Huperzine A attenuates epileptic seizures via enhancing dCA1-projecting septal cholinergic transmission

Abstract

Cholinergic transmission, independent of classical glutamatergic and GABAergic signaling, critically plays a crucial role in epilepsy. Huperzine A (Hup A), an acetylcholinesterase (AChE) inhibitor, exerts potent anticonvulsant activity, but its mechanism of action within cholinergic circuits remains unclear. Here, we show that Hup A mitigates epileptic seizures by enhancing hippocampal dorsal CA1 (dCA1)-projecting cholinergic transmission. We found that systemic injection of Hup A not only reduces seizures in acute models, including the maximal-electroshock seizure (MES), pentylenetetrazol (PTZ), and kainic acid (KA) models but also alleviates the seizure severity in chronic epilepsy models induced by kindling and KA, indicating a broad-spectrum anti-seizure efficacy. Interestingly, using immunohistochemistry, viral tracing, and in vivo fiber photometry, we found that Hup A selectively inhibits AChE in the dCA1 rather than in other hippocampal subregions or cortex, enhancing dCA1-projecting septal cholinergic transmission. Significantly, selective ablation of septal ChAT⁺ neurons reversed the anti-seizure effects of Hup A. We further identified that α7 nicotinic acetylcholine receptors in the dCA1 region mediate the anti-seizures cholinergic circuit modulated by Hup A. Together, our results demonstrate that Hup A exerts broad-spectrum anti-seizure efficacy via modulating dCA1-projecting septal cholinergic transmission, providing potential therapeutic avenues for epilepsy through targeted cholinergic modulation.

Introduction

Epilepsy is a circuit-level brain disorder, characterized by recurrent seizures resulting from abnormal synchronous neuronal activity, affecting nearly 1% of the global population [1]. Antiseizure medications (ASMs) remain the primary therapeutic strategy for epilepsy, mainly by suppressing glutamatergic neurotransmission or enhancing GABAergic neurotransmission. However, ASMs typically reduce seizures rather than prevent epileptogenesis, and around 30% of patients develop drug resistance [2, 3]. Temporal lobe epilepsy (TLE) is the most common form of drug-resistant epilepsy. As such, exploring for drugs that target alternative pathways, beyond traditional glutamatergic and GABAergic signaling, is critical to advancing in TLE treatments.

Huperzine A (Hup A), a natural sesquiterpene alkaloid derived from Huperzia serrata (Thunb. ex Murray) Trevis., is a potent and highly selective reversible inhibitor mainly targeting acetylcholinesterase (AChE). In China, Hup A is used to treat Alzheimer’s disease [4], and in the United States, it is approved as a nutritional supplement for cognition enhancement [5]. Preclinical studies suggest that Hup A exerts potent anticonvulsant activity in animal models [6, 7], especially showing robust protection against seizures in Scn8a^R1620L mutants, Scn1a^+/− mouse models of Dravet syndrome (a severe developmental and epileptic encephalopathy characterized by early-onset febrile and afebrile seizures as defined by the ILAE) and Scn1a^R1648H/+ mouse model of generalized epilepsy with febrile seizures plus [7–9], suggesting a favorable safety profile for continuing clinical development. However, whether Hup A exhibits broad-spectrum anti-seizure properties in other epilepsy models, particularly in TLE, has not yet been fully elucidated. More importantly, the cholinergic network contributes to epilepsy in a circuitry heterogeneous manner [10–13]. We previously demonstrated that activation of cholinergic neurons in the medial septum (MS), which directly project to the dorsal CA1 (dCA1) of the hippocampus alleviated TLE effectively [14, 15], while other cholinergic circuits have little effect or even exacerbate seizures [12]. As an AChE inhibitor, the anti-seizure mechanism of Hup A within cholinergic circuits has still not been revealed.

Here, we used a combination of genetic and pharmacological approaches, along with immunohistochemistry, viral tracing, and in vivo calcium fiber photometry, to demonstrate that Hup A exerts broad-spectrum anti-seizure effects by enhancing dCA1-projecting septal cholinergic transmission. Our results provide a potential therapeutic strategy for epilepsy targeting cholinergic transmission and underscore the therapeutic significance of Hup A in treating TLE.

Materials and methods

Animals

Male (8–16 weeks of age) C57BL/6 (WT, RRID: IMSR_JAX:000664, SLAC or Zhejiang) and Cholineacetyl-transferase (ChAT) Cre-recombinase mice (ChAT-Cre, male, strain no. 006410, RRID: IMSR_JAX:006410) were used. All mice were housed in a specific-pathogen-free facility (five mice per cage) at 22 ± 2 °C and 50% ± 10% relative humidity under a 12-h light-dark cycle (lights on from 08:30 to 20:30) with food and water available ad libitum. All behavior experiments were performed between 8:30 a.m. and 5:30 p.m. The use and care of the mice followed the standards of Institutional Animal Care and were approved by the ethical committee of Zhejiang Chinese Medical University.

Drugs

Hup A injection ((−)-Hup A, Hup A, Cat# 1N220901, Wepon, Zhejiang, China), its enantiomer (+)-Hup A (Wepon, Zhejiang, China), pentylenetetrazole (PTZ, Cat# P6500, Sigma, SL, USA), and kainic acid (KA, Cat# ab120100, Abcam, Cambridge, UK) were dissolved in 0.9% sterile saline. 0.1 mg/kg, 0.3 mg/kg, and 1.0 mg/kg Hup A and (+)-Hup A was injected intraperitoneally for the efficacy assessments, 100 nM Hup A or 20 μM acetylcholine chloride (ACh, Cat# HY-B0282, MedChemExpress, NJ, USA) were locally injected into the dorsal CA1 (dCA1) (250 nL) to investigate the underlying mechanisms.

To assess the release of ACh in the dCA1 following cholinergic receptors blockade, atropine sulfate monohydrate (Atr, 25 mg/kg, Cat# A0257, Sigma-Aldrich, SL, USA), mecamylamine hydrochloride (MEC, 3 mg/kg, Cat# M9020, Sigma-Aldrich, SL, USA), and methyllycaconitine citrate (MLA, 3 mg/kg, Cat# ab120072, Abcam, Cambridge, UK) were injected intraperitoneally. For in situ injection of cholinergic receptor antagonists, 250 nL MLA (20 μM) dissolved in saline, was administered to the dCA1 through the cannula. Vehicle groups received an equivalent volume of saline either intraperitoneally or in situ.

Virus vectors

AAV2/2Retro-hEF1a-DIO-mCherry-WPRE-pA (viral titers: 1.69E + 13 VG/mL) and AAV2/9-hSyn-FLEX-ChrimsonR-mCherry (viral titers: 1.90E + 13 VG/mL) were purchased from Taitool Bioscience Co., Ltd (Shanghai, China). AAV2/9-hSyn-ACh3.0 (viral titers: 5.68E + 12 VG/mL) and AAV2/9-EF1a-flex-taCasp3-TEVP (viral titers: 5.62E + 12 VG/mL) were purchased from BrainVTA Co., Ltd (Wuhan, China). All viral vectors were aliquoted and stored at −80 °C until use.

Stereotaxic surgery

Surgical procedures in this study were aligned with our previous study [16]. Mice were deeply anesthetized with isoflurane (3% for induction and 1.5% for maintenance) and head-fixed in a stereotaxic frame (68043, RWD Life Science Co., Ltd, Shenzhen, China). The electrodes (791500, each 0.127 mm in diameter, 0.5 mm tip distance; A-M Systems, WA, USA) were implanted in the right ventral hippocampus CA3 (vCA3, AP −2.9 mm, ML −3.2 mm, DV −3.2 mm) for kindling and EEGs recording. Two screws were placed over the cortex and cerebellum to record EEGs and serve as the reference electrodes, respectively. Another two screws were placed over the right and left brain to fix and affixed with dental cement.

For virus delivery and optic fiber or cannula implantation, the following coordinates were measured from Bregma according to the brain atlas of mice: MS: AP + 1.1 mm, ML 0.0 mm, DV − 4.3 mm; right dCA1: AP − 2.0 mm, ML − 1.3 mm, DV − 1.6 mm. The virus was injected into each site via a 1-μL microliter syringe (Gaoge Industrial and Trading Co. Ltd, Shanghai, China) attached to an ultra micro pump (Micro 4, World Precision Instruments, USA) at a rate of 100 nL/min. The injection needle was withdrawn 5 min after the end of the injection. The virus was allowed to express for a minimum of 4 weeks for sufficient expression in both the somata and axons. Only data from mice with correct injection sites were included.

For dCA1-projecting cholinergic neuronal population input mapping, 250 nL AAV2/2Retro-hEF1a-DIO-mCherry was injected into the dCA1 of ChAT-Cre mice. Mice were sacrificed after 2 weeks of virus expression.

Maximal-electroshock seizure (MES) model

The MES model was performed as our previous study [17]. We clamped the ears of mice through a crocodile clip with saline, and then a stimulus current was delivered after 1 h of drug intervention. Electroshock seizures (maximal output voltage 750 V, 50 Hz sine wave pulse, 0.2 s stimulus duration) were generated via rodent shocker (HSE-HA, Hugo Sachs Elektronik, Freiburg, Germany). The initial stimulus current intensity was 6 mA and raised with 1 mA every 1 min. The threshold and the duration of tonic-clonic seizures were analyzed.

Pentylenetetrazol (PTZ)-induced seizure model

The PTZ model was performed as in our previous study [18]. One hour after drug injection, mice were given 100 mg/kg PTZ to induce acute seizure, and then were observed for the occurrence of seizure activity for 30 min after PTZ administration. The latency to stage 2 (nodding), stage 4 (bilateral forelimb clonus and rearing), and stage 6 (stiffened) was analyzed.

Hippocampal kindling

The kindling model was constructed as indicated in our previous study [19]. One week after surgery recovery, mice received kindling stimulations ten times daily (400 μA, 20 Hz, 2 s trains, 1 ms monophasic square wave pulses) with 30 min intervals continuously for 3 to 4 days until the mice were fully kindled, the drug was applied once daily before the first kindling stimulation. EEGs of the right ventral hippocampal CA3 (vCA3) and cortex were recorded by a digital amplifier (NuAmps, Neuroscan System, VA, USA). Behavioral seizure stages were scored according to the Racine scale (1972): stage 1 (ear and facial movement), stage 2 (nodding), stage 3 (unilateral forelimb clonus), stage 4 (bilateral forelimb clonus and rearing), stage 5 (falling and jumping with bilateral forelimb clonus). Stages 4–5 were defined as secondary generalized seizures (sGS). We identified the kindled states of mice when they underwent stage 5 three times.

KA-induced acute seizure model

A cannula was implanted into the right dCA1 for KA delivery, and an electrode was implanted into the right vCA3 for EEG recording. One week after surgery, mice were monitored via EEG for 10 min as the baseline. Then, KA (0.5 ng/nL, 500 nL) was given via a cannula into dCA1 over 5 min after 1 h of drug intervention, and the mice behaviors were observed in the next 90 min synchronized with EEG recording by a PowerLab system (AD Instruments, Australia) at a sample rate of 1 kHz. Behavioral seizure stages were scored according to the modified Racine scale: stage 1 (facial movement), stage 2 (head nodding), stage 3 (unilateral forelimb clonus), stage 4 (bilateral forelimb clonus and rearing), stage 5 (rearing and falling), and stage 6 (generalized tonic-clonic seizures or death). Stages 1–3 were defined as focal seizure (FS), and stages 4–6 were considered to be sGSs. The seizure stage, latency to FS, and the number of GS were analyzed.

KA-induced chronic TLE model

KA-induced chronic TLE model was performed as in our previous studies [20]. Briefly, KA (0.6 ng/nL, 500 nL) was injected stereotactically into the dCA1 of WT mice. After 2 months of injection (spontaneous seizure period), an electrode was implanted in the right vCA3 of KA mice to record EEGs. Then, EEGs were continuously recorded in KA mice for 8 h daily after 7 days for surgery recovery. The spontaneous seizure event was defined as regular spike clusters with a duration of ≥10 s, spike frequency of ≥3 Hz, and at least triple amplitude to the baseline EEG, consistent with previous research. The number and total duration of seizure activity were analyzed.

Rota-rod test

Motor performance was evaluated using an accelerating Rota-rod (LE8205, Panlab Harvard Apparatus, Barcelona, Spain). One hour following Hup A treatment, mice were placed on the accelerating rod, which accelerated from 5 to 40 rpm over a period of 300 s. The maximum latency before falling and the speed at the time of the fall were recorded for each mouse.

Fiber photometry

Fiber photometry recording was performed as in our previous study [21]. The fiber photometry system (QAXK-FPS-TC-MC-LED, ThinkerTech, Nanjing, China) emitted 488 nm excitation light coupled into a 200 μm optical fiber, and the light reaching the corresponding area excited green fluorescence signals. The GFP fluorescence signals were collected by a photomultiplier tube, then converted the current output to voltage signals through an I-V amplifier, further filtered by a low-pass filter (40 Hz cut-off). The 470 nm and 405 nm lights were alternatively given during recording, in which the 405 nm was used as the internal reference control. The signaling that was responded to was quantification to heatmap and peri-event. The baseline average fluorescence change was denoted as F₀. ΔF/F₀ was calculated by the fluorescence change (F-F₀)/F₀ value.

For optogenetic light delivery in mice expressing ChrimsonR in the MS, a subregion of septum, a 635-nm (20 Hz, 10 ms/pulse, 5 mW) light laser (QAXK-LASER-R1, ThinkerTech, Nanjing, China) was delivered immediately after drug injection for 30 min.

The event started with electrical or light stimulation. For recording the ACh probe fluorescence signals after a single pulse stimulation in MS, “Pre” was defined as the last 5 s baseline, and “Post” was defined as the post-stimulation activity for 5 s. For recording the ACh probe signals during light stimulation, “Pre” was defined as the last 20 s baseline, and “Light” was defined as the post-stimulation activity for 30 s.

Measurement of AChE activity

The deeply anesthetized mice were perfused with 0.1 M phosphate-buffered saline (PBS) and separated the hippocampus and cortex regions. Tissues were weighed and homogenized by PBS, then samples were centrifuged at 1000 × g for 10 min at 4 °C, and supernatants were aliquoted and stored at −20 °C until quantification. A mice AChE ELISA kit (MB-5926A, Meibiao Biotechnology Co., Ltd, Jiangsu, China) was used according to the manufacturer’s instructions.

Immunohistochemistry and imaging

Mice were deeply anesthetized with pentobarbital sodium and were perfused with cold PBS, followed by cold 4% (w/v) paraformaldehyde (PFA) in PBS. The whole brain was dissected and post-fixed for 4 h in 4% PFA at 4 °C, then transferred to 30% (w/v) sucrose until up to the bottom. Coronal brain sections (40 μm) were frozen in optimum cutting temperature (OCT) compound and cut with a cryostar NX70 (Thermo Fisher Scientific, CA, USA), then floated in PBS. At room temperature, sections were blocked with 5% donkey serum/0.3% Triton X-100/PBS for 2 h. Free-floating sections were then incubated overnight at 4 °C with the following primary antibodies: anti-ChAT (Cat# AB144P, RRID: AB_2079751, Millipore, MA, USA) and AChE (Cat# ab2803, RRID: AB_303316, Abcam, Cambridge, UK). Sections were then washed with PBS three times and incubated with Alexa Fluor 647-conjugated donkey anti-goat IgG (Cat# ab150131, RRID: AB_2732857, Abcam, Cambridge, UK) or Alexa Fluor 488-conjugated donkey anti-mouse IgG (Cat# ab150105, RRID: AB_2732856, Abcam, Cambridge, UK) for 2 h at room temperature in the dark. DAPI (Cat# P0131, Beyotime, Shanghai, China) was used to stain the nucleus. A Leica TCS SP8 confocal laser scanning microscope (Leica Microsystems, Mannheim, Germany) captured the image using LAS X software (Version 3.7.1).

Quantification and statistical analysis

Data were presented as means ± SEM. The number of experimental replicates (n) was indicated in each figure legend. Statistical comparisons were performed using GraphPad Prism (version 10.1.2, GraphPad Software, CA, USA) with appropriate methods indicated in figure legends. Photometry data were analyzed with MATLAB R2020b (The MathWorks Inc., MA, USA). Probability (P) values less than 0.05 were considered statistically significant, and significance levels are presented as *P < 0.05, **P < 0.01, ***P < 0.001, or ****P < 0.0001 in the manuscript and all figures.

Results

Hup A attenuates epileptic seizures in various acute seizure models

To determine whether Hup A influenced acute seizure activity, we first used a MES model, which clinically resembles generalized tonic-clonic seizures and is commonly used to test anti-seizure effects [17]. We found Hup A dose-dependently increased threshold and reduced seizure durations in the MES-evoked tonic-clonic seizure behaviors (Fig. 1b, c). Mainly, 0.3 mg/kg Hup A exhibits the most favorable performance (threshold: P < 0.001; durations: P < 0.0001) compared to the vehicle group, and 0.1 mg/kg (threshold: P < 0.001; durations: P < 0.01), 1.0 mg/kg (durations: P < 0. 01) Hup A also mitigated tonic-clonic seizure in a degree. Hup A at all tested dosages lowered the mortality rate in MES-induced seizure model (Fig. S1a). Next, we addressed whether Hup A provides similar protection in the PTZ seizure model, which represents acute generalized myoclonic seizure behaviors [22]. We found that a single administration of Hup A dose-dependently alleviated the seizure activity induced by PTZ to varying degrees. There was a moderate increased latency to stage 2 from Hup A-treated mice compared to the vehicle (Fig. 1b, P < 0.05), while 1.0 mg/kg Hup A significantly prolonged the latency to stage 2 (Fig. 1b, P < 0.05) and stage 6 (Fig. 1f, P < 0.01). In addition, Hup A slightly lowered mortality rate in PTZ-induced seizure model (Fig. S1b). These findings suggest that systemic administration of Hup A holds potential as an effective treatment of acute seizures.

Fig. 1. Hup A attenuates epileptic seizures in various acute seizure models.

a Chemical structure of Hup A. Effects of Hup A on the seizure threshold (b) and duration (c) in the MES test. n = 11. d–f Effects of Hup A on the latency to stage 2 (d) stage 4 (e) and stage 6 (f) in the PTZ test. n = 10. Effects of Hup A on seizure stage (g) latency to focal seizure (FS) (h) number of generalized seizures (GS) (i) and power ratio (j) in the acute KA test. n = 8, 9. Veh, vehicle. One-way ANOVA with post hoc Dunnett’s test for (b–h) ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, and ^****P < 0.0001 compared with “Veh” group. Two-way repeated measures ANOVA with post hoc Bonferroni’s test for (j) ^*P < 0.05, ^****P < 0.0001 compared to the “Vehicle” group. k Representative seizure EEGs were recorded from CA3 after the Hup A treatment.

Considering the presence of enantiomers in Hup A (Fig. S2a), which may exhibit distinct pharmacological effects, we also investigated the effect of (+)-Hup A on acute seizure. In sharp contrast, (+)-Hup A exacerbated MES-induced tonic-clonic seizures, manifested as a decreased threshold (Fig. S2b, 0.1 mg/kg: P < 0.05) and a slightly increased seizure duration (not significantly, Fig. S2c) to MES. In the PTZ model, only 0.1 mg/kg (+)-Hup A expended the latency to stage 2 (Fig. S2d, P < 0.05), 4 (Fig. S2e, P < 0.05) and 6 (Fig. S2f, P < 0.05). Accordingly, these results suggest that both enantiomers of Hup A display anti-seizure activity, among which Hup A shows the most robust effect in acute seizure models.

Finally, we assessed the acute antiepileptic actions of Hup A in acute hippocampal seizures. Treatment with 0.1 mg/kg Hup A significantly delayed the latency to FS (Fig. 1h, P < 0.05) and the power ratio (Fig. 1j, P < 0.05) during the 90 min following KA injection. Notably, declined seizure stage (Fig. 1g, P < 0.001), number of generalized seizures (GS) (Fig. 1i, P < 0.0001), and relative power ratio (Fig. 1j, P < 0.0001) from Hup A-treated mice were observed in comparison with the vehicle at 1.0 mg/kg. Typical hippocampal EEGs for the vehicle and 0.1 mg/kg Hup A groups were shown in Fig. 1k. To assess potential sedative effects and impacts on motor balance, we then performed the Rota-rod test. We found that 0.1 mg/kg Hup A did not impair motor activity, as evidenced by increased motor speed and latency to fall off. In contrast, treatment with 1.0 mg/kg Hup A significantly decreased motor speed and latency, indicating a potential sedative effect (Fig. S3). Based on these findings, we selected 0.1 mg/kg Hup A for subsequent experiments. These results suggested that Hup A exerts dose-dependent effects in mitigating acute seizure activity across different seizure models, with optimal efficacy observed among its enantiomers.

Hup A attenuates seizure in chronic epilepsy models

Next, we explored the impact of Hup A on the hippocampal seizure severity in the hippocampal kindling model. As shown in Fig. 2, treatment with 0.1 mg/kg Hup A delayed the development of the behavioral seizure stage (Fig. 2d, F (49, 686) = 1.654, P < 0.01) and shortened after-discharge durations (ADDs) (Fig. 2e, F (49, 686) = 1.969, P < 0.001) during kindling acquisition, compared with the vehicle group. Furthermore, in fully kindled mice (with stable sGS), 0.1 mg/kg Hup A inhibited the sGS severity, as indicated by a decreased seizure stage (Fig. 2f, P < 0.01), shortened ADDs (Fig. 2g, P < 0.001) and GSDs (Fig. 2h, P < 0.001). Typical hippocampal EEGs during the kindled seizure and corresponding power spectra are shown in Fig. 2i. These results suggest that Hup A robustly ameliorated the hippocampal seizure severity in the kindling-induced epilepsy model.

Fig. 2. Hup A ameliorates the development of hippocampal seizures.

Schematic paradigm (a) and illustration (b) of the kindling-induced seizure model. c Representative fluorescence image stained with DAPI. White dotted lines indicate electrode traces. Scale bar, 100 μm. Effects of Hup A on the development of seizure stage (d) and ADDs (e) during kindling acquisition. Two-way repeated measures ANOVA with post hoc Bonferroni’s test, ^**P < 0.01, ^***P < 0.001 compared to the “Vehicle” group. Effects of Hup A on the seizure stage (f), ADDs (g), and GSDs (h) in the fully kindled state. n = 8, Mann–Whitney test for (f–g) and Unpaired t-test for (h) ^**P < 0.01, ^***P < 0.001 compared to the “Veh” group. Veh vehicle. i Typical GS EEGs recorded from the cortex.

To further elucidate the effects of Hup A on chronic hippocampal seizures, we adopted an intrahippocampal KA-induced TLE model, which closely resembles the clinical chronic spontaneous seizure syndrome and is often resistant to various ASMs [23]. In this model, chronic spontaneous seizures gradually appeared ~2 weeks after KA injection and persisted for months. The experimental schedule was listed as following: saline-treated 3 days for the “Pre” stage, 0.1 mg/kg Hup A-treated 3 days for intervention (once daily), then Hup A-withdrew and saline-treated 3 days again for the “Post” stage, along with the 8 h hippocampal EEGs recording (Fig. 3a). We found that Hup A remarkably attenuated spontaneous seizure during the 3-day intervention, reflected by reduced seizure number (Fig. 3b, d, P < 0.001) and duration (Fig. 3c, e, P < 0.01). In contrast, the seizure severity returned to baseline after Hup A withdrawal (Fig. 3b–e). Typical hippocampal EEGs during the different stages are shown in Fig. 3f. Overall, Hup A effectively protects against hippocampal seizures in a chronic TLE model.

Fig. 3. Hup A relieves the chronic spontaneous seizures in KA-induced TLE.

a Experimental paradigm in KA-induced chronic TLE mice. Hup A (0.1 mg/kg) was intraperitoneally injected. Total number (b) and duration (c) of seizures per 8 h during different periods. n = 10, Paired t-test. ^***P < 0.001 compared with “Pre” group, ^##P < 0.01 compared with “Hup A” group. d Heatmap of the number of seizures per 8 h recorded daily in each mouse. e Statistic analysis of the total duration of seizures detected on each day. f Typical seizure EEGs recorded from vCA3.

Hup A increases the MS-dCA1 cholinergic transmission

Previous studies have reported that AChE is expressed in the various brains of healthy rodents [24]. Hup A, an AChE inhibitor, has been shown to reduce both the expression and activity of AChE. To correlate the anti-seizure mechanisms of Hup A with cholinergic signaling, we further investigated the characteristics of Hup A on AChE in the epilepsy model. First, we evaluated the concentration of AChE in the hippocampus and cortex of kindled mice and found that 0.1 mg/kg Hup A administration selectively decreased the hippocampal AChE level (Fig. 4a, P < 0.05) without altering the content within the cortex (Fig. S4a). Next, we used immunofluorescence approach to examine the distribution of AChE in the hippocampus of kindled mice. Intriguingly, AChE was highly expressed in all subregions of the hippocampus and cortex (Figs. 4b and S4b). However, 0.1 mg/kg Hup A selectively reduced AChE expression in the dCA1 region (P < 0.01), while no significant changes were observed in the dorsal dentate gyrus (dDG), dCA3 or cortex regions (Figs. 4c and S4c), suggesting the anti-seizure mechanisms of Hup A may be linked to the dCA1 region.

Fig. 4. Hup A increases the MS-dCA1 cholinergic transmission.

a Quantification of AChE in the dCA1 by ELISA. n = 5, unpaired t-test, ^*P < 0.05 compared with “Vehicle” group. Veh vehicle. Representative fluorescence images (b) and corresponding quantification (c) of AChE in dCA1, dDG and dCA3. Scale bar, 100 μm. n = 3, unpaired t-test. ^**P < 0.01 compared with “Vehicle” group. d Illustration of AAV2 virus for retrograde tracing of cholinergic transmission. e Representative images of dCA1-projecting (Red) MS cholinergic neurons and ChAT⁺ neurons (Green) within different coronal sections from the anterior to the posterior axis. Scale bar, 50 μm. f Schematic of virus injection and fiber photometry recordings of ACh3.0. g Representative coronal image showing the expression of ACh3.0 in the dCA1. White dotted lines indicate tracks. Scale bar, 100 μm. h Average fluorescence signals (ΔF/F₀) of ACh3.0-expressed CA1 terminal in response to MS stimulation. Heatmap (i “Base” states; j “Kindled” states; each row represents one mouse) and corresponding quantification (k) of average ΔF/F₀. n = 8, Mann–Whitney test, ^**P < 0.01 compared with “Base” group, ^#P < 0.05 compared with “Kindled” group.

To investigate whether Hup A increases the concentration of dCA1 ACh during epileptic seizures by inhibiting the neural activity, we first mapped the location of dCA1-projecting cholinergic neurons. Using AAVretro-expressing Cre-dependent mCherry (AAV2/2-hEF1a-DIO-mCherry) [25] injected into the dCA1 region of ChAT-Cre mice (Fig. 4d), we observed that input cells were concentrated in the MS, between bregma positions from +1.34 mm to +0.5 mm along the anterior-posterior axis (Fig. 4e), which was consistent with previous reports [14], suggesting that MS-dCA1 cholinergic inputs.

Next, we explored the behavioral relevance of Hup A and MS-dCA1 cholinergic transmission in vivo, by expressing ACh3.0 (AAV2/9-hSyn-ACh3.0), an ACh sensor [26] in the dCA1 region and implanted an optical fiber above the injection site. An electrode was also implanted in the MS to elicit ACh release via a single-pulse stimulation (500 ms, 0.2 mA). Fluorescence dynamics of ACh3.0 in response to Hup A (Fig. 4f–g) were recorded under two conditions: (a) in naive mice, where MS stimulation caused a slight increase in ACh fluorescence in dCA1, which was significantly raised by Hup A administration (Fig. 4h–k, P < 0.01); and (b) in kindled epileptic mice, where MS stimulation induced a much higher ACh release, further amplified by Hup A treatment (Fig. 4h–k, P < 0.05). Together, our anatomical and functional investigations confirm that MS-dCA1 cholinergic transmission is enhanced by Hup A.

Hup A attenuates epileptic seizures by enhancing the MS-dCA1 cholinergic transmission

As Hup A modulated AChE in the dCA1 region, we next sought to address whether a direct infusion of Hup A into the dCA1 has an antagonism on the development of hippocampal seizure (Fig. 5a), we implanted a cannula in the dCA1 region and found 100 nM Hup A in dCA1 region significantly delayed the development of the behavioral seizure stage (Fig. 5b, F (34, 336) = 1.800, P < 0.01) and reduced ADDs (Fig. 5c, F (34, 336) = 1.589, P < 0.05) compared with the vehicle group. The above results reveal the anti-seizure effects of dCA1-localized Hup A administration, which are likely to be MS-dCA1 cholinergic transmission that processes the effect on epilepsy of Hup A.

Fig. 5. Hup A attenuates epileptic seizures by enhancing the MS-dCA1 cholinergic transmission.

a Illustration of Hup A interferes with the dCA1. The cannula was implanted in the dCA1 for Hup A intervention. Effects of Hup A injection in dCA1 on the development of seizure stage (b) and ADDs (c) during kindling acquisition, n = 6. Two-way repeated measures ANOVA with post hoc Bonferroni’s test, ^*P < 0.05 and ^**P < 0.01 compared to the “dCA1-Saline” group. d Schematic of viral injection, stimulation, and EEG recording in ChAT-taCasp3^MS-dCA1 mice. AAV-EF1α-FLEX-taCasp3 virus was injected in the MS and the cannula was implanted in the dCA1 for Hup A intervention. Effects of Hup A on the development of seizure stage (e) and ADDs (f) during kindling acquisition, in the condition of AAV-taCasp3-mediated apoptosis of the ChAT⁺ neurons in MS, n = 9. Two-way repeated measures ANOVA with post hoc Bonferroni’s test, ^**P < 0.01 and ^****P < 0.0001 compared to the “ChAT-Saline” group; ^####P < 0.0001 compared to the “ChAT::taCas3-Hup A” group. g Immunostaining showing the expression of ChAT in the MS. White arrow indicates the site of AAV-taCasp3 virus injection. Scale bar, 30 mm. h Typical sGS EEGs recorded from the cortex are shown.

To address whether blocking MS-dCA1 cholinergic transmission disrupts the anti-seizure actions of Hup A, we capitalized an AAV-taCaspase3-mediated (AAV2/9-EF1a-flex-taCasp3-TEVP) [27] apoptosis approach to selectively ablate ChAT⁺ neurons in the MS of ChAT-Cre mice (Fig. 5d). Immunochemistry confirmed a 97.11% reduction of cholinergic neurons in the MS (Fig. 5g), verifying the efficiency of viral expression. Ablation of MS ChAT⁺ neurons significantly promoted the development of the hippocampal seizures, as evidenced by increased behavioral seizure stages (Fig. 5e, F (68, 840) = 1.650, P < 0.0001) and amplified ADDs (Fig. 5f, F (68, 840) = 1.253, P < 0.0001) compared to the “ChAT-Saline” group, suggesting MS-dCA1 cholinergic transmission is anti-seizure in nature. Notably, in mice expressing the taCaspase3 virus in MS ChAT⁺ neurons, Hup A administration did not confer any protective effect on the seizure stage (Fig. 5e) or ADDs (Fig. 5f) compared to the “ChAT::taCas3-Saline” group. Further, exogenous supplementation of acetylcholine chloride (ACh) following the elimination of MS ChAT-positive neurons partially restored the anticonvulsant effects of Hup A, as indicated by the reduced seizure stage (Fig. 5e) and shortened ADDs (Fig. 5f) compared to the “ChAT::taCas3-Hup A” group. Representative hippocampal EEG recordings during the kindled seizures were shown in Fig. 5h. Together, these results demonstrate that the anti-seizure effects of Hup A rely on strengthening MS-dCA1 cholinergic transmission.

α7 nACh receptors mainly mediate anti-seizure cholinergic circuit modulated by Hup A

Finally, to identify which the cholinergic receptors involved in the anti-seizure cholinergic circuit modulated by Hup A, we sought to explore the influence of systemic Hup A administration combined with different cholinergic receptor antagonists on cholinergic transmission using fiber photometry. We expressed Cre-dependent ChrimsonR AAVs (AAV2/9-hSyn-FLEX-ChrimsonR-mCherry) [28] to the MS of ChAT-Cre mice and labeled ACh release by ACh3.0 sensor (AAV2/9-hSyn-ACh3.0). After recovery from surgery, 635-nm laser light stimulation was delivered into the dCA1 region for activation of the dCA1-projecting MS cholinergic fibers and the release of ACh in the dCA1 region was simultaneously recorded (Fig. 6a).

Fig. 6. α7 nACh receptors mainly mediate anti-seizure cholinergic circuit modulated by Hup A.

a Schematic of virus injection and stimulation in ChAT-Cre mice. AAV-hSyn-FLEX-ChrimsonR-mCherry virus was injected in the MS, AAV-hSyn-ACh3.0 virus in the dCA1, and optic fiber was implanted in the dCA1 for light stimulation and fluorescence signals recording. b Representative fluorescence image of ChrimsonR (red) showing co-localization with ChAT⁺ cholinergic neurons (green) in the MS. Scale bar, 60 μm. c Representative fluorescence signals of the average ΔF/F₀ with Hup A and different AChR antagonist interventions. Sal, saline. d–f Heatmap (d, e each row represents one mouse) and corresponding quantification (f) of average ΔF/F₀. n = 6, One-way ANOVA with post hoc Dunnett’s test, ^*P < 0.05 and ^****P < 0.0001 compared with “Sal + Sal” group, ^#P < 0.05 and ^####P < 0.0001 compared with “Sal + Hup A” group. g Illustration of Hup A and AChR antagonists interferes with the dCA1. The cannula was implanted in the dCA1 for Hup A and AChR antagonist intervention. Effects of Hup A injection in dCA1 on sGS seizure stage (h) and ADDs (i) in fully kindled state. n = 11, One-way ANOVA with post hoc Dunnett’s test, ^**P < 0.01 and ^****P < 0.0001 compared to the “Sal” group; ^##P < 0.01 and ^###P < 0.001 compared to the “Sal + Hup A” group. j Typical sGS EEGs recorded from the cortex are shown.

Immunofluorescence staining analysis demonstrated that 90.14% of mCherry labeled neurons were ChAT positive (Fig. 6b). Continuous pulses of photo-stimulation (30 s) of the dCA1-projecting MS cholinergic circuit elicited a mild increase in the release of ACh (ΔF/F₀, Sal + Sal). Administration of Hup A significantly enhanced this effect (Hup A + Sal, P < 0.0001), with similar amplification observed when muscarinic AChR (mAChR) antagonist Atr was co-administered (Hup A + Atr, P < 0.05). Notably, combining Hup A with nicotinic (nAChR) antagonists MEC tended to decrease the enhanced ACh signals (Hup A + MEC, P < 0.05), suggesting nAChRs rather than mAChRs as the primary contributor in the Hup A-enhanced cholinergic transmission. Comparative analysis of RNA expression levels showed that α7 nAChRs are highly expressed in hippocampal dCA1, which is also critical for synaptic plasticity [29] and highly sensitive to increased ACh levels during heightened cholinergic activity [30]. Accordingly, we combined α7 nAChR antagonists MLA with Hup A (Hup A + MLA) and found that MLA completely reversed the increase of ΔF/F₀ signals in this cholinergic transmission compared with the “Hup A + Sal” group (P < 0.0001, Fig. 6c–f). These results imply that the Hup A enhanced MS-dCA1 cholinergic ACh signaling is mainly mediated by α7 nACh receptors.

We next used a pharmacological approach to examine the contribution of α7 nACh receptors in the dCA1 in mediating the anti-seizure effects of Hup A. MLA was directly delivered into the dCA1 region before local Hup A administration, followed by behavioral analysis on the kindled mice (Fig. 6g). We observed an alleviation in seizure severity in Hup A-treated mice compared with the “Sal” group, as noted with a significant decrease in sGS seizure stage (Fig. 6h, P < 0.0001) and ADDs (Fig. 6i, P < 0.01). By contrast, MLA pre-treatment reversed the anti-seizure effects of Hup A (Fig. 6h, seizure stage: P < 0.001; Fig. 6i, ADDs: P < 0.01). Typical hippocampal EEGs during the kindled seizure were shown in Fig. 6j. Collectively, the anti-seizure actions of Hup A involve a crucial contribution of the α7 nACh receptors in the dCA1 region.

Discussion

This study presents a series of experiments to identify the specific cholinergic circuits and receptors modulated by the AChE inhibitor Hup A, and to demonstrate its potential in preventing epileptic seizures. Here, we first confirmed that a single systemic administration of Hup A attenuates seizures in the MES, PTZ, and KA-induced seizure models, which phenocopy generalized tonic-clonic seizures and generalized clonic seizures in preclinical tests [31]. These results, combined with prior studies showing that Hup A protects against PTZ-triggered seizures [32] and corneal stimulation-induced seizures in Scn1a^+/− mutants mice [7], highlight the broad-spectrum anti-seizure activities of Hup A. Intriguingly, we observed the satisfying efficacy of Hup A on acute seizure activity even at 0.1 mg/kg, offering the benefit of minimizing cholinergic drug buildup. Of the approved anticonvulsants tests, Hup A processes the satisfying potency with the lowest ED₅₀ across all seizure stimulus current levels measured, being 57 times more potent than levetiracetam and 301 times more potent than valproic acid at 32 mA [33, 34]. Additionally, given the two enantiomers of Hup A, the naturally occurring Hup A and the synthetic (+)-Hup A enantiomer [35], the difference in their anti-seizure actions remain unclear. Our findings reveal that (+)-Hup A only moderately retards PTZ-induced acute seizures, indicating that the naturally occurring form, Hup A, maybe a more optimal anti-seizure compound. This is likely due to Hup A being 38–50 times more potent in inhibiting AChE than its (+)-Hup A enantiomer [36]. Moreover, Hup A treatment significantly prevented seizure development and attenuated GS expression in the kindling-induced hippocampal seizure model. Hup A also reduced the chronic spontaneous seizure activity within 8 h per day in the KA-induced TLE model, which is often resistant to multiple ASMs. Supported by previous finding that no spontaneous seizure activity was observed for up to 4 h following a single intracerebroventricular administration of Hup A [5, 37]. Our results further confirm that Hup A exhibits broad-spectrum anti-seizure effects across various seizure and epilepsy models, with its long-lasting and high-potency therapeutic features.

To investigate the anti-seizure mechanisms of Hup A in the circuit level, we focused on its well-established ability to inhibit AChE with high specificity and efficacy [38]. There is considerable regional variation in ACh elevation after Hup A injection in naive mice, mainly concentrated in the frontal and parietal cortex, nucleus accumbens, hippocampus, and striatal cortex [24]. However, there is a paucity of information regarding the effects of Hup A under epileptic conditions. Our findings suggest that the dCA1 region of the hippocampus, rather than the cortex, or dCA3, DG, is critical in mediating the anti-seizure effects of Hup A. Combining electrical stimulation of the MS with ACh3.0 sensor-based fiber photometry in the dCA1 region, we demonstrated that Hup A significantly enhances MS-dCA1 cholinergic transmission in both naive and epileptic mice. Direct infusion of 100 nM Hup A into the dCA1 region retarded the development of hippocampal seizure. Additionally, we provide evidence that AAV-taCaspase3-mediated ablation of septal cholinergic neurons significantly reversed Hup A-mediated seizure protection. Notably, exogenous supplementation of ACh following the elimination of MS ChAT-positive neurons restored the seizure protection conferred by Hup A. These findings suggest that the anti-seizure action of Hup A involves recruitment of MS-dCA1 cholinergic transmission. Previous research has found that Hup A exerts its anti-seizure effects by enhancing GABA signaling through suppression of AChE activity. Our results are consistent with prior findings that hippocampus-projecting septal cholinergic neurons directly target GABAergic interneurons in the hippocampus, and that activation of GABA neurons is essential for the anti-seizure effects of MS-hippocampus cholinergic circuit [39].

Finally, we identify the subtype of cholinergic receptors contributing to Hup A-evoked cholinergic transmission during epilepsy. Prior evidences have suggested that atropine completely blocks the anticonvulsant effect of Hup A in the 6 Hz-induced seizure model, proposing a vital role for the mAChR receptor [33]. However, we found that atropine only slightly reversed Hup A-evoked the increases in cholinergic transmission, suggesting the involvement of a mAChR-irrelevant mechanism. In contrast, we found Hup A modulated cholinergic transmission mainly mediated by α7nAChRs within the dCA1 region. Importantly, we used a pharmacological approach to verify the contribution of α7nAChRs in the dCA1 to the anti-seizure effects of Hup A. In fact, α7nAChRs have been considered as a candidate target for the epilepsy phenotype. Encoding and transcription of α7nAChRs are particularly abundant in rat hippocampus, amygdala, and cortex, as shown in situ by hybridization [40], especially in hippocampal GABAergic neurons [29], which overlap with the regions where AChE predominates. The α7nAChRs play an essential role in inhibitory neural transmission, promoting Ca²⁺-dependent vesicular GABA release and increasing postsynaptic GABAergic currents [41, 42]. Therefore, Hup A exerts its anti-seizure effects by inhibiting AChE in these regions, thereby contributing to the MS-hippocampal cholinergic circuit-mediated suppression of seizures via α7nAChRs. While MEC targets several neuronal nAChR subtypes, including α2, α4, α7 and β2 nAChRs, α4β2 nAChRs in the dCA1 region may also contribute to seizure modulation is less well-defined compared to α7 nAChRs [30]. Future studies could explore whether α4β2 nAChRs contribute complementary or synergistic effects when modulated alongside α7 nAChRs, providing a broader understanding of cholinergic involvement in seizures. Moreover, it has shown that nACh and mACh receptor agonists, along with AChE inhibitors, contribute to a shared signaling pathway that amplifies NR2B-NMDA receptor responses at Schaffer collateral synapses in the hippocampal CA1 region [43]. Hup A has also been shown to decrease KA-induced microglial NLRP3 expression and caspase-1 activity in the hippocampus by diminishing oxidative stress to ameliorate seizure severity [6]. Collectively, NR2B-NMDA receptors, GABAergic transmission, or neuroinflammation could serve as convergence points of cholinergic downstream pathways, warranting further investigation. This at least, indicates that cholinergic neurotransmission targeted by Hup A may be a more suitable drug for treating TLE.

Together, this study assesses the broad-spectrum efficacy of Hup A, especially in hippocampal seizures, and dissects the anatomical underpinnings of the antiepileptic actions by enhancing MS-dCA1 cholinergic transmission, providing potential drug targets for the precise treatment of TLE through modulation of cholinergic neurotransmission.

Declaration of transparency and scientific rigor

This declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigor of preclinical research recommended by funding agencies, publishers, and other organizations engaged with supporting research.

Author contributions

The main idea of this study was from Yi W, ZC, and Yu W. Yu W, KYH, QYZ, YJS, LJL, and FW conducted the experiments. Yu W, KYH, and QYZ conducted the data analysis. GT, FF, CLX, JJF, XHJ, and JNW provided technical guidance and contributed to the data discussion. Yu W and Yi W wrote the manuscript. WLL and Yi W edited the manuscript. Yi W and ZC supervised all aspects of the work.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41401-025-01522-w.