2.1 Chemicals

Phencyclidine (PCP) was generously gifted by the Changchun Institute of Applied Chemistry, Chinese Academy of Science. LT-102 was synthesized in our laboratory, as outlined in our previous publication.²⁶ TAK-137 was prepared according to methods previously reported (patent WO2012020848). All other materials were procured from standard commercial sources.

2.2 Animals

We sourced specific pathogen-free (SPF) embryonic day 18 (E18) pregnant Sprague Dawley rats, and male C57BL/6J mice weighing 20–22 g (6–7 weeks old) from the Animal Laboratory Center of Sichuan University (Chengdu, China). Mice were accommodated in clear polypropylene cages, with five animals per cage, on corn cob bedding. These conditions were environmentally controlled at 22°C with a 12-h dark/light cycle and a relative humidity of 60%. All animals had free access to standard laboratory chow and water ad libitum. All experimental procedures were performed in accordance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health). The study protocols adhere to the ARRIVE guidelines and received approval from the Institutional Animal Care and Use Committee of Sichuan University (Approval No. 2020135A).

2.3 Primary culture of hippocampal neurons

Primary hippocampal neurons were prepared as described earlier.²⁴ Briefly, the hippocampi of rat embryos were isolated from Sprague Dawley rats on embryonic day 18 in ice-cold Hanks' Balanced Salt Solution and dispersed into single cells. Dissociated cells were plated on poly-D-lysine-coated 96-well plates (Corning, New York) at a density of 5 × 10⁴ cells/well for a Ca²⁺ influx assay, 1.5 × 10⁴ cells/well in a 35-mm dish for whole-cell voltage-clamp recordings, and 3 × 10⁵ cells per well in 12-well plates (Corning, USA) for western blot analysis. Cells were then cultured in Neurobasal Medium (Thermo Fisher Scientific, USA) supplemented with 1% B27 (Thermo Fisher Scientific, USA), 2 mM L-glutamine (Thermo Fisher Scientific, USA), 100 U/mL penicillin, and 100 μg/mL streptomycin (Sigma-Aldrich, USA). These cells were incubated for 5 days at 37°C in 5% CO₂ for the Ca²⁺ influx assay and western blot analysis.

2.4 Ca²⁺ influx assay using primary neurons

The Ca²⁺ influx assay was carried out as recommended by the manufacturer (Dojindo, Japan), in line with previously described procedures.²⁴ Briefly, cultured neurons were incubated with Fluo-4 AM fluorescent calcium indicator dye solution with 1.25 mM Probenecid (TCI, Japan) for 60 min at 37°C in 5% CO₂. This was followed by two washes with D-Hanks Balanced Salt Solution (Ca²⁺ free) to remove the extracellular Fluo-4 AM. The relative intracellular Ca²⁺ levels triggered by a compound were monitored using a fluorometric imaging plate reader (Molecular Devices SpectraMax i3, USA). The activity of each compound was defined as the fluorescence intensity integrated over the total measurement period, with 0% activity defined in the presence of only 0.1% DMSO and 100% activity in the presence of 5 μM s-AMPA and 10 μM TAK-137.

2.5 Western blot analysis

Following five days in culture, neuronal cells were exposed to compounds with or without s-AMPA (1 μM) and further incubated at 37°C for 24 h in a 5% CO₂ humidified incubator. Subsequently, total protein was harvested using Cell Lysis Buffer (Invent Biotechnologies; SD-001) in compliance with the manufacturer's guidelines. Electrophoresis was performed using AnyKD-PAGE gel (ATGene, Chongqing, China, Cat# ATG0046). Polyvinylidene fluoride membranes with transferred proteins were blocked with Protein Free Rapid Blocking Buffer (EpiZyme, Shanghai, Cat# PS108) for 15 min and left overnight at 4°C with primary antibodies: phospho-GluA1 (at Ser831, Cell Signaling Technology, USA, Cat# 75574S, 1:1000), GluA1 (Cell Signaling Technology, USA, Cat# 2983, 1:1000), brain-derived neurotrophic factor (BDNF; Abcam, UK, Cat# ab108319, 1:1000), and Beta-Actin (Proteintech, USA, Cat# 66009, 1:10,000), Membranes were incubated the next day with horseradish peroxidase-linked secondary antibodies (Cell Signaling Technology, Danvers, USA; polyclonal anti-rabbit IgG, Cat# 7074 or polyclonal anti-mouse IgG, Cat# 7076, 1:2000) for 1 h at room temperature and then exposed to Clarity Western ECL Substrate (Cat #1705061; Bio-Rad, USA) or Super ECL Detection Reagent (Yeasen, Shanghai, China, Cat# 36208ES76) with a ChemiDoc MP system (Bio-Rad, USA). Optical densities were analyzed using ImageJ software (Bethesda, Maryland, USA).

2.6 AMPA-evoked currents in rat primary hippocampal neurons

After 14 days of culture, we employed whole-cell voltage-clamp recordings on hippocampal neurons. We used patch electrodes with a tip resistance of 3–5 MΩ, filled with an intracellular solution. This solution was composed of 5 mM NaCl, 140 mM K-Gluconate, 1 mM MgCl₂•6H₂O, 0.1 mM CaCl₂•6H₂O, 1 mM EGTA, 10 mM HEPES, and 2 mM Mg-ATP and was adjusted to pH 7.2 with KOH. The extracellular solution contained 140 mM NaCl, 3.5 mM KCl, 1 mM MgCl₂•6H₂O, 2 mM CaCl₂•2H₂O, 10 mM D-glucose, 10 mM HEPES, and 1.25 mM NaH₂PO₄•2H₂O and was adjusted to pH 7.4 with NaOH. We maintained the neurons at a voltage clamp of 80 mV. Steady-state inward currents were prompted via the application of specific agonists and compounds through a carefully controlled perfusion system. We collected and evaluated the experimental data using the pClamp 10 software (Axon Instruments, Foster City, CA, USA). Current magnitudes were normalized to the currents stimulated by 10 μM s-AMPA.

2.7 hERG tail current in HEK293 cells

To record hERG tail currents, a patch-clamp assay was performed using a QPatch 48X automated electrophysiology platform (Sophion Bioscience, Ballerup, Denmark).²⁷ The extracellular solution was the same as the one used in section 2.6. The intracellular solution employed consisted of 20 mM KCl, 115 mM K-aspartate, 1 mM MgCl₂•6H₂O, 5 mM EGTA, 10 mM HEPES, and 2 mM Na₂-ATP. The pH of the intracellular solution was adjusted to 7.2 using KOH.

2.8 Kinase-inhibition assays

The in vitro kinase inhibition potency of LT-102 against a panel of 310 kinases was measured at 10 μmol/L, using the Homogeneous Time-Resolved Fluorescence (HTRF) method. This service was commercially provided by ICE Bioscience Co., Ltd., China. Kinase inhibition assays were performed based on previously described methods.²⁸ In brief, all reagents and samples were prepared, which included kinase substrate solution, ATP solution, kinase enzyme solution, and LT-102, into a 384-well microplate. After incubating this setup at 25°C for 1 h, allowing the kinase reaction to transpire, we terminated the reaction with a stop solution. We then added HTRF detection reagents, which include two specific antibodies marked with donor and acceptor fluorophores, into the wells. Another round of incubation allowed for the development of the HTRF signal. The fluorescence signal was subsequently measured using a compatible HTRF plate reader (BMG Labtech). The signal was expressed as the ratio of acceptor fluorescence (measured at 665 nm) to donor fluorescence (measured at 620 nm) and multiplied by 10,000 to enhance sensitivity.

2.9 Preparation of hippocampal slices

Mice were deeply anesthetized using isoflurane and subsequently decapitated. Mouse hippocampal slices (300 μm) were prepared using a vibratome (Leica, VT 1000 S, Germany). The slicing procedure was performed in ice-cold artificial cerebrospinal fluid that was oxygenated to maintain optimal conditions. The ACSF composition consisted of 125 mM NaCl, 2.5 mM KCl, 1.25 mM NaH₂PO₄, 25 mM NaHCO₃, 25 mM D-Glucose, 2 mM CaCl₂, and 1.5 mM MgCl₂. The ACSF solution was continuously bubbled with a mixture of 95% O₂ and 5% CO₂ to maintain a pH of 7.4. After slicing, the hippocampal slices were transferred to oxygenated ACSF at room temperature and recovered for 60 min before initiating further experiments.

2.10 LTP recordings

The protocol for recording field excitatory postsynaptic potentials (fEPSPs) was described previously.²⁹ Field recordings were conducted in the CA1 region of the hippocampus, and the Schaffer collateral fibers were stimulated using a bipolar electrode. To induce long-term potentiation (LTP), a theta-burst stimulation (TBS) protocol was employed. This protocol consisted of five trains of bursts, each containing four pulses at a frequency of 100 Hz. The stimulation was repeated four times with intervals of 10 s between each train. The magnitude of the LTP was calculated by comparing the average slopes of the fEPSPs during the last 10 min of a recording with the baseline slopes recorded before stimulation. Data acquisition and analysis were performed using pClamp 11 software (Axon Instruments, Foster City, CA, USA).

2.11 Miniature excitatory postsynaptic current recordings

To investigate miniature excitatory postsynaptic currents (mEPSCs), whole-cell patch-clamp recordings were conducted on mouse hippocampal slices following a previously established protocol.³⁰ The recordings were performed specifically on CA1 pyramidal neurons. The components of the intracellular solution included the following: 140 mM K-gluconate, 2 mM MgCl₂, 10 mM HEPES, 8 mM KCl, 2 mM Na₂-ATP, and 0.2 mM Na₂-GTP. The pH of the solution was adjusted to 7.2 using KOH. The extracellular solution, which bathed the hippocampal slices, consisted of 125 mM NaCl, 2.5 mM KCl, 1.25 mM NaH₂PO₄•2H₂O, 25 mM NaHCO₃, 10 mM D-Glucose, 2 mM CaCl₂•2H₂O, and 1.5 mM MgSO₄. This solution was equilibrated with a mixture of 95% O₂ and 5% CO₂ for at least 30 min, until its pH was 7.4. The resistance of the pipettes was 4–6 MΩ. Additionally, the extracellular solution contained antagonists: bicuculline (10 μM, Sigma-Aldrich, USA) to block GABA_A receptors, 2-amino-5-phosphonopentanoic acid (20 μM, Sigma-Aldrich, USA) to block NMDA receptors, and tetrodotoxin (TTX, 1 μM, Sigma-Aldrich, USA) to block action potentials. During the mEPSC recordings, CA1 pyramidal neurons were voltage-clamped at −70 mV without synaptic stimulation. The recordings were performed using pClamp11 software. Baseline recordings were obtained for 10 min before the administration of LT-102. Following LT-102 administration, recordings were performed for an additional 10 min.

2.12 AMPAR-mediated excitatory synaptic current recordings

The resistance of pipettes and the intracellular and the extracellular solution were the same as in the section 2.11, with the exception that TTX was not added to the extracellular solution. The AMPAR-mediated EPSC amplitude was measured as the peak EPSC amplitude at a holding potential of −70 mV. The paired-pulse ratio (PPR) was calculated as the ratio of the amplitude of the second response to the first at each interpulse interval (20, 50, 100, and 200 ms).

2.13 Drug administration

PCP was prepared by dissolving it in normal saline, and the solution was freshly prepared daily. LT-102 or normal saline was dissolved in a solution containing 0.1% DMSO and 0.4% Tween 80. All treatments were administered intraperitoneally (i.p.) at 10 mL/kg. The dosing regimen followed an established protocol.³¹ Forty mice, each averaging 8 weeks of age, were randomly assigned to four groups, with ten mice per group. To ensure consistency in behavioral outcomes, mice were selected from the same batch to mitigate bias from repeated testing. Age and weight matching was conducted, and random group allocation was performed prior to the experiment to bolster result reliability. The animals in the vehicle-treated group received saline (0.9% NaCl), while the PCP treatment group received PCP at 10 mg/kg twice daily for seven consecutive days. Following a 7-day wash-out period, the treatment group received a single dose of either 0.01 mg/kg or 0.05 mg/kg LT-102, while the control and PCP model groups received saline injection daily for 7 days. Behavioral testing was initiated at the conclusion of drug treatment.

2.14 Behavioral tests

2.15 Crystallization of hGluA2 with compounds

DNA encoding residues 413–795 of human GluA2 (hGluA2), with amino acids 528–652 replaced by the amino acid GT, was synthesized by Shanghai Sangon Biotech Inc. The synthesized DNA was then cloned into a PET28a expression vector containing an N-terminal His-MBP tag followed by a TEV protease site. The expression vector was introduced into Escherichia coli BL21(DE3) competent cells. The transformed cells were allowed to undergo overnight expression at 16°C. After the expression period, the cells were harvested and disrupted by sonication in a lysis buffer containing 20 mM Tris–HCl, 250 mM NaCl, and pH 8.0. The cell lysate was centrifuged at 12,000 × g for 30 min at 4°C. The protein of interest, tagged with the His-MBP fusion, was purified using NTA affinity beads (GE Healthcare, Little Chalfont, UK). Subsequently, TEV protease treatment was employed to remove the N-terminal His-MBP tag. To exclude the TEV protease and the released His-MBP tag, additional NTA affinity beads (GE Healthcare) and MBP resin (GE Healthcare) were utilized. Finally, the tag-free hGluA2 was purified through size-exclusion chromatography utilizing a Superdex 200 column (GE Healthcare). The final buffer for this purification step contained 10 mM HEPES, pH 7.0, 20 mM NaCl, and 1 mM EDTA.

The purified hGluA2 protein was concentrated at 15 mg/mL. It was then incubated with compound LT-102 at a molar ratio of 1:5 on ice for 0.5 h to allow binding. Crystals of the hGluA2/LT-102 complex were obtained using the vapor diffusion hanging drop method, with the reservoir buffer consisting of 22% (w/v) PEG8000, 0.1 M NaAC at pH 5.0, and 0.2 M (NH₄)₂SO₄. A cryoprotectant solution containing 20% glycerol was added to the reservoir solution. The crystals were then flash-frozen in liquid nitrogen. Diffraction data for the hGluA2/LT-102 complex were collected at the beamline BL19U1 of the Shanghai Synchrotron Radiation Facility at 100 K. The collected data were processed using XDS software to obtain the diffraction intensities. To determine the structure of the hGluA2/LT-102 complex, a molecular replacement was performed using Phaser. A previously reported structure with the PDB code 2XHD was used as the search model.³⁴ Model building was carried out using Coot, and the complex structure was refined using PHENIX.

2.16 Statistical analysis

All data were analyzed using Prism 8 (GraphPad Software Inc., San Diego, CA). The results were expressed as the mean ± standard error of the mean (SEM). Nonlinear regression analyses were performed to generate dose–response curves and calculate EC₅₀ values. They evaluated the normality of the data distribution using the Shapiro–Wilk test. To determine differences among groups, we employed either one-way or two-way ANOVA with Tukey's post hoc test for parametric data, or Kruskal-Wallis test for nonparametric data, followed by Dunn's post hoc test. Student's t-test was used to compare differences between two groups. Statistical significance was defined as a p-value less than 0.05 (p < 0.05).

3.1 LT-102 is an AMPAR potentiator with little agonistic effect

To identify novel AMPAR potentiators with reduced agonistic effects, we performed compound screening using a self-designed chemical library, using TAK-137 as a reference compound. Through a Ca²⁺ influx assay performed on primary rat hippocampal neurons, we identified a novel AMPAR potentiator called LT-102 (Figure 1A).²⁶ TAK-137 has previously been reported as an efficient and selective AMPAR potentiate with minimal agonistic effect on primary neurons.²⁵ Therefore, we compared the effects of TAK-137 and LT-102 using a Ca²⁺ influx assay. The EC₅₀ and E_max values of LT-102 in the presence of an agonist s-AMPA (5 μM) were determined to be 0.169 μM and 129.6%, respectively, while those of the positive control TAK-137 were 1.323 μM and 106.8%, respectively (Figure 1B). Similar to TAK-137, LT-102 exhibited minimal agonistic effect in the absence of s-AMPA (Figure 1B). Furthermore, pretreatment with the competitive AMPAR antagonist NBQX significantly suppressed the agonist-dependent activity of LT-102, indicating that LT-102 acted on AMPARs (Figure 1C). X-ray crystallography was employed to investigate further the binging interactions of LT-102 with the ligand-binding domain (LBD) of the AMPAR subunit GluA2. X-ray data collection and refinement statistics are listed in Table S1. The co-crystal structure revealed that the LBD of GluA2 adopts a classic LBD fold and that LT-102 binds to a hydrophobic pocket formed by LBD residues, similar to other AMPAR potentiators (Figure 1D). LT-102 was observed to bind to allosteric sites on the LBD, forming two hydrogen bonds with residues Gly 219 of S1 and Pro 105 of S2 in the LBD (Figure 1E). The two hydrogen bonds between LT-102 and the LBD likely contributed to the high potentiation of LT-102, leading to increased AMPAR activity. These findings demonstrate that LT-102 is an AMPAR potentiator with minimal agonistic effect and exhibits higher potency and efficacy.

3.2 LT-102 selectively potentiates AMPA-evoked currents in primary rat hippocampal neurons

To further evaluate the efficacy and selectivity of LT-102, we performed whole-cell patch-clamp recordings of primary rat hippocampal neurons. LT-102 treatment significantly amplified AMPA-evoked currents in a dose-dependent manner in the presence of 10 μM s-AMPA while exhibiting virtually no agonistic activity in the absence of s-AMPA. The half-maximal effective concentration (EC₅₀) for LT-102 was determined to be 8.558 μM, and the maximal response (E_max) values represented an 813.8-fold augmentation in the presence of 10 μM s-AMPA (Figure 2A). Remarkably, LT-102 at a concentration of 10 μM considerably amplified the potency of s-AMPA, which led to a decline in its EC₅₀ from 11.357 to 4.375 μM (Figure 2B). We further investigated the selectivity of LT-102 and found that AMPA-induced currents were almost entirely inhibited in the presence of 10 μM NBQX (Figure 2C). Given the often coupled and interdependent functions of AMPAR and NMDAR, we examined the effects of LT-102 on NMDA-mediated currents in the presence of NMDA (30 μM) and its co-agonist glycine (1 μM). Delightfully, LT-102 at 10 μM or 30 μM had no discernible enhancing or inhibiting effect on NMDA-activated currents (Figure 2D). Additionally, patch-clamp recordings demonstrated that LT-102 treatment did not impact the hERG tail current in hERG-HEK293 cells, implying a low risk of cardiotoxicity (IC₅₀ > 30 μM, Figure 2E). To extend our understanding of LT-102's selectivity, we examined its effects against a panel of 310 kinases. The results revealed that LT-102 exerted less than 50% inhibitory activities against all tested kinases, indicating its high specificity (Figure S1; Table S2). These findings strongly suggest that LT-102 is a highly selective AMPAR potentiator.

3.3 LT-102 increased BDNF expression and GluA1 phosphorylation in an agonist-dependent manner

Given the crucial role of BDNF and GluA1 phosphorylation in synaptic plasticity modulation,³⁵ we examined the impact of LT-102 treatment on these factors in primary rat hippocampal neurons. The results revealed that LT-102 treatment at 10 or 30 nM significantly enhanced BDNF protein expression and GluA1 phosphorylation in the presence of 1 μM s-AMPA (Figure 3A). Conversely, no significant changes were observed in the absence of s-AMPA (Figure 3B). Additionally, pretreatment with 10 μM NBQX effectively inhibited the LT-102-induced BDNF expression and GluA1 phosphorylation in the presence of 1 μM s-AMPA (Figure 3C). These findings demonstrate that LT-102 treatment promotes BDNF expression and GluA1 phosphorylation through the activation of AMPAR activation in an agonist-dependent manner in vitro.

3.4 LT-102 treatment enhanced long-term potentiation (LTP) at CA1 hippocampal synapses

To evaluate the effect of LT-102 treatment on LTP in mouse hippocampal slices, field excitatory postsynaptic potentials (fEPSPs) were measured in CA1 pyramidal neurons through stimulation of the Schaffer collaterals. The results demonstrated a significant enhancement of the average fEPSP slope during the last 10 min following theta burst stimulation (TBS) compared to the vehicle control (Figure 4A,B). Additionally, the impact of LT-102 treatment on synaptic transmission's intrinsic properties was investigated. The findings revealed a substantial increase in mEPSC frequency following LT-102 administration (Figure 4C–E), while mEPSC amplitude remained unaffected (Figure 4F,G). Furthermore, the evoked EPSC (eEPSC) amplitude was significantly elevated (Figure 4H). To evaluate the potential contribution of presynaptic neurotransmitter release to the improvement of LTP induced by LT-102, paired-pulse facilitation (PPF) was assessed. PPF was significantly reduced from 2.13 ± 0.31 to 1.16 ± 0.07 at an interstimulus interval (ISI) of 20 ms (Figure 4I,J). Nevertheless, we found no significant difference at ISIs of 50–200 ms (Figure 4I). These data suggested that the facilitated effects of LT-102 treatment on LTP may be attributed to an AMPAR-mediated increase in presynaptic transmitter release.

3.5 LT-102 treatment ameliorates PCP-induced spatial memory deficits in mice

To investigate the effect of LT-102 treatment on schizophrenia-associated spatial learning and memory dysfunction, we established a PCP-induced mouse model based on previous protocols.^{31, 36} The Morris water maze and novel object recognition tests were used to gauge the efficacy of LT-102 treatment. The dosing regimens are depicted in Figure 5A. Overall, the latencies to reach the platform decreased in all groups over the five training days in the Morris water maze (Figure 5B). Compared with the vehicle-treated mice, PCP-treated mice required significantly more time to reach the platform, and LT-102-treated mice found the platform faster from day 3 onwards (Figure 5B). In the probe test, the number of virtual platform crossings (Figure 5C) and time spent in the target quadrant (Figure 5D) were significantly lower for PCP-treated mice compared to vehicle-treated mice, indicating spatial learning and memory function deficits. This impairment was dose-dependently mitigated by LT-102 treatment (Figure 5C,D). Representative swimming traces from the probe trials are depicted in Figure 5F. Swimming speed did not differ significantly between the groups (Figure 5E). We further explored the effects of LT-102 treatment on recognition memory using the novel object recognition test. During the habituation trial, we found no differences in exploratory preferences between the groups (Figure 5G). However, during the novelty recognition trial, PCP-treated mice showed an apparent reduction in the discrimination index compared to the control, suggesting impaired recognition memory (Figure 5H). LT-102-treated mice demonstrated a significantly higher discrimination index in a dose-dependent manner (Figure 5H). These results suggest that LT-102 treatment dose-dependently counteracts the spatial and recognition memory impairments observed in PCP-treated mice. We also observed a significant decrease in the basal expression level of BDNF and the phosphorylation level of GluA1 in the hippocampus of PCP-treated mice compared to the control group (Figure 5I). Interestingly, LT-102 administration effectively reversed these reductions, suggesting that BDNF and the GluA1 subunit play crucial roles in mediating the treatment effect of LT-102.