NRSF regulation in in vitro models of seizures

(1)

COGNITIVE SCIENCE CENTER AMSTERDAM – BRAIN AND COGNITIVE SCIENCES

NRSF regulation in in vitro

models of seizures

Sicco D. de Knecht, B.Sc.

Supervisor: Tallie Z. Baram, M.D. Ph.D. September 2011 – May 2012

NRSF is a transcription factor known to regulate epilepsy-related genes after seizures and to play a role in epileptogenesis. We investigated the regulation of NRSF in organotypic hippocampal slice cultures using two models of seizures. In the kainic acid model for seizures we found no regulation of NRSF protein and mRNA levels. In a model of altered metabolism, using 2-deoxy-D-glucose we found

preliminary evidence of NRSF regulation as a result of inhibited glycolysis. Also, we report a difference in HCN1 channel isoform distribution in an in vitro and in vivo model.

(2)

1

Introduction

Epilepsy, seizures and NRSF

Epilepsy is the third most common brain disorder, affecting 1% of the US population (Yoon et al. 2009). Temporal lobe epilepsy (TLE) correlates highly with incidents such as brain trauma, brain infection and complex febrile seizures in humans. These complex febrile seizures can initiate epileptogenesis by altering gene expression in affected tissue. Hyperpolarization-activated cyclic nucleotide-gated channel 1 (HCN1) is important in neuronal excitability and is known to be downregulated in several models of TLE (Brewster et al. 2002; Richichi et al. 2008; Shah et al. 2004; Jung et al. 2007; Marcelin et al. 2009), reducing the threshold for excitation and increasing the likeliness of epileptic activity.

Neuron-Restrictive Silencer Factor (NRSF) a repressive transcription factor involved in downregulation of a number of genes that are repressed after seizures (S. McClelland et al. 2011; Ballas & Mandel 2005). Targets of NRSF include HCN1, but also the AMPA receptor subunit GluR2 (Calderone et al. 2003), growth factor BDNF and its receptor TrkB (Garriga-Canut et al. 2006). NRSF exerts its repressive function by binding to the Neuron Restrictive Silencer Element (NRSE) in the DNA. When NRSF and its cofactor(s) bind to the NRSE, the complex remodels the chromatin and maintains a repressive environment on the (promoter of) the gene. This repressive environment inhibits transcription of the gene downstream, resulting in downregulation of the gene (Calderone et al. 2003; Roopra et al. 2004).

Said changes alter the properties of neurons and make the network more susceptible to seizures, paving the road for the development of epilepsy. Previous research in an in vivo animal model of epilepsy has shown that preventing NRSF binding to the NRSE mitigates epileptogenesis and blocks the

downregulation of HCN1 (S. McClelland et al. 2011). This finding strengthens the hypothesis that NRSF plays a role in the regulation of HCN1-channel expression after KA induced seizures (S. P. McClelland 2011) and suggests a role for NRSF in epileptogenesis. The ultimate goal of our study is to identify the mechanisms that are crucial to the regulation of NRSF by seizures.

NRSF regulation

Although much is known about the downstream effects of NRSF, little is known about the regulation of NRSF expression itself (Kreisler et al. 2010). NRSF is known to play a role in development, repressing the neuronal phenotype in cells and is highly expressed in non-neuronal cells (Schoenherr & Anderson 1995). In differentiated neuronal populations the expression of NRSF is heavily reduced and expression is absent in glial cells (Palm et al. 1998).

(4)

3 Following seizures NRSF protein and mRNA levels have been shown to be upregulated in the

hippocampus in various models of epilepsy (Palm et al. 1998; Graef & Godwin 2010; Calderone et al. 2003; Hu et al. 2011) suggesting that NRSF is regulated as a response to seizures. This upregulation is visible at different time points: typically mRNA levels are increased as soon as 24 hours after seizures and protein levels are upregulated after 48 hours.

It remains unclear what the crucial mechanisms and pathways are that alter NRSF expression. Regulation could be driven by a plurality of signaling pathways activated during seizures, like the Ca2+/Zn2+-dependent signaling or glutamate receptor-dependent signaling known to influence HCN1 expression (Richichi et al. 2008). In a physiological sense seizures have a profound impact on the metabolic state of neurons. Altering the metabolic state of the brain after seizures, for example by inhibition of glycolysis, can mitigate epileptogenesis (Stafstrom et al. 2009). It is likely that NRSF regulation is dependent on the metabolic state of neurons (Qazi & Qureshi 2007; Roopra et al. 2004).

A suitable seizure model to study NRSF expression

In order to be able to distinguish between the many processes that might contribute to the regulation of NRSF after seizures a suitable model is required to study these contributions. Our model should allow for targeted manipulation of specific intra- and extracellular processes in the hippocampus, without the confounding effects of systemic responses or limitations to applying specific (pharmacological) agents. Since network functioning is crucial to study epilepsy, the model should salvage as much of the

functional connectivity of the hippocampus as possible. Finally, epileptogenesis is a time dependent process: the chosen model should provide a sufficient time frame to study the result of manipulations. The organotypic hippocampal slice culture model (Stoppini et al. 1991) is a reduced in vitro model that allows the experimenter to study epileptogenesis in a controlled manner. Organotypic cultures can be kept in culture for an extended period of time, offering a clear advantage over acute slices.

Furthermore, this model offers high control over the contents of the extracellular milieu and pharmacological agents can be introduced into the system by addition to the culture medium. In organotypic cultures network organization is kept relatively intact preserving the lamellar organization. Although the organotypic culture model offers a lot of advantages it is important to regard its

limitations. Systemic responses to seizures that might contribute in epileptogenesis are absent in this model. Also, even though the lamellar organization (Andersen et al. 2000) of the hippocampus is

(5)

4

Kainic acid seizure model

The kainic acid (KA) seizure model (Routbort et al. 1999) is reliable way to evoke seizures in vitro and in vivo. KA is a glutamate receptor agonist which triggers widespread glutamate receptor activation

directly upon application, this (reversibly) drives seizure-like activity followed by calcium influx (Routbort et al. 1999; Richichi et al. 2008; Zha et al. 2008). In vivo application of KA upregulates NRSF protein levels (Mcclelland et al. 2011; Hu et al. 2011) and NRSF mRNA levels have been shown to be downregulated in organotypic cultures (McClelland, Richichi et al., in press).

Altering metabolic state

Previous research has shown that altering the metabolic state of neurons can influence epileptogenesis. Application of 2-deoxy-D-glucose (2DG) to the bath medium is an established method to arrest glycolysis in in vitro models. 2DG is a glucose molecule in which a 2 hydroxyl-group is replaced by a single

hydrogen group. 2DG can enter the cell but cannot undergo further glycolysis (Lazzarini et al. 1988). When added to the medium 2DG blocks glycolysis, creating a metabolic state similar to that of the ketogenic diet, and downregulates NRSF targets such as BDNF and TrkB (Garriga-Canut et al. 2006). However, it remains unclear whether NRSF protein and mRNA levels are regulated by this metabolic state.

(6)

5

Methods

Organotypic hippocampal slice cultures

Shortly after dissection and slicing, coronal hippocampal slices (400 µm) from P8 Sprague Dawley rat pups (Harlan, CA, average weight=16 grams, 50% male/50% female) were positioned on moist mesh inserts and placed in culture medium (50% minimal essential medium, 25% Hank's balanced salt solution, 20% inactivated horse serum, 30 mM HEPES, 30 mM glucose, 3 mM glutamine, 0.5 mM ascorbic acid, 1 mg/ml insulin, 5 mM NaHCO3, pH 7.3, added Penicillin/Streptomycin). Cultures were kept in an incubator at 35 °C in a 5% CO2, 95 % O2 enriched atmosphere.

To control for anatomy care was taken to pair adjacent slices from each hippocampus over the control and experimental conditions. The use of these sister slices limited the variability in network organization and cellular buildup. All slices were taken from the same (middle) region of the hippocampus, resulting in a sum of 2 experimental pairs per animal (n=2).

The medium was refreshed the first day (DIV 1) to wash out the primary inflammatory response caused by culturing. Seizure-like events were induced after 4 days in vitro (DIV4) by incubating the cultures in medium containing 6 µM of kainic acid (KA) for three hours. Seizure-like activity was arrested by replacing the medium with regular culture medium (Fig. 1). Glycolytic inhibition was induced after 4 days in vitro by incubating the cultures in medium containing 20 µM of 2DG. Slices were harvested 48 hours after KA/2DG application for protein analysis and 24 hours after KA application for RT-qPCR.

Fig 1. Experimental protocol. KA induced seizure-like activity protocol in upper panel, 2DG metabolic state protocol in lower

(7)

6

Western Blot

Whole cell

Slices were snap frozen and immediately stored at – 80 °C prior to protein extraction. Samples were dissolved in 90 – 360 µL radioimmunoprecipitation assay (RIPA) buffer with a Protease Inhibitor Cocktail (PIC 50X Complete™; Roche, Almeda, CA) diluted according to the manufacturer's instructions. Slices were dounced by the use of a pestle and vortexed briefly to dissolve the tissue. Successively, the sample was sonicated for 5 minutes on a high setting (30sec on/30sec off) using a Bioruptor sonicator

(Diagenode, Sparta, NJ) and spun down at 16,000 rcf for 20 min. Nuclear enrichment

In order to separate the nuclear fraction of proteins slices were suspended in 250 µL hypotonic buffer (10mM KCl, 10mM Tris pH 8, 1.5mM MgCl2) with 5 µL Protein Inhibitor Cocktail (PIC, Roche) and vortexed before placing them on a rotator for 20 min, this caused the cells to swell due to osmotic stress. Next, 10% NP40 was added to the solution for 1 minute to lyse the cell membrane, keeping the nuclear envelope intact. The solution was spun down at 22,000 rcf for 10 min and the supernatant was aspirated (or kept for the membrane fraction), and the pellet dissolved in 30 µL of RIPA with 1 µL PIC for 10 min. The dissolved pellet was sonicated for 5 min and spun down at 23,000 rcf for 20 min.

Membrane proteins

The cytosolic/membrane fraction typically contains a lower concentration of protein. In order to reach sufficient protein levels for detection of HCN1, 6 slices were pooled per condition and dissolved in 600 µL sucrose solution with 10 µL PIC. The homogenate was spun down at 1000 rcf for 10 min to separate the nuclear debris and the membrane fraction. The supernatant was transferred to a new tube and spun down at 20,000 rcf for 25 min and the remaining pellet was dissolved in 50 µL RIPA and 1 µL PIC. For GluR2(3) detection 2 sliced were pooled and a whole cell extraction was performed as described above. Gel Loading, Running and Transfer

A Bradfort assay (Bio-Rad, Hercules, CA) was performed in order to determine the protein

concentrations of all samples. The concentration of all samples was normalized by dilution (10 µg total protein) and samples were suspended and heated (95 °C, 10 min) in Leamlli buffer prior to loading the samples on 4 – 12 % Tris-Glycene gels (Lonza, PAGEr, Rockland, ME). Electrophoresis separated protein bands were then transferred to PVDF membranes via wet transfer.

(8)

7 Protein detection

Prior to immunoreactive protein detection with antibodies membranes were blocked with 10% non-fat milk in PBS-T to reduce non-specific binding of the primary antibody. Primary antibody incubation occurred overnight at 4 °C on a mechanical shaker; concentrations used were 1:50,000 for NRSF (α-rabbit, Santa Cruz, H290X, 1:10000 – 1:5000), 1:11,000 for HCN1 (α-mouse, Neuromab, clone N70/28), 1:20,000 for RNA Polymerase II – subunit A mouse, Millipore, clone 8wg16), 1:100,000 for actin (α-rabbit, Sigma-Aldrich,) and 1:2000 GluR2(3) (rabbit Chemicon, Cat. #07-598). Secondary antibody (TruBlot® α-rabbit, eBioscience at 1:50,000) was applied to the bath for 30 min at RT on a mechanical shaker and washed off with PBS-T for 1.5 hours. Chemiluminescence detection was performed using the ECL Plus kit (Pierce®).

Optical Density

Optical density was determined within the linear range (0.2 – 0.8) using the ImageTool software set (UTHSCSA, TX). Values of the immunoreactive bands of the target protein (NRSF, HCN1 or GluR2) were normalized to the reference household protein actin.

Real Time quantitative Polymerase Chain Reaction

RNA isolation

Single hippocampal slices were harvested in 750 µL of Trizol (TRI Reagent, Sigma-Aldrich) and dissolved completely before use or storage at -80 °C. RNA isolation was performed using the DirectZol™ RNA Miniprep RNA R2050 purification kit (Zymo Research, USA) and DNA traces were removed by incubation with DNase I for 15 min at 37 °C. Total mRNA was eluted in 25 µL DNase and RNase free water. RNA concentration measurement was performed at 260/280 nm using a Nanodrop machine (Nanodrop 2000C, Thermo Scientific). Samples were diluted to ensure that total mRNA concentrations were similar across samples (typically 50 ng/ µL).

Amplification and detection

Real Time quantitative Polymerase Chain Reactions were performed on a Roche 480 Thermocycler® using the Roche Lightcycler® 480 RNADirect RNA amplification set (Roche Lightcycler® 480 RNA master hydrolysis probes), using 2 µL and 18 µL of premix for every reaction. Each reaction was run in triplicate and a negative control (RNA/DNA free H2O instead of sample) was included for every assay.

(9)

8 Cycle threshold determination

The cycle threshold in the RT-qPCR experiments was determined with the use of the Ct analysis integrated in the accompanying Lightcycler® 480 software set (Roche). The ∆∆Ct method (Livak & Schmittgen 2001) was used to calculate the differences in NRSF mRNA levels between conditions, comparing the target gene (NRSF) to the reference gene (GFAP) mRNA levels.

Primer assay selection

Primer assay selection for Nrsf (exon location 2-3) and Gapdh (exon location, 1-3) mRNA was performed using the IDT DNA assay design tool. The primer-probe combinations (PrimeTime™ Mini/Midi qPCR assays, IDT DNA) were optimized for detection in the Roche Lightcycler® system and amplicon length was restricted to a maximum of 90 base pairs. Primer assay validation was performed by running a dilution series for every assay and calculating the primer assay efficiency.

Immunocytochemistry – Fos staining

Right after the 3 hour treatment with KA, hippocampal cultures were fixed in freshly prepared paraformaldehyde solution (4 %) for 30 min at room temperature to ensure high structural integrity. After washing out the PFA with Phosphate Buffered Saline (PBS, 1% Triton X) the free floating cultures were quenched with H2O2 for 30 min to reduce background signal.

Cultures were blocked in 1% BSA for 30 minutes preceding primary antibody incubation (Fos α-rabbit, Calbiochem, C#PC38 (Ab-5)(4-17), 1:30,000) for three days at 4 °C. Secondary antibody incubation (biotynilated anti-rabbit IgG (H + L), BA-1000, Vector, Burlingame, CA, 1:600) was performed for 90 min at RT prior to diamino-bezidine (DAB) visualization of binding using the avadin/biotin system (Vector Vectastain ABC kit, (Vector, Burlingame, CA, PK-6100) – 1:300 A + B)).

Analysis

Values are given as mean (+/- the standard error of the mean) as compared to control. Unless otherwise indicated values were compared using Student’s t-test. P<0.05 was assumed to indicate a significant difference.

(10)

9

Experiments

Detection of NRSF protein

The first question we set out to answer was whether it was possible to detect NRSF protein in the extracts of single hippocampal slices. Single hippocampal slice protein extracts were run for WB alongside the extract of samples of (2) pooled slices. Chemiluminescent detection revealed that it is possible to detect NRSF protein levels extracted from a single slice (Fig. 2), opening up the possibility to subject multiple adjacent slices from a single hippocampus over various experimental conditions.

Fig 2. NRSF in whole cell protein extracts. Single slices extracts (first two from the left) revealed detectable levels of NRSF

similar to extracts from multiple slices (first two from the right).

The kainic acid model

NRSF protein levels detected in the whole cell extract did not differ between the KA and the control condition (Fig. 2). The differences between whole cell protein levels could have been too small to be detected by WB. Since NRSF is a transcription factor that predominantly functions in the cell nucleus, we hypothesized that a difference extracted the fraction of proteins would be detectable in the nuclear protein fraction. The success of the nuclear enrichment was confirmed by co-staining for RNA Polymerase II – Subunit A, a protein found only in the cell nucleus.

Nuclear NRSF could be reliably detected in the WB run on the nuclear enriched fraction but no difference in nuclear NRSF was found between the control and KA condition (Fig. 3) in our in vitro model. Slices harvested at different time points after KA treatment also did not show a difference in NRSF levels at any of the time points investigated (12, 24, 48 and 72 hours).

(11)

10 Fig 3. – NRSF expression in nuclear enriched protein fraction. Nuclear enrichment success was demonstrated by detection of

RNA Polymerase II – Sub A (at ~220 kDa).

Kainic Acid reliably drives seizure-like activity in the in vitro model

Expression of the early gene c-fos, as detected by immunoreactive staining of Fos, reveals the efficacy of KA in driving seizure-like activity in vivo (Labiner et al. 1993) and in vitro (Richichi et al. 2008). After three hours of treatment with KA (6 or 12 µM) cFos protein was abundant in the KA conditions and absent or only sparsely expressed in the control condition (Fig. 4). This affirms the assumption that the application of KA in the medium reliably drove neuronal activity in vitro.

The ionic composition of the medium used during KA treatment might be of influence to the severity and distribution of the seizure-like activity. In order to test this hypothesis KA induced seizure-like events were generated in slices under serum free and serum included conditions. We could not detect a difference in cFos expression due to serum addition. Additionally in neither the condition with and the one without serum, was there a significant difference in NRSF protein expression in the WB.

The functional anatomy influences the KA induced activation but not NRSF expression

When comparing the KA induced cFos activation pattern between posterior section, containing part of the entorhinal cortex (EC), and middle sections of the hippocampus, without EC, we found a distinct difference in the cFos expression pattern in the DG (fig. 4). As a result of KA-induced activation the DG in the middle sections of the hippocampus contained a high number of cFos expressing cells as opposed to the DG in the posterior sections which hardly contained any cFos expressing cells. However, regional differences in KA induced activation did not correlate with the expression of NRSF protein or mRNA, it was similar at both locations in the hippocampus.

(12)

11 Fig 4 cFos staining in hippocampal cultures. A)Fos staining in control condition B) KA induced activity in sections from the

middle of the hippocampus C) the posterior hippocampus (lower panel). Activity related cFos expression was absent in control and highly abundant in KA treated cultures. Note the difference in DG cFos expression in the posterior and middle sections.

(13)

12

Detection of NRSF mRNA

Since the generation of NRSF protein is dependent on the translation and transcription of the NRSF gene one would expect to detect changes in expression in messenger RNA levels following KA induced

seizure-like events. In order to study the transcriptional activity of NRSF we used quantitative Real Time Polymerase Chain Reaction (RT-qPCR) to determine mRNA differences between conditions. In our experiments we successfully detected NRSF mRNA (average cycle threshold Ct=24) as well as GAPDH (cycle threshold Ct=19) with high accuracy (standard deviation of Ct < 0.1). However, we found no differences in the NRSF mRNA levels 24 hours after KA induced seizure-like activity.

NRSF target gene regulation in the KA model

KA induced seizure-like activity is known to regulate HCN1 channels and GluR(3)2 receptor subunits in organotypic cultures (Richichi et al. 2008). In an attempt to replicate these findings the same

experimental setup was used as described before (Fig. 1). In this experiment WB was run with

antibodies for HCN1 and GluR2(3). The application of KA did not induce detectable changes in the levels of GluR2(3) subunit nor HCN1 channel expression; protein levels per se crossed the detection threshold.

Metabolic state – The 2DG model

NRSF protein levels and glycolysis

Next, we set out to determine whether NRSF protein levels could be altered by changing the metabolic state. In this experiment we studied the effect of inhibiting glycolysis on NRSF protein levels. Glycolytic inhibition was performed by adding 2DG (Sigma-Aldrich, St. Louis, MO) to the medium at a

concentration of 10 µM (Stafstrom et al. 2008) for two days before extraction of the nuclear protein fraction (for experimental protocol see Fig 1.). Nuclear NRSF protein levels were lower after addition of 2DG to the medium but this difference did not reach significance (p=0.22).

(14)

13

HCN1 subunit composition differences between the in vitro and in vivo model

In our experiments we observed differences in the molecular weights of HCN1 isoforms between the in vivo and the in vitro models. In order to reach the detection threshold for HCN1 in the in vitro model the membrane protein extract of a 6 hippocampal slices (from 3 animals) were pooled. The in vivo sample consisted of the extract of a single (whole) hippocampus from a comparable developmental state as our in vitro model (P12 rat in vivo, as compared to P8 + DIV4 in vitro).

Comparing the optical density of the HCN1 fractions revealed that the upper HCN1 band at ~125 kDa had a lower intensity in the in vitro model than the lower band at ~105 kDa. In the in vivo samples the distribution of HCN1 subunits was mirrored, with a stronger upper and weaker lower band. This effect was independent from both KA and 2DG.

Fig 6. Membrane protein fractions in in vivo and in vitro. The in vitro HCN1 fraction (left) with a stronger lower band at ~105

(15)

14

Discussion

In our experiments we used two different in vitro models of epilepsy/seizures to study the regulation of the transcription factor NRSF. Although it was possible to detect NRSF mRNA and protein reliably in our experiments, no regulation of NRSF was detected in either the kainic acid or the 2-deoxy-D-glucose treatment conditions. However, a number of important technical improvements and important observations have been made. These findings, and their respective considerations will be discussed below.

NRSF protein and mRNA detection

In this study we have shown that it is possible to detect NRSF in the nuclear enriched protein fraction and whole cell mRNA isolated fraction of single slices from the rat hippocampus (P8 + DIV6). These are novel findings that provide the possibility to study NRSF regulation in an in vitro system. This model provides high control over the parameters of the system and allows the experimenter to isolate specific variables. In this approach adjacent slices taken from the same hippocampus can be compared; forming anatomically optimally matched pairs.

Kainic acid model

Seizure-like activity was confirmed by ICC detection of the neuronal activity marker cFos after 3 hours of application of KA. This suggests that KA successfully drives widespread neuronal activity comparable to seizure-like events. The comparison between the sections from the middle and posterior hippocampus (including part of the EC) shows that the functional connectivity of the hippocampus is preserved in the in vitro model used. The gating function of the dentate gyrus (DG) (Heinemann et al, 1992, Lothmann et al., 1992), is preserved in the organotypic cultures, filtering the spread of activation from the entorhinal cortex (EC) to the hippocampus (Iijima et al. 1996; Coulter & Carlson 2007) by inhibition.

NRSF protein regulation was absent in our KA model at the time points studied, in the whole cell and nuclear fraction. Also, NRSF mRNA detection performed with RT-qPCR revealed no significant

differences due to KA application. These findings suggest that, contrary to earlier findings, NRSF levels themselves might not be regulated after KA induced seizure-like activity in the in vitro model used. Accordingly, the expression levels of HCN1 and GluR2(3) were unaltered by KA induced seizure-like events. When the membrane and nuclear fraction of the same sample were run side by side, no regulation in both NRSF and HCN1 was detected. These finding contrasts with earlier findings in the in vitro KA model of seizure (Richichi et al. 2008). However, the fact that neither NRSF mRNA or protein

(16)

15 and henceforth its downstream targets was regulated does not falsify the central thesis that NRSF is a transcription factor that regulates expression of GluR2(3) and HCN1.

One restriction of the approaches used is that they do not account for cell type/structure specific expression differences in NRSF levels. Due to the fact that protein and mRNA extracts are isolated from an entire slice, altered expression in one structure (e.g. CA1 vs. DG) might not be high enough to cross the detection threshold. Previously reported regulation of NRSF mRNA was restricted to the CA1-3 regions and absent in other domains within the culture (Palm et al. 1998, McClelland and Richichi, in press). Also contrary regulated NRSF protein and mRNA expression in different regions within the section could cancel each other out.

The given data does not provide information on the activity of NRSF itself in performing its role as a transcription factor. An in vivo chromatin immunoprecipitation (ChIP) study has shown that the binding of NRSF to the NRSE is dependent on KA induced seizure-like activity (S. McClelland et al. 2011). Future research could focus on the NRSF-NRSE binding in the in vitro KA model as a measure of NRSF activity rather than the abundance of the protein/mRNA.

Metabolic state – the 2DG model

To our knowledge this is the first report of glycolysis as a possible regulator of NRSF protein levels in the in vitro hippocampal cultures model. The finding that an altered metabolic, state rather than generalized KA induced seizure-like activity, influences NRSF expression suggests that the metabolic properties of seizures play an important role in NRSF regulation. This effect should be further verified in a follow up study but suggests a role of metabolism in the regulation of the transcription factor NRSF. A power analysis revealed1 that a total of 11 experiments (pairs of hippocampal slices) would be needed to sufficiently probe the difference between the control group and the treatment group.

The anti-epileptogenic effects of 2-Deoxy-D-glucose have been studied previously, revealing that 2DG inhibits epileptogenesis by increasing the afterdischarge threshold for seizures (Garriga-Canut et al. 2006; Stafstrom et al. 2009). Whether this adaptation in electrophysiological properties is dependent on changes in ion channel or receptor expression remains unclear. Future research could focus on the relationship between 2DG application, NRSF expression and membrane build up.

1

A priori power analysis performed with G-power 3.1.4 (Faul et al. 2007) based on data from the described experiment (with calculated effect size of 0.472).

(17)

16

HCN1 glycosylation

In our experiments we observed a difference in HCN1 subunit composition between the in vitro and in vivo models. HCN1 channels glycosylation is known to alter the weight of HCN1 isoforms as a result of differential subunit heteromerization of the channel (Proenza et al. 2002). These adaptations can result in a different molecular weight of the HCN-channels that can be visualized by electrophoresis.

The post-translational differences in the molecular weight of HCN1, between the in vitro and in vivo samples of hippocampal slices of similar ‘age’, might be the result of different metabolic states in the in vivo and in vitro situation. Glycosylation of HCN1 channels has been shown to be activity and glycolysis dependent (Zha et al. 2008; Proenza et al. 2002) and could influence the molecular buildup of HCN1 channels. Since this effect was robust and independent of KA or 2DG application it is most likely that these changes are inherent properties of the in vitro model.

However, the different samples in both conditions (in vivo/in vitro) cannot be compared directly due to the considerable differences in origin and composition of the samples. A way to create a more valid comparison would be to pool single hippocampal slices from different animals in the in vivo model, that have a similar anatomy to the slices from the in vitro model, and compare these to the in vitro model.

(18)

17

Refrences

Andersen, P., Soleng, a F. & Raastad, M., 2000. The hippocampal lamella hypothesis revisited. Brain

research, 886(1-2), pp.165-171.

Ballas, N. & Mandel, G., 2005. The many faces of REST oversee epigenetic programming of neuronal genes. Current opinion in neurobiology, 15(5), pp.500-6.

Brewster, A. et al., 2002. Developmental Febrile Seizures Modulate Hippocampal Gene Expression of Hyperpolarization-Activated Channels in an Isoform- and Cell-Specific Manner. The Journal of

Neuroscience, 22(11), pp.4591-4599.

Calderone, A. et al., 2003. Ischemic insults derepress the gene silencer REST in neurons destined to die.

The Journal of neuroscience : the official journal of the Society for Neuroscience, 23(6), pp.2112-21.

Coulter, D. a & Carlson, G.C., 2007. Functional regulation of the dentate gyrus by GABA-mediated inhibition. Progress in brain research, 163(07), pp.235-43.

Faul, F. et al., 2007. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior research methods, 39(2), pp.175-91.

Garriga-Canut, M. et al., 2006. 2-Deoxy-D-glucose reduces epilepsy progression by NRSF-CtBP-dependent metabolic regulation of chromatin structure. Nature neuroscience, 9(11), pp.1382-7. Graef, J.D. & Godwin, D.W., 2010. Intrinsic plasticity in acquired epilepsy: too much of a good thing? The

Neuroscientist : a review journal bringing neurobiology, neurology and psychiatry, 16(5),

pp.487-95.

Hu, X.-L. et al., 2011. Conditional Deletion of NRSF in Forebrain Neurons Accelerates Epileptogenesis in the Kindling Model. Cerebral cortex (New York, N.Y. : 1991), 21(9), pp.2158-65.

Iijima, T. et al., 1996. Entorhinal-Hippocampal Interactions Revealed by Real-Time Imaging. Science, 272(5265), pp.1176-1179.

Jung, S. et al., 2007. Progressive dendritic HCN channelopathy during epileptogenesis in the rat pilocarpine model of epilepsy. The Journal of neuroscience : the official journal of the Society for

Neuroscience, 27(47), pp.13012-21.

Kreisler, a et al., 2010. Regulation of the NRSF/REST gene by methylation and CREB affects the cellular phenotype of small-cell lung cancer. Oncogene, 29(43), pp.5828-38.

Labiner, D.M. et al., 1993. Induction of c-fos mRNA by Kindled Seizures : with Neuronal Burst Firing. The

Jounal of Neuroscience, 73(February), pp.744-751.

Lazzarini, S.J., Schneider, J.E. & Wade, G.N., 1988. Inhibition of fatty acid oxidation and glucose metabolism does not affect food intake or hunger motivation in Syrian hamsters. Physiology &

behavior, 44(2), pp.209-13.

Livak, K.J. & Schmittgen, T.D., 2001. Analysis of relative gene expression data using real-time

quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods (San Diego, Calif.), 25(4), pp.402-8.

(19)

18 Marcelin, B. et al., 2009. H Channel-Dependent Deficit of Theta Oscillation Resonance and Phase Shift in

Temporal Lobe Epilepsy. Neurobiology of disease, 33(3), pp.436-47.

McClelland, S. et al., 2011. Neuron-restrictive silencer factor-mediated cyclic nucleotide gated

channelopathy in experimental temporal lobe epilepsy. Annals of Neurology, 00000, p.n/a-n/a. McClelland, S.P., 2011. Neuron Restrictive Silencer Factor-mediated epigenetic processes in

neuroplasticity. Dissertation.

Mcclelland, S. et al., 2011. Epileptogenesis after prolonged febrile seizures: Mechanisms, biomarkers and therapeutic opportunities. Neuroscience Letters.

Palm, K. et al., 1998. Neuronal expression of zinc finger transcription factor REST/NRSF/XBR gene. The

Journal of neuroscience : the official journal of the Society for Neuroscience, 18(4), pp.1280-96.

Proenza, C. et al., 2002. Different roles for the cyclic nucleotide binding domain and amino terminus in assembly and expression of hyperpolarization-activated, cyclic nucleotide-gated channels. The

Journal of biological chemistry, 277(33), pp.29634-42.

Qazi, R. & Qureshi, S.A., 2007. The ketogenic diet in epilepsy - Molecular basis of anti-seizure efficacy.

Issues in basic neuroscience, 2(3), pp.171-175.

Richichi, C. et al., 2008. Mechanisms of seizure-induced “transcriptional channelopathy” of

hyperpolarization-activated cyclic nucleotide gated (HCN) channels. Neurobiology of disease, 29(2), pp.297-305.

Roopra, A. et al., 2004. Localized domains of G9a-mediated histone methylation are required for silencing of neuronal genes. Molecular cell, 14(6), pp.727-38.

Routbort, M.J., Bausch, S.B. & McNamara, J.O., 1999. Seizures, cell death, and mossy fiber sprouting in kainic acid-treated organotypic hippocampal cultures. Neuroscience, 94(3), pp.755-65.

Schoenherr, C.J. & Anderson, D.J., 1995. The neuron-restrictive silencer factor (NRSF): a coordinate repressor of multiple neuron-specific genes. Science (New York, N.Y.), 267(5202), pp.1360-3. Shah, M.M. et al., 2004. Seizure-induced plasticity of h channels in entorhinal cortical layer III pyramidal

neurons. Neuron, 44(3), pp.495-508.

Stafstrom, C.E. et al., 2009. Anticonvulsant and antiepileptic actions of 2-deoxy-Dglucose in epilepsy models. Annals of neurology, 65(4), pp.435-447.

Stafstrom, C.E., Roopra, A. & Sutula, T.P., 2008. Seizure suppression via glycolysis inhibition with 2-deoxy-D-glucose (2DG). Epilepsia, 49 Suppl 8, pp.97-100.

Stoppini, L., Buchs, P. a & Muller, D., 1991. A simple method for organotypic cultures of nervous tissue.

Journal of neuroscience methods, 37(2), pp.173-82.

Yoon, D. et al., 2009. Economic impact of epilepsy in the United States. Epilepsia, 50(10), pp.2186-91. Zha, Q. et al., 2008. Activity-dependent heteromerization of the hyperpolarization-activated,

cyclic-nucleotide gated (HCN) channels: role of N-linked glycosylation. Journal of neurochemistry, 105(1), pp.68-77.

NRSF regulation in in vitro models of seizures