Seizure reduces AP half-width via BK channels
Seizure-induced changes in AP waveform were analyzed using whole-cell recording of layer 2/3 cortical pyramidal neurons from young (P13-P16) mouse somatosensory cortex. The GABAA
receptor antagonist picrotoxin was injected in mice to induce generalized seizures, a type that has a strong neocortical component (Willoughby et al., 1995
Comparison of AP waveform in cells from control and 24 hr post-seizure animals (“control” and “post-seizure” groups) showed a significant change in AP half-width (control: 2.93 ± 0.07 ms, n=25 cells; post-seizure: 2.44 ± 0.06 ms, n=26 cells; p<0.001) that could be entirely attributed to decrease in decay time (), suggesting the involvement of K+ channels. Analysis of other intrinsic and active membrane properties did not show any significant differences between control and post-seizure neurons ().
Figure 1 Seizure increases BK channel contribution to AP waveform. (A) Minimal current injection (1 s duration) elicits a single AP from a representative control cell (gray) and 24 hrs post-seizure cell (black). (B) Overlay of AP waveforms aligned to threshold (more ...)
Passive and active membrane properties of control and post-seizure cells.
Inclusion of the intracellular Ca++ chelator BAPTA normalized the difference in AP decay and half-width (), consistent with the activation of a Ca++-gated K+ channel (half-width in control in BAPTA 3.71 ± 0.21 ms, n=5; seizure in BAPTA 3.52 ± 0.07 ms, n=5; p>0.3 for BAPTA in control vs. post-seizure). Application of the BK channel antagonists paxilline or Ibtx induced a significant increase in AP half-width in post-seizure, but not control neurons (; Ibtx data not shown), indicating that after seizure, BK channels have a prominent role in shaping the AP (half-width in control: 2.97 ± 0.18 ms before paxilline versus 3.2 ± 0.11 ms after paxilline, n=5, p>0.3. post-seizure: 2.43± 0.12 ms before paxilline versus 3.1 ± 0.23 ms after paxilline, n=6, p<0.05). In no case did we observe that drug application significantly altered input resistance; thus we conclude that BK channel currents are not a significant source of currents at resting membrane potentials.
BAPTA appeared to have a slightly larger effect on increasing AP half-width compared to either paxilline or Ibtx in both control and post-seizure cells (). Other voltage-gated conductances, such as voltage-gated Ca++ channels, or Ca++-gated conductances, such as IK or SK channels, that might be activated during the repolarizing phase of the AP could be responsible for the relative increase in spike half-width in BAPTA, though activation of these conductances is typically not observed during the AP itself. Another possibility is that under our experimental conditions, BAPTA provides a more complete blockade of BK channels than either paxilline or Ibtx alone.
The reduction in AP half-width required the induction of prior seizures, since administration of a subconvulsive dose of picrotoxin (1 mg/kg, no seizures observed) did not significantly change spike width (half-width after subconvulsive dose 3.19±0.13ms; n= 14 cells; p>0.5 vs control).
Whole-cell BK channel currents are increased after seizure
To further verify that BK channel currents were enhanced following seizures, we pharmacologically isolated BK channel currents from both control and post-seizure cells using paxilline. The amplitude of steady-state BK currents were examined by comparing currents after a 180 mV voltage step applied from an initial holding potential of -80 mV, where averaged traces were subtracted before and after paxilline application to yield the BK channel current. On average, total K+-channel currents were not significantly larger 24 hrs after picrotoxin-induced seizures compared to control (control 1294 ± 312 pA, n=9 versus post-seizure 2144 ± 429 pA, n=9; p>0.1).
We found that BK current amplitude was highly variable amongst layer 2/3 pyramidal neurons in both control and post-seizure conditions. However, on average, BK current amplitudes significantly increased after seizure (; amplitude of paxilline-sensitive current in control 133.5 ± 61.1 pA, n=9; post-seizure 547.78 ± 195.5 pA, n=9, p<0.05). In BK channel currents from some cells, there appeared to be a clear reduction in activation kinetics after seizure (see for example, ). However, this finding was not consistent between cells within a group and failed to reach statistical significance in a comparison between control and post-seizure cells (p>0.5). These data indicate that concurrent with the seizure-dependent reduction in AP half-width, somatic BK channel currents are also enhanced, and that K+ influx via BK channels may comprise a larger fraction of total cellular K+-currents.
Figure 2 BK channel currents are altered after seizure. (A) Representative potassium channel currents from a control neuron (P14) before (left) and after (right) paxilline application. Cells were held at -80 mV and an 80-160 ms voltage step (+100 to +180 mV in (more ...)
BK channels increase evoked firing frequency
Does a gain-of-function in BK channels influence firing output and network excitability after seizure? It has been shown experimentally that BK channel antagonists can reduce high firing rates in both neocortical (Jin et al., 2000
) and hippocampal pyramidal neurons (Gu et al., 2007
). A gain-of-function in BK channels could increase spike output directly by reducing early interspike intervals (ISIs) during a period of strong synaptic activity, or indirectly by increasing Na+
channel deinactivation that occurs during the fast after-hyperpolarization (fAHP) and thus enhancing the availability of Na+
channels to participate in later APs during high frequency firing (Gu et al., 2007
To examine whether enhanced BK channel activity after seizure might increase spike output at the level of an isolated neuron, we compared spike trains elicited by current injection in control and post-seizure cells in the presence or absence of the BK channel antagonist paxilline (). Instantaneous firing frequency (IFF; inverse of the first ISI) was significantly increased in post-seizure neurons (), despite comparable minimal current required for a single AP and input resistance values (see ). Paxilline application to post-seizure neurons normalized IFF to control values, suggesting that this was primarily due to BK channels ().
Figure 3 Evoked firing is increased after seizure in a BK channel -dependent manner. (A) Evoked firing during a 200 pA current injection (1 sec) in control neurons. (B) Same as (A) but after paxilline application. (C) Evoked firing as in (A) but in a post-seizure (more ...)
The post-seizure increase in IFF was accompanied by an increase in the total number of spikes elicited by a sustained 200 ms current injection (. Number of spikes elicited by a 200 pA current injection: control 9.75 ± 1.38, n=5; control in paxilline 9.33 ± 0.92, n=6; seizure 14.20 ± 1.36, n=6; seizure in paxilline 8.2 ± 1.11, n=6). This increase was sensitive to BK channel antagonists, since paxilline application reduced firing output in post-seizure cells (). Paxilline induced a significant increase in ISI for post-seizure neurons, but had no effect on total spike output, ISI (), or IFF (data not shown) for control cells. This finding is consistent with results from , where BK channel antagonists had no significant effect on AP half-width in control neurons and support the conclusion that BK channels do not significantly influence AP half-width and firing output in layer 2/3 pyramidal neurons in slices from control animals, at least under our recording conditions. Furthermore, these results indicate that BK channels can increase spike output in a cell-autonomous manner, independent of changes in synaptic drive that may occur after seizure.
Increased BK channel function does not alter neurotransmitter release
Under some conditions, BK channel activity can reduce neurotransmitter release at excitatory synapses in the hippocampus by reducing the duration of depolarization at the axon terminal (Hu et al., 2001
; Stewart and Foehring, 2001
; Raffaelli et al., 2004
). A seizure-dependent gain-of-function in BK channels at the axon terminal could conceivably result in reduced neurotransmitter release and thus decrease cortical excitability. In contrast, activation of BK channels in some cell types can enhance neurotransmitter release (Pattillo et al., 2001
; Xu and Slaughter, 2005
To more directly examine the contribution of BK channel to regulating neurotransmitter release, we examined the amplitude of paired-pulse-evoked EPSCs at excitatory synapses in layer 2/3. One measurement of neurotransmitter release efficacy is to examine how two closely timed stimuli affect vesicle release, and thus, the amplitude of the post-synaptic response. Excitatory synapses in this layer exhibit paired-pulse facilitation (PPF), indicating that an initial stimulus primes neurotransmitter release to a second, closely spaced stimulus. If a gain-of-function in BK channel was apparent at axon terminals after seizure, PPF might be changed as release probability to the initial stimulus was altered.
In order to more directly examine a role for BK channels in mediating neurotransmitter release after seizures, we examined the paired pulse ratio (PPR) at excitatory synapses within layer 2/3 in both control and post-seizure cells. We found that under control conditions, BK channels have little effect in mediating the PPR, since bath application of paxilline did not alter the PPR (amplitude of 2nd/1st synaptic response) in control neurons (baseline: 1.26 ± 0.09 pA; post-paxilline: 1.35 ± 0.02 pA; n=4 cells, p>0.3). After seizure, PPR ratios were similar to control and were not altered by paxilline application (baseline: 1.21 ± 0.04 pA; post-paxilline 1.25 ± 0.03 pA; n=4 cells, p>0.4), suggesting that a BK channel gain-of-function does not change neurotransmitter release properties at synapses after seizure (). Thus, we find that a seizure-dependent gain-of-function in BK channels does not influence glutamate release at synapses onto layer 2/3 pyramidal neurons, at least as determined under our recording conditions.
Figure 4 BK channels do not contribute to short-term facilitation. (A-B) Paired pulse synaptic currents evoked by layer 2/3 stimulation (interstimulus interval = 50 ms) in the presence of 50 uM picrotoxin before (baseline) and after bath application of 10 nM paxilline (more ...)
BK channel antagonists reduce spontaneous firing after seizure
We analyzed spontaneous firing activity of single neurons in acute brain slices, a preparation that approximates the intact cortical network in vivo
(Sanchez-Vives and McCormick, 2000
). Spontaneous activity under these conditions is the sum of many different variables, such as total inhibitory and excitatory drive and intrinsic excitability; as such, it can be a useful indicator of seizure-dependent changes in average network activity. Furthermore, whole-cell recording in acute brain slices provides excellent pharmacological access and enables the examination of the role of specific ionic conductances on network excitability. Since a gain-of-function in BK channels enhanced spike output at the level of a single cell, we expected that these effects might be magnified in a semi-intact network in brain slices, resulting in an overall increase in firing rates.
Because spontaneous firing rates are extremely low (<0.01Hz) in the presence of ACSF containing elevated Ca++
, we used a solution with lower Ca++
that more closely resembles CSF in vivo
; Sanchez-Vives and McCormick, 2000
) to examine spontaneous firing activity in acute brain slices. Under these conditions, spontaneous firing was observed in slices from control animals (). Twenty-four hours after the initial seizure, we observed a significant increase in firing activity (; control 0.078 ± 0.012 Hz, n=13; post-seizure 0.173 ± 0.033, n=14, p<0.05 between post-seizure versus all other groups by ANOVA) suggesting that seizures may initiate a cascade of changes that result in an increase in network activity in the cortex.
Figure 5 BK channel antagonists reduce spontaneous activity after seizure. (A) Spontaneous firing activity over the course of ~8 minutes from a representative control cell. (B) Representative example of spontaneous firing in the presence of paxilline in a control (more ...)
Elevated spontaneous firing rates in post-seizure neurons could be reduced to near control levels by application of BK channel antagonists (; post-seizure in paxilline 0.040 ± 0.015 Hz, n=7; post-seizure in iberiotoxin 0.031 ± 0.010 Hz, n=8). Although BK channel antagonists induced a reduction in firing rates when applied to control slices, this difference was not significant (; control in paxilline 0.031 ± 0.008 Hz, n=7; control in iberiotoxin 0.054 ± 0.030 Hz, n=11).
Firing activity was not due to intrinsic bursting of layer 2/3 neurons but was dependent on synaptic transmission, since bath application of the AMPA receptor antagonist NBQX (10 μM), the NMDA receptor antagonist D-APV (50 μM), and the GABAA receptor antagonist Picrotoxin (100 μM) eliminated all firing (, n=4, p<0.01 versus all other groups by ANOVA). These data indicate that antagonism of BK channels is sufficient to reduce abnormal firing activity after chemoconvulsant-induced seizures in semi-intact cortical networks.