microRNA‐132 is overexpressed in glia in temporal lobe epilepsy and reduces the expression of pro‐epileptogenic factors in human cultured astrocytes

Abstract Temporal lobe epilepsy (TLE) is a chronic neurological disease in humans, which is refractory to pharmacological treatment in about 30% of the patients. Reactive glial cells are thought to play a major role during the development of epilepsy (epileptogenesis) via regulation of brain inflammation and remodeling of the extracellular matrix (ECM). These processes can be regulated by microRNAs (miRs), a class of small non‐coding RNAs, which can control entire gene networks at a post‐transcriptional level. The expression of miRs is known to change dynamically during epileptogenesis. miR‐132 is one of the most commonly upregulated miRs in animal TLE models with important roles shown in neurons. However, the possible role of miR‐132 in glia remains largely unknown. The aim of this study was to characterize the cell‐type specific expression of miR‐132 in the hippocampus of patients with TLE and during epileptogenesis in a rat TLE model. Furthermore, the potential role of miR‐132 was investigated by transfection of human primary cultured astrocytes that were stimulated with the cytokines IL‐1β or TGF‐β1. We showed an increased expression of miR‐132 in the human and rat epileptogenic hippocampus, particularly in glial cells. Transfection of miR‐132 in human primary astrocytes reduced the expression of pro‐epileptogenic COX‐2, IL‐1β, TGF‐β2, CCL2, and MMP3. This suggests that miR‐132, particularly in astrocytes, represents a potential therapeutic target that warrants further in vivo investigation.


| INTRODUCTION
Epilepsy is a common chronic neurological disease that is estimated to affect more than 65 million people worldwide (WHO [Epilepsy fact sheet], n.d.). Temporal lobe epilepsy (TLE) is the most common form of epilepsy in adults, which is refractory to pharmacological treatment in about 30% of patients. TLE accounts for over 70% of all surgical cases of epilepsy and is often associated with hippocampal sclerosis (HS), tumors, or malformations of cortical development (Blumcke et al., 2017). Neuropathological features associated with TLE include disruption of the blood-brain barrier (BBB), chronic neuroinflammation, gliosis, and remodeling of the extracellular matrix (ECM; Gorter, van Vliet, & Aronica, 2015). Previous studies in animal TLE models have demonstrated that these pathogenic and adaptive events occur early after an epileptogenic insult in the rodent brain and are present throughout epileptogenesisthe process by which normal brain tissue becomes susceptible to generation of recurrent spontaneous seizures (Loscher, 2011;Pitkanen, Lukasiuk, Dudek, & Staley, 2015). Currently, treatments that prevent or modify epileptogenesis in patients do not exist, thus there is a need for new therapies with anti-epileptogenic and disease-modifying properties (Loscher, Klitgaard, Twyman, & Schmidt, 2013).
Astrocytes play a key role in this process: under normal physiological conditions, they provide trophic support and maintenance to neurons, but under pathological conditions they undergo morphological and functional changes and acquire reactive properties (Sofroniew & Vinters, 2010). Reactive astrocytes have been shown to display both pro-inflammatory neurotoxic and anti-inflammatory neuroprotective phenotypes, which activate various signaling pathways (Anderson et al., 2016;Zamanian et al., 2012). Among the key pathogenic signaling pathways in TLE are the pro-inflammatory pathway mediated by cytokine interleukin-1 beta (IL-1β; Vezzani & Baram, 2007) and the multifunctional transforming growth factor beta (TGF-β) pathway (Heinemann, Kaufer, & Friedman, 2012). Under pathological conditions these pathways are responsible for the production of the ECM components and remodeling of the ECM structure in the brain (Lau, Cua, Keough, Haylock-Jacobs, & Yong, 2013), which largely depends on the production and activation of the key proteolytic enzymes of the extracellular space-matrix metalloproteinases (MMPs; Rempe, Hartz, & Bauer, 2016). Chronic neuroinflammation and remodeling of the ECM lead to neurodegeneration and establishment of aberrant synaptic connections, which may ultimately result in epileptogenesis. The modulation of the pathways regulating inflammation and the ECM may be a potential strategy to prevent epileptogenesis. microRNAs (miRs) is a class of small non-coding RNAs, which can control entire gene networks at a post-transcriptional level (Ambros, 2004; Bartel, 2018). miRNAs have been shown to regulate versatile cellular processes both in healthy conditions and in neurological disorders (Cao, Li, & Chan, 2016), including epilepsy (Brennan & Henshall, 2018), which makes them candidate agents for novel antiepileptogenic treatments. miRNAs display a dynamic expression profile in the hippocampus of TLE rats and are associated with the regulation of multiple signaling pathways (Gorter et al., 2014). Among the key miRNAs deregulated in epilepsy are miRNAs enriched in astrocytes and associated with brain inflammation (Aronica et al., 2010;Iyer et al., 2012;Korotkov et al., 2018;van Scheppingen et al., 2016;van Scheppingen et al., 2018). In a recent meta-analysis of differentially expressed miRNAs across rat TLE models, we identified miR-132 as one of the most commonly upregulated miRNAs (Korotkov, Mills, Gorter, van Vliet, & Aronica, 2017). miR-132 has been associated with the innate immune response and brain inflammation (Soreq & Wolf, 2011); however, its functions in the brain have been studied mostly in neurons (Aten, Hansen, Hoyt, & Obrietan, 2016). Given the multitude of reactive glial responses in chronic neurological diseases, such as epilepsy, and the inducible nature of miR-132 expression, we hypothesized that reactive glia may express miR-132 in TLE. The aim of this study was to characterize the cell-type specific expression of miR-132 in human and rat TLE hippocampus and investigate the effects of miR-132 on the expression of pro-epileptogenic factors in cultured human astrocytes.

| Human brain tissue
The cases included in this study were obtained from the archives of the department of Neuropathology of the Amsterdam UMC, the Netherlands. A total of 16 brain specimens were examined from patients who underwent surgery for drug-resistant TLE. Tissue was obtained and used in accordance with the Declaration of Helsinki and the Amsterdam UMC Research Code provided by the Medical Ethics Committee. All cases were reviewed independently by two neuropathologists and the classification of HS was based on analysis of microscopic examination as described by the International League Against Epilepsy (HS ILAE Type 1, n = 12; HS ILAE Type 2, n = 4; Blumcke et al., 2013). Control material was obtained during autopsy of people without a history of seizures or other neurological diseases (n = 10).
Brain tissue was fixed in 10% buffered formalin and embedded in paraffin or directly frozen and stored at -80 C. The clinical findings of human samples are presented in Table 1.

| Experimental animals
Adult male Sprague Dawley rats (Harlan Netherlands, Horst, the Netherlands) were used in this study which was approved by the University Animal Welfare committee and performed in accordance with the guidelines of the European Community Council Directives 2010/63/ EU. The rats were housed individually in a controlled environment (21 ± 1 C; humidity 60%; lights on 08:00 a.m. -8:00 p.m.; food and water available ad libitum).

| Electrode implantation and status epilepticus induction
Rats were anesthetized with an intramuscular injection of ketamine (74 mg/kg; Alfasan, Woerden, the Netherlands) and xylazine (11 mg/kg; Bayer AG, Leverkusen, Germany), and placed in a stereotactic frame. In order to record hippocampal EEG, a pair of insulated stainless steel electrodes (70 μm wire diameter, tips 800 μm apart) was implanted into the left dentate gyrus (DG) under electrophysiological control, as described previously (Gorter et al., 2006). A pair of stimulation electrodes was implanted in the angular bundle. Two weeks after recovery from the operation, each rat was transferred to a recording cage (40 × 40 × 80 cm) and connected to a recording and stimulation system (NeuroData Digital Stimulator, Cygnus Technology Inc., Delaware Water Gap, PA) with a shielded multi-strand cable and electrical swivel (Air Precision, Le Plessis Robinson, France). A week after habituation to the new condition, rats underwent tetanic stimulation (50 Hz) of the hippocampus in the form of a succession of trains of pulses every 13 s. Each train was of 10 s duration and consisted of biphasic pulses (pulse duration 0.5 ms, maximal intensity 700 μA). Stimulation was stopped when the rats displayed sustained forelimb clonus and salivation for several minutes, which usually occurred within 1 hr. Immediately after termination of the stimulation, periodic epileptiform discharges occurred at a frequency of 1-2 Hz which lasted on average 10.0 ± 0.4 hr (status epilepticus, SE). Rats had frequent seizures during this period as observed by both their behavior and EEG.
To determine seizure frequency, continuous EEG recordings (24 hr/day) were made in all rats. Hippocampal EEG signals were amplified (10×) by a high impedance headstage connected to an amplifier (20x; CyberAmp, Axon Instruments, Burlingame, CA), bandpass filtered (1-60 Hz) and digitized by a computer. A seizure detection program (Harmonie, Stellate Systems, Montreal, Canada) sampled the incoming signal at a frequency of 200 Hz per channel. All EEG recordings were visually screened and seizures were confirmed by trained human observers. Seizures were characterized by synchronized high-voltage amplitude oscillations and were scored when the amplitude increased more than twofold and lasted for at least 10 s.

| Tissue preparation
For in situ hybridization, rats were deeply anesthetized with pentobarbital (Euthasol, AST Farma, Oudenwater, the Netherlands, 60 mg/kg ip) and perfused via the ascending aorta (300 ml 0.37% Na 2 S / 300 ml 4% paraformaldehyde in 0.1 M phosphate buffer, For RT-qPCR analysis, rats were decapitated 1 day after SE (acute phase, n = 5), 1 week after SE (latent phase, n = 6) or 3-4 months after SE (chronic phase, n = 5). Electrode-implanted control rats were also included (n = 5). The brain was dissected and the parahippocampal cortex, which includes mainly the entorhinal cortex and parts of the perirhinal and posterior piriform cortex, was removed by incision at the ventro-caudal part underneath the rhinal fissure until approximately 5 mm posterior to bregma, as well as the hippocampus. The hippocampus was sliced into smaller parts (200-300 μm) and the DG and Cornu Ammonis (CA1) regions were cut out of the slices in 4 C saline solution under a dissection microscope. All material was frozen on dry ice and stored at −80 C until use.

| Cell cultures
Primary fetal astrocyte-enriched cell cultures were derived from human fetal brain tissue (14-20 weeks of gestation) obtained from medically induced abortions. All material was collected from donors from whom a written informed consent for the use of the material for Island, NY). Cell isolation was performed as follows: meninges and blood vessels were removed, tissue was minced and dissociated by incubation with 2.5 mg/ml trypsin at 37 C for 20 min, followed by inactivation of trypsin with astrocyte medium. The tissue was passed through a 70 μm mesh filter and the cell suspension was transferred into a flask with fresh astrocyte medium and maintained in a 5% CO 2 incubator at 37 C. After 48 hr incubation, the medium was replaced with fresh medium and was subsequently refreshed twice a week.
Cultures reached confluence after 2-3 weeks. Astrocytes were used at passages 2-5. More than 98% of the cells in primary culture, as well as in the successive passages were strongly immunoreactive for the astrocytic marker glial fibrillary acid protein (GFAP) and S100β as previously reported (Iyer et al., 2012).

| Treatment of cell cultures
Cells were plated in poly-L-lysine coated plates (5 × 10 4 cells/well in 12-well plates for RNA analysis or 2 × 10 5 cells/well in 6-well plates for protein analysis) and were transfected with miR-132 mimic, ratios of all samples were higher than 2.0 (tissue material) or higher than 1.7 (cell culture material) and did not differ between autopsy control and surgical tissue, indicating that the material was of good quality. To evaluate mRNA expression, oligo dT primers (2.5 nmol) were annealed to 250 ng total RNA (cell culture material) or 2,000 ng (tissue material) in a total volume of 12.5 μl by incubation at 72 C for 10 min, and cooled to 4 C. Reverse transcription was performed by the addition of 12.5 μl RT-mix, containing: First Strand Buffer Quantification of data was performed using LinRegPCR in which a baseline correction and window-of-linearity were determined for each sample separately, followed by a linear regression analysis on the Log (fluorescence) per cycle to fit a straight line through the PCR data set.
The slope of this line is used to determine the PCR efficiency of each individual sample. The mean PCR efficiency per amplicon and the Ct value per sample are used to calculate a starting concentration N0 per sample, which is expressed in arbitrary fluorescence units (Ruijter et al., 2009). The starting concentration N0 of each specific product was then divided by the geometric mean of the starting concentrations N0 of the reference genes and this ratio was compared between groups.

| In situ hybridization on human and rat brain tissue
Paraffin-embedded brain tissue was deparaffinised in xylene and rinsed in ethanol (2× 100%, 1× 70%) and sterile water. Antigen retrieval was performed using a pressure cooker in sodium citrate buffer, pH 6.0, at 121 C for 10 min. The oligonucleotide probe against miR-132-3p (Table S1)

| Evaluation of in situ hybridization
The expression of miR-132 was analyzed in the hippocampus of human and rat brain tissue using two different approaches: the optical density (OD) approach and the in situ reactivity score (IRS) approach.
For the OD approach, the OD above threshold of the hybridization signal in microphotographs obtained by in situ hybridization was measured using ImageJ. Measurements were performed in the granule cell layer (GCL), molecular layer, and hilus of the DG, the pyramidal cell layer (PCL) of CA1, as well as the stratum radiatum and stratum oriens. For the in situ reactivity approach, the intensity of the hybridization signal was evaluated in neurons and glia in the DG, hilus, and CA1 using a scale of 1-4 (1: no; 2: weak; 3: moderate; 4: strong hybridization signal). The score represents the predominant signal intensity found in each case. Furthermore, the relative number of positive cells (0: no; 1: single to 10%; 2:11-50%; 3: >50%) was also evaluated in these areas. Then the in situ reactivity score (IRS) was calculated by multiplying the intensity score by the relative number score as described previously (Broekaart et al., 2017).

| Statistical analysis
Statistical analyses were performed using IBM SPSS Statistics 21. Comparisons between groups were done using the Mann-Whitney U test.
A value of p < .05 was assumed to indicate significant difference. OD analysis demonstrated that the expression of miR-132 during the acute stage was~5-8-fold higher in the DG ML, CA1 SR, and CA1 SO as compared to control (Figure 1h, all p < .05). The IRS approach, in which neurons were discriminated from glial cells, showed that neuronal expression of miR-132 was higher in the DG, CA1, and hilus at the acute stage (Table 2, p < .05) as compared to control. Higher expression at the acute stage was also observed in the cells with glial morphology in the DG (p < .05), in the hilus (p < .01), and in the CA1 (p < .01; Table 2). During the latent stage, the OD did not differ from control in any of the analyzed hippocampal layers (Figure 1h); however, a higher expression in the cells with glial morphology was observed in the CA1 (Table 2  numerous stained cells were observed in CA1 SO and SR (Figure 1g).
The OD during the chronic stage was~2-4-fold higher in the SR (p < .01) and SO (p < .05), while twofold lower in hilus (p < .05) as compared to control (Figure 1h). Higher expression at the chronic stage was also observed in the cells with glial morphology in the DG (p < .05), hilus (p < .05), and CA1 (all p < .01; Table 2). The OD did not change in the GCL and PCL at any time point as compared to control ( Figure 1h). Double labeling of miR-132 with cell-type specific markers revealed that the expression of miR-132 was co-localized with the neuronal marker NeuN (Figure 1i), the astrocyte marker GFAP (Figure 1j), and the microglia marker Iba1 (Figure 1k).

| Higher miR-132 expression in the human hippocampus
Following the findings in the rat hippocampus, we investigated the expression of miR-132 in resected hippocampi from patients with drug-resistant TLE as well as from autopsy controls (Table 1) (Table 3).
Double labeling of miR-132 with cell-type specific markers showed that miR-132 co-localized with virtually all NeuN + cells in the human TLE-HS hippocampus, including granule cells, pyramidal cells and neurons in the hilus (Figure 3a). In addition to neuronal expression, miR-132 was found in astrocytes based on co-localization with GFAP (Figure 3b), excitatory amino acid transporter 1 (EAAT1; Figure 3c) and glutamine synthetase (GS; Figure 3d). The expression in astrocytes was found throughout hippocampus, in the CA1 and hilus, as well as in the DG ML. The co-localization of miR-132 with the marker of microglial cells Iba1 (Figure 3e) was also observed, which was most prominent in the CA1. Additionally, miR-132 expression was closely associated with the brain endothelial cell marker CD34 (Figure 3f). We further assessed whether miR-132 expression was co-localized with markers of astrocyte and microglia reactivity.
to a 30% lower expression of TGFB2 (p < .05) under TGF-β1 stimulation ( Figure 4b). The expression of other predicted miR-132 target genes SMAD2 (Figure 4c) and THBS1 (Figure 4d) did not change following miR-132 modulation. Overexpression of miR-132 also resulted in 2.2-fold higher expression of TGFBR2 (p < .05, Figure 4e). Western blot analysis for TGF-β2 protein confirmed a 20% lower expression after miR-132 overexpression (Figure 4f, g). These data shows that miR-132 may be involved in the regulation of the pro-epileptogenic TGF-β pathway in human astrocytes.

| miR-132 overexpression downregulates IL-1βinduced factors
Next, we investigated the potential role of miR-132 in astrocytes after stimulation of cells with IL-1β, another pro-epileptogenic pathway associated with neuroinflammation. As expected, the expression of a number of inflammatory genes was more than twofold higher after IL-1β stimulation: prostaglandin-endoperoxide synthase 2 (PTGS2), IL1B, C-C Motif Chemokine Ligand 2 (CCL2), IL6 (all p < .001), as well as MMP3 (p < .001) and MMP9 (p < .01; Figure 5a). The expression of miR-132 did not change following IL-1β treatment ( Figure S1b). The expression of the miR-132 predicted target gene PTGS2 was 20% lower (p < .01) after miR-132 overexpression and 50% higher (p < .001) after miR-132 inhibition ( Figure 5b). The same pattern was found for IL1B (Figure 5c). CCL2 gene expression was also lower (p < .01) after miR-132 overexpression (Figure 5d). The expression of IL6 was not modulated by miR-132 (Figure 5e). The expression of MMP3 was modulated in the same way as PTGS2 and IL1B, with miR-132 mimic decreasing MMP3 by 40% (p < .001) and miR-132 antagomir further increasing F I G U R E 2 Expression of miR-132 in the human hippocampus. (a)-Taqman RT-qPCR showed a 3.4-fold higher (p < .001) miR-132 expression in the hippocampus of patients with TLE-HS; (b-j)-in situ hybridization for miR-132 in the hippocampus of TLE without HS (e-g) and TLE with HS (h-j) patients compared to autoptic control (b-d); mostly neuronal expression of miR-132 was observed in the DG (b), hilus (c), and CA1 regions of control hippocampus. Higher expression was observed in cells with glial morphology in the DG ML (e, here and further indicated by black arrowheads), hilus (f) and CA1 (g) which was most prominent in the DG ML of TLE-HS (h); hilus (i) and CA1 (j) and was associated with GCL dispersion and severe neuronal loss in the hilus and CA1 PCL; (k)-OD analysis showed a higher expression in CA1 SR (p < .05) and a lower expression in the GCL and PCL (both p < .05) of the TLE-HS hippocampus relative to control; DG, dentate gyrus; GCL, granule cell layer; Hil, hilus; ML, molecular layer of the dentate gyrus; PCL, pyramidal cell layer; SO, stratum oriens; SR, stratum radiatum; scale bar 100 μm; Mann-Whitney U test *p < .05, ***p < .001 [Color figure can be viewed at wileyonlinelibrary.com] (p < .001) MMP3 expression 2.2-fold (Figure 5f). The expression of MMP9 was also 1.8-fold higher after miR-132 inhibition (p < .001, Figure 5g), although a trend toward upregulation was also seen for miR-132 overexpression. TargetScan prediction showed a predicted binding site for miR-132 in the 3 0 UTR of PTGS2 mRNA ( Figure S1c); therefore, we performed a western blot analysis, which confirmed that COX-2 protein expression was also 40% lower after miR-132 overexpression in astrocytes (p < .01; Figure 5h,i), confirming mRNA data. Thus, miR-132 may act as a negative regulator of pro-epileptogenic gene expression in human astrocytes during inflammatory conditions.

| DISCUSSION
We investigated the expression of miR-132 in the epileptogenic rat and human hippocampus as well as the effects of miR-132 modulation in human cultured astrocytes after stimulation with epilepsy-T A B L E 3 In situ reactivity scores for miR-132 + cells in the human hippocampus DG Hilus CA1
F I G U R E 3 Double labeling of miR-132 with cell type-specific markers in human TLE-HS. (a)-miR-132 was expressed in neurons throughout hippocampus, including granule cells, pyramidal cells and hilar neurons; colocalization of miR-132 was also found with markers of astrocytes GFAP (b), EAAT1 (c) and GS (d); miR-132 was also co-localized with Iba1 + microglial cells (e) in the areas of glial scar in CA1 and with CD34 (f) associated with blood vessels; the co-localization with reactivity markers showed miR-132 expression in vimentin + cells with astrocytic morphology (g), HLA-DR/DP/DQexpressing cells with microglial morphology (h) and STAT3 + cells in the CA1; black arrows indicate cells shown in higher magnification in insets; scale bar 50 μm; scale bar in inset a =10 μm and applies for insets a-i [Color figure can be viewed at wileyonlinelibrary.com] associated cytokines TGF-β1 and IL-1β. We showed a persistently increased expression of miR-132 in the rat and human epileptogenic hippocampus, particularly in glial cells. Furthermore, we showed that miR-132 can negatively regulate the expression of pro-epileptogenic factors in cultured human astrocytes. These findings will be further discussed in more detail in the following paragraphs.

| miR-132 expression is increased in the rat and human epileptogenic hippocampus
miR-132 is a cAMP-response element binding protein (CREB)regulated miRNA, enriched in neurons (Jovicic et al., 2013), and involved in the regulation of neuronal morphogenesis (Vo et al., 2005) and dendritic spine density (Hansen, Sakamoto, Wayman, Impey, & Obrietan, 2010;Wayman et al., 2008), which implicates miR-132 in the regulation of neuronal plasticity, memory, and learning (Aten et al., 2016). The expression of miR-132 is dysregulated in a number of neurological pathologies (Soreq & Wolf, 2011) and increased miR-132 expression was also found in resected brain tissue of children with TLE (Peng et al., 2013;Ren, Zhu, & Li, 2016). Here, we report that expression of miR-132 is increased in the hippocampus of adult TLE-HS patients. Moreover, miR-132 was found to be increased in the epileptogenic DG of rats already at the acute stage of electricallyinduced epileptogenesis, which confirmed our previous microarray findings (Gorter et al., 2014), as well as reports from other large-scale gene expression studies (Bot, Debski, & Lukasiuk, 2013;Kretschmann et al., 2015). The increased miR-132 expression in the rodent hippocampus has also been shown following administration of other proepileptogenic stimuli including pilocarpine (Nudelman et al., 2010;Peng et al., 2013) and kainic acid (Jimenez-Mateos et al., 2011), and a meta-analysis of differentially expressed miRNAs across various animal TLE models identified miR-132 as one of the most commonly upregulated miRNAs during the acute, latent and chronic stages of epileptogenesis (Korotkov et al., 2017), highlighting its potential involvement in the pathophysiology of epilepsy. Our histological analysis showed that miR-132 expression in the rat hippocampus was increased in previously quiescent glial cells, including astrocytes and microglia. However, previous studies both in healthy and pathological tissue have focused exclusively on neuronal miR-132 expression and little is known about potential miR-132 expression and function in F I G U R E 4 miR-132 overexpression modulates TGF-β pathway. The stimulation of primary human fetal astrocytes with TGF-β1 resulted in higher expression of TGFB1 (p < .01), TGFB2 (p < .001), TGFBR1 (p < .001), THBS1 (p < .01), MMP2 (p < .001), MMP14 (p < .001), and miR-132 (p < .001) as compared to non-stimulated control (a); the higher expression of TGFB2 was attenuated by miR-132 mimic transfection (p < .05) after TGF-β1 stimulation (b); no change was found in expression of SMAD2 (c) and THBS1 (d); the expression of TGFBR2 was higher after miR-132 transfection (p < .001) (e); western blot analysis showed that the expression of TGF-β2 protein was lower (p < .01) after miR-132 transfection (f, g); the data was normalized to the stimulated control condition and the dotted line indicates non-stimulated control; veh, vehicle (Lipofectamine 2000); α-132 indicates antagomir; n = 5 for RT-qPCR, n = 3 for western blot, Mann-Whitney U test, *p < .05, **p < .01, ***p < .001 glial cells. Therefore, we further investigated whether this phenomenon could also be observed in human TLE-HS tissue.

| miR-132 is expressed in glial cells in human TLE-HS
Increased expression of miR-132 was detected in astrocytes and microglia within the hippocampus of TLE patients. At the same time, the expression of miR-132 in the principal neuronal layers of the hippocampus was reduced, consistent with the phenomena of granule cell dispersion in the DG and neuronal damage in the CA1. We observed moderate-to-severe astrogliosis in the hippocampus of TLE patients.
However, it has been shown that various subtypes of reactive astrocytes may exist based on their gene expression profiles (Zamanian et al., 2012). The pro-inflammatory astrocyte response may lead to a loss of normal physiological functions by astrocytes, which become a source and mediators of chronic neuroinflammation (Aronica, Ravizza, Zurolo, & Vezzani, 2012;. On the other hand, reactive astrocytes may participate in glial scar formation and produce neurotrophins and thrombospondins, aiding synapse formation and repair Sofroniew, 2015), which may lead to epilepsy. Therefore, we investigated whether miR-132 expression was co-localized with certain markers of reactive glial cells.
miR-132 expression in the hippocampus of patients with TLE was observed in association with various astrocyte markers, including GFAP, EAAT1, GS, and vimentin, a marker of reactive astrocytes.
Notably, the increased expression of miR-132 in the endothelium of human tumors and hemangiomas has been shown to promote neovascularization (Anand et al., 2010). It is worth noting that neurons-the primary source of miR-132 in the brain-may be able to secrete miR-132-carrying exosomes, which could influence vascular integrity (Xu et al., 2017). This suggests that miR-132 in non-neuronal cells could have an exogenous origin, however we observed a mostly perinuclear miR-132 hybridization signal in astrocyte somas and no hybridization signal in astrocytic processes. Thus, miR-132 is expressed in a wide range of cell types in the rat and human TLE hippocampus and could be associated with diverse functions, extending beyond the known functions in neurons.

| miR-132 acts as a negative regulator of proepileptogenic factors induced by TGF-β1 and IL-1β
In order to study the potential functions of miR-132 in astrocytes, we performed experiments in primary human astrocyte cultures after stimulation with cytokines associated with epileptogenesis. The analysis of predicted target genes identified the TGF-β pathway as one of the most enriched pathways associated with miR-132. The activation of the TGF-β pathway by TGF-β1 in astrocytes not only increased the transcription of the genes directly involved in the pathway, such as TGFB1, TGFB2, and TGFBR1, but also led to an increase in ECMrelated genes THBS1 and matrix metalloproteinases MMP2 and MMP14. Interestingly, the stimulation with TGF-β1 also led to an increased expression of miR-132. Previously, miR-132 induction has been shown in human astrocytic cell line by a pro-inflammatory myeloid-related protein (MRP) 8 (Kong et al., 2015) and in primary rat astrocytic cultures by basic fibroblast growth factor (bFGF; Numakawa et al., 2011), confirming the inducible nature of miR-132 expression in glial cells. We investigated the effects of miR-132 overexpression on the expression of its predicted targets within the TGF-β pathway and found that the expression of TGFB2 as well as TGF-β2 protein could be attenuated by miR-132. TGF-β signaling can have pro-epileptogenic functions after a brain insult. The disruption of the BBB and extravasation of blood albumin has been shown to activate TGF-β signaling in astrocytes and contribute to epileptogenesis (Cacheaux et al., 2009;Ivens et al., 2007), potentially through induction of synaptogenesis (Weissberg et al., 2015). Moreover, the treatment of rats with an angiotensin II Type 1 (AT1) receptor antagonist losartan can block TGF-β signaling and prevent the development of recurrent spontaneous seizures and seizure severity following the BBB breakdown (Bar-Klein et al., 2014). In this regard, the modulation of TGF-β2 by miR-132 could be beneficial for prevention of epileptogenesis. However, miR-132 overexpression lead to an increased expression of TGFBR2, which is also increased upon suppression of miR-211, and associated with a nonconvulsive electrographical seizures and increased convulsive seizure susceptibility in mice (Bekenstein et al., 2017). Moreover, TGF-β signaling in astrocytes has been also shown to be necessary for limiting neuroinflammation in a mouse experimental stroke model (Cekanaviciute et al., 2014), whereas treatment of rat primary astrocytes with TGF-β1 could completely reverse the pro-inflammatory astrocytic phenotype caused by IL-1α, tumor necrosis factor alpha (TNF-α) and complement component 1q (C1q) stimulation .
We next assessed the effect of miR-132 on pro-epileptogenic factors associated with a pro-inflammatory cytokine IL-1β. Among the predicted targets of miR-132 are PTGS2 and its product COX-2, a key regulator of inflammation mediated by IL-1β. We found that both PTGS2 and COX-2 were decreased following miR-132 overexpression.
COX-2 is normally expressed in neurons, but it has also been found in astrocytes in human TLE (Desjardins et al., 2003) and in an animal TLE model (Holtman et al., 2009). Moreover, the expression of other proinflammatory genes associated with epilepsy, such as IL1B and CCL2, was decreased by miR-132 overexpression in astrocytes. This suggests an anti-inflammatory mode of action by miR-132. This is in accordance with previous studies, since miR-132 has been shown to have anti-inflammatory action in the brain via targeting acetylcholinesterase (Shaked et al., 2009). Moreover, inhibition of miR-132 in a human astrocytic cell line under stimulation with MRP8 promoted the expression of pro-inflammatory genes IL1B, IL6, and TNF, whereas upregulation of miR-132 suppressed the transcription of these genes (Kong et al., 2015).
In addition, both TGF-β1 and IL-1β are responsible for transcription of genes, which take part in ECM remodeling. We found that TGF-β1 stimulation increased the constitutively expressed MMP2 and MMP14, whereas the inducible MMP3 and MMP9 could be increased by IL-1β stimulation in human astrocytes. MMPs are the major enzymes responsible for the remodeling of the ECM (Rempe et al., 2016) and we previously demonstrated the dynamic expression of Mmp2, Mmp3, Mmp9, and Mmp14 during epileptogenesis in the rat hippocampus (Gorter et al., 2007;Korotkov et al., 2018). miR-132 overexpression negatively regulated MMP3 and inhibition of miR-132 led to a further increase in MMP9, a metallopeptidase implicated in epileptogenesis (Wilczynski et al., 2008). In summary, miR-132 may act as a negative regulator of various pro-epileptogenic factors induced by TGF-β1 and IL-1β stimulation in human astrocytes.

| Therapeutic potential of miR-132
Due to the suppression of pro-epileptogenic factors (including proinflammatory and ECM factors), it would be interesting to investigate whether miR-132 could be used as a novel strategy to inhibit epileptogenesis. It has been previously shown that intracerebroventricular injection of an antagomir of miR-132 protected against hippocampal CA3 neuronal death 24 hr after SE induced via intra-amygdala injection of kainic acid (Jimenez-Mateos et al., 2011).
This suggests that miR-132-overexpressing therapy can aggravate neuronal damage, however, the authors did not investigate the later time points after SE, at which brain inflammation and ECM alterations are much more prominent. It is interesting to note that inhibition of miR-132 in astrocytes further augmented the expression of inflammatory genes in our study, which may result in exacerbation of astrocyte-mediated inflammatory response in case of using miR-132 inhibitors in vivo. Therefore, this needs to be studied and the optimal time window for miR-132 therapy has to be determined.

| CONCLUSIONS
Our experiments showed that miR-132 is persistently increased in glial cells in the human and rat epileptogenic brain. Since miR-132 can act as a negative regulator of epilepsy-associated factors in cultured human astrocytes, the therapeutic potential of miR-132, specifically in astrocytes, needs to be further studied in in vivo experiments.

ACKNOWLEDGMENTS
The research leading to these results has received funding from the The Netherlands) for providing fetal tissues.