Recovery of stress-impaired social behavior by an antagonist of the CRF binding protein, CRF_6-33, in the bed nucleus of the stria terminalis of male rats

1. Introduction

Social stress exposure can lead to a wide range of neurologic behavioral disturbances relevant to psychiatric disorders [1]. Social defeat of an intruder confronted by a resident is a potent social stressor in rodents. This model promotes several behavioral, endocrine and physiological changes in the intruder animal, which may result in anxiety- or depressive-like behaviors [2,3]. Social stress produces neuroendocrine responses characterized by increased activity of the hypothalamus-pituitary-adrenal axis which is mediated by a cascade of hormones, such as CRF, adrenocorticotropic hormone and glucocorticoids [4]. Although these mechanisms are protective, prolonged exposure to stress may engender pathophysiological consequences [5].

The effects of CRF and the structurally related Urocortins are mediated by two CRF receptor subtypes: CRF type 1 (CRFR1) and type 2 (CRFR2). Additionally, CRF and Urocortins interact with a CRF binding protein (CRFBP), which binds CRF with high affinity (K_i values for CRF and Urocortin 1 are 217 and 77.2 pM, respectively) compared to CRF receptors [6,7]. CRF is bound with a 40-fold higher affinity for CRFR1 than CRFR2, while CRFR2 and the CRFBP are bound with high affinity to Urocortin in rats [8,9]. Although the CRFBP function in the brain is largely unknown, distinct mechanisms have been proposed. One possibility is that CRFBP is responsible for capturing the CRF available extracellularly, preventing CRF receptor activation [10]. Alternatively, there is evidence that the function of the CRFBP may be more complex and may extend beyond mere sequestration of CRF or CRF-like ligands [11]. Based on ex vivo experiments in mice, Ungless and colleagues proposed that CRF enhances N-methyl-D-aspartate (NMDA) receptor activity through activation of CRFR2 in neurons of the ventral tegmental area (VTA), an effect that requires the action of the CRFBP [11]. Similarly, our group recently demonstrated that the co-infusion of astressin-2B, a CRFR2 antagonist, with a sub-effective dose of CRF_6–33 in the VTA attenuates alcohol binge drinking in mice [12]. These studies not only indicate that the CRFBP may actively participate in CRF signaling, but also suggest that the CRFBP may act in combination with CRFR2. More evidence, however, is needed to clarify the role of this binding protein in the brain, specifically in discrete regions involved in the stress regulation.

The BNST connects the amygdaloid complex with the hypothalamus and has been implicated in the control of neuroendocrine functions and behavioral responses to stress [13,14]. Both, inhibitory and facilitatory roles of the BNST in these responses have been reported, depending on the type of the aversive stimulus and subregion of the BNST [15]. For example, the anterior portion has been described as an important site of action for extrahypothalamic CRF in the modulation of anxiety-like behaviors [16]. In vivo experiments showed that intracerebroventricular (i.c.v.) infusion of CRF increased startle responses and this effect can be blocked by CRF antagonism or lesions of the anterior BNST in rodents [17,18]. Moreover, intra-BNST CRF increased anxiety-like behaviors in the elevated plus-maze [19], produced place conditioning aversion to a CRF-paired environment [19], and reinstated cocaine seeking in rats during drug abstinence [20]. The BNST contains CRF-producing neurons which seem to be sensitive to stress [21,22]. In fact, exposures to corticosterone and the pharmacological stressor yohimbine upregulate CRF mRNA expression in the BNST [23,24].

Exposure to intermittent episodes of social defeat seems to reduce responses to hedonic stimuli; defeated animals show a lack of sexual interest [25], decreased preference for sweet solutions [26,27], reduction of locomotor and exploratory behaviors in novel environments [27], and reduced preference for social interaction [28]. The current experiments were designed to test the hypothesis that four brief encounters with an aggressor rat can elicit dysregulated behaviors in adult Wistar rats, including anxiety- or depressive-like symptoms. Further, we investigated the role of BNST CRFBP in the modulation of stress-induced anxiety responses.

2. Methods

2.1. Animals

Adult male Wistar rats were obtained at approximately postnatal day (PND) 50 and housed at the Animal Experimental Unit of Hospital de Clínicas de Porto Alegre (Porto Alegre, RS, Brazil) for 2 weeks before starting the experiments. At PND 70, rats were singly housed in custom-built acrylic cages (15 × 25 × 20 cm). Separate male Wistar rats, weighing 465 ± 10 g, with a reliable history of aggressive behavior in confrontations with intruders, termed stimulus ‘resident’ rats, were housed in pairs with sterile female Wistar rats in large custom-built acrylic cages (46 × 71 × 46 cm). Resident rats and intruder rats were kept in separate rooms with controlled environmental conditions: 21 ± 1° C temperature, 40–60 % humidity and 12 h/12 h light-dark cycle (lights on at 7:00 AM). The cages were lined with sawdust bedding and rats had free access to food and water. A total of N = 50 rats was used in this study (n = 7 residents; n = 7 females paired with residents; n = 3 ovariectomized females for social interaction test; n = 33 intruders). This study was carried out in accordance with the Brazilian Federal Law No11.794/2008 for the scientific use of animals. The protocol was approved by the Ethics Committee on Animal Use of Animal Experimentation Unit from Hospital de Clínicas de Porto Alegre.

2.2. Drug

The selective CRFBP antagonist CRF_6–33 (Tocris Bioscience; Ellisville, MD, USA) was dissolved in sterile saline solution (NaCl 0.9%) at concentrations of 0.25 and 0.5 µg in 0.2 µL.

2.3. Procedures

The stress protocol was conducted following procedures established by Miczek and colleagues [26,29]. Rats were subjected to four brief intermittent episodes of social defeat stress over the course of 10 days, or served as contemporary home cage controls. After concluding the stress protocol, rats were tested for sweet solution preference, had bilateral cannulae implanted into the BNST, received local injections of CRF_6–33 and were examined for social approach to an unfamiliar ovariectomized female (Figure 1.a). All procedures occurred during the first four hours of the dark phase. Rats were monitored daily for health conditions and were weighed before the first and the last social defeat session.

Figure 1 –

a. Experimental procedure. Stress: rats exposed to intermittent episodes of social defeat; Non-stress: control animals; n = 13–14 rats per group; b. Schematic representation of the social approach apparatus in a three-dimensional view; c. Length and height of the chambers and walls. The central chamber has openings that allow the rat to move freely throughout the whole apparatus. After the habituation period (5 min), an unfamiliar ovariectomized female was placed inside a small acrylic cage in one of the side chambers and an identical empty cage was placed in the opposite side chamber. The walls were removed and the rat could move freely throughout the three chambers. The positions of the female and the empty cage were alternated between side chambers and upper and lower corners between sessions.

2.4. Intermittent episodes of social defeat stress

Intruders were exposed to a different resident for four sessions with 72 h interval between them [30]. The protocol consists of three phases: in the first phase, ‘pre-defeat’, the female was removed and a protective cage containing the intruder was placed into the large resident home cage for 10 min. The intruder was threatened and investigated by the resident through the perforated acrylic walls of the protective cage. In the second phase, ‘defeat phase’, the intruder was removed from the protective cage and directly confronted the aggressor resident. The fight was terminated after the intruder displayed supine posture for 5 s, or 5 min after the resident’s first biting attack, or after 5 bites, whichever occurred first. In the third phase, ‘subordination’, the intruder was placed back into the protective cage within the resident’s home cage for 10 min [30,31]. Control rats were handled and weighed, without being submitted to the stress protocol.

2.5. Preference for sweet solution

At PND 85 and 95, rats were removed from their home cages and assessed for sweet solution preference (0.8% saccharin and sodium cyclamate; Zero·Cal, Hypermarcas; São Paulo, SP, Brazil) during one hour in a test cage (24 × 38 × 15 cm) in an appropriated room for behavioral testing. First, at PND 71, rats were exposed overnight to one bottle containing the sweet solution and another bottle containing water. During the following four consecutive days, PND 72–75, rats were given a daily 2-bottle choice (sweet solution and water) for one hour, one week before the stress protocol (baseline condition). To counteract side preference, the position of the bottles was switched between trials. Animals were not food or water deprived, before or after the procedure. Fluid consumption was measured by weighing the bottles, and preference was expressed as percentage of total intake, calculated by dividing the total amount of sweet solution consumed by the total fluid intake (water plus sweet solution) [32]. An empty “drip” cage served as control for evaporation and spillage due to handling of bottles.

2.6. Stereotaxic Surgery

Rats were anesthetized with isoflurane prior to and during the surgery. Pre-surgical analgesia was induced with tramadol (20 mg/kg, i.p.) and local anesthesia with bupivacaine (8 mg/kg, intradermal). Rats were placed on a stereotaxic frame (Kopf Instruments; Tujunga, CA, USA) and guide cannulae were bilaterally implanted into the BNST (from bregma, AP: 1.0; ML: +/− 2.5; DV: −8.0 with 10° angle) [33,34]. Guide cannulae were fixed with stainless steel screws and dental cement, and were sealed with custom made dummy cannulae. Immediately after surgery, rats received dipyrone (500 mg/kg, i.p.), and subsequent pain control was provided with tramadol (20 mg/kg, i.p.), every 12 h for 2 days. Surgeries were conducted at least 5 days before the first infusion.

2.7. Intra-BNST CRF_6–33 infusions

Starting ten days after the last social defeat encounter (PND 96), rats received intra-BNST CRF_6–33 three times, according to a design that counterbalanced saline and drug doses (once a day, with 48 h intervals between the infusions). Microinjections were performed with an infusion pump (Insight EFF 311; Ribeirão Preto, SP, Brazil) and were delivered bilaterally and simultaneously at a constant volume of 0.2 μL/side over a period of 2 min. The injector needles extended 1 mm beyond the tip of the guide cannulae and were left in place for 1 additional minute after the end of the infusion to avoid reflux and allow for diffusion. During the microinjection procedure, rats were allowed to move freely in a small cage (30 × 20 × 13 cm). Infusions started five minutes before the social approach test. Doses were scheduled as follows: saline solution, a lower dose of CRF_6–33 (0.25 μg/0.2 μL), or a higher dose of CRF_6–33 (0.5 μg/0.2 μL). These doses have been previously used [12].

2.8. Social Approach Test

The social approach apparatus consists of an acrylic box (40 × 50 × 120 cm) with removable walls separating the box into three chambers (Figure 1.b-c). The walls have openings that allow the animal to move freely throughout the whole apparatus. Rats were initially confined to the middle chamber for 5 min. After this habituation period, an unfamiliar ovariectomized female was placed inside a small acrylic cage (15 × 25 × 20 cm) in one of the side chambers. An identical empty cage was placed in the opposite side chamber. Then, the walls were removed, allowing the rat to explore all three chambers over a period of 10 min. The small cages did not allow physical contact between the rats, preventing the female from any aggressive interaction. The location of the female and the empty cage were alternated between the upper and lower corners of the side chamber in consecutive trials. All sessions were video recorded and total distance travelled was measured using a behavioral tracking system (ANY-Maze; Wood Dale, IL, USA); additional behaviors were coded using event-logging software [35]. During the habituation period distance traveled was measured for 5 min, and frequency and duration of entries into the side chambers containing the unfamiliar female (interaction zone) or the empty cage (object zone) were measured for 10 min [36]. The apparatus was cleaned with 70% alcohol between trials. Animals were tested three times, with a 48 h-interval between sessions.

2.9. Histology

One day after the last social approach test (PND 101), rats were deeply anesthetized with an overdose of isoflurane 5% for more than 3 min and were perfused with 0.9% saline followed by 4% buffered paraformaldehyde solution prior to brain removal. The fixed brains were sliced in 80 µm coronal sections using a cryostat (Microm HM 505 E; Waltham, MA, USA). Slices were mounted on glass slides and stained with hematoxylin-eosin, and injector placements were verified using light microscopy, according to the rat brain atlas [34]. Diagrammatic representations of bilateral BNST injection sites are shown in Figure 2. Rats with injector tracks that did not reach the BNST were excluded from the analysis (~18%).

Figure 2 –

a. Correct cannulae placements into the bed nucleus of the stria terminalis; b. Diagram and representative photomicrograph after hematoxylin-eosin staining. The number of points in the figures is less than the total number of animals because of overlapping injection sites.

2.10. Statistical analysis

To assess the effects of social stress on preference for sweet solution, a two-way repeated measures analysis of variance (ANOVA) was conducted (Statistica 6.0, Dell Software; Round Rock, TX, USA) using stress vs. non-stress as the condition factor, and baseline and two additional observation times as the sessions factor. To assess social approach, frequency of entries and time spent on either interaction zone (IZE and IZT) or object zone (OZE and OZT) were analyzed. A two-way repeated measures ANOVA was performed, followed by a priori hypothesis driven one-way ANOVA to evaluate the effects of CRF_6–33 treatments (0.25 μg/0.2 μL, and 0.5 μg/0.2 μL) using each condition (stress or non-stress) as a single factor. When indicated by a significant main effect, post hoc comparisons were performed using Bonferroni correction for multiple comparisons. T-tests were conducted to compare stressed vs. non-stressed groups treated with saline. The statistical significance was set at p < 0.05.

3. Results

3.1. Effects of intermittent social defeat stress on sweet solution preference and total distance travelled

Intermittent exposure to social defeat stress did not produce significant effects on preference for sweet solution as revealed by two-way repeated measures ANOVA (F (1, 20) = 0.33, p > 0.05). Moreover, rats did not show significant differences in preference across sessions (F (1, 20) = 0.16, p > 0.05) (Table 1). Intermittent exposure to social stress did not affect total distance travelled (F (1, 23) = 0.97, p > 0.05), as well as the treatments (F (1, 23) = 0.09, p > 0.05) (Figure 3e).

Table 1 –

Data and summary of repeated measures ANOVA for preference (mean ± SEM) for sweet solution during baseline and two additional measurements. The significance threshold was set at p < 0.05 for all comparisons.

	Sweet Solution Preference (%)
	PND 72–75 (Baseline)	PND 85	PND 95
Non-stressed	93.0 ± 3.0	96.0 ± 3.0	96.0 ± 7.0
Stressed	93.0 ± 2.0	94.0 ± 3.0	90.0 ± 6.0

Figure 3 -.

Effects of corticotropin releasing factor (CRF) fragment 6–33, a CRF binding protein antagonist, administered into the bed nucleus of the stria terminalis of rats exposed to intermittent social defeat stress or non-stressed controls. a. Number of entries into the interaction zone; b. Time spent in the interaction zone; c. Number of entries into the object zone; d. Time spent in object zone. The interaction zone contained an unfamiliar female placed inside a small acrylic cage and the object zone contained an empty small acrylic cage. Data are mean ± SEM; * stressed vs. non-stressed rats treated with saline; # saline infusion vs. CRF_6–33 0.25 µg/0.2 µl in stressed rats; p < 0.05, n = 13–14 rats per group. e. Total distance travelled measured in meters. The animals were tested during the habituation period (5 min) in the central chamber of the three-chamber apparatus. The central chamber measures 0.5 × 0.5 m. Data are mean ± SEM; n = 12 rats per group

3.2. Effects of intra-BNST CRF_6–33 on social approach

The two-way repeated measures ANOVA did not reveal significant differences between conditions (IZE: F (1, 75) = 1.28; IZT: F (1, 75) = 0.98; OZE: F (1, 75) = 0.30; OZT: F (1, 75) = 2.12; p > 0.05 in all cases), treatments (IZE: F (2, 75) = 1.52; IZT: F (2, 75) = 0.64; OZE: F (2, 75) = 0.13; OZT: F (2, 75) = 0.78; p > 0.05 in all cases), and interactions between conditions and treatments (IZE: F (2, 75) = 2.10; IZT: F (2, 75) = 1.82; OZE: F (2, 75) = 0.09; OZT: F (2, 75) = 0.14; p > 0.05 in all cases).

Pairwise comparisons between conditions (stress vs. non-stress) after saline infusions showed that socially stressed animals presented lower levels of social interaction compared to non-stressed rats (IZE: t = 2.39 and IZT: t = 2.46; p < 0.05 in both cases) (Figure 3a-b). As the two-way repeated measures ANOVA failed to demonstrate significant differences between conditions we performed a one-way ANOVA based on a priori hypotheses. Stressed rats treated with intra-BNST CRF_6–33 entered the interaction zone significantly more than saline-treated animals (IZE: F (2, 39) = 4.14, p < 0.05). Post hoc analysis indicated that this difference is related to an increase in social approach promoted by the low dose of CRF_6–33 (Figure 3b). Drug treatments did not alter the time spent in the interaction zone (IZT: F (2, 39) = 1.91, p > 0.05) (Figure 3b) or the exploratory behaviors of non-stressed animals (IZE: F (2, 36) = 0.07; IZT: F (2, 36) = 0.20; OZE: F (2, 36) = 0.20; OZT: F (2, 36) = 0.17; p > 0.05 in all cases) (Figure 3).

General exploratory behaviors were not altered by exposure to intermittent social stress in saline-treated animals (OZE: t = 0.48; OZT: t = 0.99; p > 0.05 in both cases) (Figure 3c-d). Moreover, intra-BNST CRF_6–33 did not affect the exploration of a new object (OZE: F (2, 39) = 0.01; OZT: F (2, 39) = 0.72; p > 0.05 in all cases) (Figure 3c-d), as well as activity in the central chamber (data not shown), indicating that the drug has a selectively action on social approach.

4. Discussion

The current experiments provide the first evidence for selective improvements on social behaviors after exposure to brief intermittent episodes of social defeat stress following microinjection of the CRFBP antagonist CRF_6–33 into the BNST. The drug did not affect general activity or exploration. The lack of changes on sweet solution preference across sessions and between non-stressed controls and stressed rats suggests that the deficits in social approach exhibited by stressed animals are not related to the development of depressive-like symptoms.

To investigate the stress effect on behaviors with hedonic motivation, we tested the preference of the animals for a palatable sweet solution prepared with 0.8% saccharin and sodium cyclamate. This non-stressful, non-invasive procedure, allows multiple tests without compromising the animal’s behavior [32]. Classically, the reduction in preference for sweet solutions and palatable food has been interpreted as an index of anhedonia, the lack or disruption of the ability to experience pleasure [37]. Anhedonia is considered one of the core symptoms of major affective disorders according to the DSM-5 [38]. In this study, we did not find differences in the preference for sweet solution after acute experience of social defeat or 10 days after the last session of social stress. It seems that chronic exposure to social stress is required in order to engender hedonic and motivational deficits in rats [37,39], although under specific conditions acute social defeat stress may produce suppression of preference for sweet solutions [40]. Thus, the lack of changes in preference for sweet solution found in the present study is in accordance with evidence contrasting the effects of chronic and intermittent exposures to stress [26].

Stressed animals were less likely to approach a conspecific female in the three-chamber social approach and interaction test 10–13 days after the last confrontation. Deficits in social interaction have been reported after intermittent exposure to social defeat, with impaired interactions persisting up to seven weeks after the last social defeat experience [41]. Reduced social interaction is associated with several psychiatric conditions, such as depression, social anxiety and autism spectrum disorders, which are characterized by dysfunctional reciprocal social interactions and altered coping with social environments [42,43]. It is tempting to suggest that the neuroadaptations promoted by social defeat stress may lead to difficulties in coping with social challenges in defeated animals. In a non-threatening environment, impaired social interaction may be interpreted as a maladaptive effect as well as a sign of anxiety [43,44].

In our study, the antagonism of CRFBP located in the BNST selectively re-established social approach in stressed rats, without modifying general exploratory behaviors. We used the same doses of CRF_6–33 reported as effective in a previous study (0.25 and 0.5 µg) [12]. Our data showed that CRF_6–33 promoted a significant increase in the frequency of entries in the interaction zone in socially stressed animals to levels similar to non-stressed controls.

It is worth mentioning that CRF_6–33 has already been tested in behavioral cognitive and pre-clinical studies and has also shown performance-enhancing effects. Animals tested in the Morris water maze exhibited significant improvements on performance after i.c.v. infusion of CRF_6–33 [45,46]. Additionally, these animals presented an increase in duration of searching the target quadrant compared to vehicle-treated animals, indicating an enhancement of spatial learning [46]. However, in the same set of experiments, a similar effect was seen after infusion of CRF_1–41, a CRFR agonist, indicating that the improvements reported may be due an increase in the availability of CRF due to CRFBP antagonism.

There is evidence, however, supporting the idea that CRFBP is not only a sequestering protein, but may also possess additional functions [11,12]. Increasing data indicate that CRFR2 and CRFBP seem to act together, at least in the VTA, to promote the CRF effects [11]. More precisely, Slater and colleagues [47] showed that CRFBP forms a protein complex with the CRFR2α isoform. These authors suggest that the CRFBP could act as an escort protein regulating the access of CRFR2α located intracellularly to the cell surface. These findings strengthen the idea of a synergic mechanism of action between the CRFBP and CRFR2.

Interestingly enough, the CRFBP has been demonstrated to present spontaneous cleavage, originating two distinct fragments: a 27 kD and a 10 kD [50]. Haas-Koffler and colleagues [51] verified recently that the CFRBP (full length) can stably be expressed on the plasma membrane and only the 10 kD fragment is able to potentiate CRF signaling properties. Thus, taking into account that the 27kD fragment retains the active picomolar affinity binding site for CRF [52], it is possible that the distinct role of the CRF_6–33 may rely on the fact that blocking the binding site of the 27 kD fragment would suppress the CRFBP sequestrating action and induce an increase in free CRF levels. On the other hand, the 10 kD fragment could act potentiating the CRFR2α. Further studies, however, will help to clarify the significance of the CRFBP fragments in in vivo experiments.

The BNST is known as one of the most complex structures in the central nervous system [15], whose heterogeneous nature is due to its division into different subregions and cell types, known to modulate anxiety in opposing ways [15]. The BNST is densely connected with the hypothalamus, amygdala, midbrain, and lower brainstem regions, with different afferents [53,54] and distinct projections [55,56]. Moreover, BNST nuclei connect each other intensely, and the γ-aminobutyric acid (GABA) seems to be the primary neurotransmitter, although a small number of glutamatergic neurons are also located in its ventral portion. BNST neurons also express a variety of neuropeptides, including CRF, enkephalin, neuropeptide Y, neurotensin and somatostatin [57]. In fact, CRF is produced by anterolateral BNST neurons and seems to act in this same region, potentiating anxiety-like behavior and stress response [18,21,22]. In our study, CRF_6–33 infusions occurred mostly in the anterior portion of the BNST. Interestingly, the inhibition of CRF neurons in the anterior BNST promoted anxiolytic-like effects in the elevated plus-maze and open field tests [58]. Contrarily to our results, however, Klampfl and colleagues [59] reported no effect of CRF_6–33 into the anterior BNST, and anxiogenic-like effects accompanied by increased locomotion when CRF_6–33 was injected into the posterior BNST. Nonetheless, important differences between these studies are that Klampfl and colleagues [59] used female rats from a different strain and during their lactating period.

A limitation of the present study is that we did not measure the content of CRF in the BNST. However, results from our group indicate that repeated social defeat stress promotes increase in CRF levels in the BNST of mice [60]. Direct injection of CRF in BNST subregions will also help to identify specificities of the CRF mechanisms, and future studies using CRFR1 and CRFR2 antagonists will bring light to the role of the CRFBP in the central nervous system.

5. Conclusions

The present study brings new findings about the involvement of the CRF system in anxiety-related states, implicating the BNST CRFBP in the modulation of social behaviors in male rats previously exposed to a social stressor. Further studies will help to elucidate how the CRFBP promotes its action in discrete brain regions, as well as the role of stress in this circuitry, and whether the CRFBP could be considered a target candidate for the development of novel therapeutic treatments of stress-related disorders.

Highlights.

• Intermittent social defeat did not promote anhedonic-like symptoms

• Intra-BNST CRF_6–33 infusions restored social approach impaired by social defeat

• CRF mechanisms in the BNST modulate anxiety responses in stressed rats

Abbreviations

BNST

bed nucleus of the stria terminalis

CRF

corticotropin releasing factor

CRFBP

CRF Binding Protein

CRFR

CRF receptor

EPSC

Excitatory postsynaptic currents

i.c.v.

intracerebroventricular

IZE

Interaction Zone - Entries

IZT

Interaction Zone – Time

NMDA

N-methyl-D-aspartate

OZE

Object Zone - Entries

OZT

Object Zone - Time

PND

postnatal days

VTA

ventral tegmental area