Clock Gene Regulates Kainic Acid-Induced Seizures Through Inhibiting Ferroptosis in Mice

Temporal lobe epilepsy (TLE) is a common and intractable form of epilepsy. There is a strong need to better understand molecular events underlying TLE and to nd novel therapeutic agents. Here we aimed to investigate the role of Clock gene and ferroptosis in regulating TLE. TLE model was established by treating mice with kainic acid (KA). Regulatory effects of Clock gene on KA-induced seizures and ferroptosis were evaluated using Clock knockout (Clock −/− ) mice. mRNA and protein levels were determined by quantitative real-time PCR and western blotting, respectively. Ferroptosis was assessed by measuring the levels of iron, GSH and ROS. Transcriptional regulation was studied using a combination of luciferase reporter, mobility shift and chromatin immunoprecipitation (ChIP) assays. We found that Clock ablation exacerbated KA-induced seizures in mice, accompanied by enhanced ferroptosis in the hippocampus. Furthermore, Clock ablation reduced the hippocampal expression of GPX4 and PPAR-γ, two ferroptosis-inhibitory factors, in mice and in N2a cells. Moreover, Clock regulates diurnal expression of GPX4 and PPAR-γ in mouse hippocampus and rhythmicity in KA-induced seizures. Consistently, Clock overexpression up-regulated GPX4 and PPAR-γ, and protected against ferroptosis in N2a cells. In addition, based on luciferase reporter, mobility shift and ChIP assays, we uncovered that CLOCK protein trans-activated Gpx4 and Ppar-γ through specic binding to an E-box element in gene promoters. In conclusion, CLOCK protests against KA-induced seizures through promoting expression of GPX4 and PPAR-γ and inhibiting ferroptosis. Regulatory effects of Clock gene on KA-induced seizures and ferroptosis were evaluated using Clock knockout (Clock -/- ) mice. mRNA and protein levels were determined by quantitative real-time PCR (qPCR) and western blotting, respectively. Ferroptosis was assessed by measuring the levels of iron, GSH and ROS. Transcriptional regulation was studied using a combination of luciferase reporter, mobility shift and chromatin immunoprecipitation (ChIP) assays. We have demonstrated for the rst time that CLOCK protests against KA-induced seizures through promoting expression of GPX4 and PPAR-γ and inhibiting ferroptosis.


Introduction
Epilepsy is a brain disease characterized by unprovoked recurrent seizures, affecting over 70 million people worldwide. Temporal lobe epilepsy (TLE) is a common and intractable form of epilepsy and associated with pathologic changes in hippocampal physiology and morphology , . Approximately 30% of patients with epilepsy remain intractable to antiseizure drugs, and this subset of patients is usually diagnosed with TLE. Thus, there is a strong need to better understand molecular events underlying TLE and to nd novel therapeutic agents. Kainic acid (KA)-induced seizure is regarded as a good model of TLE as the behavioral seizures and neuropathological lesions in the animals are highly similar to those of TLE patients.
Ferroptosis is an iron-dependent, oxidative form of non-apoptotic regulated cell death, characterized by an increase in free iron and accumulation of lipid peroxides. GPX4 (glutathione peroxidase 4) is considered as a key regulator of ferroptosis. It converts the toxic phospholipid hydroperoxides (lipid-OOH) to nontoxic phospholipid alcohols (lipid-OH) by utilizing an electron donated by GSH (glutathione). Thus, ferroptosis can be triggered by inhibition of GPX4 or disruption of GSH synthesis 7, . Ferroptosis has been implicated in the development of various pathological conditions such as cancers, neurodegeneration, ischemia reperfusion injury and acute kidney injury. Moreover, various ferroptosis inducers and inhibitors have been shown to modulate disease progression in preclinical models. Therefore, targeting ferroptosis may provide a new avenue for disease management.
PPARs (peroxisome proliferator-activated receptors) are a family of ligand-responsive nuclear receptors consisting of three members, namely, PPAR-α, PPAR-β (or PPAR-δ) and PPAR-γ. PPARs can be activated by endogenous ligands such as fatty acids and their derivatives. Upon ligand binding, PPARs form heterodimers with retinoid X receptors and bind to a speci c DNA response element (PPRE) in promoter to regulate gene transcription. PPAR isoforms show tissue-speci c differences in their expression and functions 12 . Although PPAR-γ is abundantly present in adipose tissues, it can be found in many other tissues such as the liver and hippocampus. PPAR-γ has a critical role in regulating lipid metabolism, insulin sensitivity, tumor cell growth, apoptosis and differentiation. In addition, PPAR-γ has the potential to activate defensive mechanisms against lipoperoxidative reactions by controlling the expression of antioxidant enzymes, thereby inhibiting ferroptosis , .
Various aspects of physiology and behaviors in mammals are subjected to circadian rhythms that are driven and maintained by the circadian clock system. Disruption of circadian rhythms is linked to various disorders such as depression, diabetes, cancers, and cardiovascular diseases. At the molecular level, the circadian clock system contains multiple transcriptional-translational feedback loops in which various positive and negative components regulate the expression of clock-controlled genes (CCGs). Clock (circadian locomotor output cycles kaput) is one of core positive regulatory genes in the circadian clock system and is expressed in virtually all tissues. It binds to a speci c response element (called "E-box") in the promoters of target genes and activates gene transcription. In addition to regulating circadian rhythms, Clock is also involved in regulation of many other physiological processes including cell cycle, lipid metabolism, glucose metabolism, and immune responses. It is therefore of no surprise that mutations in Clock gene are associated with pathological conditions such as osteoarthritis, atherosclerosis, heart failure, tumorigenesis and bipolar disorder -.
In the present study, we aimed to investigate the role of Clock gene and ferroptosis in regulating TLE on the basis of KA-induced seizure model. Regulatory effects of Clock gene on KA-induced seizures and ferroptosis were evaluated using Clock knockout (Clock -/-) mice. mRNA and protein levels were determined by quantitative real-time PCR (qPCR) and western blotting, respectively. Ferroptosis was assessed by measuring the levels of iron, GSH and ROS. Transcriptional regulation was studied using a combination of luciferase reporter, mobility shift and chromatin immunoprecipitation (ChIP) assays. We have demonstrated for the rst time that CLOCK protests against KA-induced seizures through promoting expression of GPX4 and PPAR-γ and inhibiting ferroptosis.

KA-induced seizures
Clock -/mice and wild-type littermates (WT/Clock +/+ ) were treated with KA (i.p., 20 mg/kg) to induce acute seizures (status epilepticus). The stages in the status epilepticus were recorded according to the Racine scale as previously described: 0 (no response), 1 (staring and reduced locomotion), 2 (head nodding), 3 (unilateral forelimb clonus), 4 (bilateral forelimb clonus), 5 (rearing and falling), and 6 (status epilepticus and death). Seizure severity was assessed by integrating individual scores per mouse over the duration of the experiment. 24 h after KA treatment, mice were sacri ced to collect hippocampus. Hippocampal iron, MDA and 4-HNE were measured using their assay kits. In order to collect electroencephalogram (EEG), mice were subjected to electrode implantation prior to KA induction as described in our recent publication.

H&E, FJB and TUNEL staining
Mouse brain tissues were xed in 4% paraformaldehyde and embedded in para n. 4-µm-thick coronal sections from the hippocampus were prepared for conventional H&E, FJB and TUNEL staining as described 30, . Images were acquired using a Nikon Optiphot uorescent microscope (Tokyo, Japan). Degenerate/dead neurons, FJB-and TUNEL-positive cells were identi ed and counted using ImageJ (National Institutes of Health, Bethesda, MD). At least three regions from each section and three sections were imaged for each animal.

Immuno uorescence
Immuno uorescence staining for NeuN and GFAP was performed to examine the neuronal apoptosis and astrocytes in mouse hippocampus as described in our recent publication 30 .

Quantitative real time PCR (qPCR)
Total RNA was isolated using RNAiso Plus (Takara, Otsu, Japan), and subjected to reverse transcription with HiScript II Q RT SuperMix (Vazyme, Nanjing, Jiangsu, China). qPCR was performed using SYBR Green Master Mix (Vazyme, Nanjing, China) and determined using previously described reaction conditions. The sequences of all primers are shown in Table 1. The speci city of all primers was checked by BLAST analysis (http://www.ncbi.nlm.nih.gov/BLAST/). Relative gene expression levels were calculated by the equation 2 -∆∆Ct and normalized to housekeeping gene peptidylprolyl isomerase B (Ppib).

Western blotting
Total protein was extracted with RIPA lysis buffer. Protein concentration was measured by using a BCA kit (Beyotime, Shanghai, China). The protein samples were separated by 10% SDS-PAGE and then transferred to PVDF membranes. Membranes were blocked with 5% skimmed milk for 1 h, and incubated with primary antibodies (anti-CLOCK, anti-GPX4, anti-PPAR-γ and anti-GAPDH) over night at 4°C. On the next day, the membranes were incubated with secondary antibody at room temperature for 1 h. Protein bands were imaged by using Omega Lum G imaging system (Aplegen, San Francisco, CA) and quanti ed by uorchem 5500 software (Alpha Innotech, San Leandro, CA). GAPDH was used as a loading control.

Electrophoretic mobility shift assay (EMSA)
EMSA was performed using an EMSA kit (Beyotime, Shanghai, China) according the manufacturer's instruction. In brief, nuclear extract from N2a cells were incubated with biotin-labeled probe (unlabeled probe or unlabeled mutated probe was added for competitive experiments) in EMSA binding buffer. The mixture was subjected to 4% nondenaturing PAGE and transferred onto a Hybond-N + membrane (Amersham, Buckinghamshire, UK). The membrane was incubated with enhanced chemiluminescent and visualized by using Omega Lum G imaging system (Aplegen, San Francisco, CA). Oligonucleotide probes are provided in Table 1.

ChIP
ChIP assays were performed using a SimpleChip plus Enzymatic Chromatin IP kit according to the manufacturer's protocol (Cell Signaling Technology, Beverly, MA). In brief, hippocampal tissues were cross-linked with 1.5% formaldehyde and terminated by glycine. After digestion with micrococcal nuclease and sonication, the sheared chromatin was immunoprecipitated overnight with anti-CLOCK antibody or normal rabbit IgG (as a control). Immunoprecipitated chromatin was decross-linked at 65℃ and puri ed by spin columns. Puri ed DNAs were used as templates for qPCR with speci c primers (Table 1).

Statistical analysis
All data are presented as mean ± standard error of the mean (SEM). Statistical differences between two groups were analyzed by Student's t-test. One-way or Two-way ANOVA followed by Bonferroni post hoc test was used for multiple group comparisons. The level of signi cance was set at p < 0.05 (*).

Clock ablation exacerbates KA-induced seizures in mice
Mice were treated with KA to induce acute seizures (status epilepticus). Clock ablation in mice resulted in accelerated progression of behavioral seizure stages, and increased seizure severity and duration ( Figure  1A/B/C). Moreover, loss of Clock in mice increased the frequency of seizures after KA induction based on the EEG recordings ( Figure 1D). These ndings indicated a critical role of Clock gene in KA-induced seizures. The pathological hallmarks of epileptic seizures include neuronal loss and gliosis. Based on H&E, FJB and TUNEL staining, Clock -/mice showed a higher level of neuron death in the hippocampus ( Figure 1E). Supporting this, Clock -/mice had a reduced number of living neurons (indicated by NeuN + cells) and increased astrogliosis (indicated by GFAP + cells) in the hippocampus ( Figure 1F). Taken together, these data indicated that Clock ablation exacerbates KA-induced seizures in mice.

Clock ablation promotes ferroptosis in mice
Ferroptosis has been implicated in epileptogenesis , . We showed that the potent iron chelator DFO, an inhibitor of ferroptosis, signi cantly alleviated KA-induced seizures in wild-type mice (Figure 2A

Clock ablation reduces hippocampal expression of GPX4 and PPAR-γ, two ferroptosis-inhibitory factors, in mice
We next investigated the mechanisms by which Clock gene regulates ferroptosis. We analyzed the expression of ferroptosis-related genes (Acsl4, Bid, Dmt1, Dpp4, Fth, Ftl, Gpx4, Hspa5, Lpcat3, Nfs1, Nox1, Nrf2, Ppar-γ, Se, Slc7a11, Tfrc and Trp53) in the hippocampus in Clock -/versus control mice after KA induction. Of these tested genes, Gpx4 and Ppar-γ (two ferroptosis-inhibitory genes) mRNAs were considerably lower in Clock -/than in control mice ( Figure 3A). Consistently, Clock -/mice had lower levels of GPX4 and PPAR-γ proteins ( Figure 3B). We also assessed the effects of Clock on GPX4 and PPAR-γ in mice without KA treatment. Likewise, Clock ablation in mice reduced the mRNA and protein expression of both GPX4 and PPAR-γ in the hippocampus ( Figure 3C/D). Altogether, Clock ablation reduces hippocampal expression of GPX4 and PPAR-γ, two ferroptosis-inhibitory factors, in mice.

Clockregulates diurnal expression of GPX4 and PPAR-γ in mouse hippocampus and rhythmicity in KA-induced seizures
Clock is a circadian clock gene, whose expression oscillates with time of the day 28 . We con rmed that CLOCK protein was rhythmically expressed in mouse hippocampus with a nadir at ZT10 in wild-type mice ( Figure 4A). Interestingly, GPX4 and PPAR-γ expression also varied according to the circadian time in the hippocampus, and their diurnal patterns were similar to that of CLOCK protein ( Figure 4A). However, Clock ablation reduced the expression levels of both GPX4 and PPAR-γ, and blunted their diurnal rhythms ( Figure 4A). Therefore, Clock gene had an important role in regulating the rhythmicity in GPX4 and PPAR-γ expression. Previous studies have shown that epileptic seizures in rodents display circadian rhythms, and seizures tend to occur more frequently in the later light than in the dark phase -. Consistently, we found that in wild-type mice the seizures were more severe when KA was injected at ZT10 and less severe when 3.5. Clock overexpression up-regulates GPX4 and PPAR-γ, and protects against ferroptosis in N2a cells RSL3, a known inducer of ferroptosis, was used to induce ferroptosis in N2a cells. As expected, RSL3treated cells showed reduced viability, increased ROS and decreased GSH ( Figure 5A/B). Clock silencing (by siRNA) resulted in more extensive ferroptosis as evidenced by lower levels of cell viability and GSH as well as higher ROS accumulation ( Figure 5A). In contrast, Clock overexpression led to increased cell viability and GSH, as well as decreased ROS accumulation, suggesting attenuation of RSL3-induced ferroptosis ( Figure 5B). Moreover, we found that the mRNAs of Gpx4 and Ppar-γ were reduced in Clocksilenced cells, but increased in Clock-overexpressed cells after RSL3 induction ( Figure 5C). Likewise, Clock showed a similar effects on the expression of GPX4 and PPAR-γ in normal N2a cells (without RSL3 induction). To be speci c, Clock silencing led to decreased mRNA and protein expression of both GPX4 and PPAR-γ ( Figure 6A/B). Overexpression of Clock resulted in elevated mRNA and protein levels of both GPX4 and PPAR-γ ( Figure 6C/D). Altogether, Clock gene regulates ferroptosis in N2a cells probably through modulating the expression of GPX4 and PPAR-γ.

CLOCK regulates Gpx4 and Ppar-γ transcription via a Ebox element
Since CLOCK protein functions as a transcriptional activator, we next investigated whether it regulates GPX4 and PPAR-γ expression via a transcriptional mechanism. Clock overexpression plasmid dosedependently increased the Gpx4 (-1974/+38 bp)-Luc and Ppar-γ (-1900/+100 bp)-Luc reporter activities according to luciferase reporter assays ( Figure 7A). Consistently, the siRNA targeting Clock gene reduced the Gpx4-Luc and Ppar-γ-Luc reporter activities ( Figure 7B). Based on sequence analysis with in silico algorithm (Jasper), we found three E-boxes (putative motif for CLOCK binding and action) in Gpx4 promoter and in Ppar-γ promoter ( Figure 7C/D). Truncation and mutation experiments demonstrated that -63 bp E-box of Gpx4 and 67 bp E-box of Ppar-γ were required for CLOCK actions, while other predicted Eboxes were not ( Figure 7C/D). EMSA assays further con rmed direct interactions of CLOCK protein with the identi ed E-boxes in Gpx4 and Ppar-γ ( Figure 7E). According to ChIP assays, CLOCK protein can be recruited to Gpx4 and Ppar-γ promoter sequences (containing the E-box element) in the hippocampus of wild-type mice ( Figure 7F). However, such recruitment was lost in the hippocampus of Clock -/mice ( Figure 7F). Taken together, CLOCK protein trans-activated Gpx4 and Ppar-γ through speci c binding to an E-box element in gene promoters.

Discussion
In this study, we have de ned a protective role of Clock gene in KA-induced seizures in mice (Figure 8).
More importantly, we have uncovered that Clock regulates epileptic seizures through inhibiting ferroptosis in the hippocampus. The evidence for the links between Clock, ferroptosis and epileptic seizures is strong. First, DFO, an inhibitor of ferroptosis, signi cantly alleviated KA-induced seizures in both wild-type and Clock −/− mice ( Figure 2). Second, Clock ablation promoted ferroptosis in mice and in N2a cells, whereas Clock overexpression protected against ferroptosis in N2a cells ( Figure 5). Third, Clock positively regulated the transcription and expression of both GPX4 and PPAR-γ, two ferroptosis-inhibiting factors, through speci c binding to an E-box element in target gene promoters ( Figure 7). Therefore, Clock gene protests against KA-induced seizures through promoting expression of GPX4 and PPAR-γ and inhibiting ferroptosis. Our ndings provide increased understanding of the complex pathways for regulation of epileptic seizures by circadian clock.
Previous studies have shown that epileptic seizures in rodents display circadian rhythms, and seizures tend to occur more frequently in the later light than in the dark phase 35,37 . It is consistent to nd that herein in wild-type mice the seizures were more severe when KA was injected at ZT10 and less severe when KA was injected at ZT22 (Figure 4). Interestingly, the circadian time-dependency of seizure severity was considerably attenuated in Clock -/mice (Figure 4), indicating involvement of Clock in circadian regulation of epileptic seizures. Therefore, the circadian rhythm in seizures appears to arise from the rhythmicity in the extent of ferroptosis caused by circadian expression of GPX4 and PPAR-γ that are directly driven by the CLOCK protein. This is supported by the fact that GPX4 and PPAR-γ share similar diurnal patterns with that of the CLOCK protein ( Figure 4).
Considering that Clock gene has a protective role in epileptic seizures and is down-regulated in epileptic tissues, it is most likely involved in epileptogenesis. We have previously shown that REV-ERBα (a direct target of CLOCK) drives the expression of GABA transporters and enhances GABA reuptake, thereby alleviating GABA-mediated inhibition and promoting epileptic seizures 30 . Li et al found that deletion of Clock in pyramidal cells causes seizures during sleep in mice 39 . This phenotype was linked to the alteration of cortical circuits 39 . Therefore, although we have shown that Clock gene protests against KAinduced seizures through inhibiting ferroptosis, there is a possibility that other mechanisms are involved, such as disruption of GABAergic function and alteration of cortical circuits.
KA is an excitatory amino acid. Treatment with KA can cause epileptic seizures in the hippocampus.
These seizures propagate to other limbic structures and the induced neuropathlogical changes in the hippocampus are comparable to those of patients with TLE 5 . Thus, the Clock gene may be a promising drug target for treating TLE as small molecules targeting CLOCK such as CLK8 are being identi ed and synthesized. We found that KA-induced seizures were associated with ferroptosis. This support the notion that ferroptosis contributes to epileptic seizures 33,34 . We showed that the ferroptosis inhibitor DFO reduced (a ~50% reduction) the severity of seizure in wild-type mice ( Figure 2). The reducing effect of DFO was also observed in Clock knockout mice. However, the reduction (~30%) is less evident (Figure 2). This was probably because a higher level of ferroptosis caused by Clock deletion was more di cult to be repressed by DFO.
We have identi ed CLOCK as a positive regulator of PPAR-γ based on several lines of evidence. First, Clock ablation reduces hippocampal expression of PPAR-γ at both mRNA and protein levels in mice ( Figure 3). Second, Clock knockdown reduces, whereas Clock overexpression increases, the expression of PPAR-γ in N2a cells ( Figure 6). Third, based on a combination of luciferase reporter, mobility shift and ChIP assays, CLOCK protein trans-activated Ppar-γ through speci c binding to an E-box element in gene promoter ( Figure 7). However, Reitz et al found that Ppar-γ mRNA expression was up-regulated in the heart in Clock ▲19/▲19 (exon 19 deletion of Clock gene and a 51-amino-acid deletion of CLOCK protein) mice, suggesting a negative regulatory effect of Clock on Ppar-γ 25 . Although the exact reason for this contradiction was unknown, there was a possibility that the positive versus negative action of CLOCK was tissue dependent. This is because the activity of CLOCK is strongly affected by the cellular microenvironments such as the redox state and the types of cofactors.
It was noteworthy that mouse Ppar-γ gene generates two different mRNA transcripts (i.e., Ppar-γ1 and Ppar-γ2) using two distinct promoters. The primer for qPCR assay of Ppar-γ in current study cannot discriminate these two transcripts. Thus, mRNA expression of Ppar-γ here measures the total level of Ppar-γ1 and Ppar-γ2. Although both transcripts are expressed in mouse hippocampus, Ppar-γ1 is the dominant form as the expression ratio of Ppar-γ1 over Ppar-γ2 is more than 7 and a speci c antibody fails to detect Ppar-γ2 protein in the hippocampus. Therefore, the western blots most likely detected the Ppar-γ1 protein in the hippocampus, and the diurnal pattern may be the authentic rhythm of Ppar-γ1 protein ( Figure 4).
In summary, we have demonstrated that Clock gene protests against KA-induced seizures through promoting expression of GPX4 and PPAR-γ and inhibiting ferroptosis. Our ndings enhance a deeper understanding of the crosstalk between circadian clock and epileptic seizures. Targeting CLOCK protein may provide a promising approach for management of epileptic seizures.  images and quantitative analysis of NeuN+ cell and GFAP+ cell for hippocampus from Clock-/-and WT mice. Data are mean ± SEM (n = 6). *p < 0.05 (t-test). KA, kainic acid.  Data are mean ± SEM (n = 6). *p < 0.05 (t-test). KA, kainic acid.    promoters. In all panels except E, data are mean ± SEM (n = 6). *p < 0.05 (t-test).

Figure 8
Schematic diagram showing the mechanism for CLOCK regulation of GPX4 and PPAR-γ. CLOCK inhibits ferroptosis by promoting the transcription of Gpx4 and Ppar-γ via direct binding to an E-box cis-element, thereby protecting against KA-induced epileptic seizures.