Antiseizure drugs differentially modulate theta-burst induced long-term potentiation in C57BL/6 mice

Authors

  • Peter J. West,

    Corresponding author
    1. Department of Pharmacology and Toxicology, University of Utah, Salt Lake City, Utah, U.S.A
    2. Anticonvulsant Drug Development Program, University of Utah, Salt Lake City, Utah, U.S.A
    3. Interdepartmental Neuroscience Program, University of Utah, Salt Lake City, Utah, U.S.A
    • Address correspondence to Peter J. West, Department of Pharmacology and Toxicology, Anticonvulsant Drug Development Program, University of Utah, 417 Wakara Way, Suite 3211, Salt Lake City, UT 84108, U.S.A. E-mail: peter.west@utah.edu

    Search for more papers by this author
  • Gerald W. Saunders,

    1. Anticonvulsant Drug Development Program, University of Utah, Salt Lake City, Utah, U.S.A
    Search for more papers by this author
  • Gregory J. Remigio,

    1. Department of Pharmacology and Toxicology, University of Utah, Salt Lake City, Utah, U.S.A
    2. Interdepartmental Neuroscience Program, University of Utah, Salt Lake City, Utah, U.S.A
    Search for more papers by this author
  • Karen S. Wilcox,

    1. Department of Pharmacology and Toxicology, University of Utah, Salt Lake City, Utah, U.S.A
    2. Anticonvulsant Drug Development Program, University of Utah, Salt Lake City, Utah, U.S.A
    3. Interdepartmental Neuroscience Program, University of Utah, Salt Lake City, Utah, U.S.A
    Search for more papers by this author
  • H. Steve White

    1. Department of Pharmacology and Toxicology, University of Utah, Salt Lake City, Utah, U.S.A
    2. Anticonvulsant Drug Development Program, University of Utah, Salt Lake City, Utah, U.S.A
    3. Interdepartmental Neuroscience Program, University of Utah, Salt Lake City, Utah, U.S.A
    Search for more papers by this author

Summary

Objective

Cognitive comorbidities are increasingly recognized as an equal (or even more disabling) aspect of epilepsy. In addition, the actions of some antiseizure drugs (ASDs) can impact learning and memory. Accordingly, the National Institute of Neurological Disorders and Stroke (NINDS) epilepsy research benchmarks call for the implementation of standardized protocols for screening ASDs for their amelioration or exacerbation of cognitive comorbidities. Long-term potentiation (LTP) is a widely used model for investigating synaptic plasticity and its relationship to learning and memory. Although the effects of some ASDs on LTP have been examined, none of these studies employed physiologically relevant induction stimuli such as theta-burst stimulation (TBS). To systematically evaluate the effects of multiple ASDs in the same preparation using physiologically relevant stimulation protocols, we examined the effects of a broad panel of existing ASDs on TBS-induced LTP in area CA1 of in vitro brain slices, prepared in either normal or sucrose-based artificial cerebrospinal fluid (ACSF), from C57BL/6 mice.

Methods

Coronal brain slices containing the dorsal hippocampus were made using either standard or sucrose-based ACSF. Recordings were obtained from four slices at a time using the Scientifica Slicemaster high throughput recording system. Slices exposed to ASDs were paired with slices from the opposite hemisphere that served as controls. Field excitatory postsynaptic potentials (fEPSPs) were recorded, and all ASDs were applied to slices by bath perfusion for 20 min prior to the induction stimulus. LTP was induced by TBS or by high-frequency stimulation (HFS). The following ASDs were examined: 100 μM phenobarbital (PB), 80 μM phenytoin (PHT), 50 μM carbamazepine (CBZ), 600 μM valproate (VPA), 60 μM topiramate (TPM), 60 μM lamotrigine (LTG), 100 μM levetiracetam (LEV), 10 μM ezogabine (EZG), and 30 μM tiagabine (TGB).

Results

Among voltage-gated sodium channel inhibitors, CBZ significantly attenuated TBS-induced LTP, PHT attenuated both TBS-induced LTP and post–tetanic potentiation (PTP), and LTG failed to affect LTP but did attenuate PTP. ASDs that modulate γ-aminobutyric acid (GABA)ergic synaptic transmission, such as PB and TGB, significantly attenuated LTP in brain slices prepared in sucrose-based ACSF but not standard ACSF. Third generation ASDs, such as LEV and TPM, did not affect LTP in ACSF- or sucrose-prepared brain slices. Although EZG failed to affect LTP, it did significantly attenuate PTP under both slicing conditions. VPA failed to affect LTP in area CA1, both in C57BL/6 mice and Sprague-Dawley rats, using TBS or HFS. However, VPA did attenuate TBS-induced LTP in the dentate gyrus (DG).

Significance

The results of experiments describe herein provide a comprehensive summary of the effects of many commonly used ASDs on short- and long-term synaptic plasticity while, for the first time, using physiologically relevant LTP induction protocols and slice preparations from mice. Furthermore, methodologic variables, such as brain slice preparation protocols, were explored. These results provide comparative knowledge of ASD effects on synaptic plasticity in the mouse hippocampus and may ultimately contribute to an understanding of the differences in the cognitive side effect profiles of ASDs and the prediction of cognitive dysfunction associated with novel investigational ASDs.

image

Dr. Peter J. West is a research assistant professor in the ADD Program at the University of Utah.

Patients with temporal lobe epilepsy (TLE) often experience cognitive comorbidities that are increasingly recognized as an equal (or even more disabling) part of epilepsy.[1] These cognitive symptoms do not universally disappear once seizures are well controlled, and in some cases can progressively worsen, which leads to a poor quality of life for patients and their families.[2] In addition, antiseizure drugs (ASDs) can be a source of cognitive disturbances; some first- and second-generation ASDs can impact learning and memory in both experimental animals and humans.[3] Accordingly, the National Institute of Neurological Disorders and Stroke (NINDS) Epilepsy Research Benchmarks recognizes this as a significant challenge and calls for increased efforts aimed at the development and implementation of standardized protocols for screening pharmacologic treatments for epilepsy for their amelioration or exacerbation of cognitive comorbidities.[4]

Long-term potentiation (LTP) is a form of activity dependent synaptic plasticity that can be experimentally observed at a large number of synapses, both in vivo and in vitro.[5] Because this plasticity is generally agreed to be necessary for information storage, LTP has gained acceptance as a widely used model of learning and memory at the level of the synapse.[6] The usefulness of this model is exemplified by its utility in cognitive drug discovery and better understanding the pathologic impairments in cognition associated with diseases such as Alzheimer's disease and epilepsy.[7]

ASDs generally exert their effects by suppressing neuronal excitation or enhancing inhibition.[8] These mechanisms, in addition to the known effects of ASDs on cognition in humans and experimental animals, suggest that ASDs may suppress LTP and other forms of synaptic plasticity. Indeed, this has been shown for several ASDs.[9-14] However, these earlier studies have inconsistent protocols that limit direct comparison between ASDs. Furthermore, all prior studies used high-frequency stimulation (HFS) as the LTP induction stimulation. HFS is considered to have less physiologic relevance than other stimulation protocols such as theta-burst stimulation (TBS).[15] Notably, the influences of treatments that modulate γ-aminobutyric acid (GABA)ergic neurotransmission are undetected when HFS is used but readily apparent with TBS.[16] Given the prominent role of GABA in the mechanism of action of several of the clinically available ASDs,[8] TBS may be the preferred method for LTP induction when examining the effects of ASDs.

The aim of the present study was to examine the effects of ASDs on TBS-induced LTP (TBS-LTP) in area CA1 of brain slices containing the C57BL/6 mouse hippocampus. To our knowledge, this study is the first to examine the acute effects of these commonly used ASDs on LTP in mice while using TBS. In addition, we examined how methodologic differences in brain slice preparation affect the modulatory action of ASDs on LTP. Finally, in the case of valproic acid, where our initial results were inconsistent with several LTP studies reported by others[10, 12, 14, 17] and its known cognitive effects in vivo,[18] we examined how a number of factors including induction stimulation, rodent model, and hippocampal subregion affect its modulatory action on synaptic plasticity. These comprehensive data are expected to serve as the foundation upon which novel investigational ASDs may be screened for their potential to impair cognitive function.

Methods

Hippocampal brain slice preparation

All experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the University of Utah's Animal Care and Use Committee. All efforts were made to minimize the number and suffering of animals used. Male C57BL/6 mice (20–25 g; Charles River, Raleigh, NC, U.S.A.) or Sprague-Dawley Rats (100–150 g) were anesthetized with sodium pentobarbital (60 mg/kg, i.p.), and brains were rapidly removed and placed in ice-cold (4°C) oxygenated artificial cerebral spinal fluid (ACSF) solution (95% O2/5% CO2) containing (in mM): NaCl (126.0), KCl (3.0), Na2PO4 (1.4), MgSO4 (1.0), NaHCO3 (26.0), glucose (10.0), and CaCl2 (2.5). The pH (7.30–7.40) and osmolarity (290–300 mOsm) of the ACSF were verified prior to each experiment. Where indicated, this ACSF was substituted with one where NaCl was replaced by sucrose (200 mM), MgSO4 concentration was raised to 3 mm, and CaCl2 concentration was lowered to 0.5 mm (indicated as sucrose-based ACSF). Brains were then submerged in oxygenated ACSF or sucrose-based ACSF maintained at 4°C where coronal brain slices (350 μm) containing the dorsal hippocampus were cut. Slices were then transferred to an incubation chamber containing oxygenated ACSF at room temperature for 2 h prior to recording.

Long-term potentiation of excitatory synaptic transmission

Extracellular field excitatory postsynaptic potentials (fEPSPs) were recorded using a Slicemaster high-throughput brain slice recording system (Scientifica, Uckfield, East Sussex, United Kingdom). Slices were continuously perfused with oxygenated ACSF (2.5 ml/min). Recordings were performed at 30–31°C. fEPSPs from four independent brain slices were recorded simultaneously, thereby allowing two slices to be tested for ASD effects while the other two slices served as matched controls obtained from the same experimental animal.

Concentric bipolar stimulating electrodes (MCE-100; Rhodes Medical Instrument, Summerland, CA, U.S.A.) were placed in either the stratum radiatum of CA1 or the inner molecular layer of the DG. Recording microelectrodes were filled with ACSF and placed within 250–500 μm of the stimulating electrodes. Placement of stimulating and recording electrodes in the inner molecular layer of the DG was confirmed by the verification of paired-pulse (50 msec interpulse interval) depression. Picrotoxin (10 μm) was present at all times when LTP was measured in the DG.

Data were acquired using pClamp 10 interfaced to a Digidata 1440A data acquisition board (Molecular Devices, Sunnyvale, CA, U.S.A.) at a sampling rate of 10 kHz, low-pass filtered at 1 kHz, and high-pass filtered at 3 Hz. One hundred microsecond stimuli ranging from 1 to 40 V were used to evoke fEPSPs, and the magnitude of the fEPSP was determined by measuring the 20–80% slope of the rising phase.

Input-output curves were generated and the stimulation strength was set to 50% of the range between the minimum and maximum fEPSP. Slices were then stimulated every 30 s for a 30 min baseline period. Drugs were applied by bath exchange for 20 min prior to the LTP induction stimulus, and changes in baseline synaptic transmission were quantified. LTP, in either area CA1 or the DG, was induced using TBS (four trains of four pulses at 100 Hz separated by 200 msec and repeated once with a 20 s interval), or by HFS (1 s/100 Hz), where indicated. Low frequency stimulation was resumed for 60 min at which point LTP was quantified relative to baseline. In addition, post–tetanic potentiation (PTP) time constants (tau) were measured; these were represented as best-fit values (95% confidence interval [CI]) from one-phase exponential decay regression functions. In some cases, additional experiments, where LTP was not induced, were performed to evaluate the effects of ASDs on basal synaptic transmission.

All representative traces and measurements are the average of data from five consecutive traces taken at the end of the indicated condition. LTP (% control) was quantified by normalizing the % change over baseline in ASD-treated slices to the average of the same measurement in control slices generated from the same animals. All data are presented as mean ± standard error of the mean (SEM), and all N values are presented as number of slices/number of animals. Unless otherwise noted, all statistical comparisons of LTP were made using the unpaired two-tailed Student's t-test, and significance was determined at p < 0.05. PTP was compared using an extra sum-of-squares F test (†p < 0.05, ‡p < 0.0001) to compare fitted rate constant values for one-phase exponential decay curves. Data were excluded if the slope of fEPSPs during the 30 min baseline changed by >20%.

Chemicals and antiseizure drugs

Retigabine dihydrochloride (ezogabine) was obtained from Santa Cruz Biotechnology (Dallas, TX, U.S.A.). Tiagabine was kindly provided by Cephalon, Inc. (Frazer, PA, U.S.A.). Topiramate was kindly provided by Johnson and Johnson Pharmaceutical Research and Development (Springhouse, PA, U.S.A.). All other chemicals and ASDs were obtained from Sigma (St. Louis, MO, U.S.A.). Unless otherwise noted, the ASD concentrations tested in these studies were chosen from the high end of their therapeutic plasma concentrations.[19, 20]

Results

In control brain slices, TBS induced both a short-term PTP and LTP. The PTP time constant (tau) was determined to be 14.0 min (95% CI 13.1–15.1, N = 208 slices/77 mice), and LTP was 142.3 ± 2.2% of the baseline (N = 208 slices/77 mice). Brain slices treated with ASDs were statistically compared only to matched control slices from the same experimental animals. Using a one-way ANOVA to compare across 14 independent control groups, LTP was not significantly different (F13,193 = 0.6435, p = 0.82), indicating that control LTP was highly reproducible under these experimental conditions. A summary of the effects of ASDs on TBS-induced PTP and LTP in area CA1 is presented in Table 1.

Table 1. Summary of ASD effects on baseline fEPSPs, TBS-induced PTP, and TBS-induced LTP in area CA1
ASD (μM)Plasma conc.a (μM)NbASD baselinec(% control)PTP time constantd (min)LTPe (% baseline)ASD LTPf (% control LTP)
ControlASDControlASD
  1. Experiments where brain slices were prepared using sucrose-based ACSF are indicted.

  2. a

    ASD concentrations tested in these studies were chosen from the high end of their therapeutic plasma concentrations (“Plasma Conc.”; Bialer et al.,[19], Large et al.,[20]) with the exception of TGB.

  3. b

    “N” values represent (no. control slices/no. ASD treated slices/no. mice).

  4. c

    ASD effects on baseline fEPSPs (“ASD Baseline”) are represented as a % change (±SEM) after a 20 min ASD exposure as compared to the control fEPSP. A paired two-tailed Student's t-test (*p < 0.05) was used to compare fEPSP slope before and after ASD exposure.

  5. d

    PTP time constants (tau) represent best-fit values (95% CI) from one-phase exponential decay regression functions. Statistics were performed using an extra sum-of-squares F test (p < 0.05, p < 0.0001) to compare fitted rate constant values for one-phase exponential decay curves.

  6. e

    LTP values (60 min after TBS) represent mean ± SEM.

  7. f

    An unpaired two-tailed Student's t-test (*p < 0.05, **p < 0.01) was used to compare LTP values between ASD-treated brain slices and control brain slices obtained from the same mice.

Phenytoin (PHT)        
8040–7914/11/4−6.1 ± 3.714.3 (10.5, 22.2)4.3 (3.5, 5.6)146.6 ± 6.6125.0 ± 7.853.7 ± 16.8 *
Carbamazepine (CBZ)        
5013–5112/13/43.5 ± 3.114.3 (11.2, 19.9)14.9 (11.4, 21.6)143.0 ± 5.4119.4 ± 4.645.0 ± 10.8 **
Lamotrigine (LTG)        
6010–5922/21/7−3.3 ± 1.619.8 (14.9, 29.9)11.2 (9.8, 13.1)145.6 ± 8.1140.5 ± 4.888.9 ± 10.5
Phenobarbital (PB)        
10043–10814/18/6−2.6 ± 2.712.9 (10.9, 15.8)13.1 (10.5, 17.4)135.8 ± 6.5139.8 ± 9.5110.9 ± 26.4
100: sucrose 15/17/9−3.3 ± 1.710.9 (8.6, 14.7)7.5 (6.2, 9.3)142.7 ± 10.0115.2 ± 5.435.5 ± 12.7 *
Tiagabine (TGB)        
300.05–0.2729/24/9−0.5 ± 1.717.6 (15.1, 21.6)13.8 (11.3, 17.6)144.1 ± 6.7144.7 ± 5.3101.3 ± 12.0
30: sucrose 18/19/6−2.2 ± 2.911.8 (9.9, 14.6)13.4 (11.8, 15.5)131.3 ± 7.0106.2 ± 4.219.6 ± 13.5 **
Ezogabine (EZG)        
100.4–210/15/4−8.2 ± 2.4 *12.3 (9.3, 18.4)3.4 (2.5, 5.2)152.7 ± 11.3146.4 ± 5.988.0 ± 11.2
10: sucrose 14/12/4−16.8 ± 2.4 *20.4 (15.5, 29.6)3.6 (2.8, 4.9)152.4 ± 7.9144.9 ± 5.485.7 ± 10.2
Topiramate (TPM)        
6015–6015/14/5−2.9 ± 1.811.6 (9.6, 14.8)18.1 (11.4, 43.5)138.2 ± 8.5141.7 ± 13.9109.1 ± 36.4
60: sucrose 13/14/40.3 ± 2.012.0 (8.7, 19.4)7.8 (6.3, 10.1)159.5 ± 15.1163.7 ± 9.6107.0 ± 16.1
Levetiracetam (LEV)        
10035–11813/13/61.1 ± 1.415.3 (11.4, 23.5)10.6 (8.4, 14.6)155.4 ± 7.0145.2 ± 8.681.7 ± 15.5
100: sucrose 12/11/61.9 ± 2.79.2 (7.4, 12.2)16.5 (13.6, 20.8)139.2 ± 8.9141.2 ± 6.1105.0 ± 15.6
Valproic acid (VPA)        
600277–69313/12/4−3.8 ± 2.114.1 (11.3, 18,7)14.0 (11.6, 17.8)131.1 ± 6.7131.6 ± 6.4101.5 ± 20.5
600: sucrose 11/11/53.7 ± 2.614.8 (12.1, 18.9)7.4 (5.8, 10.3)133.0 ± 6.3146.2 ± 10.8139.7 ± 32.6
1,000 16/16/7−2.3 ± 2.310.9 (9.1, 13.5)10.7 (9.2, 12.9)136.8 ± 7.3134.7 ± 5.794.5 ± 15.5

Sodium channel blocking ASDs differentially affect TBS-LTP

As illustrated in Figure 1, the effects of 80 μm PHT (Fig. 1A), 50 μm CBZ (Fig. 1B), and 60 μm LTG (Fig. 1C) on baseline synaptic transmission, PTP, and TBS-LTP were compared. PHT, CBZ, and LTG did not significantly affect baseline synaptic transmission; PHT's 6.1 ± 3.7% attenuation of the fEPSP was not significant (p = 0.150, two-tailed paired student t-test). Both PHT and LTG significantly reduced the PTP decay time constant compared to matched controls. However, CBZ was without a PTP effect. Both PHT and CBZ significantly attenuated TBS-LTP, whereas LTG had no significant effect on TBS-LTP in area CA1.

Figure 1.

Sodium channel blocking ASDs differentially attenuate TBS-LTP. All traces are the average of five sweeps taken before TBS (gray) or 60 min after TBS (black). Scale bars representing 0.25 mV and 2 msec apply to both control and ASD traces. In all summary time course figures, horizontal red lines from −20 to 0 min represent the duration of drug perfusion in ASD-treated slices (closed red circles) or ACSF perfusion in control slices (open circles). All error bars represent SEM. The following summary time courses are illustrated: (A) 80 μm PHT, (B) 50 μm CBZ, and (C) 60 μm LTG. *p < 0.05, **p < 0.01.

GABA-enhancing ASDs attenuate TBS-LTP in brain slices prepared in sucrose-based ACSF

It has been suggested that inhibitory synaptic transmission is compromised in brain slices prepared using ACSF-based cutting solutions; conversely, preparing brain slices in a sucrose-based ACSF better preserves inhibitory interneurons.[21] The consequence of this preservation of inhibition is greatly attenuated LTP.[22] To assess this under our experimental conditions, we evaluated the impact of sucrose-based ACSF on the magnitude of LTP in brain slices prepared from mice. TBS-LTP in sucrose-based ACSF prepared slices (141.1 ± 3.3%, N = 96 slices/40 mice) was not significantly different from that in ACSF prepared slices. Using a one-way ANOVA to compare across seven independent experiments, LTP in matched-control slice groups prepared in sucrose-based ACSF was not significantly different (F6,90 = 1.280, p = 0.27). This indicates that control LTP was highly reproducible under these experimental conditions and, unlike previous reports,[22] was indistinguishable from LTP in ACSF prepared brain slices.

Exposing ACSF-prepared brain slices to 100 μm PB (Fig. 2A) did not affect PTP or LTP. In contrast, when tested with slices prepared in sucrose-based ACSF, 100 μm PB significantly attenuated LTP (Fig. 2B). In addition, a saturating concentration of TGB was tested for its effects on TBS-induced PTP and LTP. Similar to PB, TGB failed to affect either PTP or LTP in ACSF-prepared brain slices (Fig. 2C), but did significantly attenuate LTP in sucrose-prepared slices (Fig. 2D).

Figure 2.

Representative GABA-enhancing ASDs attenuate TBS-LTP in brain slices prepared in sucrose-based ACSF. All traces are the average of five sweeps taken before TBS (gray) or 60 min after TBS (black). Scale bars representing 0.25 mV and 2 msec apply to both control and ASD traces. Horizontal red lines from −20 to 0 min represent the duration of drug perfusion in ASD-treated slices (closed red circles) or ACSF perfusion in control slices (open circles). All error bars represent SEM. The following summary time courses are illustrated: (A) 100 μm PB in ACSF-prepared slices or (B) sucrose-prepared slices. (C) 30 μm TGB in ACSF-prepared slices or (D) sucrose-prepared slices. *p < 0.05, **p < 0.01.

Representative third-generation ASDs do not affect TBS-LTP

Using brain slices prepared in ACSF, 10 μm EZG (Fig. 3A), 60 μm TPM (Fig. 3C), and 100 μm LEV (Fig. 3E) failed to significantly affect LTP. However, although there was no effect of TPM or LEV (Fig. 3C,E, respectively), EZG significantly attenuated baseline synaptic transmission by 8.2 ± 2.4% (p = 0.004, two-tailed paired Student's t-test) and reduced the decay time constant of PTP (Fig. 3A). These ASDs were further evaluated in sucrose-prepared brain slices. Unlike PB and TGB (Fig. 2), the effects of these ASDs in sucrose-prepared brain slices were largely indistinguishable from their effects in ACSF-prepared brain slices. EZG again attenuated baseline synaptic transmission (16.8 ± 2.4% reduction, p < 0.0001, two-tailed paired Student's, t-test) and reduced the decay time constant of PTP (Fig. 3B). EZG's attenuation of baseline synaptic transmission was reversible within 60 min as demonstrated in slices where LTP was not induced (Fig. 3B, N = 16 slices/2 mice); a 20 min exposure to 10 μm EZG significantly reduced the fEPSP slope by 11.2 ± 3.7%, and a 60 min wash in ACSF reversed the majority of this attenuation (7.6 ± 3.4% increase in the fEPSP slope). TPM again did not significantly affect any quantified measurement (Fig. 3D). Finally, although there was a significant increase in the PTP time constant, LEV did not significantly affect either baseline synaptic transmission or TBS-LTP (Fig. 3F).

Figure 3.

Representative third-generation ASDs do not affect TBS-LTP. All traces are the average of five sweeps taken before TBS (gray) or 60 min after TBS (black). Scale bars representing 0.25 mV and 2 msec apply to both control and ASD traces. Horizontal red lines from −20 to 0 min represent the duration of drug perfusion in ASD-treated slices (closed red circles), ASD treated slices that did not experience TBS (closed black circles), or ACSF perfusion in control slices (open circles). All error bars represent SEM. The following summary time courses are illustrated: (A) 10 μm EZG in ACSF-prepared slices or (B) sucrose-prepared slices. (C) 60 μm TPM in ACSF-prepared slices or (D) sucrose-prepared slices. (E) 100 μm LEV in ACSF-prepared slices or (F) sucrose-prepared slices.

Valproic acid does not affect LTP in area CA1

As illustrated in Figure 4A, 600 μm VPA failed to significantly alter either PTP or TBS-LTP. Furthermore, VPA continued to have no effect on TBS-LTP in brain slices prepared in sucrose-based ACSF (Fig. 4B). However, VPA did significantly reduce the time constant of PTP in slices prepared using sucrose-based ACSF.

Figure 4.

Valproic acid does not affect LTP in area CA1, but does attenuate TBS-LTP in the dentate gyrus. All traces are the average of five sweeps taken before TBS or HFS (gray) or 60 min after TBS (black). Scale bars representing 0.25 mV and 2 msec apply to both control and ASD traces. Horizontal red lines from −20 to 0 min represent the duration of drug perfusion in ASD-treated slices (closed red circles) or ACSF perfusion in control slices (open circles). All error bars represent SEM. The following summary time courses are illustrated: (A) 600 μm VPA in ACSF-prepared slices or (B) sucrose-prepared slices. (C) 600 μm VPA in ACSF-prepared slices where LTP was induced by HFS. (D) 600 μm VPA in ACSF-prepared slices from Sprague-Dawley rats and LTP was induced by HFS. (E) 1 mm VPA in ACSF-prepared slices. (F) 600 μm VPA's effect on TBS-LTP in the dentate gyrus of ACSF-prepared slices. Ten micromolar picrotoxin was present at all times in panel (F). **p < 0.01.

These results are inconsistent with several VPA studies reported by others.[10, 12, 14, 17] Because methodologic differences could explain these inconsistencies, we examined how a number of factors including induction stimulation, rodent model, and concentration impact VPA's modulatory action on synaptic plasticity. Using HFS, VPA significantly altered the time course of PTP (Fig. 4C). The PTP time constant in VPA-treated slices was 7.4 min (95% CI 6.3–8.8, N = 16 slices/4 mice) and in control slices was 11.2 min (95% CI 9.4–13.9, N = 11 slices/4 mice). However, VPA failed to alter HFS-LTP in mouse brain slices. To rule out a species-specific effect of VPA, we also examined its effects on HFS-LTP in brain slices from Sprague-Dawley rats (Fig. 4D). Consistent with findings from C57BL/6 mice, VPA significantly altered the time course of PTP and failed to alter HFS-LTP; the PTP time constant in VPA-treated slices was 4.3 min (95% CI 3.7–5.2, N = 14 slices/5 mice) and in control slices was 7.7 min (95% CI 6.8–8.9, N = 16 slices/5 mice). VPA's effects on synaptic plasticity were also assessed at 1 mm, but this higher concentration did not significantly alter TBS-induced LTP (Fig. 4E).

Valproic acid attenuates LTP in the dentate gyrus

Because VPA's lack of effect on LTP in area CA1 is inconsistent with its reported in vivo effects,[18] we hypothesized that VPA may affect synaptic plasticity at synapses important for learning and memory other than area CA1. Therefore, VPA's effects on TBS-induced PTP and LTP were assessed at the medial perforant path to DG synapse (Fig. 4F). VPA failed to significantly alter PTP; the PTP time constant in VPA-treated slices was 15.3 min (95% CI 11.8–21.5, N = 14 slices/5 mice) and in control slices was 9.9 min (95% CI 7.1–16.3, N = 11 slices/5 mice). However, 600 μm VPA did significantly attenuate TBS-LTP at this synapse; LTP was 148.5 ± 10.0% in control slices and 119.0 ± 4.4 in VPA-treated slices. Therefore, LTP in VPA-treated slices was significantly attenuated (39.1 ± 9.0% of control, p = 0.0075, unpaired two-tailed Student's t-test).

Discussion

The present study is the first to comprehensively examine the effects of ASDs on PTP and LTP in mouse in vitro brain slices while using the physiologically relevant TBS induction protocol. This study also examined the impact of brain slice preparation protocol on the pharmacologic modulation of synaptic plasticity. Although concentration-response relationships remain to be determined, the results demonstrate that ASDs, at the high end of their therapeutic plasma concentration range, possess a rich and diverse assortment of modulatory actions on synaptic plasticity in the mouse hippocampus that, in most cases, mirror their effects on learning and memory in rodent behavioral models. These effects on synaptic plasticity may contribute to ASDs' impairing effects on learning, memory, and possibly other aspects of cognition.

Sodium channel blocking ASDs

First- and second-generation sodium channel blocking ASDs, such as PHT and CBZ, adversely affect many measures of cognition in healthy volunteers to similar degrees.[23] On the other hand, the third-generation ASD LTG is generally better tolerated and produces fewer or less severe impairments.[24] When examined in rodents, PHT has been shown to disrupt spatial memory,[25] whereas CBZ's effects on memory have been mixed.[25, 26] Some of these discrepancies might be explained by differences in the concentrations employed. Consistent with its benign profile in humans, LTG has been shown to have no negative effect on learning and memory in rodents.[27]

It has long been appreciated that PHT affects HFS-induced PTP in rat area CA1.[28] Similarly, our data reveal that PHT reduces the time course of PTP after TBS in mice. However, PHT has been shown by others to have no effect on HFS-LTP, despite its consistent impairment of learning and memory in humans and experimental animals.[29, 30] In contrast with these prior studies, our data show PHT attenuated TBS-LTP in area CA1 of mice. This inconsistency may be related to the use of TBS or possibly other methodologic differences. Regardless, our data are more consistent with PHT's well-characterized in vivo effects on cognition in both rodents and humans.

Unlike PHT, which affected both PTP and LTP, CBZ suppressed only LTP and LTG suppressed only PTP. These data are consistent with the LTP profiles of CBZ and LTG in other preparations.[9, 11] Although LTG had a small but significant effect on PTP in our experiments, these short-term plasticity effects, which may impact information processing and rapid presynaptic regulation,[31] are less predictive of behavioral cognitive effects. Instead, LTG's lack of effect on LTP is more predictive of its favorable cognitive side-effect profile. Likewise, both PHT's and CBZ's attenuation of LTP is consistent with their impairing effects on learning and memory in rodents and humans.

GABA-enhancing ASDs and the influence of slice-preparation methodologies

Barbiturates, such as PB, have long been associated with a cognition impairing side-effect profile.[18] Additionally, PB was also shown to impair HFS-LTP in area CA1 of rat hippocampal brain slices.[12] However, our initial PB experiments failed to find an effect on synaptic plasticity. Although this incongruity may be due to induction-stimulus differences between these studies, we explored other possible factors that may influence the outcome of these experiments. PB increases the open time of GABA-A receptors and thereby enhances synaptic inhibition[8]; therefore, its lack of effect on TBS-LTP may be due to insufficient GABAergic tone or impaired inhibitory interneuron function. Because others have shown that the use of sucrose-based ACSF better preserves GABA-mediated synaptic transmission,[21] we tested the effect of brain slice preparation protocols on PB's LTP effects. Indeed, when brain slices were prepared in sucrose-based ACSF, 100 μm PB significantly suppressed TBS-LTP.

To further examine the impact of brain slice preparation methodology on GABAergic function, we examined TGB's effects on TBS-LTP. At concentrations of TGB that potently and selectively inhibit GAT1,[32] TGB failed to affect TBS-LTP in brain slices prepared using ACSF. However, this same concentration of TGB significantly attenuated TBS-LTP in brain slices prepared in sucrose-ACSF. These data strongly support the proposition that GABAergic synaptic transmission is better preserved in brain slices prepared in sucrose-ACSF. This, along with the observation that GABA-modulating treatments do not affect LTP when induced by HFS,[15] suggests that the optimal methodology for studying the modulation of LTP by compounds with GABAergic mechanisms of action should utilize TBS in slices prepared in sucrose-based ACSF.

Third-generation ASDs

Third-generation ASDs are generally better tolerated and produced fewer or less severe impairments in cognition,[24] with TPM being the notable exception.[33] However, TPM fails to affect spatial learning and memory in rodents.[34] Similarly, LEV and EZG were also found to have no effect on learning and memory in rodents.[35, 36] Accordingly, these third-generation ASDs may be expected to have a similarly benign profile with regard to their effects, or lack thereof, on synaptic plasticity. Consistent with this hypothesis, others have shown that LEV does not affect HFS-LTP in area CA1,[10] and TPM failed to have any effect on LTP induced at the perforant path–DG synapses in vivo.[13] To our knowledge, the effects of EZG on LTP have not been reported under any experimental conditions.

In our experiments, LEV, TPM, and EZG all failed to significantly affect TBS-LTP in area CA1 of mice, in both ACSF- and sucrose-prepared brain slices. These results are consistent with their lack of effect on learning and memory in behavioral experiments performed in rodents. Furthermore, their lack of effect on LTP in sucrose-prepared slices suggests that LEV, TPM, and EZG do not substantially affect GABAergic synaptic transmission at the concentrations used in these studies. Indeed, because TPM's cognitive effects seem to be on attention, word fluency, verbal memory, and psychomotor speed in humans,[3] it is possible that TPM may preferentially affect synaptic plasticity at synapses outside the hippocampus. Because short-term plasticity effects may play a role in information processing and rapid presynaptic regulation,[31] it remains to be determined if EZG's effects on PTP translate to cognitive impairments in healthy rodents and humans.

Valproic acid

VPA adversely affects many measures of cognition in healthy humans.[18] In rodents, VPA has been shown to disrupt spatial working memory in the eight-arm radial maze,[37] but not in other models.[25] Furthermore, in several studies, approximately 500–1,000 μm of VPA impaired HFS-LTP in area CA1 of the rat hippocampus.[10, 12, 14, 17] In contrast with these previous reports, 600 μm VPA failed to significantly affect TBS-LTP in area CA1 of mice, in both ACSF- and sucrose-prepared brain slices, in our experiments. Because the pharmacologic modulation of LTP can be influenced by induction stimulus, VPA's effect on HFS-LTP in area CA1 of both mouse and rat was also tested. Our data show that VPA's lack of effect on LTP in area CA1 was independent of induction stimuli and rodent species. Furthermore, concentration had no effect, since 1,000 μm VPA failed to modulate TBS-LTP. Therefore, our experiments failed to reproduce the attenuating effects of VPA on LTP shown by others in area CA1 under a variety of conditions. This discrepancy may be due to methodologic differences that have not yet been examined.

We additionally examined VPA's effects on TBS-LTP in the DG in order to test the hypothesis that VPA impairs synaptic function and/or plasticity at synapses important for spatial learning and memory other than area CA1. The DG was chosen because it has been shown to be necessary for encoding and separating spatially similar events into distinct representations.[38] Indeed, 600 μm VPA significantly attenuated TBS-LTP in the DG. These findings suggest that VPA may selectively impair pattern separation and spatial learning. Given VPA's broad-spectrum mechanism of action, the underlying explanation for these differential effects on hippocampal subregions remains to be determined. Be that as it may, these data suggest that a lack of drug effect on LTP in area CA1 alone may not be fully predictive of a benign cognitive profile for new investigational compounds; instead, examination of synaptic plasticity effects at additional synapses important for learning and memory may be necessary.

Conclusion

Cognitive side effects associated with epilepsy pharmacotherapy have been noted since the first ASDs were introduced. The deleterious effects of ASDs on cognition can have varying degrees of impact on the patient's quality of life.[3] Further defining the impact of various ASDs on cognition is particularly important considering that the newer ASDs can have similar seizure-control efficacy.[39] This makes their tolerability profiles, which are largely affected by cognitive side effects, an important differentiating factor.[40] Therefore, the studies presented here may ultimately contribute to our understanding of the differences in cognitive side effect profiles of ASDs and the prediction of cognitive dysfunction associated with novel investigational ASDs.

Acknowledgments

The authors would like to acknowledge Dr. Roy M. Smeal for numerous helpful conversations. This work was supported by NINDS Contract 271201100029C.

Disclosure

H. Steve White has received research grants and a contract from the National Institute of Neurological Disorders and Stroke (NIH). He has also received speaker's or consultancy fees from Insero Health, Janssen Pharmaceuticals, Inc., Upsher-Smith Laboratories, UCB Pharma, and Citizen's United for Research in Epilepsy. Dr. White is also one of two scientific co-founders of NeuroAdjuvants, Inc., Salt Lake City, UT. No other authors have any conflicts of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Ancillary