Pentylenetetrazole kindling induces dynamic changes in GAD65 expression in hippocampal somatostatin interneurons

Introduction: One of the mechanisms of epileptgenesis is impairment of inhibitory neural circuits. Several studies have compared neural changes among subtypes of gamma-aminobutyric acid-related (GABAergic) neurons after acquired epileptic seizure. However, it is unclear that GABAergic neural modifications that occur during acquisition process of epileptic seizure. Methods: Male rats were injected with pentylenetetrazole (PTZ kindling: n = 30) or saline (control: n = 15) every other day to observe the development of epileptic seizure stages. Two time points were identified: the point at which seizures were most difficult to induce, and the point at which seizures were most easy to induce. The expression of GABAergic neuron-related proteins in the hippocampus was immunohistochemically compared among GABAergic subtypes at each of these time points. Results: Bimodal changes in seizure stages were observed in response to PTZ kindling. The increase of seizure stage was transiently suppressed after 8 or 10 injections, and then progressed again by the 16th injection. Based on these results, we defined 10 injections as a short-term injection period during which seizures are less likely to occur, and 20 injections as a long-term injection period during which continuous seizures are likely to occur. The immunohistochemical analysis showed that hippocampal glutamic acid decarboxylase 65 (GAD65) expression was increased after short-term kindling but unchanged after long-term kindling. Increased GAD65 expression was limited to somatostatin-positive (SOM + ) cells among several GABAergic subtypes. By contrast, GAD, GABA, GABA A R α 1, GABA B R1, and VGAT cells showed no change following short-or long-term PTZ kindling. Conclusion: PTZ kindling induces bimodal changes in the epileptic seizure stage. Seizure stage is transiently suppressed after short-term PTZ injection with GAD65 upregulation in SOM + cells. The seizure stage is progressed again after long-term PTZ injection with GAD65 reduction to baseline level.


Introduction
Gamma-aminobutyric acid (GABA) is a widely distributed inhibitory neurotransmitter in the central nervous system that is required for the suppression of neuronal excitability.Changes in the balance of excitatory/inhibitory neuronal input are considered the underlying mechanism of epileptogenesis.GABA-related (GABAergic) neurons are classified into several subtypes, including parvalbumin-positive (PV + ), somatostatin-positive (SOM + ), calretinin-positive (CR + ), and cholecystokinin-positive (CCK + ) neurons, based on their morphological and electrophysiological characteristics (Kubota and Kawaguchi, 1997;Kawaguchi and Kondo, 2002;Gonzalez-Burgos and Lewis, 2008).Studies of the chemogenetic recruitment of specific GABAergic subtypes have indicated that these GABAergic subtypes play distinct roles in the suppression of neuronal excitation (Cǎlin et al., 2018).GABA subtypes exhibit variation in their vulnerability to neuroexcitotoxicity; for example, PV + and CCK + perisomatic inhibitory neurons are more vulnerable to epileptic seizures or traumatic brain injury than SOM + and CR + dendritic inhibitory neurons in the hippocampus (Huusko et al., 2015).
Pentylenetetrazole (PTZ) kindling is a chronic epilepsy model that is used to observe the onset of progressive seizures (Cain, 1982;Rubio Abbreviations: ANOVA, analysis of variance; CCK, cholecystokinin; CR, calretinin; DAPI, 4,′6-diamidino-2-phenylindole dihydrochloride; DG, dentate gyrus; GABA, gamma-aminobutyric acid; GAD, glutamic acid decarboxylase; H, hilus; NPY, neuropeptide Y; PBS, phosphate-buffered saline; PTZ, pentylenetetrazole; PV, parvalbumin; SEM, standard error of the mean; SL, stratum lucidum; SO, stratum oriens; SOM, somatostatin; SP, stratum pyramidale; SR, stratum radiatum. et al., 2023).Several previous studies have reported changes in GABAergic neurons using a PTZ kindling model.For example, in one study, glutamic acid decarboxylase (GAD) 65, which is responsible for activity-dependent phasic GABA synthesis (Tian et al., 1999;Kanaani et al., 2010) was decreased in the hippocampus of rats (Hao et al., 2016).However, both increases (Szyndler et al., 2018) and decreases (Hao et al., 2016) in GABA A receptor expression have been reported in the hippocampus of rats.PTZ kindling in animals has led to changes in GABAergic neuron-related proteins in the hippocampus, with less neurodegeneration and cell death than kainic acid or pilocarpine models (Pisa et al., 1980;Turski et al., 1984;Hu et al., 2013).Therefore, the PTZ kindling model is suitable for analyzing changes in epileptic susceptibility without serious neuronal damage, in a manner not directly associated with the acquisition of epileptogenicity (Samokhina and Samokhin, 2018).However, most previous studies based on PTZ kindling have focused on neuronal changes in fully kindled animals, or after epileptogenesis, whereas changes to GABAergic neurons during the kindling process or during changes in epileptic sensitivity remain unclear.Investigation of the changes in GABA synthesis-or receptorrelated molecules among GABA subtypes involved in changes in epileptic sensitivity will help to elucidate the mechanisms of epileptogenesis.
In this study, we classified the PTZ kindling process into short-and long-term treatment durations, in which PTZ-induced seizures were less and more effective, respectively.Then we examined the effects of shortand long-term PTZ injection duration on GABAergic neuron-related proteins in the hippocampus to explore the responses of abnormal GABAergic subtypes during changes in epileptic sensitivity.

Animals
Experiments were performed using male, 5-month-old Long-Evans rats (Institute for Animal Reproduction, Kasumigaura, Ibaraki, Japan, http://www.iar.or.jp).The rats were housed in groups in a temperatureand humidity-controlled room under a 12 h/12 h light/dark cycle with food and water provided ad libitum.Efforts were made to minimize the number of animals used, as well as their pain and discomfort.For statistical analysis of the behavioral experiments, a minimum of five animals per group was required, and three experiments were conducted in three groups to confirm reproducibility (5 rats × 3 groups × 3 times = 45 rats).All animal experiments were approved by the Tohoku University Committee for Animal Research, Seiryo Campus (Sendai, Japan).

PTZ injection
PTZ kindling is an experimental model for studying epilepsy-related neural changes.On alternating days, intraperitoneal injections of 40 mg/kg PTZ (Tocris Bioscience, Bristol, UK) dissolved in saline were performed under inhalation anesthesia with isoflurane (1-3 %; Pfizer, Tokyo, Japan).The control group (CTL) was injected with an equal volume of saline solution under inhalation anesthesia with isoflurane.These rats were awake and moving within at least 1 min after PTZ or saline injection.Thirty rats were used in the PTZ group, and 15 rats were used in the CTL.PTZ or saline injections were performed between 1:00 p.m. and 4:00 p.m.

Seizure stage
A maximum of 20 PTZ injections were administered to any single rat.Seizures were monitored after each injection and classified into six stages according to the Racine staging system (Racine, 1972), as follows: Stage 1, facial movements; Stage 2, rhythmic head movements and nodding; Stage 3, bilateral forelimb clonus; Stage 4, rearing; Stage 5, rearing and falling; Stage 6, death.These rats' behaviors were monitored until the seizure stage decreased to Stage 1 or less.Seizure stages were analyzed to assess the scores by three independent blinded observers.The staging-specificity in kindled rats is proved by electroencephalographic analysis (Lüttjohann et al., 2009;Jalilifar and Yadollahpour, 2018;Van Erum et al., 2019).Based on the seizure intensity after 20 PTZ injections, we classified the kindling process as either short-term kindling, which seizures are less likely to occur, or long-term kindling, which continuous seizures are likely to occur.Short-term kindling was defined as the first day of five continuous seizures of less than Stage 3, and long-term kindling was defined as the last day of five continuous seizures of more than Stage 3. Similarly, daily (48 h interval) changes in seizure stage from the previous injection were compared individually for each rat.The latency to seizure onset was also measured.

Tissue preparation
The rats were perfused at 24 h after the last PTZ injection (of 10 or injections).Under deep nembutal anesthesia (Somnopentyl, 250 mg/kg body weight, intraperitoneally injected, Kyoritsu Seiyaku Co., Ltd., Tokyo, Japan), the rats were transcardially perfused with phosphatebuffered saline (PBS) containing 2 IU/mL heparin, and then with 4 % paraformaldehyde in 0.1 M phosphate buffer (pH 7.4).For GABA staining, the rats were perfused with 4 % paraformaldehyde and 0.1 % glutaraldehyde in 0.1 M phosphate buffer (pH 7.4).The brains were removed, post-fixed in a perfused solution for 24 h at 4 • C, and then immersed in 30 % sucrose solution in PBS for at least 24 h.Coronal sections (30 μm thick) were cut using a cryostat (Leica CM1950, Leica Microsystems, Wetzlar, Germany).

Expression measurements in neuropil regions
To measure GAD65, GAD67, GABA, GABA A R α1, GABA B R1, VGAT, and NPY expression in the neuropil regions, we examined six hippocampal sections per animal.Avoiding somatic staining, we randomly set 12 squares (10 μm × 10 μm) in each hippocampal region (SO, SP, SR, SL, and H), and measured the average intensity within the squares in each layer (range: 0-65,535).Average intensity values were used as data points for GAD65, GAD67, GABA, GABA A R α1, GABA B R1, VGAT, and NPY expression after subtraction of the background signals.Among these protein expressions, NPY expression were used to confirm the presence of epileptic seizures.NPY expression accumulates in mossy fibers with post-status epilepticus (Schwarzer et al., 1995), and the intensity is considered to reflect the intensity of epileptic seizure (Nagaki et al., 2000).

Expression measurements in the soma
To measure GAD65 expression in each GABAergic subtype (PV, SOM, CR, and CCK), we examined eight hippocampal sections per animal and measured the intensity of GAD65 staining in the soma.To distinguish GAD65 staining in the soma from that in the synaptic button, we used serial z-stack sections, as described by Kajita and Mushiake (2021).We randomly set four squares within each soma, avoiding the nucleus and buttons, and measured the average intensity within each square.The average intensity over the four squares was used as a data point for GAD65 intensity after subtraction of the background signals.To avoid measuring out-of-focus cells, we selected only those cells showing clear DAPI staining.

Cell density measurements
To measure the cell density of GABAergic subtypes (PV, SOM, CR, and CCK), we examined eight hippocampal sections per animal, counting the cells of each subtype in each hippocampal layer (SO, SP, and SR), and dividing the result by the area of the layer to determine the cell density (cells/mm 2 ).

Statistical analyses
Data were collated in Excel (Microsoft, Redmond, WA, USA) for further analyses.Numbers were truncated after the third or fourth decimal place.We used IBM SPSS Statistics for Windows v29.0 (IBM Corp., Armonk, NY, USA) for the following statistical analyses: Kolmogorov-Smirnov test for data normality, Levene's test for equal variances, Student's t-test, the single-sample t-test, one-way analysis of variance (ANOVA; for datasets with a normal distribution), Kruskal-Wallis test (for datasets lacking a normal distribution), Bonferroni's post hoc test (for datasets with equal variance), and Dunnett's T3 post hoc test (for datasets lacking equal variance).Data are presented as means ± standard errors of the mean (SEMs), and significance was determined at thresholds of P < 0.05 and P < 0.01.

Bimodal changes in seizure stage following PTZ kindling
The rats received repeated PTZ injections and their seizure stages were scored according to the Racine staging system.Following PTZ injection, the seizure stage first increased to above Stage 3, gradually decreased to approximately Stage 2, and then increased and remained above Stage 3. Kindling stage did not reach the Stage 4 or 5, tonic-chronic seizure after 20 injections, but based on our criteria (described in Materials and Methods), we prepared two groups of PTZkindled rats, as shown in Fig. 1 A Seizure onset was observed from 5 min to 56 min after PTZ injection, and there was no regularity in the latency to seizure onset depending on the number of injections.Five rats died of PTZ-induced epileptic seizures within 10 injections (three in the short-term and two in the long-term kindling groups); all were excluded from data analyses.Y. Kajita et al.

GAD65 expression increased in the hippocampal CA1 region after short-term PTZ kindling
The brains were fixed 24 h after the last PTZ injection (of 10 or 20 injections).Typical confocal microscopic images captured in this study are shown in Fig. 2. In both the short-and long-term injection groups, enhanced NPY expression was detected in the SL of the CA3 region and the H of the DG, presumably labeling sprouted mossy fibers.Localized NPY change indicated that PTZ-induced seizures originated in the brain (epileptic seizure) in both the short-and long-term kindling groups.NPY expression in the hippocampus is considered to reflect the intensity of epileptic seizure; therefore, we compared NPY immunostaining intensity between the short-and long-term groups.In the SL and H neuropil regions, the long-term group showed significantly higher NPY intensity than the short-term group [for SL, short-term: 3474.88 ± 461.38 and long-term: 6001.52 ± 716.758, t (6) = 2.964, P = 0.025] [for H, short-term: 1820.01 ± 280.654 and long-term: 3678.52 ± 526.349, t (3.92) = 3.282, P = 0.031].These results were consistent with seizure intensity classifications based on the Racine stage.
Next, we examined the expression of GABAergic interneuron-related proteins (GAD65, GAD67, GABA, GABA A R α1, GABA B R1, and VGAT) via immunohistochemistry.All proteins were mainly detected in the neuropil (synaptic) region in the SO, SP, and SR of the hippocampal CA1 region (Fig. 2).Notably, no obvious damage to the brain tissue was observed in either group.
To compare GAD65, GAD67, GABA, GABA A R α1, GABA B R1, and VGAT expression between the short-and long-term kindling groups, we plotted immunostaining intensity in the neuropil region in the SO, SP, SR, and total hippocampal CA1 region (Fig. 3).Among these six proteins, GAD65 showed significant changes in staining intensity In the SO, the short-term group had significantly higher GAD65 intensity than the CTL and long-term groups (short-term vs. CTL, P = 0.015; short-term vs. long-term, P = 0.009).In the SP, the short-term group showed significantly higher GAD65 intensity than the CTL and long-term groups (short-term vs. CTL, P < 0.001; short-term vs. long-term, P = 0.045).In the SR, the shortterm group showed significantly higher GAD65 intensity than the long-term group (short-term vs. long-term, P = 0.017).Overall, the short-term group showed significantly higher GAD65 intensity than the CTL and long-term groups (short-term vs. CTL, P = 0.003; short-term vs. long-term, P = 0.034).Other GABAergic neuron-related proteins (GAD67, GABA, GABA A R α1, GABA B R1 and VGAT) did not show significant difference in expression levels in the three groups.These results indicate that GAD65 expression increased after short-term PTZ kindling, but not after long-term PTZ kindling.

GAD65 increased in SOM + cells but not in other subtypes in the hippocampal CA1 region
Next, we investigated which subtype of GABAergic interneuron increases GAD65 expression in short-term kindling.We analyzed the following four subtypes using four chemical markers: PV + , SOM + , CR + , and CCK + cells.Typical confocal microscopic images are shown in Fig. 4. All GABAergic subtypes were detected in the CTL, short-term, and long-term groups.
To compare GAD65 expression among the GABAergic subtypes, we plotted soma immunostaining intensity in the SO, SP, SR, and total hippocampal CA1 region (Fig. 5).Among these subtypes, only SOM + cells showed significant changes in GAD65 intensity in the SO, SP and In the SO, the short-term group showed significantly higher GAD65 intensity than long-term group (short term vs. long term: P = 0.039).In the SP, the short-term group showed significantly higher GAD65 intensity than the long-term group (short-term vs. long-term: P = 0.011).Overall, the short-term group showed significantly higher GAD65 intensity than the CTL and long-term groups (short-term vs. CTL, P = 0.037; short-term vs. long-term, P = 0.010).Other GABAergic subtypes (PV + , CR + , and CCK + cells) showed equivalent GAD65 expression among the three groups.Thus, short-term kindling upregulates GAD65 expression in SOM + cells, but not cells of the other subtypes.

Neither short-nor long-term kindling altered the cell density of GABAergic subtypes
Next, we investigated whether cell death of GABAergic neurons through excitotoxicity was a causal factor in seizure stage progression.Typical confocal microscopic images for each subtype in hippocampal CA1 neurons are shown in Fig. 6.
To compare the cell numbers of each subtype, we plotted the cell density of immunopositive cells in each layer (SO, SP, SR, and Total).The cell densities of the GABAergic subtypes (PV, SOM, CR, and CCK) did not significantly differ among the three groups in each hippocampal layer.Therefore, these GABAergic subtypes did not cause serious cell death after short-or long-term PTZ kindling in hippocampal CA1 neurons.

Discussion
PTZ kindling triggered bimodal changes in seizure stages, characterized by a temporary decline in seizures following short-term kindling Y. Kajita et al. and stable seizures following long-term kindling.GAD65 expression was elevated in short-term but not long-term kindling.GAD65 upregulation was detected in SOM + neurons but not in other GABAergic subtypes.

Short-term kindling enhances epileptic resistance
In PTZ kindling rats, seizure stage was suppressed after 8 or 10 injections.This suggests that a transient epileptic resistant state is induced through repeated PTZ injection.To the best of our knowledge, no study has reported a bimodal change in the epileptic seizure stage in PTZinduced kindling, although some studies have shown similar findings, in that seizure stages were decreased following approximately 10 injections in rats (Nassiri-Asl et al., 2010;Wen et al., 2017) and mice (Fleck et al., 2016;Nieoczym et al., 2021).
PTZ was once applied as a nervous system stimulant in therapy for psychiatric disorders based on the theory of biological antagonism  between epilepsy and schizophrenia (Fink, 1984;Kalinowsky, 1986), but this treatment has since shifted to electrical stimulation through electroconvulsive therapy (ECT) (Cooper and Fink, 2014).The voltage required to induce a seizure increases over the routine course of ECT (Scott and Boddy, 2000); one study reported that 94.4 % of patients demonstrated an increase in the seizure threshold (Duthie et al., 2015).Fig. 5. Expression of GAD65 among GABAergic subtypes in short-and long-term PTZ-kindled rats.Average intensity levels of GAD65 immunostaining in (A) PV + neurons (n = 7, 7, and 6 rats in the CTL, short-term, and long-term groups), (B) SOM + neurons (SO and total: n = 7, 7, and 6 in the CTL, short-term, and long-term groups, respectively; SP: n = 7, 6, and 6 for the CTL, short-term, and long-term groups, respectively; SR: n = 7, 5, and 6 for the CTL, short-term, and long-term groups, respectively), (C) CR + neurons (SO, SR, and total: n = 6 for all groups; SP: n = 5, 6, and 6 for the CTL, short-term, and long-term groups, respectively), and (D) CCK + neurons (SO, SP, and total: n = 7, 7, and 6 for the CTL, short-term, and long-term groups, respectively; SR: n = 7, 6, and 6 for the CTL, short-term, and long-term groups, respectively).Error bars indicate SEM.The Kruskal-Wallis test or one-way ANOVA followed by Bonferroni's or Dunnett's T3 test were used to compare mean expression levels (*P < 0.05).Our results indicate that repetitive stimulations other than electrical stimulation may also enhance epileptic resistance, and that the same endogenous mechanism may be involved in both PTZ kindling and ECT.Recent clinical studies report the cases of prolonged seizures after ECT, suggesting our hypothesis (Shin et al., 2023;Saltoglu et al., 2023).

Expression increases in GAD65 but not GAD67
The expression of GAD65, but not GAD67, was increased by shortterm kindling, which suggests that short-term kindling increases GAD65-dependent GABA synthesis.GABA synthesis is primarily dependent on GAD67 in the normal state (Rimvall et al., 1993;Dicken et al., 2015).By contrast, GAD65 has high affinity for the cofactor pyridoxal 5′-phosphate and is responsible for rapid, activity-dependent GABA synthesis (Martin and Rimvall, 1993).Notably, transgenic rat studies have demonstrated that a lack of GAD67 expression leads to greater reduction in GABA content (Fujihara et al., 2020) than a lack of GAD65 expression (Kakizaki et al., 2021).By contrast, the survival ratio is lower in GAD65 knockout rats than in GAD67 rats, with >80 % of GAD65 knockout rats dying due to spontaneous seizures.Therefore, increased GAD65 expression may be crucial for normal neural function under pathological conditions such as epileptic seizure.This hypothesis is supported by previous studies that have reported that repeated electroconvulsive shock (Jinno and Kosaka, 2009) or transcranial magnetic stimulation (Volz et al., 2013) increased GAD65, but not GAD67, in the hippocampus or cortex.Therefore, we speculate that seizure resistance associated with short-term kindling may be mediated by the upregulation of GAD65 expression.

Increased GAD65 expression is not accompanied by GABA upregulation
GAD65 expression increased following short-term PTZ kindling, suggesting that repeated stimulation strengthens GABAergic inhibitory pathways in the brain.This finding is consistent with previous research; electrical kindling in the performant path temporarily enhances the synaptic inhibition of granule cells in the DG (Maru and Goddard, 1987), and electrical kindling in the amygdala and hippocampus increases benzodiazepine receptors (Tuff et al., 1983).However, our results did not show an increase in GABA synthesis associated with the increase in GAD65.We hypothesize that GABA synthesis via GAD65 may increase immediately after PTZ injection and then quickly return to baseline levels within 24 h after injection.This hypothesis is supported by previous findings that a single electroconvulsive shock increases hippocampal GABA expression within 1 h after stimulation but returns to basal levels within 2 h after stimulation (Green, 1986) and that GABA upregulation can inhibit GABA synthesis through feedback inhibition in the hippocampus and striatum (Green et al., 1978;Green and Vincent, 1987).Although GABA synthesis is sensitive to acute stimulation, it is maintained at constant levels through adaptive responses to normal conditions.Such a feedback system may prevent the detection of increased GABA expression.

SOM + cells are related to epileptic susceptibility
There was increased GAD65 expression in SOM + cells after shortterm kindling, which returned to the CTL level after long-term kindling.In a previous study, SOM + cells had lower GAD65 levels than other GABAergic subtypes under normal conditions (Kajita and Mushiake, 2021).Therefore, it is possible that epileptic progression is suppressed by the upregulation of SOM + cells; however, such suppression would be transient, disappearing with repeated kindling.
SOM + cells are a GABAergic subtype targeting the distal part of dendrites; they exhibit the selective gating of excitatory inputs into pyramidal neurons (Taniguchi, 2014).In a previous study, silencing SOM + cells strongly increased burst firing without altering the spike phase, whereas PV + cell silencing had no effect on burst firing but shifted the spike phase preference in the hippocampal CA1 region (Royer et al., 2012).Compared to other subtypes, SOM + cells are selectively decreased in the hippocampus of epileptic rats (Sloviter, 1987;Buckmaster and Jongen-Rêlo, 1999;Cossart et al., 2001;Wyeth and Buckmaster, 2021).Therefore, SOM + cells appear to be particularly sensitive to excitotoxicity, such that the reduction of SOM + cells in epileptic animals reduces dendritic inhibition and contributes to the pathogenesis of epilepsy.Dendritic, but not somatic, GABAergic inhibition is decreased in the hippocampal CA1 region of epileptic rat models (Cossart et al., 2001), supporting this hypothesis.

Long-term PTZ injection does not produce fully kindled states
The expression of GAD65 or GABA A and GABA B receptors in the hippocampus did not change after long-term kindling, inconsistent with the findings of previous PTZ kindling studies (Getova et al., 1998;Hao et al., 2016;Szyndler et al., 2018).One possible reason for this discrepancy is that our rats did not reach Stage 5 seizure after long-term PTZ kindling (20 injections).Generally, two successive Stage 5 seizures after a single PTZ injection is considered to indicate a fully kindled state (Dhir, 2012).It is possible that further PTZ injections could enable the detection of these changes, because a recent study on PTZ kindling show that transient epileptic resistant states (no significant) in rats disappear with subsequent successive injections, and they acquired fully kindling states (Alam et al., 2023).Besides the number of injection, the dose of the PTZ could be the factor to determining whether the fully kindled state likely to induce or not.Although our dose (40 mg/kg) is in the commonly used range (Han et al., 2000;Acar et al., 2021), the seizure stages were started from Stage 2 after the first injection in short-and long-term kindling.These results indicate that some level of seizers (not tonic-clonic seizures) were induced after the first injection.These phenomena are similar to the concept in ECT, which defines the treatment threshold as the amount of stimulated electricity above the seizure threshold.The reason why our rats did not reach fully kindled states may be that enhanced epileptic resistance, such as that observed with ECT (Scott and Boddy, 2000;Bajbouj et al., 2005), was also induced in our PTZ kindling procedure.
On the other hand, activating the endogenous inhibitory system may be necessary to induce the high susceptibility to epileptic seizure.The following reports suggest that heightened GABAergic tonus is the predecessor of consequent increasing seizure susceptibility.Hyperexcitable interneurons responsible for recurrent inhibition contribute to slowwave genesis in the row of spike-wave discharges and the absence epilepsy manifestations (Gloor, 1979).In addition, induction of seizures after cessation of long-lasting artificial (iontophoretic) raising GABA levels in neuronal structures (the syndrome of GABA withdrawal) is known (Montiel et al., 2000).Our findings of bimodal character of PTZ kindling are consistent with these previous reports.

GAD65-mediated SOM release may cause bimodal changes in seizure stage
In contrast to other subtype markers such as PV, CR, and CCK, SOM acts as an anti-epileptic neuropeptide in the hippocampus (Boehm and Betz, 1997).The SOM release-induced hyperpolarization of CA1 neurons is not caused by changes in GABA release or effects on GABA receptors (Xie and Sastry, 1992).Previous report shows that SOM + cells increase in the hippocampal CA1 region in electrical kindling rats (Botterill et al., 2017).However, our results did not show increases in SOM + cells in the hippocampal CA1 region of PTZ-kindled rats.Because neurogenesis is restricted to the subgranular zone of the DG or subventricular zone of the lateral ventricle in the adult brain (Ming and Song, 2005), we speculate that higher numbers of SOM + cells were identified due to increased somatostatin expression.Spontaneous or stimulation-evoked SOM release in the DG is enhanced following electrical kindling (Marti et al., 2000).Increased SOM expression mediated by adeno-associated virus vector suppresses kindling development (Zafar et al., 2012;Leibowitz et al., 2019).In addition, increased SOM detected in the hippocampus in Stage 2 is equivalent to that in Stage 5 in electrically kindled rats (Vezzani et al., 1996).These findings suggest that SOM upregulation during the initial phase of kindling suppresses kindling stage progression.
SOM and NPY, another anti-epileptic neuropeptide, are coexpressed in a subset of GABAergic neurons in the hippocampus (Köhler et al., 1987).The depolarizing action of GABA may synchronize the rapid release of SOM and NPY (Fu and van den Pol, 2007).According to our findings, GAD65 upregulation may facilitate SOM release from presynaptic terminals via auto-receptors, and neuropeptide/GABA co-release provides epileptic-resistant states in the initial phase of pathological conditions such as epileptic seizure.However, this possibility remains in the realm of speculation and requires further verification.
A limitation of this study is that our finding based on only immunohistochemical staining in the hippocampus.In the future, the accuracy of our hypothesis should be examined by complementary data through other molecular biological techniques in multiple regions other than hippocampus.

Fig. 4 .
Fig. 4. Immunostaining of GAD65 among GABAergic subtypes in short-and long-term PTZ-kindled rats.Triple immunofluorescence images show GAD65 (green); PV, SOM, CR, or CCK (magenta); and DAPI (blue) signals in the (A-D) CTL, (E-H) short-term, and (I-L) long-term groups.White squares indicate regions enlarged in panels on the right.Arrowheads indicate somata of each subtype.

Fig. 6 .
Fig. 6.Cell density levels of GABAergic subtypes in short-and long-term PTZ-kindled rats.Immunostaining results show PV + , SOM + , CR + , and CCK + signals in the (A-D) CTL, (E-H) short-term, and (I-L) long-term groups.White squares indicate the enlarged SP regions.(M-P) Average cell density levels for each subtype (n = 7, 6, and 5 rats for the CTL, short-term, and long-term groups, respectively).Error bars indicate SEM.One-way ANOVA followed by Bonferroni's or Dunnett's T3 test was performed to compare mean expression levels.