Loss of CREBRF Reduces Anxiety-like Behaviors and Circulating Glucocorticoids in Male and Female Mice

Abstract

Glucocorticoid signaling controls many key biological functions ranging from stress responses to affective states. The putative transcriptional coregulator CREB3 regulatory factor (CREBRF) reduces glucocorticoid receptor levels in vitro, suggesting that CREBRF may impact behavioral and physiological outputs. In the present study, we examined adult male and female mice with global loss of CREBRF (CrebrfKO) for anxiety-like behaviors and circulating glucocorticoids in response to various acute stress conditions. Results demonstrate that both male and female CrebrfKO mice have preserved locomotor activity but reduced anxiety-like behaviors during the light–dark box and elevated plus maze. These behavioral phenotypes were associated with lower plasma corticosterone after restraint stress. Further studies using unhandled female mice also demonstrated a loss of the diurnal circulating corticosterone rhythm in CrebrfKO mice. These results suggest that CREBRF impacts anxiety-like behavior and circulating glucocorticoids in response to acute stressors and serves as a basis for future mechanistic studies to define the impact of CREBRF in glucocorticoid-associated behavioral and physiological responses.

Glucocorticoids mediate numerous important behavioral and physiological functions over the lifespan of an organism. In response to a physical or emotional stressor, the hypothalamic–pituitary–adrenal (HPA) axis is initiated, causing the paraventricular nucleus of the hypothalamus (PVN) to release corticotropin-releasing hormone (CRH). This stimulates the anterior pituitary to secrete adrenocorticotropin hormone (ACTH), which triggers the adrenal glands to increase release of glucocorticoids such as corticosterone in rodents (1, 2). Elevated serum glucocorticoids then provide negative feedback by activating the glucocorticoid receptor (GR) at the levels of the PVN and pituitary. Additionally, because mice are nocturnal, the diurnal rhythm of corticosterone release is characterized by an initial rise into the dark period and a gradual decrease during the light period. This pattern is instrumental in establishing circadian regulation of biological responses of various tissues (3). Thus, circulating glucocorticoids normally fluctuate by increasing during a stressor as well as rising and falling in a circadian manner.

Disruptions in glucocorticoid signaling adversely impact health and disease (reviewed in (4, 5)). Increasing glucocorticoid levels through administration of corticosterone (6) or the synthetic GR agonist dexamethasone (Dex; (7)) heightens anxiety-like behaviors in rodents. In select regions of the brain, loss of the GR decreases (8, 9) whereas overexpression increases (10) anxiety-like behaviors. These data illustrate that pharmacologically or genetically modulating glucocorticoid signaling impacts anxiety-related behaviors.

Previously, analysis of the novel putative transcriptional coregulator CREB3 regulatory factor (CREBRF), originally identified as Luman Recruitment Factor (11), in vitro may regulate stability and subcellular localization of the GR (12), and in vivo female global CREBRF knockout mice (CrebrfKO) have lower postpartum circulating corticosterone (12). These data implicate multiple levels of CREBRF control of glucocorticoid physiology, yet little is known about CREBRF even though it is highly expressed in human tissues (13). Therefore, to expand the understanding of CREBRF’s effects on glucocorticoid-mediated behavior and physiology, specifically in response to acute stressors, we evaluated anxiety-related behaviors and circulating glucocorticoids in both male and female CrebrfKO mice.

Material and Methods

Animals

Mice with global targeted deletion (knockout, KO) of the endogenous murine Crebrf gene (CrebrfKO) were generated as previously described (12). The mice used in the current study were congenic C57BL/6J and obtained from heterozygous breeding to generate litter-matched CrebrfKO and control mice. Notably, CrebrfKO female mice display maternal behavior deficits (12), but we did not observe any robust differences in generating litters or in the number of pups using heterozygous breeding. Mice were genotyped using the following primers in a single reaction: 5′-CCTTCTACATCTCCCATTCCAG, 5′-AAGAGGAGGATAAGCTGGTGTC, and 5′-GTCTTTGAGCACCAGAGGACAT, resulting in a 409 bp band for the control allele and a 571 bp band for the knockout allele. Mice were kept on a 12 hour light–dark cycle, group housed, and weaned onto a 20% kcal protein, 10% kcal fat, 70% kcal carbohydrate, 3.82 kcal/g energy density purified diet (Research Diets Inc, New Brunswick, NJ, USA, RD12450Hi) with food and water ad libitum. Female estrus stage was measured using vaginal smears (14, 15). All data were generated in a minimum of 2 separate cohorts to ensure reproducibility of results, the order of animal testing (ie, genotype) was randomized, and consideration was made to reduce animal number and suffering. To ensure each cohort had litter-matched sex and genotype mice, group size varied due to litter distribution, and the need to group house to avoid effects of isolation on behavior. Animal handling and experimentation was in accordance with Institutional Animal Care and Use Committee (IACUC) protocol approval at the University of Pittsburgh and conducted in conformity with PHS Policy for Care and Use of Laboratory Animals.

Anxiety-like behaviors.

The testing order was open field test, light–dark box, and elevated plus maze (14, 16). All mice were handled 2 days prior to each behavior test to acclimate them to mild disturbances. Apparatuses were cleaned between each animal and allowed to air dry. Behavior testing was randomized for sex and genotype, occurred between 11.00 hours and 15.00 hours in the light phase, and mice were given at least 4 rest days between tests. Behavior testing was performed by the same female experimenter, video recorded (Sony HD Handicam), and analyzed if quantified by 2 separate individuals blinded to sex and genotype.

Locomotor activity in the open field test

The open field test measured locomotor activity in a novel environment (17). Mice were placed in the center of a clear acrylic chamber with dimensions of 24 inches in height, 20 inches wide, and 20 inches in length (custom generated by shopPOPdisplays, NJ, USA) labeled with 4-inch squares for a total of 5 minutes. Total line crosses/ambulation, line crosses in the center 9 squares, and time in the center square were quantified with all 4 paws needed to count as a line crossing.

Light–dark box

The light–dark box consisted of 2 equal-sized chambers (40 cm long, 20 cm wide, and 35 cm high). One chamber was clear, the other black and enclosed with a lid, and the 2 chambers were connected with a small tunnel (Stoelting Co, Wood Dale, IL, USA). Mice were placed in the light chamber and total transitions (after all 4 paws entered the chamber) between the light and dark chambers, latency to enter the dark chamber, and time in each chamber were quantified for a total of 10 minutes (18).

Elevated plus maze

The test consisted of 2 open arms, 2 closed arms, and a central platform with dimensions of 5 cm lane width, 35 cm arm length, 15 cm wall height, and 50 cm leg height (Stoelting CO, Wood Dale, IL, USA). Mice were placed facing the open arm and allowed to explore for 5 minutes. The total closed and open arm entries, time in the closed and open arms, and open arm ratio (open arm entries/total arm entries) were quantified and analyzed. Mice that left the apparatus were placed back on the open arm and time adjusted so all mice were tested for a total of 5 minutes.

Corticosterone Measurement

CRH challenge

Control and CrebrfKO male and female mice were an administered intraperitoneal (IP) injection of 40 µg/kg CRH (Sigma Aldrich, St. Louis, MO, USA, C3042) followed by tail blood collection at 0, 15, 30, 60, and 120 minutes (19).

Dex suppression test

Mice received IP Dex (0.1 mg/kg body weight; Sigma-Aldrich D4902, less than 0.001% ethanol) at 11.00 hours. Mice were injected and placed back into their home cage. Blood was collected immediately following injection and 6 hours later via the tail for corticosterone analysis (19-21).

Restraint stress

Mice were placed into a restrainer (Braintree Scientific Inc, Braintree, MA, USA, TV-150 STD) and blood was collected from the tail at the initiation of restraint stress (time 0 – completed within 3 minutes of initial handling), after 30 minutes of restraint (time 30), and after 60 minutes of recovery in their home cage (time 90; 14, 15). Testing occurred in the light phase during the same 4-hour window as behavior testing between 11.00 hours and 15.00 hours.

Femoral catheterization

Femoral arterial catheters were chronically implanted as previously described (22, 23). Briefly, at 20 weeks of age, female mice were anaesthetized with inhaled 2% isoflurane, and microrenathane catheters were inserted in the left femoral artery, sutured in place, stabilized with superglue, tunneled subcutaneously to the upper back, attached to posterior cervical muscles, and then connected to a 360-degree dual-channel swivel. Catheters were maintained by continuously flushing saline containing 20 U/mL heparin (Baxter, Deerfield, IL, USA) using a syringe pump with multisyringe adaptor, monitored for patency daily, and kept unclogged by manual flushes using a 1-cc syringe with 26G needle when necessary. Mice were individually housed following surgery. Body weight, food intake, and blood pressure were monitored prior to surgery, during recovery, and at the conclusion of testing using a pressure transducer (Argon, Athens, TX, USA) connected to the arterial catheter. Recovery was indicated by return to preoperative body weight and normal activity (eg, nesting behavior). During blood sample collections, mice were not handled. Arterial blood was sampled over 4 days at 3 times daily: immediately after lights on at 8 am, 2 pm, and prior to lights off at 8 pm. For each blood sample, a total of 50 μL of whole blood was removed and 2 μL was used to measure blood glucose (Ascencia elite XL glucometer; Bayer, Mishawaka, IN, USA). The plasma was separated and red blood cells were resuspended with 20 μL of saline with 100 U/mL heparin and reinfused into the mouse to maintain hematocrit at baseline levels.

Corticosterone analysis

Blood samples from femoral catheterization or the tail during restraint stress, CRH challenge, and Dex suppression test were collected in heparin coated tubes and centrifuged at 14 000 relative centrifugal force (RCF) for 1 minute. Plasma was then transferred to a new tube and stored at –80°C. Corticosterone was determined using DetectX® Corticosterone Enzyme Immunoassay Kit (Arbor Assays, Ann Arbor, MI, USA, K014) and analyzed using Myassays.com.

Statistical analysis

Results are presented as mean ± standard error of the mean (SEM). Data were analyzed using 2-way analysis of variance (ANOVA) for effects of genotype (control vs CrebrfKO) and sex (male versus female) or with repeated-measures ANOVA for CRH challenge, Dex suppression test, and restraint stress using Prism 8 (GraphPad Software Inc, San Diego, CA, USA). Once an overall significance was established, post hoc comparisons were performed using the Tukey test. In all cases, P < .05 was considered statistically significant.

Results

As an initial assessment of the impact of global loss of CREBRF on behavior in both male and female mice, we measured mobility in an open field. There was no differences for total line crosses (Fig. 1A), time in the center (Fig. 1B), or center crosses (Fig. 1C). These data indicate that male and female CrebrfKO mice do not display differences in exploratory or locomotor ability compared with controls.

Figure 1.

Locomotor activity in the open field test in male and female control and CrebrfKO mice. Line crosses/ambulation (A), time in the center (B), and center crosses (C) are shown. Data are mean ± SEM. Groups: control male (n = 9), CrebrfKO male (n = 10), control female (n = 16), CrebrfKO female (n = 17).

To evaluate anxiety-like behaviors, we performed the light–dark box in both male and female control and CrebrfKO mice. There was no main effect on the number of total transitions (Fig. 2A) or the latency to enter the dark (Fig. 2B). For time spent in the light, two-way ANOVA revealed a significant effect for genotype (F(1, 44) = 7.07, P = .01) but not for sex (F(1, 44) = 0.64, P = .43). These results indicate that loss of CREBRF does not impact locomotor ability, yet there was a decrease in anxiety-like behaviors in CrebrfKO mice that was not sex specific.

Figure 2.

Exploratory activity as a measure of anxiety-like behaviors in the light–dark box in male and female control and CrebrfKO mice. The total number of transitions between chambers (A) latency to dark (B), and time spent in the light (C) are shown. Data are mean ± SEM. *P = .01. Groups: control male (n = 9), CrebrfKO male (n = 10), control female (n = 15), CrebrfKO female (n = 15).

Since there was a significant decrease in anxiety-like behaviors in CrebrfKO compared with controls using the light–dark box, we next performed the more stressful elevated plus maze (14). There was no effect of genotype or sex on total arm entries (Fig. 3A), further supporting that loss of CREBRF does not disrupt locomotor activity. Upon examination of open versus closed arm preference, 2-way ANOVA showed that time spent in the closed arm was significantly influenced by genotype (F(1, 48) = 20.46, P < .0001) and post hoc analysis revealed that this was driven by an effect in control females versus CrebrfKO females (P = .0003) rather than males (P = .11). For open arm ratio, 2-way ANOVA analysis showed an effect for genotype (F (1, 48) = 14.94, P = .003) and for sex (F(1, 48) = 12.04, P = .0011). Post hoc comparisons revealed this was driven by an increase in the open arm ratio in CrebrfKO females versus control males (P < .0001), versus control females (P < .001), versus CrebrfKO males (P < .05, Fig. 3C) but not for control males versus CrebrfKO males (P = .16). These results in the elevated plus maze, along with the light–dark box, demonstrate a reduction in anxiety-like behaviors under modest and more robust anxiety-inducing stimuli, mainly driven by genotype.

Figure 3.

Anxiety-like behavior on the elevated plus maze in male and female control and CrebrfKO mice. The total arm entries (A), time spent in the closed arm (B), and open arm ratio (C) are shown. Data are mean ± SEM. *P < .0001. Groups: control male (n = 9), CrebrfKO male (n = 10), control female (n = 16), CrebrfKO female (n = 17).

Prior work in postpartum female CrebrfKO mice identified a decrease in plasma corticosterone (12). Thus, we sought to perform a more extensive characterization of circulating glucocorticoids and HPA axis function in male and female CrebrfKO mice. Because CRH is a critical coordinator of the HPA axis in response to stress, we performed the CRH challenge to measure pituitary-mediated corticosterone release. Two-way ANOVA showed a significant effect of CRH administration on corticosterone among the groups (F(4, 174) = 82.02, P < .0001). These data indicate that the pituitary is able to respond to CRH by releasing ACTH to promote adrenal corticosterone release into the circulation similarly in CrebrfKO and control mice. Next, we tested the ability of Dex to provide negative feedback at the level of the hypothalamus and pituitary to inhibit CRH and ACTH release, respectively, and ultimately adrenal corticosterone release into the circulation. Two-way ANOVA revealed a significant difference of Dex administration on corticosterone in both genotypes (baseline compared with 6 hours; F(1, 54) = 84.41 P < .0001; Fig. 4B). Post hoc analysis showed that Dex was effective in suppressing circulating corticosterone compared to baseline for control males (P < .05), control females (P < .001), CrebrfKO males (P < .001), and CrebrfKO females (P < .0001). There was no effect of genotype or sex on adrenal weights (data not shown). Taken together, these data indicate that adrenal glucocorticoid release and HPA axis feedback loops (which require GR action) remain intact in CrebrfKO mice.

Figure 4.

Circulating corticosterone in response to a CRH challenge (A) and Dex suppression test (B) in male and female control and CrebrfKO mice. Data are mean ± SEM. *P < .0001. Groups for CRH challenge: control male (n = 10), CrebrfKO male (n = 9), control female (n = 13), CrebrfKO female (n = 7). Groups for Dex suppression test: control male (n = 3), CrebrfKO male (n = 4), control female (n = 10), CrebrfKO female (n = 14).

Because anxiety-like behaviors were lower in both male and female CrebrfKO mice than in control mice (Figs. 2 and 3), but no overall defects were identified in corticosterone release in response to the CRH challenge (Fig. 4A), we next measured glucocorticoid secretion in response to an acute stressor. Using restraint stress, 2-way ANOVA revealed a significant interaction for time (F(2, 75) = 24.07, P < .0001). Post hoc analysis showed restraint stress at 30 minutes significantly increased corticosterone compared to baseline for control males (P < .05), control females (P < .01), CrebrfKO males (P < .05), and CrebrfKO females (P < .05). At 90 minutes, post hoc analysis revealed there was a significant increase compared with baseline only in control males (P < .05) and control females (P < .05) but not CrebrfKO males (P = .73) or CrebrfKO females (P > .99). These data suggest that the initial activation of the HPA axis is not impacted by genotype, but that there is an early termination of the stress response in CrebrfKO mice (Fig. 5).

Figure 5.

Circulating corticosterone in response to restraint stress in male and female control and CrebrfKO mice. Corticosterone levels in males (A) and females (B) are shown at baseline, after 30 minutes of restraint stress, or after 60 minutes of recovery. Data are mean ± SEM. *P < .01 and #P < .05 versus respective baseline group. Groups: control male (n = 6), CrebrfKO male (n = 6), control female (n = 11), CrebrfKO female (n = 7).

To further characterize the HPA axis in CrebrfKO mice, we measured the natural diurnal corticosterone rhythm in undisturbed conditions using 2 separate cohorts of litter-matched mice that had not undergone prior testing (Fig. 6A). On day 4, the first day of blood collection, circulating corticosterone was elevated and highly variable at all 3 time points in both control and CrebrfKO mice (Fig. 6B). By the fifth day, control mice exhibited a normal diurnal corticosterone pattern with highest levels immediately prior to the dark cycle, a decrease at the start of the light cycle, and the lowest levels at the midpoint of the light cycle, which was subsequently maintained on days 6 and 7 (Fig. 6B). CrebrfKO mice, on the other hand, at the start of the light cycle on day 2 had their highest corticosterone levels recorded and showed an overall decline rather than the expected diurnal pattern on days 6 or 7 (Fig. 6B). Integration of data over the entire testing period analyzed using 2-way ANOVA revealed a significant interaction for time and genotype (F(2, 86) = 4.7, P = .0116) = P < .0001) as well as effect for time (F(2, 86) = 11.9, P < .0001) but not genotype (F(1, 86) = 2.43, P = .1227). Post hoc comparison showed in control females there was a significant increase at 8 pm compared with 8 am (P = .0006) and 2 pm (P < .001). This was not present in CrebrfKO mice at 8 pm compared with 8 am (P = .97) or 2 pm (P = .78). Additionally at 8 pm, there was a significant decrease in CrebrfKO female mice compared with control female mice (P = .0256), suggesting the normal diurnal rhythm was blunted in CrebrfKO mice with significantly lower circulating corticosterone prior to the dark cycle.

Figure 6.

Diurnal circulating corticosterone in unhandled female control and CrebrfKO mice. Schematic protocol and timeline of experiment (A), diurnal corticosterone rhythm over time (B), and integration of daily corticosterone levels at 8 am, 2 pm, and 8 pm over the entire testing period (C). Data are mean ± SEM. *P < .0001 versus 8 am control, $P < .001 versus 2 pm control, #P < .05 versus respective genotype at 8 pm. Groups: control female (n = 6), CrebrfKO female (n = 5).

Discussion

In this study, we examined the impact of CREBRF on behavior and physiology using a global CrebrfKO mouse model. Our findings indicate that loss of CREBRF (1) does not alter exploratory or ambulatory behavior (2) reduces anxiety-like behaviors using the light–dark box and elevated plus maze, (3) does not alter circulating corticosterone in response to a CRH challenge or the Dex suppression test, (4) results in an early termination of circulating corticosterone after restraint stress, and (5) blunts the circulating corticosterone peak entering the dark cycle. These results demonstrate that CREBRF contributes to stress-induced behavioral and physiological phenotypes.

Despite being widely expressed in humans and mice, CREBRF’s physiological function remains largely unrecognized. Behavioral effects observed in CrebrfKO mice suggest CREBRF may exert direct or indirect effects in the brain. Specifically, female CrebrfKO mice exhibit changes in behavior such as an increase in maternal neglect and a decrease in anxiety-like behaviors (12). Therefore, we examined anxiety-related behaviors in both male and female CrebrfKO mice using tests under varying levels of inherent stress. There were no differences in the number of line crosses during the open field test, transitions during the light–dark box, or total arm entries during the elevated plus maze, indicating no gross motor defects in CrebrfKO mice. In the light–dark box, we detected an increase in time spent in the light portion, and in the elevated plus maze, we detected a decrease in the closed arm as well as preference for the open arm in CrebrfKO males and females. These results are consistent with previous findings in CrebrfKO female mice (12), aligns with findings that overall females compared with males display an increase in activity (24), and more importantly demonstrated that CREBRF impacts anxiety-like behaviors.

Previous work in vitro demonstrated that ectopic overexpression of CREBRF decreased GR protein levels and global deletion of CREBRF in female mice lowered circulating corticosterone in the postpartum period (12). These limited data suggested a possible link between CREBRF and glucocorticoid physiology that warranted further analysis. Due to the established role of glucocorticoids in anxiety-like behavior (6-9) and our findings that CrebrfKO mice are more resilient to stress, we examined whether the loss of CREBRF impacted circulating glucocorticoids. Both CrebrfKO and control mice exhibited similar circulating corticosterone at baseline and initially after restraint stress. This was consistent with results during the CRH challenge in which male and female CrebrfKO and control mice had similar elevations in circulating corticosterone. However, during recovery from restraint stress, there was a significant decrease in circulating corticosterone in CrebrfKO compared with control mice, suggesting a more robust and/or early termination of the stress response to a physical stressor. There were no defects in negative feedback during the Dex suppression test, suggesting that the reduction in circulating corticosterone may result from a different rate of negative feedback in CrebrfKO mice compared with controls and unlikely to result from CREBRF effects on corticosterone metabolism and elimination. In addition to CREBRF’s potential function in the dynamics of glucocorticoid response to a stressor, CREBRF appears to also impact the kinetic profile of the glucocorticoid circadian rhythm. Specifically, there was a reduction in corticosterone immediately prior to the dark cycle in CrebrfKO female compared to the normal diurnal peak in control female mice. Given that the glucocorticoid secretion pattern is regulated by the circadian system (reviewed in (25)), and strong evidence in humans and rodents for circadian rhythm disruption impacting anxiety (26), alterations in the diurnal glucocorticoid rhythm suggests a potential mechanisms for the reduction in anxiety-like behaviors observed in CrebrfKO mice. Taken together, these findings provide key evidence that CREBRF plays a role in the dynamic regulation of circulating glucocorticoids under distinct physiological conditions (ie, circadian rhythm and restraint stress) and is not sex specific. Our studies do not rule out the possibility that CREBRF impacts intracellular glucocorticoid action, either through its influence on ligand metabolism via 11β-hydroxysteroid dehydrogenase enzymes, or GR activity.

Similar to GR expression, CREBRF has been identified in regions implicated in anxiety and stress responses including the amygdala (27), hippocampus (12, 27), cortex (27), and hypothalamus (28, 29). Prior work has determined within these specific regions that the overexpression of the GR results in delayed termination of the HPA axis after restraint stress (19) and increases anxiety-like behaviors in the light–dark box and elevated plus maze (10), whereas disrupting GR conversely reduces anxiety-related behaviors (8). Therefore, our findings that CrebrfKO mice have an early termination after restraint stress and decreased anxiety-like behaviors would suggest a reduction of GR levels in these key brain regions. This is inconsistent with in vitro work that found increasing CREBRF significantly reduced GR (12) and instead indicates CREBRF may function as a coregulator (12) to mediate GR transcriptional responses in these brain regions. This would account for similar corticosterone levels between groups before and immediately after restraint stress, during the CRH challenge and Dex suppression test, and the diurnal rhythm at 8 am and 2 pm, yet differences in a reduction of circulating corticosterone during the initial rise into the dark period, early termination of the HPA axis after an acute stressor, and behavioral responses to low and high stress situations in CrebrfKO mice.

In summary, global loss of CREBRF resulted in decreased anxiety-like behaviors and circulating glucocorticoids that was not sex specific. This study corroborates prior findings that female CrebrfKO mice display a reduction in anxiety-like behavior and demonstrate comparable responses in male mice. In addition, circulating corticosterone was reduced both during recovery from an acute stressor and prior to the dark cycle in CrebrfKO mice. Collectively, these results expand the role of CREBRF in stress-related behaviors and regulation of circulating glucocorticoids.

Glossary

Abbreviations

ACTH

adrenocorticotropic hormone

ANOVA

analysis of variance

CREBRF

CREB3 regulatory factor

CRH

corticotropin-releasing hormone

Dex

dexamethasone

GR

glucocorticoid receptor

HPA

hypothalamic–pituitary–adrenal

PVN

paraventricular nucleus of the hypothalamus

SEM

standard error of the mean

Additional Information

Disclosure Summary: The authors have nothing to declare.

Data Availability

The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.