- Email: [email protected]
Characterization of rebound depolarization in hippocampal neurons
BBRC Biochemical and Biophysical Research Communications 348 (2006) 1343–1349 www.elsevier.com/locate/ybbrc
Characterization of rebound depolarization in hippocampal neurons Rainer Surges *, Monika Sarvari, Marc Steffens, Thomas Els Department of Neurology and Clinical Neurophysiology, University Clinics of Freiburg, Germany Received 26 July 2006 Available online 7 August 2006
Abstract Rebound depolarization (RD) following hyperpolarizing pulses is found in several neuronal cell types where it takes part in the regulation of neuronal firing behavior. During whole-cell current and voltage clamp recordings in slice preparations, we investigated the modulation of RD by different stimulation patterns and its underlying ionic currents in rat CA1 pyramidal cells. RD was mainly carried by the hyperpolarization-activated cation current Ih (about two-third) and T-type calcium currents (about one-third), respectively. RD increased with increasing hyperpolarizing amplitude and stimulation frequency, whereas RD substantially decreased with longer pulse duration and, less pronounced, with increasing pulse number. The pulse duration-related decrease of RD was due to a decrease of the driving force of Ih. In conclusion, we showed that RD is differentially modulated by precedent hyperpolarization. Since RD amplitude was high enough to generate action potentials, RD may serve, even under physiologic conditions, as an inhibition-excitation converter. 2006 Elsevier Inc. All rights reserved. Keywords: CA1; Ih; HCN channels; T-type calcium channels; Postinhibitory rebound depolarization; Hyperpolarization-activated cation current; Epilepsy; ZD7288
Neurons of different brain regions, e.g., within the thalamus, cerebellum, and hippocampus [1–3], display a transient depolarization following hyperpolarizing pulses, a phenomenon called rebound depolarization (RD). RD contributes, for example, to the regulation of the neuronal firing mode in cerebellum [1] and to oscillations in thalamocortical networks [3]. RD is, at least in the cell types mentioned above, substantially carried by the hyperpolarization-activated cation current (Ih). Ih is a mixed, slowly developing inward current consisting of Na+- and K+-ions that is activated by hyperpolarization more negative than 60 mV and which does not inactivate during sustained hyperpolarization. Upon repolarization Ih slowly deactivates, leading to an inward current producing RD. In hippocampal CA1 pyramidal cells, the hyperpolarization activated, cyclic nucleotide-gated (HCN) channels underlying Ih are expressed on dendrites as well as on the somatic cell membrane with a sevenfold higher current density on distal dendrites where it
*
Corresponding author. Fax: +49 761 2705281. E-mail address: [email protected] (R. Surges).
0006-291X/$ - see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2006.07.193
takes part in spatial and temporal integration of synaptic input [4]. However, Ih is of robust amplitude on neuronal cell soma and participates in the regulation of intrinsic membrane properties such as resting potential, input resistance [5], and resonance behavior [6]. CA1 neurons receive inhibitory signals from diverse interneurons that project onto dendrites as well as to the somatic and perisomatic cell membrane [e.g., 7–9] where they can activate somatic Ih [2]. Previously, RD was shown to be essentially involved in the enhanced excitability of hippocampal circuitry in a rat model of febrile seizures [2,8], suggesting a considerable role of RD in the regulation of neuronal firing behavior. We therefore characterized the modulation of RD and its underlying ionic currents in CA1 pyramidal cells of rat hippocampal slices using the whole-cell patch clamp technique. Materials and methods Slice preparation. Hippocampal slices were prepared from 14 to 22 postnatal (P) days old Wistar rats of either sex. Experiments were performed according to national and institutional guidelines. Rats were decapitated, the brains were rapidly removed, hemisected, and submerged in ice-cold
1344
R. Surges et al. / Biochemical and Biophysical Research Communications 348 (2006) 1343–1349
physiological Ringer’s solution (Biometra, Go¨ttingen, Germany) which was continuously bubbled with 95% O2 and 5% CO2 and contained (in mM): NaCl, 125; NaHCO3, 25; KCl, 2.5; NaH2PO4, 1.25; MgCl2, 1; glucose, 25; CaCl2, 2. The hemispheres were blocked in the horizontal plane and glued to the stage of a vibratome (Leica VT 1000S, Wetzlar, Germany). Horizontal sections through the hippocampus (350–400 lm thick) were cut, and slices were transferred to a holding chamber, where they were submerged in Ringer’s solution. This solution was continuously bubbled with 95% O2 and 5% CO2 and was maintained at 35 C for the first 30 min and then at room temperature (20–22 C). Slices were kept in the holding chamber for at least 1 h before recordings were performed. Electrophysiology. Individual slices were transferred into the recording chamber, fixed with a nylon-grid and continuously perfused with oxygenated Ringer’s solution at room temperature (20–22 C). Neurons were visualized by infrared differential interference contrast video-microscopy with a Newvicon camera (C2400, Hamamatsu, Hamamatsu City, Japan) and an infrared filter (RG9, Schott, Mainz, Germany) to an upright microscope (Axioskop 2 FS, Zeiss, Oberkochen, Germany) equipped with a 40· water immersion objective. Recordings at room temperature (20– 22 C) were made in the whole-cell configuration of the patch-clamp technique using an Axopatch 200B amplifier and pClamp 8 (Axon Instruments, Foster City, CA, USA). For current clamp recordings, signals were low-pass filtered at 1 or 10 kHz and sampled at 2.5, 3.3 or 50 kHz, and during voltage-clamp experiment signals were low-pass filtered at 1 kHz and sampled at 5 kHz, respectively (Digidata 1200, Axon Instruments). Pipettes were pulled from non-filamented borosilicate glass capillaries (Hilgenberg, Malsfeld, Germany; outer diameter 1.5 mm, inner diameter 1.0 mm) on a P-97 Flaming/Brown horizontal puller (Sutter Instruments, Novato, CA, USA) and had a mean resistance of 4.8 ± 0.1 MX (n = 29) with the pipette solution. Standard pipette solution contained (in mM): Kgluconate, 120; KCl, 20; MgCl2, 2, CaCl2, 1; Hepes, 10; EGTA, 10; ATP– Tris, 2; pH 7.35, adjusted with KOH. The offset potential between the patch-pipette and the reference electrode was zeroed before the tight-seal (>2 GX) was established. Series resistance was determined by using the ‘‘membrane-test’’ protocol provided by the software pClamp 8.0 and amounted to 13–20 MX (compensation 40–50% in the voltage clamp experiments). A liquid junction potential of +13.1 mV was calculated and all values given are corrected for this voltage offset so that without correction the voltages given in this paper would be 13 mV more positive. Solutions and chemicals. The standard bath-solution consisted of (in mM): NaCl, 120; NaHCO3, 25; KCl, 2.5; MgCl2, 1; glucose, 25; CaCl2, 2. In most current clamp experiments, 20 lM bicuculline and 10 lM NBQX (Tocris, Biotrend Chemikalien, Ko¨ln, Germany) were added to the bath solution to block spontaneous synaptic events. To isolate Ih during voltage-clamp experiments, the following substances were added (in mM): Tetrodotoxin (TTX; Tocris, Biotrend Chemikalien, Ko¨ln, Germany), 0.005–0.01; NiCl2, 1; TEA, 10; 4-AP, 2; BaCl2, 0.5. ZD7288 [4-(N-ethyl-Nphenylamino)-1,2-dimethyl-6-(methylamino) pyrimidium chloride] (Tocris) was dissolved in H2O in a stock solution of 25 mM. CsCl was dissolved in H2O in a stock solution of 100 mM. Unless indicated otherwise, drugs were purchased from Sigma (Taufkirchen, Germany). Analysis. Data were acquired with the help of pCLAMP 8.0 and analyzed offline using the software packages Clampfit 8.0 (Axon Instruments) and GraphPad PRISM 2.0 (GraphPad Software Inc., San Diego, CA, USA). Linear regression analysis to obtain the reversal potential of Ih was performed with GraphPad PRISM. Statistical significance (p < 0.05) was assessed, if not indicated otherwise, using the two-tailed paired student’s t test. All data were expressed as means ± SEM.
ing potential (RP) was determined immediately after establishing the whole-cell mode and amounted to 71 ± 1 mV. At the onset of each experiment, a family of current pulses was applied in the current clamp mode. Prolonged depolarizing current injection evoked a train of action potentials (AP) displaying spike adaptation (not shown), a behavior that is typical for CA1 pyramidal neurons [e.g., 10,11]. During hyperpolarizing current injections, a slowly developing, pronounced ‘voltage sag’ occurred (see Fig. 1A, left panel), indicative of the presence of Ih. Upon repolarization to RP, one can observe a substantial transient depolarization called rebound depolarization (RD; see Fig. 1A, left panel). To examine the influence of precedent hyperpolarization on RD amplitude, the following characteristics of the hyperpolarizing pulses were varied: (a) amplitude, (b) duration, (c) pulse number, and (d) frequency. To avoid AP’s due to RD, TTX (0.5 lM) was added to the bath solution or alternatively, a slight hyperpolarizing current was constantly delivered via the patch clamp amplifier (in order to increase the difference between membrane potential and AP-threshold). First, hyperpolarizing current pulses (2 s) of different amplitudes were injected into the neurons starting from the resting potential (or 1–2 mV more negative due to constantly applied negative current injections via the amplifier). With increasing current amplitude, RD amplitude substantially increased from 1.2 ± 0.3 mV to 9.6 ± 1.6 mV (Fig. 1A, n = 8). Next, hyperpolarizing pulses ( 200 pA) of different duration ranging from 1 to 10 s were administered. Surprisingly, we observed a small but robust decline in the RD amplitude with increasing pulse duration, from 9.1 ± 1.0 mV at 1 s to 7.3 ± 1.0 mV at 10 s (Fig. 1B, n = 8). Upon variation of the pulse number (from 5 to 45 pulses), RD amplitude was only slightly affected within the given conditions with a relative maximum of RD about 8.6 ± 0.5 mV following 10 pulses and a tendency to decline upon higher pulse numbers to 8.1 ± 0.4 at 45 pulses (Fig. 1C, n = 6). Finally, we delivered 30 pulses (120 ms; 200 pA) at different stimulation frequencies ranging from 1 to 8 Hz (Fig. 1D). The relative increase in RD was highest at 2 Hz as compared to 1 Hz (RD at 1 Hz 4.5 ± 0.8 mV to 6.0 ± 0.6 mV at 2 Hz, p < 0.05) reaching a plateau between 2 and 6 Hz with a further increase at 7 Hz (Fig. 1D, right panel, n = 8). At a shorter pulse duration (60 ms) and higher frequencies, we even observed a 3.6-fold increase of the RD amplitude at 16 Hz (RD amplitude 5.5 ± 1.2 mV) as compared to 1 Hz (RD amplitude 1.5 ± 0.3 mV, n = 4, p < 0.05; Fig. 1D, right panel inset).
Results
Generation of AP’s upon RD is decreased by increasing duration of the precedent hyperpolarization
RD amplitude depends on the characteristics of the precedent hyperpolarization The present study includes information from 29 visually selected cells within the CA1 pyramidal cell layer. The rest-
In the previous experiments the generation of AP was inhibited by pharmacologic or electrophysiologic means. In this section, we investigated the dependence of AP on the amplitude and duration of the precedent
R. Surges et al. / Biochemical and Biophysical Research Communications 348 (2006) 1343–1349
1345
hyperpolarization. We first applied 2 s hyperpolarizing currents and recorded the membrane potential in the absence of TTX. In analogy to the dependence of the RD amplitude on the current amplitude, the AP number (or the AP probability) also increased with increasing current amplitude of the precedent hyperpolarization (Fig. 2A; n = 9). Upon variation of the pulse duration, we observed a reduction of the AP number from 1 at a 1 s pulse to about 0.6 ± 0.2 AP’s following a pulse of 10 s, in accordance with the pulse duration-related decrease of RD amplitude (Fig. 2B; n = 5). Nature of RD: Ih and T-type calcium channels are both involved To assess the nature of the underlying channel activity of RD, the changes of RD amplitudes as well as of sag amplitudes (displaying an estimation of Ih channel activation) were recorded in the presence of different channels blockers. We first applied Ni2+ (0.1 mM) to block T-type calcium channels, leading to a decrease of RD amplitude by 36% from 6.9 ± 1.0 mV to 4.4 ± 0.3 mV (Fig. 3A; n = 4; p < 0.05) without alterations of the sag amplitude. Additional application of the Ih channel blocker ZD7288 (50 lM) nearly abolished the RD amplitude by 94% as compared to control to 0.4 ± 0.4 mV (n = 4; p < 0.005) with an accompanying reduction in sag amplitude by
Fig. 1. Modulation of RD amplitude by different stimulation patterns. (A) In current clamp mode, hyperpolarizing currents of 2 s duration with different amplitudes (from 30 to 360 pA in 30 pA-steps, 0.15 Hz) were injected (left panel: original recordings; grey trace 120 pA; black trace 240 pA; sag and RD as indicated) and the resulting RD amplitudes were plotted vs. the corresponding current amplitudes (right panel; n = 8). Scaling 25 mV/1 s, RP 71 mV. (B) Pulses ( 200 pA) of different duration ranging from 1 to 10 s were administered (left panel: original recordings) and corresponding RD amplitudes were plotted vs. the pulse duration (right panel, n = 8). Scaling 25 mV/2 s, RP 69 mV. (C) Different number of identical pulses ( 200 pA, 120 ms, 4 Hz) were applied (left panel: original recordings with stimulation by 10 pulses) and the RD amplitudes following the last pulse were plotted vs. the pulse no. (right panel; n = 6). Scaling 25 mV/1 s, RP 66 mV. (D) Thirty identical pulses ( 200 pA, 120 ms) were applied at different stimulation frequencies (left panel: original recordings at 2 Hz) and RD amplitudes after the last pulse were plotted vs. the stimulation frequency (right panel; n = 8). Inset shows RD amplitudes at shorter pulse duration (60 ms) and higher stimulation frequency (up to 16 Hz; n = 4). Scaling 25 mV/5 s, RP 69 mV.
Fig. 2. Number of action potentials upon RD is decreased by increasing duration of hyperpolarization. (A) Two-second pulses of different amplitude were applied in the absence of TTX (left panel: original recordings; grey trace 120 pA; black trace 240 pA) and AP no. upon RD was plotted vs. the current amplitude (right panel, n = 9). Scaling 25 mV/1 s, RP 78 mV. (B) Pulses ( 200 pA) of different duration ranging from 1 to 10 s were administered (left panel: original recordings) and corresponding AP no. was plotted vs. the pulse duration (right panel, n = 5). Scaling 25 mV/2 s, RP 65 mV.
1346
R. Surges et al. / Biochemical and Biophysical Research Communications 348 (2006) 1343–1349
administered Cs+ (5 mM) to block Ih channels with similar results (Fig. 3C). Especially upon wash-out of ZD7288, both the sag amplitude as well the RD amplitude only partially recovered, as known by previously published voltageclamp experiments (e.g., [13]; data not shown). Note that blockade of Ih channels substantially increases the neuronal input resistance with an additional slight hyperpolarization of the membrane potential [5]. Therefore, in order to correct for errors due to higher voltage deflections upon the same current amplitude (because of Ih channel blockade in the presence of ZD7288 or Cs+, respectively), only the traces of similar initial voltage response to the current injection were compared. For instance, the RD amplitudes under control conditions upon a 200 pA pulse were compared to the RD amplitudes during ZD7288-application upon a 150 pA pulse. The duration-related inhibition of RD is due to Ih channel activity
Fig. 3. Both T-type calcium currents and Ih contribute to RD. (A) Hyperpolarizing pulses ( 150 to 200 pA) were applied under control conditions, in the presence of 0.1 mM Ni2+ alone and during additional application of 50 lM ZD7288 (left panel: original recordings; upper black trace control, grey trace under Ni2+ application; lower black trace in the presence of Ni2+ and ZD7288) and the resulting sag/RD amplitudes were compared under the different conditions (black bars: control; grey bars: Ni2+; white bars: Ni2++ZD7288, recordings from the same 4 cells). RP 78 mV. (B) Hyperpolarizing pulses ( 150 to 200 pA) were applied under control conditions and during application of 50 lM ZD7288 (left panel: original recordings; upper black trace control, grey trace in the presence of ZD7288) and the resulting sag/RD amplitudes were compared (black bars: control, grey bars: ZD7288, n = 4). RP 73 mV. (C) Hyperpolarizing pulses ( 150 to 200 pA) were applied under control conditions and during application of 5 mM Cs+ (left panel: original recordings; upper black trace control, grey trace in the presence of Cs+) and the resulting sag/RD amplitudes were compared (black bars: control, grey bars: Cs+, n = 4). RP 80 mV. Note that because of possible errors due an increase of input resistance in the presence of ZD7288/Cs+, the control values (at a 200 pA pulse) were compared to traces upon lower current injections that displayed similar initial voltage responses (e.g., at 150 pA pulses). Statistical significance as indicated by asterisks. Scaling for all recordings 25 mV/1 s.
80% as compared to control from 9.8 ± 1.8 mV to 1.9 ± 0.4 mV (Fig. 3A; n = 4, p < 0.02). Application of ZD7288 (50 lM) alone reduced the sag amplitude by 72% from 13.6 ± 1.9 mV to 3.7 ± 1.1 mV (n = 4, p < 0.02), and the RD amplitude by 65% from 10.8 ± 0.7 mV to 3.8 ± 1.2 mV (n = 4, p < 0.01). As ZD7288 may also block neuronal T-type calcium channels [12], we additionally
As shown above, RD amplitude was reduced upon sustained hyperpolarization. Upon prolongation of the current pulse, RD was diminished by 15% from 9.1 ± 0.9 mV at a 1 s pulse to 7.7 ± 0.8 mV at a 4 s pulse (Fig. 4A, left traces; Fig. 4B, n = 8, p 0.02). In the presence of the T-type calcium channel blocker Ni2+ (0.1 mM), this pulse-related RD-inhibition was still present (Fig. 4A, middle traces; Fig. 4B, n = 4; p < 0.01), whereas it was abolished during Ih channel blockade by Cs+ (5 mM; Fig 4A, right traces; Fig. 4B, n = 4, ns).
Fig. 4. Inhibition of RD amplitude upon sustained hyperpolarization is due to Ih channel activity. (A) Original RD traces following a hyperpolarizing pulse ( 150 or 200 pA) of 1 s (black trace) and 4 s (grey trace) in the absence of channel blockers (left panel) and in the presence of 0.1 mM Ni2+ (middle panel) or 5 mM Cs+ (right panel) are shown. Scaling 4 mV/ 0.5 s. (B) The RD amplitudes upon a 1 s pulse (black bars) were compared with those upon a 4 s pulse (grey bars) under the different conditions (control n = 8; Ni2+ n = 4; Cs+ n = 4). Statistical significance as indicated by asterisks.
R. Surges et al. / Biochemical and Biophysical Research Communications 348 (2006) 1343–1349
The contribution of Ih to RD in current clamp mode underlies the so-called tail currents in the voltage-clamp mode. To examine the influence of pulse duration on these tail currents, we performed Ih recordings in the voltageclamp mode by application of voltage pulses from a holding potential from 63 mV to 103 mV for 1 and 4 s, respectively, followed by depolarizing steps back to the holding potential (Fig. 5A). The instantaneous current amplitudes upon the steps to more depolarized levels (=tail currents, indicated with asterisks in Fig. 5A) were plotted versus the corresponding command potentials after leakage
1347
subtraction (Fig. 5B). First, we compared the tail current amplitudes upon relaxation to the holding potential after a precedent hyperpolarization of 1 and 4 s, respectively. Indeed, tail currents were reduced by 10% by prolongation of the precedent hyperpolarization (inset of Fig. 5B; tail current amplitude following a 4 s pulse: 75 ± 5 pA; 1 s pulse: 83 ± 5 pA; n = 5; p < 0.005). To possibly explain the duration-related decrease in tail current amplitudes, we analyzed the reversal potential of Ih as well as the whole-cell conductance under both stimulation conditions. To this end, the tail current amplitudes were submitted to linear regression analysis (Fig. 5B). The intercept of the resulting lines with the X-axis represents the reversal potential, the respective slope is a measure of whole-cell conductance due to Ih (Fig. 5B, solid and dashed lines). In accordance to the observed reduction in tail current amplitude upon longer voltage pulses, the reversal potential was shifted by 7 mV to more hyperpolarized potentials at 4 s pulses, corresponding to a reduction of the driving force by 15% (driving force calculated after a 1 s pulse: 48 ± 12 mV; 4 s pulse: 41 ± 11 mV; n = 5, p < 0.005; Fig. 5C, left panel). In contrast, whole-cell conductance due to Ih channel activity was increased by 20% at a 4 s pulse, from 2.4 ± 0.7 nS at a 1 s pulse to 2.9 ± 0.8 nS at a 4 s pulse (Fig. 5C, right panel; n = 5; p < 0.05). Discussion The findings of the present study are that (1) RD is mainly carried by Ih (two-third) and by Ni2+-sensitive T-type calcium currents (one-third), that (2) RD is differentially modulated by precedent hyperpolarization with increasing RD amplitude upon increasing hyperpolarization and stimulation frequency, whereas (3) RD amplitude substantially decreases with a longer pulse duration and, less pronounced, with increasing pulse number, and that (4) this pulse duration-related decrease of RD amplitude is due to changes in Ih channel activity (mainly via a decrease of the driving force). (5) Postinhibitory AP number is also decreased upon sustained precedent hyperpolarization.
Fig. 5. Ih tail current amplitudes are impaired upon longer voltage pulses, going along with a decrease of the driving force and an increase in Ih whole-cell conductance. (A) In the voltage-clamp mode, pulses from a holding potential of 63 to 103 mV were applied for 1 or 4 s, followed by depolarizing steps back to the holding potential (upper panel stimulation protocol; lower panel original current traces after off-line leakage subtraction). The instantaneous current amplitudes upon the depolarizing steps represent the tail currents that were subsequently analyzed (indicated with asterisk). Scaling 100 pA/1 s. (B) Tail current amplitudes upon a 1 s (closed circles) and 4 s pulse (open circles) were plotted vs. the corresponding depolarizing steps and data points were submitted to linear regression analysis revealing the reversal potential/ whole-cell conductance. Inset shows the tail current amplitudes upon a 1 s (black bar) and 4 s pulse (grey bar, n = 5 each). (C) The driving force (left panel) decreased upon a 4 s pulse (grey bar) as compared to a 1 s pulse (black bar, n = 5), whereas the whole-cell conductance increased upon longer pulses (right panel, same symbols as left panel). Statistical significance as indicated by asterisks.
Ih and T-type calcium channels are involved in the generation of RD Neurons were hyperpolarized from a RP around 70 mV up to levels around 110 mV (at 200 pA) with subsequent rebound to RP. Within this voltage range, Ih channels as well as T-type calcium channels are the most probable candidates to be involved in the generation of RD. Ih channels are largely activated in the given voltage range, what leads to an inward current that itself partially depolarizes the neuronal membrane to around 100 mV. Since Ih channels do not inactivate, a substantial amount of Ih channels are still open upon repolarization, inducing an inward current depolarizing the membrane beyond the RP. Ih channel blockade by application of ZD7288 has markedly reduced the voltage
1348
R. Surges et al. / Biochemical and Biophysical Research Communications 348 (2006) 1343–1349
sag by more than 70% and RD amplitude by 65%, indicating that RD is largely carried by Ih channel activity. Since ZD7288 may also inhibit neuronal T-type calcium channels [12], we also applied Cs+ to block Ih channels, with a similar reduction in RD amplitude by 60%. Therefore, we suggest that ZD7288, at least at 50 lM, has only minor influence on neuronal T-type calcium channels. T-type calcium channels are activated upon depolarization beyond 80 to 70 mV with a relatively rapid inactivation. The inactivated calcium channels need hyperpolarization to recover and to be available for a further activation. Application of the Ni2+ did not affect the initial voltage response neither the voltage sag, but reduced the RD amplitude by about 35%. Additional application of ZD7288 nearly abolished RD amplitude, indicating that RD is largely carried by these two voltage-gated cationic channels. Differential modulation of RD amplitude by variation of stimulation protocols We have shown that increasing hyperpolarizing current amplitudes amplify RD amplitude. This relationship can be explained by voltage-dependent recruitment of Ih channels as well as promoted recovery of T-type calcium channels at more hyperpolarized levels. Since Ih channels do not inactivate but slowly deactivate upon repolarization with a time constant of 150 ms [14], the frequency-dependence of RD amplitude is due an increasing amount of activated Ih channels that are not yet deactivated. Interestingly, increasing pulse duration markedly reduced RD amplitude. This pulse-related decrease of RD was still present during Ni2+-application (Fig. 4), in accordance with the idea that hyperpolarization-dependent recovery of T-type calcium channels leads to an increase rather than decrease of RD amplitude. In contrast, the pulse-related RD reduction was abolished in the presence of Cs+, suggesting the involvement of Ih channels. Pulse duration-related decrease of RD is due to Ih channel activity Ih recordings in the voltage-clamp mode essentially showed the same phenomenon, that is to say a reduction of tail current amplitudes upon longer test pulses. As expected, the Ih whole-cell conductance increased upon longer test pulses, due to a recruitment of slowly activating Ih channel subunits. As previously shown, Ih substantially contributes to the membrane (input) resistance [5]. An increase of Ih conductance leads to a decrease of the membrane (input) resistance that could contribute to the impairment of RD upon longer test-pulses (in current clamp mode). Additionally, the driving force decreased by 15% during longer hyperpolarization, what fits well with the observed decrease of tail current amplitude as well as RD amplitude. In the present paper, hyperpolarization levels in current clamp as well as in voltage clamp recordings were close to the equilibrium potential of potassium (under the
current experimental conditions around 102 mV). Hence, at the hyperpolarization level that preceded RD/tail currents there was virtually no or only very small K+ net flux across the membrane, but a solid Na+ influx. This led most probably to an accumulation of Na+ ions over the time, what in turn caused a reduction of the driving force for Na+ ions upon relaxation, and thereby a reduction of RD and tail current amplitudes, respectively. Functional implications of RD modulation within the hippocampus GABAA receptor activation is generally thought to inhibit neuronal activity by hyperpolarization of the membrane potential, thereby shifting it away from the APthreshold. Here, we have showed that relatively short hyperpolarizing pulses at frequencies between 8 and 16 Hz can induce a RD that is sufficiently strong to generate action potentials (see Fig. 1D inset and Fig. 2A). Thus, a hyperpolarizing, inhibitory input can be converted into neuronal excitation even in physiologic tissue. Interestingly, GABAA receptor activity is increased in some animal models of temporal lobe epilepsy [15,16] and in a rat model of febrile seizures [8]. Interestingly in the latter, GABAergic input of CA1 pyramidal cells is substantially enhanced that, together with a concomitant augmentation of somatic Ih, led to a hyperexcitability and increased susceptibility to convulsant substances [2,8]. Additionally, Ih (and RD) were shown to contribute to the generation of focal paroxysmal activity in neocortical neurons [17]. Hence, RD together with Ih as the major underlying ionic current may serve as an inhibition-excitation converter under physiologic and pathophysiologic conditions. In contrast, the pulse duration-related decrease of RD reported in the present paper has probably no functional importance within the hippocampal circuitry, since the pulses that we used are beyond known physiologic signals. However, the phenomenon is interesting and might be of importance when considering biophysical properties of Ih. In conclusion, we characterized the modulation of RD by different stimulation patterns as well as its underlying ionic currents in rat CA1 pyramidal cells and showed that RD may serve as an inhibition-excitation converter even under physiologic conditions. References [1] C.D. Aizenman, D.J. Linden, Regulation of the rebound depolarization and spontaneous firing patterns of deep nuclear neurons in slices of rat cerebellum, J. Neurophysiol. 82 (1999) 1697–1709. [2] K. Chen, I. Aradi, N. Thon, M. Eghbal-Ahmadi, T.Z. Baram, I. Soltesz, Persistently modified h-channels after complex febrile seizures convert the seizure-induced enhancement of inhibition to hyperexcitability, Nat. Med. 7 (2001) 331–337. [3] A. Luthi, D.A. McCormick, Periodicity of thalamic synchronized oscillations: the role of Ca2+-mediated upregulation of Ih, Neuron 20 (1998) 553–563.
R. Surges et al. / Biochemical and Biophysical Research Communications 348 (2006) 1343–1349 [4] J.C. Magee, Dendritic hyperpolarization-activated currents modify the integrative properties of hippocampal CA1 pyramidal neurons, J. Neurosci. 18 (1998) 7613–7624. [5] R. Surges, T.M. Freiman, T.J. Feuerstein, Input resistance is voltage dependent due to activation of Ih channels in rat CA1 pyramidal cells, J. Neurosci. Res. 76 (2004) 475–480. [6] H. Hu, K. Vervaeke, J.F. Storm, Two forms of electrical resonance at theta frequencies, generated by M-current, h-current and persistent Na+ current in rat hippocampal pyramidal cells, J. Physiol. 545 (2002) 783–805. [7] E.H. Buhl, S.R. Cobb, K. Halasy, P. Somogyi, Properties of unitary IPSPs evoked by anatomically identified basket cells in the rat hippocampus, Eur. J. Neurosci. 7 (1995) 1989–2004. [8] K. Chen, T.Z. Baram, I. Soltesz, Febrile seizures in the developing brain result in persistent modification of neuronal excitability in limbic circuits, Nat. Med. 5 (1999) 888–894. [9] B. Santoro, T.Z. Baram, The multiple personalities of h-channels, Trends Neurosci. 26 (2003) 550–554. [10] P.F. Costa, M.A. Ribeiro, A.I. Santos, Afterpotential characteristics and firing patterns in maturing rat hippocampal CA1 neurones in in vitro slices, Brain Res. Dev. Brain Res. 62 (1991) 263–272. [11] R. Azouz, G. Alroy, Y. Yaari, Modulation of endogenous firing patterns by osmolarity in rat hippocampal neurones, J. Physiol. 502 (1997) 175–187.
1349
[12] R. Felix, A. Sandoval, D. Sanchez, J.C. Gomora, J.L. De la VegaBeltran, C.L. Trevino, A. Darszon, ZD7288 inhibits low-threshold Ca(2+) channel activity and regulates sperm function, Biochem. Biophys. Res. Commun. 311 (2003) 187–192. [13] N.C. Harris, A. Constanti, Mechanism of block by ZD7288 of the hyperpolarization-activated inward rectifying current in guinea pig substantia nigra neurons in vitro, J. Neurophysiol. 74 (1995) 2366– 2378. [14] R. Surges, A.L. Brewster, R.A. Bender, H. Beck, T.J. Feuerstein, T.Z. Baram, Regulated expression of HCN channels and cAMP levels shape the properties of the h current in developing rat hippocampus, Eur. J. Neurosci. 24 (2006) 94–104. [15] J.W. Gibbs 3rd, M.D. Shumate, D.A. Coulter, Differential epilepsyassociated alterations in postsynaptic GABA(A) receptor function in dentate granule and CA1 neurons, J. Neurophysiol. 77 (1997) 1924– 1938. [16] A.R. Brooks-Kayal, M.D. Shumate, H. Jin, T.Y. Rikhter, D.A. Coulter, Selective changes in single cell GABA(A) receptor subunit expression and function in temporal lobe epilepsy, Nat. Med. 4 (1998) 1166–1172. [17] I. Timofeev, M. Bazhenov, T. Sejnowski, M. Steriade, Cortical hyperpolarization-activated depolarizing current takes part in the generation of focal paroxysmal activities, Proc. Natl. Acad. Sci. USA 99 (2002) 9533–9537.
Characterization of rebound depolarization in hippocampal neurons Rainer Surges *, Monika Sarvari, Marc Steffens, Thomas Els Department of Neurology and Clinical Neurophysiology, University Clinics of Freiburg, Germany Received 26 July 2006 Available online 7 August 2006
Abstract Rebound depolarization (RD) following hyperpolarizing pulses is found in several neuronal cell types where it takes part in the regulation of neuronal firing behavior. During whole-cell current and voltage clamp recordings in slice preparations, we investigated the modulation of RD by different stimulation patterns and its underlying ionic currents in rat CA1 pyramidal cells. RD was mainly carried by the hyperpolarization-activated cation current Ih (about two-third) and T-type calcium currents (about one-third), respectively. RD increased with increasing hyperpolarizing amplitude and stimulation frequency, whereas RD substantially decreased with longer pulse duration and, less pronounced, with increasing pulse number. The pulse duration-related decrease of RD was due to a decrease of the driving force of Ih. In conclusion, we showed that RD is differentially modulated by precedent hyperpolarization. Since RD amplitude was high enough to generate action potentials, RD may serve, even under physiologic conditions, as an inhibition-excitation converter. 2006 Elsevier Inc. All rights reserved. Keywords: CA1; Ih; HCN channels; T-type calcium channels; Postinhibitory rebound depolarization; Hyperpolarization-activated cation current; Epilepsy; ZD7288
Neurons of different brain regions, e.g., within the thalamus, cerebellum, and hippocampus [1–3], display a transient depolarization following hyperpolarizing pulses, a phenomenon called rebound depolarization (RD). RD contributes, for example, to the regulation of the neuronal firing mode in cerebellum [1] and to oscillations in thalamocortical networks [3]. RD is, at least in the cell types mentioned above, substantially carried by the hyperpolarization-activated cation current (Ih). Ih is a mixed, slowly developing inward current consisting of Na+- and K+-ions that is activated by hyperpolarization more negative than 60 mV and which does not inactivate during sustained hyperpolarization. Upon repolarization Ih slowly deactivates, leading to an inward current producing RD. In hippocampal CA1 pyramidal cells, the hyperpolarization activated, cyclic nucleotide-gated (HCN) channels underlying Ih are expressed on dendrites as well as on the somatic cell membrane with a sevenfold higher current density on distal dendrites where it
*
Corresponding author. Fax: +49 761 2705281. E-mail address: [email protected] (R. Surges).
0006-291X/$ - see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2006.07.193
takes part in spatial and temporal integration of synaptic input [4]. However, Ih is of robust amplitude on neuronal cell soma and participates in the regulation of intrinsic membrane properties such as resting potential, input resistance [5], and resonance behavior [6]. CA1 neurons receive inhibitory signals from diverse interneurons that project onto dendrites as well as to the somatic and perisomatic cell membrane [e.g., 7–9] where they can activate somatic Ih [2]. Previously, RD was shown to be essentially involved in the enhanced excitability of hippocampal circuitry in a rat model of febrile seizures [2,8], suggesting a considerable role of RD in the regulation of neuronal firing behavior. We therefore characterized the modulation of RD and its underlying ionic currents in CA1 pyramidal cells of rat hippocampal slices using the whole-cell patch clamp technique. Materials and methods Slice preparation. Hippocampal slices were prepared from 14 to 22 postnatal (P) days old Wistar rats of either sex. Experiments were performed according to national and institutional guidelines. Rats were decapitated, the brains were rapidly removed, hemisected, and submerged in ice-cold
1344
R. Surges et al. / Biochemical and Biophysical Research Communications 348 (2006) 1343–1349
physiological Ringer’s solution (Biometra, Go¨ttingen, Germany) which was continuously bubbled with 95% O2 and 5% CO2 and contained (in mM): NaCl, 125; NaHCO3, 25; KCl, 2.5; NaH2PO4, 1.25; MgCl2, 1; glucose, 25; CaCl2, 2. The hemispheres were blocked in the horizontal plane and glued to the stage of a vibratome (Leica VT 1000S, Wetzlar, Germany). Horizontal sections through the hippocampus (350–400 lm thick) were cut, and slices were transferred to a holding chamber, where they were submerged in Ringer’s solution. This solution was continuously bubbled with 95% O2 and 5% CO2 and was maintained at 35 C for the first 30 min and then at room temperature (20–22 C). Slices were kept in the holding chamber for at least 1 h before recordings were performed. Electrophysiology. Individual slices were transferred into the recording chamber, fixed with a nylon-grid and continuously perfused with oxygenated Ringer’s solution at room temperature (20–22 C). Neurons were visualized by infrared differential interference contrast video-microscopy with a Newvicon camera (C2400, Hamamatsu, Hamamatsu City, Japan) and an infrared filter (RG9, Schott, Mainz, Germany) to an upright microscope (Axioskop 2 FS, Zeiss, Oberkochen, Germany) equipped with a 40· water immersion objective. Recordings at room temperature (20– 22 C) were made in the whole-cell configuration of the patch-clamp technique using an Axopatch 200B amplifier and pClamp 8 (Axon Instruments, Foster City, CA, USA). For current clamp recordings, signals were low-pass filtered at 1 or 10 kHz and sampled at 2.5, 3.3 or 50 kHz, and during voltage-clamp experiment signals were low-pass filtered at 1 kHz and sampled at 5 kHz, respectively (Digidata 1200, Axon Instruments). Pipettes were pulled from non-filamented borosilicate glass capillaries (Hilgenberg, Malsfeld, Germany; outer diameter 1.5 mm, inner diameter 1.0 mm) on a P-97 Flaming/Brown horizontal puller (Sutter Instruments, Novato, CA, USA) and had a mean resistance of 4.8 ± 0.1 MX (n = 29) with the pipette solution. Standard pipette solution contained (in mM): Kgluconate, 120; KCl, 20; MgCl2, 2, CaCl2, 1; Hepes, 10; EGTA, 10; ATP– Tris, 2; pH 7.35, adjusted with KOH. The offset potential between the patch-pipette and the reference electrode was zeroed before the tight-seal (>2 GX) was established. Series resistance was determined by using the ‘‘membrane-test’’ protocol provided by the software pClamp 8.0 and amounted to 13–20 MX (compensation 40–50% in the voltage clamp experiments). A liquid junction potential of +13.1 mV was calculated and all values given are corrected for this voltage offset so that without correction the voltages given in this paper would be 13 mV more positive. Solutions and chemicals. The standard bath-solution consisted of (in mM): NaCl, 120; NaHCO3, 25; KCl, 2.5; MgCl2, 1; glucose, 25; CaCl2, 2. In most current clamp experiments, 20 lM bicuculline and 10 lM NBQX (Tocris, Biotrend Chemikalien, Ko¨ln, Germany) were added to the bath solution to block spontaneous synaptic events. To isolate Ih during voltage-clamp experiments, the following substances were added (in mM): Tetrodotoxin (TTX; Tocris, Biotrend Chemikalien, Ko¨ln, Germany), 0.005–0.01; NiCl2, 1; TEA, 10; 4-AP, 2; BaCl2, 0.5. ZD7288 [4-(N-ethyl-Nphenylamino)-1,2-dimethyl-6-(methylamino) pyrimidium chloride] (Tocris) was dissolved in H2O in a stock solution of 25 mM. CsCl was dissolved in H2O in a stock solution of 100 mM. Unless indicated otherwise, drugs were purchased from Sigma (Taufkirchen, Germany). Analysis. Data were acquired with the help of pCLAMP 8.0 and analyzed offline using the software packages Clampfit 8.0 (Axon Instruments) and GraphPad PRISM 2.0 (GraphPad Software Inc., San Diego, CA, USA). Linear regression analysis to obtain the reversal potential of Ih was performed with GraphPad PRISM. Statistical significance (p < 0.05) was assessed, if not indicated otherwise, using the two-tailed paired student’s t test. All data were expressed as means ± SEM.
ing potential (RP) was determined immediately after establishing the whole-cell mode and amounted to 71 ± 1 mV. At the onset of each experiment, a family of current pulses was applied in the current clamp mode. Prolonged depolarizing current injection evoked a train of action potentials (AP) displaying spike adaptation (not shown), a behavior that is typical for CA1 pyramidal neurons [e.g., 10,11]. During hyperpolarizing current injections, a slowly developing, pronounced ‘voltage sag’ occurred (see Fig. 1A, left panel), indicative of the presence of Ih. Upon repolarization to RP, one can observe a substantial transient depolarization called rebound depolarization (RD; see Fig. 1A, left panel). To examine the influence of precedent hyperpolarization on RD amplitude, the following characteristics of the hyperpolarizing pulses were varied: (a) amplitude, (b) duration, (c) pulse number, and (d) frequency. To avoid AP’s due to RD, TTX (0.5 lM) was added to the bath solution or alternatively, a slight hyperpolarizing current was constantly delivered via the patch clamp amplifier (in order to increase the difference between membrane potential and AP-threshold). First, hyperpolarizing current pulses (2 s) of different amplitudes were injected into the neurons starting from the resting potential (or 1–2 mV more negative due to constantly applied negative current injections via the amplifier). With increasing current amplitude, RD amplitude substantially increased from 1.2 ± 0.3 mV to 9.6 ± 1.6 mV (Fig. 1A, n = 8). Next, hyperpolarizing pulses ( 200 pA) of different duration ranging from 1 to 10 s were administered. Surprisingly, we observed a small but robust decline in the RD amplitude with increasing pulse duration, from 9.1 ± 1.0 mV at 1 s to 7.3 ± 1.0 mV at 10 s (Fig. 1B, n = 8). Upon variation of the pulse number (from 5 to 45 pulses), RD amplitude was only slightly affected within the given conditions with a relative maximum of RD about 8.6 ± 0.5 mV following 10 pulses and a tendency to decline upon higher pulse numbers to 8.1 ± 0.4 at 45 pulses (Fig. 1C, n = 6). Finally, we delivered 30 pulses (120 ms; 200 pA) at different stimulation frequencies ranging from 1 to 8 Hz (Fig. 1D). The relative increase in RD was highest at 2 Hz as compared to 1 Hz (RD at 1 Hz 4.5 ± 0.8 mV to 6.0 ± 0.6 mV at 2 Hz, p < 0.05) reaching a plateau between 2 and 6 Hz with a further increase at 7 Hz (Fig. 1D, right panel, n = 8). At a shorter pulse duration (60 ms) and higher frequencies, we even observed a 3.6-fold increase of the RD amplitude at 16 Hz (RD amplitude 5.5 ± 1.2 mV) as compared to 1 Hz (RD amplitude 1.5 ± 0.3 mV, n = 4, p < 0.05; Fig. 1D, right panel inset).
Results
Generation of AP’s upon RD is decreased by increasing duration of the precedent hyperpolarization
RD amplitude depends on the characteristics of the precedent hyperpolarization The present study includes information from 29 visually selected cells within the CA1 pyramidal cell layer. The rest-
In the previous experiments the generation of AP was inhibited by pharmacologic or electrophysiologic means. In this section, we investigated the dependence of AP on the amplitude and duration of the precedent
R. Surges et al. / Biochemical and Biophysical Research Communications 348 (2006) 1343–1349
1345
hyperpolarization. We first applied 2 s hyperpolarizing currents and recorded the membrane potential in the absence of TTX. In analogy to the dependence of the RD amplitude on the current amplitude, the AP number (or the AP probability) also increased with increasing current amplitude of the precedent hyperpolarization (Fig. 2A; n = 9). Upon variation of the pulse duration, we observed a reduction of the AP number from 1 at a 1 s pulse to about 0.6 ± 0.2 AP’s following a pulse of 10 s, in accordance with the pulse duration-related decrease of RD amplitude (Fig. 2B; n = 5). Nature of RD: Ih and T-type calcium channels are both involved To assess the nature of the underlying channel activity of RD, the changes of RD amplitudes as well as of sag amplitudes (displaying an estimation of Ih channel activation) were recorded in the presence of different channels blockers. We first applied Ni2+ (0.1 mM) to block T-type calcium channels, leading to a decrease of RD amplitude by 36% from 6.9 ± 1.0 mV to 4.4 ± 0.3 mV (Fig. 3A; n = 4; p < 0.05) without alterations of the sag amplitude. Additional application of the Ih channel blocker ZD7288 (50 lM) nearly abolished the RD amplitude by 94% as compared to control to 0.4 ± 0.4 mV (n = 4; p < 0.005) with an accompanying reduction in sag amplitude by
Fig. 1. Modulation of RD amplitude by different stimulation patterns. (A) In current clamp mode, hyperpolarizing currents of 2 s duration with different amplitudes (from 30 to 360 pA in 30 pA-steps, 0.15 Hz) were injected (left panel: original recordings; grey trace 120 pA; black trace 240 pA; sag and RD as indicated) and the resulting RD amplitudes were plotted vs. the corresponding current amplitudes (right panel; n = 8). Scaling 25 mV/1 s, RP 71 mV. (B) Pulses ( 200 pA) of different duration ranging from 1 to 10 s were administered (left panel: original recordings) and corresponding RD amplitudes were plotted vs. the pulse duration (right panel, n = 8). Scaling 25 mV/2 s, RP 69 mV. (C) Different number of identical pulses ( 200 pA, 120 ms, 4 Hz) were applied (left panel: original recordings with stimulation by 10 pulses) and the RD amplitudes following the last pulse were plotted vs. the pulse no. (right panel; n = 6). Scaling 25 mV/1 s, RP 66 mV. (D) Thirty identical pulses ( 200 pA, 120 ms) were applied at different stimulation frequencies (left panel: original recordings at 2 Hz) and RD amplitudes after the last pulse were plotted vs. the stimulation frequency (right panel; n = 8). Inset shows RD amplitudes at shorter pulse duration (60 ms) and higher stimulation frequency (up to 16 Hz; n = 4). Scaling 25 mV/5 s, RP 69 mV.
Fig. 2. Number of action potentials upon RD is decreased by increasing duration of hyperpolarization. (A) Two-second pulses of different amplitude were applied in the absence of TTX (left panel: original recordings; grey trace 120 pA; black trace 240 pA) and AP no. upon RD was plotted vs. the current amplitude (right panel, n = 9). Scaling 25 mV/1 s, RP 78 mV. (B) Pulses ( 200 pA) of different duration ranging from 1 to 10 s were administered (left panel: original recordings) and corresponding AP no. was plotted vs. the pulse duration (right panel, n = 5). Scaling 25 mV/2 s, RP 65 mV.
1346
R. Surges et al. / Biochemical and Biophysical Research Communications 348 (2006) 1343–1349
administered Cs+ (5 mM) to block Ih channels with similar results (Fig. 3C). Especially upon wash-out of ZD7288, both the sag amplitude as well the RD amplitude only partially recovered, as known by previously published voltageclamp experiments (e.g., [13]; data not shown). Note that blockade of Ih channels substantially increases the neuronal input resistance with an additional slight hyperpolarization of the membrane potential [5]. Therefore, in order to correct for errors due to higher voltage deflections upon the same current amplitude (because of Ih channel blockade in the presence of ZD7288 or Cs+, respectively), only the traces of similar initial voltage response to the current injection were compared. For instance, the RD amplitudes under control conditions upon a 200 pA pulse were compared to the RD amplitudes during ZD7288-application upon a 150 pA pulse. The duration-related inhibition of RD is due to Ih channel activity
Fig. 3. Both T-type calcium currents and Ih contribute to RD. (A) Hyperpolarizing pulses ( 150 to 200 pA) were applied under control conditions, in the presence of 0.1 mM Ni2+ alone and during additional application of 50 lM ZD7288 (left panel: original recordings; upper black trace control, grey trace under Ni2+ application; lower black trace in the presence of Ni2+ and ZD7288) and the resulting sag/RD amplitudes were compared under the different conditions (black bars: control; grey bars: Ni2+; white bars: Ni2++ZD7288, recordings from the same 4 cells). RP 78 mV. (B) Hyperpolarizing pulses ( 150 to 200 pA) were applied under control conditions and during application of 50 lM ZD7288 (left panel: original recordings; upper black trace control, grey trace in the presence of ZD7288) and the resulting sag/RD amplitudes were compared (black bars: control, grey bars: ZD7288, n = 4). RP 73 mV. (C) Hyperpolarizing pulses ( 150 to 200 pA) were applied under control conditions and during application of 5 mM Cs+ (left panel: original recordings; upper black trace control, grey trace in the presence of Cs+) and the resulting sag/RD amplitudes were compared (black bars: control, grey bars: Cs+, n = 4). RP 80 mV. Note that because of possible errors due an increase of input resistance in the presence of ZD7288/Cs+, the control values (at a 200 pA pulse) were compared to traces upon lower current injections that displayed similar initial voltage responses (e.g., at 150 pA pulses). Statistical significance as indicated by asterisks. Scaling for all recordings 25 mV/1 s.
80% as compared to control from 9.8 ± 1.8 mV to 1.9 ± 0.4 mV (Fig. 3A; n = 4, p < 0.02). Application of ZD7288 (50 lM) alone reduced the sag amplitude by 72% from 13.6 ± 1.9 mV to 3.7 ± 1.1 mV (n = 4, p < 0.02), and the RD amplitude by 65% from 10.8 ± 0.7 mV to 3.8 ± 1.2 mV (n = 4, p < 0.01). As ZD7288 may also block neuronal T-type calcium channels [12], we additionally
As shown above, RD amplitude was reduced upon sustained hyperpolarization. Upon prolongation of the current pulse, RD was diminished by 15% from 9.1 ± 0.9 mV at a 1 s pulse to 7.7 ± 0.8 mV at a 4 s pulse (Fig. 4A, left traces; Fig. 4B, n = 8, p 0.02). In the presence of the T-type calcium channel blocker Ni2+ (0.1 mM), this pulse-related RD-inhibition was still present (Fig. 4A, middle traces; Fig. 4B, n = 4; p < 0.01), whereas it was abolished during Ih channel blockade by Cs+ (5 mM; Fig 4A, right traces; Fig. 4B, n = 4, ns).
Fig. 4. Inhibition of RD amplitude upon sustained hyperpolarization is due to Ih channel activity. (A) Original RD traces following a hyperpolarizing pulse ( 150 or 200 pA) of 1 s (black trace) and 4 s (grey trace) in the absence of channel blockers (left panel) and in the presence of 0.1 mM Ni2+ (middle panel) or 5 mM Cs+ (right panel) are shown. Scaling 4 mV/ 0.5 s. (B) The RD amplitudes upon a 1 s pulse (black bars) were compared with those upon a 4 s pulse (grey bars) under the different conditions (control n = 8; Ni2+ n = 4; Cs+ n = 4). Statistical significance as indicated by asterisks.
R. Surges et al. / Biochemical and Biophysical Research Communications 348 (2006) 1343–1349
The contribution of Ih to RD in current clamp mode underlies the so-called tail currents in the voltage-clamp mode. To examine the influence of pulse duration on these tail currents, we performed Ih recordings in the voltageclamp mode by application of voltage pulses from a holding potential from 63 mV to 103 mV for 1 and 4 s, respectively, followed by depolarizing steps back to the holding potential (Fig. 5A). The instantaneous current amplitudes upon the steps to more depolarized levels (=tail currents, indicated with asterisks in Fig. 5A) were plotted versus the corresponding command potentials after leakage
1347
subtraction (Fig. 5B). First, we compared the tail current amplitudes upon relaxation to the holding potential after a precedent hyperpolarization of 1 and 4 s, respectively. Indeed, tail currents were reduced by 10% by prolongation of the precedent hyperpolarization (inset of Fig. 5B; tail current amplitude following a 4 s pulse: 75 ± 5 pA; 1 s pulse: 83 ± 5 pA; n = 5; p < 0.005). To possibly explain the duration-related decrease in tail current amplitudes, we analyzed the reversal potential of Ih as well as the whole-cell conductance under both stimulation conditions. To this end, the tail current amplitudes were submitted to linear regression analysis (Fig. 5B). The intercept of the resulting lines with the X-axis represents the reversal potential, the respective slope is a measure of whole-cell conductance due to Ih (Fig. 5B, solid and dashed lines). In accordance to the observed reduction in tail current amplitude upon longer voltage pulses, the reversal potential was shifted by 7 mV to more hyperpolarized potentials at 4 s pulses, corresponding to a reduction of the driving force by 15% (driving force calculated after a 1 s pulse: 48 ± 12 mV; 4 s pulse: 41 ± 11 mV; n = 5, p < 0.005; Fig. 5C, left panel). In contrast, whole-cell conductance due to Ih channel activity was increased by 20% at a 4 s pulse, from 2.4 ± 0.7 nS at a 1 s pulse to 2.9 ± 0.8 nS at a 4 s pulse (Fig. 5C, right panel; n = 5; p < 0.05). Discussion The findings of the present study are that (1) RD is mainly carried by Ih (two-third) and by Ni2+-sensitive T-type calcium currents (one-third), that (2) RD is differentially modulated by precedent hyperpolarization with increasing RD amplitude upon increasing hyperpolarization and stimulation frequency, whereas (3) RD amplitude substantially decreases with a longer pulse duration and, less pronounced, with increasing pulse number, and that (4) this pulse duration-related decrease of RD amplitude is due to changes in Ih channel activity (mainly via a decrease of the driving force). (5) Postinhibitory AP number is also decreased upon sustained precedent hyperpolarization.
Fig. 5. Ih tail current amplitudes are impaired upon longer voltage pulses, going along with a decrease of the driving force and an increase in Ih whole-cell conductance. (A) In the voltage-clamp mode, pulses from a holding potential of 63 to 103 mV were applied for 1 or 4 s, followed by depolarizing steps back to the holding potential (upper panel stimulation protocol; lower panel original current traces after off-line leakage subtraction). The instantaneous current amplitudes upon the depolarizing steps represent the tail currents that were subsequently analyzed (indicated with asterisk). Scaling 100 pA/1 s. (B) Tail current amplitudes upon a 1 s (closed circles) and 4 s pulse (open circles) were plotted vs. the corresponding depolarizing steps and data points were submitted to linear regression analysis revealing the reversal potential/ whole-cell conductance. Inset shows the tail current amplitudes upon a 1 s (black bar) and 4 s pulse (grey bar, n = 5 each). (C) The driving force (left panel) decreased upon a 4 s pulse (grey bar) as compared to a 1 s pulse (black bar, n = 5), whereas the whole-cell conductance increased upon longer pulses (right panel, same symbols as left panel). Statistical significance as indicated by asterisks.
Ih and T-type calcium channels are involved in the generation of RD Neurons were hyperpolarized from a RP around 70 mV up to levels around 110 mV (at 200 pA) with subsequent rebound to RP. Within this voltage range, Ih channels as well as T-type calcium channels are the most probable candidates to be involved in the generation of RD. Ih channels are largely activated in the given voltage range, what leads to an inward current that itself partially depolarizes the neuronal membrane to around 100 mV. Since Ih channels do not inactivate, a substantial amount of Ih channels are still open upon repolarization, inducing an inward current depolarizing the membrane beyond the RP. Ih channel blockade by application of ZD7288 has markedly reduced the voltage
1348
R. Surges et al. / Biochemical and Biophysical Research Communications 348 (2006) 1343–1349
sag by more than 70% and RD amplitude by 65%, indicating that RD is largely carried by Ih channel activity. Since ZD7288 may also inhibit neuronal T-type calcium channels [12], we also applied Cs+ to block Ih channels, with a similar reduction in RD amplitude by 60%. Therefore, we suggest that ZD7288, at least at 50 lM, has only minor influence on neuronal T-type calcium channels. T-type calcium channels are activated upon depolarization beyond 80 to 70 mV with a relatively rapid inactivation. The inactivated calcium channels need hyperpolarization to recover and to be available for a further activation. Application of the Ni2+ did not affect the initial voltage response neither the voltage sag, but reduced the RD amplitude by about 35%. Additional application of ZD7288 nearly abolished RD amplitude, indicating that RD is largely carried by these two voltage-gated cationic channels. Differential modulation of RD amplitude by variation of stimulation protocols We have shown that increasing hyperpolarizing current amplitudes amplify RD amplitude. This relationship can be explained by voltage-dependent recruitment of Ih channels as well as promoted recovery of T-type calcium channels at more hyperpolarized levels. Since Ih channels do not inactivate but slowly deactivate upon repolarization with a time constant of 150 ms [14], the frequency-dependence of RD amplitude is due an increasing amount of activated Ih channels that are not yet deactivated. Interestingly, increasing pulse duration markedly reduced RD amplitude. This pulse-related decrease of RD was still present during Ni2+-application (Fig. 4), in accordance with the idea that hyperpolarization-dependent recovery of T-type calcium channels leads to an increase rather than decrease of RD amplitude. In contrast, the pulse-related RD reduction was abolished in the presence of Cs+, suggesting the involvement of Ih channels. Pulse duration-related decrease of RD is due to Ih channel activity Ih recordings in the voltage-clamp mode essentially showed the same phenomenon, that is to say a reduction of tail current amplitudes upon longer test pulses. As expected, the Ih whole-cell conductance increased upon longer test pulses, due to a recruitment of slowly activating Ih channel subunits. As previously shown, Ih substantially contributes to the membrane (input) resistance [5]. An increase of Ih conductance leads to a decrease of the membrane (input) resistance that could contribute to the impairment of RD upon longer test-pulses (in current clamp mode). Additionally, the driving force decreased by 15% during longer hyperpolarization, what fits well with the observed decrease of tail current amplitude as well as RD amplitude. In the present paper, hyperpolarization levels in current clamp as well as in voltage clamp recordings were close to the equilibrium potential of potassium (under the
current experimental conditions around 102 mV). Hence, at the hyperpolarization level that preceded RD/tail currents there was virtually no or only very small K+ net flux across the membrane, but a solid Na+ influx. This led most probably to an accumulation of Na+ ions over the time, what in turn caused a reduction of the driving force for Na+ ions upon relaxation, and thereby a reduction of RD and tail current amplitudes, respectively. Functional implications of RD modulation within the hippocampus GABAA receptor activation is generally thought to inhibit neuronal activity by hyperpolarization of the membrane potential, thereby shifting it away from the APthreshold. Here, we have showed that relatively short hyperpolarizing pulses at frequencies between 8 and 16 Hz can induce a RD that is sufficiently strong to generate action potentials (see Fig. 1D inset and Fig. 2A). Thus, a hyperpolarizing, inhibitory input can be converted into neuronal excitation even in physiologic tissue. Interestingly, GABAA receptor activity is increased in some animal models of temporal lobe epilepsy [15,16] and in a rat model of febrile seizures [8]. Interestingly in the latter, GABAergic input of CA1 pyramidal cells is substantially enhanced that, together with a concomitant augmentation of somatic Ih, led to a hyperexcitability and increased susceptibility to convulsant substances [2,8]. Additionally, Ih (and RD) were shown to contribute to the generation of focal paroxysmal activity in neocortical neurons [17]. Hence, RD together with Ih as the major underlying ionic current may serve as an inhibition-excitation converter under physiologic and pathophysiologic conditions. In contrast, the pulse duration-related decrease of RD reported in the present paper has probably no functional importance within the hippocampal circuitry, since the pulses that we used are beyond known physiologic signals. However, the phenomenon is interesting and might be of importance when considering biophysical properties of Ih. In conclusion, we characterized the modulation of RD by different stimulation patterns as well as its underlying ionic currents in rat CA1 pyramidal cells and showed that RD may serve as an inhibition-excitation converter even under physiologic conditions. References [1] C.D. Aizenman, D.J. Linden, Regulation of the rebound depolarization and spontaneous firing patterns of deep nuclear neurons in slices of rat cerebellum, J. Neurophysiol. 82 (1999) 1697–1709. [2] K. Chen, I. Aradi, N. Thon, M. Eghbal-Ahmadi, T.Z. Baram, I. Soltesz, Persistently modified h-channels after complex febrile seizures convert the seizure-induced enhancement of inhibition to hyperexcitability, Nat. Med. 7 (2001) 331–337. [3] A. Luthi, D.A. McCormick, Periodicity of thalamic synchronized oscillations: the role of Ca2+-mediated upregulation of Ih, Neuron 20 (1998) 553–563.
R. Surges et al. / Biochemical and Biophysical Research Communications 348 (2006) 1343–1349 [4] J.C. Magee, Dendritic hyperpolarization-activated currents modify the integrative properties of hippocampal CA1 pyramidal neurons, J. Neurosci. 18 (1998) 7613–7624. [5] R. Surges, T.M. Freiman, T.J. Feuerstein, Input resistance is voltage dependent due to activation of Ih channels in rat CA1 pyramidal cells, J. Neurosci. Res. 76 (2004) 475–480. [6] H. Hu, K. Vervaeke, J.F. Storm, Two forms of electrical resonance at theta frequencies, generated by M-current, h-current and persistent Na+ current in rat hippocampal pyramidal cells, J. Physiol. 545 (2002) 783–805. [7] E.H. Buhl, S.R. Cobb, K. Halasy, P. Somogyi, Properties of unitary IPSPs evoked by anatomically identified basket cells in the rat hippocampus, Eur. J. Neurosci. 7 (1995) 1989–2004. [8] K. Chen, T.Z. Baram, I. Soltesz, Febrile seizures in the developing brain result in persistent modification of neuronal excitability in limbic circuits, Nat. Med. 5 (1999) 888–894. [9] B. Santoro, T.Z. Baram, The multiple personalities of h-channels, Trends Neurosci. 26 (2003) 550–554. [10] P.F. Costa, M.A. Ribeiro, A.I. Santos, Afterpotential characteristics and firing patterns in maturing rat hippocampal CA1 neurones in in vitro slices, Brain Res. Dev. Brain Res. 62 (1991) 263–272. [11] R. Azouz, G. Alroy, Y. Yaari, Modulation of endogenous firing patterns by osmolarity in rat hippocampal neurones, J. Physiol. 502 (1997) 175–187.
1349
[12] R. Felix, A. Sandoval, D. Sanchez, J.C. Gomora, J.L. De la VegaBeltran, C.L. Trevino, A. Darszon, ZD7288 inhibits low-threshold Ca(2+) channel activity and regulates sperm function, Biochem. Biophys. Res. Commun. 311 (2003) 187–192. [13] N.C. Harris, A. Constanti, Mechanism of block by ZD7288 of the hyperpolarization-activated inward rectifying current in guinea pig substantia nigra neurons in vitro, J. Neurophysiol. 74 (1995) 2366– 2378. [14] R. Surges, A.L. Brewster, R.A. Bender, H. Beck, T.J. Feuerstein, T.Z. Baram, Regulated expression of HCN channels and cAMP levels shape the properties of the h current in developing rat hippocampus, Eur. J. Neurosci. 24 (2006) 94–104. [15] J.W. Gibbs 3rd, M.D. Shumate, D.A. Coulter, Differential epilepsyassociated alterations in postsynaptic GABA(A) receptor function in dentate granule and CA1 neurons, J. Neurophysiol. 77 (1997) 1924– 1938. [16] A.R. Brooks-Kayal, M.D. Shumate, H. Jin, T.Y. Rikhter, D.A. Coulter, Selective changes in single cell GABA(A) receptor subunit expression and function in temporal lobe epilepsy, Nat. Med. 4 (1998) 1166–1172. [17] I. Timofeev, M. Bazhenov, T. Sejnowski, M. Steriade, Cortical hyperpolarization-activated depolarizing current takes part in the generation of focal paroxysmal activities, Proc. Natl. Acad. Sci. USA 99 (2002) 9533–9537.