PLPP/CIN Regulates Seizure Activity by the Differential Modulation of Calsenilin Binding to GluN1 and Kv4.2 in Mice

Calsenilin (CSEN) binds to Kv4.2 (an A-type K+ channel) as well as N-methyl-D-aspartate receptor (NMDAR), and modulates their activities. However, the regulatory mechanisms for CSEN-binding to Kv4.2 or NMDAR remain elusive. Here, we demonstrate the novel role of pyridoxal-5′-phosphate phosphatase/chronophin (PLPP/CIN), one of the cofilin-mediated F-actin regulators, in the CSEN binding to Kv4.2 or GluN1 (an NMDAR subunit). PLPP/CIN dephosphorylated CSEN in competition with casein kinase 1, independent of cofilin dephosphorylation. As compared to wild-type mice, PLPP/CIN transgenic (PLPP/CINTg) mice showed the enhancement of Kv4.2–CSEN binding, but the reduction in CSEN–GluN1 binding. In addition, PLPP/CINTg mice exhibited the higher intensity (severity), duration and progression of seizures, but the longer latency of seizure on-set in response to kainic acid. PLPP/CIN knockout mice reversed these phenomena. Therefore, we suggest that PLPP/CIN-mediated CSEN dephosphorylation may play an important role in the functional coupling of NMDAR and Kv4.2, which regulates the neuronal excitability.


INTRODUCTION
Epilepsy is one of the common neurological disorders, suffering about 1% of all people, which is characterized by the periodic occurrence of seizures exhibiting abnormal synchronized neuronal discharges. A delicate balance between excitatory and inhibitory inputs in neurons and impairment of this balance lead to the uncontrolled neuronal hyperexcitability, which initiates neuronal damage and symptomatic seizure activity. Unprovoked recurrent seizures contribute to a cyclical or progressive process of worsening epilepsy and neurological deficits including learning disabilities and memory problems (Elger et al., 2004;Kim J.E. et al., 2007;Kim et al., 2008;Seeburg and Sheng, 2008;Mantegazza et al., 2010;Kim and Kang, 2011). Although the pathophysiology underlying seizure susceptibility is relevant to channelopathy, aberrant synaptic organization, impaired glial function, inflammation, and neuronal loss (Kang et al., 2006;McNamara et al., 2006;Wetherington et al., 2008;Vezzani et al., 2011), the cellular and molecular mechanisms in epilepsy remain unclear.
Pyridoxal-5 -phosphate phosphatase/chronophin (PLPP/CIN) is a serine (S) protein phosphatase, which activates cofilin-mediated F-actin depolymerization (Gohla et al., 2005;Kim et al., 2008Kim et al., , 2016. We have reported that PLPP/CIN expression is up-regulated in the rat hippocampus following pilocarpine-induced status epilepticus (SE), which is closely relevant to increased neuronal excitability . Recently, we have also reported that PLPP/CIN transgenic (PLPP/CIN Tg ) mice show the enhancement of N-methyl-D-aspartate receptor (NMDAR)-mediated synaptic plasticity, but PLPP/CIN knockout (PLPP/CIN −/− ) mice show reduced synaptic strength without the synaptic NMDAR distribution and the heterotrimerization (Kim et al., 2016). Therefore, PLPP/CIN would be a potentially important factor to regulate seizure activity via interaction with NMDAR. However, the PLPP/CIN-mediated signal pathways available to identify seizure susceptibility are less defined.
Calsenilin (CSEN) is first identified as a Ca 2+ -regulated transcriptional repressor, contains four Ca 2+ -binding EF hand domains and belongs to the neuronal calcium sensor family (Burgoyne, 2007). CSEN has multifunctional properties to block gene expression via direct binding with the downstream regulatory element sequence in target genes (Gomez-Villafuertes et al., 2005), and to act as a partner for Kv4.x subunits (Morohashi et al., 2002). Thus, CSEN is also named downstream regulatory element antagonist modulator (DREAM) or potassium channel interacting protein 3 (KChIP3). Indeed, CSEN regulates the neuronal excitability by diminishing dendritic excitability and back-propagated action potentials accompanied by Kv4.x channels (Hoffman et al., 1997;An et al., 2000;Cai et al., 2004). Furthermore, CSEN expression is markedly decreased in the hippocampi of experimental animal model and human epilepsy patient (Hong et al., 2003). Interestingly, CSEN is one of the negative modulators of NMDAR (Zhang et al., 2010). Since NMDAR and Kv4.2 reciprocally regulate each other Jung et al., 2008) and CSEN activity is modulated by the casein kinase 1 (CK1)-mediated phosphorylation (Choi et al., 2003), it is likely that PLPP/CINmediated CSEN dephosphorylation would be involved in the pathophysiology of epilepsy by modulating the bindings of CSEN to NMDAR or Kv4.2 channels. Here, we demonstrate the novel role of PLPP/CIN in CSEN dephosphorylation that regulated seizure activity induced by kainic acid (KA) via the modulation of Kv4.2-CSEN or GluN1-CSEN bindings, independent of cofilinmediated F-actin depolymerization. Therefore, we suggest that PLPP/CIN may play an important role in the functional coupling of Kv4.2 and NMDAR via CSEN, which regulates the neuronal excitability.

Experimental Animals and Chemicals
PLPP/CIN Tg (C57BL/6J background) and PLPP/CIN −/− mice (129/SvEv-C57BL/6J background) were used in the present study. The same genetic background mice were used as control animals for each group (Kim et al., 2016). Animals were provided with a commercial diet and water ad libitum under controlled conditions (22 ± 2 • C, 55 ± 5% and a 12:12 light/dark cycle). All reagents were obtained from Sigma-Aldrich (St. Louis, MO, United States), except as noted. The antibody used in the present study is listed in Supplementary

Surgery
Under anesthesia with isoflurane (3% induction and 1.5% maintenance in a 65:35 mixture of N 2 O:O 2 ), animals were positioned over a heated pad, and core temperature was maintained 37-38 • C. An infusion guide cannula with osmotic pump connector (3260PGA, Plastics One, Roanoke, VA, United States) was implanted into the ventricle (2 mm depth from bregma), and connected to an osmotic pump (1003D, Alzet, Cupertino, CA, United States) containing (1) vehicle, (2) IC261 (3 µM), or (3) CKI-7 (100 µM). The pump was subcutaneously placed in the interscapular region Kang, 2011, 2015;Kim et al., 2014). In some animals, a standard infusion guide cannula (C315GA, Plastics One, Roanoke, VA, United States) was implanted into the same site. Thereafter, the cannula was sealed with a dummy cannula. Animals were also implanted monopolar stainless steel electrode (Plastics One, Roanoke, VA, United States) into the left dorsal hippocampus (2 mm posterior; 1.25 mm lateral; 2 mm depth from bregma). Three days after surgery, freely moving animals were given KA as followed.

Seizure Induction and EEG Recording
After baseline recording for at least 30 min, an internal infusion cannula (C315IA, Plastics One, Roanoke, VA, United States) was inserted into the lumen of the guide cannula to inject KA (0.15 µg/µl) or 4-aminopyridine (4-AP; 0.1 µg/µl) into the ventricle over a 1-min period using a microinjection pump (1 µl/min, KD Scientific, Holliston, MA, United States). Control animals received an equal volume of normal saline instead of KA or 4-AP. EEG signals were recorded with a DAM 80 differential amplifier (0.1-3000 Hz bandpass; World Precision Instruments, United States) and the data were digitized (1000 Hz) and analyzed using LabChart Pro v7 software (AD Instruments, NSW, Australia). Time of seizure onset was defined as the time point showing paroxysmal depolarizing shift, defined as lasting more than 3 s and consisting of a rhythmic discharge between 4 and 10 Hz with amplitude of at least two times higher than the baseline EEG Kang, 2011, 2015). Total power was measured during the 2h recording session and normalized by the baseline value obtained from each animal. Spectrograms were automatically calculated using a Hanning sliding window with 50% overlap by LabChart Pro v7. Diazepam (Valium; Hoffman la Roche, Neuilly-sur-Seine, France; 10 mg/kg, i.p.) was administered 2 h after KA injection and repeated, as needed. After recording, animals were used for anatomical or biochemical experiments.

Co-immunoprecipitation and Western Blot
As described previously (Kim et al., 2016), animals were quickly decapitated, and their hippocampi were dissected out in the presence of cooled artificial cerebrospinal fluid (in mM: 124 NaCl, 5 KCl, 1.25 NaH 2 PO 4 , 26 NaHCO 3 , 10 dextrose, 1.5 MgCl 2 , and 2.5 CaCl 2 ). The hippocampal tissues were lysed in radioimmunoprecipitation assay buffer (RIPA: 50 mM Tris-HCl pH 8.0; 1% Nonidet P-40; 0.5% deoxycholate; 0.1% SDS, Thermo Fisher Scientific, United States) containing protease inhibitor cocktail (Roche Applied Sciences, United States), phosphatase inhibitor cocktail (PhosSTOP R , Roche Applied Science, United States) and 1 mM sodium orthovanadate. Protein concentrations were determined by BCA protein assay (Pierce, United States) and equal amounts of total proteins were precipitated with the appropriate primary antibodies and protein G sepharose at 4 • C overnight. Beads were collected by centrifugation, eluted in 2× SDS sample buffer, and boiled at 95 • C for 5 min. Next, western blotting was performed according to standard procedures. The rabbit anti-β-actin primary antibody was used as internal reference. The signals were scanned and quantified on ImageQuant LAS 4000 system (GE health, United States). The values of each sample were normalized with the corresponding amount of β-actin.

Immunohistochemistry, FJB Staining, Golgi Impregnation, and Analysis of Spine Morphology
Under urethane anesthesia (1.5 g/kg, i.p.), animals were perfused transcardially with 4% paraformaldehyde in 0.1 M phosphate buffer (PB, pH 7.4). Brains were post-fixed in the same fixative overnight and then cryoprotected and sectioned at 30 µm with a cryostat. Free-floating coronal sections were incubated in a mixture of primary antibodies in PBS containing 0.3% Triton X-100 and 2% normal chicken serum overnight at room temperature. Sections were also incubated in a mixture of FITCand Cy3-conjugated secondary antisera (Amersham, Piscataway, NJ, United States, 1:200) for 1 h at room temperature (Kim et al., 2014). Some tissues were used for a conventional Fluoro-Jade B (FJB) staining (Kim et al., 2014;Kim and Kang, 2015;Hyun et al., 2016). To establish the specificity of the immunostaining, a negative control test was carried out with preimmune serum instead of the primary antibody. No immunoreactivity was observed for the negative control in any structures. Golgi impregnation was performed using FD Rapid GolgiStain TM kit (FD NeuroTechnologies, Inc., MD, United States), according to the manufacturer's instructions. All images and dendritic spine morphology were analyzed using an AxioImage M2 microscope and AxioVision Rel. 4.8 software or a confocal laser scanning microscope (LSM 710, Carl Zeiss Inc., Oberkochen, Germany).

Statistical Analysis
After evaluating the values on normality using Shapiro-Wilk W-test, data were analyzed by Student's t-test or ANOVA followed by Newman-Keuls post hoc test. A p < 0.05 is considered to be statistically different. Quantitative data were expressed as mean ± standard error of the mean.

PLPP/CIN Decreases Seizure
Susceptibility, but Increases the Progression of Seizures and Neuronal Damage in Response to KA Consistent with our previous study (Kim et al., 2016), average spine width was ∼0.40 µm and head/neck ratio was >2 in wildtype (WT, PLPP/CIN +/+ ) mice. As compared to WT animals, the spines in PLPP/CIN Tg mice showed thin necks with very small heads (head/neck ratio: 1.13), while PLPP/CIN −/− mice showed gigantic spines with a head/neck ratio of ∼4 (p < 0.05 vs. WT animals, respectively; Figures 1A,B). Since dendritic spines are the major integral site excitatory input, changed spine structures affect the neuronal excitability and seizure susceptibility (Wong and Guo, 2013). Thus, we measured the seizure susceptibility of PLPP/CIN Tg and PLPP/CIN −/− mice in response to KA.
In WT animals, the first seizure occurred 11.3 min after KA injection, progressing with time to prolonged episodes of high-frequency and high-amplitude discharges by 120 min post-KA injection (Figures 1C-E). EEG analyses revealed no difference in the latency of seizure on-set, duration and intensity (total EEG power) of seizures between two different genetic background WT mice. PLPP/CIN Tg and PLPP/CIN −/− mice showed the first seizure activity within 18.7 and 5 min after KA injection, respectively (p < 0.05 vs. WT animals, respectively; Figures 1C-E). PLPP/CIN Tg mice showed increase in seizure intensity (severity) to 1.5∼3-fold of that observed in WT animal. In PLPP/CIN −/− mice, seizure activity reduced ∼60 min after KA injection (p < 0.05 vs. WT animals, respectively;

Figures 1C-E).
Having observed that PLPP/CIN deletion interrupted the progression of seizures in response to KA, we investigated SEinduced neuronal damage in the hippocampus 1 day post-KA injection. WT mice displayed typical CA1 and CA3 pyramidal cell loss for FJB staining (Figures 1F,G). As compared to WT animals, PLPP/CIN Tg mice showed profound neuronal injury in the CA1 and CA3 regions, but PLPP/CIN −/− mice demonstrated the overt attenuation of neuronal damage in these regions (p < 0.05 Frontiers in Molecular Neuroscience | www.frontiersin.org vs. WT animals, respectively; Figures 1F,G). These findings indicate that PLPP/CIN may increase seizure intensity, duration, and the latency of seizure on-set.

F-Actin Polymerization Affects Seizure Intensity, But Not Latency of Seizure On-Set
To elucidate the role of PLPP/CIN-mediated F-actin depolymerization in response to KA, we measured cofilin phosphorylation level 2 h after KA injection. In control group of in PLPP/CIN Tg and PLPP/CIN −/− mice, phospho (p)-Cofilin/Cofilin ratio were 0.8-and 1.22-fold of WT animals, respectively (n = 7, respectively, p < 0.05, respectively; As compared to WT animals, cofilin phosphorylation is lower in PLPP/CIN Tg mice, but is higher in PLPP/CIN −/− mice. Western blot shows the up-regulation of PLPP/CIN expression and the reduction in pCofilin level without altered cofilin expression in WT animals. The reduced cofilin phosphorylation in PLPP/CIN Tg mice is more efficient, while that in PLPP/CIN −/− mice is less, as compared to WT animals. (A) Western blot of PLPP/CIN, cofilin, and pCofilin. (B) Quantification of PLPP/CIN expression, cofilin expression, and cofilin phosphorylation level based on western blot data. Open circles indicate each individual value. Horizontal bars indicate mean value (mean ± SEM; * , # p < 0.05 vs. saline-treated and WT animals, respectively; n = 7, respectively). (C-E) The effect of F-actin dynamics on seizure activity in response to KA. Jasplakinolide (an F-actin stabilizer) decreases seizure intensity (total EEG power) in PLPP/CIN Tg mice, but latrunculin A (an F-actin depolymerizer) enhances it in PLPP/CIN −/− animals. However, neither jasplakinolide nor latrunculin A affects the latency of seizure on-set in both groups. (C) Representative EEG traces and frequency-power spectral temporal maps in response to KA. (D) Quantification of latency of seizure on-set. Open circles indicate each individual value. Horizontal bars indicate mean value (mean ± SEM; n = 7, respectively). (E) Quantification of total EEG power (seizure intensity) in response to KA (mean ± SEM; p < 0.05 vs. vehicle; n = 7, respectively).
These findings indicate that PLPP/CIN-mediated F-actin depolymerization may contribute to the seizure intensity (severity), but not seizure threshold.

PLPP/CIN Regulates CSEN Phosphorylation
Since seizure activity affects CSEN expression in the mouse hippocampus following KA injection (Matsu-ura et al., 2002;Hong et al., 2003), we explored the effect of KA-induced seizure on CSEN expression. In control animals, double immunofluorescent study revealed that PLPP/CIN was colocalized with CSEN in hippocampal neurons, and that PLPP/CIN overexpression or deletion did not affect CSEN expression. However, PLPP/CIN overexpression decreased pCSEN level, but PLPP/CIN deletion increased it (Figures 3A-C and Supplementary Figure 4).
As compared to WT animals, pCSEN levels in PLPP/CIN Tg and PLPP/CIN −/− mice were 0.72-and 1.44-fold of WT animal levels, respectively (n = 7, respectively, p < 0.05, respectively; To investigate whether other protein phosphatases affect pCSEN levels, we examined the expressions and phosphorylation levels of protein phosphatase 1A (PP1A), protein phosphatase 2A (PP2A), and protein phosphatase 2B (PP2B) after KA injection. However, KA injection did not influence on protein phosphatase expressions and phosphorylations (Supplementary Figures 1, 9). Together with increases in PLPP/CIN expression and its activity, our findings demonstrate that PLPP/CIN may dephosphorylate CSEN.
To clarify the direct phosphatase activity of PLPP/CIN on CSEN, we applied in vitro assay using recombinant proteins. Co-immunoprecipitation revealed that PLPP/CIN bound to CSEN, and effectively reduced CSEN phosphorylation to 62% of vehicle level (n = 7, p < 0.05 vs. vehicle; Figure 4). Taken together, our findings suggest that PLPP/CIN may directly regulate CSEN phosphorylation.

Seizure Activity Increases the Binding of PLPP/CIN to CSEN
Next, we investigated whether seizure activity affects the binding activity of PLPP/CIN to CSEN in vivo. Under physiological condition (control animals), PLPP/CIN-CSEN binding in PLPP/CIN Tg mice was 1.33-fold of WT animal level (n = 7, p < 0.05, respectively; Figures 4A,B and Supplementary Figure 5). Two hours after KA injection, the PLPP/CIN-CSEN binding in WT animals was increased to 1.5-fold of control level (n = 7, p < 0.05, Figures 4A,B and Supplementary Figure 5). In PLPP/CIN Tg mice, the PLPP/CIN-CSEN binding was 1.7-fold of control level in WT animal (n = 7, p < 0.05, Figures 4A,B and Supplementary Figure 5). These findings indicate that seizure activity may increase the binding of PLPP/CIN with CSEN, which may represent activity-dependent CSEN dephosphorylation by PLPP/CIN.
To identify the role of PLPP/CIN in the binding of CSEN to Kv4.2, we applied in vitro assay using crude extracts from the PLPP/CIN −/− brains. PLPP/CIN treatment significantly increased CSEN co-precipitation with Kv4.2 to 1.5-fold of vehicle level (n = 7, p < 0.05 vs. vehicle; Figures 5C,D and Supplementary Figure 6). Together with PLPP-mediated CSEN dephosphorylation (Figure 3), these findings suggest that the affinity of intact (dephosphorylated) CSEN to Kv4.2 may higher than that of pCSEN.
In vitro assay using crude extracts of the PLPP/CIN −/− brains also revealed that the binding of CSEN to GluN1 was decreased to 0.71-fold of vehicle level in the presence of PLPP/CIN (n = 7, respectively, p < 0.05; Figures 6C,D and Supplementary Figure 7). However, the binding of CSEN to Kv4.2 was increased (n = 7, respectively, p < 0.05; Figures 5C,D, 6E). Because PLPP/CIN overexpression increases NMDAR functionality without changed NMDAR distribution and the heterotrimerization (Kim et al., 2016), our findings suggest that PLPP/CIN may potentiate NMDAR functionality by inhibiting the CSEN-GluN1 binding.

CSEN Phosphorylation Is Regulated by the Competitive Binding with PLPP/CIN and CK1
Next, we investigated the relationship between PLPP/CIN and CK1, since CK1 phosphorylates CSEN (Choi et al., 2003). In vitro binding assay using crude extracts from the PLPP/CIN −/− brains revealed that PLPP/CIN inhibited the binding of CSEN with CK1 to 0.68-fold of vehicle level (n = 7, respectively, p < 0.05; Figures 6E-G and Supplementary Figure 7). In vitro binding assay using recombinant proteins also demonstrated that PLPP/CIN reduced CK1-mediated CSEN phosphorylation to 0.56-fold of vehicle level (n = 7, respectively, p < 0.05; Figures 6H,I). Taken together, these findings indicate that PLPP/CIN may be one of the counterparts of CK1 for CSEN phosphorylation.

CSEN Phosphorylation Directly Regulates Seizure Activity in Response to KA Independent of Cofilin Phosphorylation
The remaining question is whether CSEN phosphorylation affects seizure activity in response to KA by regulating GluN1-CSEN The comparison of responsiveness to 4-AP. As compared to WT animals, PLPP/CIN Tg mice show the higher responsiveness to 4-AP, but PLPP/CIN −/− animals reveal lower responsiveness, as compared to WT animals. (E) Representative EEG traces and frequency-power spectral temporal maps in response to 4-AP.

DISCUSSION
The present study investigated the involvement of PLPP/CIN in the seizure activity in response to KA in mice. This study demonstrated four major findings: (1) PLPP/CINmediated F-actin depolymerization correlated with the seizure intensity (total EEG power) in response to KA.
(2) PLPP/CIN dephosphorylated CSEN, reflecting that CSEN is the novel    Turner and Happold, 1961), and later identified as a serine specific phosphatase of the non-thiol-based haloacid dehalogenase superfamily of hydrolases and a modulator for cofilin activity (Gohla et al., 2005;Kim et al., 2008Kim et al., , 2016. In our previous study , Seizure activity elevates PLPP/CIN expression in the rat hippocampus accompanied by increase in neuronal excitability. In the present study, the genetic PLPP/CIN manipulation changed seizure activity in response to KA. PLP plays a role as a cofactor in γ-aminobutyric acid (GABA) synthesis by glutamate decarboxylase (Miller et al., 1978). Thus, genetic defect of PLP metabolism results in vitamin B 6 -dependent epilepsy, which is an autosomal recessive epileptic encephalopathy responsive to pharmacological dosages of pyridoxine (a PLP precursor) and resistant to antiepileptic drugs (Stockler et al., 2011). With respect to these reports, altered seizure activity in PLPP/CIN Tg and PLPP/CIN −/− mice would be affected by PLPP/CIN-mediated alterations in GABAergic inhibitory neurotransmission via changed PLP concentration. However, this hypothesis can be excluded because the PLP concentration in the brain is tightly regulated by two rate-limiting synthetic enzymes, pyridoxal kinase and pyridoxine-5 -phosphate oxidase (Choi et al., 1983;Ubbink et al., 1990). Indeed, pyridoxine could not inhibit seizure activity in rats . Furthermore, both PLPP/CIN Tg and PLPP/CIN −/− mice show the similar GABAergic transmission under physiological condition (Kim et al., 2016).
In the present study, we found that F-actin polymerization by jasplakinolide significantly attenuated seizure intensity in PLPP/CIN Tg mice, but F-actin depolymerization by latrunculin A increased it in PLPP/CIN −/− animals. These findings indicate that PLPP/CIN-mediated F-actin depolymerization itself affects KA-induced seizure intensity. However, neither jasplakinolide nor latrunculin A affected the latency of seizure on-set in both groups. Thus, it is likely that the unknown PLPP/CIN-mediated mechanism may underlie changed seizure susceptibility and its duration. In the present study, PLPP/CIN bound to CSEN and dephosphorylated it in vivo and in vitro. In addition, these interactions between PLPP/CIN and CSEN were in competition with CK1 that phosphorylated CSEN. These findings are the first evidence providing the role of PLPP/CIN as a CSEN dephosphorylating enzyme. Furthermore, PLPP/CIN-mediated CSEN dephosphorylation inhibited the GluN1-CSEN binding, but enhanced the Kv4.2-CSEN binding in vivo and in vitro. These findings also provide the novel biological functions of PLPP/CIN-mediated CSEN dephosphorylation.
CSEN modulates multiple intracellular events related to pain (Cheng et al., 2002), long-term potentiation , learning and memory (Lilliehook et al., 2003;Alexander et al., 2009). In particular, CSEN plays an important role in maintenance of Kv4.2 channel functionality (Hoffman et al., 1997;An et al., 2000). Kv4.2 channels are pore-forming subunits of the neuronal somatodendritic I A , which inhibits neuronal hyperexcitability by dampening dendritic excitability and back-propagated action potentials (Hoffman et al., 1997;Cai et al., 2004). CSEN binds to Kv4.2 channels and promotes Kv4.2 channel functionality . Indeed, CSEN deletion reduces I A density in hippocampal dentate granule cells in mice (Lilliehook et al., 2003). Furthermore, reductions in Kv4.2 channel expression and a mutation in the gene encoding Kv4.2 (KCND2) are reported in animal models of epilepsy (Tsaur et al., 1992;Monaghan et al., 2008) and patients with temporal lobe epilepsy (Singh et al., 2006). With respect to these previous reports, it is likely that PLPP/CIN-mediated changes in seizure susceptibility would be relevant to the role of CSEN as an auxiliary subunit for Kv4.2 channels, although we did not investigate the Kv4.2 functionality by a direct electrophysiological approach. Indeed, the present study demonstrates that PLPP/CIN Tg mice showed significantly increases in CSEN coprecipitation with Kv4.2 and responsiveness to 4-AP. In addition, PLPP/CIN −/− mice reversed to these phenomena. Therefore, our findings provide the possibility that PLPP/CIN may decrease seizure susceptibility via the enhancing CSEN-Kv4.2 interactions.
A local increase in dendritic excitability favors the backpropagated action potentials with a subsequent boost in the NMDAR-mediated Ca 2+ influx (Frick et al., 2004). Indeed, Kv4.2 regulates Ca 2+ signaling through spontaneous NMDAR activation to control synaptic NMDAR expression and plasticity (Jung et al., 2008), which in turn leads to activity-dependent Kv4.2 internalization (inactivation) . Thus, it is likely that Kv4.2 and NMDAR may reciprocally regulate each other under normal condition. In the present study, we found that CSEN co-precipitation with GluN1 was decreased in PLPP/CIN Tg mice under normal condition. Furthermore, PLPP/CIN Tg mice showed the prolonged seizure duration and the severe progression of seizure activity in response to KA, accompanied by the rapid dissociation of CSEN from GluN1.
In contrast, PLPP/CIN −/− mice exhibited the spontaneous abrogation of KA-induced seizure activity with less reduced CSEN co-precipitation with GluN1. Furthermore, inhibition of CSEN phosphorylation by CK1 inhibitors delayed the latency of seizure on-set in response to KA, accompanied by reduced CSEN-GluN1 binding. Since CSEN (Zhang et al., 2010) and PLPP/CIN deletion (Kim et al., 2016) inhibit NMDAR functionality, our findings suggest that PLPP/CIN may increase KA-induced seizure duration and neuronal death due to the NMDAR activation by the reduction of the CSEN-GluN1 binding. In light of the functional coupling of NMDAR and Kv4.2 (Frick et al., 2004;Jung et al., 2008), the present study clearly demonstrates that PLPP/CIN-mediated CSEN regulation may maintain appropriate neuronal activity under physiological condition ( Figure 8A) and have a key role in seizure susceptibility, duration, and progression ( Figure  8B).
PLPP/CIN-mediated F-actin dynamics play an important role in the maturation of dendritic spines (Kim et al., 2016). Dendritic spines are critical for synaptic transmission, representing the primary location of excitatory synaptic inputs to neurons (Kasai et al., 2010). Thus, dendritic spine pathology (loss of dendritic spines and alterations in spine shape/size) seems to be an important pathophysiological mechanism in epilepsy (Ferhat et al., 2003;Wong, 2005). Indeed, a significant decrease in dendritic spine density is frequently seen in hippocampal pyramidal neurons and dentate granule cells in patients with temporal lobe epilepsy (Freiman et al., 2011) and epilepsy animal models (González-Burgos et al., 2004;Ampuero et al., 2007). However, the ictogenic role of dendritic spine morphology has been still controversial. For example, enhanced LIM kinase 1 (LIMK1, one of counterparts of PLPP/CIN for cofilin phosphorylation) protein translation reduces the seizure susceptibility accompanied by dendritic spine loss as well as paradoxical increased spine volume (Jimenez-Mateos et al., 2012. In the present study, PLPP/CIN Tg mice were less susceptible to KA-induced seizures with small and immature dendritic spines. In contrast, PLPP/CIN −/− mice were more sensitive with abnormal gigantic spines. Furthermore, jasplakinolide abolished seizure intensity (not seizure susceptibility or its duration) in PLPP/CIN Tg mice, and latrunculin A enhanced it in PLPP/CIN −/− animals. Since small spines are less efficient in the compartmentalization of Ca 2+ from the dendritic shaft than large spines (Majewska et al., 2000), the present data provide the possibility that the small spine size or F-actin depolymerization may enhance seizure intensity due to increased Ca 2+ flux from spines to the dendritic shaft. Further studies are needed to elucidate the correlation between ictogenesis and spine pathology.

CONCLUSION
We found that PLPP/CIN-mediated F-actin depolymerization increased seizure intensity in response to KA. In addition, PLPP/CIN dephosphorylated CSEN, which differently affected seizure susceptibility and its duration by regulating the binding of CSEN to Kv4.2 and NMDAR. Thus, we suggest that PLPP/CIN may play an important role in the regulation of neuronal excitability via modulating CSEN phosphorylation under physiological and pathological conditions (Figure 8).

AUTHOR CONTRIBUTIONS
T-CK designed and supervised the project. J-EK performed the experiments described in the manuscript with H-WH, S-JM, D-SL, AJ, and MK. J-EK and T-CK analyzed the data and wrote the manuscript.