BK potassium channels facilitate high-frequency firing and cause early spike frequency adaptation in rat CA1 hippocampal pyramidal cells

Abstract

Neuronal potassium (K(+)) channels are usually regarded as largely inhibitory, i.e. reducing excitability. Here we show that BK-type calcium-activated K(+) channels enhance high-frequency firing and cause early spike frequency adaptation in neurons. By combining slice electrophysiology and computational modelling, we investigated functions of BK channels in regulation of high-frequency firing in rat CA1 pyramidal cells. Blockade of BK channels by iberiotoxin (IbTX) selectively reduced the initial discharge frequency in response to strong depolarizing current injections, thus reducing the early spike frequency adaptation. IbTX also blocked the fast afterhyperpolarization (fAHP), slowed spike rise and decay, and elevated the spike threshold. Simulations with a computational model of a CA1 pyramidal cell confirmed that the BK channel-mediated rapid spike repolarization and fAHP limits activation of slower K(+) channels (in particular the delayed rectifier potassium current (I(DR))) and Na(+) channel inactivation, whereas M-, sAHP- or SK-channels seem not to be important for the early facilitating effect. Since the BK current rapidly inactivates, its facilitating effect diminishes during the initial discharge, thus producing early spike frequency adaptation by an unconventional mechanism. This mechanism is highly frequency dependent. Thus, IbTX had virtually no effect at spike frequencies < 40 Hz. Furthermore, extracellular field recordings demonstrated (and model simulations supported) that BK channels contribute importantly to high-frequency burst firing in response to excitatory synaptic input to distal dendrites. These results strongly support the idea that BK channels play an important role for early high-frequency, rapidly adapting firing in hippocampal pyramidal neurons, thus promoting the type of bursting that is characteristic of these cells in vivo, during behaviour.

Figure 10. Removing the inactivation of I_BK in the model reduced the adaptation and increased the excitability

A, diagram of the I_BK Markov model states; C, closed state; O, open state; I, inactivated state. B, the rate of the I→C transition was increased sufficiently to fully prevent cumulative I_BK inactivation, thus producing a ‘non-inactivating’ BK current (red plot) (see Methods). Left panel: f–I relation of the first interspike interval was not changed by removing I_BK inactivation. Right panel: f–I relation for the first 4 spikes (average frequency) was increased by removing I_BK inactivation. Ca, top trace: spike train in response to a depolarizing current pulse for I_BK with (black) and without (red) inactivation (scale bar, 30 mV). Note the fast afterhyperpolarization (fAHP, arrows) following every spike when inactivation was removed. Middle trace: I_BK during the spike train (scale bars: 5 ms and 3 nA). Cb, plot of adaptation based on the simulations shown in Ca. D, in the model, a synapse on the apical dendrite (left cartoon) was activated strongly so that the EPSP evoked a burst of spikes at the soma. Top panels: a series of somatic voltage responses of the model; the recovery rate of I_Na inactivation was reduced stepwise from fast (left panels) to slow (right panels; see Methods). Bottom panels: same as above, but without inactivation of I_BK.

Figure 1. Tetraethylammonium (TEA) reduced high-frequency discharge

Sharp intracellular electrode recordings from rat CA1 pyramidal cells. Aa, repetitive firing at two different frequencies was evoked by somatic injection of depolarizing current pulses. Left: a strong (0.8 nA) 50 ms long pulse evoked 5 action potentials (APs). Right: a weaker (0.2 nA) 100 ms long pulse evoked 4 APs. The current pulses, which were separated by a 6 s interval, were repeated at constant intensities throughout the experiment, while the background membrane potential was kept at −60 mV by injecting steady depolarizing current. Ab, bath application of TEA (100 μm) clearly reduced the high-frequency firing rate and spike number (left: the number of APs was reduced from 5 to 4), but had no significant effect on low firing frequency (right). Ac, washout of TEA for 10 min reversed the effect. B, time course of the 1st ISI firing frequency. (•: high frequency; ○: low frequency). C, summary data showing that 100 μm TEA significantly reduced the 1st ISI high firing frequency (*: n = 7; washout: n = 4; P < 0.05), but had no effect on low-frequency firing (NS: n = 7; washout: n = 4; P > 0.05). D, time course of the averaged firing frequency of the 1st–3rd interspike intervals (ISIs): •, high frequency; ○, low frequency. E, summary data showing that 100 μm TEA significantly reduced the mean firing frequency of 1st–3rd ISI in response to the strong (0.8 nA, 50 ms) pulse (**: n = 7; washout: n = 4; P < 0.01), but had no effect on the low firing frequency in response to the weak (0.2 nA, 100 ms) pulse (NS: n = 7; washout: n = 4; P > 0.05).

Figure 5. BK channel effects on spike frequencies: results from IbTX experiments and computer simulations

A, action potential frequency of the first interspike interval (1st ISI) in response to a series of 1.0 s long depolarizing current pulses (0.1–0.6 nA, in steps of 0.05 nA) in normal medium (•) and after adding 100 nm IbTX (○). The plot shows the mean firing frequencies from all CA1 hippocampal pyramidal cells tested (n = 7). IbTX clearly reduced the high-frequency firing in response to strong current pulses (≥ 0.2 nA; *P < 0.05; **P < 0.01), but had no significant effect on the low firing rate in response to 0.15 nA pulses (NS: P > 0.05). B, similar to A, but showing the average frequencies of the first three intervals (1st–3rd ISI). IbTX significantly reduced the spike frequencies in response to current pulses > 0.35 nA (*: n = 7, P < 0.05), but had no significant effect at the lower current intensities (0.15–0.3 nA; NS: n = 7, P > 0.05). C and D, computer simulations with a CA1 pyramidal cell model showed that BK channels promote high-frequency firing. The f–I relations of the 1st ISI (C) and of the average of the first 3 interspike intervals (D, 1st–3rd ISI) are plotted for different BK channel conductance densities (G_BK: 0, 10, 20, 40, 80, 130, 200, 400, 600 and 800 pS μm⁻²). Current pulses of increasing amplitude were applied at the soma (pulse duration: 1 s). The membrane potential prior to each current pulse was held at −60 mV by steady current injection. The continuous lines represent the standard model with I_BK present (130 pS μm⁻²) or blocked (0 pS μm⁻²).

Figure 7. Blockade of BK channel reduced the burst firing frequency in the presence of XE991 and forskolin

A, sample traces from a sharp-electrode current-clamp recording showing the first three action potentials (1st–3rd spikes) evoked by a 0.8 nA current pulse, before and after bath-application of 100 nA IbTX. The perfusion media contained XE991 and forskolin, to block Kv7/KCNQ/M- and sAHP-currents, respectively. Ba–Da, time courses of the effects of IbTX (100 nA, same experiment as in A) on the decay slope of the 1st spike (Ba), the rise slope of the 2nd spike (Ca), and the instantaneous firing frequency of 1st ISI (Da). Similar results were observed in all the cells tested with IbTX (100 nA, n = 3), as shown by the column graphs in Bb, Cb and Db (*P < 0.05, **P < 0.01).

Figure 9. Experiments and computer modelling showed reduced spike-frequency adaptation after BK channel blockade

Aa, sample trace showing an early part of a spike train in response to a brief current injection. Before the pulse, the cell was held at −60 mV by DC current injection. Note that in normal medium (Control), the cell showed strong spike-frequency adaptation. Ab, after bath-application of 100 nm IbTX, a current step with stronger intensity was injected to the cell in order to achieve the same firing frequency of the 1st ISI as before IbTX (vertical dashed lines). Note that the 2nd and 3rd interspike intervals (ISIs) were clearly reduced compared with the control condition. B and C, summary plots showing reduced spike-frequency adaptation after BK channel blockade by 100 nm IbTX: B, the 2nd and 3rd ISIs were clearly reduced by IbTX when the 1st ISI was kept constant (NS: n = 6, P > 0.05; **: n = 6, P < 0.01). C, spike frequencies of the 2nd and 3rd ISIs were clearly increased by IbTX when the frequency of the 1st ISI was kept constant (NS: n = 6, P > 0.05; **: n = 6, P < 0.01). D–F, results from computer simulations of the experiments shown in panels A–C, respectively.

Figure 2. The BK channel blocker IbTX reduced early high-frequency firing

A and B, typical sharp intracellular electrode recordings from a rat CA1 pyramidal cell. Ba, a high-frequency train of 5 action potentials (APs) was evoked by a 0.9 nA, 50 ms long depolarizing current pulse, starting from a membrane potential of −60 mV (maintained by steady depolarizing current injection). Bb, 13 min after application of iberiotoxin (IbTX, 100 nm), the cell fired only 4 APs in response to an identical current pulse. A shows the expanded traces of the 1st and 2nd APs taken from B, showing that the 2nd was significantly delayed after application of IbTX (Ac). Blockade of BK channel broadened the action potentials and eliminated the fast afterhyperpolarizations (fAHPs; arrow). Summary time course showing that 100 nm IbTX reduced the instantaneous firing frequency of the 1st ISI from 169 ± 5 Hz to 146 ± 7 Hz (C, n = 6), and the averaged 1st−3rd ISI frequency from 122 ± 4 Hz to 103 ± 3 Hz (D, n = 6). E, summary data from all cells tested (n = 6): comparison of the 1st ISI frequencies and averaged 1st−3rd ISI frequencies before and after application of IbTX (n = 6, each. *P = 0.016, **P = 0.001).

Figure 3. Application of IbTX had little or no effect on low-frequency firing

A and B, typical sharp intracellular electrode recordings from a rat CA1 pyramidal cell. Ba, the cell fired a train of 4 APs in response to a 0.4 nA, 100 ms long depolarizing current pulse, with the preceding membrane potential held at −60 mV by steady-state current. Aa–c show expanded traces taken from 1st to 2nd action potentials of Ba–c. Ab and Bb were recorded 14 min after bath application of 100 nm IbTX. Bc, overlaid traces showing that IbTX did not significantly change the firing rate of the cell. The current pulse was constant (0.4 nA). Ac, IbTX caused spike broadening of the two first spikes, but the interspike interval (ISI) was not changed. C, summary time course from all the cells tested shows that IbTX did not significantly change the instantaneous firing frequency of the 1st ISI: (before IbTX: 68.1 ± 8.5 Hz; after application of IbTX: 65.2 ± 8.5 Hz; n = 4). D, time course of averaged 1st–3rd ISI frequency (41.3 ± 4.1 and 40.8 ± 3.4 Hz, before and after application of IbTX, respectively) (n = 4). E, summary graph showing no significant IbTX effect on either the 1st ISI or the averaged 1st–3rd ISI frequency (NS: P > 0.05).

Figure 4. Effects of IbTX on the 1st and 2nd action potential slope during early high-frequency firing

Aa–c, recordings of the first two action potentials during a high frequency spike train, elicited by a brief, depolarizing current step (0.8 nA, 50 ms). Aa, in normal medium, the ISI between the 1st and 2nd spikes was 5.7 ms (firing rate: 175 Hz). Note the fAHP following the 1st AP (arrow). Ab, application of the BK channel blocker IbTX (100 nm) delayed the onset of the 2nd action potential (ISI: 7.2 ms; frequency: 139 Hz). Ac, overlaid traces of Aa and Ab, comparing the 1st and 2nd APs evoked by identical current pulses (0.8 nA) before and after application of IbTX. Ba, the decay slope of the 1st AP was reduced after application of IbTX (dashed line), whereas the rising slope remained unchanged (see inset at expanded time scale, X: 0.2 ms; Y: 50 mV). Bb, both the rising slope (see inset) and the repolarization slope of the 2nd AP were slowed down by IbTX (dashed line), and there was a slight elevation of the spike threshold (▴). Ca–Fa, normalized average time courses (%) from all the cells tested: IbTX effects on the rising slope (Ca) and decay slope (Da) of the 1st AP, and the rising slope (Ea) and decay slope (Fa) of the 2nd AP. Cb–Fb, summary graph from all the cells tested, showing that BK channel blockade significantly reduced rising slope of the 2nd AP· Eb (*: n = 5, P < 0.05), as well as the decay slopes of both 1st AP (Db,**: n = 5, P < 0.01) and 2nd AP (Fb, **: n = 5, P < 0.01). In contrast, IbTX had no significant effect on the rising slope of the 1st AP (Cb, NS: n = 5, P > 0.05).

Figure 6. Computer simulations: reduction in firing frequency by BK channel blockade is not dependent on activation of I_M or I_sAHP

A and B, f–I relations of the first interspike interval (A, ISI-1) and of the average of the first 3 interspike intervals (B, ISI-1–3) from simulations with I_sAHP present (dashed lines) or blocked (continuous lines). The f–I relations were simulated as described in Fig. 5. C and D, similar to A and B, but here the f–I relations were plotted from simulations with I_M present (dashed lines) or blocked (continuous lines).

Figure 8. Computer simulations of a spike train and the response of various ionic currents, while varying the BK channel density

A, somatic responses to a current pulse (50 ms, 0.4 nA). The membrane potential prior to the pulse was held at −60 mV by steady current. The simulations were repeated for 3 different BK channel densities: 0 (red), 130 (black), and 600 pS μm⁻² (dashed black). The first four spikes in the train are shown. Note that the delayed rectifier potassium current (I_DR) increases considerably during the interspike intervals when the spikes are broadened as a consequence of a reduced BK channel density (0 pS μm⁻², red). Also, I_Na does not recover completely from inactivation when the spikes get broader (I_Na, I-state). B, phase plot of the second spike. Note the increased threshold and reduced rise slope and spike peak when I_BK is blocked (red arrow).

Figure 11. Blocking BK channels with IbTX reduced the firing frequency of population spikes in striatum pyramidale

A, a tungsten electrode was positioned in the middle of stratum radiatum to stimulate the Schaffer collaterals (stimulation duration, 150 μs; stimulation interval, 30 s). A field recording electrode was positioned in stratum pyramidale. The perfusion medium contained 4.0 mm K⁺, bicucculine (15 μm) and APV (100 μm). B, under these conditions, a single stimulus (arrow) evoked 3 population spikes (PS). The stimulation artifact is indicated by an arrow. The amplitude of the stimulus current was adjusted so the interval between the first two population spikes was ∼5 ms (i.e. ∼200 Hz). Bath application of 100 nm IbTX reduced the frequency of the population spikes. C, the average (n = 4) time course of the IbTX effect on the PS latencies following the stimulation artifact (t = 0). D, average time course of the IbTX effect on the frequency of the first and second PS interval (ISI1 and ISI2). E, average time course of the normalized change in decay slope of the three population spikes. F and G, summary data of the IbTX effect on the rise slopes and decay slopes of the three population spikes (n = 4, NS: P = not significant, *0.01 < P < 0.05, **P < 0.01).