A novel social fear conditioning procedure alters social behavior and mTOR signaling in differentially housed adolescent rats.

Abstract

Vulnerabilities to fear-related disorders can be enhanced following early life adversity. This study sought to determine whether post-weaning social isolation (PSI), an animal model of early life adversity, alters the development of social fear in an innovative model of conditioned social fear. Male and female Sprague-Dawley rats underwent either social rearing (SR) or PSI for 4 weeks following weaning. Rats were then assigned to groups consisting of either Footshock only, Social conditioned stimulus (CS) only, or Paired footshock with a social CS. Social behavior was assessed the next day. We observed a novel behavioral response in PSI rats, running in circles, that was rarely observed in SR rats; moreover, this behavior was augmented after Paired treatment in PSI rats. Other social behaviors were altered by both PSI and Paired footshock and social CS. The mammalian target of rapamycin (mTOR) pathway was assessed using immunohistochemistry for phosphorylated ribosomal protein S6 (pS6) in subregions of the prefrontal cortex (PFC) and amygdala. Paired treatment produced opposite effects in the PFC and amygdala in males, but no differences were observed in females. Conditioned social fear produced alterations in social behavior and the mTOR pathway that are dependent upon rearing condition and sex.

1. Introduction.

Stress-related disorders pose a major health problem, and those that involve social components are particularly problematic due to the necessity of a social structure for an ideal quality of life. Vulnerability to fear-related disorders such as post-traumatic stress disorder (PTSD) and social anxiety disorder (SAD) is enhanced in individuals with histories of early life adversity (Maercker et al., 2013). Early emotional neglect (Solomon & Mikulincer, 2007) and early social adversity are particularly potent risk factors for anxiety disorders including SAD (Kertz et al., 2017). Social fear can profoundly and persistently impair the quality of an individual’s functioning across many domains, as social interactions are a vital component of almost every aspect of life (Karatzias, 2017; Herman,1992).

Previous studies in our laboratory (Wall et al., 2012; Grotewold et al., 2014; Goodell et al., 2017; Fontenot et al., 2018) and others (Ferdman et al., 2007; Toth et al., 2008; Zhao et al., 2009) have used post-weaning social isolation (PSI) as an animal model of early life adversity. PSI consists of depriving adolescent rats of social experience by housing them individually (as compared to housing in same-sex groups [Socially Reared], typically 3 or 4 per cage) for a period of 4 to 8 weeks after weaning (Fone and Porkess, 2008). This period encompasses the adolescent period, which in the rat has been described as lasting roughly from P28 to P60 (Spear, 2000). Adolescence in the rat has also been defined as beginning as early as P21 and consisting of early (P21-P30), middle (P30-P45), and late (P45-P60) periods (Tirelli et al., 2003). Social isolation during this period produces persistent behavioral changes known as the isolation syndrome (Hatch et al., 1965) that includes outcomes such as increased anxiety-like behavior (Einon and Morgan, 1977) and overreaction to novelty (Hall et al., 1998). We (Wall et al., 2012, Grotewold et al., 2014, Goodell et al., 2017) and others (Toth et al., 2008, 2011; Wongwitdecha and Marsden, 1996) have observed increased aggressive behavior as well as overall increases in social interaction after PSI, and we have observed this in both males and females (Wall et al., 2012, Grotewold et al., 2014; Fontenot et al., 2017). We have also observed that after PSI rats spend less time escaping a stimulus rat, in spite of being more aggressive (Goodell et al., 2017). This suggests that a stimulus rat may have greater salience to a previously isolated rat, and that consequently a socially conditioned fearful experience would have a more pronounced effect after PSI, resulting in greater and/or longer lasting fear of the stimulus rat or even of a rat that is similar to the stimulus rat. It is important to note that PSI did not alter the acquisition of conditioned fear in a standard conditioned fear paradigm, although extinction was impaired (Skelly et al., 2015).

Conditioned fear and its extinction are frequently used to model normal fear learning as well as PTSD and other stress-related disorders. The canonical conditioned fear paradigm involves the pairing of a context or one or more footshocks with a tone (LeDoux, 2000). However, traumas involving interpersonal violence may be more likely to lead to PTSD than those that do not, such as accidents or natural disasters (Yehuda & LeDoux, 2007), thus a context or a tone may not capture the emotional nature of these types of trauma. Social trauma can be investigated using the social defeat model but there are potential benefits of using a footshock/social cue conditioned fear approach. For example, female rats are less aggressive than males and are difficult to equate with males using social defeat models (Solomon, 2017; Haller et al., 1999). Moreover, unlike social defeat, the use of a footshock unconditioned stimulus (US) can be equated across animals, does not cause wounding, and can be controlled by the experimenter.

Toth et al. (2012, 2013) have reported a method of social fear conditioning in which rats received a footshock when they initiated social interaction with a novel conspecific that had been placed in the apparatus within a mesh cage. In that model, animals received footshocks that depended on how often social contact was made, thus active approach behavior was punished. Our procedure differs from that of Toth et al. (2012, 2013) by allowing for standardization of the number of shocks delivered, by explicit pairing of the conditioning stimulus (CS) and US, by ensuring that the number and timing of shocks received does not vary according to the sociability of the subject, and by assessing social interaction during re-exposure. The work of Toth et al. (2012, 2013) was the first to provide a valuable model of social fear learning using operant conditioning procedures, it is of interest to develop a Pavlovian model as the two types of conditioning differ in a number of dimensions. For example, operant conditioning requires a voluntary response, and some types of learned social fear may be better modeled using involuntary responses to aversive conditioning stimuli. In addition, the neural substrates of these different types of learning are distinct (Campese et al., 2016; LeDoux et al., 2017), thus it is of value to develop an additional fear conditioning paradigm for assessing the neural underpinnings of classically conditioned social fear.

A rich literature exists describing the behavioral consequences of Pavlovian fear conditioning, and the neural pathways that mediate it are well characterized (LeDoux, 2012). Freezing behavior is the most widely used measure of conditioned fear, however, it is only one component of a repertoire of defensive behaviors (Blanchard et al., 1975). The premise of cued fear conditioning is that neutral stimuli present during a stressful event can subsequently produce a conditioned fear response. This response may vary depending on the nature of the specific stimulus as well as the context in which the stressful event occurs. Conditioned fear responses comprise a continuum of adaptive behaviors including freezing (the canonical dependent measure in auditory fear conditioning studies), fighting, or flight (Blanchard, 1975), which become maladaptive if fear pathologies develop. However, most studies of conditioned fear use auditory stimuli as the CS and freezing as the conditioned response. We have developed a novel fear conditioning paradigm in which a footshock US is explicitly paired with a social stimulus (a novel same-sex conspecific) as the CS. As the primary deficits following PSI are social in nature, social fear conditioning may illuminate a vulnerability in rats exposed to PSI and may have utility in other models of social anxiety as well.

The neural circuitry thought to mediate learned fear in classical fear conditioning paradigms has been well established and includes the medial prefrontal cortex (mPFC) and amygdala, although a role for this circuitry has not been explored in classically conditioned social fear. This conditioned fear circuitry includes the ventral aspect of the mPFC which comprises the prelimbic (PL) and infralimbic (IL) subregions. The PL and IL have opposing and dissociable functions in processes important for conditioned fear and its persistence. The PL is especially important for the acquisition and expression of conditioned fear, while the IL is critical for fear extinction and extinction memory (Sierra-Mercado et al., 2011). The ventral orbital (VO) region of the PFC may also play a role (Pujara et al., 2019). The amygdaloid complex, including the basolateral (BLA) and central (CeA) nuclei, play an essential role in fear conditioning and extinction (Maren & Fanselow, 1996), in part due to their reciprocal connections with the mPFC (Giustino & Maren, 2015). There is evidence that the medial amygdala (MeA) is involved in fear conditioning as well (Tsuda et al., 2015), and because of its important role in social behavior (Vochteloo & Koolhaas, 1987) the MeA may be particularly important for social fear.

The development and recall of conditioned fear requires neural plasticity, which can be provided by intracellular signaling pathways such as the mammalian target of rapamycin (mTOR) cascade. mTOR is a serine/threonine protein kinase that forms the catalytic subunit of two protein complexes, mTOR Complex 1 (mTORC1) and 2 (mTORC2) that regulate distinct cellular processes. Ribosomal protein S6, a downstream target of mTOR, is a component of Complex 1 and phosphorylated (p)S6 has been used as a specific marker of mTORC1 activation (Li et al. 2019, Yuskaitis et al., 2018). The mTORC1 cascade triggers mRNA translation and protein synthesis important for neuronal plasticity, while mTORC2 is associated with cell survival and proliferation (Saxton et al. 2017; Hoeffer and Klann, 2010).. (von Manteuffel et al., 1997), Evidence supports the involvement of mTOR and pS6 with fear memory processes. For example, acute tailshock stress induced the phosphorylation of mTOR and several downstream signaling molecules in the hippocampus (the only region assessed) when measured by western blot (Yang et al., 2008). Moreover, blockade of mTOR with rapamycin prevented the potentiating effect of tailshock stress on long-term depression (LTD; Yang et al., 2008), and rapamycin prevented the ability of corticosterone administration to enhance memory of contextual fear conditioning (Xiong et al., 2015). Phosphorylation of p70s6K, an enzyme in the mTOR pathway immediately upstream of S6, was reduced in the IL and increased in the BLA during retrieval of a fear memory (Levin et al., 2017). Consolidation of extinction memory for contextual conditioned fear was blocked by intra-mPFC microinjection of the mTOR inhibitor rapamycin in rats (Penha-Farias et al., 2019). Interestingly this effect was observed whether a social support animal was present or not (Penha-Farias et al., 2019).

Here we used PSI in male and female rats to explore social fear vulnerability in a novel model of conditioned social fear: Pavlovian conditioned fear to a footshock US paired with a social CS in male and female rats. We hypothesized that rats that had experienced PSI would exhibit more conditioned social fear than socially reared rats. We then assessed pS6 expression in subdivisions of the PFC and amygdala using immunohistochemistry.

2. Material and methods.

2.1. Animals.

Male (N=60) and female (N=60) Sprague-Dawley rats (Envigo, Indianapolis, IN, USA) were purchased at post-natal day (P) 21 and housed for 4 weeks in either individually (PSI) or in same-sex groups of 3 (SR); during experimentation rats were returned to the same housing conditions. PSI rats were exposed to the sight, sound, and smell of other rats within the colony room. All rats were on a 12:12 hr light/dark cycle maintained at 23 ͦC in standard Plexiglas cages with food and water freely available. Experimental procedures were conducted during the light phase (lights on at 0700 h). All rats were unhandled and did not receive enrichment during 4 weeks of housing as additional handling and enrichment attenuate the effects of social isolation (Krebs-Thomson et al., 2001; Rosa et al., 2005; Sciolino et al., 2010; Gentsch et al., 1982). Estrous cycles were not monitored in females to avoid the stress of additional handling (Sfikakis et al., 1996). Younger (P35) naïve socially reared rats were used as stimulus (CS) rats during the experiments in order to be less threatening to the experimental rats. CS rats were each used twice during the experiment to minimize use of rats and were counterbalanced by experimental rat group. All experiments were approved by the University of Colorado Institutional Animal Care and Use Committee.

2.2. Conditioned Social Fear: Day 1 in Context A.

A timeline of all experimental procedures is shown in Figure 1A. Following 4 weeks of PSI or SR, social fear conditioning took place. Following 4 weeks of differential housing, experimental rats were assigned to three groups (n = 10 rats per group) based on Day 1 conditions: social CS only, footshock only, or paired footshock with a social CS. Conditioned social fear took place in a Plexiglas operant conditioning chamber (17 ¾ L, 12 ¼ W, and 6 ¼ H) with a Coulbourn grid floor connected to a current generator. A platform consisting of a Plexiglas floor (on the CS side) with a non-conductive mesh divider was inserted into the chambers to separate the CS rat from the experimental rat while protecting the CS rat from the shock (the apparatus is shown in Figure 1B). The mesh (Home Depot) consisted of non-conductive aluminum with 1 cm grid allowed the CS rat and experimental rat a chance to interact through sniffing, nose pokes, sight, and smell prior to the shock. Conditioning occurred under white light identical to standard housing lights. The chambers were washed with Parsley Plus all-purpose cleaner (Earth Plus) to provide a novel and distinct odor to distinguish it from Context B (see below). Experimental rats were placed into the chamber on the side with the exposed shock bars for 5 minutes to habituate to the chamber. Following habituation, CS rats were placed through the front door on the side with the Plexiglas flooring for 30 seconds to allow for the experimental rat to smell, see, hear, and nose poke the CS rat. After 30 seconds, a single 2 second 0.8 mA shock was delivered to the experimental rat. Previous research indicates that a single footshock CS is sufficient to produce a conditioned fear response (Fanselow, 1990; LeDoux, 1993; Woodruff et al., 2015). Immediately following the footshock, rats were returned to their previous housing conditions.

Figure 1.

Timeline of the experimental procedures (A). Fear conditioning was performed in a standard conditioning chamber that was modified with a wire divider and Plexiglas platform to accommodate a naive stimulus rat that did not receive a footshock (B). This allows for sensory, but not physical contact between the naïve stimulus rat and the experimental rat.

2.3. Social interaction testing: Day 2 in Context B.

The following day, experimental rats were given 10 minutes to socially interact with the same CS rats they were exposed to on Day 1 (Footshock Only rats were exposed to a novel stimulus rat). The social interaction chamber was a clear Plexiglas box (14 ¾ L, 14 ¾ W, 14 ¾ H) under red light conditions. Either cedar or pine bedding was used to provide a novel, distinctive odor to distinguish it from Context A and minimize any possible contextual memory. Experimental rats were allowed 5 minutes to explore and habituate to the chamber before the stimulus rat was placed in the chamber for a 10-minute social interaction trial. All interactions were video recorded using a Logitech HD720p webcam. Immediately following the social interaction test, experimental rats were placed into a cage and transferred to a procedure room for perfusion. We have performed a pilot study (data not shown) to determine the optimal time point for pS6 expression and have found that pS6 expression is very transient with peak expression observed immediately after a challenge.

2.4. Behavioral Analysis.

Social and avoidance behaviors (descriptions of social behaviors that were assessed are provided in Table 1) were analyzed from digital video files by researchers blind to experimental conditions using JWatcher v1.0 software (http://www.jwatcher.ucla.edu/).

Table 1.

Descriptions of social behaviors of the experimental rats during re-exposure to the stimulus rat on Day 2.

Behavior	Characteristic
Total Social Behavior	The overall amount of time that the experimental rat spends actively engaged in any type of social interaction with the novel conspecific (Fontenot et al., 2018).
Aggressive Grooming	Vigorous grooming where fur is being pulled and teeth are used by the experimental rat toward the stimulus rat (Hurst et al., 1999).
Other Social Behavior	Other social behavior was calculated as Total Social Interaction minus Aggressive Grooming (Fontenot et al., 2018).
Running Away	The experimental rat avoids and runs or darts away from the stimulus animal.
Running in Circles	The experimental rat runs in an open circle or spins on an axis when approached by the stimulus rat.

2.5. Immunohistochemistry.

Rats were deeply anesthetized with Fatal Plus (Vortech Pharmaceutical Dearborn, Michigan). After intracardiac perfusion with 4% paraformaldehyde in 0.01 M PBS, brains were removed and postfixed for 24 hours with 4% paraformaldehyde. Afterwards, the brains were transferred and cryoprotected in 30% sucrose for at least 72 hours until immunohistochemistry was performed (Wall et al., 2012).

Prior to slicing, brains were removed from the 30% sucrose solution and rapidly frozen using isopentane at −20 C. Coronal brain sections (30 μm) of the mPFC and VO (3.72 mm to 2.52 mm anterior to bregma) and Amygdala ( −2.38 mm to −3.24 mm) were collected using a precision cryostat (Leica CM1850) and the atlas of Paxinos and Watson, 2006 for coordinates. Sections were collected directly into a set of six wells filled with cryoprotectant (300 g Sucrose, 9g NaCl, 500 mL 0.1 M PB, 300 mL propylene glycol) and stored at 4 ͦC.

The mPFC, VO, and amygdala were immunolabeled for pS6. We used immunohistochemistry (IHC) rather than Western blot in order to obtain adequate spatial resolution to detect small subdivisions of specific brain regions. Due to the large size of the experiment, male and female tissues were run in separate assays. A subset of male and female tissues (SR, Social Only males and females) were run side by side in a separate IHC to determine sex differences in pS6 expressionThree sections were selected for each rat and placed in 0.01M PBS in a 25-well plate. Sections were washed 3 times with 0.01M Tris-Buffered Saline (TBS) between steps on the first day and in 0.01 M PBS on the second day of immunohistochemistry unless otherwise stated. Sections were incubated in 0.3% H₂O₂ (TBS) for 40 minutes. Sections were incubated in normal goat serum (NGS) blocking solution (TBS, 30% Triton X, NGS) for 1 hour then incubated in primary rabbit anti-Phosphor-S6 Ribosomal (1:500; Rabbit Ab S235/236; Cell Signaling Technology, Lot: 23) overnight at room temperature. The next day, sections were washed 3 times in 0.01 PBS then incubated in secondary biotinylated goat anti-rabbit (Biotin-Sp-conjugated AffiniPure Goat Anti-Rabbit IgG Jackson ImmunoResearch Laboratories INC. Lot: 128074) IgG 1:200 in NGS blocking solution for 2 hours. Sections were then incubated in ABC solution (Vectastain ABC Kit Lot # ZD0426 Vector Laboratories Inc. Burlingame, CA 94010) for 2 hours prior to staining with SG Substrate (Vectastain SK-4700 Vector Laboratories INC. Lot: ZE0613). Sections were incubated in SG for 20–30 minutes. Sections were then mounted onto slides using a 0.02% gelatin solution, dehydrated using an ethanol procedure, and then cover slipped using Permount (Sigma-Aldrich) for microscopy.

2.6. Microscopy.

Microscopy was done using an Olympus BX51 and VisioPharm software (Horscholm, Denmark). Researchers blind to experimental procedures counted immunolabeled neurons using Paxinos and Watson (2006) as a guide. Within the mPFC, PL, IL and VO subregions were analyzed. Within the amygdala, the CeA, BLA, and medial ventral (MeV) subregions were analyzed. Neuronal cell counts were performed at 20x for all subregions. Six hemispheres (mPFC: 3.72 mm to 2.52 mm, VO: 4.20 mm to 2.7 mm and Amygdala: −2.38 mm to −3.24 mm) were analyzed per rat. Cell counts were performed within the amygdala in real time using VisoPharm software and within the mPFC using ImageJ software (https://imagej.nih.gov/ij/index.html) and a particle counting macro written in house and used on saved image files captured using VisoPharm software. Figure 2 shows representative photomicrographs of pS6 immunolabeling (left) and placement of the counting frames in subregions of the mPFC (middle) and amygdala (right).

Figure 2.

Left, representative photomicrographs of phosphorylated ribosomal S6 (pS6) immunoreactivity in subregions of the medial prefrontal cortex and amygdala. P, Prelimbic region; I Infralimbic region; O, Ventral orbital region of the prefrontal cortex; C, Central nucleus of the amygdala B, Basolateral amygdala; V, Medial ventral amygdala. White arrows indicate examples of pS6 labeled cells. Right, location of the field of view (not to scale), plates 9 (left) and 30 (right) from Paxinos & Watson (2006).

2.7. Statistics.

Behavioral and immunohistochemistry results were analyzed using SPSS Statistics v24. Behavioral data (either time [seconds] or frequency) were analyzed using a 3 (condition: social only, footshock only, or paired footshock and social cue) x 2 (rearing: SR or PSI) x 2 (sex: Male or Female) factorial ANOVA. Where significant interactions were observed, Fisher’s PLSD post-hoc tests were used to determine if there were differences between groups. Immunohistochemistry results were analyzed using 3 (condition) x 2 (rearing) factorial ANOVAs. Male and female immunohistochemistry data were analyzed separately because male and female tissues were run in different assays. Data from the side by side immunohistochemistry experiment comparing a subset of male and female rats (SR social only) were analyzed using t-tests. Alpha was set at 0.05 for all experiments.

3. Results.

3.1. Social and Avoidance Behavior.

To assess whether pairing a footshock US with a CS rat produced alterations in responding to the CS rat when tested in a different environment the next day, we analyzed several avoidance and social behaviors in male and female rats that had experienced either PSI or SR. We assessed two types of avoidance behavior: Running in Circles and Running Away, and 2 types of social behavior, Aggressive Grooming and Other Social Interaction (calculated as Total Social Interaction minus Aggressive Grooming).

3.1.1. Running in Circles.

In both males and females, Paired US + CS rat (Paired) on Day 1 produced an increase in Running in Circles on Day 2 that was specific to PSI rats (Figure 3, top). This was reflected by a significant Conditioning by Housing interaction, F (2, 108) = 4.27, p < 0.05. ANOVA revealed a significant main effect of Housing on Running in Circles, F (1, 108) = 48.89, p < 0.0001; PSI rats ran in circles much more frequently than SR rats did. ANOVA also revealed a significant main effect of Conditioning, F (2, 108) = 4.63, p < 0.05. Post hoc analysis revealed that Paired rats ran in circles more frequently than both US only and CS rat only rats, p < 0.05. There was no main effect of Sex nor were there any interactions involving Sex (all p values > 0.50).

Figure 3.

Paired footshock and social cue altered social avoidance behavior in male and female rats that had experienced post-weaning social isolation (PSI) or social rearing (SR). Top, running in circles was increased in Paired PSI rats (males and females) compared to all other groups (@ p < 0.05). Running in circles was greater in PSI rats (males and females) compared to SR rats (**p < 0.001). Bottom, running away from the stimulus rat was greater in PSI rats than SR rats (**p < 0.001), with only a trend for increased running away in Paired rats.

3.1.2. Running Away.

Rats that had experienced PSI exhibited a much greater frequency of Running Away from the CS rat (Figure 3, bottom). ANOVA revealed a significant main effect of Housing, F (1, 108) = 43.84, p < 0.0001. There was only a slight trend for Conditioning, p = 0.13, such that Paired rats ran away with somewhat greater frequency than the other Conditioning groups. There was also a modest trend for Sex; p = 0.10; males ran away somewhat more frequently than females. There were no significant interactions (all interaction p values > 0.33).

3.1.3. Aggressive Grooming.

We have previously observed that PSI treatment dramatically increased Aggressive Grooming in males and females, and here we analyzed the effects of pairing of an aversive US with a CS rat on aggressive grooming after PSI or SR (Figure 4, top). There was a significant Housing by Sex interaction, F (1, 108) = 7.19, p < 0.01. Post hoc analysis revealed that although PSI males and females did not differ from each other, all other Housing by Sex groups differed from each other, p < 0.05. This result was likely driven by the trend for a Housing by Conditioning interaction, F (1, 108) = 2.71, p < 0.07, such that SR males in the Paired condition had increased Aggressive Grooming relative to SR males in both the US only and CS rat only groups (both p < 0.05). PSI produced a robust increase in Aggressive Grooming, as revealed by a significant main effect of Housing, F (1, 108) = 44.03, p < 0.001.

Figure 4.

Aggressive and other social behavior in male and female rats exposed to either social rearing (SR) or post-weaning social isolation (PSI), then conditioned social fear procedures. Top, Aggressive grooming was greater overall in PSI females than SR females (** p < 0.01). SR males in the Paired condition engaged in somewhat more aggressive grooming than Footshock only and Social only controls and were indistinguishable from PSI males. Bottom, other social behavior (total social interaction minus aggressive grooming) was greater in females than males (*** p < .001). Non-aggressive behavior was greater in PSI rats compared to SR rats (** p < 0.001). In SR males, Paired treatment produced a decrease in non-aggressive social interaction relative to Footshock only (# p < 0.05).

3.1.4. Other Social Behavior.

Non-Aggressive Social Behavior (Figure 4, bottom) was calculated by subtracting Aggressive Grooming from total social interaction time. ANOVA revealed a significant Conditioning by Sex interaction, F (2, 108) = 4.20, p < 0.05. Post hoc analysis revealed that Paired males exhibited less Other Social Behavior than US only males and Paired females, p < 0.05. In addition, CS rat only males exhibited less Other Social Behaviorthan CS rat only females, p < 0.05. There was a Housing by Sex interaction, F (2, 108) = 5.35, p < 0.05. Post hoc analysis revealed that PSI produced an increase in Other Social Behavior in females relative to males, p < 0.05.

3.2. Expression of pS6 in Subregions of the Prefrontal Cortex and Amygdala.

To assess plasticity in brain regions known to be involved in learned fear, we used IHC to measure the expression of pS6. Because male and female brains were run in separate IHC assays, they were analyzed using separate ANOVAs.

3.2.2. Prefrontal Cortex.

Male PFC data are shown in Figure 5 (left; PL:top, IL:middle, and VO:bottom

Figure 5.

Phosphorylated ribosomal S6 (pS6) expression in subregions of the prefrontal cortex and amygdala in male rats exposed to either social rearing (SR) or post-weaning social isolation (PSI), then conditioned social fear procedures. Prelimbic (PL), there were no group differences in pS6 expression. Infralimbic (IL), there was less pS6 expression in Paired rats (both SR and PSI) compared to rats that only received a footshock (* p < 0.05). Ventral Orbital (VO), there was less pS6 expression in Paired rats (both SR and PSI) compared to rats that only received a footshock (* p < 0.05). Basolateral Amygdala (BLA), there was more pS6 expression in Paired rats in the SR group compared to Footshock only rats (* p < 0.05). Central Amygdala (CeA), there were no effects of housing or condition. Medial Ventral Amygdala (MeV), there was more pS6 expression in Paired rats in the PSI group compared to footshock only rats (* p < 0.05).

Prelimbic Cortex.

Neither Conditioning nor Housing impacted pS6 expression in the PL. There were no significant main effects and no interactions (all p values > 0.80).

Infralimbic Cortex.

pS6 expression in the IL depended upon the conditioning group. ANOVA revealed an overall effect of Conditioning on pS6 expression in the IL, F (2,54) = 5.53, p < 0.01. Post hoc analysis revealed that there was less pS6 expression in the Paired group compared to the US only group regardless of Housing condition, p < 0.01, but there were no significant differences between Paired and CS rat only or US only and CS rat only, both p = 0.10. There was no effect of Housing (p = 0.29) and no Housing by Conditioning interaction (p = 0.52) on pS6 expression in the IL.

Orbital Cortex.

Overall, there was less pS6 expression in VO in Paired rats. ANOVA revealed a significant main effect of Conditioning, F (2, 54) = 5.99, P < 0.01. Post hoc analysis revealed that there was significantly less pS6 expression in the VO of Paired rats compared to both US only and CS rat only regardless of housing condition, both p < 0.05. There was no effect of Housing (p = 0.29) and no Housing by Conditioning interaction (p = 0.52) on pS6 expression in the VO.

3.2.3. Amygdala.

Male Amygdala data are shown in Figure 5 (right; BLA: top, CeA: middle, and MeV: bottom).

Basolateral Amygdala.

Overall, there was more pS6 expression in the BLA of Paired rats. This was supported by a significant main effect of Conditioning, F (2, 54) = 3.92, P < .05. Post hoc analysis revealed greater pS6 expression in Paired vs US only, p < .05; and a strong trend for Paired vs CS rat only, p = .07.

Central Amygdala.

There were no main effects of Housing or Conditioning on pS6 expression in the CeA and there were no interactions, all p values > 0.12.

Medial Amygdala.

In the MeA, there was less pS6 expression in rats that received US only. There was a significant main effect of Conditioning, F (2, 54) = 3.74, P < .05. Post hoc analysis revealed less pS6 in US only vs Paired and US only vs CS rat only, both p < .05.

3.3.2. Prefrontal Cortex.

Female PFC data are shown in Figure 6 (left; PL: top, IL: middle, and VO: bottom).

Figure 6.

Phosphorylated ribosomal S6 (pS6) expression in subregions of the prefrontal cortex and amygdala in female rats exposed to either social rearing or postweaning social isolation, then conditioned social fear procedures. There were no differences between groups in any brain regions.

In contrast to the effects of Housing and Conditioning that we observed in male rats, there were no effects of our manipulations on pS6 expression in the PFC of females. There were no significant effects of Housing or Conditioning and no interactions in PL, IL, or VO; all p values > 0.45.

3.3.3. Amygdala.

Female Amygdala data are shown in Figure 6 (right; BLA: top, CeA: middle, and MeV: bottom).

In the BLA, there were no effects of Housing or Conditioning and no interactions, all p values > 0.16. There was a tendency for PSI females to have greater pS6 expression in the CeA; as revealed by a trend for Housing effect, F (2, 54) = 3.54, P = .06. In the MeV, there were no effects of Housing or Conditioning and no interactions, all p values > 0.27.

3.4. Male and female comparison.

Data from a side by side comparison of a subset of male and female rats (SR social only rats) are shown in Table 2. In the IL, pS6 expression was greater in males than in females, t (15) = 4.78, p < .05. There were no other significant effects of sex in subregions of the PFC. In the MeV, pS6 expression was greater in males than in females, t (15) = 8.08, p < .05. There were no other effects of sex in subregions of the amygdala.

Table 2.

Expression of pS6 in subregions of the PFC and amygdala in male and female rats. A side-by-side immunohistochemistry assay was performed in a subset of male and female rats in which tissue was assayed from all rats in the Social Only groups.

Prefrontal Cortex	Male	Female
Prelimbic	1,179.89 (65.5)	1,044.13 (52.5)
Infralimbic	1,180.27 (42.9)	1,037.38 (48.4) *
Ventral Orbital	1,033.75 (48.0)	931.69 (32.8)
Amygdala
Central Nucleus	930.07 (64.1)	824.25 (42.6)
Basolateral Nucleus	435.00 (37.0)	386.68 (24.7)
Medial Ventral	422.17 (45.9)	233.99 (44.0) *

Values are means and SEMs of cells/mm².

^*p < 0.05, difference between males and females.

3.5. Data availability statement.

All data can be found at https://doi.org/10.7910/DVN/NXH9RU.

4. Discussion.

In the present study, we have described a novel model of conditioned social fear. To the best of our knowledge, this is the first report of a Pavlovian model of social fear, and the first to test conditioned social fear using a rat model of early social adversity, PSI. While there is evidence that PTSD is more likely to occur after traumatic events with social elements (including assault, rape, and combat) than traumatic events without social elements (Yehuda & LeDoux, 2007), conditioned fear using nonsocial CSs (such as a tone) are standard. We hypothesized that conditioned social fear would alter subsequent behavior when the experimental rat was re-exposed to the CS rat, and that this would depend on the social history of the experimental rat. (running in circles, running away) as well as social behaviors (aggressive grooming and other social behavior) in an open field, and observed that alterations in these behaviors by paired footshock and novel same-sex conspecific CS rat differed depending on whether rats had experienced PSI or SR.

PSI resulted in increased social interaction and aggressive grooming in both male and female rats, consistent with previous work from our laboratory (Wall et al., 2012; Grotewold et al. 2014; Fontenot et al. 2017). Our finding of robust increases in social interaction in both male and female rats after PSI contrasts with results of Ferdman et al. (2007) in which males, but not females, engaged in increased social interaction after isolation rearing. Methodological differences such as the testing apparatus, the length of the isolation period, and the age of the animals when tested may account for this difference. There were also large increases in aggressive grooming in both male and female isolation-reared rats. Aggressive grooming has been associated with increased corticosterone (Hurst et al., 1999), suggesting that isolates were more stressed by the novel conspecific than SR rats. This may also explain the increase in running away observed in PSI rats. Interestingly, more social behavior (excluding aggressive grooming) was observed in females that had experienced PSI, suggesting a greater drive toward prosocial behavior.

Experiencing a single trial of paired footshock with a CS rat increased circling behavior in PSI rats, which were, even in the absence of a social stressor, more likely to turn in circles when exposed to the CS rat in an open field. Circling behavior, which we observed in both male and female isolates but infrequently in SR rats, may be an anxiety-like behavior (Pezzato et al., 2015; Waters et al., 2012). Pezzato et al., (2015) describe an increase in circling, which they term “circling stereotypy”, in rats with lesions of the median raphe nucleus. Because this behavior was attenuated with lithium treatment, the authors hypothesized that circling is a manifestation of behavioral disinhibition associated with mania (Pezzato et al., 2015). As additional support for the hypothesis that circling is associated with mania or other types of anxiety, increased circling corresponded with increased anxious-like behavior in both the elevated plus maze and the light-dark box (Pezzato, 2015). Waters et al. (2012) have shown that antagonism of the 5-HT7 receptor with the anxiolytic drug SB-269970 reduced circling behavior produced by both amphetamine and PCP in (unlesioned) rats and suggested that this reflects reduced anxiety associated with psychosis. In a study comparing effects of social isolation on circling in mice and rats, spontaneous nocturnal circling increased in isolated mice, but not rats (Jerussi & Hyde, 1985). Circling may not be a simple escape behavior, because while PSI rats ran away from the CS rat more frequently than SR rats did, this was not differentially impacted by Conditioning. Thus, circling behavior may represent a different emotional state and rely on a different neural pathway than running away. It is interesting to note that PSI rats exhibited both increased social interaction and increased running away, suggesting pressure to both approach and avoid the CS rat.

We also demonstrate that a single pairing of a footshock with a novel conspecific alters subsequent social behavior, particularly in males that had been socially reared. When SR males were re-exposed to the same stimulus rats that had been paired with a footshock the following day in a novel context, aggressive behavior by the experimental rat against the stimulus rat increased while non-aggressive social interaction decreased. We observed dramatically increased aggressive grooming behavior in our PSI males and females, consistent with our previous work (Wall et al., 2012; Grotewold et al., 2014; Goodell et al., 2017; Fontenot et al., 2017) and that of others (Ferdman et al., 2007; Toth et al., 2008; Zhao et al., 2009). However, although paired footshock and social cue reduced aggression slightly in PSI males, this did not reach statistical significance. Rather, paired footshock and social cue increased aggression in SR males such that they were indistinguishable from PSI males. Paired footshock and social cue increased aggression in SR females only modestly and had no effect on non-aggressive social behavior in females. It is notable that PSI treatment increased both social interaction and running away, suggesting increased approach and avoidance of a social encounter.

Fear-related disorders are frequently modeled using Pavlovian fear conditioning; a rich literature exists describing the behavioral consequences of auditory fear conditioning, and the neural pathways that mediate it are well characterized (LeDoux, 2012). Freezing behavior is the most widely used measure of conditioned fear; however, it is only one component of a varied repertoire of defensive behaviors (Blanchard et al., 1975). In the present study, freezing behavior was not observed (data not shown) in the open field during exposure to the CS rat. Freezing has been described as a “flight-or-fight response put on hold” (Kozlowska et al., 2015) and may not be the most adaptive strategy when danger is imminent. When an individual is confronted with physical contact with a threatening stimulus, a burst of activity that may include escape (this may be especially true in females; see discussion below) or defensive aggression may be a more successful strategy (Fanselow, 1994). Other behavioral responses may be selected depending on the environment. For example, freezing may be the dominant response only in situations when no opportunity for flight or fighting exists. In studies of cue conditioned fear memory and extinction, testing for the response to an auditory cue is usually done in a separate context than that used in the fear conditioning session (Baran et al., 2009; Gruene et al., 2016; Sierra-Mercado et al., 2011). These contexts are inescapable and typically small, thus freezing is a highly adaptive response, but this is not always the case in the natural world. If given the choice to escape, flight or avoidance may be more adaptive; however, if given a concrete CS (such as a conspecific) and no opportunity to escape, fighting may be a more adaptive strategy.

We observed differential effects of conditioned social fear on social interaction in males and females. Although men may be more likely to experience traumatic events, women are more likely than men to develop several stress-related disorders including PTSD (Breslau et al., 1999). While it remains unclear if this sex difference is strictly biological or if it reflects sociocultural features, animal models can help tease these dimensions apart. Several studies of sex differences using traditional conditioned fear paradigms currently exist. For example, Keiser et al. (2017) reported both stronger context fear conditioning and greater generalization to a similar context in females, consistent with the enhanced generalization of a fearful stimulus in PTSD patients (Briscione et al., 2014). In contrast, Baran et al. (2009) reported no sex differences in the acquisition of auditory fear; however, in that study females extinguished more slowly than males. Gruene et al. (2016) did not observe sex difference in auditory cued conditioned fear, although if extinction occurred during proestrus females spent less time freezing during extinction and extinction memory (an additional day of testing). Interestingly, that group (Gruene et al., 2015) also observed that females exhibited more active responses (darting) during fear conditioning than males, and those that were more active during conditioning spent less time freezing during extinction. It is possible that our findings reflect greater social buffering in females, because non-aggressive social interaction was not reduced by Paired treatment in females. However, social buffering is most robust when the CS rat is familiar (Kiyokawa, 2018), unlike in our experiment. Alternatively, female rats may be more influenced by prosocial behavior toward the CS rat; this sex difference has been shown to be true in rodent models (Bartal et al., 2011).

Toth et al. (2012, 2013) have developed a method of social fear in which rats received a footshock when they initiated social interaction with a novel conspecific that had been placed in the apparatus within a mesh cage. During testing and extinction on subsequent days the stimulus animal was presented in the mesh cage and proximity to the cage was assessed (Toth et al., 2012, 2013). The present study differs from that procedure in that the CS and US were explicitly paired on Day 1. Moreover, in the present study social interaction was measured in freely interacting animals rather than assessing proximity to the CS rat.

The neural circuitry thought to mediate learned fear includes the ventral aspect of the mPFC which comprises the PL and IL subregions. The PL and IL have opposing and dissociable functions in processes important for conditioned fear and its persistence. The PL is especially important for the acquisition and expression of conditioned fear, while the IL is critical for fear extinction and extinction memory (Sierra-Mercado et al., 2011). The amygdaloid complex, including the BLA and CeA, play an essential role in fear conditioning and extinction (Maren & Fanselow, 1996), in part due to their reciprocal connections with the mPFC (Giustino & Maren, 2015). There is evidence that the medial amygdala (MeA) is involved in fear conditioning as well (Tsuda et al., 2015), and because of its important role in social behavior (Vochteloo & Koolhaas, 1987) the MeA may be particularly important for social fear. The MeA receives direct projections from olfactory regions and is responsive to chemosignals from conspecifics (Samuelsen & Meredith, 2009). Moreover, and consistent with the present findings in males, the posterior ventral MeA was preferentially activated in that study by exposure to threatening chemosignals, cat odor (Samuelsen & Meredith, 2009). Less pS6 was expressed in both the BLA and MeV of male rats that were first exposed to the CS rat on Day 2 (Footshock only on Day 1), and this may reflect social memory when rats were exposed to the CS rat on both Day 1 and Day 2. Moreover, we observed that more pS6 expression was observed in the BLA of Paired males compared to the other conditioning groups, suggesting that activation of the mTOR pathway in this region is particularly important for conditioned social fear, at least in males. It is important to note that in this experiment pS6 was assessed after exposure to the CS rat on Day 2, thus the results may reflect retrieval of the fear memory, consistent with the results of Levin et al. (2017).

Less is known about a role for the OFC in fear, anxiety, and social interaction. Lesions of the OFC increased aggressive behavior in male rats (females were not tested), although overall time spent engaged in social interaction was not affected (Rudebeck, 2007). In macaques, excitotoxic lesions of the OFC increased defensive responding to both neutral and threatening stimuli, and the authors suggest the OFC could be important in anxiety disorders (Pujara, 2019). Interestingly, the OFC/amygdala neural circuit appears to regulate action-outcome decisions (Fiuzat, 2017), which are likely to be critical for emotional regulation during social interactions. Additional support for “top-down” regulation of the amygdala by the OFC during emotional regulation is provided by a human imaging study of patients with major depressive disorder, (Zheng et al., 2018) in which disrupted white matter between OFC and amygdala was normalized by antidepressant treatment. This OFC/amygdala connection is critical for healthy social interactions, as functional connectivity of this pathway is reduced in social anxiety disorder both during resting state (Hahn, 2011) and during an emotion discrimination task (Sladky, 2015). Although the relationships between amygdala and frontal cortex pS6 expression did not reach significance (data not shown), it is noteworthy that the alteration of pS6 by our manipulations in these brain regions were in opposite directions. Imaging studies have shown that the brain regions assessed in the present study are functionally altered in PTSD patients (Etkin & Wager, 2007), and altered plasticity within this network may underpin the impaired extinction that characterizes PTSD (Radley et al., 2004). Notably, the mPFC/amygdala pathway is also critical for sociability as optogenetic activation of this pathway reduced social interaction (Felix-Ortiz et al., 2016).

5. Conclusion.

In conclusion, the present study demonstrates that social fear conditioning alters social behavior and pS6 expression in brain regions involved in emotion regulation. A novel avoidant social behavior, running in circles, was enhanced in rats that had experienced social isolation during adolescence, and was even greater in these rats after social fear conditioning. Sex differences in the effects of these manipulations on pS6 expression were observed, suggesting that plasticity related intracellular signaling pathways may play a different role in the behavioral effects observed during fear memory retrieval in both males and females.