Clock knockout in inhibitory neurons reduces predisposition to epilepsy and influences anxiety-like behaviors in mice

Epilepsy is a brain disorder affecting up to 1 in 26 individuals. Despite its clinical importance, the molecular mechanisms of epileptogenesis are still far from clarified. Our previous study showed that disruption of Clock in excitatory neurons alters cortical circuits and leads to generation of focal epilepsy. In this study, a GAD-Cre; Clock flox/flox mouse line with conditional Clock gene knockout in inhibitory neurons was established. We observed that seizure latency was prolonged, the severity and mortality of pilocarpine-induced seizure were significantly reduced, and memory was improved in GAD-Cre;Clock flox/flox mice. We hypothesize that mice with CLOCK knockout in inhibitory neurons have increased threshold for seizure, opposite from mice with CLOCK knockout in excitatory neurons. Further investigation showed Clock knockout in inhibitory neurons upregulated the basal protein level of ARC, a synaptic plasticity-associated immediate-early gene product, likely through the BDNF-ERK pathway. Altered basal levels of ARC may play an important role in epileptogenesis after Clock deletion in inhibitory neurons. Although sEPSCs and intrinsic properties of layer 5 pyramidal neurons in the somato-sensory cortex exhibit no changes, the spine density increased in apical dendrite of pyramidal neurons in CLOCK knockout group. Our results suggest an underlying mechanism by which the circadian protein CLOCK in inhibitory neurons participates in neuronal activity and regulates the predisposition to epilepsy.


Introduction
Epilepsy is a common neurological disorder (Scharfman, 2007) caused by an imbalance between neuronal excitation and inhibition, resulting in abnormally high synchronized neuronal firing in the brain (Fisher et al., 2017;Guerrini et al., 2019;Liu et al., 2007).Only portion of patients are responsive to anti-epileptic drugs or become responsive (Brodie et al., 2012) after surgical treatments (Perucca et al., 2007).The high prevalence, recurrence, and the development of refractory disease with multiple complications associated with epilepsy pose significant danger and difficulties to patients (Hussain et al., 2020;Singh and Sander, 2020).Therefore, more research is still needed to explore the pathogenesis and potential therapeutic targets for epilepsy.
The circadian nature of seizures has long been recognized in humans and rodents.Circadian rhythm is a biological rhythm driven by a series of biological clock proteins.Two of the key components, Circadian Locomotor Output Cycles Kaput (CLOCK) and Brain and Muscle ARNT-Like1 (BMAL1), form heterodimers and translocate to the nucleus where they promote gene transcription by interacting with the E-box promoter.PER1/2 (Period1/2), CRY1/2 (Cryptochrome 1/2), and NR1D1 (Nuclear Receptor) are involved in the positive and negative regulation of circadian oscillations by BMAL1 and CLOCK (Gachon et al., 2004;Leite Goes Gitai et al., 2019;Zhang et al., 2021).The core clock components and CLOCK control genes constitute the basic structure of the biological clock system (Patke et al., 2020).
Circadian rhythm disturbance in patients with epilepsy may build a vicious cycle leading to disease progression (Gibbon et al., 2019;Maganti and Jones, 2021).Some circadian rhythm signaling proteins were found to be reduced after seizures (Li et al., 2017;Wan et al., 2020;Zhang et al., 2021).It has also been reported that Bmal1 knockout lowered the seizure threshold (Barca-Mayo et al., 2017;Gerstner et al., 2014;Wu et al., 2021).Several studies have suggested the Rev-erbα gene promotes seizures (Griffin et al., 2019;Griffin et al., 2020;Zhang et al., 2021), whereas others have suggested that it plays a protective role in epilepsy (Andre et al., 2007;Steve et al., 2014;Yue et al., 2020).CLOCK may maintain neuronal homeostasis and regulate synaptic plasticity by altering the morphology and number of spines (Kobayashi et al., 2015;Musiek et al., 2013) which is closely associated with epilepsy (Marrone and Petit, 2002;Peebles et al., 2010;Phillips and Pozzo-Miller, 2015).Our previous study showed CLOCK reduction and neuronal dendritic spine defects in epileptogenic cortical tissues surgically removed from patients (Li et al., 2017).Loss of CLOCK in tested human samples was more obvious in excitatory neurons than in inhibitory neurons.Mice with a CLOCK-specific knockout in excitatory neurons showed decreased seizure thresholds and defects in dendritic spines (Li et al., 2017).Thus, the CLOCK dysfunction in excitatory neurons may affect predisposition to seizure by altering synaptic plasticity.
Interestingly, we observed an increased trend (although not significant) of the latency to generalized tonic-clonic seizures in mice with Clock deletion in PV inhibitory neurons compared to control mice (Li et al., 2017).This result implies that the effects of CLOCK on seizure in inhibitory neurons were different from its effects in excitatory neurons.Since PV positive neurons only account for ~40% inhibitory neurons (Hafner et al., 2019;Lee et al., 2010;Pfeffer, 2014), we hypothesized that deletion of Clock in all inhibitory neurons will significantly increase seizure threshold.To inspect this hypothesis, we established a GAD-Cre; Clock flox/flox mouse line with conditional knockout of Clock in GAD2positive inhibitory neurons, which account for all inhibitory neurons.With seizure induction by intraperitoneal pilocarpine injection, seizure behavior observation and EEG recording were used to examine and compare GAD-Cre;Clock flox/flox mice with control Clock flox/flox mice.Additionally, changes in the levels of BDNF, ERK and ARC proteins were also investigated as possible mechanisms.

Mice and genotyping
All mice were housed in a specific pathogen-free (SPF) environment with a 12 h light/ 12 h dark cycle, room temperature of 22-25 • C, 40-60% humidity, and free access to food and water.The Wenzhou Medical University Institutional Animal Care (Number: wydw2022-0466) and Use Committee approved all animal care and experimental procedures.Mice with conditional deletion of Clock in GAD2-positive neurons (GAD-Cre;Clock flox/flox ) were obtained by crossing Clock flox/flox (B6.129S4-Clock tm1Rep /J, The Jackson Laboratory, 010490) females with GAD2-Cre (B6J.Cg-Gad2 tm2(cre)Zjh /J, The Jackson Laboratory, 010802) males.The primers 9910,9911,19048,28223,28224 were used for genotyping as suggested by the Jackson Laboratory.All mice from the same litter were used in the experiments on a C56BL6 background.

Behavioral tests
We used 9-to 12-week-old age-matched male mice for behavioral tests which were performed between 9:00 AM and 5:00 PM in a sound attenuation booth and were recorded and analyzed with the ANY-maze system.During these behavioral tests, the two groups of mice were crosstested to ensure that the test time was as consistent as possible.All behavioral assays were done randomized and double-blind to genotypes.

Open-field test
Animals were individually placed in the center of an open-field arena (50 × 50 × 50 cm), with their backs facing the investigator, and allowed to explore freely for 10 min.The arena was cleaned with 75% ethanol between the trials.The experiment was carried out in a soundproof environment with suitable and consistent light intensity, temperature, and humidity.

Tail-suspension test
Mice were immobilized on a suspension bar with the tail and nose tip approximately 30 cm above the ground.After 2 min of accommodation, the time spent immobile was assessed for 4 min by an observer blinded to the treatment.

Y-maze test
The spontaneous alternation Y-maze test is mainly used to assess short-term spatial working memory.It comprises three identical opaque arms (A, B, and C).Mice are gently placed in the center of the maze and allowed to explore freely for 5 min.Waste was cleaned and wiped with alcohol to reduce interference between each test.The mice could enter three different arms (ABC, ACB, BAC, BCA, CAB, CBA) in turn.The following formula was used to calculate the spontaneous alternation rate (Miedel et al., 2017): number of spontaneous alternations total number of arms − 2 × 100

Novel object recognition test
Habituation period: Mice were allowed to move freely within the experimental setup (without objects) for 10 min.Training period: two identical fixed odorless objects were placed about 10 cm from the walls on both sides (to allow space and exploration for the mice) and the mice were allowed explore for 10 min.Test period: memory was tested 24 h after completion of the training stage.One object was replaced with a new object and the mice were allowed to explore for 10 min.The paths taken and the times taken for exploration were recognized and recorded by software.The recognition index (RI) was calculated as follows: Time to explore new object Time to explore new and old object × 100

Morris water maze test
The Morris water maze is the most widely used behavioral experiment for the evaluation of spatial learning ability and memory.It consists of a circular pool, a moveable platform hidden underwater (placed in the second quadrant), and an automatic image acquisition and processing system (ANY-Maze).The experimental procedure included (1) Positioning navigation experiment to evaluate the learning and memory acquisition of the mice.Training was conducted four times daily in all four quadrants, with a 30-min interval between each session.Each trial lasted 60 s and ended when the mouse climbed up and remained on the hidden platform for 10 s.If the mouse failed to find the platform, it was guided to the platform and kept there for 10 s.The time taken by the mouse to reach the platform was recorded as the incubation period.The times taken for the four trials were averaged for each mouse and were reported as results.The experiment was performed for five days.(2) The probe trial was used to measure the ability of the mice to remember the position of the platform after learning how to find it.The platform was removed and the mouse was placed in the water in the quadrant opposite the position of the original platform.Each mouse was tested only once, and the latency to reach the platform and the times of crossing the platform were recorded and analyzed.

Pilocarpine model
Mice (8-to 10 weeks old) were injected intraperitoneally with lithium chloride (127 mg/kg bodyweight).After 16-18 h, scopolamine (1 mg/kg body weight) was injected intraperitoneally 30 min before the pilocarpine (250 mg/kg body weight) injection to reduce the peripheral cholinergic reaction caused by pilocarpine (Hoehna et al., 2012;Lenz et al., 2017).The sham mice received the equivalent volume of saline.The mice in all groups received pilocarpine at the same time of day, between 14:00 and 17:00.Seizures were monitored using the Racine scale (Borges et al., 2003;Racine, 1972) (Table 1).Diazepam (3 mg/kg body weight) was given to terminate status epilepticus 1.5 h after the onset.Seizure latency was defined as the time from pilocarpine injection to the first stage IV and above seizure.Mice with seizures of grade IV and above that survived were used for subsequent immunofluorescence staining analysis.Lithium chloride (L9650) and Diazepam (D9900) were purchased from Sigma.Scopolamine (S0231) and Pilocarpine (P344660) were purchased from TCI chemicals and Aladdin, respectively.All the drugs were dissolved in sterile saline (0.9% NaCl).

Electroencephalograph (EEG)
Male mice (8 weeks old) were weighed, anesthetized with 1% amobarbital 0.1 ml/10 g, and placed in a stereotaxic frame with ear strips and bite plate.Veterinary ointment was applied to both eyes for protection and the mice were shaved.A longitudinal midline incision was made in the skin to expose the lambda (where the sagittal plane and lambda sutures intersect), the bregma (where the coronal suture intersects with the sagittal suture), and the target position.When the skull had dried, four holes were slowly drilled, and four sterile stainless-steel screws were placed in the epidural (two front, 2 mm and two back, mm) recording electrodes.Then myoelectric electrodes were placed under the neck muscle; the posterior was sutured and fixed with dental cement to avoid exposure to metal parts.Animals were recovered for days after electrode implantation surgery before proceeding to the next experiment.Brain electrical activity before and after pilocarpine injection (as above method) was recorded.EEG signals and synchronized video were recorded using the Medusa EEG&EMG System (Bio-Signal Technologies).EEG signals were sampled at 500 Hz.The data was analyzed using Sirenia Seizure Pro software.

Electrophysiological recordings
Mice with 3-4 weeks of age were deeply anesthetized with isoflurane and then decapitated.Brains were removed and transferred to an icecold and oxygenated high sucrose slicing solution for 2-3 min (in mM): 234 sucrose, 11 glucose, 26 NaHCO 3 , 2.5 KCl, 1.25 NaH 2 PO 4 , MgSO 4 , and 0.5 CaCl 2 .Then the brains were sectioned to crownal slices (300 μm in thickness) in slicing solution using a Leica VT1200S vibratome and recovered at 34 • C for one hour in the oxygen-saturated artificial cerebral spinal fluid (ACSF) containing the following chemicals (in mM):126 NaCl, 26 NaHCO 3 , 10 glucose, 2.5 KCl, 1.25 NaH 2 PO 4 , 2 MgCl 2 , and 2 CaCl 2 (pH 7.4).Before recording, crownal slices were first to be placed in the recording chamber and visualized with a fixedstaged upright microscope (Olympus MVX10).All recordings of layer somatosensory cortical pyramidal cells were performed with EPC-10 Patchmaster (HEKA, Germany), and with continuous ACSF perfusion (2 ml/min at room temperature), Glass pipettes (nonfilament, Garner Glass Company, CA, USA) were pulled to obtain electrodes(Model P-97, Sutter Instruments) that generated resistances between 5 and 7 MΩ after being filled with intracellular solution containing (in mM): 130 K-gluconate, 10KCl, 10 HEPES, 4 Mg-ATP, 0.3 Na 2 -GTP, and Phosphocreatine.For whole-cell electrophysiological recordings, a gigaohm seal between the cell and the glass pipette was formed first, and a brief suction was applied to break into the cell.Pyramidal cell action potential firing properties were recorded under current clamp mode.In order to investigate the morphology of each recorded pyramidal cell, biocytin (1%, B1592, ThermoFisher Scientific) was added to the intracellular solution, and cells were kept under whole-cell recording mode for at least 10 min, allowing the biocytin to diffuse well into the cell.Slices containing recorded cells were then fixed with 4% paraformaldehyde overnight and processed for post-hoc morphological characterization.All data were analyzed off-line using clampfit 10.6 software.

Statistical analysis
Intergroup differences in gene/protein expression and fluorescence intensity were analyzed by t-tests and two-way ANOVA followed by Holm-Sidak's multiple comparisons test were used to compare the intergroup seizure process.Data on epilepsy severity and mortality were compared using Fisher's exact test.The results are expressed as mean ± SEM.P-values <0.05 were considered statistically significant.Details of the statistical tests used for each analysis are provided in the Figure legends.All statistical analyses were performed using GraphPad Prism (San Diego, CA, USA) software.

Generation and identification of GAD-Cre;Clock flox/flox mice
Glutamate decarboxylase-2 (GAD2) is expressed in most inhibitory neurons.To further investigate the association of CLOCK with epilepsy, Clock knockout was introduced in GAD2-positive neurons.GAD2-Cre; Clock flox/flox mice were generated by the crossbreeding of Clock flox/flox and GAD2-Cre mice via a Cre-loxP-mediated recombinant system (Fig. 1A).GAD-Cre;Clock flox/flox mice was assessed by genotyping (Fig. 1B).Immunofluorescence staining results showed that CLOCK levels were clearly lower in the inhibitory interneurons in the cerebral cortex of GAD-Cre;Clock flox/flox mouse compared with those in the Clock flox/flox mouse (Fig. 1E-G).Western blotting showed a small but significant decrease in CLOCK protein levels in the GAD-Cre;Clock flox/flox group (t (10) = 6.062, p = 0.0001, unpaired t-test; Fig. 1C,D), comparable to the relative percentage of GABAergic interneurons in cortical neurons, which is about 20% (Tamamaki et al., 2003).

Knockout of CLOCK in inhibitory neurons decreased anxiety-like behavior
The effects of Clock knockout on the behavioral performance were assessed by multiple behavioral tests.The open field test (OFT) and the tail suspension test (TST) were used to examine the exploratory and anxiety-like behavior in mice.In OFT test, mice prefer exploring in surrounding areas in a new environment, while the inquiring nature of animals will prompt them to explore the central area.Mice with lower anxiety-like behavior tend to spend more time in the central area (Kraeuter et al., 2019).OFT tests showed that the distance ratio in the central region did not change significantly (Fig. 2A), whereas the time spent in the central region increased (t (32) = 2.773, p = 0.009, unpaired t-test; Fig. 2B) in GAD-Cre;Clock flox/flox group compared to control group.In the TST, a mouse changes from struggling to break free from a hook where it is suspended by the tail to enter a desperate state of immobility.The immobility time is a main factor in assessing the level of depressionlike behavior of mice in the TST test (Gao et al., 2023).The immobility time in the GAD-Cre;Clock flox/flox group was significantly reduced (t (20) = 6.950, p < 0.0001, unpaired t-test; Fig. 2C) compared to control Clock flox/flox mice.OFT and TST results suggest that GAD-Cre; Clock flox/ flox mice exhibit less anxiety-like or depression-like behavior.
Y-maze, the novel object recognition (NOR), and the Morris water maze tests were employed to evaluate the memory retention.In the Ymaze test, short-term spatial working memory was assessed by the spontaneous alternation rate of a mouse.The results showed no significant difference between the two groups (Fig. 2D).In the NOR test, GAD-Cre;Clock flox/flox mice showed improved memory as shown by an increased novel object recognition index after 24 h compared with the Clock flox/flox mice (t (18) = 2.385, p = 0.028, unpaired t-test; Fig. 2E, F).In the Morris water maze test, the average escape latency decreased with the increase of training days and no significant differences were observed between the two groups, manifesting no difference in the learning ability between the two groups (Fig. 2G).In the probe trial (platform withdrawal), the number of platform crossings showed no difference (Fig. 2I), whereas both the percentage of time (t (16) = 2.892, p = 0.01, unpaired t-test; Fig. 2J) and percentage of distance spent in the target quadrant (t (16) = 2.512, p = 0.023, unpaired t-test; Fig. 2K) were significantly increased in GAD-Cre;Clock flox/flox mice, indicating that the GAD-Cre;Clock flox/flox mice have improved memory.
Video EEG is the gold standard for detecting seizures with high sensitivity and precision.To confirm that GAD-Cre;Clock flox/flox mice mitigate seizures, the EEG activities of five mice from each group at baseline (before pilocarpine injection) and 30 min after pilocarpine injection were shown.Fig. (3E-G) shows the EEG trajectory of one representative mouse of each group, including the activities at baseline (Fig. 3E) and 30 min after pilocarpine administration (Fig. 3F-3G).Fig. 3H depicts the time-frequency domain analysis of 30 min after pilocarpine injection.Fig. 3I (t (8) = 3.504, p = 0.008, unpaired t-test) and Fig. 3J show the total power and spectrum power data for each frequency band (five brainwaves) using Fourier transform.The Clock flox/ flox mice showed abnormal brain waves and high power following pilocarpine injection, while the GAD-Cre;Clock flox/flox mice did not show abnormal EEG activity and increased power.EEG recording results are consistent with our observations based on Racine scores.

GAD-Cre;Clock flox/flox mice have less activated astrocytes than control group after pilocarpine injection
Progressive neuronal death and glial cell proliferation are common pathological changes observed in epilepsy (Jimenez-Mateos et al., 2012;Zhang et al., 2021).Seizure severity is indirectly reflected by glial activation.Activation of glial fibrillary acidic protein (GFAP) acts as a marker for astrocyte differentiation and its upregulation is often accompanied by a reactive response to central nervous system injury.Here, GFAP immunofluorescence staining was conducted on Clock flox/flox and GAD-Cre;Clock flox/flox mice brain tissues 24 h after seizure induction by pilocarpine.In saline group, GFAP levels were low and no significant difference was found between Clock flox/flox and GAD-Cre;Clock flox/flox mice (Fig. 4A and B).After pilocarpine injection, GFAP levels were significantly increased in Clock flox/flox mice but not in GAD-Cre;Clock flox/ flox mice (Fig. 4A-4B).Therefore, GFAP levels in Clock flox/flox mice were significantly higher than those in GAD-Cre;Clock flox/flox mice after pilocarpine induction of seizure (Fig. 4B).Decreased GFAP induction in GAD-Cre;Clock flox/flox mouse brain suggests that the mutant brain is less responsive to pilocarpine treatment or has less neuronal damage.

Cortical pyramidal neurons exhibit higher apical dendritic spine density in GAD-Cre;Clock flox/flox mice
To examine whether Clock knockout in inhibitory neurons will affect brain circuitry and neuronal morphology, electrophysiological recordings were performed on layer 5 pyramidal neurons from somatosensory cortex and these neurons were labeled with biocytin during recording for post hoc morphological study.No significant changes were found in neuronal properties (Table 2), neither in their sEPSCs between Clock flox/flox mice and GAD-Cre;Clock flox/flox mice (Fig. 6A-E).However, higher apical spine density was found in the pyramidal neurons in the GAD-Cre;Clock flox/flox mice compared with the Clock flox/flox mice (P = 0.024, unpaired t-test, Fig. 6F and G).

Discussion
Previous studies reported that the CLOCK levels in epileptic brain tissues from patients were significantly reduced (Li et al., 2017;Zhang et al., 2021), and the reductions of CLOCK were more pronounced in excitatory neurons (Li et al., 2017).Consistent with this, deletion of Clock in excitatory neurons in mouse brains resulted in spontaneous seizures and shortened latency (Li et al., 2017).To further refine the effect of CLOCK on seizures, GAD-Cre;Clock flox/flox mice with Clock gene knockout in inhibitory neurons were constructed in this study.To our surprise, GAD-Cre;Clock flox/flox mice showed prolonged seizure latency, reduced severity, and mortality in the acute phase of pilocarpineinduced seizures.Therefore, deletion of the Clock gene in different types of neurons causes different effects on epilepsy.GABAergic inhibitory neurons occupy about 20% of the total neuronal population in mammals and play a critical role in connecting upstream and descending neurons via feedback or feed-forward inhibition to maintain the excitation/inhibition balance in brain circuits (Tamamaki et al., 2003).This complex synaptic network forms the structural basis of a highly complex nervous system.One possible interpretation is that Clock deletion in inhibitory neurons increased network inhibitory inputs and the mice were more resistant to drug induced seizures than control mice.Another possibility is the reduction of excitatory inputs in GAD-Cre;Clock flox/flox mice, causing the neuronal excitation/inhibition imbalance.In this study, we examined the excitatory inputs in GAD-Cre;Clock flox/flox mice and didn't see significant changes in sEPSCs and properties in pyramidal neurons in somatosensory cortex(Fig.6A-E).It could be possible that inhibitory drives were affected in GAD-Cre;Clock flox/flox mice.A previous study found that Clock genes intrinsic to PV cells coordinate postnatal maturation of their networks and consequently the onset of critical period plasticity in visual cortex V1 (Kobayashi et al., 2015), suggesting the importance of CLOCK in critical period plasticity.Also in this study, the pyramidal neurons in layer 4 visual cortex showed reduction in the frequency of mIPSCs, suggesting the Clock knockout in PV neurons alone will affect the inhibitory drive in V1 cortex.Our study used GAD-Cre; Clock flox/flox mice and knocked down Clock in all inhibitory neurons, including the majority of neuron population in the SCN.Such change could affect inhibitory drive and plasticity in different cortical regions via different inhibitory circuitry, which remains to be clarified.No significant changes in excitatory inputs were observed in the pyramidal neurons from GAD-Cre;Clock flox/flox mice, despite the pyramidal neurons have increased spine density.Since these mice exhibit seizure resistance, it is possible that Clock knockout in inhibitory neurons alone could shift the overall excitation/inhibition balance.
Increased basal ARC protein level was observed in the GAD-Cre; Clock flox/flox mice.The Arc was initially discovered as a gene that was strongly induced by epilepsy and expressed in the cell soma and dendrites (Janz et al., 2018;Peebles et al., 2010;Sibarov et al., 2023;Yu et al., 2022;Zhang and Bramham, 2021).Moreover, ARC was identified as a candidate gene involved in the pathogenesis of many neurological diseases, including epilepsy (Korb and Finkbeiner, 2011;Shepherd and Bear, 2011;Sibarov et al., 2023;Yu et al., 2022).ARC regulates synaptic plasticity and neuronal excitability by altering spine morphology and mediating the endocytosis of AMPA receptors (Chowdhury et al., 2006;Peebles et al., 2010;Shepherd et al., 2006).Synaptic plasticity is associated with many brain activity and pathological processes, including epilepsy (Peebles et al., 2010).The pathogenesis of epilepsy is closely related to excitatory and inhibitory synaptic imbalances and abnormalities (Risher et al., 2018).Our previous study indicated that the defects in the dendritic spines of neurons were associated with decreased CLOCK expression in excitatory neurons (Li et al., 2017).CLOCK dysfunction regulates spines morphology and alters synaptic function (Kobayashi et al., 2015;Li et al., 2017;Musiek et al., 2013).The regulation of the number and morphology of spines influences synaptic plasticity.Neuronal plasticity may be the bridge between the biological clock and epilepsy.Arc − /− mice is susceptible to seizures induced by convulsant drugs (Peebles et al., 2010).Arc − /− mice showed an epilepticlike network hyperexcitability and altered dendrite spine morphology with fewer thin spines and more mushroom spines.,These results suggest that ARC is important in maintaing excitatory-inhibitory balance in a neuronal network and preventing hyperexcited network activity.Therefore, we hypothesize that the anticonvulsant effects observed in GAD-Cre;Clock flox/flox mice at least partially contributes from the increased basal ARC levels, which has prevention effects on hyperexcitability.
Our study also showed that both BDNF and p-ERK levels were upregulated in the GAD-Cre;Clock flox/flox mice.ARC can be induced by BDNF activation of the mitogen-activated protein kinase (MAPK) signaling pathway (Rao et al., 2006;Ying et al., 2002).Studies suggested that BDNF could be used to treat epilepsy, and its continuous infusion in the hippocampus delays epilepsy latency, reduces seizure frequency, and improves cognitive performance (Falcicchia et al., 2018).Viral vector-mediated BDNF supplementation in rat hippocampal regions  alleviates neuronal death, favors neurogenesis, and reduces both seizure frequency and severity (Paradiso et al., 2011).Previous study showed that ERK phosphorylation of Kv4.2 (a fast-inactivating A-type potassium channel) may represent a protective mechanism by which hyperexcitability in hippocampal neurons is limited (Berkeley et al., 2002;Ying et al., 2002).Therefore, ERK might limit the hyperexcitability of hippocampal neurons through phosphorylating ion channels to have protective effect against pilocarpine-induced seizures.The basal levels of ARC protein are upregulated likely by activated BDNF and the ERK signaling pathway.These proteins together might play protective roles and contribute to the seizure resistant response observed in GAD-Cre; Clock flox/flox mice.
In addition to their effects on seizure, BDNF and ARC are also involved in the memory and pathogenesis of depression (Aboul-Fotouh, 2015;Chiba et al., 2012;Dunham et al., 2009;Kunugi et al., 2010;Mizuno et al., 2000;Palop et al., 2005;Ying et al., 2002).Decreasing BNDF synthesis leads to memory impairment, which can be reversed after increasing BDNF expression (Mizuno et al., 2000).In a rat depression model induced by chronic restrain stress, BDNF expression in the prefrontal cortex is decreased (Chiba et al., 2012).When rats were treated with mecamylamine, a nicotinic acetylcholine receptor antagonist, prefrontal cortex level of BDNF was increased and chronic restrain stress-induced depressive-like behaviors were prevented (Aboul-Fotouh, 2015).These studies suggest that BDNF has promoting effects on memory and inhibiting effects on depression.BDNF enhances LTP by activating MEK-ERK and inducing ARC expression (Ying et al., 2002).ARC itself also plays a role in consolidating memory and maintaining LTP (Guzowski et al., 2000;Palop et al., 2005;Plath et al., 2006b;Zhang and Bramham, 2021).In Arc knockout mice or in mice whose basal levels of ARC protein were decreased by 60% through intrahippocampal infusions of antisense oligodeoxynucleotides specific for Arc mRNA, both LTP and Long-term memory were impaired, suggesting that Arc plays a role in memory formation (Guzowski et al., 2000;Plath et al., 2006a).Based on these publications, the elevated BDNF and ARC levels could contribute to the improved memory and reduced depressive-like behavior observed in GAD-Cre;Clock flox/flox mice.
In this study, we found that CLOCK deletion in inhibitory neurons alleviated seizures and improved mood and memory, complementing our previous study in which CLOCK deletion in excitatory neurons promoted seizures and impaired dendrite spines.This study provides a novel finding that loss of CLOCK in inhibitory neurons decreases the susceptibility to epilepsy.We only showed that GAD-Cre;Clock flox/flox mice were resistance to seizure induced by pilocarpine in the acute phase, further studies are needed to evaluate CLOCK's effects on the chronic phase of seizure induction.More work is needed to fully elucidate the mechanism of CLOCK roles in different types of neurons to regulate neuronal activities associated with seizure.

Fig. 2 .
Fig. 2. Knockout of the Clock gene in inhibitory neurons affects mood and memory capacity, shown by behavioral tests.(A-B) The ratio of distance (A) and time (B) spent at the center in an open field test (n = 19 for Clock flox/flox group, n = 15 for GAD-Cre;Clock flox/flox group).(C) Immobility times in the tail suspension test (n = 12 for Clock flox/flox group, n = 10 for GAD-Cre;Clock flox/flox group).(D) Spontaneous alternations rate in the Y-maze test (n = 12 for Clock flox/flox group, n = 10 for GAD-Cre;Clock flox/flox group).(E) Simple diagram of the test process and representative trajectory heatmap of the testing period for NOR test.(F) During the training time of NOR test, the time ratio of exploration the object was calculated.The object was replaced by new object on day 3 for NOR test.The right panel shows the recognition index in the novel object recognition (NOR) test (n = 10 for each group).(G-K) Morris water maze test results.Escape latency during training days (G) and representative trajectory heatmap (H), platform-crossing frequency (I), percentage of time (J), percentage of distance (K) spent in the target quadrant in the probe trial (without a platform) in the Morris water maze test (n = 9 mice for each group).Data are represented as means ± SEM (each dot represents one mouse), repeated measures two-way ANOVA was used for (G), with unpaired Student's t-tests used for the rest.Ns, not significant, *P < 0.05, **P < 0.01 and ****P < 0.0001.

Fig. 4 .
Fig. 4. GAD-Cre;Clock flox/flox mice showed reduced astrocyte proliferation compared with Clock flox/flox mice 24 h after pilocarpine injection.(A) GFAP and NeuN staining of the hippocampus, 24 h after saline or pilocarpine (250 mg/kg) injection in Clock flox/flox and GAD-Cre;Clock flox/flox mice.GFAP (green): Glial fibrillary acidic protein; NeuN (red): neuronal nuclei; scale bar = 100 μm.(B) Quantification of GFAP in hippocampus 24 h after saline or pilocarpine-injection in Clock flox/flox and GAD-Cre;Clock flox/flox mice.Data are presented as mean ± SEM and analyzed by two way ANOVA and Sidak's post-test.N = 4 for saline group, n = 5 for pilocarpine group, ns: not significant, *P < 0.05, **P < 0.01.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 1
Racine scale is used to monitor the seizures.

Table 2
Active and passive membrane properties of excitatory neurons in mouse somatosensory cortex.