Male C57BL/6J mice (10–12 weeks, 20–22 g) were provided by the Experimental Animal Center of Southern Medical University. The mice in each group could drink and eat freely in a standard specific pathogen-free environment, with alternating 12 h light and dark cycles (lighting interval: 7:00–19:00). All animal procedures were performed according to the National Institutes of Health guide for the care and use of animals for scientific purposes and preapproved by the Institutional Animal Care and Use Committee of Southern Medical University. To simulate chronic MA consumption, MA (purity ≥99.1%, provided by the National Institute Control of Pharmaceutical and Biological Products, Beijing, China) was administered for 4 weeks as previously described (Keshavarzi et al., 2019) (10 mg/kg i. p. daily). Ethology tests were performed after MA administration. At the end of the experiment, mice were euthanized and their brain tissues were collected for analysis. Every effort was made to minimize animal pain, suffering, and distress and reduce the number of animals used.

The FST was used to evaluate depressive-associated behavior in the animals (Cao et al., 2013). The experimental device was a transparent plexiglass hollow cylinder (25 cm high and 12 cm in diameter). The water level was at an approximate height of 20 cm and the water temperature was 25 ± 1 °C. During the experiment, the mice were placed gently into the water and allowed to swim freely for 6 min. Their cumulative immobility time was recorded for the last 4 min. After each experiment, the mice were removed, dried with a towel, and put back into the cage and the water in the device was changed.

The tail end of the mouse was fixed with tape and hung above the ground for 6 min to suspend the head downward. The body of the mice did not contact the tail suspension instrument, except for the tail. The immobility time of mice was recorded in the last 4 min.

SPT was used to evaluate anhedonia during MA-induced depression in mice (Qin et al., 2019). In this experiment, the mice were fed in a single cage. Pure water was placed on one side of the cage and 1% sucrose solution was placed on the opposite side; the mice could freely choose between the pure water and sugar solution. The relative positions of each were randomly decided and we recorded the consumption of both. Sugar preference is expressed as a percentage of the total liquid consumption.

The EPM consisted of an open arm and closed arm. The mice were gently placed in the central area with their back to the experimenter. Video was recorded for 5 min. After the test of each mouse was completed, the device was cleaned thoroughly with 75% alcohol. The total time of open arm entry and the number of open arm entries for each mouse were counted.

After anesthetization with 1% pentobarbital sodium (40 mg/kg), the heart was exposed through a thoracotomy, while the brain was extracted after perfusion. Brain tissue was immersed in 4% paraformaldehyde in saline for 24 h; subsequently, we performed gradient dehydration of the sucrose solution and immersed the brain tissue in a 30% sucrose solution. Then, we performed serial coronal sectioning (40 μm). The brain slices were rinsed with 0.1 M phosphate buffer saline (PBS) three times on a shaking table (5 min each time). A 3% bovine serum albumin (BSA) solution containing 0.5% Triton X-100 was added before drilling for 40 min. Primary antibodies (Mouse anti-GAD67, MAB5406, 1:1,000, Merck Millipore, Billerica, Mass., USA; rabbit anti-TPH2, ab184505, 1:1,000, Abcam, Cambridge, United Kingdom) were added after dilution with 0.1 M PBS containing 3% BSA and incubated in a shaker at 4 °C for 24 h. After rinsing three times with 0.1 M PBS, the corresponding fluorescent secondary antibody was added and incubated for 1 h in the dark at 25 °C. Then, the brain slices were sealed with a mounting medium (Vector Laboratories, Inc., Burlingame, CA), containing 4ʹ,6-diamidino-2-phenylindole. Images were captured under a laser confocal microscope (LSM 710; Carl Zeiss Microscopy, Thornwood, NY, United States).

Total RNA was extracted from the tissue samples using TRIzol. The transcripts of the target genes in each sample were normalized with GAPDH/β-actin and expressed as a fold change using the 2^−ΔΔCt equation. The PCR primers used in this study are listed as follows:

TPH2, forward, 5′ AGC​ATT​TGG​ACG​GAG​GAA​GA 3′, reverse, 5′ TGT​ACT​CGA​CCC​TGG​GAA​TG 3′;

All brain tissues were homogenized in a protein extraction buffer (Beyotime, Shanghai, China) containing protease and phosphatase inhibitors at 4 °C for 30 min. The supernatant was collected after the lysates were centrifuged. Protein concentrations were measured using the BCA Protein Assay Kit (Beyotime, Shanghai, China). The samples were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes (Millipore, Billerica, MA, United States). The membranes were blocked in a blocking buffer at room temperature for 1 h and incubated overnight at 4 °C with anti-TPH2 (1:1,000, ab184505, Abcam), anti-5-HTR2C (1:1,000, ab137529, Abcam), and anti-GAPDH (1:5,000, ab125247, Abcam). Furthermore, the membranes were washed with a Tris-buffered saline with 0.1% Tween 20 Detergent buffer and incubated with corresponding secondary antibodies at room temperature for 1 h. The membranes were detected using electrochemiluminescence reagents (Bio-Rad, Hercules, CA, United States) and visualized using a Tanon Imaging system (Tanon, Shanghai, China). Band densities were measured using the ImageJ software and normalized to GAPDH expression. This experiment was performed in triplicate, and the representative images are presented.

Monoamine transmitters in the brain homogenate were detected by HPLC. The HPLC system included a Sykam high-performance liquid chromatograph, C18 reverse-phase analytical column (DIAMONSIL, 2.1 × 100 mm, 2.5 μm), column temperature chamber, SenCell electronic flow cell (Antec, Netherlands), ADF filter (0.05 Hz), 25 μm electrode, in situ Ag/AgCl reference electrode, VT 03 glass carbon working electrode (3 mm), and pulse damper. The mobile phase consisted of a 15% methanol aqueous solution containing 0.74 mmol/L sodium octane sulfonate, 80 mmol/L sodium dihydrogen phosphate, 0.027 mmol/L disodium ethylenediaminetetraacetic acid, and 2 mmol/L potassium chloride adjusted to pH 3.0 with phosphoric acid. After all solutions were prepared, a 0.22-μm organic phase filter membrane was used to remove bacteria, followed by ultrasonic degassing for 30 min. The column temperature was maintained at 40 °C. Fresh standards were prepared before the experiment. The working voltage of the electrochemical detector electrode and its attenuation were 0.56 V and 200 nA, respectively.

Stereotactic surgery was conducted as previously described (Wang et al., 2019b). Briefly, after mice were anesthetized using an intraperitoneal injection of 40 mg/kg pentobarbital sodium, erythromycin eye ointment was applied to their eyes, and the hair on the top of the skull was cut off using eye scissors. Following iodine disinfection, the scalp was cut, exposing the anterior fontanelle. The position of the mouse head was adjusted to have both the front and back as well as the right and left levels of the mouse brain in the same line. The location of the BLA was: AP = −1.5 mm, ML = ±2.7 mm, and DV = −4.5 mm. The skull was drilled with a dental drill, and the drug delivery cannula (purchased from RWD) was embedded and fixed on the top of the mouse head with dental cement. It was used for drug administration and related behavior tests after 1 week of recovery.

To inject the virus, the microinjection needle (Hamilton, 5 μL, 33 g) was used to puncture to the corresponding depth and 100 nL of adeno-associated virus (AAV) was injected into the bilateral brain regions at a speed of 50 nL/min. After 5 min, the microinjection needle was slowly withdrawn to ensure that the AAV solution was fully absorbed. After suturing, kanamycin lidocaine was applied to the wound to prevent infection.

In the last week of MA administration, the relevant drug intervention experiments began. The catheter was inserted into the cannula and the infusion of the 5-HT_2AR receptor antagonist M100907 (0.1 μmol in 100 nL) (Adrielle and Helio, 2012) and the 5-HT_2CR antagonist sb242084 (10 nmol in 100 nL) (Ceglia et al., 2010) was performed using a microsyringe (701-RN, Hamilton, USA) connected to a microinfusion pump (KD Scientific, United Ststes) within 2 min. After the injection, the catheter was removed after 1 min. The related behavioral experiments were performed 30 min after the last administration.

Small hairpin RNA (shRNA) directed against 5-HT2CR was adapted from Anastasio et al. (2015) and verified by RT-qPCR. The shRNA sequence was synthesized and packed into rAAV-U6-shRNA (5-HT2CR)-CMV-EGFP (rAAV2) by BrainVTA (Wuhan, China). Using a similar process, rAAV-U6-shRNA (scramble)-CMV-EGFP was produced as a scrambled control. The virus (titer: 1 × 10¹³ vg/mL) was injected into the mice 2 weeks before the end of MA administration.

After decapitation, the brain tissue was quickly removed and placed in precooled artificial cerebrospinal fluid (ACSF). After soaking for 1 min, the brain slices containing BLA tissue were transferred to an incubator in a constantly warm ACSF bath and incubated for 1 h. A glass microelectrode was used, with a tip resistance of 3.5 MQ when filled with an internal solution. We adjusted the perfusion speed to 1.5 ml/min and circulated the ACSF (124 mmol/L NaCl, 24 mmol/L NaHCO₃, 5 mmol/L KCl, 2.4 mmol/L CaCl₂, 21.3 mmol/L MgSO₄, 1.2 mmol/L KH2PO₄, and 10 mmol/L glucose, pH 7.35–7.45), continuously saturating the solution with a mixture of 95% O₂ and 5% CO₂. The incubated brain slices were carefully transferred to the recording tank with a pipette and fixed with a cover net. The electrode tip was adjusted to the center of the field of vision under a low-power microscope, to slowly drop the solution onto the cell and form a high-resistance sealing state. The hold current was within 100 pA. Pyramidal neurons and interneurons are distinguished by morphology and action potential (AP), as previously described (Rainnie, 1999). The SF-77 multichannel perfusion system was used for fast liquid exchange (Warner Instrument Corporation). When miniature inhibitory postsynaptic currents (mIPSCs) were recorded, magnesium-free ACSF, TTX, DNQX, and APV were added to the perfusion system to block the action potentials produced by the Na²⁺ channel and the excitatory current mediated by the N-methyl-d-aspartate (NMDA)- and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid-type glutamate receptors. The frequency and amplitude of mIPSCs were recorded with the membrane potential held between −70 and −80 mV. When AP was recorded, the recording mode was adjusted to the current-clamp mode. APs were evoked by a 600-ms, 100 pA depolarizing current pulse. Signals were acquired using a Multiclamp 700 b amplifier (Cellular Devices, Sunnyvale, CA, United States) and pClamp 10 (Molecular Devices, Sunnyvale, CA, United States). The recorded data were analyzed using MiniAnal software.

The data are expressed as mean ± standard error (SEM). SPSS (Released 2007. Version 16.0. Chicago, SPSS Inc.) was used for statistical analysis. Student’s t-test and one-way analysis of variance were used to analyze the differences between groups. Differences with p values <0.05 were considered statistically significant.

After 4 weeks of administration, depression-like behaviors were measured using the FST, SPT, and TST. In the FST experiment, the immobility time in the MA group was significantly higher than that in the control group (Figure 1A). In the SPT, the sucrose preference of mice in the MA group was significantly lower than that in the control group (Figure 1B). In the TST, significantly higher mean immobility times were observed in the MA group than in the control group (Figure 1C). These three tests consistently showed the depression-like state of the animals in the MA group. To assess anxiety-like behavior, mice were placed in an EPM; those in the MA group spent significantly less time in the open arm (Figure 1D) and entered less frequently than those in the control group (Figure 1E), which indicated anxiety-like behavior. We then examined the biosynthesis of 5-HT in the DRN during the behavioral changes. The results of TPH immunofluorescence staining showed that the intensity of TPH2 immunoreactivity in the MA group was significantly lower than that in the control group (Figure 1F). The results from the qPCR and WB analysis of TPH2 showed the same tendency and verified this result (Figures 1G,H). The expression of TPH2 in the DRN was significantly lower in the MA group than in the control group and accompanied the occurrence of emotional disorder-related symptoms.

It is known that DRN is the main area in the brain where 5-HT neurons are distributed. The 5-HT neurons of the DRN project to the BLA; the terminals of 5-HT neurons have been extensively marked in the BLA area (http://connectivity.brain-map.org/). BLA tissue was collected, and HPLC analyses were performed to assay serotonergic metabolism in the BLA regions of the mice. MA administration significantly decreased the levels of 5-HT and its metabolite 5-hydroxyindoleacetic acid (5-HIAA) in the BLA (Figures 2A,B). 5-HT exerts its role by binding with different receptor subtypes, so we examined the mRNA levels of the 5-HT receptors in the BLA. The qPCR results showed that the expression of 5-HT_1AR was significantly decreased, that of 2A and 2C was significantly, and that of 5-HT₃R and 5-HT₄R did not significantly change after MA administration (Figure 2C). The 5-HT_2CR immunofluorescence and WB analysis confirmed that the expression of 5-HT_2CR in the MA group was significantly higher than that in the control group (Figures 2D,E).

5-HTRs are potential therapeutic targets for the treatment of emotional disorders (Amidfar and Kim, 2018). We investigated how changes in 5-HTR affected behavioral changes using a neuropharmacological strategy. During MA administration, we implanted a drug delivery cannula on the bilateral BLA and either the 5-HT_2CR (sb242084) or the 5-HT_2AR (M100907) antagonist was infused (Figure 3A). After 1 week of antagonist administration, we performed the behavioral tests. In the FST experiment, the immobility time of the 5-HT_2CR antagonist group was significantly lower than that in the vehicle-treated group, whereas the 5-HT_2AR antagonist exerted little effect (Figure 3B). In the SPT, the 5-HT_2CR antagonist group showed higher sucrose preference than the vehicle-treated group (Figure 3C). The anxiety-like behavior was evaluated by EPM; the time spent investigating the open arm (Figure 3D) and number of entries into the open arm (Figure 3E) were significantly higher in the 5-HT_2CR antagonist group than in the vehicle-treated group, whereas the 5-HT_2AR antagonist treatment had little effect. These pharmacological blockade experiments show that BLA 5-HT_2CR is a valid target against emotional symptoms when ceasing chronic MA administration.

To confirm the role of 5-HT_2CR in the treatment of emotional symptoms, we packed recombinant AAV2 vector-expressing shRNA against 2C (5-HT_2CR-shRNA1/2) and nonsense shRNA as a negative control (Con-shRNA). The AAV contains a GFP tag to visualize in vivo expression. After AAV-5-HT_2CR-shRNA was injected into the BLA, the virus expressed its GFP in the BLA (Figure 4A), with WB results confirming its interference efficiency (Figure 4B). We then conducted depression-screening behavioral tests. In the FST experiment, the knockdown of 5-HT_2CR significantly increased immobility time (Figure 4C). In the SPT, the knockdown of 5-HT_2CR reversed the decreased sucrose preference compared with that in the Con-shRNA-transfected group (Figure 4D). In the EPM test, the knockdown of 5-HT_2CR resulted in anxiolytic-like effects on behavior in chronic MA administration mice, with significantly increased open arm stay time (Figure 4E) and number of entries (Figure 4F). These knockdown strategy results further show that BLA 5-HT_2CR is a valid target against the emotional symptoms caused by chronic MA administration.

5-HT_2CR-positive neurons have been shown to be co-expressed with GABAergic interneurons in other brain areas (Liu et al., 2007). We performed immunofluorescent co-staining between 5-HT_2CR and GAD67 in the BLA and found that most 5-HT_2CR-positive neurons co-labeled with GAD67 (>93%, Figure 5A), suggesting that 5-HT_2CR-positive neurons in the BLA are GABAergic interneurons. 5-HT_2CR has also been found to be coupled with the GABA-A receptor (Burke et al., 2015). GABA-A is an ion channel with an inhibitory function, indicating that the function of interneurons in the local brain region of the BLA are inhibited during chronic MA administration. We detected mIPSCs in the BLA. The mIPSC frequency was lower in the MA group than in the control group; however, the amplitude remained unchanged (Figures 5B,D). After 5-HT_2CR interference, the function of mIPSCs was partially recovered—the frequency significantly increased, which implies that there is a disinhibitory mechanism in BLA pyramidal neurons. We speculated that 5-HT_2CR positive interneurons regulate the excitability of pyramidal neurons in the local BLA region, thereby affecting neuronal activity. Therefore, we recorded the AP of excitatory principal neurons in BLA; the frequency of BLA AP significantly increased in the MA group but decreased in the 5-HT_2CR knockdown group (Figures 5C,E). Thus, 5-HT_2CR-positive interneuron likely mediates neuronal activities in the BLA.