Orexin 1 Receptor Antagonism in the Basolateral Amygdala Shifts the Balance from Pro- to Anti-stress Signaling and Behavior

Abstract

BACKGROUND:

Stress produces differential behavioral responses through select molecular modifications to specific neurocircuitry elements. The orexin system targets key components of this neurocircuitry in the basolateral amygdala (BLA).

METHODS:

We assessed the contribution of intra-BLA Orexin 1 receptors (Orx₁R) in the expression of stress-induced phenotypes of mice. Using the Stress Alternatives Model (SAM), a social stress paradigm that produces two behavioral phenotypes, we characterized the role of intra-BLA Orx₁R using acute pharmacological inhibition (SB-674042) and genetic knockdown (AAV-U6-Orx₁R-shRNA) strategies.

RESULTS:

In the BLA, we observed that Orx₁R (HCRTR1) mRNA is predominantly expressed in CamKIIα⁺ glutamatergic neurons and rarely in GABAergic cells. While there is a slight overlap in HCRTR1 and Orexin 2 receptor (Orx₂R; HCRTR2) mRNA expression in the BLA, we find that these receptors are most often expressed in separate cells. Antagonism of intra-BLA Orx₁R after phenotype formation shifted behavioral expression from stress sensitive (Stay) to resilient (Escape) responses, an effect that was mimicked by genetic knockdown. Acute inhibition of Orx₁R in the BLA also reduced contextual and cued fear freezing responses in Stay animals. This phenotype-specific behavioral change was accompanied by biased molecular transcription favoring HCRTR2 over HCRTR1, and MAPK3 over PLCB1 cell signaling cascades and enhanced BDNF mRNA.

CONCLUSIONS:

Functional reorganization of intra-BLA gene expression is produced by antagonism of Orx₁R, which promotes elevated HCRTR2, greater MAPK3, and increased BDNF expression. Together, these results provide evidence for a receptor-driven mechanism that balances pro- and anti-stress responses within the BLA.

INTRODUCTION

Stress-induced alterations in neurocircuitry result in divergent behavioral responses. Enhanced stress reactivity (pro-stress) in rodent models is similar to human affective dysfunction in mood disorders like depression, fear-/anxiety-related disorders, or post-traumatic stress disorder (PTSD) (1). Current pharmacotherapies for affective disorders have limited success, and a mechanistic understanding remains elusive.

Balance within key stress circuits may be disrupted during periods of intense or prolonged stress to shift signaling dynamics in pro- or anti-stress pathways (2-4). Stressful stimuli are interpreted, in part, through converging signals in the basolateral amygdala (BLA), where glutamatergic projection neurons are influenced by distinctive GABAergic interneurons, to direct behavioral responses (5). Additionally, activity in the BLA is modified by hypothalamic orexinergic neurons, which are critical for panic (6, 7) and motivation (8, 9).

Orexin A (Orx_A) and orexin B (Orx_B), neuromodulators derived from a single pre-propeptide, activate two G-coupled protein receptors: orexin 1 receptors (Orx₁R) bind Orx_A and Orx_B (EC₅₀=30nM vs 2,500nM), as do orexin 2 receptors (Orx₂R; EC₅₀=38nM vs 36nM)(10, 11). These receptors stimulate G_q proteins which increase intracellular Ca²⁺ (11) to activate phospholipase C (PLC) pathways (12). The PLC_β1 isozyme variant is transcribed in the amygdala (13), and its dysfunction is linked to psychopathologies like depression (14), bipolar disorder (15), addiction (16), and schizophrenia (17, 18).

Stimulation of Orx₁R can also activate extracellular signal-regulated protein kinase (ERK). In the amygdala, recruitment of ERKs is important for consolidation, reconsolidation, and extinction of fear memories (19, 20). While Orx₁R in the BLA are important in regulating fear (21, 22), depression (23, 24), and anxiety (25), it is unclear how shifts in molecular signaling cascades mediate such responses and initiate stress-induced phenotype development.

Utilizing the Stress Alternatives Model (SAM), a behavioral paradigm that separates individuals into social stress resilient (Escape; as validated by Social Interaction/Preference [SIP] test) and vulnerable (Stay) populations (26), we explored how Orx₁R activity in the BLA is involved in the formation of stress-related phenotypes. As a social interaction and avoidance paradigm in which smaller subjects encounter intense attacks from larger novel aggressors over a four-day period, the SAM produces two separate subsets of animals, exhibiting social avoidance or enhanced fear conditioned responses (27, 28). Unlike a traditional social defeat outcome, the SAM provides mice an opportunity to avoid social aggression by exiting the arena through escape tunnels only large enough for the smaller mouse. By the end of the second day of social interaction, test subjects commit to a phenotype: Escape or Stay. These stable phenotypes may be altered through pharmacological manipulations (Escape reduced by anxiogenic drugs, Stay reduced by anxiolytic drugs) administered on the third day of the SAM (28-30). Thus, the SAM is a useful tool for studying the development of stress-induced phenotypes, while providing an opportunity to explore physiological and clinically relevant molecular mechanisms.

We investigated if inhibition of intra-BLA Orx₁R, alters the formation of social stress-induced behavioral phenotypes. We predict that pharmacological inhibition or genetic knockdown will shift behavioral patterns in vulnerable (Stay) populations toward resilience (Escape). Further, we explored if Orx₁R inhibition affects conditioned fear responses and alters expression of genes responsible for balancing signaling in pro- and anti-stress neurocircuitries. Together, these results allow us to propose a neurocircuit model that defines the role of intra-BLA Orx₁R signaling in the balance of pro- and anti-stress states.

METHODS AND MATERIALS

Social Stress and Choice Paradigm

Aggressive social interactions between larger novel CD1 and smaller male C57BL/6NHsd mice dyads in the SAM apparatus (Fig. 1) involves four trials, lasting up to five minutes each, allows test animals the opportunity to shorten stressful encounters by making use of size-restricted tunnels at the apical end of the oval open field interaction arena. A tone given during isolation in the SAM apparatus prior to social interaction permits comparisons between cued and contextual fear conditioning. The escape routes provide a choice, producing two stable phenotypes: active avoidance (Escape) and accepting confrontation (Stay), which may be modified by drug treatment on Day 3. The treatment regimen allows for statistical comparisons between groups, and within subjects, by comparing responses to SAM interactions before and after treatment. All procedures were performed in accordance with the Guide for the Care and Use of Laboratory Animals (NIH Publications No. 80-23) and approved by the University of South Dakota Institutional Animal Care and Use Committee.

Figure 1.

The Stress Alternatives Model (SAM) is used to assess the development of stress-induced phenotypes. A) The SAM is a 4-day behavioral paradigm in which (I) a test mouse is placed into an opaque cylinder, (II) presented a tone, (III) exposed to social aggression, and commits to a phenotype: (IV) Escape or (V) Stay. B) The behavioral timelines for (I) pharmacology and (II) genetic knockdown experiments (mice are the same age at testing) include surgeries targeting the BLA, SAM exposure (Days 1-4), and the testing of contextual and cued fear responses (Day 5).

Experimental Overview (see also Supplemental Information)

The primary treatments for these experiments is inhibition of BLA Orx₁R, via the antagonist SB-674042 (0.3 nmol/0.3 μL delivered bilaterally intra-BLA, 1h prior to interaction on Day 3), contrasted with Orx₁R stimulation (accomplished by Orx_A + Orx₂R antagonism), or short-hairpin knockdown (bilateral intra-BLA transfection beginning 30 days prior to SAM interaction). Considering the difference in timing of delivery, these treatments were done and analyzed separately, with a priori hypotheses. All behavioral measures were performed during the dark cycle when the animals are active, and included Escape (use of the apical tunnels), Stay (remaining in the SAM arena with the novel aggressor), time spent attentive to the escape hole, latency to escape (for Escape mice), fear conditioned freezing (measured in response to the tone [CS] and context, prior to the social interaction unconditioned stimulus [US], and as a conditioned response [CR on Day 5] in the absence of the US), and food intake. Thus, treatment groups included home cage controls, and intra-BLA SB-674042 (or vehicle, Orx_A, Orx_A + MK-1064, MK-1064) injection of Escape and Stay mice. In addition, transgenic treatment groups included home cage controls, intra-BLA AAV-Orx₁R-shRNA injection, and intra-BLA AAV-scramble-shRNA injection. Brains and blood were collected for visual representations of gene expression (using RNAscope) of HCRTR1, HCRTR2, calbindin (CALB1), Ca⁺⁺/Calmodulin Kinase type 2 alpha (CAMKIIα), Glutamate Decarboxylase (GAD1), and parvalbumin (PVALB) in BLA, as well as to measure plasma concentrations of the stress hormone corticosterone (by enzyme linked immunosorbent assay). Gene expression (using RT-qPCR) of HCRTR1, HCRTR2, PLCB1, MAPK1, MAPK3, BDNF, and GAPDH (housekeeping gene) were measured in BLA tissue. All experimental designs and statistical analyses were based on a priori hypotheses, using two-way repeated measures ANOVA, two-way ANOVA, one-way ANOVA, Regression analyses, and t-test, followed (where appropriate) by post hoc analyses.

Results

OrxR Expression in BLA

The glutamatergic marker, CamKIIα identified the vast majority of BLA neurons (~80%; Fig. S2) as well as those expressing HCRTR1 (31, 32) (also in some calbindin-GABAergic neurons; Fig. 2). Few (<20%) BLA HCRTR1-possessing cells express GAD1 (GABAergic marker) and co-express parvalbumin (PV, ~10%; Figs. 2G-K). Our results suggest HCRTR1 is expressed in 10-15% of BLA glutamatergic neurons and ~5% of GABA cells (Fig. 2K). In BLA cells, mRNA for HCRTR1 and HCRTR2 largely do not overlap, ~80% of HCRTR1⁺ cells do not co-express HCRTR2 (Figs. 2L-O). Specific BLA GABAergic neurons may predominantly localize Orx₂R (Fig. 2P) (28).

Figure 2.

In the untreated BLA, Orx₁R are expressed predominantly in glutamatergic neurons and are rarely co-expressed with Orx₂R. A) Imaged sections containing BLA cells (LA = lateral amygdala) stained with probes targeting mRNA of B) HCRTR1 (red), C) CamKIIα (green), and D) Calb (Magenta) revealed when E) merged (with DAPI) that F) Orx₁R⁺ cells mostly co-express the glutamatergic cell marker, CamKIIα (N = 4, F_2,9 = 54.4, p < 0.001; CamKIIα⁺ vs Calb⁺: t₆ = 10.4, p < 0.001; CamKIIα⁺ vs Other: t₆ = 5.2, p < 0.001; Calb⁺ vs Other: t₆ = 5.2, p < 0.001; bars are statistically different from one another as illustrated with unique letters, e.g. A is significantly different from B and C; p < 0.001). G) Expression of HCRTR1 (red) GAD₆₇ (GAD1) mRNA (yellow) infrequently overlap with H) most HCRTR1⁺ cells being absent of the GABAergic marker (N = 5, t₈ = 29.5, *p < 0.001). I) While a subset of BLA GABAergic neurons produce the calcium-binding protein parvalbumin (Pvalb⁺), J) HCRTRV (red) cells are mostly absent of Pvalb expression (light blue) with K) less than 10% being both HCRTR1⁺ and Pvalb⁺ (N = 4, t₆ = 23.1, *p < 0.001). K) Further, more BLA glutamatergic (CamKIIα⁺) neurons (compared to GABAergic → GAD1⁺) also express HCRTR1 (N = 9, t₇ = 3.2, *p < 0.015). L) Images of BLA cells with fluorescent markers labeling M) HCRTR1 mRNA (red) and HCRTR2 mRNA (green) demonstrate N) most BLA cells express neither HCRTR1 nor HCRTR2 (N = 4, F_2,9 = 42.1, p < 0.001; HCRTR1⁺ vs Other, t₆ = 7.5, p < 0.001; HCRTR2⁺ vs Other, t₆ = 8.4, p < 0.001; bars are statistically different from one another as illustrated with unique letters, e.g. A is significantly different from B). O) Most HCRTR1⁺ cells in the BLA do not express HCRTR2 (N = 4, t₆ = 10.1, *p < 0.001), as depicted in P) showing Orx₁R on glutamatergic neurons.

Motivation for Active Avoidance (Escape)

In the SAM, animals evenly choose one of two stable (27-29, 33) behavioral phenotypes, Escape (44.7%) or Stay (55.3%; Figs. 1A, S1B, C)(26, 27, 29, 33, 34). Time spent investigating escape routes predict active avoidance and indicates motivation to escape (28). Time spent attentive to the hole was significantly greater in vehicle-treated Escape mice (Fig. 3A), but intra-BLA infusion of the Orx₁R antagonist (Escape: Fig. 3B, C; Stay: Fig. 3B, D) or AAV-U6-Orx₁R-shRNA (Fig. 3E) increases attention to the escape route. Further, receptor activation with Orx_A reduced time Escape mice spent investigating the escape route (Fig. S3).

Figure 3.

Motivation toward Escape behavior is impacted through inhibition of intra-BLA Orx₁R. A) Escape mice, as compared to those expressing the Stay phenotype, spend a greater % of time investigating the SAM escape routes (N = 19, Phenotype Effect: F_1,51 = 16.4, p < 0.001; Escape vs Stay: Day 1, t₁₇ = 2.6, *p ≤ 0.018; Day 2, t₁₇ = 2.5, *p ≤ 0.017; Day 4, ti₇ = 4.2, *p < 0.001). B) While Escape mice, in general, explore the escape routes more often, C) inhibition of intra-BLA Orx₁R promotes even more attention toward the escape tunnels (N = 34, Treatment Effect: F_1,30 = 7.7, p ≤ 0.019; Day 3 Vehicle Escape vs Orx₁R Ant. Escape, t₁₀ = 2.5, ⁺p ≤ 0.018). D) Antagonism of intra-BLA Orx₁R only slightly stimulates escape route exploration in Stay mice (Day 4 Vehicle x Orx₁R Ant., t₂₀ = 2.1, ⁺p ≤ 0.05). E) Knockdown of intra-BLA Orx₁R temporarily and minimally increases attention toward escape on Day 3 of the SAM (N = 22, Day 3 Scramble vs AAV-Orx₁R-shRNA, t₂₀ = 2.4, ⁺p ≤ 0.024). F) Illustration demonstrating inhibition of intra-BLA Orx₁R on predominantly on glutamatergic neurons promotes attention toward the escape route in the SAM arena. In Pharmacological Experiments, drug treatment is administered on Day 3 as designated by the bold square. Note, the data plotted in A, C, and D are the same as those graphed in B; we have separated out these individual comparisons for the sake of clarity.

Avoidance (Escape)

Upon intra-BLA injections of an Orx₁R antagonist on SAM Day 3, a substantial number of Stay mice exhibited Escape behavior (Fig. 4A), with a 30% shift that day, and a significant increase the day after (Day 4 = 70% increase). Interestingly, intra-BLA activation of both Orx receptors with Orx_A or biased activation of Orx₁R (Orx_A + Orx₂R antagonist) blocked Escape behavior in a small, but not statistically significant, proportion of mice on Days 3 and 4 (Fig. S4), exhibiting deviation from stable phenotype behavioral patterns.

Figure 4.

Intra-BLA Orx₁R mediates stress-related behavioral phenotype development. A) Infusion of an Orx₁R antagonist (SB-674042) into the BLA promotes Escape behavior in Stay mice (N = 22, Day 4, χ²: F₁ = 9.3, *p < 0.001). B) Knockdown of Orx₁R (AAV-Orx₁R-shRNA) upsets normal Day 2 phenotype commitment behavior (as observed with AAV-Scramble-shRNA controls), inducing more Escape behavior on Days 3 and 4 (N = 22). C) Escape animals learn to efficiently utilize the escape route to avoid social aggression (Escape latency = Time with Social Aggressor) over the course of 4 days while Stay mice remain with the aggressor (N = 19, Phenotype Effect: F_1,45 = 175.3, p < 0.001; Time Effect: F_3,45 = 26.1, p < 0.001; Interaction Effect: F₃,45 = 26.1, p < 0.001; Escape vs Stay: Day 2, t₁₇ = 5.8, *p < 0.001; Day 3, t₁₇ = 10.6, *p < 0.001; Day 4, t₁₇ = 11.9, *p < 0.001; Within Escape phenotype comparison, F_3,18 = 17.8, p < 0.001, Day 1 vs Day 3, t₆ = 5.7, p < 0.001; Day 1 vs Day 4, t₆ = 6.5, p < 0.001; Day 2 vs Day 3, t₆ = 2.9, p ≤ 0.009; Day 2 vs Day 4, t₆ = 3.7, p ≤ 0.002; p < 0.05 for Days marked with unique lettering, e.g. A is different from B and C). D) Antagonizing intra-BLA Orx₁R promotes aggressor avoidance in Stay mice (N = 34, Time Effect: F_3,54 = 2.9, p ≤ 0.043; Interaction Effect: F_3,54 = 2.9, p < 0.043; Day 4 Vehicle Stay vs Orx₁R Ant. Stay, t₂₀ = 3.4, ⁺p < 0.001), but has no effect on those animals exhibiting the Escape phenotype. E) Knockdown of intra-BLA Orx₁R does not impact the overall latency of aggressor avoidance (N = 22). Overall, F) inhibition of Orx₁R in the BLA appears to prompt Escape behavior. In Pharmacological Experiments, drug treatment is administered on Day 3 as designated by the bold square.

As knockdown reduced Orx₁R expression prior to stressful interactions, we did not expect a dramatic change in behavior over the course of SAM trials, but AAV-U6-Orx₁R-shRNA yielded incrementally (though not significantly) more escape on the last two days (Fig. 4B). By Day 4, 72.7% of AAV-U6-Orx₁R-shRNA-treated mice displayed Escape compared to 54.5% of the scramble control.

Escape mice spent significantly less time in the SAM arena with the CD1 mouse on Days 2 through 4 (26, 27, 29, 33, 35), thus escape latency was reduced (Fig. 4C). Stay mice remain for the entire 5 min period, unless treated with Orx₁R antagonist, significantly reducing time spent with aggressive CD1 mice on Day 4 (Figs. 4D, F). Inhibition of Orx₁R did not influence escape latency in Escape animals (Figs. 4D). Importantly neither of the Orx₁R manipulations, antagonist or knockdown treatments, influenced arousal/locomotion (Figs. S5), but did result in small but significant decreases in food intake and body weight (Figs. S6).

Cued and Contextual Fear Conditioning

Cued fear responses significantly enhanced freezing in both Escape and Stay phenotypes, and Stay mice displayed heightened freezing behavior to context (opaque cylinder divider) as well (Figs. 5A, B). Although inhibition of intra-BLA Orx₁R did not affect the fear freezing profile in Escape mice, antagonist-treated Stay mice exhibited significantly reduced contextual and cued fear responses (Figs. 5B, S7; Table S1). Like mice of the Escape phenotype, knockdown of BLA Orx₁R did not affect conditioned freezing behavior (Fig. S8). Importantly, activation of intra-BLA Orx receptors with Orx_A did not change the fear freezing profile in Escape or Stay mice compared to vehicle (Figs. 5B, S7; Table S1). However, biased stimulation of Orx₁R in the BLA with a combination of Orx_A + Orx₂R antagonist eliminated the conditioned response in Escape, but not Stay mice (Figs. 5B, S7; Table S1). Further, acute inhibition of Orx₂R in the BLA eliminated the cued freezing response in Escape mice and significantly reduced freezing during the post-tone period (Figs. 5B, S7; Table S1). Stay mice treated with an Orx₂R antagonist displayed no statistical differences in the levels of contextual and cued freezing (Figs. 5B, S7; Table S1).

Figure 5.

Inhibition of Orx₁R in the BLA reduces contextual/cued fear responses and stress hormone concentrations. A) Although both Escape and Stay phenotypes learn to associate a cue (tone, CS⁺) with social aggression (Phenotype Effect: F_1,17 = 7.6, p < 0.013; CS Effect: F_1,17 = 47.7, p < 0.001; Escape CS⁻ vs CS⁺, t₆ = 3.9, ^#p < 0.008; Stay CS⁻ vs CS⁺, t₁₁ = 5.7, ^#p < 0.001), Stay mice exhibit heightened freezing behavior to both context (CS⁻; t₁₇ = 2.8, *p < 0.011) and tone (CS⁺; t₁₇ = 2.3, *p ≤ 0.033). Baseline measurements of freezing are represented by a dotted line. B) Antagonism of intra-BLA Orx₁R reduces conditioned fear responses in Stay animals while Orx₂R inhibition diminishes fear freezing in Escape mice (N = 71; * represents significant differences compared to Escape mice in the same treatment group; + signifies significance compared to Vehicle-treated animals in the same phenotype group; ! identifies significant differences compared to OrxA-treated mice; $ denotes significant differences compared to Orx₂R antagonist-treated animals). See Supplemental Information Figure S7 for specific a priori hypotheses comparisons. C) Mice exposed to social stress produce elevated levels of stress hormone (N = 39, F_2,12 = 24.3, p < 0.001; Cage Control vs Vehicle Escape, t₅ = 3.1, ^p ≤ 0.028; Cage Control vs Vehicle Stay, t₉ = 9.9, ^p < 0.001); however, Stay animals have the highest concentration (Vehicle Escape vs Stay, t₁₀ = 2.6, p ≤ 0.025). Inhibition of intra-BLA Orx₁R reduces corticosterone levels in Stay mice (Vehicle Stay vs Orx₁R Ant. Stay, t₁₀ = 5.1, ⁺p < 0.001; Orx₁R Ant. Stay vs Orx_A Stay, t₆ = 3.3, ^!p ≤ 0.002).

Corticosterone Concentrations

Social stress in SAM interactions increases corticosterone concentrations in both Escape and Stay animals (27, 28, 33), although Stay mice have higher levels of corticosterone compared to Escape. Inhibition of BLA Orx₁R decreased Stay corticosterone concentrations compared to vehicle-treated Stay animals; and did not differ significantly from non-stressed mice (Fig. 5C). Treatments with OrxA or the combination of Orx_A and an Orx₂R antagonist did not change corticosterone levels relative to vehicle-treated controls, however, the differences between Escape and Stay were eliminated and levels were elevated compared to Orx₁R antagonist-treated mice (Fig. 5C). Inhibition of BLA Orx₁R not only reduces social fear responses, but also reverses social stress responsiveness.

Antagonism of intra-BLA OrxR Recruits Alternative Signaling

Although HCRTR1 expression was unaltered following vehicle treatment, Orx₁R antagonism reduced intra-BLA HCRTR1 in Escape mice compared to non-stressed cage controls (Fig. 6A), and simultaneously elevated HCRTR2 expression in Stay mice compared to Escape and vehicle-treated Stay mice (Fig. 6B; Table S2). In vehicle controls HCRTR2 expression was higher in Escape mice compared to both Stay and Orx₁R antagonist-treated Escape mice (Fig. 6B; Table S2). A reduction in HCRTR1 gene expression after Orx₂R antagonism was observed, but only in Stay animals relative to vehicle (Fig. 6A; Table S2). Expression of HCRTR2 in the BLA was reduced in both Escape and Stay phenotypes after blocking Orx₂R, contrasting with Orx₁R antagonism, which enhanced HCRTR2 mRNA levels in Stay mice (Fig. 6B; Table S2).

Figure 6.

Transcriptional changes (relative to home-cage naïve controls) in the BLA after Orx₁R or Orx₂R antagonism shifts signaling profile. A) Antagonism of Orx₁R in the BLA reduces HCRTR1 expression (N = 45, Treatment Effect: F_2,27 = 3.5, p ≤ 0.043), but only significantly so in animals expressing the Escape phenotype (Cage Control vs Orx₁R Ant. Escape, t₁₁ = 2.2, ^p ≤ 0.050); whereas infusion of an Orx₂R antagonist in the BLA reduces HCRTR1 expression in Stay mice compared to vehicle animals of the same phenotype (t₁₀ = 2.2, ⁺p < 0.044). B) While Escape mice (Treatment Effect: F_2,27 = 9.8, p < 0.001; Interaction Effect: F_2,27 = 8.6, p < 0.001) treated with vehicle express higher HCRTR2 levels compared to Stay mice (t₉ = 3.0; *p < 0.016) and Orx₁R- or Orx₂R-antagonist-treated Escape animals (Vehicle vs Orx₁R Ant., t₇ = 2.6, ⁺p < 0.035; Vehicle vs Orx₂R Ant.: t₇ = 4.5, ⁺p < 0.001; Orx₁R Ant. vs Orx₂R Ant.: t₈ = 3.5, ^!p < 0.001), Orx₁R antagonism results in elevated levels (Escape vs Stay, t₁₀ = 2.2, *p ≤ 0.05; Vehicle vs Orx₁R Ant., t₁₂ = 2.4, ⁺p < 0.034) while Orx2R inhibition leads to a reduction (Vehicle vs Orx₂R Ant.: t₁₀ = 3.5, ⁺p < 0.002; Orx₁R Ant. vs Orx₂R Ant.: t₁₀ = 4.7, ^!p < 0.001) of HCRTR2 in Stay mice. C) A reduction of PLCB1 (Phenotype Effect: F_1,27 = 19.1, p < 0.001; Interaction Effect: F_2,27 = 4.3, p ≤ 0.023) that is found in Escape mice under control conditions (Cage control vs Vehicle Escape, t₁₀ = 5.1, ^p < 0.001; Escape vs Stay, t₉ = 5.0, *p < 0.001) and Orx₁R antagonism (Escape vs Stay, t₁₀ = 3.1, *p < 0.012; Cage Control vs Orx₁R Ant., t₁₁ = 3.3, ^p ≤ 0.007) is eliminated with intra-BLA Orx₂R antagonism (Vehicle vs Orx₂R Ant.: t₇ = 2.8, ⁺p < 0.017). D) While Stay mice treated with an Orx₁R antagonist express higher levels of MAPK3 (Phenotype Effect: F_1,27 = 11.3, p ≤ 0.002; Treatment Effect: F_2,27 = 4.3, p ≤ 0.023; Interaction Effect: F_2,27 = 5.1, p ≤ 0.013) in the BLA compared to Vehicle controls (t₁₂ = 3.1, ⁺p < 0.001), administration of an Orx₂R antagonist does not induce the same transcriptional response (Orx₁R Ant. vs Orx₂R Ant.: t₁₀ = 2.7, ^!p ≤ 0.022). E) Expression of BDNF in the BLA after treatment (Interaction Effect: F_2,27 = 10.6, p < 0.001) with an Orx₂R antagonist is enhanced in Escape mice (Orx₂R Ant. Escape vs Stay: t₈ = 2.9, *p ≤ 0.019; Vehicle vs Orx₂R Ant.: t₇ = 2.7, ⁺p ≤ 0.013; Orx₁R Ant. vs Orx₂R Ant.: t₈ = 2.5, ^!p ≤ 0.017) and reduced in Stay animals (Vehicle vs Orx₂R Ant.: t₁₀ = 2.2, ⁺p < 0.05; Orx₁R Ant. vs Orx₂R Ant.: t₁₀ = 3.9, ^!p < 0.001); a phenotypically opposite effect is observed after Orx₁R antagonism (Escape vs Stay, t₁₀ = 2.8, *p ≤ 0.018; Orx₁R Ant. Stay vs Vehicle Stay, t₁₂ = 2.2, ⁺p ≤ 0.049). Transcriptional changes after F) intra-BLA Orx₁R antagonism and G) Orx₂R inhibition are differentially regulated in a phenotype-dependent fashion.

Transcription of BLA PLC_β1 (PLCB1) mRNA (13) is important for Orx₁R signaling (36). We predicted Orx₁R antagonist might limit PLCB1 expression levels (Fig. 6C). Escape mice in both vehicle and Orx₁R antagonist groups expressed lower amounts of PLCB1 compared to Stay and cage control animals (Fig. 6C). Further, greater PLCB1 followed intra-BLA Orx₂R inhibition compared to vehicle-treated Escape mice (Fig. 6C; Table S2).

Alternative molecular pathways recruited during G_q activation are driven by ERK genes (MAPK1 & MAPK3). In Stay mice Orx₁R antagonism resulted in a significant increase in MAPK3 expression (MAPK1 mRNA was unaffected, Fig. S9) compared to similarly treated Escape, vehicle-treated Stay, and non-stressed cage control mice (Fig. 6D; Table S2). Inhibition of intra-BLA Orx₂R did not alter MAPK3 gene expression (Fig. 6D; Table S2).

The transcription of brain-derived neurotrophic factor (BDNF) is tied to neuroplasticity (37, 38) and behavioral changes like extinction of fear memories (39), so we predicted an increase in BDNF might be associated with intra-BLA Orx₁R inhibition (Fig. 6E). As hypothesized, intra-BLA Orx₁R antagonism resulted in elevated BDNF in Stay compared to Escape mice and vehicle-treated Stay mice (Fig. 6E; Table S2). Finally, Orx₂R antagonist treatment enhanced BDNF expression in Escape mice, while diminishing transcription in Stay animals, an effect that is phenotypically opposite to that observed after Orx₁R inhibition (Fig. 6E; Table S2). As Stay mice treated with an Orx₁R antagonist experienced shifts from stress-vulnerable to resilient behavioral responses, the alterations in gene expression reported here (Figs. 6F, G) may be implicit in this behavioral plasticity.

Molecular Restructuring is Related to Fear Responsiveness

Expression levels of HCRTR2, but not HCRTR1, in both vehicle- and Orx₁R antagonist-treated mice are negatively correlated with cued freezing (Figs. 7A, B). Relative expression levels of PLCB1 were positively correlated with cued freezing behavior in vehicle-treated mice (Fig. 7C); however, this relationship is not observed after intra-BLA Orx₁R inhibition (Fig. 7D). Contextual freezing behavior was associated with MAPK3 expression in only vehicle-treated mice (Figs. S10I). By contrast, intra-BLA antagonism of Orx₁R cued freezing behavior was negatively correlated to MAPK3 expression (Fig. 7F), but not in vehicle-treated mice (Fig. 7E). The lack of gene expression correlations with cued fear freezing when phenotypes were assessed independently (Figs. S11), indicates that behavioral and transcriptional relationships exist within collective operational adaptations that link behavioral change to molecular modification. Importantly, no relationships between gene expression and conditioned fear freezing were observed for any of the tested cell signaling markers after Orx₂R antagonism except for BDNF (not Orx₁R antagonism), in which a significant negative correlation was revealed (Fig. S12E). Together, these results suggest a functional connection between Orx₁R antagonist-induced shifts in gene expression and fear-related behaviors.

Figure 7.

Conditioned fear freezing response is related to gene expression changes (fold change relative to home-cage naïve controls) resulting from intra-BLA Orx₁R antagonism. In both A) vehicle- (N = 11, F_1,9 = 16.1, R² = 0.6419, p ≤ 0.003) and B) Orx₁R antagonist-treated animals (N = 12, F_1,10 = 7.2, R² = 0.4197, p ≤ 0.023) a negative correlation exists between HCRTR2 expression and cued fear freezing. C) With vehicle treatment, relative PLCB1 expression is positively associated with cued fear freezing (F_1,9 = 6.4, R² = 0.417, p ≤ 0.0319). D) This relationship is not observed in mice that were administered an Orx₁R antagonist (F_1,10 = 0.7, R² = 0.0625, p ≥ 0.4333). E) While there is not a significant association between MAPK3 expression and cued fear freezing after vehicle treatment (F_1,9 = 3.8, R² = 0.2973, p ≥ 0.0828), F) a significant negative correlation is observed after Orx₁R antagonism (F_1,10 = 6.3, R² = 0.3877, p < 0.0306).

Potential molecular mechanism behind intra-BLA Orx₁R antagonism

To help generate a theoretical mechanism to explain the physiological basis surrounding the observed behavioral (and phenotypic) shifts resulting from intra-BLA inhibition of Orx₁R, we explored transcriptional relationships in systems that exhibited similar regression patterns (Fig. 8). With antagonism of Orx₁R, there is a strongly positive relationship between HCRTR2 and MAPK3 expression (Fig. 8A). Importantly, this association does not exist after vehicle or Orx₂R antagonist treatment (Fig. S13). While there are no observed relationships between BDNF and HCRTR2 expression levels (Figs. 8B, S13), BDNF expression is positively correlated to MAPK3 expression in animals treated with an Orx₁R antagonist (Fig. 8C). Notably, no relationships exist between FICRTR1 expression and the other genes of interest (Figs. S13). These data allowed us to predict a working model to explain how BLA Orx₁R may function to establish behavioral patterns consistent with stress-induced phenotype development (Fig. 9).

Figure 8.

The BLA transcriptional changes (relative to home-cage naïve controls) that result from Orx₁R antagonism form relationships that hint at molecular timelines and signaling dynamics. A) While relative gene expression of MAPK3 is positively correlated to the transcriptional changes of HCRTR2 (N = 12, F_1,10 = 8.3, R² = 0.4532 p ≤ 0.0164), B) there is no association between BDNFand HCRTR2 (F_1,10 = 0.3, R² = 0.0313, p ≥ 0.5822). However, C) a positive relationship emerges when comparing BDNF expression to that of MAPK3 (F_1,10 = 8.2, R² = 0.4517, p < 0.0167).

Figure 9.

Predicted circuit demonstrates the influence of intra-BLA Orx₁R antagonism, during endogenous stimulation through Orx_A and Orx_B release, on microcircuit dynamics in a phenotype-dependent fashion. A) Escape mice treated with an Orx₁R Antagonist (SB-674042) undergo molecular shifts, including a feedforward reduction of FICRTR1 and reduced PLCB1 transcription, leading to diminished orexin activity on glutamatergic neurons in the BLA. Escape mice also have a feedforward decrease in FICRTR2 expression, potentially via (un-diagrammed) negative circuit feedback, even while Orx₂R are stimulated. B) While Orx_B and Orx_A maintain stimulation of some GABAergic neurons through Orx2R, antagonism of some pyramidal neurons via intra-BLA Orx₁R inhibition differentially modifies molecular mechanisms in Stay mice through enhancement of Orx₂R (HCRTR2), ERK₁ (MAPK3), and BDNF transcription and increased orexin activity in Orx₂R-containing neurons (likely GABAergic cells).

Discussion

Antagonism of Orx₁R in the BLA can reverse or diminish expression of stress-related behavior. Our results suggest BLA Orx₁R play a central role in stress responsiveness (40, 41) and related behavioral, physiological, and molecular outcomes that are important components of affective disorders (42, 43), such as anxiety (7), depression, and PTSD. Acute inhibition of intra-BLA Orx₁R promotes Escape over Stay responses and limits freezing during fear conditioning in a phenotype-dependent way. Further, inhibition of Orx₁R alters gene expression associated with critical signaling cascades. Following intra-BLA Orx₁R antagonism, transcription for receptors and intracellular signaling becomes biased toward Orx₂R (FICRTR2) over Orx₁R (FICRTR1), and ERKi (MAPK3) over PLC_β1 (PLCB1) pathways. Importantly, even when BLA Orx₁R are inhibited, native Orx_A and Orx_B will bind Orx₂R. The relationship of these behavioral and molecular changes to enhanced expression of FICRTR2 mRNA, potentially in BLA neurons that do not contain Orx₁R (Figs. 2L-O), suggests receptor-mediated mechanisms that balance pro- and anti-stress responses in BLA microcircuits.

Aggressive social interactions in the SAM produced two behavioral phenotypes that represent risk assessment and choice: Escape and Stay. These phenotypes, like those exposed to social defeat paradigms (44, 45), exhibit resilience (tightly linked to Escape) and susceptibility (highly correlated with Stay) in the SIP test (28). However, unlike traditional social defeat, SAM-separated phenotypes are expressed early in the behavioral paradigm, providing insight into the development and progression of stress-induced behavior and pathophysiology. Anxiolytic drugs (such as CRF₁ receptor antagonist antalarmin and the Orx₂R agonist [Ala¹¹, d-Leu¹⁵]–Orx_B) promote escape, while anxiogenic drugs (such as the α₂ antagonist yohimbine and the Orx₂R antagonist MK-1064) delay and/or block escape behavior (28, 29). Surprisingly, neither the Orx₁R antagonist (Fig. 4D) nor knockdown (Fig. 4E) influenced escape latency, although it is reduced by anxiolytic factors such as exercise, Neuropeptide S, antalarmin, and increased by anxiogenic factors like yohimbine (29). We posit that enhanced escape on Day 4, following BLA Orx₁R inhibition (on Day 3, drug treatment), is a reflection of the shift toward anti-stress signaling indicated by downregulation in pro-stress signaling (HCRTR1), and upregulation of anti-stress systems (HCRTR2, MAPK3, BDNF). Thus, BLA dual Orx₁R/Orx₂R inhibition (DORA) may not promote behavioral change. These stress-induced effects are paired with important learning and motivational components during SAM interactions (27, 29, 33, 35), and in human affective disorders (46).

In addition to species-specific anxious behavior and learning, social stress promotes behavioral inhibition, depressed behavioral drive and motivation in some individuals (47), plus a lower rate of adaptive behavior (48). Behavioral depression reveals two distinctive phenotypes related to stress responsiveness in humans and other animals (45, 49, 50). In SAM social interactions, Stay animals do less exploration of the escape route (Fig. 3A) and show indecisiveness relative to escape (35). Measuring motivation in the SAM is derived from a simple choice process, Escape or Stay (26, 27). Antagonism and knockdown of Orx₁R increases interest in the escape route for both Stay and Escape mice (Figs. 3C, D). Thus, BLA Orx₁R regulate stress-induced motivational behaviors; greatest in Escape mice, but marking a dramatic behavioral reversal in Stay mice that typically avoid the escape route (Figs. 3B, C). Attention to the escape route happens prior to escape, and is thus the first evidence of phenotypic differentiation in the SAM (28, 35). Latency to escape, and escape behavior also are influenced by motivation, although as previously demonstrated, these behaviors are strongly affected by stress and fearfulness associated with familiarity of the SAM or social interaction (27-29, 33, 35). Our results, like those of others, suggest Orx activity plays a fundamental role in motivation (8, 51), and in this case, specifically in the BLA for behaviors associated with stress-related motivation and choice.

Understanding the development of choice and motivation in the SAM is enhanced by pairing aversive aggression (US) with a non-threatening stimulus (tone CS) prior to interaction, promoting potent cued and contextual conditioned responses (CR) similar to standard fear conditioning approaches that utilize foot shock as a US (52). While the CRs elicited are similar, e.g. freezing (53), the ethological and ecological relevance of the US to the subject are not. By associating naturally aversive US with a benign stimulus (54), the SAM allows views into development of fear learning as it relates to the etiology of stress-provoked neurocircuitry changes, and demonstrates a connection between stress-induced fear expression and phenotype (Fig. 5). While early work suggested only Stay mice exhibited cued fear learning (27, 33), it is now clear both Stay and Escape respond to auditory cues with enhanced freezing compared to pre-tone freezing, and Stay mice also show contextual (prior to the cue) fear conditioning (Fig. 5A).

Fear responses are mediated through Orx₁R activity in the amygdala, and in locus ceruleus (LC), which connects to the amygdala (22, 55-57). Our results similarly demonstrate Orx₁R, but not Orx₂R, inhibition diminishes both contextual and cued conditioned fear freezing in Stay animals (Figs. 5B; Table S1). While antagonizing Orx₁R reduces fear/panic-induced freezing (7, 56, 58), Orx₂R antagonism appears to eliminate fear learning in Escape mice, suggesting a phenotype-dependent effect (Figs. 5B; Table S1). Importantly, although Orx₂R antagonism in the BLA reduced cued freezing only in Escape mice, we have previously demonstrated a potential anxiogenic effect of blocking receptor function (25, 28). This response may be dependent on brain region, as Orx₂R activity in the nucleus accumbens shell and prelimbic prefrontal cortex may enhance anxious behavior (59, 60). Further, Orx₂R antagonism has demonstrated antidepressive capabilities in a clinical setting (61).

Stimulation of intra-BLA Orx₁R and Orx₂R receptors using Orx_A in Stay mice produces no reduction in contextual or cued fear conditioning (Fig. 5B; Table S1), suggesting the inhibition of both types of learned fear responses result specifically from Orx₁R inhibition in Stay mice. To clarify the roles of Orx₁R and Orx₂R, we administered Orx_A while concurrently inhibiting Orx₂R (MK1064), leaving Orx₁R stimulated, and again there was no statistically significant reduction in either type of fear conditioning response (Fig. 5B; Table S1). Interestingly, knockdown of Orx₁R did not affect the fear freezing profile (Fig. S8). As knockdown occurred before the introduction of social stress, activity levels of Orx₁R after SAM exposure allowed for fear learning (higher freezing after CS), but did not diminish freezing as observed with acute antagonism after stress and phenotype development (Fig. 5B).

Molecular gene expression during SAM fear conditioning and phenotype development indicated potential shifts in receptor-linked intracellular signaling cascades (Fig. 6). Acute inhibition of intra-BLA Orx₁R, by means of a feedforward rather than feedback mechanisms, lowered HCRTR1 expression in Escape mice while enhancing HCRTR2 mRNA in Stay animals (Figs. 6A, B). Antagonism of Orx₂R in BLA did the opposite, reducing HCRTR1 mRNA only in Stay mice, and in a similar feedforward way, decreasing HCRTR2 expression in both phenotypes (Figs. 6A, B). Mice exhibiting escape and reduced fear freezing, expressed lower PLCB1 compared to the Stay phenotype; an effect unaltered by SB-674042 treatment, but reversed by Orx₂R antagonism (Fig. 6C). However, intra-BLA Orx₁R antagonism increased MAPK3 and BDNF expression in Stay animals only, with Orx₂R inhibition having no effect on expression of MAPK3, and enhancing BDNF, but only in Escape mice, while reducing BDNF in Stay mice (Figs. 6D-G; Table S2). These results suggest social stress disrupts gene expression, and potentially alters BLA signaling pathways depending on an individual’s stress state. Therefore, pharmacological interventions (like acute Orx₁R antagonism) may functionally amend behavior through signaling adaptations that are phenotype dependent.

Fear conditioning responses appear to be related to specific transcriptional reorganization taking place during/after intra-BLA Orx₁R inhibition (Fig. 7). In treated animals, negative regressions exist between cued fear freezing behavior and FICRTR2 as well as MAPK3 (62) transcriptional changes (Figs. 7B, F). Without treatment (vehicle), cued freezing was positively linked to PLCB1 gene expression (Fig. 7C), an effect not observed with Orx₁R antagonism (Fig. 7D). These associations provide evidence for potential mechanistic remodeling (Fig. 9) in the BLA during periods of stress that is tied to phenotype formation and involves Orx receptor activity. This balancing act between Orx₁R and Orx₂R creates an influence over BLA microcircuits, which further defines downstream signaling dynamics, in a way that can modify stress-induced behavior (2). Since changes in HCRTR2 expression after intra-BLA Orx₁R inhibition are positively associated with MAPK3 but not BDNF transcription levels (Figs. 8A, B), it appears the adjusted bias of Orx₂R over Orx₁R activity favors ERKi signaling (Fig. 9). Amplification of ERK₁, in turn, may lead to enhanced BDNF expression (Fig. 8C) and plastic changes within BLA microcircuits (Fig. 9) (62, 63). Importantly, these findings highlight a role of intra-BLA Orx₁R in establishing pro-stress behavioral states; but exposes a receptor-driven balance that takes part in the fluid, not static, appearance of phenotype-specific behavior.

Conclusions

Modulation of BLA stress-regulatory pathways via Orx₁ receptors found predominantly on glutamatergic pyramidal neurons modifies gene expression and behavior. Modulation of pro-stress BLA microcircuits via Orx₁R inhibition reduces stress-induced behavior. In the process, Orx₁R BLA inhibition modifies gene expression of HCRTR2 which impedes pro-stress responses. Concurrently, transcription levels for downstream molecular signaling systems associated with Orx receptor signaling are also tilted toward increased ERK₁ (MAPK3), rather than PLC_β1 (PLCB1) signaling pathways, potentially altering behavior.

KEY RESOURCES TABLE

Resource Type	Specific Reagent or Resource	Source or Reference	Identifiers	Additional Information
Add additional rows as needed for each resource type	Include species and sex when applicable.	Include name of manufacturer, company, repository, individual, or research lab. Include PMID or DOI for references; use “this paper” if new.	Include catalog numbers, stock numbers, database IDs or accession numbers, and/or RRIDs. RRIDs are highly encouraged; search for RRIDs at https://scicrunch.org/resources.	Include any additional information or notes if necessary.
Chemical Compound or Drug	SB-674042 Selective Orx1R Antagonist	MedChemExpress, Monmouth Junction, NJ	Catalog Number HY-10898
Chemical Compound or Drug	MK-1064 Selective Orx2R Antagonist	MedChemExpress, Monmouth Junction, NJ	Catalog Number HY-19914
Commercial Assay Or Kit	RNAscope Fluorescent Multplex Detection Reagents	Advanced Cell Diagnostics, Newark, CA	Catalog Number 320851
Commercial Assay Or Kit	RNAscope HCRTR1 Probe	Advanced Cell Diagnostics, Newark, CA	Catalog Number 466631
Commercial Assay Or Kit	RNAscope HCRTR2 Probe	Advanced Cell Diagnostics, Newark, CA	Catalog Number 460881
Commercial Assay Or Kit	RNAscope CAMKIIα Probe	Advanced Cell Diagnostics, Newark, CA	Catalog Number 445231
Commercial Assay Or Kit	RNAscope CALB1 Probe	Advanced Cell Diagnostics, Newark, CA	Catalog Number 428431
Commercial Assay Or Kit	RNAscope GAD1 Probe	Advanced Cell Diagnostics, Newark, CA	Catalog Number 400951
Commercial Assay Or Kit	RNAscope PVALB Probe	Advanced Cell Diagnostics, Newark, CA	Catalog Number 421939
Commercial Assay Or Kit	Corticosterone ELISA Kit	Enzo Life Sciences, Farmingdale, NY	Catalog Number ADI-900-097
Commercial Assay Or Kit	One-step RT-qPCR Kit	Thermo Fisher Scientific, Waltham, MA	Catalog Number 4392653
Commercial Assay Or Kit	RT-qPCR HCRTR1 Assay	Thermo Fisher Scientific, Waltham, MA	Catalog Number 4351370, Mm01185776_m1
Commercial Assay Or Kit	RT-qPCR HCRTR2 Assay	Thermo Fisher Scientific, Waltham, MA	Catalog Number 4351370, Mm01179312_m1
Commercial Assay Or Kit	RT-qPCR PLCB1 Assay	Thermo Fisher Scientific, Waltham, MA	Catalog Number 4351370, Mm01329382_m1
Commercial Assay Or Kit	RT-qPCR MAPK1 Assay	Thermo Fisher Scientific, Waltham, MA	Catalog Number 4448892, Mm00442479_m1
Commercial Assay Or Kit	RT-qPCR MAPK3 Assay	Thermo Fisher Scientific, Waltham, MA	Catalog Number 4351370, Mm01278702_gH
Commercial Assay Or Kit	RT-qPCR BDNF Assay	Thermo Fisher Scientific, Waltham, MA	Catalog Number 4351370, Mm04230607_s1
Commercial Assay Or Kit	RT-qPCR GAPDH Assay	Thermo Fisher Scientific, Waltham, MA	Catalog Number 44533200, Mm99999915_g1
Organism/Strain	Male C57BL/6NHsd Mice	Envigo (an inotiv company), Indianapolis, IN	Item Number 4405M
Organism/Strain	Male Hsd:ICR (CD-1) Retired Breeder Mice	Envigo (an inotiv company), Indianapolis, IN	Item Number 3017M
Peptide, Recombinant Protein	Orexin A (human, rat, mouse)	Tocris, Minneapolis, MN	Catalog Number 1455
Software; Algorithm	ImageJ	doi:10.1038/nmeth.2089	RRID: SCR_003070
Software; Algorithm	ANY-maze Video Tacking Software (Version 6.0)	Stoelting Co., Wood Dale, IL	RRID: SCR_014289
Transfected Construct	AAV-U6-Orx1R-shRNA, AAV-scramble-shRNA	DiLeone Lab, Yale University, New Haven, CT, https://doi.org/10.1016/j.psyneuen.2013.10.010
Other	Stress Alternatives Model (SAM)	Summers Lab, University of South Dakota, Vermillion, SD, https://doi.org/10.3389/fnbeh.2014.00121