Dysfunction of γ-aminobutyric acid A (GABAA) receptors (GABAARs) is a prominent factor affecting intractable epilepsy. Plic-1, an ubiquitin-like protein enriched in the inhibitory synapses connecting GABAARs and the ubiquitin protease system (UPS), plays a key role in the modification of GABAAR functions. However, the relationship between Plic-1 and epileptogenesis is not known. In the present study, we aimed to investigate Plic-1 levels in patients with temporal lobe epilepsy, as well as the role of Plic-1 in regulating onset and progression of epilepsy in animal models. We found that Plic-1 expression was significantly decreased in patients with epilepsy as well as pilocarpine- and pentylenetetrazol (PTZ)-induced rat epileptic models. Intrahippocampal injection of the PePα peptide, which disrupts Plic-1 binding to GABAARs, significantly shortened the latency of seizure onset, and increased the seizure severity and duration in these two epileptic models. Overexpressed Plic-1 through lentivirus transfection into a PTZ model resulted in a reduction in both seizure severity and generalized tonic–clonic seizure duration. Whole-cell clamp recordings revealed that the PePα peptide decreased miniature inhibitory postsynaptic currents (mIPSCs) whereas overexpressed Plic-1 increased mIPSCs in the pyramidal neurons of the hippocampus. These effects can be blocked by picrotoxin, a GABAAR inhibitor. Our results indicate that Plic-1 plays an important role in managing epileptic seizures by enhancing seizure inhibition through regulation of GABAARs at synaptic sites.
- protein linking integrin-associated protein to cytoskeleton 1 (ubiquitin-like protein 1 or Plic-1)
• GABAAR dysfunction plays a significant role in intractable epilepsy but the underlying mechanisms are not well understood, although we have previously demonstrated that Plic-1, a regulator of GABAAR, is significantly decreased in TLE patients and in two different rat epilepsy models.
• In the present study, we found that disruption of Plic-1 binding to GABAARs significantly shortened the latency of seizure onset and increased seizure severity, whereas overexpression of Plic-1 reduced seizure severity and GTC duration in epileptic models. Further support for these findings was obtained by measuring mIPSCs and eIPSCs recorded in pyramidal neurons of epilepsy models.
• Our work indicates that Plic-1 plays an important role in managing epileptic seizures through regulation of GABAA-ergic synaptic current and suggests that Plic-1 may serve as a potential therapeutic target for epilepsy.
Reduced expression of γ-aminobutyric acid A (GABAA) receptors (GABAARs) corresponds to enhanced excitation of neuronal circuits and abnormal network oscillations involved in epileptogenesis [1–3]. Dysfunction of GABAARs also affects the efficacy of GABA-targeting drugs in patients with chronic temporal lobe epilepsy (TLE) [4–6]. Newly synthesized GABAARs are assembled in the endoplasmic reticulum (ER), trafficked through the Golgi complex, and recycled back to the cell surface . Changes in GABAARs may contribute to short- or long-term synaptic plasticity in epilepsy [8,9]. Dysfunction of GABAARs results in the failure of dendritic inhibition in the hippocampus of TLE patients and epileptic rodent models [4,10–12], e.g. a significant decrease in GABAAR-gated chloride currents in hippocampal neurons under low-magnesium conditions is due to the reduction in GABAARs . In addition, studies indicate that the β2/3 and γ2 subunits of GABAARs are internalized or dephosphorylated during or after an epileptic seizure [6,13].
The protein linking integrin-associated protein to cytoskeleton-1 (Plic-1) is a 67-kDa ubiquitin-like protein enriched at inhibitory synapses . The N-terminus of Plic-1 contains a proteasome-interacting motif that is homologous to ubiquitin and interacts with the ubiquitin protease system (UPS). The C-terminus contains an ubiquitin-associated domain, which binds directly to the GABAARs . Plic-1 has been identified as a protein that interacts with α1–3, α6 and β1–3 subunits of GABAARs . Plic-1 regulates GABAARs by stabilizing their residence in the ER through inhibition of ER-associated degradation of GABAARs. Plic-1 also facilitates the insertion of newly synthesized or recycled receptors into the plasma membrane of neurons, thereby increasing the level of GABAARs on the cell surface [17,18].
However, the role of Plic-1 during onset and progression of epilepsy has not yet been determined. In the present study, we first investigated Plic-1 expression levels in patients with epilepsy and rat models of epilepsy. We found that Plic-1 was decreased in brain tissues in both. We then injected a cell-permeable PePα peptide into the hippocampus of two rat models with seizure, to examine the connection between Plic-1 and GABAARs . Intrahippocampal injection of PePα peptide, which disrupts Plic-1 binding to GABAARs, significantly shortened the latency of seizure onset, and increased the seizure severity and duration in the two rat models of seizure. Furthermore, we administered lentivirus (LV) that overexpressed Plic-1 to the hippocampus to investigate the impact of Plic-1 on the onset of epileptic seizures. Overexpressed Plic-1 through LV transfection in the pentylenetetrazol (PTZ) model resulted in a reduction of seizure severity and duration of generalized tonic–clonic seizures (GTCs). Finally, using whole-cell patch–clamp recordings we revealed variation in the miniature inhibitory postsynaptic current (mIPSC) in pyramidal neurons of the hippocampus after Plic-1 inhibition. The present study is the first to provide evidence that Plic-1 plays an important role in managing epileptic seizures through regulation of GABAAR function.
MATERIALS AND METHODS
A total of 40 human brain tissue samples (30 samples from TLE patients and 10 from non-TLE patients) were randomly selected from our brain tissue bank . Informed consent was set by the National Institutes of Health and the Committee on Human Research at Chongqing Medical University, Chongqing, China, and established based on the guidelines of the Declaration of Helsinki of the World Medical Association and the guidelines for the conduct of human research.
Neocortical tissues adjacent to epileptic foci at the temporal region from TLE patients were collected. All patients had at least 4 years of a continuous history of TLE, and had been diagnosed via: typical clinical manifestations (according to the 2001 International Classification of Epileptic Seizures); 24-h abnormal electroencephalogram (EEG); magnetic resonance imaging; and pathological histological changes (Table 1). The selected patients had been taking antiepileptic drugs (AEDs) including: levetiracetam, valproate, phenobarbital, carbamazepine, topiramate, phenytoin, lamotrigine, oxcarbazepine and clonazepam. Control neocortical tissue samples were obtained from non-TLE patients who needed therapeutic surgical tissue resection as a result of head trauma but had no obvious histological seizure manifestations or usage of any AEDs. There were no statistically significant differences in age of onset, frequencies or durations of seizures, and sex between the TLE patients (P>0.05).
A total of 96 Sprague Dawley male rats (220–280 g, 2–3 months of age) from the Experimental Animal Center of Chongqing Medical University were used for the present study. All animal experimental procedures were approved by the Ethics Committee of Chongqing Medical University (CQMU 0002648), as well as by international guidelines on the ethical use of animals that all efforts should be made to minimize the number of animals used and their suffering. The rats were habituated to the laboratories (room temperature: 23±1°C; 12-h light and dark cycles, free access to food and water) before the experiments were conducted. The pilocarpine model of seizure, which mimics the features of human TLE [20,21], was established by intraperitoneal (IP) injection of lithium chloride (127 mg/kg, Sigma-Aldrich Co.) into the rats. IP pilocarpine (50 mg/kg at first dose, Sigma-Aldrich Co.) was administered 24 h after injection of lithium chloride, followed by repeated IP injection (at 10 mg/kg) every 30 min until the rats developed status epilepticus. To establish the pilocarpine epilepsy model, 1 week after pilocarpine injection, if the rats displayed spontaneous recurrent seizures they were considered as the chronic phase pilocarpine epilepsy model. For controls, the rats were injected intraperitoneally (i.p.) with an equal volume of saline rather than pilocarpine.
For the PTZ model of seizure, which is considered to be a proper animal model for preclinical epilepsy research on GABAARs [22,23], the rats were injected with PTZ (65 mg/kg, i.p., Sigma-Aldrich Co.) to induce seizures. To stop the continuous seizures, diazepam (10 mg/kg) was given by IP injection in rats 1 h after the first onset of seizures. To establish the PTZ epilepsy model, the animals were given a subconvulsive dose of PTZ (40 mg/kg, i.p.) every day until the first unprovoked seizure was observed (usually 1–2 weeks after PTZ administration). If the rats displayed spontaneous seizure, they were considered as the chronic phase PTZ epilepsy model. In our experiments, only the chronic phases of the pilocarpine and PTZ epilepsy models were used for detection of Plic-1 expression and on EEG recordings. The pilocarpine and PTZ models of seizure were used to investigate the mechanism of Plic-1. For controls, the rats were injected intraperitoneally with an equal volume of saline rather than PTZ.
A magnesium-free epilepsy model that induces spontaneous recurrent epileptiform discharges (SREDs) in brain slices was prepared for electrophysiological evaluation (see Electrophysiological assessments).
The general behavioural activities induced by either pilocarpine or PTZ were recorded by video camera and analysed by two researchers blinded to treatment conditions. For the pilocarpine model of seizure the behaviours were scored as Racine's scale evaluation : 0: arrest, wet dog shakes and normal behaviour; 1: facial twitches (nose, lips and eyes); 2: chewing and head nodding; 3: forelimb clonus; 4: rearing and falling on forelimbs; and 5: imbalance and falling on side or back. For the PTZ model of seizure, the behaviours were scored as follows : 0: no further seizures; 1: a generalized clonic seizure; 2: a generalized clonic seizure with loss of righting reflexes; 3: a generalized clonic seizure with loss of righting reflexes plus running and bouncing; and 4: score of 3 plus forelimb tonus. Latency was defined as the time after the pilocarpine or PTZ injection to the first seizure onset. The duration of GTCs after PTZ was evaluated among different groups. To minimize possible complicating effects on the behaviour of the animals’ circadian rhythms, the seizures were induced between 13:00 and 16:00 hours. For the pilocarpine and the PTZ epilepsy models, the spontaneous recurrent seizures (SRSs) were scored according to Veliskova's criteria  and recorded with an EEG during the epilepsy period after the seizure-free latent periods. The score criteria are as follows: 1: staring and mouth clonus; 2: automatisms; 3: monolateral forelimb clonus; 4: bilateral forelimb clonus; 5: bilateral forelimb clonus with rearing and falling; and 6: tonic–clonic seizure. In pilocarpine epilepsy models, rats with scores above stage 4 and EEG recordings of the neocortex showing bursts of spiking activity were divided into a spontaneous seizure group. Rats without SRSs and with no electrographic seizures during the same chronic period were included in a non-spontaneous group. Rats without seizure were included in a control group. In PTZ epilepsy models, rats with scores above stage 4 and EEG recordings of the neocortex showing bursts of spiking activity were divided into a spontaneous seizure group. Rats without seizure during the same chronic period were included in a control group.
PePα peptide synthesis and stereotactic injection
A PePα peptide (NYFTKRGYAW) was synthesized (ChinaPeptides Co. Ltd and Wuhan More Biotechnology Co. Ltd) for intrahippocampal injection to interrupt the interaction of Plic-1 with GABAARs. The amino acid sequence NYFTKRGYAW is the same as the α1 subunit of GABAAR, and can also bind to Plic-1 via the ubiquitin-associated A (UBA) domain . Therefore, PePα competitively binds to GABAARs, and antagonizes the interaction between Plic-1 and GABAARs . A scrambled peptide (sequence AGKFNWYRTY) was also synthesized and used as the control to elucidate any possible disturbances caused by chemicals in the models . To make the peptides cell permeable, trans-activator protein (TAT) (YGRKKRRQRRR) was used as the internalization sequence. The purity of the two peptides used was evaluated by HPLC and MS analysis (Peptides Co. Ltd). To facilitate microscopic peptide visualization, FITC was inserted into the N-terminus of the peptide before injection to serve as a fluorescent label. The fluorescence images were obtained using an excitation wavelength of 494 nm for FITC (emission wavelength 518 nm).
The rats from each group (n=8) were anaesthetized with 3.5% chloral hydrate (1 ml/100 g i.p.), then placed in a stereotaxic instrument. Surgery began by exposing the dorsal surface of the skull of the rat and using a dental drill to create two parallel burr holes in the skull, according to the stereotactic coordinates at the CA1 region: anterior–posterior (AP) 3.2 mm, medial–lateral (ML) 2.5 mm and dorsal–ventral (DV) 2.6 mm. PePα peptide or scrambled peptide stock solutions were prepared freshly in saline at a concentration of 1 mg/kg in a 10-μl volume for each rat. A Hamilton syringe filled with 10 μl of a stock solution of the PePα peptide or scrambled peptide was inserted 3.7 mm in depth below the surface of the skull. The solution containing the PePα peptide was smoothly injected at a rate of 2 μl/min, and the needle was left in place for 5 min before being steadily and slowly retracted. For the control groups, an equal volume of saline or scrambled peptide solution was injected using the same approach. In addition, we used two different concentrations (1 mg/kg and 5 mg/kg per animal) of peptide solutions to examine whether there were any dose-dependent effects for the synthesized peptides in the PTZ seizure models. We also tested whether either the PePα peptide or the scrambled peptide could have any direct effect on seizure onset without chemoconvulsant. Seizures were induced in all rats 3 h after the injection using a pilocarpine or PTZ injection.
In vivo multichannel EEG recording and local field potential analysis
After PePα or scrambled peptide or saline was injected into the CA1 region (AP 3.2 mm, ML 2.5 mm and DV 2.6 mm), a microwire array (a 2×8 array of platinum–iridium alloy wire, each with a 25-μm diameter) was implanted into the same CA1 region of the hippocampus. A convulsant drug (PTZ) was given intraperitoneally 2 hours later. In vivo multichannel EEG recording was continuously recorded over 10 min, after administering the convulsant drug. An electrophysiological seizure was defined as manifesting seizure behaviour (from stage 1 to stage 4 according to PTZ scored criteria and from stage 1 to stage 5 according to Racine's scale evaluation for pilocarpine, including any epileptiform discharge in EEG which does not show in the rats’ behaviour). EEG recordings (of the intrahippocampal local field potential or LFP) showing high amplitude discharge (three times more than the baseline) were initiated from the hippocampus and spread to the cortex with high frequency (>5 Hz). The latency and the total duration of GTCs in an electrophysiological seizure in the 5 min during status epilepticus were evaluated across the groups. LFPs from the CA1 region of the hippocampus were preamplified (1000×), filtered (0.1–1000 Hz) and digitized at 4 kHz through an OmniPlex D Neural Data Acquisition System (Plexon). EEG recordings from both pilocarpine and PTZ epilepsy models in the chronic phase were also captured using the same method, without giving peptide or saline but with electrode fixing on rats in different groups. The behaviours related to spontaneous seizure were simultaneously recorded by video camera observation 24 h every day. After the seizures were recorded, the animals were sacrificed (pentobarbital 80 mg/kg, i.p.). The brain tissues were collected for morphological and biochemical studies.
Lentivirus vector generation and stereotactic injection
The LV vector Plic-1 [LV-Plic-1, Plic-1 sequence data from Ubqln1 (NM_053747)], encoding the amplified sequence of Plic-1, was incorporated into the vector as Ubi-MCS-3FLAG-SV40-EGFP (GV287 carrier, AgeI/AgeI digestion), along with the transgene for green fluorescent protein (GFP), and driven by a ubiquitin promoter (GeneChem Inc.). The same LV vector expressing GFP only (LV-GFP) was used as a control. The estimated titres of these LV vectors were determined using an RNA slot–blot technique in combination with serial dilution transduction on 293T cells. The final titre was 2×108 TU/ml for LV-Plic-1 and 2×109 TU/ml for LV-GFP (GeneChem Inc.).
Animals were randomly divided into three groups: LV-Plic-1 group, LV-GFP group and control group. We conducted bilateral, dorsal hippocampus, CA1 region microinjection at a total volume of 20 μl (10 μl per side) with LV-Plic-1, LV-GFP or saline for each animal using a glass pipette (0.2 μl/min). The pipette was left in place for at least 5 min after injection to prevent backflow. Successful LV infection could be detected after 5 days of injection. The LV transfection lasted for the lifespan of the rat after successful inoculation into host cells. We induced seizures with PTZ and pilocarpine 2 weeks after administration of the LV vectors. An additional LV-Plic-1 group (n=5) was not given any chemoconvulsant pharmaceuticals to exclude the possibility that Plic-1 overexpression alone may induce seizures. Behaviour was scored for all groups as described above. The in vivo multichannel EEG recording started 2 weeks after LV transfection. The procedure was the same as described above in the section on in vivo multichannel EEG recording and LFP analysis.
Immunofluorescence double labelling
Immunofluorescence double labelling was conducted on frozen sections of the neocortical tissue from TLE patients or control groups, and from the neocortical and hippocampal tissue of rat brains. Rabbit anti-Plic-1 (N-terminus, 1:100, Sigma), mouse anti-microtubule-associated protein 2 (MAP2, 1:50, Santa Cruz Biotechnology), mouse anti-glial fibrillary acidic protein (GFAP, 1:50, BosterBio), mouse anti-GABAAβ2/3 (anti-GABAARβ2,3 chain antibody, clone bd17; 1:100, Millipore) and mouse anti-GAD67 (synaptic vesicle membrane of GABA-ergic synapses, 1:100, Millipore). For double immunofluorescence labelling, frozen sections were first immersed in 100% acetone for dehydration for 30 min at room temperature and then incubated in 1% Triton-X for at least 30 min at 37°C. Next, sections were washed in ice-cold PBS before being blocked with 10% goat serum (Beijing Zhongshan Golden Bridge Co., Ltd.) for 30 min at 37°C. Sections were then incubated with a mixture of polyclonal rabbit anti-Plic-1 with either a monoclonal mouse MAP2 or a mouse anti-GFAP, or with mouse anti-GABAAβ2/3 at 4°C overnight. Sections were washed and incubated with Alexa Fluor 647-labelled goat anti-mouse IgG (1:200, Beyotime, Inc.), DYlight549 (549)-conjugated goat anti-mouse IgG (1:100, Zhongshan Golden Bridge, Inc.), or DYlight488-conjugated goat anti-rat IgG (1:100, Zhongshan Golden Bridge, Inc.) in the darkroom for 60 min at 37°C. After being washed with ice-cold PBS+Triton-X, sections were mounted using 80% glycerol. The fluorescence images were examined by laser scanning confocal microscopy (Leica Microsystems) on an Olympus IX 70 inverted microscope equipped with a Fluoview FVX confocal scan head.
Human neocortical tissues from TLE patients and controls, as well as rat cortex in different groups, were homogenized in RIPA lysis buffer (Beyotime Institute of Biotechnology) to separate the protein. Supernatants were removed after centrifugation at 4°C (16000 g for 10 min). Membrane-associated proteins were separated according to the transmembrane protein extraction protocol using the ProteoExtract Native Membrane Protein Extraction Kit (M-PEK Kit, Merck Millipore) The membrane proteins from fresh tissue were extracted based on the previous protocol [26,27] Briefly, the membrane proteins from fresh tissue were extracted using extraction buffers I and II, protease inhibitor cocktail and wash buffer (Merck Millipore). The procedure was as follows: the tissue pieces in ice-cold PBS were first centrifuged at 100 g twice, for 2 min at 4°C, to ensure that tissue pieces had pelleted. The supernatant was then removed and the tissue transferred to a pre-cooled homogenizer. Merck Millipore Corp. was then added to the homogenizer, followed by ice-cold extraction buffer I. After complete homogenization, the tissue was transferred to small cell clumps, then incubated for 10 min at 4°C with gentle agitation and centrifuged at 1000 g for 5 min at 4°C. After centrifuging, the supernatant was removed and the pellet washed with ice-cold PBS briefly and resuspended in 5 ml of fresh PBS; the suspension was centrifuged again at 1000 g for 5 min at 4°C. The supernatant with 0.2 ml of extraction buffer IIA and 5 μl of the Protease Inhibitor Cocktail Set III was added to the supernatant and the solution incubated for 15 min at 4°C, then centrifuged at 16000 g for 15 min at 4°C. Finally the supernatant, which was enriched in integral membrane proteins, was collected.
The concentration of pure protein was determined using the bicinchoninic acid protein (BCA) assay (Beyotime Institute of Biotechnology). Prepared proteins (50 μg per lane) were separated by (SDS/PAGE: 5% spacer gel, 90 V, 40 min; 10% separating gel, 100 V, 60 min) before being transferred to a PVDF membrane (250 mA, between 90 and 120 min according to different molecular mass of the protein). Next, the PVDF membrane was incubated at 37°C for 1.5 h in 5% skim milk or freshly prepared 5% BSA with 0.05% Tween-20 for blocking. PVDF membranes were then incubated with polyclonal rabbit anti-Plic-1 (1:500, Sigma-Aldrich) diluted in 5% skim milk or mouse anti-GABAAβ2/3 (1:500, Millipore) diluted in freshly prepared 5% BSA at 4°C overnight. The membranes were washed with Tween-20/TBS for 30 min, with 10-min changes, and then incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG antibody or horseradish peroxidase-conjugated goat anti-mouse IgG antibody (1:3000, Beyotime, Inc.) for 1.5 h at 37°C. A mouse anti-GAPDH antibody (1:1000, Santa Cruz Biotechnology) was used as a loading control. Densitometry quantification was measured using the Quantity One 1D Analysis Software (Bio-Rad Laboratories) as absorbance values, and Plic-1 or GABAAβ2/3 levels were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH).
Sprague Dawley rats aged 14–28 days or 2–3 months were used. Experimental rats were anaesthetized with 3.5% chloral hydrate (1 ml/100 g, i.p.) and then quickly decapitated. Cortical brain slices (400-μm thick) were cut in the coronal plane with a Vibratome (NVSLM1, Camden Instruments), and then immersed in an ice-cold dissection buffer. Brain sections were immediately removed and placed for 1 h at 37°C in artificial cerebrospinal fluid (ACSF) containing (in millimoles): NaCl, 124; KCl, 3; NaH2PO4·2H2O, 1.23; NaHCO3, 26; CaCl2, 2; MgCl2, 2; glucose, 10 (pH 7.3–7.4). The sections were then transferred to ACSF at 23°C for 30 min before being used for electrophysiological recordings (the dissection buffer and ACSF were equilibrated with 95% O2:5% CO2). Cells were visually chosen for recording by inverted-phase contrast microscopy (Nikon) of the hippocampal CA1 region. Whole-cell, membrane-ruptured, patch–clamp recordings of mIPSCs were carried out on the pyramidal neurons of the CA1 region in hippocampal slices.
All the electrophysiological recordings in ACSF were critically controlled and performed at 24.5°C. Membrane potential was maintained at −70 mV in voltage-clamp mode. For a Mg-free epileptic model in vitro, the brain sections were first exposed to Mg/ACSF for >5 min, then the bathing solution was replaced by ACSF without MgCl2 (Mg-free ACSF). Epileptiform discharge was characterized by activity potential manifesting as continuous high-frequency spike discharges compared with activity potential in Mg/ACSF . For mIPSC recording, a patch pipette (3–5 MΩ) including internal electrode solution was used containing (in millimoles): CsCl, 100; Hepes, 10; MgCl·26H2O, 1; EGTA, 1; NMG, 30; MgATP, 5; Na3GDP, 0.5; two creatine phosphate salts, 1; pH 7.2–7.3, 270–280 mosmol/l. We infused the solution with 20 μM DNQX (6, 7-Dinitroquinoxaline-2,3(1H,4H)-dione), 40 μM APV (D-2-amino-5-phosphonovalerate) and 1 μM TTX (tetrodotoxin) to block additional currents such as excitatory postsynaptic currents (EPSCs) and evoked inhibitory postsynaptic currents (eIPSCs). After 10 min of stable mIPSCs, Mg/ACSF was transferred into the brain sections. PePα peptide (200 μg/ml) or scrambled peptide (200 μg/ml) stock solutions were added to the Mg/ACSF medium. When stable mIPSCs were again obtained, recordings were collected. After the mIPSCs were recorded, Mg-free ACSF, without drugs, was used as the wash-out medium. PePα peptide (200 μg/ml) stock solutions were added into normal ACSF with magnesium to exclude any direct impact of PePα peptide on mIPSCs in the pyramidal neurons. GABA (100 or 500 μM, Sigma-Aldrich) was then added to the brain slices to ascertain whether the mIPSCs were GABA active-dependent currents. Finally, picrotoxin (PTX, 100 μM), a GABAA antagonist, was added to assess whether the mIPSC was conducted via the GABAAR.
To determine whether Plic-1 was able to alter phasic GABA-ergic synaptic current or tonic GABA-ergic current, a lower concentration of gabazine (200 nM) was infused into Mg-free ACSF to selectively inhibit the phasic GABA-ergic synaptic current, which left only a tonic GABA-ergic current, and then PePα peptide (200 μg/ml) was added into Mg-free ACSF. Variation of mIPSCs in the pyramidal neurons was recorded before and after both gabazine and PePα peptide treatment.
For eIPSCs, a bipolar-stimulating electrode was used to stimulate afferent fibres and the current pulses (0.1-ms duration) were set at 0.05 Hz, at a holding potential of 0 mV. Access resistance was continuously monitored throughout the process. Recordings with series resistance <20 MΩ were included for analysis. Phasic currents in mIPSCs and eIPSCs in all experimental procedures were measured, and at least one or two cells from ten brain slices in each group were randomly selected to exclude cell variability. In addition, we directly compared the impact of PePα peptide or scrambled peptide on mIPSCs in pyramidal neurons of the CA1 region, in the hippocampal slices, in normal ACSF and in Mg-free ACSF. The mIPSCs caused by PePα peptide only or scrambled peptide only were also recorded before changing to Mg-free ACSF.
To analyse the functional consequences of Plic-1 overexpression with respect to GABAARs, the hippocampal slices of seizure rats were collected 24 h after PTZ injection from LV-Plic-1, LV-GFP or control groups. The mIPSCs were recorded in pyramidal neurons of the CA1 region in the hippocampal slices incubated in Mg-free ACSF. All signals were amplified using a MultiClamp 700B Amplifier (Axon) and digitized with Digidata 1322 A. Data for mIPSCs were filtered at 10 kHz and those for eIPSC at 2 kHz, digitized at 10 kHz and analysed using pClamp 9.2 software (Molecular Devices) and MiniAnalysis Software. (Version 6.0.3; Synaptosoft).
All data are shown as means±S.E.M.s. Analyses of significance were performed using Student's t-test when the means of two experimental groups were compared and least significant difference (LSD) tests if more than two groups were compared. ANOVA repeated measures were used to compare sequential measurements with a single control measurement, followed by Dennett's test. ANOVA for non-parametric data (Kruskal–Wallis H test) was used for behaviour scores. Electrophysiological data were analysed using the Kolmogorov–Smirnov (KS) two-sample test using MiniAnalysis Software. Data for eIPSC properties were performed using Student's t-test for two independent experimental groups. P<0.05 was considered statistically significant (SPSS 16.0).
Plic-1 expression in the brain samples of epileptic patients and animal models
In human neocortical tissue, Plic-1 expression was significantly lower in the TLE group compared with the controls (0.47±0.03 in TLE and 1.14±0.16 in controls; Student's t-test; Figure 1A). In the cortex of the pilocarpine epilepsy model, Plic-1 expression in neocortical tissue was significantly decreased 2 months after seizure onset in the spontaneous seizure group (0.03±0.02) compared with the non-spontaneous seizure (0.22±0.06) and control groups (0.23±0.07; ANOVA followed by LSD tests; Figure 1B). In addition, in the hippocampus of the pilocarpine epilepsy model, there was significantly less Plic-1 expression in the spontaneous seizure group (0.08±0.03) compared with the non-spontaneous seizure (0.22±0.04) or control group (0.26±0.05; ANOVA followed by LSD tests; Figure 1C). In the hippocampus of the PTZ epilepsy model, Plic-1 expression was also significantly decreased in the spontaneous seizure group (0.25±0.05) compared with the control group (0.68±0.10; LSD tests; Figure 1D)
Bursts of spiking activity lasting <1 min, followed by depressed background activity, could be detected in the spontaneous groups with an EEG recording in chronic periods after the latency period, but could not be detected in the non-spontaneous groups during the same periods. The burst of spiking activity corresponded to the seizures at different stages including monolateral forelimb clonus or bilateral myoclonic twitch, a chronic generalized seizure. This feature of EEGs can be observed in both the pilocarpine epilepsy model and the PTZ epilepsy model (Figure 1E, upper panel for pilocarpine epilepsy model, lower panel for PTZ epilepsy model). Representative EEG recordings obtained from the start of the induction of acute seizures with pilocarpine to a fully generalized tonic–clonic seizure (stage 5) was displayed (Figure 1F).
Confocal microscopy revealed that Plic-1 was located both at the membrane and in the cytoplasm of neurons, with a little in the nuclei. Plic-1 was in MAP2-positive neurons in the hippocampus and cortex of the normal rats (Figure 1G, upper panel). Plic-1 also co-localized with GAD67-positive neurons in the cortex of normal rats (Figure 1G, lower panel). In addition, Plic-1 co-localized with GABAAβ2/3 on the cell membrane in the cortical neurons of TLE patients, and in the cortex and hippocampal regions (CA1, CA2, CA3, CA4, dentate gyrus) of the epileptic rat model (Figure 1H). There was an agglomerate distribution of GABAAβ2/3 on the membrane and in the cytoplasm, with the strongest immunoreactivity at the initial segment of neuron axons (Figure 1H). This is consistent with the previous studies which showed enrichment of these receptors at inhibitory synapses . However, Plic-1 did not co-localize with GFAP (Figure 1G, middle panel).
Administration of the PePα peptide increases seizure severity in animal models
To investigate the effect of the PePα peptide on seizure severity in animal models, we first assessed whether it was successfully administered in the hippocampus. The FITC-positive cells were visualized in the CA1 region 2 h after injection (Figure 2B), indicating that the peptide had been successfully delivered into the hippocampus.
In behaviour evaluation of the pilocarpine model of seizure, we found that the PePα peptide significantly enhanced seizure severity in the PePα peptide-treated group compared with the two control groups under Racine's scale evaluation: [X2(H)=7.20, P<0.05 (after 10 min); X2(H)=7.29, P<0.05 (after 50 min); Kruskal–Wallis test; Figure 2C, left] The PePα peptide-treated group also displayed shortened latency of seizure onset in the pilocarpine model of seizure (5.80±2.06 min). In contrast, the scrambled peptide-treated group (32.60±2.18 min) showed no significant difference compared with the control (37.80±2.71min, P<0.01; LSD tests; Figure 2C, right). Multichannel hippocampal EEG recording of electrographic seizure in the pilocarpine seizure model showed that the PePα peptide-treated group displayed shortened latency (8.22±1.19 min) compared with the scramble peptide-treated (21.13±1.57 min) and control groups (19.16±2.37 min, P<0.05; LSD tests; Figure 2E).
The results of behaviour evaluation and multichannel EEG recording have also been confirmed in a PTZ model of seizure. In behaviour evaluation of the PTZ model of seizure: [seizure severity: X2(H)=7.63, P<0.05; Kruskal–Wallis test; Figure 2D, left; seizure onset latency: P<0.05; LSD tests; Figure 2D, middle; duration of GTCs: P<0.05; LSD tests; Figure 2D, right].
Multichannel hippocampal EEG recording of electrographic seizure in the PTZ model of seizure showed that the PePα peptide-treated group displayed shortened latency (0.46±0.07 min) compared with the scramble peptide-treated (1.61 ±0.16 min) and control groups (2.11±0.21 min, P<0.05; LSD tests; Figure 2F). The PePα-peptide treated group also displayed prolonged total duration of electrophysiological seizures (3.38±0.14 min) compared with the scramble peptide-treated (1.71±0.26 min, P<0.05; LSD tests) and control groups (1.88±0.30 min; P<0.05; LSD tests; Figure 2F).
However, there was no difference in seizure severity between the different doses of PePα peptide administered for behaviour evaluation in the PTZ model: seizure severity: P>0.05; Kruskal–Wallis test; seizure onset latency: 0.77±0.18 min (1 mg/kg) and 0.69±0.17 min (5 mg/kg); P>0.05; LSD tests.
Administration of PePα peptide reduces expression of GABAARs in the hippocampus of seizure rats
In the pilocarpine model of seizure, Western blot analysis indicates that cell membrane-associated GABAARβ2/3 subunit expression in the hippocampus was significantly reduced in the PePα peptide-treated group (0.44±0.27) compared with the scramble peptide-treated (1.80±0.45) and control groups (1.91±0.48, P<0.05; LSD tests; Figure 2G, left). However, Western blotting for cell membrane-associated Plic-1 among these three groups did not show any significant differences (P>0.05; LSD tests; Figure 2G, right). A significant decrease in membrane-associated GABAARβ2/3 subunit expression could also be seen in the PePα peptide-treated group (0.35±0.07) in the PTZ model of seizure compared with the scramble peptide-treated (1.11±0.14) and control groups (1.18±0.13, P<0.05; LSD tests; Figure 2H, left), whereas cell membrane-associated Plic-1 among these three groups did not show any significant differences (P>0.05; LSD tests; Figure 2H, right)
The PePα peptide reduces inhibition of pyramidal neurons in CA1
When the PePα peptide was transferred into Mg-free ACSF, whole-cell current-clamp recordings from CA1 pyramidal neurons showed a higher firing of action potentials (APs, frequency of APs: 3.09±0.54 Hz) compared with the neurons with Mg/ACSF (APs: 0.31±0.04 Hz; P<0.05; n=8 cells per group; Kolmoforov–Smirnov or KS test). After infusion of PePα peptide into the hippocampal slices for 30 min in Mg-free ACSF, a significant reduction in amplitude of mIPSCs (14.02±0.74 pA) was recorded in the pyramidal neurons of CA1 compared with controls (21.57±1.54 pA, P<0.05; KS test; Figure 3A). The scrambled peptide-treated group showed no significant difference from the control group (P>0.05; KS test; Figure 3B). The PePα peptide-treated, but not the scrambled peptide-treated, group also decreased current frequency (P<0.05; KS test; Figures 3C and 3D). However, before and after transferring the Mg-free ACSF into the medium, the mIPSC amplitudes did not change significantly (20.10±1.42 pA) (Figure 3E), suggesting that the Mg-free model itself will not induce any significant changes in mIPSCs. Incubating PePα peptide alone also does not reduce the amplitude of mIPSCs (Figure 3F). Administration of a lower (100 μM, 16.02±0.83 pA) and a higher (500 μM, 23.65±1.12 pA) concentration of GABA demonstrated that the mIPSCs were GABA dose dependent (Figure 3G). Treatment with 100 μM PTX abolished the mIPSC, indicating that it is mediated by GABAARs (Figure 3H). After selectively blocking the phasic GABA-ergic synaptic current with 200 nM gabazine, the PePα peptide failed to alter tonic GABA-ergic mIPSC amplitude, suggesting that Plic-1 affects phasic GABA-ergic synaptic but not tonic GABA-ergic current.
In addition, after perfusion of 200 μM PePα peptide, the amplitude of the eIPSCs decreased significantly (96.90±5.46 pA) compared with the Mg-free controls (181.94±22.63 pA) in the pyramidal neurons of the CA1 region (P<0.05; KS test; Figure 3I). The scrambled peptide treatment did not reduce GABA-ergic-mediated eIPSCs (P>0.05; KS test; Figure 3J). Decreased amplitude of an eIPSC could be washed out (178.23±5.27 pA). This reversed eIPSC could also be abolished by 100 μM PTX.
Overexpression of Plic-1 increases GABAAR expression and decreases seizures
To investigate whether Plic-1 overexpression could have the opposite effect to PePα inhibition, we administered recombinant overexpressed lentivirus-Plic-1-sh gene into the hippocampus of the PTZ and pilocarpine epilepsy models. After injection of the exogenous recombinant LV-GFP with overexpressed Plic-1, we visualized GFP-positive cells in the CA1, CA3, CA4 and dentate gyrus regions after 14 days in control animals (Figure 4A). The total Plic-1 levels in the CA1 region increased steadily from control (0.58±0.11), 3 days (0.64±0.16), 7 days (1.15±0.06) to 14 days (1.72±0.16) after the initial injection (P<0.05; ANOVA repeated measures followed by Dennett's test; Figure 4B), indicating that the recombinant lentivirus has been successfully transfected into the neurons in the hippocampus.
In the PTZ model of seizure, behavioural evaluation indicated that rats in the LV-Plic-1 group had less seizure severity than rats in either the LV-GFP or the control group (X2(H)=9.45, Kruskal–Wallis test; P<0.05; Figure 4E, left upper panel). However, the latency of seizure onset was not as significantly increased in the LV-Plic-1 group compared with the LV-GFP (P>0.05) and control groups (P>0.05; LSD tests; Figure 4E, right upper panel). The duration of GTCs was significantly shorter in the LV-Plic-1 compared with the control group (P<0.01; LSD test; Figure 4E, lower panel). Similar results of behavioural evaluation had been confirmed in a pilocarpine model of seizure in three groups (data not shown).
Multichannel EEG recording revealed that the LV-Plic-1 group had a significantly shorter total duration of electrophysiological seizure (0.80±0.14 min) compared with either the LV-GFP (2.08±0.35 min, P<0.05; LSD tests) or the control (2.26±0.43min, P<0.05; LSD tests; Figure 4D) group in the PTZ model of seizure. However, the LV-Plic-1 group showed no significant difference in the latency compared with other groups (P>0.05; LSD tests). In the pilocarpine model of seizure, multichannel EEG recording revealed that the total duration of electrophysiological seizure in the LV-Plic-1 group (12.81±4.72 min) was also shorter than either the LV-GFP (38.69±3.49 min, P<0.05; LSD tests) or the control (39.06±6.62 min, P<0.05; LSD tests) group.
Western blot analysis of the cell membrane-associated Plic-1 level in the PTZ model of seizure of the LV-Plic-1 group (0.33±0.03) was significantly higher than in the LV-GFP (0.17±0.02) and control groups (0.16±0.03, P<0.05; Figure 4C). Plic-1 expression in the pilocarpine model of seizure was also significantly increased in the LV-Plic-1 group (1.15±0.17) compared with that in the LV-GFP (0.49±0.14) and control groups (0.62±0.18,*P<0.05).
Overexpression of Plic-1 enhances synaptic inhibition of pyramidal neurons in CA1
Western blot analysis of brain slices indicated that cell membrane-associated GABAAβ2/3 expression in the LV-Plic-1 group (1.97±0.47) was significantly elevated compared with both the LV-GFP (0.60±0.23) and the control groups (0.73±0.29, P<0.05; LSD tests; Figure 5A) in the PTZ model of seizure. In the pilocarpine model of seizure, cell membrane-associated GABAAβ2/3 levels in the LV-Plic-1 group (1.58±0.09) were also significantly elevated compared with both the LV-GFP (0.60±0.17) and the control groups (0.37±0.04, P<0.05; LSD tests). In addition, mIPSCs in the LV-Plic-1 group had a significantly higher amplitude (27.49±2.94 pA) than in the LV-GFP (17.89±1.62 pA) and control (15.20±1.51 pA) groups 24 h after seizure, recorded in Mg-free ACSF (P<0.01; KS tests; Figures 5B–5D). The mIPSCs in the LV-Plic-1 group also had a significantly higher mIPSC frequency compared with the other two groups (P<0.01; KS test; Figure 5E). Finally, 100 μM PTX abolished the mIPSC in all groups, indicating that the mIPSCs are mediated by GABAARs.
To exclude any changes of endogenous Plic-1 during the experimental procedure, we investigated the Plic-1 levels at 6, 24 and 72 h in the hippocampus of the pilocarpine seizure model. Our results showed that the Plic-1 protein levels in the seizure group were significantly increased at 24 h (0.96±0.06) compared with the control group (0.32±0.04), but not at 6 h (0.45±0.03, P<0.05; ANOVA repeated measures followed by Dennett's test). Plic-1 levels in the CA1 area (0.92±0.10) were not significantly different from those for the other areas of the hippocampus (including the dentate gyrus and the CA3 area, 0.68±0.11, P>0.05; LSD tests) in the seizure model.
In the present study, we were the first to report that Plic-1 plays an important role in controlling epileptic seizures. In particular, we found that Plic-1 expression was decreased in the neocortex of TLE patients compared with controls. Plic-1 co-localizes with GABA-ergic synapses specifically associated with GABAAβ2/3. We also assessed the dynamic changes of Plic-1 expression in both the cortex and the hippocampus during seizure development in two epileptic rat models. A transient elevation of Plic-1 during the acute phase of seizure in the animal models may be due to its endogenous anti-seizure function. To investigate the role of Plic-1 in seizure and the mechanisms underlying these actions, we used both synthesized peptides to block Plic-1 and a lentivirus overexpressing Plic-1 to examine the relationship between Plic-1 and GABAAR in two classic seizure rat models. We conducted the behaviour tests and multichannel EEG recording, and observed that seizure severity was aggravated whereas the latency of seizure was shortened when PePα peptide, an inhibitor that blocked connections between GABAAR and Plic-1, was administrated to both the pilocarpine and the PTZ models of seizure. However, overexpression of Plic-1 through LV transfection attenuated the severity of seizure and these changes are associated with up-regulated GABAARβ2/3 expression. In vitro, whole-cell patch–clamp recording revealed that Plic-1 affects seizure onset through regulation of GABAARs on phasic GABA-ergic synaptic currents but not tonic GABA-ergic currents in CA1 pyramidal neurons.
A functional association between Plic-1 and GABAAR has been reported in in vitro studies on cortical neurons or HEK-293 cells [7,16]. These studies indicated that Plic-1 increases the levels of GABAARβ2/3 subunits by prolonging the half-life of polyubiquitinated GABAARβ2/3 and its residence in the ER . However, a decrease of Plic-1, or other scaffolding proteins connected to GABAAR and its different subunits, could affect GABAAR expression in neurons [30,31]. Plic-1 might also stabilize GABAARs from endocytosis, ubiquitination  or degradation. To our knowledge, the role of Plic-1 in neurological disorders such as epilepsy has not been investigated. In the present study, we found that Plic-1 could regulate the strength of synaptic inhibition in an epileptic model by modulating the mIPSCs of GABAARs. Our electrophysiology results from whole-cell patch–clamp of pyramidal neurons in CA1 from both seizure experimental rats and the Mg-free epileptic model further support our in vivo findings that Plic-1 plays an important role in regulation of seizures of epilepsy through GABAARs.
It is generally considered that GABAARβ2/3 is the predominant receptor among GABAARs in dendritic regions of the hippocampus and dentate gyrus, which mediates decaying mIPSCs of inhibitory synapses . In contrast, some studies have reported that deletion of the β2 subunits has no effect on mIPSCs of dentate gyrus granule cells [34,35]. In CA1 pyramidal neurons, GABAARβ3 displayed a strong and long-lasting inhibition [36,37], via local dendritic shunting and somatic hyperpolarization, to control hyperexcitability, thus preventing seizures . Absence of the β3 subunit would not only perturb θ–γ coordination, but also weaken GABAARs’ mIPSCs in the CA1 pyramidal cells, thereby impairing GABA-ergic inhibition . In addition, the GABAARβ3 subunits also interact with benzodiazepine-sensitive synaptic receptors (α5β1/3γ2), but not with benzodiazepine-insensitive extrasynaptic receptors (α4β2δ) . These receptors are internalized during prolonged status epilepticus, thus compromising synaptic inhibition . Clinically, lack of the GABAARβ3 subunit could cause Angelman's syndrome, a neurodevelopmental disorder with the typical characteristics of epilepsy . Our results from both human and animal models of epilepsy indicated that GABAARβ2/3 was significantly decreased which may be due to dysfunction of Plic-1.
It has been proved that the CA1 and CA3 regions are the most vulnerable areas of epilepsy in both human and pilocarpine rat models [42–45]. Moreover, the CA1 region has always been the epicentre of synaptic morphological alteration and epileptogenic network origin in either inhibitory or excitatory cells . Notably, when seizures occurred in animals, Plic-1 protein levels increased transiently (significantly increased at 24 h but not at 6 h after first seizure onset) in the hippocampus, perhaps reflecting the functionality of endogenous Plic-1 in response to inhibiting hyperexcitation during seizures [46,47]. Therefore, overexpression of Plic-1 in the CA1 region is involved in the regulation of GABAAR expression in this area.
Previous evidence has shown that the synaptic GABAARs, which conducted phasic synaptic current, but not the extrasynaptic GABAARs, which conducted tonic synaptic current, could be stimulated by seizures , although the control of tonic inhibition of GABAARs has been the target in some AEDs in clinical trials of the treatment of specific types of epilepsy . In the present study, we demonstrated that overexpression of Plic-1 in pyramidal neurons in the CA1 region enabled GABAARs, possibly the β3 subunits, to conduct higher mIPSC amplitude and frequency, thus attenuating the severity of seizures during status epilepticus. Infusion of gabazine selectively inhibited phasic GABA-ergic synaptic current [49,50] and disabled the effect of Plic-1 on mIPSC alteration, which indicated that Plic-1 regulates the activity of GABAARβ2/3 and mediates mIPSCs on phasic GABA-ergic synaptic current, at the CA1 hippocampal synapses of pyramidal neurons during and after seizure.
In summary, our studies indicate that Plic-1 plays an important role in the regulation of epileptic seizure onset during and after status epilepticus. It is likely to function through the following mechanisms: (a) regulation of GABAAR expression at synaptic sites during epileptic seizure; and (b) an effect on the inhibitory function by changing the mIPSCs and eIPSCs of phasic GABA-ergic synaptic current, which would compensate for the fast depression of inhibition in epileptic seizure. As an endogenous anticonvulsant system in epilepsy, Plic-1, together with GABAAR-associated proteins, may provide a useful medical target in future experimental or clinical therapy in patients with epilepsy.
Yujiao Zhang designed and conducted the experiments and wrote the manuscript. Xuefeng Wang designed and edited the manuscript. Yujiao Zhang, Juan Gu, Yanke Zhang and Wei Wang conducted the experiments. Zengyou Li and Juan Gu performed statistical analysis, the table and figures. Hui Shen and Guojun Chen analyzed electrophysiology experiments. All authors have involved in the preparation of the manuscript and approved the contributions.
This research was supported by the National Natural Science Fund .
We would like to thank Beijing Tiantan Hospital of the Capital University of Medical Sciences, Xuanwu Hospital and Daping Hospital of the Third Military Medical University for providing the patients, and for the patients’ voluntary participation and the provision of surgery samples in our previous lab work on Plic-1 expression in TLE patients. We also acknowledge Dr Jiang for providing the OmniPlex D Neural Data Acquisition System (Plexon) for conduction of the in vivo multichannel EEG recordings of epileptic rats, and her technical assistance.
Abbreviations: ACSF, artificial cerebrospinal fluid; AED, antiepileptic drug; AP, anterior–posterior or action potential; DV, dorsal–ventral; EEG, electroencephalogram; eIPSC, evoked inhibitory postsynaptic current; EPSC, excitatory postsynaptic current; ER, endoplasmic reticulum; GABAA, γ-aminobutyric acid A; GABAAR, GABAA receptor; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GFAP, glial fibrillary acidic protein; GFP, green fluorescent protein; GTC, generalized tonic–clonic seizure; i.p., intraperitoneally; IP, intraperitoneal; KS test, Kolmoforov–Smirnov test; LFP, local field potential; LSD, least significant difference; LV, lentivirus; MAP2, microtubule-associated protein 2; mIPSC, miniature inhibitory postsynaptic current; ML, medial–lateral; PTX, picrotoxin; PTZ, pentylenetetrazol; SRS, spontaneous recurrent seizure; TLE, temporal lobe epilepsy
- © 2015 Authors; published by Portland Press Limited