Abstract
Depression is a frequent and serious illness, and stress is considered the main risk factor for its onset. First-line antidepressants increase serotonin (5-hydroxytryptamine; 5-HT) levels in the brain. We previously reported that an N-acetyltransferase, Shati/Nat8l, is upregulated in the dorsal striatum (dSTR) of stress-susceptible mice exposed to repeated social defeat stress (RSDS) and that dSTR Shati/Nat8l overexpression in mice (dSTR-Shati OE) induces stress vulnerability and local reduction in 5-HT content. Male mice were used in this study, and we found that dSTR 5-HT content decreased in stress-susceptible but not in resilient mice. Moreover, vulnerability to stress in dSTR-Shati OE mice was suppressed by the activation of serotonergic neurons projecting from the dorsal raphe nucleus (dRN) to the dSTR, followed by upregulation of 5-HT content in the dSTR using designer receptors exclusively activated by designer drugs (DREADD). We evaluated the role of GABA in modulating the serotonergic system in the dRN. Stress-susceptible after RSDS and dSTR-Shati OE mice exhibited an increase in dRN GABA content. Furthermore, dRN GABA content was correlated with stress sensitivity. We found that the blockade of GABA signaling in the dRN suppressed stress susceptibility in dSTR-Shati OE mice. In conclusion, we propose that dSTR 5-HT and dRN GABA, controlled by striatal Shati/Nat8l via the dSTR-dRN neuronal circuitry, critically regulate stress sensitivity. Our study provides insights into the neural processes that underlie stress and suggests that dSTR Shati/Nat8l could be a novel therapeutic target for drugs against depression, allowing direct control of the dRN serotonergic system.
Significance Statement
Given that 30% of depression patients have resistant to conventional antidepressants, finding novel therapeutic strategies for its disease is required. We previously demonstrated that the overexpression of Shati/Nat8l, N-acetyltransferase, in the dorsal striatum (dSTR) of mice induces stress vulnerability. dSTR 5-HT (5-hydroxytryptamine) levels are downregulated in stress-susceptible, nonresilient mice exposed to repeated social defeat stress (RSDS). Stress vulnerability in dSTR Shati/Nat8l overexpression mice was suppressed by the activation of serotonergic neurons projecting from the dorsal raphe nucleus (dRN) to the dSTR. We discovered that dRN GABA content correlated with stress sensitivity and inhibited GABA signaling in dRN-induced stress resilience. We suggest that novel bidirectional dSTR-dRN circuits determine the stress sensitivity underlying depression pathology.
Introduction
Mood disorders, including bipolar disorder and depressive disorders (depression), are frequent and serious illnesses. Between 1990 and 2019, the number of patients with depression increased by 64.4% (GBD 2019 Mental Disorders Collaborators, 2022). Its characteristics include a high prevalence, a lifetime prevalence of 16%, and resistance to treatment (Kessler et al., 2003; Mata et al., 2015; Pandarakalam, 2018). One in ten people has a decreased quality of life because of a wide array of depressive symptoms, such as emotional suffering, cognitive dysfunction, and social impairment (Hauenstein, 2003). Moreover, ∼30–50% of patients have forms of treatment resistance to conventional antidepressant drugs (Dudek et al., 2021). Therefore, novel therapeutic strategies and targets for drugs to treat depression are required. However, numerous unclear aspects of depression pathogenesis have prevented progress in this field, particularly in providing effective therapies for refractory forms of the disease.
Stress is linked to depressive pathology, and some studies have indicated that stressful life events have a substantial causal relationship with depression onset (Kendler et al., 1999). However, this is not always the case because some individuals are resilient (Fleshner et al., 2011). Mice can also be categorized into two groups: stress-susceptible mice, showing depression-like behaviors, and stress-resilient mice, showing no behavioral changes despite being exposed to the same stress conditions (Golden et al., 2011). Controlling stress resilience could lead to the development of new antidepressants. However, the regulatory mechanisms of stress sensitivity have not been elucidated.
An imbalance in serotonin (5-hydroxytryptamine; 5-HT) levels is associated with the pathogenesis of depression (Yuan et al., 2015; Dell'Osso et al., 2016). Over the last 50 years, 5-HT deficits have been targeted in the treatment of depression, and selective serotonin reuptake inhibitors (SSRIs) have been used as first-line antidepressants (Hofmann et al., 2017). Stress also influences 5-HT neurotransmission (Graeff et al., 1996; Hale et al., 2012). Alterations in brain 5-HT, 5-HT-related molecular content (Kang et al., 2005), and serotonergic activity (Paul et al., 2011; J. Zhang et al., 2012), have been observed in individuals exposed to acute or chronic stress. In behavioral experiments, mice with hereditary central 5-HT deficiency are more susceptible to social stress (Sachs et al., 2015), and increasing 5-HT levels with SSRIs prevents the impairment of stress adaptation (Uno et al., 2019). In addition, 5-HT is predominantly synthesized in the raphe nucleus by tryptophan hydroxylase (TPH; Ishimura et al., 1988). In the raphe nucleus of both rodents and humans, almost all serotonergic cell bodies are located in the dorsal region (dRN; Ishimura et al., 1988; Hornung, 2003).
We previously reported that dorsal striatum (dSTR)-specific overexpression of N-acetyltransferase Shati/Nat8l in mice induces a decrease in dSTR 5-HT levels (Miyamoto et al., 2017). Although 50% of control mice exposed to repeated social defeat stress (RSDS) showed a depressive phenotype, almost all dSTR Shati/Nat8l overexpression in mice (dSTR-Shati OE) showed depression-like behaviors. Furthermore, we demonstrated that Shati/Nat8l overexpression mice were vulnerable to subthreshold social stress (Uno et al., 2019). Conversely, the knock-down of striatal Shati/Nat8l induces stress resilience (Miyanishi et al., 2021). Furthermore, we suggest that Shati/Nat8l could be used as a biomarker for diagnosing depression in a clinical study (Miyanishi et al., 2020). Shati/Nat8l was previously isolated from the brains of animals with induced psychosis (Niwa et al., 2007), and exhibited N-acetyl transfer activity, catalyzing the synthesis of N-acetylaspartate (NAA) from acetyl-CoA and aspartate (Ariyannur et al., 2010). Although the expression of Shati/Nat8l in the dSTR may determine sensitivity to stress, the detailed neural mechanisms underlying this effect have not been demonstrated. Considering the suggestion that Shati/Nat8l in the dSTR contributes to stress sensitivity via the serotonergic system, we focused on the involvement of the dRN.
Materials and Methods
Animals
Male C57BL/6J (eight-week-old) and male ICR (four to five months of age) were purchased from Nihon SLC. All mice were housed at constant temperatures (25 ± 1°C) and humidity (50 ± 5%), on a 12/12 h light/dark cycle (lights were turned on at 7 A.M. and turned off at 7 P.M.), with free access to pellets and water. The experimental procedures were approved by the Committee for Animal Experiments of the University of Toyama (approval no. 2021PHA-20) and performed following the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals.
Preparation of virus vectors and microinjection
Adeno-associated virus (AAV) vectors were prepared as described previously (Krzyzosiak et al., 2010; Iida et al., 2013). The AAV-CMV-Shati/Nat8l or Mock vector (AAV2/9) contained the cytomegalovirus (CMV) promoter and cDNA encoding either the 3–6× His-tagged Shati/Nat8l or EGFP sequence, respectively. AAV-CMV-Shati/Nat8l or Mock vectors were microinjected into the dSTR, and AAV-CMV-mock vectors were used as a control for AAV-CMV-Shati/Nat8l. The AAV-pet1-hM3Dq and mock vectors (AAV2/rh10) were provided by Mieda (Kanazawa University, Kanazawa, Ishikawa, Japan; Hasegawa et al., 2014). The AAV-pet1-hM3Dq and mock vectors contained the pet1 promoter and cDNA encoding either the hM3Dq or ChR2::EYFP sequence, respectively. AAV-pet1-hM3Dq or Mock vector was microinjected into the dRN, and AAV-pet1-Mock vectors were used as controls for AAV-pet1-hM3Dq. The titers of the recombinant AAV vectors were as follows: AAV-pet1-hM3Dq, 9.8 × 1011; AAV-pet1-Mock (ChR2::EYFP), 5.2 × 1012; AAV-CMV- Shati/Nat8l, 1.0 × 1010; AAV-CMV-Mock (EGFP), 1.0 × 1010 genome copies/ml.
Bilateral microinjection of AAV-CMV-Shati/Nat8l or Mock vector into the dSTR [anteroposterior (AP), 0.5 mm; mediolateral (ML), ±2.0 mm; dorsoventral (DV), 3.0 mm] and of AAV-pet1-hM3Dq or Mock vector into the dRN (AP, −4.4 mm; ML, 0 mm; DV, 3.5 mm) were performed using stereotaxic frame (SR‐5M) following a reference image (Paxinos and Franklin, 2008). This study was approved by the Board of Safety Committee for Recombinant DNA Experiments of the University of Toyama (approval no. G2020PHA‐5).
Immunostaining
Double immunostaining was performed as described previously (Ibi et al., 2008). Brain tissue fixed with 4% paraformaldehyde (PFA) was sectioned into 30-μm slices using a cryostat (Leica). After slices were permeabilized with 0.25% Triton X‐100 and blocked with 10% goat serum, incubation with primary antibodies against mouse neuronal nuclear antigen (NeuN; 1:500, MBL), rabbit His tag (1:500, MBL), mouse tryptophan hydroxylase (TPH; 1:200, Abcom) and rabbit GFP (1:1000, Sigma-Aldrich) was performed overnight. The slices were washed with Tris-buffered saline containing Tween 20 and incubated with CFTM 594 goat anti-mouse immunoglobulin G (IgG; H+L; 1:1000; Biotium) and CFTM 488 goat anti-rabbit IgG (H+L; 1:1000, Biotium) as secondary antibodies for 2 h. After washing and mounting the sections, immunofluorescence was performed using an AxioCam ICc1 (Carl Zeiss).
Real-time RT-PCR analysis
Whole brains were placed in a mouse brain matrix (Brain Science Idea) to obtain tissue sections. Total RNA was extracted from these tissues and converted into cDNA using the Prime Script RT Reagent kit (Takara). The mRNA levels were quantified using a Thermal Cycler Dice Real-Time System (Takara) and Thunderbird Syber qPCR Mix (Toyobo). In addition, 36B4 was used as the housekeeping gene. Primer sequences for Shati/Nat8l and 36B4 mRNA were as follows (Haddar et al., 2020):
Shati/Nat8l:
forward, 5′-GTGATTCTGGCCTACCTGGA-3′;
reverse, 5′-CCACTGTGTTGTCCTCCTCA-3′;
36B4:
forward, 5′-ACCCTGAAGTGCTCGACATC-3′;
reverse, 5′-AGGAAGGCCTTGACCTTTTC-3′.
Repeated social defeat stress
ICR mice were used as stressors. Their aggressive behavior was confirmed by screening before the experiments (Golden et al., 2011). C57BL/6J mice were subjected to a 10-min daily physical attack by unfamiliar ICR mice for 10 consecutive days. ICR and C57BL/6J mice were housed in cages separated by a transparent plastic divider, allowing auditory and visual contact for 24 h. The plastic divider was removed during exposure to a physical attack.
Microdefeat stress
Microdefeat stress was induced as previously described (Golden et al., 2011). On a single day, C57BL/6J mice were subjected to three 5-min physical attacks by ICR mice, with 15 min of rest between each session.
Behavioral tests
Behavioral tests were performed 24 h after the final defeat session of the RSDS or microdefeat protocol. All tested mice (defeated mice and stress-naive control mice) were subjected to behavioral experiments during light exposure.
Social interaction test
The social interaction test was conducted in a plastic open box (40 × 40 × 30 cm) equipped with a mesh cage for the ICR mice (targets). The interaction zone (IZ) was defined as the area surrounding the mesh cage (14 × 24 cm). Avoidance zones (AZs) were defined as the corner areas on opposing sides of the mesh cage (9 × 9 cm). The testing protocol consisted of two 150-s sessions (pre-test or post-test). In the pre-test, the test mice were placed in the center of an open box with no target in the mesh cage, and their activity was recorded for 150 s. After an interval of 30 s, the test mice were returned to the open box, while the target mice were placed in the mesh cage. Their approach or avoidance activity toward the targets was recorded again for 150 s as a post-test. Social interaction ability was assessed by measuring the time spent in the IZ or AZ, and the social interaction ratio (IR). IR was calculated as follows: (time in the IZ in the post-test)/(time in the AZ in the pre-test). Mice with IR < 1.0 were classified as susceptible, while the other mice were classified as resilient (Golden et al., 2011).
Sucrose preference test
Two 15-ml water bottles were placed in the cages for 24 h for habituation. Next, one water bottle was replaced with a 1% sucrose bottle. The amounts of water and 1% sucrose consumed over 12 h were measured. Sucrose preference was calculated as (sucrose consumption)/(total water and sucrose consumption) × 100 (Golden et al., 2011).
Locomotor activity test
Test mice were placed at the center of an open-field area (40 × 40 × 30 cm) and allowed to explore freely. Locomotor activity was measured as “counts” passing a set of infrared beams in the SCANET MV-40 (MELQUEST) for 60 min (Ma et al., 2017).
Tail suspension test
The tested mice were suspended from a suspension bar (12 cm in height) with their tails using adhesive tape for 6 min. Their movements were monitored, and immobility time was measured from 1–6 min (Machado et al., 2008).
Forced swimming test
The tested mice were placed in a transparent cylinder (21 cm in diameter × 22.5 cm in height) filled with water (23 ± 1°C, 18 cm in depth), and forced to swim for 6 min. Their movements were recorded using a SCANET MV-40 (MELQUEST), and the immobility time was measured from 1 to 6 min (K. Zhang et al., 2018).
Microinfusions
After mice were anesthetized [medetomidine (0.3 mg/kg), midazolam (4.0 mg/kg), and butorphanol (5.0 mg/kg)], a guide cannula (AG-4, Eicom) was implanted into the skull with stainless steel screw and dental acrylic cement into the bilateral dSTR (AP, 0.5 mm; ML, ±2.0 mm; DV, 3.0 mm) or dRN (AP, −4.4 mm; ML, 0 mm; DV, 3.5 mm) using stereotaxic frame. Deschloroclozapine (DCZ; 0.1 μm/0.2 μl per side; Med Chem Express) or CGP36216 (GABA(B) antagonist; 3 mm/0.1 μl per side; APExBIO) were infused into dSTR or dRN, respectively, by inserting the broken dialysis probe (A-I-4-01, Eicom) through the guide cannula using an injector EPS-64 microsyringe pump (Eicom). The effect of DCZ as a selective chemogenetic ligand for designer receptors (hM3Dq) has been previously reported (Nagai et al., 2020). The dose and timing of drug administration were based on previous studies (Nagai et al., 2020; Li et al., 2021). Ringer’s solution was used as the control.
In vivo microdialysis
5-HT measurement
5-HT measurements using in vivo microdialysis were performed as previously described (Miyamoto et al., 2017). The guide cannula was implanted with stainless steel screw and dental acrylic cement into the mice dSTR (AP, 0.5 mm; ML, ±2.0 mm; DV, 2.5 mm) using a stereotaxic frame. Twenty-four hours after surgery, Ringer’s solution was perfused (flow rate: 0.5 μl/min), and the dialysate was collected in a 12 min fraction through a dialysis probe (Eicom). The collected dialysate was injected into a high-performance liquid chromatography (HPLC) system (HTEC-500; Eicom). Two hours after probe insertion, baseline 5-HT levels were calculated as the average of four consecutive fractions (the difference between each value was <10%). DCZ was microinfused when the 5-HT baseline content was stable.
GABA measurement
GABA levels were measured using in vivo microdialysis as previously reported (Fu et al., 2016). The guide cannula was implanted into the mice dRN (AP, −4.4 mm; ML, 0 mm; DV, 3.0 mm), similar to the 5-HT measurement. Twenty-four hours after surgery, a dialysis probe (FX-I-4-01, 1-mm membrane length, Eicom) was inserted into the guide cannula, and the Ringer’s solution was perfused (flow rate: 0.5 μl/min). The dialysate was collected in a 30 min fraction using a dialysis probe. The collected dialysate was injected into an HPLC system (HTEC-500; Eicom). The average of four consecutive dialysates collected 4 h after probe insertion was defined as the baseline GABA level.
Statistical analysis
All statistical analyses were performed using GraphPad Prism version 7. For comparing the results between the two groups, the Student’s t test was used. One-way ANOVA followed by Bonferroni’s post hoc test were used to compare the results of single-factor experiments between more than two groups. Two-way ANOVA followed by Tukey–Kramer’s or Bonferroni’s post hoc test, was used to compare the results of the double-factor experiments. Correlations were measured using Pearson’s correlation coefficient (r). All data are expressed as the mean ± SEM.
Result
5-HT content in the dorsal striatum and sensitivity to RSDS
Depression-like behaviors were assessed in mice after exposure to RSDS using a social interaction test. No differences in the time spent in the IZ were observed among the three groups during the pretest (Fig. 1a, left). However, the SI time decreased in susceptible mice compared with stress-naive and resilient mice during the post-test (Fig. 1a, right; F(2,22) = 15.06, p < 0.0001; one-way ANOVA). The IR was also significantly lower in susceptible mice than in stress-naive and stress-resilient mice (Fig. 1b; F(2,22) = 43.54, p < 0.0001; one-way ANOVA). In contrast, the time spent in the AZ was significantly increased in susceptible mice compared with that in stress-naive and resilient mice (Fig. 1c; F(2,22) = 24.10, p < 0.0001; one-way ANOVA). These results indicated that the RSDS was correctly performed, allowing us to obtain stress-resilient or stress-susceptible groups of animals.
Decreased 5-HT content in the dSTR is correlated with depression-like behaviors. a–c, Results of social interaction test. a, Time in the interaction zone. Naive, n = 9; Resilient, n = 8; Susceptible, n = 8; ***p < 0.005 versus susceptible mice (one-way ANOVA with Bonferroni’s post hoc tests). b, Interaction ratio. Naive, n = 9; Resilient, n = 8; Susceptible, n = 8; ***p < 0.005 versus susceptible mice (one-way ANOVA with Bonferroni’s post hoc tests). c, Time in avoidance zone. Naive, n = 9; Resilient, n = 8; Susceptible, n = 8; ***p < 0.005 versus susceptible mice, **p < 0.01 versus susceptible mice (one-way ANOVA with Bonferroni’s post hoc tests). d, Quantitative of basal 5-HT levels in the dSTR. Naive, n = 7; Resilient, n = 7; Susceptible, n = 6; **p < 0.01 versus susceptible mice, *p < 0.05 versus susceptible mice (one-way ANOVA with Bonferroni’s post hoc tests). e, f, Correlation of basal 5-HT content in the dSTR with the social interaction ability assessed by the time spent in interaction zone (e) and by interaction ratio (f) after RSDS. Resilient, n = 7; Susceptible, n = 6 (Pearson’s correlation test).
We have previously demonstrated the relationship between 5-HT in the dSTR and Shati/Nat8l-induced stress vulnerability to RSDS (Uno et al., 2019). The basal 5-HT content in the dSTR of C57BL/6J mice exposed to RSDS was measured using in vivo microdialysis after social interaction tests. They were significantly decreased in susceptible, but not resilient, mice compared with those in stress-naive mice (Fig. 1d; F(2,17) = 6.55, p = 0.0078; one-way ANOVA), suggesting the involvement of dSTR 5-HT in stress sensitivity underlying the pathogenesis of depression. These results are consistent with those of a previous study, in which a reduction in 5-HT content was observed in mice with dSTR-specific Shati/Nat8l overexpression (referred to as dSTR-Shati OE; Miyamoto et al., 2017).
Activation of serotonergic neurons from the dorsal raphe nucleus using the designer receptors exclusively activated by designer drugs (DREADD) system and 5-HT levels in the dorsal striatum
The dRN contains major serotonergic populations that project to numerous areas of the brain including the dSTR (Azmitia and Segal, 1978; Pollak et al., 2014), thus, we focused on studying the serotonergic system in dRN. The AAV-pet1-hM3Dq vector (Hasegawa et al., 2014) was microinjected into the dRN, and deschloroclozapine (DCZ) was locally administered to effectively activate serotonergic neurons from the dRN expressing hM3Dq (Fig. 2a). The expression of EYFP (green signals) was detected in serotonergic neurons (TPH-positive cells; red signals) in the dRN (Fig. 2b). We also microinjected the AAV-pet1-Mock vector into the dRN of mice in the control group. Forty-eight minutes after DCZ microinfusion, a significant increase in dSTR 5-HT content was observed for 1 h in dRN-hM3Dq mice treated with DCZ, but not in the groups injected with the AAV-pet1-Mock vector or treated Ringer’s solution (Fig. 2c; interaction effect: F(39,201) = 5.268, p < 0.0001, two-way ANOVA).
Pharmacogenetic activation of serotonergic neurons in the dRN. a, Schematic of microinjection of AAV vectors into the dRN and microinfusion of DCZ into the dSTR. b, Representative images of injection sites (dRN). TPH-positive cells: red signal; EYFP-positive cells: green signal. Entire portion (left), Scale bars: 200 μm. Magnified portion (right), Scale bars: 20 μm. c, 5-HT content in dSTR collected every 12 min for 2 h from the time DCZ was microinfused. dRN-Mock with Ringer, n = 4; dRN-Mock with DCZ, n = 5; dRN-hM3Dq with Ringer, n = 5; dRN- hM3Dq with DCZ, n = 5; ***p < 0.005 versus dRN-Mock with Ringer, **p < 0.05 versus dRN-Mock with Ringer (two-way ANOVA with Bonferroni’s post hoc tests).
Striatal Shati/Nat8l expression and vulnerability to stress
Furthermore, we investigated the contribution of serotonergic neurons projecting from the dRN to the dSTR to stress sensitivity. As mentioned above, we demonstrated that dSTR Shati/Nat8l regulates stress sensitivity and that dSTR-Shati OE mice are susceptible to stress (Uno et al., 2019). The dSTR-Shati OE mice were obtained by microinjecting the AAV-CMV-Shati/Nat8l vector into their dSTR. The AAV-CMV-Mock vector was injected into the dSTR of control mice (referred to as dSTR-Mock). As shown in Figure 3a, His-positive cells (green signals) were detected in the dSTR of the dSTR-Shati OE mice. We confirmed the overexpression of Shati/Nat8l mRNA in dSTR using real-time RT-PCR (Fig. 3b; t(13) = 4.199, p = 0.0010; Student’s t test). These mice were exposed to microdefeat stress (subthreshold social stress) to assess stress sensitivity, and depression-like behaviors were assessed following the protocol (Fig. 3c). Bilateral microinjection of AAV-CMV-Shati/Nat8l or Mock vectors into the mouse dSTR and the AAV-pet1-hM3Dq vector into the dRN was performed. DCZ was microinfused 45 min before microdefeat stress to adjust the timing during the upregulation of 5-HT content in the dSTR using a guide cannula bilaterally implanted. Twenty-four hours after microdefeat stress, behavioral tests were performed to assess the depression-like behaviors induced by RSDS exposure. There was no difference in the SI time between all groups in the pre-test (Fig. 3d, left). No difference in SI time after microdefeat stress was observed between dSTR-Mock mice treated with Ringer’s solution or DCZ (Fig. 3d, right), suggesting that DCZ had no effect on social interaction in dSTR-Mock mice. dSTR-Shati OE mice treated with Ringer’s solution showed a significant decrease in social interactions compared with dSTR-Mock mice. Activation of dRN-dSTR serotonergic neurons in dSTR-Shati OE mice by DCZ treatment prevented this decrease (Fig. 3d, right; interaction effect: F(1,39) = 27.26, p < 0.0001; two-way ANOVA). Furthermore, dSTR-Shati OE mice treated with Ringer’s solution, but not those treated with DCZ, showed lower IR after microdefeat stress (Fig. 3e; interaction effect: F(1,39) = 11.71, p = 0.0015; two-way ANOVA). Notably, the proportion of the stress-resilient group after microdefeat stress exposure was strongly elevated by DCZ treatment in dSTR-Shati cells. Although 27.2% of dSTR-Shati OE mice treated with Ringer’s solution showed stress resilience after exposure to microdefeat stress (n = resilient/susceptible: 3/8), all dSTR-Shati OE mice treated with DCZ belonged to the stress-resilient group (n = resilient/susceptible: 11/0). The decreased AZ time in dSTR-Shati OE mice treated with Ringer’s solution was prevented by microinfusion of DCZ into the dSTR (Fig. 3f; interaction effect: F(1,39) = 9.390, p = 0.0039; two-way ANOVA). In the sucrose preference test, microdefeat stress decreased sucrose preference in ringer-treated dSTR-Shati OE mice, but not in DCZ-treated dSTR-Shati OE mice (Fig. 3g; interaction effect: F(1,39) = 5.059, p = 0.0302; two-way ANOVA). In the tail suspension test (Fig. 3h; interaction effect: F(1,39) = 2.152, p = 0.1504; two-way ANOVA), No increase in the immobility time assessed in the FST was observed in DCZ-treated dSTR-Shati OE mice compared with ringer-treated dSTR-Shati OE mice (Fig. 3i; interaction effect: F(1,39) = 1.961, p = 0.1693; two-way ANOVA). To exclude the possibility that our results were affected by experimental methods, such as microinjection or microinfusion into the dSTR and dRN, a locomotor activity test was performed to assess basic motor function. We confirmed that these procedures had no influence on motor function (Fig. 3j). These results indicate that overexpression of Shati/Nat8l in the dSTR induces vulnerability to social stress, possibly by controlling striatal 5-HT levels through serotonergic neurons projected from the dRN to the dSTR.
Upregulation of 5-HT content in the dSTR induced resilience to stress in dSTR-Shati OE mice. a, Representative images of injection sites (dSTR). Shati/Nat8l (His-positive cells): green signal, NeuN-positive cells: red signal. Entire portion (right), Scale bars: 500 μm. Magnified portion (left), Scale bars: 50 μm. b, Relative expression level of Shati/Nat8l mRNA in the dSTR. dSTR-Mock, n = 8; dSTR-Shati OE, n = 7; **p < 0.01 versus dSTR-Mock mice (student t test). c, Schematic of microinjection and microinfusion and the timeline of experiments. d–f, Results of social interaction test. d, Time in the interaction zone. dSTR-Mock with ringer, n = 10; dSTR-Mock with DCZ, n = 11; dSTR-Shati OE with Ringer, n = 11; dSTR-Shati OE with DCZ, n = 11; ***p < 0.005 versus dSTR-Mock with Ringer; ###p < 0.005 versus dSTR-Shati OE with DCZ (two-way ANOVA with Bonferroni’s post hoc tests). e, Interaction ratio. dSTR-Mock with Ringer, n = 10; dSTR-Mock with DCZ, n = 11; dSTR-Shati OE with Ringer, n = 11; dSTR-Shati OE with DCZ, n = 11; ***p < 0.005 versus dSTR-Mock with Ringer; ##p < 0.01 versus dSTR-Shati OE with DCZ (two-way ANOVA with Bonferroni’s post hoc tests). f, Time in avoidance zone. dSTR-Mock with Ringer, n = 10; dSTR-Mock with DCZ, n = 11; dSTR-Shati OE with Ringer, n = 11; dSTR-Shati OE with DCZ, n = 11; **p < 0.01 versus dSTR-Mock with Ringer; ###p < 0.005 versus dSTR-Shati OE with DCZ (two-way ANOVA with Bonferroni’s post hoc tests). g, Result of sucrose preference test. dSTR-Mock with Ringer, n = 10; dSTR-Mock with DCZ, n = 11; dSTR-Shati OE with Ringer, n = 11; dSTR-Shati OE with DCZ, n = 11; ***p < 0.005 versus dSTR-Mock with ringer; ##p < 0.01 versus dSTR-Shati OE with DCZ (two-way ANOVA with Bonferroni’s post hoc tests). h, Immobility time in tail suspension test. dSTR-Mock with Ringer, n = 10; dSTR-Mock with DCZ, n = 11; dSTR-Shati OE with Ringer, n = 11; dSTR-Shati OE with DCZ, n = 11; *p < 0.05 versus dSTR-Mock with Ringer (two-way ANOVA with Tukey’s post hoc tests). i, Immobility time in a forced swimming test. dSTR-Mock with Ringer, n = 10; dSTR-Mock with DCZ, n = 11; dSTR-Shati OE with Ringer, n = 11; dSTR-Shati OE with DCZ, n = 11; *p < 0.05 versus dSTR-Mock with Ringer; #p < 0.05 versus dSTR-Shati OE with DCZ (two-way ANOVA with Bonferroni’s post hoc tests). j, Results of locomotor activity test. dSTR-Mock with Ringer, n = 10; dSTR-Mock with DCZ, n = 11; dSTR-Shati OE with Ringer, n = 11; dSTR-Shati OE with DCZ, n = 11 (two-way ANOVA with Bonferroni’s post hoc tests).
dRN GABA content in stress susceptible mice and dSTR-Shati OE mice and correlation with the social interaction behavior
Next, we investigated the regulatory mechanisms of the dSTR Shati/Nat8l on the activity of serotonergic neurons projecting from the dRN to the dSTR. In our previous study, we demonstrated the existence of neuronal projections from the dSTR to the dRN (Uno et al., 2019). Medium-sized spiny neurons (MSNs) are the most abundant cells in dSTR, accounting for 95% of all neurons (Graveland and DiFiglia, 1985; Kreitzer, 2009). Given that MSNs are GABAergic cells and major projection neurons in the dSTR (Mao et al., 2019), we investigated the involvement of GABAergic neurons in the regulation of the dRN serotonergic system. We found that after exposure to RSDS, stress-susceptible mice exhibited significantly higher basal GABA concentrations than in naive mice. This difference was not present between stress-resilient and naive mice (Fig. 4a; F(2,22) = 8.423, p = 0.0019; one-way ANOVA). Furthermore, basal dRN GABA content was negatively correlated with indicators of social interaction ability, including IZ time and IR (Fig. 4b,c; vs time in the IZ: r = −0.841, p < 0.0001; vs social IR: r = −0.796, p = 0.0002; Pearson’s correlation test). This suggests the potential involvement of dRN-GABA in the pathogenesis of depression.
GABA content in the dRN is correlated with depression-like behaviors and controlled by Shati/Nat8l in the dSTR. a, Quantitative basal GABA content in the dRN. Naive, n = 9; Resilient, n = 8; Susceptible, n = 8; **p < 0.01 versus susceptible mice; *p < 0.05 versus susceptible mice (one-way ANOVA with Bonferroni’s post hoc tests). b, c, Correlation of basal GABA content in the dRN with the social interaction ability assessed by the time spent in interaction zone (b) and by interaction ratio (c) after RSDS. Resilient, n = 8; Susceptible, n = 8 (Pearson’s correlation test). d, Quantitative basal GABA content in the dRN. Mock, n = 8; dSTR-Shati, n = 7; **p < 0.01 versus Mock mice (student t test).
We investigated the relationship between dSTR Shati/Nat8l expression and dRN GABA content. dSTR-Shati OE mice exhibited vulnerability to social defeat stress and showed an increase in basal dRN GABA content compared with dSTR-Mock mice (t(12) = 3.939, p = 0.0020; Student’s t test; Fig. 4d). These results suggest that the dSTR Shati/Nat8l regulates stress sensitivity via GABAergic neurons.
The function of GABA in the dorsal raphe nucleus on the stress sensitivity
Next, we assessed the role of dRN GABA in stress sensitivity. We microinfused CGP36216, a GABA(B) receptor antagonist, into the dRN of dSTR-Shati OE mice before microdefeat stress and depressive behaviors were investigated following the schedule (Fig. 5a). The IZ time between all groups in the pretest did not change (Fig. 5b, left). Although there was a difference in IZ time after microdefeat stress between dSTR-Mock mice and dSTR-Shati OE mice treated with Ringer’s solution, inhibition of GABA(B) signaling in the dRN by CGP36216 suppressed the decrease in IZ time observed in dSTR-Shati OE mice in the post-test (Fig. 5b, right; interaction effect: F(1,34) = 15.66, p = 0.0004; two-way ANOVA). The decrease in IR and increase in AZ time after microdefeat stress exposure in dSTR-Shati OE mice also disappeared in dSTR-Shati OE mice treated with CGP36216 (Fig. 5c,d; IR: interaction effect: F(1,34) = 5.973, p = 0.0199; AZ time: interaction effect: F(1,34) = 6.545, p = 0.0151; two-way ANOVA). Although sucrose preference in ringer-treated dSTR-Shati OE mice decreased after microdefeat stress, CGP36216-treated dSTR-Shati OE mice did not show this behavior (Fig. 5e; interaction effect: F(1,34) = 6.080, p = 0.0189; two-way ANOVA). Furthermore, dSTR-Shati OE mice treated with Ringer’s solution, but not those treated with CGP36216, showed longer immobility time after microdefeat stress exposure in the tail suspension and forced swimming tests (Fig. 5f,g; tail suspension test: interaction effect: F(1,34) = 11.01, p = 0.0022; forced swimming test: interaction effect: F(1,34) = 11.96, p = 0.0015; two-way ANOVA). Impaired motor function was not observed in the CGP36216-treated mice (Fig. 5h). These results indicate that dSTR-Shati/Nat8l-induced stress vulnerability is mediated by GABA(B) signaling in the dRN and imply that inhibition of this signaling results in stress resistance.
Blockade of GABA signaling in the dRN suppressed stress vulnerability in dSTR-Shati OE mice. a, Schematic of microinjection and microinfusion and the timeline of experiments. b–d, Results of social interaction test. b, Time in the interaction zone. dSTR-Mock with ringer, n = 10; dSTR-Mock with CGP36216, n = 11; dSTR-Shati OE with Ringer, n = 9; dSTR-Shati OE with CGP36216, n = 9; ***p < 0.005 versus dSTR-Mock with Ringer; ###p < 0.005 versus dSTR-Shati OE with CGP362160 (two-way ANOVA with Bonferroni’s post hoc tests). c, Interaction ratio. dSTR-Mock with Ringer, n = 10; dSTR-Mock with CGP36216, n = 10; dSTR-Shati OE with Ringer, n = 9; dSTR-Shati OE with CGP36216, n = 9; ***p < 0.005 versus dSTR-Mock with Ringer; ##p < 0.01 versus dSTR-Shati OE with CGP36216 (two-way ANOVA with Bonferroni’s post hoc tests). d, Time in avoidance zone. dSTR-Mock with Ringer, n = 10; dSTR-Mock with CGP36216, n = 10; dSTR-Shati OE with Ringer, n = 9; dSTR-Shati OE with CGP36216, n = 9; **p < 0.01 versus dSTR-Mock with Ringer; ##p < 0.01 versus dSTR-Shati OE with CGP36216 (two-way ANOVA with Bonferroni’s post hoc tests). e, Result of sucrose preference test. dSTR-Mock with Ringer, n = 10; dSTR-Mock with CGP36216, n = 10; dSTR-Shati OE with Ringer, n = 9; dSTR-Shati OE with CGP36216, n = 9; ***p < 0.005 versus dSTR-Mock with ringer; ##p < 0.01 versus dSTR-Shati OE with CGP36216 (two-way ANOVA with Bonferroni’s post hoc tests). f, Immobility time in tail suspension test. dSTR-Mock with Ringer, n = 10; dSTR-Mock with CGP36216, n = 10; dSTR-Shati OE with Ringer, n = 9; dSTR-Shati OE with CGP36216, n = 9; **p < 0.01 versus dSTR-Mock with Ringer; ###p < 0.005 versus dSTR-Shati OE with CGP36216 (two-way ANOVA with Tukey’s post hoc tests). g, Immobility time in a forced swimming test. dSTR-Mock with Ringer, n = 10; dSTR-Mock with CGP36216, n = 10; dSTR-Shati OE with Ringer, n = 9; dSTR-Shati OE with CGP36216, n = 9; **p < 0.01 versus dSTR-Mock with Ringer; ###p < 0.005 versus dSTR-Shati OE with CGP36216 (two-way ANOVA with Tukey’s post hoc tests). h, Results of locomotor activity test. dSTR-Mock with Ringer, n = 10; dSTR-Mock with CGP36216, n = 10; dSTR-Shati OE with Ringer, n = 9; dSTR-Shati OE with CGP36216, n = 9 (two-way ANOVA with Bonferroni’s post hoc tests).
To clarify the regulatory mechanism of GABA neurotransmitter by Shati/Nat8l, we focused on brain-derived neurotrophic factor (BDNF). The role of BDNF in the regulation of GABAergic synapse plasticity were reported (Brady et al., 2018). Furthermore, we previously reported that knock-down of Shati/Nat8l decreased BDNF expression in the dSTR, and BDNF in the dSTR determined the stress sensitivity (Miyanishi et al., 2021). BDNF mRNA in dSTR increased in dSTR-Shati OE mice compared with dSTR-Mock mice (Fig. 6; t(9) = 3.201, p = 0.0108; Student’s t test).
Overexpression of Shat/Nat8l enhanced BDNF expression in the dSTR. Relative expression level of Shati/Nat8l mRNA in the dSTR. dSTR-Mock, n = 6; dSTR-Shati OE, n = 5; *p < 0.05 versus dSTR-Mock mice (student t test).
Discussion
Our study provides evidence that the neuronal circuitry between the dSTR and the dRN, which controls 5-HT levels, is essential for the regulation of stress sensitivity. Depression was accompanied by decreased dSTR 5-HT levels in stress-susceptible mice after RSDS exposure. Therefore, the dSTR 5-HT content interferes with social interaction behaviors. Similar changes in dSTR 5-HT levels were observed in dSTR-Shati OE mice, which were vulnerable to stress. Increased vulnerability to stress induced by the overexpression of Shati/Nat8l, was prevented by the specific activation of dRN-dSTR serotonergic neurons using the DREADD system. Thus, Shati/Nat8l regulates stress sensitivity by modulating 5-HT levels in the dSTR via dRN-dSTR serotonergic neurons. Additionally, increased dRN-GABA content was observed only in RSDS stress-susceptible mice. Shati OE mice showed similar changes in dRN GABA content. dRN GABA content is correlated with social interaction behavior and the inhibition of GABA signaling in dRN-induced stress resilience, suggesting that GABA levels in the dRN determine stress sensitivity. The present study suggests that there is an interactive neuronal network connecting the dSTR and dRN, which is controlled by Shati/Nat8l, and regulates stress sensitivity by influencing GABA release in the dRN and 5-HT release in the dSTR. Thus, our study reveals the underlying molecular mechanisms that contribute to stress susceptibility.
RSDS-induced depression-like behaviors, such as impaired social interaction ability, lack of pleasure (anhedonia), and helplessness in stressful situations (Carnevali et al., 2020), reflect the symptoms of depression in humans. These depression-like behaviors are prevented by the administration of SSRIs (Bourke et al., 2014; Liu et al., 2020). SSRI microinfusion into the dSTR prevented vulnerability to RSDS and even subthreshold social stress in dSTR-Shati OE mice (Uno et al., 2019), indicating that 5-HT levels in the dSTR play an important role in stress sensitivity. These results are consistent with those obtained in the present study.
The incidence of depression is higher in women than in men (Altemus et al., 2014). Although RSDS protocols for female mice have been established (Takahashi et al., 2017), male mice were used in this study because microdefeat stress requires reasonable aggressiveness of the used ICR mice, and it was difficult to apply microdefeat stress protocols to female mice in the present study. We also investigated the role of the dSTR-dRN neural circuitry in stress sensitivity without considering the possible influence of the estrous cycle. However, an experiment including female mice should be considered in future studies using appropriate protocols of social defeat stress or other female-specific social stress models, such as social crowding stress (Furman et al., 2022), because of the need to develop treatments for all patients with depression.
Although the ventral striatum (nucleus accumbens) is involved in the pathogenesis of stress or depression (Nestler and Carlezon, 2006; Heshmati et al., 2020), the involvement of the dSTR in these pathologies was not demonstrated until our previous study (Miyamoto et al., 2017). The dSTR contributes to negative decision-making and emotional behavior (Amemori et al., 2018; Klawonn et al., 2021). Among the neuronal networks with other brain regions that modulate such functions (Lago et al., 2017; Cox and Witten, 2019), we considered the involvement of bidirectional dSTR-dRN neuronal interplay in the present study. Given that 5-HT in the dSTR plays an important role in stress sensitivity, we focused on the dRN serotonergic system. Most dRN neurons are serotonergic and project to the entire brain (Ishimura et al., 1988; Pollak et al., 2014). We demonstrated that dRN-dSTR serotonergic neurons are downstream of striatal Shati/Nat8l and that their activation induces stress resilience. To clarify the mechanisms underlying the regulation of dRN-dSTR serotonergic neurons by dSTR Shati/Nat8l, we hypothesized that dSTR-dRN GABAergic neurons were present. Medium-sized spiny neurons (MSNs) are GABAergic cells, which are the major projection neurons of the striatum (Mao et al., 2019), and are the most abundant cell type. ∼95% of the neurons in the striatum are MSN cells (Kemp and Powell, 1971; Babenko et al., 2020). Here, we demonstrated that elevated dRN GABA content was observed in stress-susceptible but not in resilient mice subjected to RSDS. Moreover, dRN GABA levels were strongly correlated with social interaction behavior indicators. These results are consistent with our other findings that dSTR-Shati OE mice exhibit increased vulnerability to social stress and show upregulation of dRN GABA content. We also demonstrated that blocking of GABA signaling in the dRN inhibited stress susceptibility in dSTR-Shati OE mice, suggesting that dRN GABA induces stress vulnerability through suppression of serotonergic neurons projecting from the dRN to the dSTR because GABA functions as an inhibitory neurotransmitter (Petroff, 2002). Decreased dSTR 5-HT content in dSTR-Shati OE mice and stress-susceptible mice after RSDS is in accordance with the fact that dRN GABA content in these mice is increased. Previous reports showed that the antidepressant effect of the potentiation of 5-HT neurons is mediated by decreased GABA signaling in the dRN (Asaoka et al., 2017), further supporting our suggestion that GABA neurotransmission suppress serotonergic neuron in the dRN. GABA(B) receptor antagonists were used in this study because previous studies have shown that blockage of GABA(B) receptors induces an antidepressant effect (Mombereau et al., 2004; Alexander, 2017) and activation of GABA(B) receptors in the dRN attenuates 5-HT neuronal activity (Tao and Auerbach, 2000), indicating the depressant effect of GABA(B) signaling in the dRN.
Although we emphasized the role of dSTR 5-HT in the present study, our results do not deny the possible involvement of 5-HT from other regions in the pathology of depression. In fact, 5-HT imbalances in the hippocampus and medial prefrontal cortex (mPFC) contribute to this imbalance (Le et al., 2016; Belleau et al., 2019). Depression-like behavior in rodents is prevented by the microinfusion of 5-HT or a 5-HT receptor agonist into the hippocampus or mPFC (Luo et al., 2008; Fukumoto et al., 2018). However, the roles of 5-HT in various regions may differ, and the dSTR 5-HT determines stress sensitivity. The dSTR-Shati OE mice, which had reduced STR 5-HT levels, did not exhibit depression-like behaviors without social defeat stress. Furthermore, activation of dSTR-dRN serotonergic neurons in mock mice did not induce antidepressant behaviors, indicating that striatal 5-HT deficits impair stress sensitivity but not depression-like behaviors. TPH-2 knock-in mice, which show a decrease in brain 5-HT by ∼70%, do not show depression-like behavior without social defeat stress either (Sachs et al., 2015). These reports support our hypothesis on the role of 5-HT in stress sensitivity.
Shati/Nat8l regulates neural circuitry between the dSTR and dRN. One possible regulatory mechanism is BDNF. dSTR Shati/Nat8l knock-down mice exhibit reduced BDNF mRNA and protein levels in the dSTR and resilience to high-intensity RSDS exposure (Miyanishi and Nitta, 2021). Furthermore, inhibition of BDNF signaling by ANA-12 (a tyrosine protein kinase inhibitor) induces stress resilience (Miyanishi et al., 2021). These results suggest that the BDNF levels in the dSTR are regulated by Shati/Nat8l and are involved in stress sensitivity. Upregulation of BDNF mRNA levels was observed in stress-susceptible but not resilient mice subjected to RSDS (Miyanishi et al., 2021), and we confirmed that dSTR Shati/Nat8l OE mice have higher BDNF mRNA levels in the dSTR compared with those in mock mice. It is possible that the BDNF increase in the dSTR induced by Shati/Nat8l overexpression could account for the high susceptibility to stress in dSTR-Shati OE mice. BDNF promotes neurogenesis, neuronal excitability, and plasticity by exerting neurotrophic functions (Ferrini and De Koninck, 2013; Colucci-D’Amato et al., 2020; Yang et al., 2020). In the present study, we observed an increase in dRN-GABA content in stress-susceptible and dSTR-Shati OE mice. The activation of GABAergic neurons projecting from the dSTR to the dRN may be induced by the upregulation of striatal BDNF levels.
While conventional antidepressants do not have efficacy in 30% of patients with treatment resistance, ketamine induces rapid effects even in those patients (Rosenblat et al., 2019). The activation of the serotoninergic system in the dRN has been reported as a therapeutic mechanism of ketamine (Fukumoto et al., 2016; Chaki and Fukumoto, 2019). Targeting dSTR Shati/Nat8l might result in effects similar to those of ketamine through direct modulation of the dRN serotonergic system. dSTR Shati/Nat8l might be a promising candidate for developing novel antidepressant therapies that are useful for all patients, including those with refractory forms of the disease.
In conclusion, we offer evidence supporting the role of dSTR 5-HT deficits in the pathogenesis of depression. We demonstrated that dSTR 5-HT and dRN GABA contents were regulated by dSTR Shati/Nat8l, and that these neurotransmitters contribute to stress sensitivity. Our results suggests that bidirectional dSTR-dRN neural circuitry determines stress resilience. Upregulation of dSTR 5-HT levels induces stress resilience, and overexpression of dSTR Shati/Nat8l results in enhanced BDNF expression in the dSTR and GABA release in the dRN and establishment of stress vulnerability through the reduction of 5-HT release in the dSTR (Fig. 7). Our study provides insights into the mechanism regulating stress sensitivity, a major contributor to depression onset, and indicate that targeting dSTR Shati/Nat8l allows direct regulation of the dRN serotonergic system. This may provide a novel therapeutic strategy for depression.
Hypothesized mechanism of establishment of stress vulnerability after RSDS. Overview of dSTR-dRN circuitry changes in the mouse brain after RSDS. Shati/Nat8l in the dSTR was increased by RSDS, and BDNF expression in the dSTR and GABA release in the dRN was increased. The serotonergic system in the dRN is deactivated by elevated GABA content, and the 5-HT content in the dSTR is decreased. Finally, stress vulnerability is established, leading to depression-like behaviors in response to stress.
Acknowledgments
Acknowledgments: We thank Naomi Takino and Mika Ito for their technical assistance in producing the AAV-CMV-Shati/Nat8l or EGFP vectors. We also thank Michihiro Mieda for providing the AAV-pet1-hM3Dq or ChR-EYFP vectors.
Footnotes
S.-i.M. has equity with the Gene Therapy Research Institution, Co., Ltd., which commercializes AAV vectors for gene therapy applications. S.-i.M. has several conflicts of interest, to the extent that the work in this manuscript increases the value of these commercial holdings. All other authors declare no competing financial interests.
This work was supported by Japan Society for the Promotion of Science (JSPS) Grants-in-Aid for Scientific Research (KAKENHI) Grant Numbers JP22J11998 (to H.M.), 26293213 (to S.-i.M.), JP21H02632 (to A.N.), JP 22H04922, and JP 16H06276 (Grant-in-Aid for Scientific Research on Innovative Areas Platform for Advanced Technologies and Research Resources, Advanced Animal Model Support; AdAMS); Nishinomiya Basic Research Foundation (H.M.); Kobayashi Foundation (A.N.); and Smoking Research Foundation Grant for Biomedical Research and Foundation (A.N.). H.M. has been supported by the Nagai Memorial Research Scholarship from the Pharmaceutical Society of Japan.
This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license, which permits unrestricted use, distribution and reproduction in any medium provided that the original work is properly attributed.