Sprague-Dawley (SD) rats (male, 220–250 g) and Kunming (KM) mice (male or female, 20–22 g) were purchased from the Beijing Animal Centre (Beijing, China). All animals were housed with food and water ad libitum. The room was maintained at 25 ± 1°C with an alternating 12 h circadian cycle. All procedures were conducted in accordance with the Animal Care and Use Committee of the Beijing Institute of Pharmacology & Toxicology (Beijing, China). The study protocols were approved according to guidelines on the ethical use of animals set by the National Institutes of Health (Bethesda, MD, United States).

LCX001 was provided by the Laboratory of Computer-Aided Drug Design & Discovery, Beijing Institute of Pharmacology and Toxicology (Xiao et al., 2016). Fentanyl (Sigma, Beijing, China) and TH-030418 were administered to generate opioid-induced respiratory depression models. TH-030418 (N-methyl-7a-[(R)-1-hydroxy-1-methyl-3-(thien-3-yl)-propyl]-6,14-endo-ethanotetrahydronororipavine) was a thienorphine derivative, synthesized by our institute. Compared with morphine or dihydroetorphine, TH-030418 displayed non-selective and high binding affinity to all subtypes of opioid receptors (Wen et al., 2011). TH-030418 also exerted a strong and long-acting analgesic effect. Ampakines, LCX001 and CX614 (Tocris, Bristol, United Kingdom), were dissolved in a 20% hydroxypropyl-β-cyclodextrin (HPCD; Sigma-Aldrich, Beijing, China) 0.45% saline solution. Propofol was purchased from AstraZeneca (Viale Dell’industria, Italy). Pentobarbital (Nembutal; Merck, Beijing, China), fentanyl, TH-030418, and morphine hydrochloride (Qinghai Pharmaceutical Factory, Xining, China) were dissolved in 0.45% saline solution at room temperature and used at different doses. The volume of each dose was 1 ml/kg for SD rats and 10 ml/kg for KM mice.

Whole-body plethysmographs (EMKA Technologies, Paris, France) of breathing frequency and depth were performed on unrestrained rodents to evaluate the protection effect of LCX001 against respiratory depression induced by fentanyl, TH-030418, propofol and pentobarbital (Nembutal). Adult male SD rats were placed in the transparent plexiglass instrument consisting of inflow and outflow ports for the continuous delivery of fresh air and removal of expired carbon dioxide. Baseline was produced approximately 30 min after acclimatization of rats to the apparatus environment. Pressure changes were measured with a pressure transducer and a signal conditioner. Each group contained six male SD rats. The control groups incorporated (1) vehicle (20% HPCD, intravenous, i.v.) at 5 min after fentanyl (120 μg/kg, subcutaneous, s.c.), TH-030418 (20 μg/kg, i.v.) or pentobarbital (Nembutal, 80 mg/kg, intraperitoneal, i.p.) separately; (2) vehicle (20% HPCD, i.v.) at 10 min before propofol (30 mg/kg, i.v.); (3) vehicle alone. The experimental groups were post-administered with LCX001 (10 mg/kg, i.v.) at 5 min after fentanyl (120 μg/kg, s.c.), TH-030418 (20 μg/kg, i.v.), or pentobarbital (Nembutal, 80 mg/kg, i.p.), and pre-treatment of LCX001 were conducted at 10 min before propofol (30 mg/kg, i.v.) because of the short-lasting propofol-induced respiratory depression (15–25 min). Respiration data were acquired and analyzed using IOX software (EMKA Technologies, Paris, France).

The analgesic effect of LCX001 was evaluated in the acetic acid-induced writhing test. Male KM mice were injected with acetic acid (0.6%, 0.4 ml/20 g i.p.) and then placed in a transparent plastic cage for observation. The number of writhes was counted and recorded 5 min after acetic acid administration for 15 min. The writhe movement was defined as contraction of the abdominal muscles causing hind limb stretching. Male KM mice (n = 70) were randomly and equally divided into seven groups. In the control groups, morphine (2.5 mg/kg, s.c.) or vehicle (20% HPCD, i.v.) was administered 30 min prior to injection of 0.6% acetic acid. The pre-treatment of LCX001 (12 mg/kg, iv) at 10 min before morphine (2.5 mg/kg, s.c.) was also set as a control group, which was followed by injection of 0.6% acetic acid after 30 min. In the experimental groups, LCX001 (1.5, 3, 6, or 12 mg/kg, i.v.) were administered 10 min before 0.6% acetic acid. Analgesia percentage was expressed using the following formula: Inhibition ratio (%) = 100 × (number of control-number of experiment)/number of control.

The hot plate test was performed to determine the anti-nociceptive effect of LCX001. The temperature of the hot plate (Hugo Sachs Elektronik-Harvard Apparatus GmbH, March-Hugstetten, Germany) was set and maintained at 55 ± 0.5°C. The latency of hind paw licking was evaluated and recorded as the response to nociception. Baseline response latencies of under 5 s or over 30 s were eliminated from the test. Female KM mice (n = 70) were randomly divided into seven groups of 10 each. The hot-plate response latency was measured at 30 min after morphine (10 mg/kg, s.c.) injection, and at 10 min after LCX001 (1.5, 3, 6, or 12 mg/kg, i.v.) or vehicle (20% HPCD, i.v.) administration. The pre-treatment of LCX001 (12 mg/kg, iv) at 10 min before morphine (10 mg/kg, s.c.) was also set as a control group to investigate the impact of LCX001 on anti-nociceptive effect of morphine. Anti-nociceptive data were quantified according to Maximum Possible Effect (MPE), and calculated as: MPE (%) = [(T1–T0)/(T2–T0)] × 100 (T0 and T1 represent latency times of hind-paw licking before and after treatment. T2 was the cut-off time, which was set at 60 s to prevent injury to the animal’s paw).

Adult male SD rats (220–250 g) were euthanized by decapitation. The brainstems were dissected and homogenized for 2 min in 15 volumes of ice-cold 50 mM Tris–HCl buffer (pH 7.4) at 4°C. The homogenates were centrifuged at 1,000 g for 10 min, and the supernatants were collected and centrifuged at 40,000 g for 20 min. The pellets were dispersed in 50 mM Tris-HCl pH 7.4 and centrifuged at 40,000 g for 20 min. The last procedures were repeated a third time with a total of four washes to eliminate endogenous glutamate from the tissue. The resulting pellets were suspended in eight volumes of 0.3 M sucrose, then stored at –70°C until use.

Tissue membranes were washed twice in 50 mM Tris-HCl pH 7.4, and the resultant membranes resuspended in 50 mM Tris-HCl pH 7.4 in the presence of 100 mM KSCN. The dispersed tissue membranes were incubated with 200 nM [³H]AMPA (58 Ci/mmol, PerkinElmer, Shanghai, China) and LCX001 at different doses (10^-9–10^-4 M) in a final volume of 1 ml. The binding reaction was performed in disposable glass culture tubes at 25°C for 60 min. Bound [³H]AMPA was separated from the free ligand by filtration under reduced pressure with GF/C filters presoaked in 0.2% polyethyleneimine. The pellets were washed with ice-cold Tris-HCl pH 7.4 three times and dried with a piece of tissue paper. Nonspecific binding was determined by 1 mM nonradioactive glutamate. The binding activity on the filters was detected using a LS6500 scintillation counter (Beckman, Shanghai, China). For saturation binding with [³H]AMPA, LCX001 (10^-4 M) was added in the resuspended assay with changing concentrations of [³H]AMPA from 1 to 400 nM.

A wild-type rat AMPA-GluA2(R)_flop cDNA corresponding to the entire coding region was generated and cloned into the pcDNA^TM3.1(+) vector (Invitrogen, China, Catalog no. V790-20) by Sangon Biotech Co., Ltd. (Shanghai, China). Human embryonic kidney 293 (HEK293) cells (purchased from the National Infrastructure of Cell Line Resource of China) were cultured in DMEM (Gibco, Shanghai, China) supplemented with 10% fetal calf serum (Hyclone, Beijing, China) and grown in an incubator at 37°C, in a 5% CO₂, 70% humidified atmosphere. For stable transfection, HEK293 cells (3 × 10⁴ cells/well) were seeded in 24-well plates the day before transfection. The cells were grown to 70–90% confluency, and the transfection was performed by adding 1 μg of GluA2 expressing vector per well. Twenty-four hours after transfection, the normal medium was replaced with DMEM/F12 (Gibco, Shanghai, China) growth medium supplemented with 1.0 mg/ml G418 (Gibco, Shanghai, China) and cultured for 7 days. The surviving cells were pelleted and dispersed (100 cells/20 ml), and then seeded into 96-well plates in DMEM/F12 medium containing 500 μg/ml G418. Discrete colonies were picked after 15 days of selection. Western blotting and whole-cell patch-clamp recording were performed to test expression of the AMPA receptor GluA2 subunit and formation of an AMPA-GluA2 ion channel. G418 was kept at 500 μg/ml for the maintenance of stably transfected clones.

HEK293 cells that stably expressed AMPA-GluA2 receptors were seeded into the recording chamber 2 days before patch-clamp recordings, and the experiments were conducted in whole-cell configuration. Thin-walled borosilicate glass (Sutter Instrument, CA, United States) was pulled to make patch pipettes. The patch pipettes had open tips with resistance of 4–8 MΩ and were filled with internal electrode solution (mM): 135 CsCl (Sigma, Beijing, China), 10 CsF (Tocris, Bristol, United Kingdom), 10 HEPES (Amresco, United States), 5 Cs4BAPTA (Tocris, Bristol, United Kingdom), 1 MgCl₂, 0.5 CaCl₂, pH 7.2. The external bath solution consisted of (mM): 135 NaCl, 5 KCl, 10 HEPES, 1 MgCl₂, 1.8 CaCl₂, pH 7.35. The holding potential was set as -80 mV by an Axopatch-700B amplifier (Axon Instruments, CA, United States), and pClamp8 software (Axon Instruments, CA, United States) was used for data acquisition. Currents were recorded and digitized at 10 kHz. A fast-solution switch system was utilized for agonist (L-glutamate, Novabiochem, Darmstadt, Germany) application at varying concentrations, which allowed constant exchange of bath solutions to the cells, and stopped momentarily on the application of agonist or drugs. LCX001 and CX614 were dissolved in dimethyl sulfoxide (DMSO) before dilution with external bath solution (DMSO final concentration 0.5%). Dose-response curves were fitted to the Hill equation using Origin 9.0 software.

Isolated brainstems were prepared from neonatal (P0-P1) SD rats using previously described methods (Mazza et al., 2000). A microscope was used for tissue dissection under aseptic conditions. The rat pups were decapitated, and the brainstems were rapidly removed. After microdissection of the cerebellum and pons from the brainstem, the medulla oblongata was separated and transferred to a new chamber. The tissue containing the medulla oblongata was minced and digested enzymatically with 0.25% trypsin for 15 min. The medulla oblongata cells were filtered, then dispersed and seeded in DMEM medium supplemented with 10% fetal calf serum and 10% equine serum (ExCell Bio, Shanghai, China). Twenty-four hours after seeding, the medium was replaced by Neurobasal (Gibco, Shanghai, China) Medium with B27 (Gibco, Shanghai, China) supplement (50:1). The medulla oblongata cells were fed on days 3–7 with a half volume change of Neurobasal/B27 medium containing Ara-C (3–5 μM) to inhibit glial growth. The cells were cultured for 10–15 days prior to use.

The GluA1 or GluA2 AMPA receptor subunits tagged with super ecliptic pHluorin (SEP) were used to investigate the trafficking and accumulation of AMPA receptors on the surface of neurons in real-time. Plasmids containing SEP-GluA1 and SEP-GluA2(R) were purchased from Addgene (plasmid #24000, plasmid #24001). Recombinant adenoviruses carrying SEP-GluA1 or SEP-GluA2(R) were constructed by Hanbio biotechnology (Shanghai, China) for neuronal transfection and SEP expression, and the plaque-forming units (PFU) were 1.26 × 10¹⁰/ml. Dissociated primary medulla oblongata neurons were prepared as described above and seeded and cultured using 20 mm Poly-D-lysine-coated dishes with inner glass coverslips (NEST, Jiangsu, China) for the preparation of live cell imaging. The primary neurons, cultured for 10 days, were transfected using recombinant adenovirus with incubation at 37°C, in a 5% CO₂, 70% humidified atmosphere for 6 h. The medium containing virus was replaced with normal medium, and imaging was performed 96 h after transfection. Artificial cerebrospinal fluid (ACSF) was prepared as bathing solution containing (mM): 120 NaCl, 5 KCl, 2 CaCl₂, 1.2 MgCl₂, 30 D-glucose, 25 HEPES, 1 tetrodotoxin, pH 7.4. For live neuron imaging, transfected cells were identified by GFP fluorescence in the bathing ACSF solution. Images were taken using the 20 × objective lens (Hamamatsu C9100-50) of an inverted confocal laser scanning microscopy system (PerkinElmer, UltraVIEW VoX, Shanghai, China) at 37°C. SEP fluorescence was achieved from a 488 nm laser line. The time-lapse imaging of SEP-GluA1 and SEP-GluA2 was set with 1s intervals and 100 ms exposure times for 10 min. For testing local stimulation, LCX001 or CX614 (47 μM final) was added to the standard solution 2 min after the commencement of imaging. Regions of interest (ROIs) were chosen from areas of diffuse fluorescence in the shaft and spines for comparison before and after drug treatment.

Dissociated primary medulla oblongata neurons were prepared as described above. The primary neurons were seeded using 20 mm Poly-D-lysine-coated dishes with inner glass coverslips (NEST, Jiangsu, China) and cultured for 15 days. On the day of the Ca²⁺ imaging experiment, Fluo-4 NW dye buffer was prepared by adding 10 ml Hank’s Balanced Salt Solution (HBSS) containing 20 mM HEPES, 2.5 mM probenecid, and 1.5 mM CaCl₂ to one bottle of Fluo-4 NW dye mix (Fluo-4 NW Calcium Assay Kits, Thermo fisher, F36206, Shanghai, China). The normal medium for neurons was replaced with Fluo-4 NW dye buffer and incubated in a 37°C atmosphere for 45 min. The assay was initiated when a baseline was produced, and followed by the addition of L-glutamate (18 μM final) at 1 min. Three minutes after the commencement of imaging, fentanyl or TH-030418 (42 μM final) was added, and the neurons were treated with LCX001 or CX614 (40 μM final) 7 min after the commencement of imaging. To evaluate the possible effect of modulators, changes in intracellular Ca²⁺ were recorded using the 20 × objective lens (Hamamatsu C9100-50) of an inverted confocal laser scanning microscopy system (PerkinElmer, UltraVIEW VoX, Shanghai, China) at 37°C. Fluorescence of Ca²⁺ imaging was obtained from a 488 nm laser line. Time-lapse imaging was set with 1 s intervals and 100 ms exposure times for 10 min.

Data are presented as the mean ± standard error of the mean (SEM). In the acetic acid writhing and hot-plate tests, experiments were performed blind. Compounds were injected by one researcher and the behavior or data was recorded by another researcher. The significance of change in respiratory parameters was evaluated by repeated measures two-way ANOVA (dose × time) followed by Bonferroni post-hoc tests for comparisons between groups. In the analysis of the acetic acid writhing test, hot-plate test and [³H]AMPA binding assay, statistical evaluation between the treated groups and control group was performed by one-way ANOVA followed by Dunnett’s test. For saturation binding with [³H]AMPA, K_d and B_max values of different groups were obtained from fits to saturation binding curves with the KaleidaGraph programme. For whole-cell patch-clamp recordings, dose-response curves were fitted to the Hill equation using Origin 9.0 software (Originlab, MA, United States) and statistical significance was determined by one-way ANOVA followed by Dunnett’s test. For live cell imaging, two-way ANOVA followed by Bonferroni post-hoc tests was used to compare the treated groups and vehicle group. A P value of less than 0.05 was considered to represent a statistically significant difference.

Whole body plethysmography measurements are used to detect spontaneous respiratory parameters generated by unanaesthetised male SD rats. As shown in Figure 1, administration of fentanyl, TH-030418, propofol and pentobarbital (Nembutal) markedly suppressed respiration. Nembutal (80 mg/kg, i.p.) and propofol (30 mg/kg, i.v.) triggered a reduction in respiratory frequency (70.3 ± 2.7% of control, 53.6 ± 5.8% of control, Figure 1F,H). The μ-opioid receptor agonist fentanyl (120 μg/kg, s.c.) and TH-030418 (20 μg/kg, i.v.) produced a more severe suppressive effect on respiratory rate (43.4 ± 5.8% of control, 28.9 ± 3.3% of control, Figure 1B,D), and the suppressed respiration lasted for more than 30 min (Figure 1B,D). 5 min after opioid and pentobarbital treatment, post-administration of LCX001 (10 mg/kg, i.v.) significantly alleviate respiratory depression by increasing the respiratory frequency and minute ventilation (Figure 1B–G, ^∗P < 0.05). In consideration of the short-lasting propofol-mediated depression (within 15–25 min), LCX001 (10 mg/kg, i.v.) was injected 10 min before propofol administration. The pre-treatment of LCX001 also markedly supported pulmonary function after propofol-induced depression (Figure 1H,I, ^∗P < 0.05).

From the Acetic acid writhing testing, as shown in Figure 2A, a 2.5 mg/kg dose of morphine (s.c.) significantly reduced the number of abdominal constrictions of male KM mice compared with the vehicle (20% HPCD, i.v.) group, and the inhibition rate reached entirely 100%. LCX001 in doses of 1.5, 3, 6, and 12 mg/kg inhibited acetic acid-induced writhing by 30.76 ± 12.75%, 41.43 ± 11.32%, 85.35 ± 9.53%, and 94.78 ± 2.74%, respectively. These results indicated that LCX001 had a significant anti-nociceptive effect at 1.5 mg/kg or higher doses (Figure 2A, ^∗P < 0.05), and a 10 mg/kg dose of LCX001, similar to morphine at 2.5 mg/kg, produced a marked analgesic effect on writhing behaviors induced by acetic acid.

In the hot plate test, Morphine (10 mg/kg, s.c.) markedly reduced hind paw licking time of female KM mice and had an increased %MPE of about 100% after hot-plate stimulation (Figure 2B). Administration of LCX001 (i.v.) at 1.5 and 3 mg/kg exerted no effect on the %MPE compared with the vehicle (20% HPCD, i.v.) treated group (Figure 2B, P > 0.05). However, intravenous injection of LCX001 at 6 and 12 mg/kg increased the level of %MPE by 30.18 ± 10.56% and 65.27 ± 12.22%, respectively (Figure 2B, ^∗P < 0.05). The treatment of LCX001 (12 mg/kg) exerted no impact on the analgesic effect of morphine (Figure 2A,B, vehicle+morphine vs. LCX001+morphine, P > 0.05). These results demonstrated that LCX001 at doses of 6 mg/kg or higher produced anti-nociceptive behavior in both acetic acid writhing and hot-plate tests.

In the [³H]AMPA binding assay, specific [³H]AMPA binding to AMPA receptors of tissue was measured as 1595 ± 30 (cpm/mg protein). Compared with the control group, specific [³H]AMPA binding was unchanged by pre-incubation with LCX001 at 10^-9–10^-7 M. However, pre-treatment of LCX001 at 10^-6, 10^-5, and 10^-4 M significantly enhanced [³H]AMPA binding affinity to tissue membranes by 1857 ± 20, 1964 ± 27 and 2095 ± 87 (cpm/mg protein) respectively (LCX001 vs. vehicle control group, ^∗∗P < 0.01 and ^∗∗∗P < 0.001, Figure 3A). These results indicate that LCX001 at 10^-6 M or higher doses increased the binding affinity of [³H]AMPA for AMPA receptors, and a 10^-4 M dose of LCX001 was chosen to generate saturation curves of [³H]AMPA binding. Several [³H]AMPA concentrations up to 400 nM were performed in the saturation binding experiments to investigate the impact of LCX001 pre-incubation (10^-4 M) on K_d and B_max parameters. The [³H]AMPA saturation binding results were analyzed using the direct-fit logistic equation, and the curve was best fitted as a one-site model. The K_d and B_max values were 171 ± 5 nM and 6.89 ± 0.11 pmol/mg of protein (Figure 3B). After pre-incubation with LCX001 at 10^-4 M, the K_d and B_max values from direct fitting were 149 ± 7 nM and 7.77 ± 0.17 pmol/mg of protein (Figure 3B). From the analysis of K_d and B_max value in saturation curves, the binding affinity and the number of binding sites were significantly improved with LCX001 treatment (vehicle group vs. LCX001, student’s paired t-test, ^∗P < 0.05).

Whole-cell voltage-clamp electrophysiology is conducted to assess the impact of LCX001 on HEK293 cells transfected with the cDNAs encoding recombinant homomeric GluA2 receptors. As a potent AMPA modulator, CX614 serves as a positive control. A concentration–response curve of glutamate evoked currents was obtained with increasing concentrations of glutamate (Figure 4A), and an effective concentration 50% of maximum response (EC₅₀) value fitted with a Hill-type equation was 3.43 mM (Figure 4D). When cells were recorded in the presence of modulators, co-application of LCX001 caused a significant leftward shift in the dose-response curve for glutamate and the EC₅₀ value of LCX001 treatment was 2.76 mM (Figure 4D). Consistent with the effect of LCX001, CX614 triggered leftward shifts in the curve and EC₅₀ values of 1.69 mM (Figure 4D). However, LCX001 alone induced no current in the absence of glutamate. These results indicated that both LCX001 and CX614 altered and promoted agonists’ affinities of glutamate to homomeric GluA2 receptors. From the concentration–response curve of glutamate evoked currents, the doses producing a half maximal response (3.5 mM) and a maximal response (10 mM) were used to evaluate the potentiation of LCX001 on glutamate-evoked amplitudes. As shown in Figure 4B,C,E,F, LCX001 (100 μM) potentiated the glutamate (3.5 mM) response by 711 ± 20 (pA) as compared with that of 432 ± 27 (pA) in the control group (Figure 4B,E, ^∗P < 0.05). However, 100 μM CX614 was more efficacious with a potentiation of 946 ± 16 (pA) on the glutamate (3.5 mM) response (Figure 4B,E, ^∗P < 0.05). At 10 mM glutamate evoked amplitudes, LCX001 and CX614 at 100 μM increased the potency of glutamate induced currents by 1120 ± 60 and 1149 ± 71 (pA), respectively, compared with that of 752 ± 35 (pA) in the control group (Figure 4C,F, ^∗P < 0.05). These data revealed that LCX001 markedly enhanced the potency of glutamate at 3.5 and 10 mM, and acted as a positive novel allosteric modulator.

The trafficking, insertion and stabilization in the recycling process are also critical components of normal AMPAR functions. To directly assess possible impact of LCX001 on altering the AMPAR recycling process and make real-time measurements of AMPA receptors on the surface of neurons, super ecliptic pHluorin (SEP) is used to visualize surface-expressed AMPA receptors containing GluA1 and GluA2(R) subunits on individual neurons. GluA2 or GluA1 tagged with super ecliptic pHluorin [SEP-GluA2(R) or SEP-GluA1] is expressed in cultured primary medulla oblongata neurons, and live cell imaging is used to record the change in absolute fluorescence in real time. When neurons, transfected with recombinant SEP-AMPAR adenovirus were prepared in the bathing solution containing artificial cerebrospinal fluid (ACSF, pH 7.4), surface stabilized SEP-GluA2 or SEP-GluA1 become relatively brightly fluorescent. In contrast, a sharp increase in fluorescence was caused by exposure to the bathing solution at a higher extracellular pH (8.5). Furthermore, the fluorescence intensity of SEP-GluA2 or SEP-GluA1 could be eclipsed after exposure to a rapidly acidified environment (pH 6.0). These data revealed that AMPARs containing SEP-GluA2 or SEP-GluA1 were successfully expressed in neurons, and that the fluorescence was sensitive to extracellular pH changes (Figure 5A). In the following tests, pHluorin fluorescence was quenched inside acidic vesicles, and they became fluorescent when the vesicles were exposed to the extracellular neutral or alkali solutions. The recycling and stabilization process could be detected as transient and dynamic changes in fluorescence at a basic extracellular pH (7.4). As a control group, application of vehicle (20% HPCD) had no significant impact on the fluorescence intensity of SEP-GluA2 or SEP-GluA1, and did not alter extracellular pH because of the buffering system in the bathing solution (Figure 5B,C). Similarly, CX614 treatment with a final dose of 47 μM resulted in no significant difference in SEP-GluA1 and SEP-GluA2 fluorescence compared with the beginning of the test (Figure 5D,E, P > 0.05). The presence of 47 μM LCX001 also exerted no impact on the fluorescence intensity of SEP-GluA1 (Figure 5D,E, P > 0.05). However, a significant and fast increase in SEP-GluA2 fluorescence was exhibited in response to application of LCX001 (final dose of 47 μM) for 6 min (Figure 5F,^∗P < 0.05). The fluorescence intensity data and the relative intensity to vehicle group showed that application of LCX001 caused significant enhancement of stabilization and accumulation of AMPARs containing GluA2 units on the surface of neurons (Figure 5D,F,^∗P < 0.05).

The previous studies demonstrate that GluA2(R) subunit can undergo special Q/R RNA editing, which effectively reduces calcium influx. The majority of GluA2 subunits ( > 95%) contain the arginine (R) residue, and the restriction of intracellular calcium concentration by GluA2(R) is deemed vital to normal neuronal function. LCX001 may mediate [Ca²⁺]_i and neuron function by means of increasing surface GluA2(R) expression. Real-time detection of [Ca²⁺]_i in Fluo-4 NW dye buffer-loaded primary cultured medullary neurons is conducted in the presence of glutamate following application of opioid drugs and AMPA modulators, and the variation of fluorescence intensity is considered to represent changes of intracellular calcium concentration ([Ca²⁺]_i). As shown in Figure 6A,B, in the control group, application of glutamate at a final dose of 18 μM stimulated a weak increase in [Ca²⁺]_i. At the time point of 3 min, the fluorescence intensity was significantly enhanced after treatment with fentanyl or TH-030418 (final dose of 42 μM). This indicated that both fentanyl and TH-030418 triggered a rapid rise in [Ca²⁺]_i in nearly all parts of the neuron soma (Figure 6C,F). The remarkable [Ca²⁺]_i bursts lasted for 7 min and did not recover back to resting levels (Figure 6C,F). Interestingly, the increased level of [Ca²⁺]_i in response to fentanyl and TH-030418 treatment was significantly blocked by the subsequent application of LCX001 (final dose of 40 μM) at the time point of 7 min (Figure 6A,B,E,H). Compared to the reducing effect of LCX001 on [Ca²⁺]_i, CX614 (final dose of 40 μM) continuously increased level of [Ca²⁺]_i (Figure 6A,B,D,G). These results revealed that LCX001 restrained calcium influx and mediated intracellular calcium concentration in response to opioid treatment.