Spike Ca²⁺ influx upmodulates the spike afterdepolarization and bursting via intracellular inhibition of K_V7/M channels

Abstract

In principal brain neurons, activation of Ca²⁺ channels during an action potential, or spike, causes Ca²⁺ entry into the cytosol within a millisecond. This in turn causes rapid activation of large conductance Ca²⁺-gated channels, which enhances repolarization and abbreviates the spike. Here we describe another remarkable consequence of spike Ca²⁺ entry: enhancement of the spike afterdepolarization. This action is also mediated by intracellular modulation of a particular class of K⁺ channels, namely by inhibition of K_V7 (KCNQ) channels. These channels generate the subthreshold, non-inactivating M-type K⁺ current, whose activation curtails the spike afterdepolarization. Inhibition of K_V7/M by spike Ca²⁺ entry allows the spike afterdepolarization to grow and can convert solitary spikes into high-frequency bursts of action potentials. Through this novel intracellular modulatory action, Ca²⁺ spike entry regulates the discharge mode and the signalling capacity of principal brain neurons.

In many types of principal brain neurons the fast spike is followed by a slow afterdepolarization (ADP) lasting tens to hundreds of milliseconds. Large ADPs cause neurons to fire in burst mode. The propensity of a neuron to burst determines its impact on target neurons, as well as the plasticity of its input and output synaptic contacts (Lisman, 1997; Kepecs & Lisman, 2003). Accordingly, intrinsically bursting neurons may serve as pacemakers of normal and abnormal rhythmic network activity (e.g. Chagnac-Amitai & Connors, 1989; Jensen & Yaari, 1997; Sanabria et al. 2001; Sipila et al. 2005; Wittner & Miles, 2007) and may promote long-term potentiation (Thomas et al. 1998; Pike et al. 1999; Fortin & Bronzino, 2001) and epileptogenesis (Yaari et al. 2007). Thus, the properties of the currents generating or abating the ADP are critical determinants of neuronal discharge behaviour at both single neuron and network levels.

Hippocampal CA1 pyramidal cells manifest a prominent spike ADP that can trigger bursting in a variety of conditions (Schwartzkroin, 1975; Masukawa et al. 1982; Jensen et al. 1994, 1996; Azouz et al. 1997; Sanabria et al. 2001; Su et al. 2001). The spike ADP comprises a passive component reflecting recharging of the membrane capacitor and an active component produced by voltage-gated conductances (Jensen et al. 1996; Metz et al. 2007). Experimental and theoretical analyses have shown that in adult CA1 pyramidal cells, persistent Na⁺ current (I_NaP; French et al. 1990) in the perisomatic region is the predominant inward current generating the active ADP (Azouz et al. 1996; Su et al. 2001; Yue et al. 2005; Golomb et al. 2006). The depolarizing action of I_NaP is counteracted primarily by perisomatic M type K⁺ current (I_M; Halliwell & Adams, 1982), generated by K_V7 (KCNQ) channels (Wang et al. 1998; Shah et al. 2002). This current prevents the escalation of the ADP into a spike burst (Yue & Yaari, 2004, 2006). Recruitment of d-type K⁺ channels in the proximal apical dendrites also contributes to curtailment of the ADP (Metz et al. 2007). The role of voltage-gated Ca²⁺ currents has been more elusive, because these currents can enhance the ADP by a direct depolarizing action and/or suppress it by activating various Ca²⁺-gated K⁺ channels (Wong & Prince, 1981; Friedman & Gutnick, 1987; Jung et al. 2001). There is compelling evidence that different Ca²⁺ currents contribute to ADP generation in developing CA1 pyramidal cells (Chen et al. 2005; Metz et al. 2005), but experimental evidence extrapolating this conclusion to normal adult neurons is meagre. In the latter neurons, Ca²⁺ currents have been implicated in ADP enhancement and bursting only in abnormal situations in which the backpropagating somatic spike initiates a Ca²⁺ spike in the apical dendrites, which, in turn, spreads to the soma, reinforcing the locally I_NaP-driven ADP (Magee & Carruth, 1999; Yaari et al. 2007). In normal conditions, however, apical dendritic Ca²⁺ currents activated by the backpropagating somatic spike are too small to ignite a local Ca²⁺ spike (Jaffe et al. 1992; Spruston et al. 1995; Hoffman et al. 1997).

Here we have used electrophysiological and pharmacological techniques to characterize the role of Ca²⁺ currents in generating the spike ADP. We report that activation of several Ca²⁺ current types during the spike strongly facilitates the spike ADP and the associated propensity for bursting. However, this enhancement is not due to a direct depolarizing action of these currents. Rather, it is due to intracellular Ca²⁺-mediated inhibition of I_M subsequent to spike Ca²⁺ influx, which unleashes the depolarizing action of I_NaP.

Methods

Slice preparation

All animal experiments were conducted in accordance with the guidelines of the Animal Care Committee of the Hebrew University. Transverse hippocampal slices were prepared from adult (125–150 g) male Sabra rats. Animals were decapitated under isoflurane anaesthesia, and transverse hippocampal slices (400 μm) were prepared with a vibrating microslicer (Leica) and transferred to a storage chamber perfused with oxygenated (95% O₂–5% CO₂) ACSF containing (mm): NaCl 124, KCl 3.5, MgCl₂ 1, CaCl₂ 1.6, NaHCO₃ 26 and d-glucose 10, pH 7.4, osmolarity 305 mosmol l⁻¹, where they were maintained at room temperature. For experiments, slices were placed one at a time in an interface chamber (33.5°C) and superfused with oxygenated ACSF. ACSFs contained also the glutamate receptor antagonists 6-cyano-7-nitro-quinoxaline-2,3-dione (CNQX; 15 μm) and 2-amino-5-phosphono-valeric acid (APV; 50 μm) to block fast excitatory postsynaptic potentials, and the GABA_A receptor antagonist bicuculline (10 μm) to block fast inhibitory postsynaptic potentials.

Electrophysiology

Intracellular recordings were obtained using sharp glass microelectrodes containing 4 m potassium acetate (90–110 mΩ) and an amplifier (Axoclamp 2A, Molecular Devices, Union City, CA, USA) allowing simultaneous injection of current and measurement of membrane potential. The bridge balance was carefully monitored and adjusted before each measurement. In some experiments, where indicated, 200 mm 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA), was included in the microelectrode filling solution. The pyramidal cells included in this study had stable resting potentials of −60 mV or more, and overshooting action potentials. The intracellular signals were filtered on-line at 10 kHz, digitized at a sampling rate of 10 kHz or more, and stored by a personal computer using a data acquisition system (Digidata 1322A) and pCLAMP9 software (Molecular Devices).

Chemicals and drugs

All chemicals and drugs were obtained from Sigma (Petach-Tikva, Israel), apart from ω-conotoxin MVIIC, ω-conotoxin GVIA, ω-agatoxin IVA (Alomone Laboratories, Jerusalem, Israel) and CNQX (RBI, Natick, MA, USA). Stock solutions of nifedipine (10 mm) were prepared in dimethyl sulfoxide (DMSO) and diluted 1 : 1000 when added to the ACSF. In experiments testing the effects of nifedipine (conducted in darkness because of its light sensitivity), equivalent amounts of DMSO were added to the control ACSFs. All other drugs were added to the ACSF from aqueous stock solutions. Measurements of drug effects were usually performed after 30 min of slice perfusion with the drug-containing ACSF.

Data analysis

Solitary spikes were evoked by injecting threshold-straddling, brief (4 ms) positive current pulses. The size of the spike ADP was measured as the integrated ‘area under the curve’ between the fast afterhyperpolarization and the point at which membrane voltage returned to resting potential. Spike width was measured at 50% of spike height. To evoke a large slow AHP (sAHP), the neurons were injected with prolonged (400 ms) suprathreshold positive current pulses. The sAHP amplitude was measured 150 ms after stimulus termination.

Results are presented as the mean ±s.e.m., unless stated otherwise. Assessment of statistical significance of differences between means was performed with one-way ANOVA or paired Student's t test, as appropriate. Significance of linear regression models was tested using the F statistic. The t statistic was used to test whether slope coefficients of linear regression lines were significantly different from zero. In all tests the significance level was set to P < 0.05.

Results

Sodium currents contribute to spike ADP generation

In order to assess the contribution of different Na⁺ and Ca²⁺ currents to ADP electrogenesis in adult CA1 pyramidal cells, we examined how selective blockers of these currents affect ADP size (measured as ‘area under the curve’; see Methods). Drugs were added to the perfusing ACSF, which in this study contained 1 mm Mg²⁺ (compared to 2 mm Mg²⁺ in our previous study of this topic; Yue et al. 2005) to approximate the more physiological divalent ion composition (McNay & Sherwin, 2004). We first reexamined the effects of tetrodotoxin (TTX), which blocks both transient Na⁺ current (I_NaT) and I_NaP (French et al. 1990). The earliest noted effect of 10 nm TTX was suppression of the ADP. After ∼20 min of exposure to TTX, the ADP was reduced to 79.7 ± 5.3% of control (n = 7; P < 0.05), while spike amplitude decreased very slightly (97.0 ± 1.1% of control; Fig. 1A). Thus, ADP suppression by TTX is not secondary to spike attenuation, but likely to be due to direct block of persistent Na⁺ channels. With time, the ADP was further reduced, but this suppression could no longer be dissociated from spike attenuation.

Figure 1. Both Na⁺ and Ca²⁺ channel blockers suppress the ADP in adult CA1 pyramidal cells.

A–H, representative recordings of spikes and ADPs. Each panel displays two overlaid traces of spikes evoked by brief (4 ms) depolarizing current pulses before (Control) and after 20 min (A) or 30 min (B–H) of exposure to the indicated drug. Each trace is an average of 5 sequential intracellular recordings. Spikes are truncated at the dashed line. Here and in following figures, the resting potential (in mV) is indicated to the left of the voltage traces. The slices were superfused with ACSFs containing 1 μm TTX (A), 10 μm riluzole (B), 1 μmω-conotoxin MVIIC (CTX-MVIIC) (C), 1 μmω-conotoxin GVIA (CTX-GVIA) (D), 0.2 μmω-agatoxin IVA (ATX-IVA) (E), 100 μm Ni²⁺ (F), 10 μm nifedipine (G) and 2 mm Mg²⁺ (H). All treatments, except for nifedipine, caused significant suppression of the spike ADP. Insets on right sides of panels A and B show overlaid traces of full spikes. I, bar graph summarizing the effects of the blockers on the spike ADP. Each bar depicts the size of the ADP after drug exposure as a percentage of control size. (TTX: n = 7; riluzole: n = 9; CTX-MVIIC: n = 7; CTX-GVIA: n = 8; ATX-IVA: n = 5; Ni²⁺: n = 6; nifedipine: n = 5; 2 mm Mg²⁺: n = 14.) *Statistically significant compared to control (P < 0.05). Error bars represent s.e.m.

In CA1 pyramidal cells, I_NaP is blocked completely by 10 μm riluzole, while I_NaT is only mildly reduced (Yue et al. 2005). At this concentration riluzole had no effect on Ca²⁺ spikes or on Ca²⁺-dependent bursting (Yue & Yaari, 2006), indicating that it does not strongly block Ca²⁺ currents in these neurons (in other types of neurons, 10 μm riluzole reduced Ca²⁺ currents by ∼20%; Huang et al. 1997; Stefani et al. 1997). At higher concentrations (> 30 μm) riluzole enhances the activation of small conductance Ca²⁺-gated K⁺ (SK) channels (Cao et al. 2002). Therefore, in experiments testing riluzole we added 100 nm apamin to the ACSF to block SK channels. Adding 10 μm riluzole to the ACSF reduced the ADP to 64.7 ± 4.4% of control (n = 9; P < 0.05; Fig. 1B), without significantly affecting spike amplitude (97.0 ± 1.4% of control).

Together, these results are consistent with previous reports showing that I_NaP provides a major depolarizing drive for ADP generation (Azouz et al. 1996; Su et al. 2001; Yue et al. 2005).

Calcium currents contribute to spike ADP generation

Next, we tested the sensitivity of the spike ADP to drugs that selectively block one or a small subset of different Ca²⁺ currents expressed by CA1 pyramidal cells. Application of 1 μmω-conotoxin MVIIC to block collectively N-type (I_CaN; McDonough et al. 1996) and P/Q-type (I_CaP/Q) Ca²⁺ currents (Randall & Tsien, 1995), significantly reduced the ADP (68.7 ± 6.3% of control; n = 7; P < 0.05; Fig. 1C). To differentiate between I_CaN and I_CaP/Q we also examined the effects of ω-conotoxin GVIA (I_CaN blocker; McCleskey et al. 1987) and ω-agatoxin IVA (I_CaP/Q blocker; Mintz et al. 1992; Moreno et al. 1997). Spike ADPs were significantly reduced by 1 μmω-conotoxin GVIA (72.6 ± 9.1% of control; n = 8; P < 0.05; Fig. 1D), as well as by 0.2 μmω-agatoxin IVA (75.5 ± 6.8% of control; n = 5; P < 0.05; Fig. 1E), implying that both I_CaN and I_CaP/Q contribute to ADP facilitation. Application of 100 μm Ni²⁺ to block T-type (I_CaT) and R-type (I_CaR) Ca²⁺ currents (Lee et al. 1999; Williams et al. 1999) reduced the ADP to 76.8 ± 8.7% of control (n = 6; P < 0.05; Fig. 1F). Further differentiation between the contributions of I_CaT and I_CaR is hampered by lack of selective and effective blockers. Application of 10 μm nifedipine to block L-type Ca²⁺ current (I_CaL; Fox et al. 1987) slightly reduced the ADP but this effect was not statistically significant (88.3 ± 8.3% of control; n = 5; Fig. 1G). It should be noted, however, that nifedipine and other dihydropyridine antagonists may act too slowly to block Ca²⁺ influx during a single spike (Helton et al. 2005). Therefore, these results do not conclusively exclude involvement of I_CaL in ADP facilitation.

Together, our data show that in addition to I_NaP, most voltage-gated Ca²⁺ currents contribute to ADP electrogenesis. These new data are inconsistent with our previous results showing that 100 μm Ni²⁺ does not affect the spike ADP in normal adult CA1 pyramidal cells (Su et al. 2002; Yue et al. 2005). This discrepancy may be due to the lower Mg²⁺ content of the ACSFs in the present study. A high Mg²⁺ concentration may interfere with Ni²⁺ binding to Ca²⁺ channels, as found in other types of channels (Lotshaw & Sheehan, 1999). We therefore reexamined in a new series of experiments the effects of 100 μm Ni²⁺ on the spike ADP in 2 mm Mg²⁺. Raising the Mg²⁺ from 1 to 2 mm by itself significantly reduced the spike ADP to 81.7 ± 4.6% of control (n = 14; P < 0.05; Fig. 1H). This effect may be due to blocking of Ca²⁺ channels by Mg²⁺, as well as to a depolarizing shift in I_NaP activation due to increased screening of membrane surface charges (Yue et al. 2005). Adding 100 μm Ni²⁺ to ACSF containing 2 mm Mg²⁺ again had no depressant effect on the spike ADP (98.9 ± 5.7% of control; n = 9), suggesting that 2 mm Mg²⁺ occludes the effect of 100 μm Ni²⁺. The effects of riluzole, nifedipine and ω-conotoxin MVIIC on the spike ADP in 2 mm Mg²⁺ ACSF were similar to those described above (data not shown). Figure 1I compares the effects of the various Na⁺ and Ca²⁺ currents blockers on the spike ADP.

Intracellular injection of BAPTA suppresses the spike ADP and occludes the effects of Ca²⁺ channel blockers

Suppression of the spike ADP by Ca²⁺ channel blockers is puzzling, because most Ca²⁺ currents are high voltage activated (threshold ∼−45 mV) and deactivate rapidly (time constant in the order of 0.3 ms; Brown et al. 1993). Therefore, it is unlikely that these currents continue to flow after the spike. Indeed, we have previously shown that following mock spikes, Ca²⁺ tail currents are practically null at potentials negative to −50 mV (Yue et al. 2005). An alternative explanation for these results is that Ca²⁺ currents facilitate the ADP by elevating intracellular Ca²⁺ concentration ([Ca²⁺]_i). To differentiate between direct versus indirect actions of Ca²⁺ currents on the ADP, we examined the effects of injecting the fast Ca²⁺ chelator BAPTA. We reasoned that if Ca²⁺ currents facilitate the ADP by a direct depolarizing action, then BAPTA would augment the ADP by increasing the driving force for Ca²⁺ currents, and perhaps also by preventing activation of opposing Ca²⁺-gated K⁺ currents (Friedman & Gutnick, 1987). On the other hand, if Ca²⁺ currents facilitate the ADP indirectly by a modulatory action of elevated [Ca²⁺]_i, then BAPTA would reduce the ADP.

We injected BAPTA (200 mm in the recording microelectrode) into 23 neurons using 100 ms long negative current pulses (−500 pA) delivered at 3 Hz for 5–10 min. To monitor the progress of cell filling with BAPTA, we also measured spike width and the amplitude of the slow afterhyperpolarization (sAHP; see Methods). Expectedly (Storm, 1987), BAPTA injection caused marked spike broadening (119.4 ± 2.0% of control; P < 0.05) and reduction in sAHP amplitude (21.6 ± 3.6% of control; P < 0.05) in all injected neurons (Fig. 2A). In addition, BAPTA injection consistently reduced the spike ADP size to 72.6 ± 2.8% of control (P < 0.05; Fig. 2A). It also variably affected the apparent input resistance (R_N) of the neurons, the average effect being a mild reduction to 85.3 ± 4.0% of control (P < 0.05; Fig. 2A). These results are summarized in Fig. 2B.

Figure 2. Intracellular injection of BAPTA suppresses the ADP.

A, representative recordings from one neuron showing the effects of intracellularly injected BAPTA on the waveform of solitary spikes (top row), the size of the sAHP (second row from top), the size of the spike ADP (third row from top) and R_N (bottom row). The left, centre and right columns of panels depict recordings obtained in control conditions, recordings obtained after ∼20 min of BAPTA injection, and overlays of expanded traces, respectively. In each panel here and in the following figures, the lower trace shows the injected current pulses and the upper trace depicts the intracellularly recorded voltage responses of the neuron. Injection of BAPTA caused spike broadening, markedly reduced the sAHP that follows repetitive discharge (evoked by a 400 ms stimulus) and markedly suppressed the spike ADP without changing passive membrane responses. B, a bar graph summarizing the effects of intracellularly injected BAPTA on spike width (n = 22), sAHP (n = 23), ADP size (n = 23) and R_N (expressed as a percentage of control; n = 17). Error bars represent s.e.m.C, plot of the BAPTA-induced change in ADP size versus the corresponding change in R_N. The data points represent 17 neurons subjected to BAPTA injection. The line through the data points was obtained by linear regression analysis and had a slope coefficient not significantly different from zero. Thus the BAPTA-induced reduction in ADP size cannot be attributed to changes in R_N. *Statistically significant compared to control (P < 0.05).

Reduction in R_N (i.e. increase in resting conductance) would be expected to shunt the ADP and thus may contribute to its reduction by BAPTA. However, the reduction in ADP size was not correlated with the reduction in R_N (r = 0.07; n = 17; P > 0.05). Likewise, analysis of the changes in spike ADP versus changes in R_N failed to disclose a significant trend (Fig. 2C). Indeed, in three neurons BAPTA injection caused an increase in R_N, yet the spike ADPs were markedly reduced (Fig. 2C). Therefore, we conclude that the BAPTA-induced suppression of the ADPs is not due to changes in R_N.

We found that the amount of BAPTA-induced ADP reduction was positively correlated with the size of the native spike ADP (r = 0.6; n = 23; P < 0.05; Fig. 3A). Moreover, injection of BAPTA reduced the variance in ADP size across the neurons. The distributions of ADP sizes before and after BAPTA injection for 23 neurons are shown in Fig. 3B. The two distributions were fitted with a Gaussian function yielding mean ±s.d. values of 318.6 ± 87.5 mV ms and 224.6 ± 50.6 mV ms for the ADPs before and after BAPTA injection, respectively.

Figure 3. Intracellular BAPTA alters the distribution of spike ADP sizes.

A, plot of the amount of BAPTA-induced ADP reduction versus the control ADP size. The data points represent 23 neurons subjected to BAPTA injection. The line through the data points was obtained by linear regression analysis and has a slope coefficient of 0.091 (significantly different from zero). The implication of this trend is that larger ADPs are more strongly suppressed by BAPTA. B, distributions of ADP sizes in a group of 21 neurons before (left panel; Control) and after BAPTA injection (right panel; BAPTA). Dashed line in right panel denotes the ADP size distribution in control to facilitate comparison.

The hypothesis that Ca²⁺ spike entry enhances the ADP via [Ca²⁺]_i increase further predicts that BAPTA injection would occlude the effects of Ca²⁺ channel blockers. We tested this prediction by monitoring the effects of Ni²⁺ (100 μm) and ω-conotoxin MVIIC (1 μm) on the ADP in BAPTA-injected neurons. Congruent with this hypothesis, neither blockers significantly affected ADP size after it was reduced by BAPTA injection. Thus, ADP size was reduced to 67.8 ± 5.9% of control by BAPTA and remained 66.1 ± 7.3% of control even after 30 min of treatment with Ni²⁺ (n = 5; Fig. 4A and D). Likewise, ADP size was reduced to 81.3 ± 8.8% of control by BAPTA and remained 82.7 ± 9.0% of control even after 30 min of treatment with ω-conotoxin MVIIC (n = 5; Fig. 4B and D). In contrast, application of 10 μm riluzole further reduced ADP size from 70.2 ± 9.8% to 43.0 ± 7.4% of control (n = 5; P < 0.05; Fig. 4C and D).

Figure 4. Spike ADP suppression by Ca²⁺ channel blockers is prevented by intracellular BAPTA.

A, effects of Ni²⁺. The spike ADP in a BAPTA-injected neuron (left panel; BAPTA) was only slightly reduced by 30 min exposure to 100 μm Ni²⁺ (middle panel; + Ni²⁺). Portions of the traces are expanded and overlaid in right panel to facilitate comparison. B, effects of ω-conotoxin MVIIC. The spike ADP in another BAPTA-injected neuron (left panel; BAPTA) was not affected by 30 min exposure to 1 μmω-conotoxin MVIIC (middle panel; + CTX MVIIC). Portions of the traces are expanded and overlaid in right panel to facilitate comparison. C, effects of riluzole. The spike ADP in yet another BAPTA-injected neuron (left panel; BAPTA) was suppressed by 30 min exposure to 10 μm riluzole (middle panel; + Riluzole). Portions of the traces are expanded and overlaid in right panel to facilitate comparison. D, bar graph summarizing the effects Ni²⁺, ω-conotoxin MVIIC (CTX) and riluzole on the spike ADP in BAPTA-injected neurons. Each bar depicts the size of the spike ADP relative to control (before BAPTA injection). (Ni²⁺: n = 5; CTX: n = 5; riluzole: n = 5.) *In all conditions ADP size was significantly smaller than in control (before injection of BAPTA; P < 0.05). †Only riluzole caused a further significant decrease in ADP size (P < 0.05). Error bars represent s.e.m.

Blocking K_V7/M channels prevents spike ADP suppression by Ca²⁺ channel blockers

How can the increase in [Ca²⁺]_i during a spike facilitate the ADP? Given that ADP size largely depends on the balance between I_NaP and I_M (Yue & Yaari, 2004, 2006; Yue et al. 2005; Golomb et al. 2006), the two mechanisms to be considered are enhancement of persistent Na⁺ channels and/or inhibition of K_V7/M channels. The effects of [Ca²⁺]_i on persistent Na⁺ channels have not been investigated, but previous studies have shown that K_V7/M channels are strongly inhibited by [Ca²⁺]_i above ∼50 nm (Selyanko & Brown, 1996; Selyanko & Sim, 1998). Therefore, we hypothesized that ADP facilitation by spike Ca²⁺ influx is caused by intracellular Ca²⁺ inhibition of I_M. A key prediction of this hypothesis is that after blocking I_M pharmacologically (which by itself would facilitate the ADP; Yue & Yaari, 2004), the ADP will no longer be suppressed by Ca²⁺ current blockers or intracellular BAPTA.

To test the latter prediction, we pretreated neurons with the selective I_M blocker linopirdine (Aiken et al. 1995; Schnee & Brown, 1998; Wang et al. 1998) for at least 50 min to obtain maximal effect. As previously described (Yue & Yaari, 2004), 10 μm linopirdine facilitated the ADP to the point of bursting (n = 17; see Fig. 5A, C and E). The appearance of bursting hindered the measurement of the first spike ADP. To unmask the ADP, we applied a brief (4 ms) negative current pulse immediately after the positive current pulse that initiated the burst response (Jensen et al. 1996). The strength of the second pulse was carefully raised until the burst response was suppressed in about 50% of the trials (see Fig. 5A, C and E). This procedure disclosed a large ADP that was 184.9 ± 9.6% of control (n = 17; P < 0.05).

Figure 5. Spike ADP suppression by Ca²⁺ channel blockers is prevented by blocking KV7/M channels.

A, effects of Ni²⁺. The spike ADP (top left panel; Control) was facilitated to the point of bursting (5 spikes) by exposure to 10 μm linopirdine (top middle panel; Linopirdine). Adding 100 μm Ni²⁺ to the ACSF exerted no effect on the burst (top right panel; + Ni²⁺). Likewise, the spike ADP underlying linopirdine-induced bursting, unmasked by appropriately placed brief negative current pulses (bottom second from left panel; Linopirdine), was unaffected by Ni²⁺ (bottom third from left panel; + Ni²⁺). The traces of the unmasked ADPs are expanded and overlaid in bottom right panel to facilitate comparison. B, plot of the time course of changes in ADP size following exposure to linopirdine and Ni²⁺. Linopirdine was applied at time zero. The dashed vertical line denotes the beginning of Ni²⁺ application. Each point represents the mean ±s.e.m. of the values obtained in 6 similar experiments. C, effects of ω-conotoxin MVIIC. In another neuron, the spike ADP (top left panel; Control) was facilitated to the point of bursting (5 spikes) by exposure to 10 μm linopirdine (top middle panel; Linopirdine). Adding 1 μmω-conotoxin MVIIC to the ACSF exerted no effect on the burst (top right panel; + CTX MVIIC). Likewise, the spike ADP underlying linopirdine-induced bursting, unmasked by appropriately placed brief negative current pulses (bottom second from left panel; Linopirdine), was unaffected by the toxin (bottom third from left panel; + CTX MVIIC). The traces of the unmasked ADPs are expanded and overlaid in bottom right panel to facilitate comparison. D, plot of the time course of changes in ADP size following exposure to linopirdine and ω-conotoxin MVIIC. Linopirdine was applied at time zero. The dashed vertical line denotes the beginning of ω-conotoxin MVIIC application. Each point represents the mean ±s.e.m. of the values obtained in 6 similar experiments. E, effects of riluzole. In yet another neuron, the spike ADP (top left panel; Control) was facilitated to the point of bursting (6 spikes) by exposure to 10 μm linopirdine (top middle panel; Linopirdine). Adding 10 μm riluzole to the ACSF converted the burst response to a single spike (top right panel; + Riluzole). Likewise, the spike ADP underlying linopirdine-induced bursting, unmasked by appropriately placed brief negative current pulses (bottom second from left panel; Linopirdine), was markedly suppressed by the drug (bottom third from left panel; + Riluzole). The traces of the unmasked ADPs are expanded and overlaid in bottom right panel to facilitate comparison. F, plot of the time course of changes in ADP size following exposure to linopirdine and riluzole. Linopirdine was applied at time zero. The dashed vertical line denotes the beginning of riluzole application. Each point represents the mean ±s.e.m. of the values obtained in 5 similar experiments. G, bar graph summarizing the effects of the three drugs on the spike ADP in linopirdine-treated neurons. Each bar depicts the size of the spike ADP relative to control (before adding linopirdine). (Ni²⁺: n = 6; CTX: n = 6; riluzole: n = 5.) *In all conditions ADP size was significantly smaller than in control (before adding linopirdine; P < 0.05). †Only riluzole caused a further significant decrease in ADP size (P < 0.05). Error bars represent s.e.m.

Exposing linopirdine-treated neurons to 100 μm Ni²⁺ did not modify the number of intraburst spikes (5.7 ± 0.3, before and after addition of Ni²⁺; n = 6; Fig. 5A). Moreover, the size of the unmasked ADP (207.4 ± 16.2% of control in drug-free ACSF) was not significantly affected by the addition of Ni²⁺ (226.5 ± 25.5% of control in drug-free ACSF; n = 6; Fig. 5A). The time course of the effects of linopirdine and Ni²⁺ on the size of the unmasked ADP are depicted in Fig. 5B(n = 6). Exposing linopirdine-treated neurons to 1 μmω-conotoxin MVIIC also did not modify the number of intraburst spikes (5.0 ± 0.8 versus 5.5 ± 1.0, before and after addition of ω-conotoxin MVIIC; n = 6; Fig. 5C). Likewise, the size of the unmasked ADP (162.8 ± 16.2% of control in drug-free ACSF) was not significantly affected by the toxin (165.5 ± 18.8% of control in drug-free ACSF; n = 6; Fig. 5C). The time course of the effects of linopirdine and Ni²⁺ on the size of the unmasked ADP is depicted in Fig. 5D (n = 6).

For comparison, we also tested how blocking I_NaP affects bursting and the unmasked ADP in linopirdine-treated neurons. As previously shown (Yue & Yaari, 2006), adding 10 μm riluzole to the ACSF converted the burst responses (mean number of intraburst spikes 6.2 ± 0.4; n = 5) to solitary spikes in all tested neurons (P < 0.05; Fig. 5E). Furthermore, the size of the unmasked ADP in linopirdine (184.4 ± 12.7% of control in drug-free ACSF) was significantly reduced by riluzole (90.2 ± 8.8% of control in drug-free ACSF; P < 0.05; Fig. 5E). The time course of the effects of linopirdine and riluzole on the size of the unmasked spike ADP is depicted in Fig. 5F(n = 5).

The results of the experiments with linopirdine-treated neurons are summarized in Fig. 5G. They support the notion that spike Ca²⁺ entry facilitates the ADP by intracellular inhibition of I_M.

Blocking K_V7 channels prevents spike ADP suppression by intracellular BAPTA

Another prediction of the latter hypothesis is that injection of BAPTA into neurons pretreated with linopirdine to block I_M also would not suppress the spike ADP. We tested this prediction in five neurons pretreated with 10 μm linopirdine for at least 50 min. All of these neurons converted to bursting mode, but the underlying ADP could be unmasked by delivering brief negative current pulses, as described above. Injection of BAPTA into linopirdine-treated neurons consistently caused marked spike broadening (117.4 ± 2.1% of control; n = 5; P < 0.05; Fig. 6) and interfered with repolarization of the intraburst spikes (Fig. 6). In congruence with the above hypothesis, the unmasked ADP was not reduced by BAPTA injection. Indeed, it was modestly enhanced in 4 of 5 neurons, perhaps due to reduced activation of Ca²⁺-gated K⁺ channels in the presence of BAPTA. However, the overall change was not statistically significant (121.2 ± 10.5% of the control ADP measured before BAPTA injection; Fig. 6).

Figure 6. Effects of BAPTA injection on the spike ADP in linopirdine-treated neurons.

Representative recordings from one neuron superfused with ACSF containing 10 μm linopirdine, which converted it into bursting mode (6 spikes). Left (linopirdine) and middle panels (+ BAPTA) show recordings before and after BAPTA injection, respectively. Traces in right panels are overlays of expanded traces in left and middle panels. BAPTA consistently caused marked broadening of the first spike (top row of panels) and reduced repolarization during the burst, transforming the burst into a plateau depolarization (middle row of panels). In addition to that, BAPTA injection resulted in an increase in the size of the large unmasked ADP (bottom row of panels).

Spike Ca²⁺ entry enables I_NaP-driven bursting by intracellular inhibition of K_V7/M channels

A fraction (15 of 117; 13%) of the sampled CA1 pyramidal cells were low-threshold bursters, generating two to three spikes in response to brief stimuli (see Fig. 7). Low-threshold bursters were seldom encountered in previous studies (Jensen et al. 1994; Su et al. 2001). We attribute the difference to the fact that in those studies the ACSF contained 2 mm Mg²⁺, causing the spike ADPs to be considerably smaller than in 1 mm Mg²⁺ ACSF (Fig. 1H). We speculated that native bursting reflects effective inhibition of I_M by spike Ca²⁺ influx, which in turn facilitates the spike ADP by unleashing I_NaP (Yue & Yaari, 2006). In line with this notion we found that in four bursters (2.1 ± 0.1 intraburst spikes), adding 1 μmω-conotoxin MVIIC to the ACSF converted the bursts to solitary spikes in all cases (Fig. 7A). Likewise, in another six bursters (2.2 ± 0.2 intraburst spikes), injecting BAPTA also consistently suppressed bursting (Fig. 7B). Finally, in four bursters (2.3 ± 0.3 intraburst spikes), adding 10 μm riluzole to the ACSF suppressed bursting in all cases (Fig. 7C). Together, these data show that native low-threshold bursting in adult CA1 pyramidal cells requires I_M inhibition by spike Ca²⁺ influx for its manifestation.

Figure 7. Spike Ca²⁺ entry enables I_NaP-driven bursting by intracellular inhibition of K_V7/M channels.

A, effects of ω-conotoxin MVIIC in a natively bursting neuron (left panel; Control). Exposure of the neuron to 1 μmω-conotoxin MVIIC converted the burst to a single spike (middle panel; CTX MVIIC). Traces in right panel are an overlay of expanded traces in left and middle panels. B, effects of BAPTA in another natively bursting neuron (left panel; Control). Injection of BAPTA converted the burst to a single spike (middle panel; BAPTA). Traces in right panel are an overlay of expanded traces in left and middle panels. Note that BAPTA also eliminated the fast AHP. C, effects of riluzole in yet another natively bursting neuron (left panel; Control). Exposure of the neuron to 10 μm riluzole converted the burst to a single spike (middle panel; Riluzole). Traces in right panel are an overlay of expanded traces in left and middle panels.

Discussion

In this study we confirm that the waveform of the spike ADP in adult CA1 pyramidal cells is to a large extent determined by interplay between I_NaP and I_M (Yue & Yaari, 2004, 2006; Yue et al. 2005; Golomb et al. 2006). Furthermore, we show for the first time that spike Ca²⁺ influx markedly facilitates the I_NaP-driven ADP, sometimes to the point of bursting, by intracellular inhibition of K_V7/M channels. The latter conclusion is based on three key findings. First, application of various Ca²⁺ current blockers suppresses the ADP. Second, intracellular injection of BAPTA suppresses the ADP. It also occludes the action of Ca²⁺ current blockers, but not that of an I_NaP blocker. Third, blocking K_V7/M channels facilitates the ADP. Yet it prevents ADP suppression by Ca²⁺ current blockers and by BAPTA, though not by an I_NaP blocker.

Inhibition of K_V7/M channels by spike Ca²⁺ influx

Several studies have shown that K_V7/M channels are modulated by changes in [Ca²⁺]_i (Delmas & Brown, 2005). In both sympathetic neurons (Selyanko & Brown, 1996) and hippocampal pyramidal cells (Selyanko & Sim, 1998), application of Ca²⁺ to the cytoplasmic face of the membrane produced a rapid and dose-dependent (EC₅₀= 100–120 nm) inhibition of K_V7/M channel activity. Furthermore, activity of K_V7/M channels in depolarized membrane patches (recorded in cell-attached configuration) was strongly inhibited following a single spike; this was prevented in Ca²⁺-free ACSF, implying involvement of Ca²⁺ influx (Selyanko & Sim, 1998). The fact that Ca²⁺ inhibition of K_V7/M channel activity persisted in the absence of ATP and GTP and was reversible upon removal of Ca²⁺ suggests that it is mediated by direct interaction with the K_V7/M channel complex, rather than by phosphorylation/dephosphorylation processes (Selyanko & Brown, 1996). Recent studies propose that calmodulin bound to the C-terminal domain of K_V7 subunits mediates Ca²⁺ inhibition of K_V7/M channel activity (Yus-Najera et al. 2002; Gamper & Shapiro, 2003; Gamper et al. 2005; Shahidullah et al. 2005).

In CA1 pyramidal cells resting [Ca²⁺]_i is about 60 nm (Jaffe et al. 1992). To obtain a significant inhibition of K_V7/M channels that would profoundly facilitate the spike ADP (> 50%; Yue & Yaari, 2004), [Ca²⁺]_i should increase to 100 nm or more (Selyanko & Brown, 1996). A single spike increases bulk [Ca²⁺]_i in the soma by only ∼2 nm (Jaffe et al. 1992), but the submembrane concentration of Ca²⁺ must be well over 100 nm. This is inferred from the fact that spike Ca²⁺ influx readily activates large conductance Ca²⁺-gated K⁺ (BK) channels, whose Ca²⁺ sensitivity is much smaller than that of K_V7/M channels (EC₅₀= 2–4 μm; Smart, 1987; Franciolini, 1988; Gong et al. 2001). Thus, to be inhibited by a single spike, K_V7/M channels also must sense the [Ca²⁺]_i within the submembrane domain.

Close inspection of the effects of Ca²⁺ channel blockers reveals that they suppress the ADP already at its onset (Fig. 1), implying that K_V7/M channel inhibition must begin within 1–2 ms after spike Ca²⁺ influx. Such a rapid action requires proximity of the sources of Ca²⁺ to the K_V7/M channels controlling the somatic spike ADP, which are distributed at or near the soma (Yue & Yaari, 2006). Congruently, there is compelling evidence for expression of all types of Ca²⁺ channels in somata and proximal apical dendrites of adult CA1 pyramidal cells (Mills et al. 1994; Magee & Johnston, 1995; Marrion & Tavalin, 1998; Su et al. 2002). However, our data so far do not indicate a preferential association of K_V7/M channels with a particular type of Ca²⁺ channel, as has been described in these neurons for BK and N-type Ca²⁺ channels (Marrion & Tavalin, 1998). Moreover, these data suggest that spike Ca²⁺ influx through either N- or P/Q-type channels is sufficient to obtain the maximal possible inhibition of K_V7/M channels by a single spike, because blocking either N- or P/Q-type Ca²⁺ channels suppressed the spike ADP by the same amount (∼30%) as blocking both of them together (Fig. 1).

Functional significance

The two actions of Ca²⁺ spike entry, namely, K_V7/M channel inhibition and BK channel activation, appear at first glance to exert opposite effects on excitability. However, they actually complement each other in controlling spike output. Facilitation of the spike ADP by K_V7/M channel inhibition increases the propensity for high-frequency bursting at the onset of a depolarization (Yue & Yaari, 2004). Likewise, enhancement of spike repolarization by BK channel activation increases early spike frequency by limiting K⁺ channel activation and Na⁺ channel inactivation (Gu et al. 2007). Thus, the two intracellular modulatory actions of Ca²⁺ spike entry synergistically promote the bursting behaviour of the neuron.

Recordings from CA1 pyramidal cells in vivo indicate that the propensity to burst and the characteristics of individual bursts (e.g. number and frequency of intraburst spikes) vary with the behavioural state of the animal (Ranck, 1973; Harris et al. 2001; Tropp Sneider et al. 2006). These state-dependent alterations may be imposed by release of neurotransmitters that modulate key currents producing the spike ADP, such as I_NaP and I_M. Our results suggest that neurotransmitter modulation of Ca²⁺ currents (Catterall, 2000) may also contribute to these alterations by affecting the Ca²⁺-mediated inhibition of K_V7/M channels. This mechanism may also contribute to the anti-epileptic efficacy of Ca²⁺ channel blockers (White et al. 2007), given that bursting neurons are essential for various forms of epileptiform discharges (Yaari & Beck, 2002).