Deletion of TRPA1 Ion Channel Modulates the Central Stress Responses in a Mouse Model of Posttraumatic Stress Disorder

TL;DR

Deleting the TRPA1 ion channel in mice changes brain responses to stress in a PTSD model, affecting noradrenergic activity and glial activation.

Contribution

Shows TRPA1 modulates PTSD-related stress adaptation in LC and PVN, independent of α2-adrenoceptors.

Findings

TRPA1 deletion increased tyrosine hydroxylase levels in the locus coeruleus.

TRPA1 deletion reduced astrocyte activation in the paraventricular nucleus.

Clonidine effects on PTSD behaviors were not influenced by TRPA1 deletion.

Abstract

What are the main findings?
Genetic deletion of Trpa1 exaggerated TH level in the LC and reduced astrogliosis in PVN in a foot shock-induced mouse model of PTSD.Genetic deletion of Trpa1 did not influence the effects of clonidine treatment on PTSD-related behavior.

Genetic deletion of Trpa1 exaggerated TH level in the LC and reduced astrogliosis in PVN in a foot shock-induced mouse model of PTSD.

Genetic deletion of Trpa1 did not influence the effects of clonidine treatment on PTSD-related behavior.

What are the implications of the main findings?
TRPA1 ion channel may support stress adaptation in PTSD through LC and PVN.This effect is not α2-adrenoceptor-mediated.

TRPA1 ion channel may support stress adaptation in PTSD through LC and PVN.

This effect is not α2-adrenoceptor-mediated.

Background: Posttraumatic stress disorder (PTSD) is a mental illness in which central stress-regulating regions, including locus coeruleus (LC) and paraventricular nucleus of hypothalamus (PVN), play key roles. Clonidine, a central sympatholytic drug, can inhibit LC activity and reduce PTSD-related symptoms, suggesting noradrenergic involvement. Glia-driven immune mechanisms may link LC activity to PVN responses. Since TRPA1 ion channel is implicated in both neuroinflammation and stress adaptation, we aimed to determine whether its presence modulates the function of brain structures contributing to PTSD-related alteration in central stress adaptation. Methods: Foot shock PTSD model was applied to Trpa1 wild-type (WT) and knockout (KO) mice, and outcomes were assessed four weeks later. Immunohistochemistry was used to evaluate tyrosine hydroxylase (TH) levels in the LC and glial activation in the PVN. Behavioral effects of clonidine and circulating corticosterone levels were also examined. Results: Stress increased LC/TH immunoreactivity and PVN glial activation. Trpa1 deletion exaggerated LC/TH responses but reduced PVN astrocyte activation. Clonidine increased freezing and decreased jumping (a hyperarousal marker). KO mice showed enhanced jumping and did not respond to clonidine. Corticosterone levels remained unchanged. Conclusions: TRPA1 may support stress adaptation in PTSD by regulating LC noradrenergic output and PVN neuroinflammation, independently of α2-adrenergic signaling.

Figures

Full text

1. Introduction

Transient receptor potential ankyrin 1 (TRPA1) is a non-selective cation channel involved in diverse patho (physiological) processes of the nervous system including pain sensation and neurodegeneration [1,2,3]. Trpa1 is expressed in several stress sensitive brain regions, such as the piriform cortex, olfactory bulb, hypothalamus and dorsal raphe nucleus [4,5,6]. Its highest mRNA level was found in the urocortinergic neurons of the centrally projecting Edinger-Westphal nucleus (EWcp), an important hub of stress adaptation [7,8,9,10]. The downregulation of Trpa1 mRNA was observed both in the EWcp neurons of mice upon chronic stress, and in the EWcp samples of human suicide victims [11]. In addition, altered stress responses were revealed in Trpa1 gene-deficient mice using diverse stress models [11,12], further supporting the role of this ion channel in stress adaptation.

Posttraumatic stress disorder (PTSD) is a mental illness induced by traumatic events leading to disturbed stress adaptation. The symptoms involve intrusive memories, flashbacks and nightmares remembering the inducing trauma. Generalized anxiety, hyperarousal and severe depression are also commonly associated with PTSD, resulting in social deprivation and precipitating, ultimately, in problems at work and in private life [13,14].

The locus coeruleus (LC) and the paraventricular nucleus of the hypothalamus (PVN) are key brain structures of stress adaptation, and their role is also extensively studied in the pathomechanism of PTSD [15,16,17,18]. LC, the center of the sympatho-adrenomedullary (SAM) system, contains primarily noradrenergic neurons, which are the major regulators of Cannon’s “fight or flight” response together with the adrenal medulla [19,20]. LC is involved in numerous other psychophysiological functions (e.g., arousal, memory formation, pain processing, behavioral flexibility), each of them severely affected by PTSD [20,21]. LC is well positioned to affect PVN function through noradrenergic projections via α_2_-receptors [22,23]. PVN is the central component of the hypothalamic–pituitary–adrenal (HPA) axis, which is also activated upon stress, leading to the secretion of corticotropin-releasing hormone (CRH). CRH stimulates the anterior pituitary to release adrenocorticotropic hormone, which finally induces the synthesis and secretion of glucocorticoids from the adrenal cortex. Glucocorticoids such as cortisol in humans or corticosterone (CORT) in mice act on various target organs to promote adaptation to stressful situations (e.g., diabetogenic effect, increased metabolism, influence on cognitive functions) and parallelly inhibit further activation of the HPA axis [24,25]. Interestingly, the reduced HPA axis and increased SAM system activity is characteristic for PTSD [15,16], accompanied by elevated CRH levels and catecholamine mobilization, as well as decreased glucocorticoid levels [17,18]. These processes may contribute to the development of vegetative symptoms and hyperarousal, typical signs of PTSD; the lacking anti-inflammatory effect of glucocorticoids may maintain the neuroinflammation [17,18,26].

Indeed, neuroinflammation is the most widely studied process in the pathomechanism of PTSD. Acute severe or chronic stress can directly activate microglia and astrocytes in the central nervous system. These glial cells produce proinflammatory cytokines contributing to the development of neuroinflammation. These glial processes can be influenced by α-adrenergic receptors [27,28]; it is highly plausible that LC-derived noradrenaline modulates PVN glia (particularly astrocytes). On the other hand, the glia cells of the PVN modulate its neuronal output and sympathetic/endocrine function [29,30,31]. Interestingly, TRPA1 ion channels were previously found on astrocytes [32,33,34,35,36] and some of the inflammatory mediators (e.g., unsaturated fatty acids, H_2_O_2_) can activate them [2,37,38]. Additionally, the increased cytokine levels may maintain the stress response via promoting the secretion of CRH from the hypothalamus [17,18].

Treatment strategies of PTSD include cognitive psychotherapy, as well as symptomatic treatment using antidepressants for mood disorders and sympatholytic agents such as clonidine (α_2_ -adrenergic receptor agonist) for vegetative symptoms, nightmares and hyperarousal [39,40,41,42]. Since the therapeutic options described above do not provide causal therapy, it is justified to explore new molecular mechanisms that could provide promising targets.

We hypothesized that the presence of TRPA1 ion channels modulates the function of key brain structures—such as LC and PVN involved in stress adaptation—and PTSD-like behavior is regulated via this. The noradrenergic outflow plays a pivotal role in the interplay between these centers via α_2_-mediated mechanisms, and we assumed that TRPA1 may indirectly affect this interaction. To test these hypotheses (i) we investigated the regulatory impact of TRPA1 channel on the noradrenergic cells of LC and on the connected neuroinflammatory responses in the PVN using an electric foot shock model of PTSD and a Trpa1 knockout (KO) mice line; then, (ii) the subsequent question was the involvement of α_2_-adrenoceptors in the process, which was studied by clonidine treatment, behavioral tests and serum CORT outcomes.

2. Materials and Methods

2.1. Animals

The studies were carried out on 3–4 months old male Trpa1^+/+^ (WT) and Trpa1^−/−^ (KO) mice obtained from Prof. P. Geppetti (University of Florence, Italy) originally generated by Bautista et al. [43]. Animals were bred on C57BL/6J background and crossed back after 10 generations. Offspring were genotyped for Trpa1 gene by PCR (sequences of primers: ASM2: ATC ACC TAC CAG TAA GTT CAT; ASP2: AGC TGC ATG TGT GAA TTA AAT).

Involvement of females from all stages of estrous cycles and/or after ovariectomy with and without hormone replacement would dramatically increase the number of groups involved. Thus, due to ethical, financial and infrastructural considerations we used only males. Animals were kept in standard polycarbonate cages (330 × 160 × 130 mm, 5–7 mice/cage or 330 × 100 × 130 mm, 2–5 mice/cage) in a temperature (20–24 °C) and humidity (50–60%) controlled 12–12 h light–dark cycle environment (lights on at 6 a.m.) at the Department of Pharmacology and Pharmacotherapy of the University of Pécs. Ad libitum standard rodent chow (LT/n, Szinbád LLC., Gödöllő, Hungary) and tap water were provided for the mice. Besides the presence of littermates, paper rolls were placed into the cages to enrich the environment. Microbiological monitoring was performed annually (FELASA “S”). All trials were carried out during the light phase of animals between 9 a.m. and 14 a.m.

All experiments were approved by the Animal Welfare Committee of the University of Pécs and by the National Scientific Ethical Committee on Animal Experimentation in Hungary (permission No: BA02/2000-46/2024), in agreement with the directive of the European Communities Council in 1986, and with the Law of XXCIII in 1998 on Animal Care and Use in Hungary. During the experiments, every effort was made to minimize the number and suffering of the animals. The authors complied with the ARRIVE guidelines.

2.2. Experimental Design

Experiment I investigated the regulatory impact of TRPA1 ion channel on stress-related brain areas involved in the pathomechanism of PTSD (Figure 1A–C). Half of Trpa1 WT and KO mice was exposed to the foot shock protocol (fear conditioning, FC), while the other half of the two genotypes were used as non-stressed controls. After, FC animals were placed back in their original cages for 4 weeks. Then, conditioned fear test (CFT) was applied followed by brain sample collection.

Experiment II analyzed the effect of TRPA1 ion channel on the behavioral responses and serum CORT levels upon the clonidine treatment (Figure 1D,E). All mice underwent FC with foot shock. After 4 weeks resting, half of the WT and KO animals were pretreated intraperitoneally (i.p.) with clonidine (0.05 mg/kg), while the other half received i.p. saline (vehicle); a subsequent CFT was performed 30 min later followed immediately by blood sampling within 5 min. Tail blood samples were collected from a cut under brief isoflurane anesthesia and local heating into heparinized hematocrit capillaries.

2.3. Behavioral Tests

2.3.1. Fear Conditioning (FC)

The mice were put into the plastic FC chamber (internal size: 25.5 (d) × 25.5 (w) × 36 (h) cm; Ugo Basile, Gemonio, Italy) through a circular front door [44] (Figure 1A). The chamber was housed in sound-attenuating cabinets with white noise (60–70 dB) and was equipped with infrared-sensitive cameras. After a stimulus-free habituation period of 150 s, seven 1 s long foot shocks (1.4 mA) with random intervals (40–60 s) were applied together with a 10 s tone (3000 Hz, 85 dB) and increased light intensity to enhance aversiveness. The shock was delivered at the end of the tone/light and co-terminated with that in each case. After 660 s the mice were gently transferred back into their home cages. The box was cleared with tap water and 20% ethanol. Non-stressed control animals were placed into the same chamber without delivering foot shock.

2.3.2. Conditioned Fear Test (CFT)

Animals were reintroduced for 660 s into the FC chamber 4 weeks after the original stress. We used the same background and cleaning materials as those used during the FC phase. Here, none of the mice underwent foot shock; after the habituation period, only the above-mentioned tone and light stimuli were applied in each animal seven times with random intervals. Thus, the first 150 s (baseline) represented the context-dependent, while the stimulus periods showed the cue-related conditioned fear response with interstimulus intervals (ISI) between the cues (context dependency with carry-over effects). The behavior was video recorded and scored later by an experimenter blinded to the treatment groups using computer-based event-recorder software (Solomon coder; version: 17.03.22). We assessed the duration of freezing (time spent immobility for more than one second) [45,46,47] and the frequency of jumping representing an active escape attempt from the threatening situation [47,48,49,50] both associated with PTSD. We measured both the total amount of the listed behavioral parameters over the 11 min period of CFT, and we also separately analyzed the contextual and the cued fear reactions.

2.4. Perfusion and Brain Sample Collection

To exclude the effect of acute stress on the neuromorphological results, perfusion and brain sample collection were performed 24 h after CFT in Exp. I. Mice were deeply anesthetized by i.p. urethane injection (2.4 g/kg) and transcardially perfused with 20 mL ice-cold 0.1 mol/L phosphate-buffered saline (PBS, pH: 7.4). This was followed by fixation with 150 mL of 4% paraformaldehyde (PFA) solution in Millonig buffer (pH 7.4) for 15 min. Subsequently, brains were removed and collected into PFA for 72 h postfixation at 4 °C, then samples were coronally sectioned using a Leica VT1000 S vibratome (Leica Biosystems, Wetzlar, Germany). Three series of 30 μm sections were collected and stored in PBS containing sodium-azide (0.01%) at 4 °C. For long term storage at −20 °C, slices were transferred into antifreeze solution (20% ethylene glycol, 30% glycerol and 0.1 mol/L sodium-phosphate buffer). Brain sections containing the LC (from Bregma −5.34 mm to −5.80 mm) or PVN (from Bregma −0.58 mm to −1.22 mm) were used for histological evaluation [51].

2.5. Immunohistochemistry

2.5.1. Tyrosine Hydroxylase (TH) Immunohistochemistry in Locus Coeruleus (LC)

Sections were first washed for 2 × 15 min with PBS and treated with citrate solution at 90 °C for 10 min. After cooling and washing the samples, 0.5% Triton X-100 (Sigma Chemical, Zwijndrecht, The Netherlands) in PBS was applied for 30 min followed by blocking of non-specific binding sites with 2% normal donkey serum (NDS) in PBS. Then, sections were incubated overnight at room temperature (RT) with a solution containing polyclonal rabbit anti-TH antibody (Abcam, Cambridge, UK, Cat. No.: ab6211) diluted to 1:4000 in PBS with 2% NDS. After 2 × 15 min washes in PBS, sections were incubated with Cy3-conjugated donkey anti-rabbit secondary antibody (Jackson, West Grove, PA, USA, Cat. No: 711-165-152) in 1:500 PBS with 2% NDS for 3 h at RT. After washing the samples, sections were mounted on gelatine-coated glass slides, then covered with glycerol-PBS (1:1) and stored at −20 °C until confocal microscopy. All samples were stained in one session.

2.5.2. Immunohistochemistry on Glial Markers in the Paraventricular Nucleus of the Hypothalamus (PVN)

Glia cells were visualized by double immunostaining using markers specific for astrocytes (glial fibrillary acidic protein; GFAP) and for microglia (ionized calcium binding adaptor molecule 1, IBA1). The pretreatment of sections and the non-specific binding site blocking was carried out as described in Section 2.5.1. Then, monoclonal mouse anti-GFAP (Novocastra, Newcastle upon Tyne, UK, Cat. No.: NCL-GFAP-GA5) and polyclonal rabbit anti-IBA1 (Fuji-Wako, Osaka, Japan, Cat. No.: 019-19741) primary antibodies both diluted to 1:2000 with 2% NDS were used overnight at RT. After washing, Cy3-conjugated donkey anti-mouse (Jackson, Cat. No.: 715-165-150) and Alexa 488-conjugated donkey anti-rabbit (Jackson, Cat. No.: 711-005-152) secondary antibodies both in 1:500 dilution with 2% NDS were applied for 3 h at RT. Sections were mounted, covered and stored according to the protocol mentioned above. All samples were stained in one session.

2.6. Image Analysis

Olympus Fluoview FV-1000 laser scanning confocal microscope (Olympus Europa, Hamburg, Germany) and FluoView FV image acquisition software system (Olympus Europa, Hamburg, Germany) were used for imaging. Digital images were acquired by sequential scanning in analog mode for the corresponding fluorophores to avoid false positive signal resulting from the slight overlap of emission spectra and to reliably quantify the fluorescent signal. The confocal aperture was set to 80 μm. The analog sequential scanning was performed using a 40× objective lens (NA: 0.75). An optical thickness of 3.5 µm was indicated by the software and the resolution was set to 1024 × 1024 pixel. The excitation time was set to 4 µs per pixel. Alexa Fluor 488 was excited at 488 nm and Cy3 at 543 nm. We assigned the virtual color green for Alexa Fluor 488 and red for Cy3. The same microscope settings were used for imaging on all samples deriving from a particular experiment.

The morphometry of LC was performed on non-edited pictures using ImageJ software (version 1.52a) by an experimenter who was blind to the experimental conditions. The intensity of TH-immunofluorescence was measured in all cells visible in the 3.5 µm thick cross-sectional images if the plane of sectioning contained the cell nucleus, (ca. 10–30 cell bodies per LC samples). The region of interest was marked out manually including only the cytoplasmic areas of neurons. The specific signal density (SSD) was determined in arbitrary units (a.u.) and corrected for the background signal. The average of SSDs was calculated in each section. The calculation was performed in four sections per animal, and the average of these four values represented the SSD of one mouse, which was finally subjected to the statistical assessment.

In case of glia activation, we determined the number and the activation score of astrocytes and microglia in the total area of PVN cross-sections with a thickness of 3.5 µm using GFAP and IBA1 immunostaining. We analyzed 60–120 cells per section for astrocytes and 10–20 cells per section for microglia and 4 slices per animal. The degree of glia activation in each counted cell was ranked from 1 (resting) to 5 (severe reactive gliosis) [52]. The main analyzed parameters were the form and enlargement of cell bodies and the complexity and thickness of processes. The glia cells were analyzed on four sections per animal, and the average of the individual scores of each glia cells on the four sections represented the activation score of the mouse.

2.7. Corticosterone Level by Radioimmunoassay

Blood samples were centrifuged at 3000 rpm for 5 min, then plasma was stored at −20 °C until further analysis. Corticosterone levels were analyzed from 10 μL sample using a specific antibody developed by the Institute of Experimental Medicine (Budapest, Hungary) [53]. The intra-assay coefficients of variation were 7.5% and 4.7%, respectively. All samples from a particular experiment were analyzed in one session.

2.8. Statistical Analysis

Data were analyzed by two-way analysis of variance (ANOVA, variables: stress and genotype in Experiment I, as well as treatment and genotype in Experiment II) and repeated measures ANOVA (within-subjects factor: cue presentation), both followed by Tukey’s post hoc test (StatSoft 13.5.0, Tulsa, OK, USA). All datasets showed normal distribution (Kolmogorov–Smirnov) and homogenous variance (Levene). Pearson’s correlation was also conducted. p < 0.05 was considered statistically significant. Data are presented as mean ± SEM. The results of post hoc test are represented in the Figures.

3. Results

3.1. The Genetic Lack of TRPA1 Does Not Influence Stress-Induced Freezing but Enhance Jumping Behavior

Investigating the behavioral responses during the FC phase, increased freezing was detected in both stressed groups compared to the controls (F_stress_ (1,26) = 1774.46, p < 0.0001; both p_WT_ and p_KO_ = 0.0002), but neither the main effect of the genotype (F_genotype_ (1,26) = 1.84, p = 0.19, p_control_ = 0.75; p_shock_ = 0.79) nor an interaction was revealed (F_interaction_ (1,26) = 0.01, p = 0.92). Similarly, shock-induced jumping behavior in both groups (F_stress_ (1,26) = 842.27, p < 0.0001; both p_WT_ and p_KO_ = 0.0002) without genotype differences (F_genotype_ (1,26) = 0.04, p = 0.85, p_control_ = 1.00; p_shock_ = 0.99; F_interaction_ (1,26) = 0.04, p = 0.85).

When we reintroduced the animals to the stress-related environment four weeks later, they still showed exaggerated total freezing time (F_stress_ (1,26) = 914.54, p < 0.0001; both p_WT_ and p_KO_ = 0.0002) independently from genotype (F_genotype_ (1,26) = 0.12, p = 0.73, p_control_ = 0.94; p_shock_ = 0.69; F_interaction_ (1,26) = 1.35, p = 0.26) (Figure 2A). Further analysis confirmed the elevation of both contextual (including the baseline and the ISIs) and cued freezing behavior in stressed groups compared to the controls (F_stress_ (1,26) = 663.12, p < 0.0001; in each comparison both p_WT_ and p_KO_ < 0.0001) without the main effect of the genotype (F_genotype_ (1,26) = 0.02, p = 0.89; F_stressgenotype_ (1,26) = 0.69, p = 0.41). Cue presentation decreased the freezing behavior of both stressed groups compared to the baseline and to the ISI values over the entire experiment (F_cue presentation_ (14,364) = 3.50, p < 0.0001, F_stresscue presentation_ (14,364) = 6.55, p < 0.0001), whose effect was not influenced by the genotype (F_genotypecue presentation_ (14,364) = 1.21, p = 0.26; F_stressgenotype*cue presentation_ (14,364) = 0.98, p = 0.48) (Figure 2B).

However, the total amount of jumping behavior (as a sign of hyperarousal) was not only influenced by previous stress (F_stress_ (1,26) = 31.11, p < 0.0001; p_WT_ = 0.09, while p_KO_ = 0.0002), but also by the genotype (F_genotype_ (1,26) = 4.53, p = 0.04; F_interaction_ (1,26) = 4.53, p = 0.04). More precisely, shocked Trpa1 KO mice jumped more than their WT counterparts (p_control_ = 1.00; p_shock_ = 0.02) (Figure 2C). Interestingly, shocked animals showed a higher jumping rate compared to the controls only until the end of the second cue (F_stress_ (1,26) = 28.86, p < 0.0001; F_cue presentation_ (14,364) = 6.45, p < 0.0001; during Cue1: p_WT_ < 0.0001 and p_KO_ = 0.004, while during the baseline and Cue2 p_KO_ < 0.0001), followed by a rapid decline (F_stresscue presentation_ (14,364) = 6.45, p < 0.0001). Here, the stress-induced jumping response was also influenced by the genotype (F_genotype_ (1,26) = 4.26, p = 0.049), as shocked KOs showed a higher jumping rate during the baseline (p = 0.02) and the second cue presentation (p = 0.002) compared to the WT counterparts (F_stressgenotype_ (1,26) = 4.67, p = 0.04), with a tendency for interaction between the genotype and cue presentation as well (F_genotypecue presentation_ (14,364) = 1.71, p = 0.051; F_stressgenotype*cue presentation_ (14,364) = 1.71, p = 0.051) (Figure 2D).

Additionally, Pearson’s test revealed that freezing positively correlated with jumping behavior (r = +0.678, p < 0.001).

3.2. The Genetic Lack of TRPA1 Is Associated with a Greater Increase in LC/TH Immunoreactivity After Foot Shock

Stress increased the TH immunopositivity in the LC of both genotypes (F_stress_ (1,26) = 40.07, p < 0.0001; p_WT_ = 0.04, p_KO_ = 0.0002) (Figure 3). Although the main genotype effect was also significant (F_genotype_ (1,26) = 13.13, p = 0.0012, p_control_ = 0.82; p_shock_ = 0.0011), but this genotype difference became apparent only after foot shock and was not detected among control, non-stressed conditions (F_interaction_ (1,26) = 5.50, p = 0.03).

3.3. The Lack of TRPA1 Diminished the Astrocyte Activation in the PVN Without Interfering with the Microglia Response in the Mouse Model of PTSD

Stress increased the number of astrocytes in the PVN (F_stress_ (1,26) = 4.50, p = 0.04) (Figure 4A,C). However, this effect was not visible during post hoc comparisons (p_WT_ = 0.41 and p_KO_ = 0.49) and was independent from the genotype (F_genotype_ (1,26) = 0.01, p = 0.92; p_control_ = 1.00 and p_shock_ = 1.00; F_interaction_ (1,26) = 0.01, p = 0.92).

In addition to the number, the activation of astrocytes was also increased in the PVN of shocked mice (F_stress_ (1,26) = 17.86, p = 0.0003), which was mainly due to an elevation in the WTs (p_WT_ = 0.0005 and p_KO_ = 0.60) (Figure 4B,C). The significant main genotype effect (F_genotype_ (1,26) = 8.49, p = 0.007) was apparent after stress, but not in control, non-shocked individuals (F_interaction_ (1,26) = 5.99, p = 0.02; p_control_ = 0.99, p_shock_ = 0.004).

In contrast to the astrocytes, we did not find any changes in the number (F_stress_ (1,26) = 3.46, p = 0.07, p_WT_ = 0.96 and p_KO_ = 0.16; F_genotype_ (1,26) = 1.75, p = 0.20; p_control_ = 0.31 and p_shock_ = 1.00; F_interaction_ (1,26) = 1.41, p = 0.25) (Figure 5A,C) or scoring (F_stress_ (1,26) = 0.04, p = 0.85, p_WT_ = 0.86 and p_KO_ = 0.96; F_genotype_ (1,26) = 0.54, p = 0.47, p_control_ = 1.00 and p_shock_ = 0.66; F_interaction_ (1,26) = 0.83, p = 0.37) (Figure 5B,C) of the microglial cells in the PVN.

3.4. Correlation Between Behavioral and Morphological Measures

There was a significant positive correlation between LC/TH immunopositivity and both observed behaviors (freezing: r = +0.716, p < 0.001; jumping: r = +0.518, p = 0.003), suggesting the involvement of LC in the development of PTSD-like behavior. However, we did not detect any significant correlation between LC and PVN measures.

3.5. Lack of TRPA1 Did Not Modulate the Effect of Clonidine Treatment on the Behavioral Responses and Serum CORT Levels in a PTSD Model

Clonidine administration significantly increased the total duration of freezing (F_treatment_ (1,36) = 12.95, p = 0.001), which was apparent in WT (p_WT_ = 0.045), but not in KO mice (p_KO_ = 0.11). However, major genotype difference was not detected (F_genotype_ (1,36) = 2.68, p = 0.11; p_vehicle_ = 0.82, p_clonidine_ = 0.47; F_interaction_ (1,36) = 0.14, p = 0.72) (Figure 6A). Further analysis of contextual and cued freezing did not reveal significant differences between the vehicle- and clonidine-treated animals (F_treatment_ (1,36) = 2.38, p = 0.13) or between the two genotypes (F_genotype_ (1,36) = 3.98, p = 0.054; F_treatmentgenotype_ (1,36) = 1.10, p = 0.30). Though cue presentation decreased the freezing response in each group compared to the baseline and to the ISI values almost throughout the entire experiment (F_cue presentation_ (14,504) = 18.64, p < 0.0001), but the treatment or the genotype did not influence this effect (F_treatmentcue presentation_ (14,504) = 0.92, p = 0.54; F_genotypecue presentation_ (14,504) = 1.22, p = 0.25; F_treatmentgenotype*cue presentation_ (14,504) = 0.93, p = 0.52). (Figure 6B).

By contrast, we found considerable genotype differences in the total amount of jumping (F_genotype_ (1,34) = 12.66, p = 0.001) both in the vehicle- (p_vehicle_ = 0.02) and clonidine-treated groups (p_clonidine_ = 0.02). Clonidine administration decreased the jumping rate (F_treatment_ (1,34) = 5.32, p = 0.03) without significant post hoc differences (p_WT_ = 0.14 and p_KO_ = 0.09) or influencing the genotype effect (F_interaction_ (1,34) = 0.02, p = 0.90) (Figure 6C). The cue presentation increased the jumping response (F_cue presentation_ (14,504) = 18.32, p < 0.0001), whose effect was influenced by the genotype (F_genotype_ (1,36) = 7.25, p = 0.011; F_genotypecue presentation_ (14,504) = 3.12, p < 0.001). More precisely, cue-induced changes in the jumping were observable only in the KOs over the entire experiment, leading to a significant difference between vehicle-treated WT and KO mice during the fourth Cue presentation (p_vehicle_ = 0.049). Though a tendency for reduced jumping was observable in the clonidine-treated mice compared to the vehicle groups (F_treatment_ (1,36) = 3.52, p = 0.07), no significant differences were revealed between the treatment groups, and this effect was not influenced by the genotype (F_treatmentgenotype_ (1,36) = 0.14, p = 0.71) or the cue presentation (F_treatmentcue presentation_ (14,504) = 0.72, p = 0.76; F_treatmentgenotype*cue presentation_ (14,504) = 1.11, p = 0.35) (Figure 6D).

Plasma CORT values at the end of CFT did not reveal any group differences (F_treatment_ (1,39) = 0.20, p = 0.66; p_WT_ = 0.98 p_KO_ = 0.73; F_genotype_ (1,39) = 0.13, p = 0.72; p_vehicle_ = 0.97 p_clonidine_ = 0.75; F_interaction_ (1,39) = 1.02, p = 0.32) (Figure 6E).

4. Discussion

Behavioral analysis revealed that foot shock increased freezing and jumping during the FC phase. However, there were no genotype differences, suggesting that the subsequent changes might not be due to altered sensation of the KO animals. In the CFT, the Trpa1 deletion had no effect on shock-induced freezing but exaggerated jumping behavior, as a sign of hyperarousal. These effects were reproducible. The behavioral changes correlated positively with LC/TH immunopositivity. The shock-induced aggravation in LC/TH levels was accompanied by astrocyte activation in the PVN, without microglial activation. The Trpa1 deletion enhanced shock-induced LC/TH elevation and reduced the astrocyte activation. Despite the previously reported positive impact on some PTSD symptoms of clonidine, in our study, manipulating the α2-adrenergic system aggravated the freezing—one of the most frequently used outcome measures of PTSD [45]. However, clonidine had a positive, protective effect on jumping behavior, which does not seem to be transmitted via the HPA axis.

The rate limiting step of noradrenaline biosynthesis is mediated by TH. Thus, the regulation of this enzyme in the LC is an important mechanism upon stress adaptation [54,55]. In line with our findings, elevated TH expression both at the mRNA and protein level was reported upon both single prolonged stress (SPS) and electric foot shock models of PTSD [55,56,57]. These data indicate the involvement of the hyperactive LC-noradrenergic system in the manifestation of PTSD. Importantly, the literature implies that noradrenergic signaling of the LC is deeply implicated in the auditory cued fear conditioning [40,58,59]. The regulatory role of the LC-noradrenergic system in fear memory consolidation depends on the prevailing level of arousal and stress: among high stress conditions its activation facilitates the consolidation and maintenance of cued fear memory, while under low stress levels it promotes fear extinction [40,60,61]. We assume that the role of TRPA1 ion channels become important under high stress conditions (i.e., after foot shock), where it moderates the TH enzyme expression in LC (Figure 3). Therefore, its genetic deletion can contribute to enhanced cue-induced jumping behavior observable in stressed KO mice (Figure 2 and Figure 6) probably via strengthening the noradrenergic outflow from LC as confirmed by positive correlation between jumping and LC/TH. Indeed, the connection between jumping and the noradrenergic system is supported by the diminishing effect of clonidine administration as well (Figure 6). Jumping might be a sign of hyperarousal [47,48,49,50,62], and clonidine is often used for the treatment of this symptom [63].

The observed genotype difference in LC/TH is contradictory in light of the fact that we did not find Trpa1 mRNA in the LC using a highly sensitive RNAscope technique [64]. It is unclear by what mechanism the TRPA1 ion channel would modulate the function of noradrenergic system. One possible explanation is the regulation of LC-input areas. It is well known that the PVN is reciprocally connected with the LC providing positive feedback loops between CRH and noradrenaline [21,40,55,61,65,66,67,68]. TRPA1 might contribute to neuroinflammation [69], especially though modulation of the astrocytes [70,71,72] that are also supported by the lower astrocyte activation in our KO mice (Figure 4). As astrocytes might have anti-inflammatory potential [73,74] and resolve neuroinflammation in the delayed phases of chronic stress [75,76,77], we might assume that their lower activation in the PVN of KO mice can increase local neuroinflammation. This neuroinflammation may elevate the CRH release from the PVN [17,18,25], which may stimulate LC/TH synthesis in KO. In support, Xie et al. reported that astrocyte activation of the hippocampus may reduce the fear responses in PTSD [78]. A further possible mechanism might be mediated through the EWcp. Indeed, we previously demonstrated the high abundance of Trpa1 mRNA in this nucleus, as well as the role of EWcp in modulating the stress response (see Introduction). We might assume that the lacking Trpa1 from the EWcp alter the interpretation of the foot shock, which changes the reactivity of the whole brain, including the LC. This is in line with the James-Lange theory of emotions [79], which states that the reaction of the whole body shapes our emotions, which may manifest itself in divergent activity of different brain areas.

In line with the reciprocal connection, LC may regulate PVN function via neuroinflammation (see Introduction). The previously described cFOS activation in PVN neurons following foot shock suggested its integrative role in FC [80,81,82]. Not only was the role of its neuropeptides (e.g., CRH, oxytocin) described among a wide range of stress conditions [10,83,84,85,86,87,88,89], but there are also studies supporting the role of their glia cells in acute and chronic stress situations [85,88,89,90,91,92,93], with limited data in conditioned fear and PTSD [81]. To our best knowledge, our current study is the first describing the activation of astrocytes in the PVN following foot shock.

The literature data suggest that astrocytes express TRPA1 ion channels (see earlier), and the mediators of oxidative stress produced excessively in PTSD [17,94,95,96,97,98,99] may stimulate TRPA1 [2,37,38] leading to astrocyte activation. Thus, we assume that the TRPA1-mediated astrocyte activation in the PVN of WT mice may play a role in the restoration of neuroendocrine and behavioral functions following stress, while the lacking TRPA1 in KOs accompanied by lacking astrocyte activation may delay recovery. Indeed, our prior study revealed both the exhaustion of CORT secretion and the impaired fear extinction with prolonged fear responses in cases of Trpa1 gene-deficiency [100].

In contrast, microglia of the PVN did not seem to be involved in the presently observed PTSD-related alterations. Interestingly, a prior study reported microglial hyper-ramification and increased soma perimeter in the PVN following contextual FC [81]. Since microglia responses depend on the modality of stressor and the genetic background of animals [81,101], this contradiction may be explained by the different experimental protocols and the diverse strain of mice. However, our previous experiments showed alterations in the microglial morphology of Trpa1 KO mice within several other brain regions implicated in the PTSD pathomechanism (central nucleus of amygdala, hippocampus, medial prefrontal cortex) suggesting the role of this ion channel in microglia-mediated processes upon conditioned fear [100]. Indeed, similar region-specific differences in stress-induced microglia responses have already been reported [102].

Our prior research revealed significantly decreased shock-induced immobility in KO mice compared to the WTs using the SPS protocol of PTSD [12]. One possible explanation is the different stress-level of the two models: the foot shock paradigm is a strong unescapable psychological stressor [45], while the SPS model is based upon mild stressors over a longer period [103,104,105]. Therefore, we suspect that TRPA1 might be important in fine-tuning the behavior to the intensity of stress. The increased jumping rate of KO animals in the foot shock model was found before [100], suggesting the involvement of TRPA1 channel in hyperarousal. However, this effect is not exclusively mediated by the adrenergic system (see ineffectiveness of clonidine). The literature indicates that clonidine may decrease the locomotor activity through presynaptic α_2_-receptor agonism among others in the LC, which may induce anxiolytic and sedative effects by the reduced central noradrenergic outflow [106,107,108,109,110]. Thus, we suspect that the elevated freezing behavior of clonidine-treated mice represents a decreased locomotion rather than a PTSD-like freezing behavior. This is supported by its jumping reducing effect as well (see reciprocal changes in cued freezing and jumping response, as well as negative correlation between freezing and jumping during Exp II; r = −0.580, p = 0.014).

Though the effect of clonidine on the function of HPA axis is debated depending on the dose and route of administration [111,112,113], a recent study showed that i.p. clonidine treatment may increase the CORT secretion [109]. However, in our case we could not find treatment or genotype difference, which let us to conclude that clonidine-mediated effects may be independent of the HPA axis activation. However, we cannot rule out confounding technical details such as the effect of i.p. injection approx. 40 min before measurement, as well as isoflurane anesthesia during sampling.

This study has several limitations. The selected stress paradigm may constrain translational relevance, while reliance on a single stressor reduces generalizability. Developmental compensation in knockout animals cannot be ruled out. Immunostaining analyses were semiquantitative, and only a limited subset of network components was assessed at discrete time points. Moreover, the observed associations are correlational; functional manipulations are needed to establish causality.

5. Conclusions

TRPA1 may be involved in stress adaptation during PTSD-related conditioned, especially in cued, fear responses via controlling noradrenergic output from the LC and the neuroinflammation in stress-sensitive brain areas like PVN, most probably not through the α_2_-adrenergic receptors. We suppose that TRPA1 activation may contribute to the recovery following strong stresses and, therefore, may be a novel drug target in the future therapy of PTSD.