attack sequence
MEASURED IN COUNTS
ATTACKS	A lunge followed by a bite followed by a roll
TAIL RATTLING	Rapid waving of the resident’s tail

Open field (males and females)

Male and female single-housed subjects were moved to the testing room and allowed to acclimate for one hour. As previously described (Rich et al., 2014a), subjects were placed in the center of the plexiglass arena (45.5 cm × 45.5 cm × 30 cm), measured at 200 lux prior to testing, and allowed to explore for 20 minutes. Movement was automatically tracked using Ethovision^® software [Noldus Information Technology Inc., Leesburg, VA, USA]. The arena was cleaned with 70% ethanol between mice and allowed to dry before another trial was initiated.

Elevated plus maze (males and females)

Male and female single-housed subjects were moved to the testing room and allowed to acclimate for one hour. Testing was performed as previously described (Rich et al., 2014b), where individual mice were placed in the center of the maze facing one of the closed arms (10 cm × 45 cm × 40 cm) and allowed to explore for 10 minutes. The center of the maze (10 cm × 10 cm) and both open arms (10 cm × 45 cm) were measured at 100 lux prior to testing. Movement was automatically tracked using Ethovision^® software. The maze was cleaned with 70% ethanol between mice and allowed to dry before another trial was initiated.

Two-trial social discrimination test (males and females)

Two weeks prior to testing, gonadally-intact C57BL/6J stimulus females were single-housed. The estrous cycles of stimulus were was not monitored, as the day of estrus does not have an effect on male investigation times when they are physically prevented from mounting (Ingersoll and Weinhold, 1987; Muroi et al., 2006). On the day of testing, male and female single-housed subjects were moved to the testing room and allowed to acclimate in red light for one hour. As previously described (Stevenson and Caldwell, 2014), thirty minutes before testing, two clean corrals were added to the left and right of experimental cages. Testing consisted of two trials, both of which were filmed. During Trial 1, one corral was removed, and a stimulus female placed under the other. The experimental animal interacted with the stimulus animal for five minutes, after which the stimulus female was removed, and the second corral replaced. Experimental mice reacclimated to both corrals for 30 minutes before the second trial. During Trial 2, the previous stimulus female (“familiar”) was placed under the corral opposite of the original placement, while a new stimulus female (“novel”) was placed under the other. The experimental animal interacted with the stimulus mice for five minutes before all mice were returned to their home cages. Video from both trials were scored at a later time using Observer^® XT software [Noldus Information Technology Inc., Leesburg, VA, USA]. Behaviors scored include duration of time spent engaged in non-social behavior (non-social is defined as facing away from a corral, digging, or self-grooming), duration of time spent investigating familiar mouse (investigating defined as nose pushed between the bars of the corral), and duration of time spent investigating the novel mouse (investigating defined as nose pushed between the bars of the corral).

Forced swim test (males and females)

Testing was performed as previously described (Caldwell et al., 2010; Rich et al., 2014b; Rodriguez et al., 2020). Briefly, twenty-four hours prior to testing, a plexiglass tank (19cm in diameter) was filled ¾ full of water and left overnight to reach room temperature (21° C). Filter topped cages of male and female single-housed subjects were moved to the testing room and allowed to acclimate in red light for one hour. Individual mice were placed gently at the top of the water and filmed for ten minutes. After testing, the mouse was dried off and returned to its home cage. Videos were scored for their entire length with the event sampling option using Observer^® XT software for time spent swimming versus floating.

Statistical analysis

Due to the known sex differences in the developing Oxt system (Cushing et al., 2003; Tamborski et al., 2016; Yamamoto et al., 2004; Young et al., 2005), differences in the behavioral testing sequence, i.e. males, but not females, undergoing resident-intruder testing, and other factors, the decision was made a priori that direct statistical comparisons between the sexes would not be appropriate. All analyses were performed using SPSS^® (IBM, Armonk, NY, USA). Resident-intruder: Individual behaviors were compared between treatments and across days using a repeated-measures analysis of variance (ANOVA). Those behaviors found to be significantly different (p<0.05) were further compared using a one-way ANOVA per day. Open field: A two-tailed t-test was used to compare time spent in inner versus outer arena within treatment groups to verify test efficacy. Total time spent in the inner versus outer part of the arena was compared between treatments using a one-way ANOVA. Elevated plus: A two-tailed t-test was used to compare time spent in open versus closed arms within treatment groups to verify test efficacy. Total time spent in the open versus closed arms was compared between treatments using a one-way ANOVA. Two-trial social discrimination: Within each treatment group the time spent investigating the familiar and novel stimulus mice in Trial 2 was compared using a paired sample t-test. Total investigation time in the second trial was compared within sex using a one-way ANOVA. Forced swim: Percent time spent swimming and floating was compared between treatments using a one-way ANOVA. Partial Eta squared (η_p²) and Eta squared (η²) were calculated using SPSS^® as measures of effect size for repeated-measures and one-way ANOVA analyses, respectively. Cohen’s d was used to estimate effect size for t-tests, using sample size and p-value (campbellcollaboration.org).

Results

Administration of an OxtrA at E16.5 does not compromise pregnancy or pup survivability

To determine if the surgery or injections caused any immediate negative impact on the pups, survival rate of the pups administered vehicle or the OxtrA 24 hours post parturition were examined. Of the 14 litters administered vehicle, 12 survived to P1 (86%), while 13 of the 13 litters administered the OxtrA survived to P1 (100%). For more detailed information on litter survival, see supplemental.

In males, administration of an OxtrA at E16.5 increased agonistic behavior

Among all agonistic behaviors analyzed, rates of anogenital contact (F_2,38=3.388, p=0.044, η_p²=0.151) and mounting (F_2,38 =4.520, p=0.017, η_p²=0.192) significantly differed across days between males administered the OxtrA and vehicle. Specifically, the rate of anogenital contact (F_1,19=9.181, p=0.007, η²=0.326) was higher in OxtrA-treated males on the third day (Figure 2A) while the rate of mounting was lower in OxtrA-treated males on the second day of testing (F_1,19=4.825, p=0.041, η²=0.203) (Figure 2B). Among the aggressive behaviors analyzed, latency to first attack significantly differed between groups across days (F_2,38=5.124, p=0.011, η_p²=0.212), and was significantly lower in males administered the OxtrA on the third day of testing (F_1,19=4.831, p=0.041, η²=0.203) (Figure 2C).

Figure 2:

Mean (+SEM) rates of anogenital contact (A) and mounting (B), as well as latency to attack (C) behaviors measured over three-days of resident-intruder testing in adult males administered vehicle (n=10) or an oxytocin receptor antagonist (OxtrA, n=11) at embryonic day 16.5. Figure 2A: The rate of anogenital contact significantly differed between groups across days (F_2,38=3.388, p=0.044), with males administered an OxtrA having higher rates of anogenital contact per-minute on the third day of testing than those administered vehicle (F_1,19=9.181, p=0.007). Figure 2B: The rate of mounting was significantly different across days between groups (F_2,38=4.520, p=0.017), and OxtrA-treated males had lower rates of mounting per-minute on the second day than those given vehicle (F_1,19=4.825, p=0.041). Figure 2C: Latency to attack was significantly different between groups across days (F_2,38=5.124, p=0.011), with OxtrA-treated males having a shortened time to attack on the third day than vehicle males (F_1,19=4.831, p=0.041).

In males and females, administration of an OxtrA at E16.5 had no effect on anxiety-like behaviors

Open field

Within group, males treated with vehicle (t(9)=15.123, p<0.001, Cohen’s d=3.188) or the OxtrA (t(10)=11.053, p<0.001, Cohen’s d=2.883), respectively, spent more time in the outer arena than the inner arena, but there were no significant differences between groups in terms of time spent in the inner (F_1,19=1.514, p=0.234, η²=0.074) or outer (F_1,19=1.154, p=0.234, η²=0.074) arena. Females treated with vehicle (t(6)=4.016, p=0.007, Cohen’s d=3.329) or the OxtrA (t(3)=4.863, p=0.017, Cohen’s d=7.574), when compared within group, similarly spent more time in the outer compared to the inner arena. When the groups were compared, neither spent significantly more time in the inner (F_1,9=0.719, p=0.418, η²=0.074) or outer (F_1,9=0.719, p=0.418, η²=0.074) arena.

Elevated plus

Males treated with vehicle (t(9)=−4.024, p=0.003, Cohen’s d=2.656) or the OxtrA (t(10)=−10.615, p<0.001, Cohen’s d=2.883) spent significantly more time in the closed arm than in the open arm respectively. Though, neither group spent significantly more time in the open (F_1,19=1.873, p=0.187, η²=0.090) or closed (F_1,19=0.697, p=0.414, η²=0.035) arm when the treatment groups were compared. Similarly, females treated with vehicle (t(8)=−9.368, p<0.001, Cohen’s d=3.605) or the OxtrA (t(6)=−7.749, p=<0.001, Cohen’s d=5.192) spent more time in the closed arm than open arm within their groups, but did not spend more time in the open (F_1,14=0.005, p=0.947, η²=0.000) or closed (F_1,14=1.163, p=0.299, η²=0.077) arms when groups were compared.

In females, administration of an OxtrA at E16.5 impaired social recognition memory

Neither males treated with vehicle (t(9)=−0.618, p=0.552, Cohen’s d=0.393) nor the OxtrA (t(10)=−0.481, p=0.641, Cohen’s d=0.291) spent more time investigating the novel stimulus over the familiar in the second trial of the social discrimination test, which suggests that the test did not work. However, males treated with the OxtrA did spent more time overall investigating the stimulus mice, as compared to control-treated males (F_1,19=4.591, p=0.045, η²=0.195) (Figure 3A). Females treated with vehicle displayed normal social recognition memory, i.e. they spent significantly more time investigating the novel stimulus over the familiar in the second trial (t(7)=−2.975, p=0.021, Cohen’s d=2.195). Conversely, females treated with the OxtrA did not have normal social recognition memory, as they did not spend significantly more time investigating the novel stimulus animal (t(6)=−0.285, p=0.785, Cohen’s d=0.218). Neither group of females spent more overall time investigating the stimulus mice (F_1,13=0.026, p=0.0876, η²=0.002) (Figure 3B).

Figure 3:

Mean (+SEM) investigation times, as measured by sniffing, of a novel and familiar stimulus animal during the second trial of a two-trial social discrimination task among males (A) and females (B) administered vehicle or an oxytocin receptor antagonist (OxtrA) at E16.5. Figure 3A: Time spent with the novel stimulus did not differ from time spent with the familiar stimulus in males treated with vehicle (n=10) (t(9)=−0.618, p=0.552) or males treated with an OxtrA (n=11) (t(10)=−0.481, p=0.641). However, total time spent investigating was significantly higher in OxtrA males compared to vehicle males (F_1,19=4.591, p=0.045). Figure 3B: In vehicle females (n=8), time spent investigating a novel stimulus was significantly different from time spent investigating a familiar one (t(7)=−2.975, p=0.021). For females treated with an OxtrA (n=7), time spent with a novel stimulus was not significantly different from time spent with a familiar stimulus (t(6)=−0.285, p=0.785). There was no difference in overall investigation time (F_1,13=0.026, p=0.0876).

Males, but not females, administered an OxtrA at E16.5 have increased depressive-like behaviors

Males treated with vehicle spent a greater duration of time swimming than those treated with the OxtrA (F_1,19=5.658, p=0.028, η²=0.229). As would be expected, the duration of time floating was the opposite (F_1,19=5.521, p=0.030, η²=0.225) (Figure 4A). Alternately, in females, there were no treatment-dependent differences in either swimming (F_1,13=0.461, p=0.509, η²=0.034) or floating (F_1,13=0.472, p=0.504, η²=0.035) (Figure 4B).

Figure 4:

Mean (+SEM) time spent swimming versus floating during a forced swim test among males (A) and females (B) administered vehicle or an oxytocin receptor antagonist (OxtrA) at E16.5. Figure 4A: Males administered vehicle (n=10) spent significantly more time swimming than those administered an OxtrA (F_1,19=5.658, p=0.028), while OxtrA-treated males (n=11) spent significantly more time floating than vehicle males (F_1,19=5.521, p=0.030). Figure 4B: Neither vehicle females (n=8) nor OxtrA-treated females (n=7) spent significantly more time swimming (F_1,13=0.461, p=0.509) or floating (F_1,13=0.472, p=0.504).

Discussion

In these experiments, we sought to functionally link the actions of embryonic Oxtr signaling to adult behavior by transiently disrupting Oxtr signaling in utero, at E16.5, followed by the examination of several behavioral endpoints in adult male and female mice. This study differs from previous work focused on the developmental role of Oxtr signaling in two important ways: 1) it is the first to centrally administer an OxtrA, as opposed to previous work, which used peripheral injections, and 2) it is the first to administer an OxtrA to mice still developing in utero. We found that adult mice administered an OxtrA during embryonic development, which transiently disrupted Oxtr signaling, had measurable sex-dependent shifts in behavior. When the Oxtr was antagonized at E16.5, adult males had increased agonistic and depressive-like behaviors, and spent more time investigating both the familiar and novel stimulus mice; there were no changes in anxiety-like behaviors or impairment of social recognition memory. On the other hand, females had disrupted social memory, but no changes in anxiety- or depressive-like behaviors in adulthood. These sex-specific effects are fairly consistent with previous experiments in which the Oxt system has been manipulated postnatally (Bales and Carter, 2003; Jia et al., 2008a; Mogi et al., 2014).

While the findings reported here contribute to the body of work that supports a role for the Oxt system in the sex-specific development of the brain, the data also suggests that there are key differences in the function of this system within the prenatal versus postnatal brain. Decreased intermale aggression was reported with a postnatal administration of an OxtrA, though that testing was in the context of parental care in a neutral arena (Bales and Carter, 2003; Jia et al., 2008b). The increase in agonistic and aggressive behaviors observed here have not been reported in any of the postnatal work, which may be due, at least in part, to our manipulation occurring prior to the presence of gonadal steroids in the embryonic brain (Konkle and McCarthy, 2011). This is of note as estrogen receptor α, as well as testosterone and its metabolites, are known to alter Oxtr binding and can drive Oxt gene expression in hypothalamus (reviewed in Jurek and Neumann, 2018).Moreover, the fact that treatment with an OxtrA on E16.5 had sex-specific effects on depressive-like, but not anxiety-like behaviors, is not a trivial finding and could be indicative of decreased resilience to stress in the males. Generally speaking, male rodents are more resilient to early life stress than female rodents (Goodwill et al., 2019). Though, following chronic stress in adulthood, they tend to have impaired cognition not observed in females (Luine et al., 2017). Moreover, chronic stress with multiple stressors can induce depressive-like behaviors in males that is associated with increased Oxtr expression (Lesse et al., 2017). So, it is possible that males administered an OxtrA were more strongly affected by repeated testing than other groups. Moving forward, it will be crucial to our understanding of the prenatal Oxt system to tease out which components of these contextual behaviors are the result of prenatal activity of the Oxt system.

The results of our social recognition memory test, however, align with the postnatal work in voles and mice, in which select, sex-specific, differences are observed during sociability tests (Bales et al., 2004b; Jia et al., 2008a; Mogi et al., 2014). These earlier reports consistently found that females with postnatal manipulations of their Oxt system had decreases in sociability that were not observed in males. Here, we found that while the two-trial social recognition memory test did not seem to “work” in males, given that males administered vehicle did not spend more time investigating the novel stimulus animal compared to the familiar stimulus, males administered the OxtrA did spent significantly more time investigating stimulus mice than males given vehicle. As we do not know if this OxtrA-facilitated shift in investigation time is reflective of impaired sociability in these mice, we plan to repeat this test in the future. Still, the fact that a sex-specific difference in social memory is observed when the manipulation is prenatal or postnatal suggests that that the neural circuitry associated with these behaviors, or the mechanism by which they are elicited, is already in place prior to birth.

Clues to the subtle alterations in behavior that resulted from the transient disruption in Oxtr signaling at E16.5 may lie in the many functions the Oxt system performs in embryonic brain development. Oxt is known to influence cell fate specification, tending to favor glial cell types as evidenced by the increased expression of mRNA and protein markers for astrocytes in rats with neonatal manipulations of the Oxt system (Bakos et al., 2016; Havránek et al., 2017). Additionally, prior to birth, Oxt contributes to the facilitation of the switch of GABA from excitatory to inhibitory (Leonzino et al., 2016; Tyzio et al., 2006). When Oxt is absent and the timing of this switch is altered, there is an increase in the likelihood of seizures and impairments in social behavior expression in mice (Ripamonti et al., 2017; Sala et al., 2011). The neonatal Oxt system also exerts an organizational effect on the expression of estrogen receptor α in a sex- and region-specific manner (reviewed in Cushing and Kramer, 2005; Yamamoto et al., 2006). It is possible that disrupting Oxtr signaling at E16.5 may impact several of these processes, resulting in the behavioral alterations described herein. This may be particularly relevant to the differences noted in the agonistic and aggressive behaviors of males administered the OxtrA who had changes in anogenital contact, mounting, and attack latency. These shifts in specific components of agonistic/aggressive behavior suggest brain region-specific effects. The amygdala is likely to be involved, as it is one of few places expressing the Oxtr at E16.5 and plays a role in inter-male aggression (Cushing and Kramer, 2005; Dhakar et al., 2012; Wang et al., 2013; Calcagnoli et al., 2015). Region-specific effects have also been found in voles when the Oxt system is manipulated in early postnatal life, resulting in differences in activation among brain regions important to neural regulation of social behavior (Cushing et al., 2003). However, to assess any region-specific effects, more work must be done to understand how disruptions of the Oxt system alter cellular and/or molecular functioning.

While the exact molecular mechanisms that contribute to our findings remain unknown, it is important to consider the development of the Oxt system, in particular, its development in males where the offset between the appearance of the Oxtr protein and transcripts for Oxt is quite dramatic. Perhaps males get their Oxt from elsewhere or use an alternate ligand for the Oxtr. While we did not test this directly, our findings do indicate that disrupting Oxtr signaling in embryonic males does result in measurable changes in behavior later in life. This suggests that the Oxtr in males is functional and may use an alternative ligand. If this is the case, one likely candidate is arginine vasopressin (Avp). Avp is present in both the male and female brain during embryonic development, including at E16.5 (Aulino and Caldwell, 2020). Evolutionarily linked, Avp and Oxt are sister hormones capable of signaling through one another’s receptors. Interestingly, in adult hamsters, this crosstalk has been found to lead to different behavioral outcomes (Song et al., 2016b, 2016a). Given that that male mice produce more Avp mRNA than females at E16.5 (Aulino and Caldwell, 2020), but do not produce their own Oxt prior to birth (Tamborski et al., 2016), the Avp system could act in a sex-dependent manner in embryonic brain development that compliments the activity of Oxt. In addition, Avp appears in its mature form sooner in embryonic development than Oxt (Whitnall et al., 1985), which suggests a different role for the peptide. Such impacts could also be due to the Oxt and Avp systems interactions with steroid hormones, which are known to have a sex-specific influence on behavioral expression in later development and adulthood (Cushing, 2013). More work must be done to confirm the specific effects of Avp on social brain development, as well as how the Oxt and Avp systems interact at this timepoint.

In summary, the data presented here provide the first functional data demonstrating that manipulation of the embryonic Oxt system, specifically Oxtr signaling, has sex-specific effects on the expression of behaviors important to social interactions. Previously, we identified E16.5 as a key timepoint for the development of the Oxt system as it is the first recorded appearance of the Oxtr (Tamborski et al., 2016). These findings build on our earlier work and suggest that the Oxt system at that time is a key player in sex-specific brain development and the later expression of behavior. While these data support previous findings in postnatal development, which have documented a sex-specific role for the Oxt system, they also suggest that Oxtr signaling prior to birth has a distinct function. However, more work must be done to understand how the Oxt system influences brain development at both the molecular and cellular levels. Finally, the interactions of the Avp and Oxt systems at E16.5, and how much the Avp system contributes to the sex-specific effects noted here, are largely unknown. Given the evolutionary history of these two neuropeptides, it is likely that understanding their interactions during this important time in development will clarify what we know of sex differences in the development of the social brain.

Highlights.

• We transiently disrupted oxytocin receptor signaling in the mouse embryonic brain

• Brief disruption of embryonic oxytocin receptor signaling affects adult behaviors

• Embryonic oxytocin receptor disruption results in sex specific behavioral effects