| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Physiology and Institute for Neuroscience, Northwestern University Feinberg School of Medicine, Chicago, Illinois 60611
Submitted 9 October 2003; accepted in final form 6 June 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
, 2000; Thompson et al. 1996), or a spatial task, watermaze training (Oh et al. 2003). We have also shown an enhancement in the sAHP in CA1 pyramidal neurons from animals at ages that show learning deficits (Moyer et al. 1992; Power et al. 2002). AHP reductions have also been reported in rat piriform cortical pyramidal neurons after odor-discrimination learning (Saar et al. 1998; Seroussi et al. 2002). Together, these data strongly indicate that reduction in the sAHP is a general mechanism underlying learning, and enhancement of the sAHP is involved in aging-related learning deficit. In support, we have shown that compounds that suppress the sAHP facilitate learning in aging animals (Kowalska and Disterhoft 1994; Moyer et al. 1992; Oh et al. 1999; Power et al. 2001, 2002).
The sAHP limits cell firing in response to sustained depolarization, a phenomenon known as spike frequency adaptation (Lancaster and Nicoll 1987; Madison and Nicoll 1984; for reviews, see Sah 1996; Storm 1990). There is evidence that the sAHP shapes temporal integration of synaptic inputs by shunting excitatory postsynaptic potentials (EPSPs) arising in the stratum radiatum (Sah and Bekkers 1996). Steady-state activation of sIAHP with intracellular application of diazo-2, a photolabile BAPTA derivative, also decreases EPSP temporal summation (Lancaster et al. 2001). Since the sAHP decreases the overall neuronal responsiveness to stimulation, it is considered an index of neuronal excitability (Sah 1996).
The sAHP can be activated by Ca2+ derived from many sources. In most studies, it was activated by Ca2+ influx associated with somatic depolarization. However, under certain conditions, voltage-gated Ca2+ influx triggered by synaptic stimuli, even in the absence of action potential generation, or N-methyl-D-aspartate (NMDA) receptor–mediated Ca2+ influx evoked by focal applications of NMDA may be sufficient to activate the sAHP (Lancaster et al. 2001; Shah and Haylett 2002). Although the functional implication for the sAHP has been generally accepted as a dampening in neuronal excitability (Sah 1996), how and to what extent this afterpotential affects the processing of different patterns of EPSPs, such as those likely to be involved during learning, are questions that remain to be determined. Thus to further our understanding of the functional significance of the sAHP reduction, our objectives in this study were 1) to characterize the sAHP activated by different patterns of synaptic stimuli and 2) to examine the interaction between the sAHP and the neuronal response to these inputs.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Hippocampus slice preparation
Six to eight mo-old male F344XBN rats were anesthetized with halothane and killed by decapitation. The brain was rapidly removed, and a block containing the left hippocampus and surrounding structures was dissected out, attached to a mounting tray with cyanoacrylate glue, and immersed in chilled artificial cerebrospinal fluid (aCSF; 
1°C) consisting of the following (in mM): 119 NaCl, 26 NaHCO3, 2.5 KCl, 1 NaH2PO4 · H2O, 1.3 MgCl2 · 6 H2O, 2 CaCl2 · 2 H2O, and 11 glucose. The aCSF used throughout the experiments was continuously aerated with carbogen (95% O2-5% CO2). Transverse hippocampus slices (300 µm) were prepared along the dorsal-ventral axis using a vibratome (TPI, O'Fallon, MO). Slices were transferred to and maintained in a holding chamber filled with aCSF at room temperature (22°C). Only slices from the middle one-third of the left hippocampus were used for this study. All experiments were conducted 
1.5 hours after slice preparation at room temperature.
Electrophysiology
Patch electrodes were made from filamented, thick-walled borosilicate glass pipettes (Sutter Instrument, Novato, CA), using a Flaming-Brown horizontal puller (P-97, Sutter Instrument), and heat-polished with a microforge (model MF-930, Narishige International, East Meadow, NY) to a resistance of 2–4 M
when filled with two internal solutions consisting of the following (in mM): 1) 140 KMeSO4, 10 KCl, 10 HEPES, 4 Mg2ATP, and 0.4 Na3GTP or 2) 130 KMeSO4, 10 KCl, 10 HEPES, 10 BAPTA, 4 Mg2ATP, and 0.4 Na3GTP. In some experiments, 30 mM BAPTA was used to chelate Ca2+ maximally. In a subset of neurons, 0.3% biocytin or 2% Lucifer yellow was added to the patch solution to label the recorded neurons for further morphological identification. The pH of these solutions was adjusted to 7.25 with KOH; the final osmolarities of these solutions were 290 mOsM. Liquid junction potential (8 mV) was not corrected.
A stimulating electrode was placed in the s. radiatum, roughly 300 µm distal and 100 µm lateral from the perpendicular axis of the cell layer and recorded neurons. Area CA3 was cut away to prevent recurrent excitation of area CA1 due to repeated synaptic stimulation. The AHP in CA1 pyramidal neurons contains a component mediated by the apamin/bicuculline-sensitive Ca2+-activated K+ current (IAHP) (Oh et al. 2000; Stocker et al. 1999). The time course of this current precedes and somewhat overlaps with the time course of sIAHP. Therefore 10 µM bicuculline methiodide was routinely added to the aCSF to block the GABAA-mediated inhibitory postsynaptic potential as well as IAHP (Debarbieux et al. 1998; Johnson and Seutin 1997; Khawaled et al. 1999). GABAB receptor activation has been shown to activate a G-protein–activated, inwardly rectifying K+ current (GIRK) that could be blocked by Ba2+ (Sodickson and Bean 1996; Takigawa and Alzheimer 2003). Thus we also tested for the presence of the GIRK-mediated hyperpolarization in response to synaptic stimuli (n = 3). A Ba2+ (200 µM)-sensitive component in the sAHP was not observed using our stimulation protocol and perfusing ACSF. Instead, we observed it only in the presence of CNQX (20 µM; which resulted in an increase of GABA release; Brickley et al. 2001; Maccaferri and Dingledine 2002) and with higher stimulation intensity than what was used to study the AHP. Therefore GABAB receptors were not blocked.
Whole cell current-clamp recordings were made with an Axopatch 1C amplifier (Axon Instruments, Union City, CA), operating in current-clamp mode, on the soma of CA1 pyramidal neurons. Synaptic stimuli were generated by a constant voltage stimulus isolator (Digitimer). CA1 pyramidal neurons were visualized using a Zeiss Axioskop (Carl Zeiss, Oberkochen, Germany) microscope equipped with a long working distance 40x water immersion objective and infrared differential interference contrast (IR-DIC) optics. Seal resistances were >2 G prior to breakthrough into the whole cell mode. All measurements were made 15 min after rupturing the membrane to allow for adequate solution equilibration.
Protocols to trigger the AHP
The AHP was triggered with trains of 0.2-ms square voltage pulses delivered to the s. radiatum via a concentric bipolar electrode (FHC, Bowdoinham, ME) at frequencies from 20 to 150 Hz. For comparison, the AHP was also triggered with trains of 1- to 2-ms somatic depolarizing steps at corresponding frequencies.
The cells were maintained at –68 mV with either depolarizing or hyperpolarizing current injection unless otherwise stated. Previous studies have shown that the AHP grows with successive stimulus presentation (Lancaster and Adams 1986; Madison and Nicoll 1984). Thus, in this study, the "threshold value" we used for suprathreshold stimulus intensity was set such that a train of five stimuli at 50 Hz consistently triggered five corresponding action potentials to insure that sufficient number of presynaptic fibers were activated so as to reduce postsynaptic action potential failure during longer stimulus trains. Data gathered from neurons with resting membrane potential < –60 mV, membrane resistance >65 M, series resistance <15 M, and spike height >110 mV from baseline potential were accepted for further analysis. Electrophysiological records were acquired using a PC in conjunction with a Digidata 1322A interface (Axon Instruments) at 5 or 10 kHz and filtered at 2 kHz with a low-pass Bessel filter. Stimulus generation and data acquisition was performed using Clampex9 (Axon Instruments).
Data analysis and statistics
The duration and integral of the total AHP were calculated from the offset of the last voltage pulse to the time that the membrane potential returned to baseline. Because the currents underlying the afterdepolarization (ADP) could be activated with just one suprathreshold stimulus, the duration and integral of the ADP, summating EPSPs triggered by subthreshold stimuli, and the depolarizing envelope triggered by short trains of suprathreshold stimuli were all calculated from the offset of the first voltage pulse to the time that the membrane potential returned to baseline. Statistical significance was determined using ANOVA, Fisher's PLSD, and Mann-Whitney U test. Pooled data from electrophysiological recordings are expressed as means ± SE. Data analysis and curve fitting were done with Clampfit9 (Axon Instruments) and Igor Pro 4.0 (WaveMetrics, Lake Oswego, OR). All statistical analyses were performed using Statview (SAS Institute, Cary, NC).
Histology for light microscopy
Some cells were recorded with 0.3% biocytin for later morphological identification. Following recording, hippocampus slices containing labeled neurons were fixed for 48 h at 4°C in 2–4% paraformaldehyde and 15% picric acid in 0.1 M phosphate buffer (pH 7.3–7.4). They were subsequently reacted in a 1:100 dilution of avidin-biotin complex conjugated to horseradish peroxidase (ABC Elite kit, Vector Laboratories, Burlingame, CA) for 2 h at room temperature and incubated in 0.1 M Tris-buffered saline containing 0.025% 3-3-diaminobenzidine tetrahydrochloride (DAB, Sigma, St. Louis, MO), 0.05% nickel chloride, and 0.006% hydrogen peroxide. Slices were dehydrated, coverslipped, and permanently mounted for morphological examination.
Drugs
KMeSO4 was purchased from ICN (Aurora, OH); D-AP5 was purchased from Tocris (Ellisville, MO). All other drugs were purchased from Sigma.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
A train of five suprathreshold stimuli delivered to the s. radiatum at 50 Hz triggered corresponding numbers of action potentials that rode on top of a depolarizing envelope, followed by a pronounced postburst AHP in CA1 pyramidal neurons (Figs. 1 and 2). Previous studies have yielded conflicting data regarding the ionic mechanisms underlying the hyperpolarizing potential following different forms of stimulation protocols (i.e., somatic depolarization, glutamate application, and synaptic stimulation) (Lancaster and Wheal 1984; Nicoll and Alger 1981). Thus we first wanted to identify the current mediating the slow component of the AHP (sAHP) activated by synaptic stimulation. The sAHP was sensitive to bath applications of isoproterenol (5–10 µM), a 
-adrenergic receptor agonist (Fig. 1A; n = 5), and carbachol (2–5 µM), a muscarinic receptor agonist (Fig. 1B; n = 5). When BAPTA (10 mM), a high affinity Ca2+ chelator, was included in the patch solution, the sAHP was never observed (Fig. 1C; n = 12). The sensitivity of the sAHP to -adrenergic and muscarinic modulations, as well as its dependence on Ca2+ for activation, indicated that the current underlying the synaptically activated sAHP is the classically defined slow Ca2+-activated K+ current, sIAHP (Benardo and Prince 1982; Cole and Nicoll 1984; Haas and Rose 1987; Hotson and Prince 1980; Lancaster and Adams 1986; Lancaster et al. 2001; Madison and Nicoll 1986; for review, see Storm 1990). Normally, the apamin/bicuculline-sensitive IAHP partially overlaps with sIAHP in CA1 pyramidal neurons. However, as bicuculline methiodide (10 µM) was a standard component of our recording aCSF, IAHP was suppressed in our recordings (Debarbieux et al. 1998; Johnson and Seutin 1997; Khawaled et al. 1999).
|
|
) or by voltage-gated Ca2+ influx when Na+ action potentials were suppressed with low concentrations of QX-314 in the patch solution (0.5–2 mM; Lancaster et al. 2001). In these studies, the membrane potentials were maintained at more depolarized levels, presumably to increase the K+ driving force and thereby optimize the conditions to trigger sIAHP. However, at a more hyperpolarized resting membrane potential, it is not clear whether a measurable sAHP can be triggered by synaptic stimuli that do not result in cell firing. To answer this question, we compared the total AHPs evoked at –68 mV by five subthreshold and five suprathreshold synaptic stimuli delivered at 50 Hz. When action potentials were not generated, only the faster AHP was observed (Fig. 2A; n = 50). In contrast, both the faster and the slow AHP were activated when suprathreshold stimuli that resulted in action potentials were presented. In most cases, a small sAHP was seen following just two suprathreshold synaptic stimuli at 50 Hz (Fig. 2B).
The sAHP evoked by somatic depolarizing steps is known to summate with successive action potentials (Lancaster and Adams 1986; Madison and Nicoll 1984), reflecting its sensitivity to intracellular Ca2+ accumulation (Sah 1992; Sah and Clements 1999). Given the heterogeneous distribution of Ca2+ channels and Ca2+-permeable receptors along the somato-dendritic axis (Benke et al. 1993; Christie et al. 1995; Yin et al. 1999; for review, Magee 1998) and the differential contribution of Ca2+ influxes via different Ca2+ channel subtypes to activate the sAHP (Shah and Haylett 2000; Tanabe et al. 1998), we next examined the summation profile for the sAHP triggered by different patterns of synaptic stimuli (Fig. 2, B and C). At all frequencies examined, the integral of the total AHP, largely reflecting the sAHP (Fig. 3), increased with each successive stimulus (Fig. 2D). The growth of the AHP integrals were all well fit with mono-exponential functions, with the AHP growth constants decreasing with increasing stimulation frequency. The maximal AHP integral reached was significantly greater for the 50- and 100-Hz trains than for a 20-Hz train (P < 0.01; ANOVA), suggesting that the limiting factors for maximal sAHP are frequency-dependent. Correspondingly, the total AHP integral triggered by a train of 15 somatically activated action potentials at 100 Hz was also significantly larger than that triggered by the same train at 20 Hz (100 Hz = 12.3 ± 1.1 mV · s; 20 Hz = 7.1 ± 0.8 mV · s; P < 0.01; ANOVA).
|
The duration of the total AHP activated by a train of five suprathreshold stimuli at 50 Hz ranged from 88.1 ms to 18.3 s (6.6 ± 0.5 s; n = 50). The distributions for the duration and integral of total AHP revealed two distinct, nonoverlapping peaks (Fig. 3A): 20% of the neurons (n = 10) showed a total AHP duration <1 s and an AHP integral <1 mV · s, whereas 80% of the neurons showed a total AHP duration >3.7 s and an AHP integral >2.3 mV · s. When BAPTA (10 mM) was included in the patch solution, the duration of the total AHP that was Ca2+-insensitive was determined to be 0.7 ± 0.03 s (n = 12). Thus we considered the presence of a hyperpolarizing potential lasting beyond 1 s after pulse offset to be indicative of the sAHP (Fig. 3B). No statistical difference was observed in the resting membrane potentials (with a sAHP = –65.9 ± 0.7 mV, n = 40; without a sAHP = –69.0 ± 1.5 mV, n = 10) or membrane resistances (with a sAHP = 120.8 ± 5.7 M, n = 40; without a sAHP = 112.6 ± 10.3 M, n = 10) between these two categories of neurons (P > 0.05, ANOVA; Table 1). The total AHP duration and integral were significantly larger for neurons with a sAHP than for those without (P < 0.0001, ANOVA), indicating that they predominantly reflect the sAHP (Fig. 3C).
|
Effect of the sAHP on temporal summation
The depolarizing integral triggered by one suprathreshold synaptic stimulus was not different for these two categories of neurons (with a sAHP = 934.9 ± 94.5 mV · ms; without a sAHP = 900.0 ± 114.4 mV · ms; P > 0.05); neither was the temporal summation of EPSPs triggered by a short train of subthreshold stimuli at 50 Hz (Fig. 4A ). In contrast, the depolarizing envelope triggered by a train of suprathreshold stimuli was much shorter for neurons with a sAHP (Fig. 4, B and C). This difference became more pronounced as the stimulus trains lengthened, with the total depolarizing integral and duration triggered by 10–15 suprathreshold stimuli significantly curtailed for neurons with a sAHP compared with neurons without (P < 0.05 and 0.005, respectively, ANOVA; Fig. 4C). For the latter group, the depolarizing envelope would overcome the fast component of the AHP and evolve into a pronounced ADP (Fig. 5A ), which could be further prolonged by lengthening the stimulus train or increasing the stimulation frequency (Fig. 5, A and B). Multiple action potential firing was commonly observed during the ADP (Fig. 6A ).
|
|
|
150 Hz; Fig. 5C). These data indicate that the development of ADP in response to high-frequency inputs was a result of ionotropic glutamate receptor activation and not due to direct electrical stimulation of the recorded neurons. Ionic mechanisms underlying ADP
To examine the ionic mechanisms underlying ADP, we searched for and recorded from additional CA1 neurons without a measurable sAHP (n = 11). Neurons were discarded if they exhibited a postburst AHP duration >1 s in response to five suprathreshold synaptic stimuli at 50 Hz. All neurons that met this selection criterion developed an ADP of varying magnitude and duration when presented with 15 suprathreshold stimuli at 150 Hz (ADP integral with 1.5 times suprathreshold stimulus intensity = 75.0 ± 17.8 mV · s; ADP duration = 7.1 ± 0.9 s). At this stimulation frequency, the magnitude and duration of the ADP increased exponentially in response to increases in stimulus intensity (Fig. 6A).
It has previously been shown that an afterdepolarizing potential can be induced by exogenous muscarinic receptor agonist application (Egorov et al. 1999; Fraser and MacVicar 1996). To determine whether the ADP reported in this study is also the result of muscarinic receptor activation, we triggered an ADP with 15 synaptic stimuli at 150 Hz, using 1.5 times suprathreshold stimulus intensity. Bath application of atropine (1 µM), an antagonist of the muscarinic receptors, had no effect on the ADP, indicating that muscarinic receptor activation is not involved in generating this afterpotential (n = 2; data not shown). In nigral dopamine neurons, blockade of an AHP current (IAHP) has been shown to lead to an afterdepolarizing potential, mediated primarily by the L-type Ca2+ channels (Ping and Shepard 1999). In the cells of the lateral geniculate nucleus, an afterdepolarizing plateau potential mediated by the L-type Ca2+ channels and NMDA receptors has also been described (Lo et al. 2002). Thus we next examined the effects of nitrendipine, an L-type Ca2+ channel blocker, and D-AP5 on the ADP. Because Ca2+ influxes through the L-type Ca2+ channels, the NMDA receptors, as well as that associated with action potential firing, have all been shown to trigger intracellular Ca2+ release (CICR; Emptage et al. 1999; Sandler and Barbara 1999), we also examined the effect of thapsigargin, an inhibitor for intracellular Ca2+ store release, on the ADP. Bath applications of thapsigargin (1 µM) caused a 20.3 ± 1.9% reduction in the ADP integral (Fig. 6, B and C; n = 5; P < 0.0001; Fisher's PLSD). Additional applications of nitrendipine (10–20 µM) to the bath caused a 52.1 ± 7.07% reduction in the ADP integral relative to control (Fig. 6, B and C; n = 5; P < 0.0001 compared with control and thapsigargin's effect; Fisher's PLSD). Further applications of D-AP5 (50 µM) caused a 89.8 ± 1.6% reduction in the ADP integral relative to control (Fig. 6, B and C; n = 5; P <0.0001 compared with control, thapsigargin's, and nitrendipine's effects, Fisher's PLSD). Thus Ca2+ influxes via the L-type Ca2+ channels and NMDA receptors, as well as CICR, all contribute to the generation of the ADP. When D-AP5 was applied alone, it caused a 90.9 ± 2.3% reduction in the ADP integral (Fig. 6, C and D; n = 4; P < 0.0001 Fisher's PLSD). In neurons that the ADPs were abolished by bath-applications of D-AP5 alone, increasing the stimulus intensity to the maximum allowed by the stimulus isolator without damaging the fiber path was ineffective in triggering an ADP (Fig. 6D), suggesting that the NMDA receptor–mediated potential is a necessary trigger to initiate the ADP.
Modulation of the sAHP leads to the development of an ADP
The depolarizing envelope associated with synaptic stimuli was much smaller when the sAHP was present. Thus we next examined the consequence of suppressing the sAHP on postsynaptic processing of high-frequency suprathreshold synaptic inputs. When BAPTA (10–30 mM) was included in the patch solution, all neurons recorded developed a pronounced ADP lasting 5 s that could be blocked by D-AP5 (50–150 Hz; n = 20; Fig. 7A ). Bath applications of carbachol (2–5 µM; n = 3; Fig. 7B) or isoproterenol (5 µM; n = 5; Fig. 7C) abolished the sAHP, and in response to a train of high-frequency synaptic stimuli, the neurons instead developed an ADP. In the presence of carbachol, two neurons showed a decrease in the number of action potentials triggered by the same train of suprathreshold stimuli compared with control. Therefore stimulus intensity was increased to maintain the same number of action potentials generated. The effect of carbachol was mostly reversed by additional application of atropine (1 µM) to the bath solution (Fig. 7B). Together, these data indicate that neuromodulation of the sAHP can powerfully modulate the postsynaptic response to synaptic inputs.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
The postburst AHP evoked by synaptic stimuli under our recording conditions contained two kinetically distinct components. The faster component is Ca2+ insensitive, as it persisted in patch solution containing BAPTA (Fig. 1C), and was activated by both sub- and suprathreshold stimuli. These characteristics suggest that the faster component reported in this study is similar to the classically defined medium AHP mediated by both IM and Ih (Takigawa and Alzheimer 2003; Williamson and Alger 1990). We did not evaluate the relative contribution of these currents to the mAHP. However, Ih channels are more densely located on the dendrite (Lorincz et al. 2002) and IM is activated at membrane potentials positive to –60 mV (Halliwell 1986). Thus at our recording potential of –68 mV, Ih deactivation most likely played a larger role in the mAHP evoked by subthreshold synaptic stimuli. The slower component of the AHP, on the other hand, is Ca2+ sensitive, and its activation from a membrane potential of –68 mV required Ca2+ influx associated with action potential generation (Fig. 2A). The sAHP was blocked by bath-applications of carbachol or isoproterenol (Fig. 1, A and B). Given that bicuculline methiodide was a standard component of our recording ACSF, the apamin-sensitive IAHP was suppressed in our recordings. Thus our data indicate that the current underlying the sAHP is the apamin-insensitive, Ca2+-activated K+ current, sIAHP (Benardo and Prince 1982; Cole and Nicoll 1984; Haas and Rose 1987; Hotson and Prince 1980; Lancaster and Adams 1986; Lancaster et al. 2001; Madison and Nicoll 1986; for review, see Storm 1990).
For a given stimulation frequency, the sAHP summated with each successive stimulus presentation (Fig. 2B); for a given number of stimuli, the sAHP increased with respect to the stimulation frequency (Fig. 2C). The AHP growth profile as plotted in Fig. 2D suggests that the degree of sAHP activation reflects the rate of rise and accumulation of intracellular Ca2+ (Sah 1992; Sah and Clements 1999). In support, a recent study has shown that the time constants for the rise in intracellular Ca2+, measured both at the soma and dendrite, decreases exponentially with respect to increasing firing rate of the cell (Abel et al. 2004). The maximal sAHP activated by a train of suprathreshold synaptic stimuli (or by corresponding number of action potentials evoked with depolarizing somatic current steps) at 50 or 100 Hz was significantly larger than that activated by the same train at 20 Hz. Given that the sIAHP does not inactivate, and its decay is thought to reflect the return of intracellular Ca2+ to baseline level by various Ca2+ buffering/clearance mechanisms (Lancaster and Zucker 1994; Sah and Clements 1999), the magnitude of the sAHP is determined by the number of the activated sAHP channels and the various Ca2+ buffering/clearance pathways in the cell. Our interpretation is that the limiting factor for the maximal AHP activated by a 50- or 100-Hz train is the saturation of the functional sAHP channels. In contrast, the maximal AHP activated by a 20-Hz train probably reflected a quasi-steady-state level of intracellular Ca2+ achieved by the activity-dependent activation of various Ca2+ buffering/clearance mechanisms to counter additional Ca2+ influx associated with subsequent stimuli.
The sAHP was not activated by subthreshold stimuli (Fig. 2A). Consistently, temporal summation of EPSPs was not different for neurons with and without a sAHP (Fig. 4A). Once the sAHP was activated by suprathreshold stimuli, it significantly curtailed the magnitude of postsynaptic depolarization triggered by subsequent stimuli (Fig. 4, B and C). Such activity-dependent activation suggests that the sAHP may be an intrinsic mechanism that enhances the signal-to-noise ratio of meaningful inputs to CA1 pyramidal neurons. Weak inputs do not activate the sAHP. However, once action potentials are generated, the sAHP would reduce subsequent, weaker inputs, thus highlighting those signals that result in cell firing from background synaptic noise.
Some CA1 neurons did not have a measurable sAHP
A small population of CA1 pyramidal neurons (20%) did not have a measurable sAHP. In response to high-frequency, intense synaptic stimulation, these cells developed a distinctive ADP (Fig. 5), mediated by NMDA receptors, the L-type Ca2+ influx, and intracellular Ca2+ release. The total depolarizing integral resulting from just one suprathreshold synaptic stimulus for both categories of neurons was not statistically different, suggesting that the number of synapses being activated was not different. Given that suprathreshold synaptic activation leads to widespread elevations of Ca2+ and Na+ throughout the dendritic tree—an effect caused by backpropagating action potentials and subsequent openings of voltage-gated Ca2+ channels as well as simultaneous activation of the NMDA receptors (Jaffe et al. 1992; Markram et al. 1995; Rose and Konnerth 2001; Spruston et al. 1995; Yuste and Denk 1995), the absence of the sAHP in these cells was unlikely the result of insufficient Ca2+ for channel activation. In the presence of atropine, these cells still developed an ADP, indicating that differential involvement of muscarinic neuromodulation cannot account for the presence or absence of the sAHP in these cells.
ADP and NMDA receptors
The ADP reported here differs mechanistically from the slow ADP and plateau potential derived from muscarinic receptor activation (Fraser and MacVicar 1996). It is mediated by influx through the NMDA receptors, the L-type Ca2+ channels, and CICR (Fig. 7, A and B). In the presence of D-AP5, further increases in the stimulus intensity failed to trigger an ADP mediated by the latter two components (Fig. 7C). These data suggest that the NMDA receptor–mediated influx alone is sufficient to support an ADP and is necessary for the activation of L-type Ca2+ influx and CICR. Similar interactions between the voltage-gated Ca2+ influx or intracellular Ca2+ release and NMDA receptors have previously been reported (Calton et al. 2000; Schiller et al. 1997; Schwindt and Crill 1998). A Ca2+-activated cation current (ICAN), activated by the muscarinic or metabotropic glutamate receptors, has been found to underlie an afterdepolarizing potential in hippocampal pyramidal neurons (Caeser et al. 1993; Congar et al. 1997; Fraser and MacVicar 1996; Greene et al. 1994; Young et al. 2004). We did not test for the involvement of ICAN in the ADP triggered by high-frequency synaptic stimuli. However, an ADP was induced in all neurons loaded with BAPTA. While this does not exclude the involvement of ICAN, it does indicate that ICAN is not necessary to support this form of ADP.
In the presence of BAPTA, all neurons developed a prolonged ADP that was sensitive to D-AP5 in response to a train of high-frequency synaptic stimuli (Fig. 7A). The inclusion of BAPTA in the patch solution most likely prevented CICR from contributing to the ADP (20%); however, it would not eliminate the charges flowing through the NMDA receptors and voltage-dependent Ca2+ channels. The pronounced ADP we observed in BAPTA-loaded neurons may also in part be caused by the removal of Ca2+-dependent inactivation of the NMDA receptors and the L-type Ca2+ channels (Hofer et al. 1997; Imredy and Yue 1994; Legendre et al. 1993; Lu et al. 2000; Neely et al. 1994; Rosenmund and Westbrook 1993).
The magnitude and duration of ADP was dependent on stimulation intensity and frequency. The rate of growth was best fit with an increasing exponential function, consistent with the active properties of the NMDA receptor channels and their activity-dependent amplification (for review, see Schiller and Schiller 2001). Action potentials are typically initiated in the axon and then backpropagate into the dendrites of CA1 pyramidal neurons, initiating dendritic Ca2+ influx (Spruston et al. 1995). Trains of action potentials initiated by somatic depolarizing steps failed to trigger an ADP at all frequencies tested, implying that backpropagating action potentials alone are insufficient. Rather, an ADP required the coincidence of repetitive backpropagating action potentials and NMDA receptor activation. Previous Na+ and Ca2+ imaging studies have shown an accumulation of Na+ transients and supralinear Ca2+ signals mediated by NMDA receptors, detected during coincident occurrence of synaptic potentials and backpropagating spikes (Koester and Sakmann 1998; Magee and Johnston 1997; Markram et al. 1997; Rose and Konnerth 2001; Schiller et al. 1998; Yuste and Denk 1995). Schiller et al. (1998) also showed that the pairing of caged glutamate release with postsynaptic action potentials selectively amplified the NMDA receptor–mediated Ca2+ signals. Similarly, when subthreshold EPSPs were paired with back-propagating action potentials, dendritic action potentials were amplified, Ca2+ influx was evoked near the site of synaptic input, and long-term potentiation (LTP) was observed (Magee and Johnston 1997). Our data suggest that the sAHP places a powerful regulation on this Hebbian-type of association through its interactions with NMDA receptors, thereby affecting synaptic efficacy. In support, an association has been observed between the reduction of the sAHP and the priming of the induction for LTP (Cohen et al. 1999).
The sAHP channels are thought to be situated near the proximal apical dendrite and/or soma (Bowden et al. 2001; Sah and Bekkers 1996; but also see Bekkers 2000). Given the sensitivity of backpropagating action potentials to hyperpolarization (Tsubokawa and Ross 1996), activation of the sAHP can shunt signals propagating both from and toward the dendrite, thereby effectively uncoupling the somato-dendritic compartments and restricting the spread of signals. In addition, the sAHP can limit temporal summation by interacting with the active properties of CA1 pyramidal neurons. During suprathreshold synaptic transmission, Ca2+ influx through the NMDA receptors and voltage-gated Ca2+ channels would activate sIAHP, thereby driving the membrane potential toward the hyperpolarizing direction that facilitates the voltage-dependent Mg2+ re-block of the NMDA receptors. Under conditions that the sAHP is suppressed (e.g., neuromodulation), temporal summation of high-frequency inputs would depolarize the neurons more easily to a membrane potential that favors the removal of the Mg2+ block of the NMDA receptors, plus keep the fast A-type K+ current (IA) in an inactivated state (Magee and Carruth 1999). Synaptic inputs occurring during this window, with NMDA receptors unblocked by Mg2+ and IA inactivated, would thus be able to trigger an amplified NMDA receptor–mediated response in the form of an ADP that is capable of initiating multiple action potentials (Fig. 5). In support, we have shown that pharmacologically blocking the sAHP with BAPTA, carbachol, or isoproterenol allowed the neurons to develop an ADP (Fig. 7). We propose that the sAHP is a postsynaptic mechanism that is critical in limiting an NMDA receptor–dependent regenerative ADP during bouts of high-frequency, intense synaptic transmission. Neuromodulation of the sAHP is thus a powerful postsynaptic mechanism in shaping the postsynaptic membrane response to strong inputs.
sAHP and ADP in the context of cellular plasticity during learning
Given that the sAHP regulates the propensity of the postsynaptic neuron to develop an NMDA receptor–dependent ADP, the reduction in the sAHP in learning (Moyer et al. 1996, 2000; Oh et al. 2003) may be playing a permissive or facilitating role for further synaptic changes that are also developing during hippocampus-dependent learning. There is evidence suggesting that the neuromodulation of the sAHP participates in establishing various forms of plasticity thought to be involved in learning. Blockade of the sAHP with isoproterenol, an agonist of the -adrenergic receptors, has been shown to convert short-term potentiation triggered by a weak tetanus train to LTP (Sah and Bekkers 1996). LTP induction triggered by a mild theta burst stimulation protocol was also shown to be "primed" by suppression of the sAHP with isoproterenol and ACPD, an mGluR agonist (Cohen et al. 1999). Together, these studies link the expression of long-term plasticity with the sAHP. We have identified neuroanatomical changes, increases in postsynaptic density area in the s. radiatum, and the total number of multiple-synapse boutons in CA1 pyramidal neurons after learning trace eyeblink conditioning (Geinisman et al. 2000, 2001). CA1 field potentials in response to s. radiatum stimulation were also increased (Power et al. 1997). Acquisition of the learned response was facilitated with cholinergic agents or Ca2+ channel antagonists that reduce the sAHP (Deyo et al. 1989; Kowalska and Disterhoft 1994; Kronforst-Collins et al. 1997; Moyer et al. 1992; Oh et al. 1999; Power et al. 2001; Weiss et al. 2000) and with co-agonists to the glycine site on the NMDA receptors in young and aging animals (Thompson et al. 1992; Thompson and Disterhoft 1997a). Conversely, trace eyeblink conditioning was impaired with antagonists to NMDA-mediated transmission, MK801 and PCP (Thompson and Disterhoft 1997a, b). Together, our data suggest that the increased information throughput occurring during learning may reflect synaptic alterations that are dependent for their development on cellular excitability changes such as a reduction in the sAHP.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: J. F. Disterhoft, Dept. of Physiology, Inst. for Neuroscience, Northwestern Univ. Feinberg School of Medicine, 303 E. Chicago Ave., Chicago, IL 60611-3008 (E-mail: jdisterhoft{at}northwestern.edu).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Bekkers JM. Distribution of slow AHP channels on hippocampal CA1 pyramidal neurons. J Neurophysiol 83: 1756–1759, 2000.
Benardo LS and Prince DA. Ionic mechanisms of cholinergic excitation in mammalian hippocampal pyramidal cells. Brain Res 249: 333–344, 1982.[CrossRef][ISI][Medline]
Benke TA, Jones OT, Collingridge GL, and Angelides KJ. N-methyl-D-aspartate receptors are clustered and immobilized on dendrites of living cortical neurons. Proc Natl Acad Sci USA 90: 7819–7823, 1993.
Bowden SE, Fletcher S, Loane DJ, and Marrion NV. Somatic colocalization of rat SK1 and D class (Ca(v)1.2) L-type calcium channels in rat CA1 hippocampal pyramidal neurons. J Neurosci 21: 1–6, 2001.
Brickley SG, Farrant M, Swanson GT, and Cull-Candy SG. CNQX increases GABA-mediated synaptic transmission in the cerebellum by an AMPA/kainate receptor-independent mechanism. Neuropharmacology 41: 730–736, 2001.[CrossRef][ISI][Medline]
Caeser M, Brown DA, Gahwiler BH, and Knopfel T. Characterization of a calcium-dependent current generating a slow afterdepolarization of CA3 pyramidal cells in rat hippocampal slice cultures. Eur J Neurosci 5: 560–569, 1993.[CrossRef][ISI][Medline]
Calton JL, Kang MH, Wilson WA, and Moore SD. NMDA-Receptor-dependent synaptic activation of voltage-dependent calcium channels in basolateral amygdala. J Neurophysiol 83: 685–692, 2000.
Christie BR, Eliot LS, Ito K, Miyakawa H, and Johnston D. Different Ca2+ channels in soma and dendrites of hippocampal pyramidal neurons mediate spike-induced Ca2+ influx. J Neurophysiol 73: 2553–2557, 1995.
Cohen AS, Coussens CM, Raymond CR, and Abraham WC. Long-lasting increase in cellular excitability associated with the priming of LTP induction in rat hippocampus. J Neurophysiol 82: 3139–3148, 1999.
Cole AE and Nicoll RA. The pharmacology of cholinergic excitatory responses in hippocampal pyramidal cells. Brain Res 305: 283–290, 1984.[CrossRef][ISI][Medline]
Congar P, Leinekugel X, Ben-Ari Y, and Crepel V. A long-lasting calcium-activated nonselective cationic current is generated by synaptic stimulation or exogenous activation of group I metabotropic glutamate receptors in CA1 pyramidal neurons. J Neurosci 17: 5366–5379, 1997.
Debarbieux F, Brunton J, and Charpak S. Effect of bicuculline on thalamic activity: a direct blockade of IAHP in reticularis neurons. J Neurophysiol 79: 2911–2918, 1998.
Deyo RA, Straube KT, and Disterhoft JF. Nimodipine facilitates associative learning in aging rabbits. Science 243: 809–811, 1989.
Egorov AV, Gloveli T, and Muller W. Muscarinic control of dendritic excitability and Ca(2+) signaling in CA1 pyramidal neurons in rat hippocampal slice. J Neurophysiol 82: 1909–1915, 1999.
Emptage N, Bliss TV, and Fine A. Single synaptic events evoke NMDA receptor-mediated release of calcium from internal stores in hippocampal dendritic spines. Neuron 22: 115–124, 1999.[CrossRef][ISI][Medline]
Fraser DD and MacVicar BA. Cholinergic-dependent plateau potential in hippocampal CA1 pyramidal neurons. J Neurosci 16: 4113–4128, 1996.
Geinisman Y. Structural synaptic modifications associated with hippocampal LTP and behavioral learning. Cereb Cortex 10: 952–962, 2000.
Geinisman Y, Berry RW, Disterhoft JF, Power JM, and Van der Zee EA. Associative learning elicits the formation of multiple-synapse boutons. J Neurosci 21: 5568–5573, 2001.
Greene CC, Schwindt PC, and Crill WE. Properties and ionic mechanisms of a metabotropic glutamate receptor-mediated slow afterdepolarization in neocortical neurons. J Neurophysiol 72: 693–704, 1994.
Haas HL and Rose GM. Noradrenaline blocks potassium conductance in rat dentate granule cells in vitro. Neurosci Lett 78: 171–174, 1987.[CrossRef][ISI][Medline]
Halliwell JV. M-current in human neocortical neurones. Neurosci Lett 67: 1–6, 1986.[CrossRef][ISI][Medline]
Hofer GF, Hohenthanner K, Baumgartner W, Groschner K, Klugbauer N, Hofmann F, and Romanin C. Intracellular Ca2+ inactivates L-type Ca2+ channels with a Hill coefficient of approximately 1 and an inhibition constant of approximately 4 microM by reducing channel's open probability. Biophys J 73: 1857–1865, 1997.
Hotson JR and Prince DA. A calcium-activated hyperpolarization follows repetitive firing in hippocampal neurons. J Neurophysiol 43: 409–419, 1980.
Imredy JP and Yue DT. Mechanism of Ca(2+)-sensitive inactivation of L-type Ca2+ channels. Neuron 2: 1301–1318, 1994.
Jaffe DB, Johnston D, Lasser-Ross N, Lisman JE, Miyakawa H, and Ross WN. The spread of Na+ spikes determines the pattern of dendritic Ca2+ entry into hippocampal neurons. Nature 357: 244–246, 1992.[CrossRef][Medline]
Johnson SW and Seutin V. Bicuculline methiodide potentiates NMDA-dependent burst firing in rat dopamine neurons by blocking apamin-sensitive Ca2+-activated K+ currents. Neurosci Lett 231: 13–16, 1997.[CrossRef][ISI][Medline]
Khawaled R, Bruening-Wright A, Adelman JP, and Maylie J. Bicuculline block of small-conductance calcium-activated potassium channels. Pfluegers 438: 314–321, 1999.
Koester HJ and Sakmann B. Calcium dynamics in single spines during coincident pre- and postsynaptic activity depend on relative timing of back-propagating action potentials and subthreshold excitatory postsynaptic potentials. Proc Natl Acad Sci USA 95: 9596–9601, 1998.
Kowalska M and Disterhoft JF. Relation of nimodipine dose and serum concentration to learning enhancement in aging rabbits. Exp Neurol 127: 159–166, 1994.[CrossRef][ISI][Medline]
Kronforst-Collins MA, Moriearty PL, Ralph M, Becker RE, Schmidt B, Thompson LT, and Disterhoft JF. Metrifonate treatment enhances acquisition of eyeblink conditioning in aging rabbits. Pharmacol Biochem Behav 56: 103–110, 1997.
; 50 Hz, 
; 100 Hz, 
). Growths of AHP integrals were fit with monoexponential functions in the following form: Amax x (1 – exp{–[(no. stimuli – 1)/k]}), where Amax is the maximal total AHP integral, and k is the growth constant.
