Sleep deprivation exacerbates seizures and diminishes GABAergic tonic inhibition.

Abstract

Patients with epilepsy report that sleep deprivation is a common trigger for breakthrough seizures. The basic mechanism of this phenomenon is unknown. In the Kv1.1^−/− mouse model of epilepsy, daily sleep deprivation indeed exacerbated seizures though these effects were lost after the 3^rd day. Sleep deprivation also accelerated mortality in ~52% of Kv1.1^−/− mice, not observed in controls. Voltage-clamp experiments on the day after recovery from sleep deprivation showed reductions in GABAergic tonic inhibition in dentate granule cells in epileptic Kv1.1^−/− mice. Our results suggest that sleep deprivation is detrimental to seizures and survival, possibly due to reductions in GABAergic tonic inhibition.

Introduction:

Sleep and epilepsy have a bidirectional relationship where seizures have sleep-wake patterns in various epilepsies and in turn, seizures disrupt sleep. Sleep deprivation (SD) is one of the most common triggers for breakthrough seizures¹. SD is often used to activate diagnostic EEGs² and in epilepsy monitoring units to trigger seizures. Some patient self-report studies showed that SD increases risk of breakthrough seizures, but a randomized controlled trial in an epilepsy-monitoring unit did not^3–4. Regardless, the basic mechanisms underlying SD-induced seizure exacerbation are unknown.

In the Kv1.1 mouse model, homozygous mice with a null mutation in the Kv1.1 gene that encodes the α-subunit of voltage-gated potassium channel⁵ have: a) spontaneous seizures with a sleep-wake or circadian pattern⁶, b) disrupted sleep and circadian rhythms⁷, c) progressive decline in time spent “resting” as mortality appoaches⁸ and d) premature mortality due to status epilepticus or SUDEP that occurs between ~P45–70⁸. In this model of temporal lobe epilepsy, we hypothesized that SD exacerbates seizures due to changes in GABAergic inhibition in the dentate gyrus.

Methods:

All procedures were performed after approval by the Institutional Animal Care and Use Committee (IACUC). A breeding colony of heterozygous Kv1.1 mice (C3HeB background) had been maintained and homozygous Kv1.1^−/− were generated using the heterozygote breeding pairs and after appropriate genotyping procedures.

Animal surgery:

EEG electrode implantation was performed at P34–36 in all mice as previously described⁶. Briefly, mice were anesthetized with isoflurane (5% induction, 1–2% maintenance) and three stainless steel screw electrodes were implanted for EEG (bregma +1.5 mm and 1 mm right, bregma −3 mm and 1 mm left, and lambda −1 mm at midline) and two stainless steel braided wires in the nuchal muscles for EMG recording. After a 1–2-day recovery, mice were transferred into individual recording chambers and allowed a ~24-hour acclimation period. Video EEG with EMG was acquired with an XLTek amplifier (XLTEK, USA) sampled at 1024 Hz.

Sleep deprivation procedure:

SD was performed for 4-hours a day x 5 days starting at lights-on (6:30 am) using gentle handling techniques (presenting novel objects for exploration or by gentle stroking with a soft paintbrush), starting at ~P37. EEG was recorded during a baseline day, during the 5-days of SD (n=19) and during one recovery day. Control animals (n=7) had seven continuous days of EEG recording with no SD. Two independent reviewers, blinded to treatment, manually analyzed EEG data for seizures (Figure 1A).

Figure 1:

Panel A: An example of a seizure is shown from a Kv1.1^−/− mouse. Each mouse had a Right frontal and Left parietal EEG and an EMG channel shown. Seizures consisted of buildup of rhythmic discharges in frontal and parietal EEG lasting 20–60 seconds. Panel B: Mean daily seizure frequency (±SEM) between SD and non-SD controls across the 7 days was 2.5±0.5 vs 0.33±0.21 for day1; 3.5±1.11 vs 1.66±0.49 for day 2; 8.21±2.51 vs 1±0.51 on day 3; 10.92±2.36 vs 1.5±0.56 on day 4; 10.42±2.27 vs 1.33±0.49 on day 5; 6.777±1.36 vs 2.83±1.60 on day 6 and 2.66±0.89 vs 3.16±1.1 on day 7 with a statistically significant interaction between day of recording and SD treatment (Two-way ANOVA: F(6, 84)=3.03; p<0.001). Post-hoc Tukey tests showed SD group had significantly higher seizure frequency on Days 2, 3 and 4 of SD compared to baseline or recovery day (p<0.05**). No difference was seen in non-SD group across days of recording. Panel C: There was ~52% (10/19) mortality in the SD group where they died of status epilepticus during window of SD whereas 100% of non-SD group (7/7) survived the 7-day recording (p=0.02 Fisher’s exact test).

Electrophysiology recordings:

In a separate group of Kv1.1^−/− with and without SD (n=6 each) and wild-type mice that had SD (n=5), hippocampal slices were obtained and patch clamp electrophysiology was performed on the recovery day (SD group) or day-7 (non-SD) according to methods described previously⁹. Briefly, hippocampal slices were submerged in aCSF in a recording chamber. DG granular cells were visually identified with an IR-DIC microscope. Recordings in the voltage-clamp configuration (−70 mV) with a Axopatch 200B amplifier (Axon Instruments; Foster City, CA), low pass filtered at 2 kHz, and digitized at 10 kHz with a Digidata 1440A (Molecular Devices, San Jose, CA) were acquired with pCLAMP 10 (Molecular Devices, San Jose, CA). Borosilicate patch pipettes with a resistance of 2–5 MΩ when filled with intracellular solution containing (in mM): 140 KCL, 10EGTA, 10HEPES, 20 phosphocreatine, 2Mg₂ATP, 0.3NaGTP (pH7.3, 301mOsm), were used to isolate GABAergic currents in a base solution containing 200 nM tetrodotoxin (TTX), 25μM (2R)-amino-5-phosphonovaleric acid (AP-5) and 10 μM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX). To measure GTI, 6-minute recordings were acquired in the base solution, followed by perfusion with 100 μM Bicuculline Methiodide(BIC), a GABA_A receptor antagonist used at a concentration that blocks all GABA_A receptors. Miniature IPSCs (mIPSCs) were detected using the MiniAnalysis program (Synaptosoft, Decatur, GA).

Analysis:

Interactions between SD-treatment and seizures across time was compared using two-way ANOVA with treatment group as a fixed variable and seizure frequency as a random variable, with post-hoc Tukey test for multiple comparisons. Differences in survival rates in SD and non-SD groups were compared using a Fisher’s exact test. Mean tonic currents (±SD) were calculated from all point amplitude histograms of 60-second segments, to which a Gaussian function was fit to the outward part of the current to avoid contamination by mIPSC currents. The magnitude of tonic current was calculated as the difference in fitted means between the baseline current and the current in the presence of BIC. Current density (pA/pF) was calculated by dividing the tonic current by cell capacitance. Differences in tonic current, current density, mIPSC amplitude and frequency were compared using one-way ANOVA with post hoc Tukey test.

Results:

Sleep deprivation exacerbated seizure frequency:

In the SD group, 9/19 Kv1.1^−/− mice survived all 7 days of EEG recording and those that survived until day 4 of SD (n=14) were included in seizure analysis. Daily seizure frequency varied from 0–10 among non-SD animals whereas it was 0–28 in the SD group. Seizure frequency increased in the SD group starting on the 2^nd day of SD, before declining after the 4^th day of SD (Figure 1B). No difference in seizure frequency across the 7-days of recording was seen for the non-SD group (p>0.05). No within group differences were seen for each day of the recording (p=0.23).

Sleep deprivation enhanced mortality due to status epilepticus:

During SD between ~P37 and 44, 10/19 (~52%) went into status epilepticus and all 10 died. Status epilepticus occurred anywhere from day 1 of SD to day 4 of SD, lasting 1–24 hours before an animal died. No mortality was seen among the non-SD or wild type controls (p=0.02, Figure 1C).

Sleep deprivation diminished the magnitude of GABAergic tonic currents in dentate granule cells:

During whole-cell voltage clamp (−70 mV) recordings, addition of the GABA_A receptor antagonist bicuculline to the perfusion solution resulted in blockade of mIPSCs as well as a reduction in holding current, signifying that GABA_A receptors contribute tonically to the resting conductance (Figure 2A). Both the tonic current magnitude and density diminished in sleep-deprived Kv1.1^−/−. Post-hoc Tukey tests showed that the Kv1.1^−/− with no-SD group had higher mean tonic current and current density than the SD group (p<0.05) (Figure 2 B&C). No differences in mean tonic current or current density were seen among wild types (Kv1.1^+/+) with and without SD. Moreover, no differences were also seen in mean mIPSC frequency or amplitude between any SD and non-SD groups (p>0.05) (Figure 2 D&E). The concentration of TTX (200nM) we used is low, but the amplitudes of mIPSCs in our sample are comparable to other studies where 200nM¹⁰ or 500 nM¹¹ concentration was utilized. Findings suggest that SD impaired GABAergic tonic inhibition in the epileptic mice but not in seizure naïve mice and phasic inhibition was unaffected by SD.

Figure 2:

Data of a tonic current experiment is shown for SD and non-SD animals. Panel A shows tracing from a voltage clamp experiment in an SD and non-SD Kv1.1^−/− animal. Tracing shows mIPSCs after application of TTX (200nM), CNQX (10μM) and AP-5 (25μM) to block sodium channels, NMDA and AMPA/kainite receptors, respectively. Then, following addition of Bicuculline, mIPSCs were abolished and a shift on the holding current is seen which is attenuated in a SD animal. Panel B & C: The mean tonic inhibitory currents were 6.23±1.15 pA (±SEM) (15 cells from 4 animals) in the no-SD Kv1.1^−/− mice, 3.16±0.82 pA (16 cells from 6 animals) in Kv1.1^−/− with SD, 2.53±0.43pA in Kv1.1^+/+ with no SD (12 cells from 5 animals) and 1.81±0.39pA in the Kv1.1^+/+ with SD (9 cells from 5 animals) (ANOVA: F(3, 48)=4.957; p=0.004). The mean tonic current density (pA/pF) was 0.116±0.09 in no-SD Kv1.1^−/− whereas it was 0.054±0.04 in SD Kv1.1^−/−, 0.540±0.009 in Kv1.1^+/+ with no-SD and 042±0.018 in Kv1.1^+/+ with SD (ANOVA: F(2 45)=3.871, p=0.01) (means and 95% CI were shown). Post-hoc Tukey test showed that no-SD Kv1.1^−/− animals had much higher tonic current or current density that SD group (p<0.05), whereas no difference was seen in Kv1.1^+/+ animals with and without SD. Panels D & E: Mean mIPSC frequency (1.64±0.17 vs 1.89±0.19 in Kv1.1^−/− animals with and without SD and it was 1.71±0.25 and 1.89±0.20 Hz in wild type Kv1.1^+/+ with and without SD respectively) or amplitudes (48.81±3.23 vs. 41.56±3.26 in Kv1.1^−/− animals with and without SD vs 36.44±1.6 vs 39.75±2.89 pA in Kv1.1^+/+ with and without SD respectively) are shown (means and 95% CI are shown) (p>0.05).

Discussion:

Our study demonstrates that SD in the Kv1.1^−/− model indeed exacerbates seizures. There was a nearly 52% mortality in the SD group that occurred between P37 and 44 whereas none occurred in non-SD controls, suggesting that SD superimposed on a condition where is sleep is already disrupted may affect survival. Interestingly, the effects of SD on seizures were lost after the 3^rd day of SD likely due to yet unknown homeostatic compensatory mechanisms, reminiscent of our prior observation where the adverse consequences of sleep fragmentation on memory function faded after the third day¹².

Mechanistically, in a normal brain, SD is known to decrease excitatory synaptic transmission in the temporal lobe with altered expression of NMDA, AMPA receptors or their ratio¹³. In Transcranial Magnetic Stimulation (TMS) studies, SD reduced intracortical inhibition and increased excitability in focal or generalized epilepsy¹⁴. SD also increased susceptibility to seizures in penicillin, amygdala kindling and electroconvulsive shock models in cats, but the mechanisms are unknown¹⁵. However, there is a dearth of studies examining changes in excitatory or inhibitory synaptic transmission and cellular or network excitability with SD in epilepsy.

GABAergic signaling is the principal mechanism of inhibition in the CNS and consists of phasic and tonic inhibition, the former mediated by the release of presynaptic GABA that activates GABA_A and GABA_B receptors within the postsynaptic and peri-synaptic membrane whereas the latter is mediated by low ambient GABA that diffuses throughout the extracellular space¹⁶. GABAergic tonic inhibition (GTI) is mediated by GABA_A receptors containing α-4 and δ-subunits in the dentate gyrus and by receptors containing α-5 and γ subunits in the CA1 region of hippocampus¹⁷. Under baseline conditions in vitro, the total charge transferred by tonic current is considerably larger than that by phasic current, and thus, reduction in GTI is expected to increase overall excitability¹⁸. In models of temporal lobe epilepsy (TLE) there is evidence of reduction in GABA_A receptor α-5 (in CA1) and δ-subunits in dentate gyrus¹⁹, though surprisingly, several studies showed that tonic currents are normal or even enhanced in the hippocampus in mouse models²⁰ or in human hippocampal tissue²¹. The higher tonic current in the epileptic mice compared to wild types is likely the result of seizures, possibly as a compensatory mechanism. SD-induced impairments in GTI likely increase excitability and contributed to seizure exacerbation. Interestingly however, SD diminished α-5 and δ-subunit containing GABA_A receptors in the hippocampus and diminished GTI in normal C57/BL6 mice²². Ganaxolone, a positive allosteric modulator of δ-subunit containing GABA_A receptors improved sleep in Kv1.1^−/− mice and enhanced survival though not significantly²². Thus, drugs that enhance GTI may have a therapeutic potential against SD-induced seizure exacerbation and needs further exploration. Overall, our results suggest that SD contributes to seizures acutely and leads to reductions in GTI that may also be relevant for epilepsy related mortality or SUDEP in humans.

Stress from gentle handing and granular cell activation by novel objects presented during SD could be confounders in our tonic current experiments and are limitations. Future studies are warranted to understand how the overall excitation-inhibition balance changes with SD. Mechanisms of diminished GTI from SD may include alterations in GABA_A receptor expression, variations in extracellular GABA concentrations or changes in endogenous neurosteroids that allosterically modulate receptors involved in GTI²³. Regardless of the mechanism, sleep is already compromised in epilepsy due to seizures and/or associated sleep disorders, and any superimposed SD may have adverse consequences as demonstrated by the acutely exacerbated seizure burden and hastened mortality in our model.

Abbreviations:

SD

Sleep deprivation

GTI

GABAergic tonic inhibition

GABA

gamma amino butyric acid

SUDEP

Sudden unexplained death in epilepsy

EEG

Electroencephalography

EMG

Electromyography

aCSF

artificial cerebrospinal fluid

NMDA

N-Methyl-D-aspartate

AMPA

α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid

TLE

Temporal lobe epilepsy