The Antagonism of 5-HT6 Receptor Attenuates Current-Induced Spikes and Improves Long-Term Potentiation via the Regulation of M-Currents in a Pilocarpine-Induced Epilepsy Model

Recent studies have documented that reduced M-current promotes epileptogenesis and attenuates synaptic remodeling. Neurite growth is closely related to the level of 5-HT6 receptor (5-HT6R) in the central nervous system. However, little research is available regarding the relation between 5-HT6R and M-current and the role of 5-HT6R in M-current regulation. Herein, we found that the expression of 5-HT6R was notably increased and the expression of KNCQ2/3, the main components of the M channel, was decreased in a time-dependent manner in pilocarpine-induced chronic epileptic hippocampus. Interestingly, antagonism of 5-HT6R by SB271046 upregulated the expression of KCNQ2 but not KCNQ3. SB271046 greatly alleviated excitatory/inhibitory imbalance and improved the impaired LTP in the chronic epileptic hippocampus. Further mechanism exploration revealed that the above effects of SB271046 can be reversed by the M-channel inhibitor XE991, which also confirmed that SB271046 can indeed improve abnormal M current. These data indicate that the antagonism of 5-HT6R may decrease the excitability of hippocampal pyramidal neurons in chronic epileptic rats and improve the impaired long-term potentiation by upregulating the expression of KCNQ2 in the M-channel.


INTRODUCTION
Cognitive impairment is common in temporal lobe epilepsy (TLE) , and the typical characteristic of pathological changes in TLE is hippocampal sclerosis, including neuronal loss in the CA3/CA1/DG region, glial proliferation, and abnormal mossy fiber sprouting (Schmeiser et al., 2017;Prada Jardim et al., 2018;Tai et al., 2018). Long-term potentiation (LTP), a neuronal model of neuroplasticity, is the physiological basis of memory formation (Kullmann and Lamsa, 2007;Bannerman et al., 2014). There are several forms of LTP, including the typical form, in which the recording electrode is on CA1 and the stimulating electrode is on CA3. Neuronal loss and lesions in chronic spontaneous recurrent seizures (SRSs) occur in CA3/CA1/DG over the course of TLE (Sloviter, 1994;Bannerman et al., 2014). Therefore, the loss of neurons in TLE is spread and brings about hippocampus-dependent learning and memory impairment, which could be detected via LTP. However, few studies have reported how TLE disturbs LTP.
The serotonin (5-hydroxytryptamine) receptor 5-HT6 (5-HT6R) is exclusively expressed in both the developing and the mature nervous systems (Roberts et al., 2002;Dayer et al., 2015;Helboe et al., 2015). It is highly expressed in mnemonic regions such as the olfactory tubercle, striatum, nucleus accumbens, and hippocampus (De Deurwaerdere et al., 2020). Studies have proven that it plays important roles in neuronal migration, neuronal differentiation, and neurite growth (Roberts et al., 2002;Dayer et al., 2015;De Deurwaerdere et al., 2020). Considering its particular distribution, 5-HT6R has recently been targeted to alleviate cognitive impairment, especially hippocampus-dependent learning and memory impairment (Seo and Tsai, 2014;Dayer et al., 2015;Karila et al., 2015;Aparicio-Nava et al., 2019). Interestingly, 5-HT6R blockers have antiepileptic effects. As early as 2000, 5-HT6R antagonists were demonstrated to increase the threshold of seizures in the maximal electroshock seizure test (Routledge et al., 2000). In 2015, increased 5-HT6R expression was found both in human TLE and kainic-acid-induced epileptic mouse models, and it was shown that 5-HT6R antagonists could alleviate seizures in the model (Wang et al., 2015). Our previous studies have demonstrated that the inactivation of 5-HT6R with SB271046 not only attenuates spontaneous recurrent seizures but also improves learning and memory performance in rats with pilocarpine-induced epilepsy (Lin et al., 2017;Liu et al., 2019). Therefore, 5-HT6R is theoretically considered to be a target for precognitive effects in an epileptic rodent model. However, the mechanism remains to be further explored.
Accumulating evidence has shown that 5-HT6R has a high level of constitutive activity established for cultured rodent brains, cultured neurons, and cell lines such as HEK-293 (). The activity of 5-HT6R was preliminarily considered to depend on its coupling to Gs, which is involved in alternative signaling mechanisms, including the rapamycin (mTOR) pathway, RhoAdependent pathway, Fyn-ERK1/2 pathway, and p-c-Jun-Jab-1 pathway . Further studies proved that 5-HT6R recruited cyclin-dependent kinase 5 (Cdk5) and is activated by the Cdc42-dependent pathway (Pujol et al., 2020). Both of the above pathways are related to neuronal migration, neurite growth, and synaptogenesis. 5-HT6R is positively linked to adenylate cyclase (AC) to increase cAMP formation, which is related to inducing LTP (Otmakhova et al., 2000). Several studies have shown that 5-HT6R antagonists increase LTP in cAlzheimer's disease/Vascular dementia (Aparicio-Nava et al., 2019;Grychowska et al., 2019;Rychtyk et al., 2019;Shahidi et al., 2019). Notably, a few studies demonstrated that 5-HT6R agonists also showed precognitive effects in cognitive disorders (Kendall et al., 2011;Vanda et al., 2018;Rychtyk et al., 2019). The contradictory mechanism remains unclear. Published data indicate that the regulatory role of 5-HT6R depends on the distribution of neurotransmitter systems. If an agonist activates 5-HT6R located directly on cholinergic and/or glutamatergic neurons, it may lead to an increase in cholinergic and glutamatergic transmission, while 5-HT6R antagonists act via GABAergic interneurons. Previous studies have indicated little expression of 5-HT6R on cholinergic and glutamatergic neurons in healthy conditions (Woolley et al., 2004). However, how 5-HT6R is expressed and how its ligands act in an epileptic model remain unclear. The KCNQ2/3 channel is the main isoform of the voltage-gated potassium channel and is expressed in both pyramidal neurons and interneurons (Soh et al., 2014a), the latter being fundamental regulators of normal brain activity. The KCNQ2/3 channel underlies the neuronal M-current. A dysfunction of the KCNQ2/3 channel in forming the Mcurrent has been shown to be responsible for multiple pediatric epileptic disorders, such as benign neonatal familial convulsion (BNFC) and early-onset neonatal epileptic encephalopathy (Surti and Jan, 2005). Several studies have demonstrated that the suppression of the M-current has antiepileptic effects in rodent models (Kay et al., 2015;Greene et al., 2018). In pyramidal neurons, the KCNQ2/3 channel primarily controls spike frequency adaptation, preventing quiescence-period neurons from entering a brief chain of activities (Rodier et al., 2018;Carver and Shapiro, 2019). Other studies have found that the deletion of either KCNQ2/3 or KCNQ2 channels from PV+ interneurons can induce elevated homeostatic potentiation of fast excitatory transmission in pyramidal neurons (Soh et al., 2018) and that decreased expression of KCNQ2 and KCNQ3 in a stress model can impair spatial learning and memory and LTP in the hippocampus (Petrovic et al., 2012;Li et al., 2014). KCNQ2/3 channels co-localize to the axon initial segment (AIS) in the hippocampus, indicating their role in action potential (AP) generation (Dai et al., 2013). Further functional studies demonstrated that the suppression of the M-current in CA1 neurons reduced the intrinsic subthreshold of theta resonance and augmented spikes after depolarization (Ford et al., 2004), inducing the burst firing mode of the hippocampus. Another behavior study showed that the M-current inhibitor XE991 could revert cognitive impairment in a rodent model (Fontan-Lozano et al., 2011). The above data indicate that the KNCQ2/3 channel plays an important role not only in network excitability but also in synaptic remodeling. Since the M-current has been proven to be a "clamper" of neuronal potentiation, we hypothesized that the M-current might be a key molecule mediating 5-HT6R ligand-induced changes in LTP in a chronic epileptic rodent model.
In the present study, acute and chronic epileptic models were established with Sprague-Dawley (SD) rats by pilocarpine injection. The pilocarpine-induced chronic epileptic rat model mimics the progress of human temporal lobe epilepsy (Kandratavicius et al., 2014). Patch-clamp was employed to detect the excitability of hippocampal pyramidal neurons, the Mcurrent, and long-term potentiation (LTP) in the hippocampus. The 5-HT6 antagonist SB271046, agonist WAY181187, Mcurrent-opening NEM, and M-current inhibitor XE991 were used to explore the mechanism of 5-HT6R ligand-induced changes in LTP via the M-current.

Animals
Adult male Sprague-Dawley rats (weighed 200-250g) were provided by the Experimental Animal Center of Fujian Medical University and housed in an SPF animal laboratory (at 22-24°C, on a 12-h light/dark cycle) with free access to food and water. The R5 rat herd cage was used (size: 545*395*200 mm), and rats were housed four or five per cage. The animals stayed in the laboratory for half an hour before the experiment to adapt to the laboratory environment. The present study was approved by the Fujian Medical University Animal Experimental Ethical Committee (No. FJMUIACUC 2019-0058). The relevant experimental protocols and procedures followed the Guidelines for the Care and Use of Laboratory Animals in Fujian Medical University.

ELECTROPHYSIOLOGY
Rats received Demerol and were anesthetized with isoflurane for the preparation of hippocampal slices. The brains were excised rapidly, placed into ACSF and coronally cut into slices (400 mm) using vibratome (Leica VT1000S) in ice-cold oxygenated ACSF (95% O 2 /5% CO 2 ), which contained NaCl (126 mM), NaHCO 3 (18 mM), KCl (2.5 mM), NaH 2 PO 4 (1.2 mM), CaCl 2 (1.2 mM), MgCl 2 (2.4 mM), and Glucose (11 mM) (pH 7.3, 325 mOsm/l). The rat brain atlas was used, and the cuts were made in Z coordinates. Slices were incubated in the ACSF at 32°C for 30 min and then at room temperature for over 30 min. Afterward, they were removed to the recording chambers for persistent perfusion (2 ml/min) with oxygenated ACSF. For field excitatory postsynaptic potential (fEPSP) recording, the external solution was adapted from Tengfei Ma et al. (2018), and the microelectrode was filled with ACSF (3-6 M resistance). fEPSPs were recorded using a stimulation protocol of 0.1 Hz with an increasing stimulation intensity from 0 to 100 mA at an increment of 10 mA and were used for input-output analysis. Paired-pulse stimulations were delivered at inter-stimulus intervals (ISI) of 10, 30, 50, 100, 200, and 500 ms. Paired-pulse ratios (PPRs) were calculated by dividing the magnitude of the response to the 2nd stimulation by the magnitude of the response to the 1st stimulation of the paired pulses. For LTP, a 20-min fEPSP baseline (measured as 30% of the maximum response) was recorded with the protocol of a stimulus pulse width of 0.1 ms at 10 Hz before the high-frequency stimulation (HFS, 100 Hz, 1s) was applied two times. Following HFS, responses to the test stimulation were recorded for another 60 min.

Statistical Analysis
All data are represented as mean ± SEM. Statistical analyses were performed with Graph Pad Prism6, SPSS20.0, Clampfit software. Student's t-tests were performed for two-group comparisons. One-way ANOVA and two-way ANOVA with Bonferroni tests were conducted for experiments with more than two groups. A value of p < 0.05 was considered statistically significant.

Identification of Significant Differentially Expressed Genes (DEGs) in Pilocarpine-Induced Chronic Epileptic Hippocampus After SB271046 Treatment
As depicted in Figure 1A, a volcano plot showed 80 upregulated (shown as red dots) and 36 downregulated (shown as green dots) DEGs between the Saline and Pilo group. Similarly, there were 45 upregulated and 190 downregulated DEGs between the Pilo and the Pilo+SB group ( Figure 1B). Using |log2Fold change| > 1 and p-value < 0.05 as the threshold, as shown in Figure 1C, there were 116 DEGs between the Saline and Pilo group and 235 DEGs between the Pilo and the Pilo+SB group. The heatmap demonstrated 32 over-lapping genes, as shown in Figure 1D. Among them, KCNQ2 was one of the overlapping genes that were downregulated obviously in the Pilo group (shown in light blue) but upregulated distinctively after SB treatment (shown in light orange).

Increased Expression of 5-HT6R and Decreased Expression of KCNQ2/3 in Rats With Pilocarpine-Induced Chronic Epilepsy
To explore the potential role of 5-HT6R and KCNQ2/3 in epilepsy, an epileptic rat model was induced by pilocarpine and the proteins in the hippocampus of the rats with spontaneous recurrent seizures (SRSs) were analyzed via Western blotting. The level of 5-HT6R significantly increased in these chronic epileptic rats when compared with that of the controls, t (10) = -6.424, p=0.0001 ( Figure 2A). Next, to observe the variation in 5-HT6R and KCNQ2/3 expression after pilocarpine-induced seizures, acute SE rats were included in the experiment. The level of 5-HT6R increased in a time-dependent manner and peaked in the chronic period (F (6,63) =30.753, p < 0.001) ( Figure 2B). However, the level of both KCNQ2 and KCNQ3 in the hippocampus decreased gradually in a time-dependent manner and dropped to a minimum in the chronic period (F (6,35) =7.426, p < 0.001 and F (6,35) =18.793, p < 0.001, respectively) ( Figures 2C, D).

SB271046 Attenuates Hippocampal Neuronal Excitability and Enhances LTP
To evaluate the neuronal excitability, sEPSCs and sIPSCs of the pyramidal neurons from the CA1 region were recorded via voltage-clamp. The number of current-induced spikes significantly increased in the Pilo group when compared with that of the Saline group (p= 0.0002) and remarkably decreased in the Pilo+SB271046 group when compared with that of the Pilo group (p= 0.0003) ( Figures 3A, B). WAY-181187 possesses high-affinity binding at the human 5-HT6 receptor and is profiled as a full receptor agonist. Both the frequency and amplitude of sEPSCs and sIPSCs increased notably in the Pilo group and significantly decreased in the Pilo+SB271046 group when compared with those of the Pilo group, while WAY181187 intervention produced no significant change ( Figures 3C-H).
These results indicate that pilocarpine-induced chronic epilepsy displays an increasing action potential (AP) and that 5-HT6R antagonist SB271046 can modulate the imbalance between excitatory and inhibitory neurotransmitters.
Next, LTP from CA3-CA1 was recorded by high-frequency stimulation (HFS). The potentiation significantly decreased in the Pilo group when compared with that of the Saline group (p < 0.001). After SB271046 intervention, the potentiation greatly increased in the SB271046 group when compared with that of the Pilo group (p < 0.001) and WAY181187 intervention produced no change ( Figures 4A, B). The amplitude of fEPSP gradually paralleled the increased stimulus intensity in the four groups, but the variation in the Pilo group decreased when compared with that of the Saline group (p < 0.001). The inputoutput response analysis suggests that chronic epileptic rats show decreased excitatory transmission ( Figure 4C). However, the paired-pulse facilitation (PPF) ratios remained unchanged among the four groups ( Figure 4D). These results suggest that 5-HT6R antagonist SB271046 improves synaptic transmission.
inhibitor XE991 was administered to the chronic epileptic hippocampal slices (Figures 6E-H). Furthermore, in the Pilo group, the LTP of the hippocampus from CA3-CA1 was significantly attenuated (p < 0.001) but was rescued by SB271046 or NEM intervention (p= 0.0003, p= 0.0000, respectively). The treatment with M-channel inhibitor XE991 produced no change. The effect of SB271046 was blocked when XE991 was added to chronic epileptic hippocampal slices (p= 0.0222) (Figures 6A-D).

Inactivated 5-HTR6 Upregulates the Expression of KNCQ2 But Not KNCQ3
As M-channel is composed of KCNQ2/3, the expression of KCNQ2/3 in the hippocampus was further analyzed by Western blotting. Notably, the level of both KCNQ2 and KCNQ3 in the Pilo group significantly decreased when compared with that in the Saline group. After the administration of SB271046, the expression of KCNQ2, but not KCNQ3, was markedly upregulated when compared with that of the Pilo group ( Figures 5C-E).

DISCUSSION
Our current study demonstrated that SRSs increased the protein expression of 5-HT6R and decreased both the RNA of KCNQ2 and protein of KCNQ2 in hippocampus. We found that, in the pilocarpine-induced epileptic rat model, SB271046 rescued the aberrant M-current, reduced the excitability of the hippocampal pyramidal neurons, and improved the impaired LTP, which were reversed by XE991, which selectively inhibits M-current. These findings indicate that the antagonism of 5-HT6R may improve LTP and alleviate the excitatory/inhibitory imbalance in the hippocampal neurons by regulating the M-channel.
In line with previous studies, the protein expression of 5-HT6R was upregulated in the epileptic brain, especially in the chronic model (Wang et al., 2015;Liu et al., 2019). Pilocarpineinduced chronic epileptic rat model mimics the progress of human temporal lobe epilepsy (Kandratavicius et al., 2014). It usually involves an acute period (< 24 hours, status epilepticus), a silent period (> 24 hours, < 7 days, no seizures, no food or water, no movement), and a chronic period (> 7 days, SRSs). In our present study, the expression of 5-HT6R in hippocampus increased gradually during the silent period and peaked in the chronic period. These data suggest that spontaneous recurrent seizures may induce overexpression of 5-HT6R in hippocampus, which plays an important role in epileptogenesis. Combined with the results of RNA-sequencing and Western blotting, we speculate that SRSs may reduce the expression of 5-HT6R protein by affecting the translation level of 5-HT6R. In the literature, few studies have focused on whether 5-HT6R is involved in epilepsy. Previous studies reported that 5-HT6R ligands regulate the imbalance between exciting and inhibiting neurotransmitters (Karila et al., 2015). A recent study found that a 5-HT6R antagonist may modulate seizure activity (Wang et al., 2015). However, the neuronal activity has not been examined.
In our present study, spontaneous EPSC and IPSC were both detected in chronic epileptic hippocampal slices, the former mainly produced by the excitatory neurotransmitter receptors, such as NMDA-R/AMPA-R, and the latter by the inhibiting neurotransmitter receptor, such as GABA-R. Other studies report that the frequency of sEPSCs increases with or without decreased frequency of sIPSCs (Soh et al., 2014b). In this work, the frequency and amplitude of both sEPSCs and sIPSCs were markedly increased in the pyramid cells of the pilocarpine-induced chronic epileptic rat model. The cells also exhibited an increased number of currentinduced spikes, further indicative of a shift towards hyperexcitability. Moreover, the antagonism of 5-HT6R by SB271046 decreased the frequency and amplitude of sEPSCs, sIPSCs, and action potentiation. These results indicate that 5-HT6R is overexpressed in the pilocarpine-induced chronic epileptic hippocampus and that the antagonism of 5-HT6R by SB271046 can modulate the imbalance between excitatory and inhibitory neurotransmitters. We further explored the mechanism of the observed effect of 5-HT6R on epilepsy. Interestingly, the gene screening spectrum showed that antagonism of 5-HT6R upregulated the expression of KCNQ2 in pilocarpine-induced chronic epileptic hippocampus. It has been reported that the loss of KCNQ2 or KCNQ3 function induces neuronal hyperexcitability and that genetic mutations in KCNQ2/3 are responsible for multiple pediatric diseases, especially during infancy (Surti and Jan, 2005). Existing studies have probed into the role of KCNQ2/3 in the developing brain. Our present study investigated the role of KCNQ2/3 in the developed or mature brain. In our experiment, M-current was aberrant in the chronic epileptic slices, indicating a malfunction but not a loss of M-current. It is known that malfunctioning of the potassium channels constituted by KV7.2 and KV7.3 may lead to abnormality of the action potential (Fisher et al., 2005). The M-channel is composed of KV7.2 and KV7.3 subunits and plays an important role in the pathogenesis of epileptic seizure. The structures of voltage-gated potassium channels encoded by KCNQ family genes are quite similar and are tetramers composed of four identical or different a subunits. The pathogenic genes of KCNQ2 and KCNQ3 are located on chromosomes 20q13.3 and 8q24, respectively.

2017
). In the current study, we found that SB271046 (a 5-HT6R antagonism) and not WAY181187 (a 5-HT6R agonist) improved the impaired hippocampal LTP in the chronic epileptic rats. Many reports have shown that the dysfunction of KCNQ/Kv7 channels is related to cognitive impairment. The reduction of KCNQ/Kv7 channels can mediate age-dependent memory decline (Wang et al., 2011;Cavaliere et al., 2013b). The Mchannel plays an important role in regulating neuronal excitability, spike generation, hippocampal theta oscillation, and neurotransmitter release (Hu et al., 2002;Yue, 2004;Peters et al., 2005;Vervaeke et al., 2006), and suppressed M-currents in mice can impair the hippocampus-dependent spatial memory (Peters et al., 2005). Thus, it is evident that KCNQ/Kv7 channels are involved in synaptic plasticity, learning, and memory. Our current data reinforce the notion that the activation of functional M-channels by NEM may facilitate the enhancement of the hippocampal synaptic plasticity in pilocarpine-induced epileptic rats. In addition, 5-HT has been shown to inhibit Mcurrents in mammalian neurons (Roepke et al., 2012). Based on our results, 5-HT6R antagonist SB271046 can rescue the impaired LTP in chronic epileptic hippocampus, and this effect can be reversed by XE991, which selectively inhibits the Mcurrent. Therefore, we speculate that the improved LTP by the antagonism of 5-HT6R may involve the regulation of M-channel. In summary, the antagonism of 5-HT6R can decrease the excitability of the hippocampal pyramidal neurons in chronic epileptic rats by modulating the imbalance between excitatory and inhibitory neurotransmitters. The antagonism of 5-HT6R improves the impaired LTP in chronic epileptic rats by regulating the KCNQ2/3-mediated M-currents, especially KCNQ2-mediated M-currents, and may serve as a promising candidate in the clinical treatment of epilepsy.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

ETHICS STATEMENT
The present study was approved by the Fujian Medical University Animal Experimental Ethical Committee (No. FJMUIACUC 2019-0058). The relevant experimental protocols and procedures followed the Guidelines for the Care and Use of Laboratory Animals in Fujian Medical University.

AUTHOR CONTRIBUTIONS
HH, WL, and CZ conceived and designed the study. CZ, RL, CL, MH, FL, GZ, YZ, and JM performed the experiments and analyzed the data. CZ wrote the paper. WL and HH revised the paper. All authors reviewed the results and approved the final version of the manuscript.