INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Epilepsy is characterized by two subtypes of electrographic discharges: asymptomatic inter-ictal spikes and symptomatic ictal seizures (for review, see Dichter et al. 1998; Neidermeyer and Da Silva 1999). In vivo recordings from various animal models of epilepsy revealed that the intracellular correlate of inter-ictal epileptic spikes is a stereotypic prolonged high-amplitude depolarizing response with overriding action potentials termed paroxysmal depolarization shift (PDS). The intracellular correlate of electrographic seizures is recurrent bursts of action potentials overriding depolarizing envelopes with variable duration and amplitude (for review, see Dichter et al. 1998; Schwartzkroin and Wyler 1980; Speckmann and Erwin 1999).

Synchronous synaptic currents and active dendritic conductances are thought to play a key role in initiation and propagation of inter-ictal spikes and electrographic seizures. According to this view, inter-ictal and ictal epileptic discharges are mediated in part by dendritic voltage-gated conductances, and glutamate-receptor channels (for review, see Crill and Schwindt 1999; McCormick and Contreras 2001; Prince and Connors 1986; Schwartzkroin and Wyler 1980; Traub and Jefferys 1994; Traub et al. 1996). As support to the preceding view, it has been shown in recent years that the dendritic membrane of pyramidal neurons contains voltage-gated channels, which participate in shaping synaptic potentials and mediate dendritic spikes (for review, see Hausser et al. 2000; Johnston et al. 1998; Yuste and Tank 1996).

Despite the putative importance of dendritic mechanisms to the formation of epileptic discharges only a handful of direct dendritic recordings have been reported in the literature during epileptic discharges in hippocampal neurons (Traub et al. 1993; Wong and Prince 1978). The majority of data concerning epileptic discharges in dendrites has been obtained indirectly using somatic recordings and computer modeling.

In this study, I used simultaneous voltage recordings from the soma and apical dendrite of the same neuron combined with calcium-imaging measurements in BCC treated brain slices to characterize inter-ictal- and ictal-like epileptic discharges in dendrites of neocortical layer-5 pyramidal neurons.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Slice preparation and electrophsiological recordings

Parasagital neocortical brain slices (350 µm thick) were prepared from 12- to 35-day-old rats, as previously described (Schiller et al. 1995; Stuart and Sakmann 1994). Thick tufted layer-5 pyramidal neurons were visualized using infrared illumination and differential interference contrast optics (IR-DIC) video microscopy. The microscope used was a fixed stage Axioscope (Zeiss) equipped with a ×60 water-immersion objective (NA 0.9). Whole cell voltage recordings from the soma, apical dendrites or simultaneous somatic and dendritic recordings from the same neuron were obtained as previously described (Stuart and Sakmann 1994). Somatic (3-5 M resistance) and dendritic (7-11 M) recording pipettes were filled with (in mM) 115 K-gluconate, 20 KCl, 2 Mg-ATP, 2 Na₂-ATP, 10 Na₂-phosphocreatine, 0.3 GTP, 10 HEPES, pH 7.2, and either 0.1 Calcium Green-1 or 0.5 Mag-fura. The extracellular solution contained (in mM) 125 NaCl, 25 NaHCO₃, 25 glucose, 3 KCl, 1.25 NaH₂PO₄, 2 CaCl₂, and 1 MgCl₂, pH 7.4. All experiments were performed in 35 ± 0.5°C. Synaptic stimulation was performed using monopolar platinum iridium electrodes (World Precision Instruments, Sarasota, FL). All chemicals were purchased from Sigma aside from the Calcium Green-1 and Mag-fura (Molecular Probes, Portland, OR) and bicuculline (BCC), 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), and DL-2-Amino-5-phosphonovalerate (APV) (Tocris, Bristol, UK). All chemical were dissolved in water aside from D600 (methoxy-verapamil), which was dissolved in DMSO. In the experiments the DMSO reached a final concentration of 0.1%.

Fluorescence measurements

High-speed [Ca²⁺]_i fluorescence-imaging measurements were performed using a frame transfer cooled CCD camera (ATC-5, Photometrics, Tucson, AZ) and the calcium-sensitive fluorescence dyes Calcium Green-1 (100 µM) and Mag-fura (400 µM). Calcium Green-1 was excited at 488 nm. Mag-fura was excited at 360 and 350 nm (the isobastic wavelength for Mag-fura in our conditions). Fluorescence measurements (acquisition rate of 40-100 Hz) were collected from small regions of interest (width: 3-5 µm, length: 10-25 µm). Background subtraction and bleach correction were performed as previously described (Schiller et al. 1995). Calcium Green-1 fluorescence values were presented as F/F in % (F/F = 100 · (F  F0)/F0, where F was the fluorescence measured at different times and F0 was the average baseline fluorescence). Mag-fura fluorescence values were converted to calcium concentrations using the isobastic ratio method (Schiller et al. 1995) (Kd of Mag-fura equals 50 µM).

Measurements of voltage responses

The peak voltage of the PDS and ictal bursts was performed after omitting the over-riding fast action potentials. In PDS responses and ictal burst in which fast action potentials were over-riding the peak of the PDS or ictal burst the measurements were performed in the segments in between the fast spikes. In double recordings, comparison of the relative timing of the somatic and dendritic PDS and ictal bursts was performed using different measuring methods. First, I compared the time in which the voltage reached the 10 and 50% value of the threshold for the first action potential initiation. Second, I compared the time in which the dendritic and somatic events started. This time was measured when the first noticeable change in the voltage was observed. In all cases, the different measurements yielded the same results.

All average results were presented as the means ± SD Statistical analysis was performed using either the Student's t-test or linear regression analysis. To determine whether r differs from 0, I used the F test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Epileptiform discharges in the soma of layer-5 pyramidal neurons

In the presence of 10 µM extracellular BCC, neocortical brain slices produced two subtypes of epileptic discharges: isolated paroxysmal depolarization shift (PDS) responses and electrographic seizures. As previously described, isolated PDS responses were evoked by extracellular synaptic stimulation in all slices (n = 90), and mimicked inter-ictal spikes (Fig. 1A).

View larger version (29K): [in this window] [in a new window]   Fig. 1. Inter-ictal and ictal epileptic discharges. An isolated paroxysymal depolarization shift (PDS) response (A) and electrographic seizure (B) recorded with a somatic whole cell pipette from 2 different layer-5 pyramidal neurons. Both responses were evoked by extracellular synaptic stimulation in the presence 10 µM bicuculline (BCC) in the bath solution.

Electrographic seizures lasted 6.6-31.2 s and consisted of recurrent bursts over-riding a prolonged depolarization waveform, which followed the entire course of electrographic seizures (Fig. 1B) (see also Hablitz 1987). The difference between the individual ictal bursts and the underlying prolonged depolarization envelope was that the individual ictal bursts were composed of a train of fast action potentials over-riding a PDS or EPSP like waveform. In contrast to the ictal bursts, the prolonged depolarization waveform followed the entire course of the seizure and, aside from infrequent instances, did not result directly in fast axonal action potential firing. Simultaneous recordings from pairs of neurons (45 neuron pairs) during isolated PDS responses and electrographic seizures revealed that the isolated PDS responses and the ictal bursts occurred concomitantly in both neurons, indicating ictal bursts probably represented a short synchronized "wave" of supra-threshold activity in the entire neuronal population of the slice.

In contrast to isolated PDS responses that were elicited in all slices of all ages examined (12-35 days), electrographic seizures were only observed in immature brain slices obtained from 12- to 17-day-old rats (Hablitz 1987). They occurred either spontaneously (71% of 90 slices examined) or in response to extracellular synaptic stimulation (88% of 90 slices examined) and mimicked epileptic seizures (Ayala et al. 1970; Matsumoto and Ajmone-Marsan, 1964; Speckmann and Elger 1999). It should be noted that repeated synaptic stimulation at constant stimulus intensities produced alternatively isolated PDS responses and electrographic seizures in the same cell.

Inter-ictal PDS responses in apical dendrites

To investigate PDS responses in the dendritic tree of neocortical neurons, whole cell recordings were performed from the apical dendrites of layer-5 pyramidal neurons. Similar to the somatic PDS response, the dendritic PDS was characterized by a large depolarizing response. To compare PDS responses in apical dendrites and the soma simultaneous whole cell recordings were performed from the soma and apical dendrite of the same layer-5 pyramidal neuron (Fig. 2). Paroxysmal depolarization shift responses were recorded in both the somatic and apical dendritic pipettes. However, the apical dendritic PDS was significantly higher in amplitude and narrower in duration than the corresponding somatic PDS (Fig. 2A). The average peak amplitude of the PDS (as measured after omitting the fast over-riding action potentials) was 68.9 ± 2.2 mV in apical dendrites and 48.3 ± 2.7 mV at the soma (n = 18 double recordings, P < 0.0001). The average half-width of the PDS was 203.8 ± 38.4 ms at the apical dendrites and 247.8 ± 57.3 ms at the soma (n = 18 double recordings, P < 0.01). The peak amplitude of the PDS recorded in the apical dendrite (100-550 µm from the soma) ranged from 7 to 18.8 mV in different experiments, and in 19 of the 22 dendritic recordings (100-550 µm from the soma) surpassed 0 mV. The reversal potential for both N-methyl-D-aspartate (NMDA) and alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) ionotropic glutamate-receptor channels is approximately 0 mV (Meyer and Westbrook 1987), and the only ions that have a positive reversal potential under physiological conditions are sodium and calcium. Hence, the positive voltage value the PDS reaches in apical dendrites indicates sodium or calcium channels participate in its formation.

View larger version (13K): [in this window] [in a new window]   Fig. 2. Comparison of PDS responses in the soma and apical dendrite of the same layer-5 neuron. A: simultaneous whole cell recordings from the soma and apical dendrite (340 µm from the soma) of the same layer-5 pyramidal neuron. Left: a schematic drawing of the neuron and the recording pipettes. Right: the simultaneous dendritic (top) and somatic (bottom) recordings of the PDS. The bottom 2 panels demonstrate the superimposed dendritic and somatic PDS responses presented at 2 different time scales. The 2 arrows marked the initiation of the somatic (gray) and dendritic (black) PDS, as measured by the first noticeable change in the voltage. Note that the dendritic PDS response was higher in amplitude and shorter in duration as compared with the somatic PDS response. In addition note that the dendritic PDS response preceded its somatic counterpart. B: the peak amplitude of PDS responses is presented as a function of the distance of the recording dendritic pipette from the soma. The values were obtained from different neurons. The somatic value is presented as mean ± SD values.

Comparing PDS responses along the apical dendritic tree revealed the peak amplitude of the PDS slightly increased along the first half of the main apical trunk. However, employing the linear regression analysis and the F statistical testing revealed this tendency failed to reach statistical significance (Fig. 2B).

Comparing the relative timing of apical dendritic and somatic PDS responses revealed that the dendritic PDS preceded the somatic PDS in all 18 double recordings (Fig. 2A). Hence, the PDS originated in the apical dendritic tree. The PDS response occurred first at the apical dendrite regardless of the location of the dendritic pipette (main apical trunk or main tuft branch, 30-550 µm from the soma). Moreover, this finding was independent of the location of the stimulating electrode, and occurred both when the stimulating electrode was placed in cortical layer 2-3 (n = 12), cortical layer 6 (n = 3) or the subcortical white matter (n = 3). It is interesting to note that in contrast to the PDS, fast action potentials overriding the PDS usually initiated at the axo-somatic region as they were observed first at the soma in 14 of 18 double recordings (for example, see Fig. 2A).

To investigate the role of voltage-gated sodium channels (VGSCs) in the formation of PDS responses, I used the intracellular VGSC blocker QX-314. In these experiments, the neurons were gradually loaded with 10 mM QX-314 via the recording pipette. Blockade of VGSCs did not eliminate the PDS responses (Fig. 3A). In fact, the amplitude of PDS responses increased, possibly due to the blockade of potassium channels by QX-314 (Andreasen and Hablitz 1993). Similar results were obtained in five additional experiments. The effectiveness of QX-314 in blocking VGSCs was monitored by the disappearance of action potentials.

View larger version (12K): [in this window] [in a new window]   Fig. 3. The role of voltage-gated calcium and sodium channels in the formation of PDS responses. A: PDS responses were recorded via a somatic whole cell pipette 1 and 40 min after institution of a whole cell recording in the presence 10 mM QX-314 in the recording pipette. Note that a PDS response was recorded in the presence of QX-314. B: PDS responses were recorded via a somatic whole cell pipette 1 and 40 min after institution of a whole cell recording in the presence 0.5 m D600 in the recording pipette. Note that the D600 markedly decreased the duration of the PDS response.

To investigate the role of voltage-gated calcium channels (VGCC) in the formation of the PDS response, I examined the effect of intracellular application of the selective L-type VGCC blocker D600 (methoxy-verapamil). In these experiments, the recording pipette contained 0.5 mM D600, and PDS responses were evoked at different time windows after establishing the whole cell recording. Loading the neuron with D600 caused the PDS response to decrease in amplitude and duration (Fig. 3B). Forty minutes after establishing the whole cell recording the peak amplitude and half-width of the PDS decreased by 9.3 ± 3.5 and 69.6 ± 9.7% as compared with the PDS response recorded 1 min after establishing the whole cell recording (n = 5). The blocker D600 was dissolved in DMSO (diluted to 0.1%). Experiments performed with similar concentrations of DMSO (0.1%) in the pipette showed no effect on the PDS (n = 3).

The addition of the NMDA receptor blocker APV (50 µM) markedly decreased the h 0 µM) to the extracellular solution eliminated all evoked and spontaneous epileptiform activity (n = 3).

PDS evoked [Ca²⁺]_i transients in the dendritic tree

The fine oblique, basal and distal tuft dendrites were inaccessible to direct intracellular voltage recordings. To investigate epileptic discharges in these fine dendritic branches, calcium-imaging measurements were performed. Figure 4 presents the [Ca²⁺]_i transients evoked by the PDS at different regions of the dendritic tree. The PDS responses evoked large calcium influx into all dendritic regions measured. Hence PDS responses spread throughout the entire dendritic tree including the fine basal, tuft and oblique dendrites. The spatial profile of the PDS evoked [Ca²⁺]_i transients showed significant nonuniformity along the dendritic tree. The peak [Ca²⁺]_i transients measured at the oblique and basal dendrites were significantly higher than those recorded at the apical trunk and reached average values of 71.0 ± 14.5 (n = 6), 29.8 ± 8.7 (n = 7), and 13.3 ± 3.3 µM (n = 9), respectively (Fig. 4, P < 0.003 for comparison between apical trunk and basal or oblique dendrites). These large differences may result in part from differences in the diameter and hence surface to volume ratio between the different dendritic branches (Schiller et al. 1995).

View larger version (28K): [in this window] [in a new window]   Fig. 4. The spatial profile of PDS evoked [Ca2+]i transients in the dendritic tree of layer-5 neocortical pyramidal neurons. A: fluorescence image of a neuron loaded with Mag-fura (0.5 mM) via the dendritic recording pipette under resting conditions. The PDS-evoked [Ca2+]i transients were recorded at different locations along the dendritic tree. Arrows mark the center of the measured dendritic regions of interest. The corresponding intra-dendritic voltage recording is presented at the bottom right panel. B: a histogram of the average peak amplitudes (mean ± SD) of PDS evoked [Ca2+]i transients at different regions along the dendritic tree (18 neurons). Left: comparison of the peak [Ca2+]i transients measured at different regions of the apical dendritic tree (P < 0.01 for comparison of tuft branches and the main apical trunk 200-300 µm from the soma). The right panel compares the peak [Ca2+]i transients measured at the main apical trunk (200-300 µm from the soma) the oblique and basal dendrites (P < 0.00001 and P < 0.003 for comparison of apical trunk and basal and oblique dendrites).

Comparison of the PDS evoked [Ca²⁺]_i transients measured along the apical dendritic tree revealed that the [Ca²⁺]_i transients developing in the distal tuft dendrites were significantly lower than those measured at the main apical trunk (Fig. 4, P < 0.01). This difference may result from reduced Ca²⁺ influx into the apical tuft as compared with the main apical trunk. However, it is possible that this finding results from increased calcium-buffering capacity in the apical tuft dendrites as compared with the apical trunk.

Ictal electrographic seizures

To investigate electrographic seizures in the dendritic tree, whole cell recordings were performed from apical dendrites of layer-5 pyramidal neurons (Fig. 5). Similar to somatic recorded seizures, electrographic seizures recorded at the apical dendrites were composed of recurrent attenuating bursts. On average there were 14.2 ± 2.4 bursts per seizure (n = 25) (see also Hablitz 1987). The initial 1-4 bursts consisted of PDS responses. As seizures progressed the amplitude of ictal bursts gradually attenuated, and they lost their PDS appearance. The peak amplitude of the sixth ictal burst, measured after omitting the over-riding fast action potentials, reached 26 ± 13% of that of the amplitude of the initial ictal PDS burst (P < 0.001, Fig. 5B). It should be noted that ictal bursts were over-riding a prolonged depolarization wave-form that followed the entire course of the seizure. The initial ictal PDS burst of electrographic seizures and isolated "inter-ictal" PDS response had a similar general appearance. However, the initial ictal PDS response was longer in duration (average half-width of 306 ± 45 and 357 ± 53 ms, P < 0.01, n = 11), with no significant difference between their peak amplitudes (P > 0.15).

View larger version (22K): [in this window] [in a new window]   Fig. 5. An electrographic seizure in the apical dendritic tree. A: a whole cell recording from the main apical trunk of a layer-5 pyramidal neuron (370 µm from the soma) during an electrographic seizure evoked by extracellular synaptic stimulation in the presence of 10 µM BCC. Top: the entire electrographic seizure. The bottom 4 panels demonstrate segments (marked by black lines) of the seizure presented in an expended time scale. The top left panel presents a schematic drawing of the neuron and the recording pipette. B: a graph plotting the peak amplitude (mean ± SD) of ictal bursts as a function of the their numerical order in the seizure. The peak amplitudes of ictal bursts were measured from the depolarization envelope of ictal bursts after truncation of action potentials. All recordings (n = 14) were performed from the apical dendrites 250-550 µm from the soma.

To further investigate electrographic seizures, I performed simultaneous whole cell recordings from the soma and apical dendrite of the same layer-5 pyramidal neuron (Fig. 6). In all 15 neurons examined, the initial 1-4 ictal bursts (average of 2.6 ± 0.4 and 2 bursts in the case of Fig. 6) were recorded first at the apical dendrite and only later observed at the soma. In contrast, the later ictal bursts of seizures were always observed first at the soma, and only later recorded at the apical dendrite (173 bursts from 15 neurons, Fig. 6). Thus during electrographic seizures the site of origin of ictal bursts shifted from the apical dendritic tree to the basal-soma region. These findings were observed in all experiments regardless of the location of the dendritic recording pipette along the apical dendritic tree (30-500 µm from the soma). Moreover, they were observed when the synaptic stimulating electrode was placed in neocortical layer 2-3 (n = 9), layer 6 (n = 3) or the subcortical white matter (n = 3).

View larger version (37K): [in this window] [in a new window]   Fig. 6. Comparison of an electrographic seizure in the soma and apical dendrite of the same layer-5 neuron. Concomitant whole-cell recordings were performed from the soma and an apical dendrite (285 µm from the soma) of the same layer-5 pyramidal neuron. The left panel presents a schematic drawing of the neuron and the recording pipettes. The right panel presents the simultaneous dendritic (top) and somatic (bottom) recordings. The lower four panels demonstrate segments (marked by black lines) of the seizure presented in an expended time scale. The arrows mark the initiation of the somatic (gray) and dendritic (black) ictal burst, as measured by the first noticeable change in the voltage. Note the shift in the dendritic site of origin of the bursts during the seizure.

Calcium-imaging experiments revealed that similar to isolated inter-ictal PDS responses the initial ictal PDS bursts evoked calcium influx into all dendritic regions measured, including the apical trunk, apical tuft, oblique and basal dendrites (n = 5). Calcium-fluorescence measurements during seizures showed that similar to the voltage recordings, [Ca²⁺]_i transients evoked by ictal bursts in apical and basal dendrites gradually attenuated during the course of the seizure (Fig. 7, A and B). In fact in seven of nine neurons examined, [Ca²⁺]_i transients evoked by the late ictal bursts were immeasurable with the fluorescence dye Mag-fura (for example see Fig. 7A). Comparable experiments performed with the high-affinity calcium indicator dyes Calcium Green-1 and Fura-2 showed that late ictal bursts evoked calcium influx into apical (n = 6) and basal (n = 3) dendrites. However, in most cases, this influx was too small to be detected by the low-affinity indicator dye Mag-fura. It should be noted that the peak [Ca²⁺]_i values presented in Fig. 7B were only measured when clear [Ca²⁺]_i transients were observed. Hence, the values presented for the last two bursts were probably overestimated (for details, see figure legend of Fig. 7B).

View larger version (29K): [in this window] [in a new window]   Fig. 7. Intracellular voltage and calcium measurements during electrographic seizures. A: concomitant intra-dendritic voltage (bottom trace) and [Ca2+]i fluorescence measurement (top trace) during an electrographic seizure. Intra-dendritic [Ca2+]i was recorded using the calcium-sensitive dye Mag-fura (0.5 mM) from a region of interest located in the apical trunk 345-365 µm from the soma. The corresponding intra-dendritic voltage recording was performed from the main apical trunk 340 µm from the soma. The seizure was elicited by extracellular synaptic stimulation in cortical layer 2-3 in the presence of 10 µM BCC. B: a graph plotting the peak of the intra-dendritic [Ca2+]i transient (mean ± SD) evoked by ictal bursts as a function of the their numerical order in the seizure in the main apical trunk 200-300 µm from the soma (left, n = 9) and basal dendrites (right, n = 4). The peak [Ca2+]i values were only measured when clear [Ca2+]i transients were observed (for apical dendrites all 9 neurons for the 1st 2 ictal bursts, 7 neurons for the 3rd burst and 5 neurons for the 4th burst; for basal dendrites all 4 neurons for the 1st 2 ictal bursts and 3 neurons for the 3rd and 4th bursts). Thus for the last 2 bursts, I probably overestimated the [Ca2+]i values.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study, I characterized "inter-ictal-" and "ictal"-like epileptic discharges in dendrites of layer-5 pyramidal neurons.

Inter-ictal-like PDS responses

The intracellular correlate of inter-ictal spikes is the PDS response (for review, see De Curtis and Avanzini 2001; Prince and Connors 1986; Schwartzkroin and Wyler 1980; Speckmann and Erwin 1999). The PDS responses developed in both the soma and apical dendrites. However, the two responses differed from one another. The PDS responses in apical dendrites had a higher amplitude and shorter duration than the corresponding somatic PDS. Regarding the origin of PDS responses, double somatic and dendritic recordings revealed that PDS responses originated from the apical dendrites and later spread anterogradly to the soma. Dendritic PDS responses preceded somatic PDS responses in all double recordings regardless of the dendritic recording site along the main apical trunk and primary tuft branches. Thus PDS responses most probably originated from the distal apical dendrites rather than the proximal apical trunk or oblique dendrites.

The majority of excitatory synapses innervate the fine basal, oblique, and tuft dendritic branches (Larkman 1991). As these dendritic branches are inaccessible to direct voltage recordings, no information exists regarding epileptic discharges in these fine dendrites. To investigate this question, I used calcium-imaging measurements. The results of these experiments indicated that PDS responses were associated with large calcium influx into all dendritic regions and, hence, spread throughout the dendritic tree including the fine basal, oblique, and distal tuft dendritic branches. From our results, it is unclear whether PDS responses originate exclusively in the apical dendritic tree and later back-propagated into basal dendrites or whether they initiate independently in the basal and apical dendritic trees.

The large calcium influx during PDS responses may convey a biochemical signal as well. The peak dendritic [Ca²⁺]_i transients, which reached values of 9.3-71.3 µM in different dendritic regions, significantly exceeded the peak [Ca²⁺]_i transient evoked under physiological conditions (Schiller et al. 1995, 1997). The large and widespread Ca²⁺ influx evoked by the PDS responses may lead to activation of various calcium-dependent processes such as opening of various calcium-dependent ionic channels, activation of different calcium dependent enzymes, or induction of long-term changes in synaptic efficiency or axonal sprouting. These processes may in turn cause secondary neuronal damage and loss and further increase the excitability of the epileptic zone.

What are the ionic mechanisms underlying the PDS response? The amplitude of PDS responses was higher in the apical dendrite than in the soma, indicating PDS responses originated in the apical dendritic tree. Moreover, the peak voltage value of PDS responses in apical dendrites usually surpassed 0 mV, which was the reversal potential for both AMPA and NMDA glutamate-receptor channels (for review, see Meyer and Westbrook 1987). Hence, from reversal potential considerations either sodium or calcium ions served as important charge carriers in the formation of the PDS response. As PDS responses persisted after blockade of VGSC, it is likely that calcium channels rather than sodium channels participated in the formation of PDS responses. This conclusion is further strengthened by the finding that intracellular application of the L-type VGCC blocker D600 decreased the amplitude and duration of the PDS response. It is important to note that the dendritic membrane contains VGCCs (Amitai et al. 1993; Kim and Connors 1993; Magee and Johnston 1995; Schiller et al. 1997; Schwindt and Crill 1997). Moreover, previous studies that used mostly somatic recordings combined with computer modeling and very few direct dendritic recordings also concluded that dendritic VGCC participated in the formation of PDS responses (De Curtis and Avanzini 2001; Prince and Connors 1986; Schwartzkroin and Wyler 1980; Straub et al. 1994, 2000; Traub et al. 1993).

What VGCC subtypes participated in the formation of the PDS response? L-type VGCC probably contributed to the PDS responses as intra and extracellular blockade of L-type VGCC reduced the amplitude and duration of PDS responses (for intracellular application see this study, for extracellular application see Straub et al. 1994, 2000). However, the apical and basal dendrites of layer-5 neocortical pyramidal neurons contain multiple subtypes of VGCC (Markram et al. 1995; Schiller et al. 1998), and thus other VGCC subtypes such as the N-, P-, Q-, R-, and T-type VGCCs probably participated in the formation of the PDS responses as well. This possibility has not been tested directly as extracellular blockers of non L-type VGCCs have been shown to reduce transmitter release (for review see Dunlop 1995; Wu and Saggau 1997), and none of them has been shown to be effective intracellularly.

Current flowing directly through glutamate-receptor channels probably also contribute to the PDS as well (Johnston and Brown 1984), possibly via the newly described dendritic NMDA spikes (Schiller et al. 2000). However, active dendritic conductances contribute 10-20 mV to the peak amplitude of the PDS, as this constitutes the difference between the peak voltage value of PDS responses and the reversal potential of ionotropic glutamate-receptor channels.

Electrographic seizures

In both the soma and apical dendrites, electrographic seizures were composed of recurrent bursts. The early bursts were composed of PDS responses. The initial PDS of seizures was similar to the isolated inter-ictal PDS response with respect to the site of origin spread along the dendritic tree and amplitude. The only difference between these two responses being that the initial ictal PDS burst of seizures was 17% wider. As the seizure progressed the voltage amplitude of ictal bursts gradually attenuated. Concomitant calcium imaging measurements showed that the calcium influx associated with ictal bursts attenuated during the seizure as well. These findings suggested that the attenuation of burst amplitude during the seizure was related in part to decreased activation of dendritic VGCC. However, other possibilities exist to explain the attenuation of [Ca²⁺]_i transient during seizures as well such as reduced activation of NMDA-receptor channels or decreased mobilization of calcium from internal stores. Previous studies have shown that calcium currents and dendritic calcium spikes undergo developmental maturation (Lorenzon and Foehring 1995a,b; Zhu 2000). In rats, the current amplitude and estimated density of VGCC increased during the first month of development (Lorenzon and Foehring 1995a,b). In addition dendritic calcium spikes in distal apical dendrites gradually developed during the first 2 mo of life. In this study the data regarding electrographic seizures was obtained from immature 12- to 17-day-old rats. Thus it is possible that in adult animals the individual ictal bursts lasted longer and showed less attenuation during the course of the seizure.

Throughout the course of electrographic seizures, the recurrent ictal bursts were over-riding a prolonged depolarization waveform. These ictal bursts resembled the bursts underlying the clonic convulsions observed in intact animals with experimental epilepsy (Ayala et al. 1970; Dichter et al. 1998; Matsumoto and Ajmone-Marsan 1964; Speckmann and Erwin 1999). It is likely that the prolonged depolarizing envelope and the over-riding ictal bursts observed during electrographic seizures constituted different phenomena with different underlying ionic mechanisms. The individual ictal bursts probably resulted from a mixture of excitatory synaptic currents and intrinsic dendritic conductances such as VGCCs evoked by synchronized suprathreshold activity sweeping throughout the neuronal population of the slice. The ionic mechanism underlying the prolonged depolarization waveform was not addressed in this study.

What is the site of origin of bursts during electrographic seizures? Simultaneous voltage recordings from the soma and apical dendrite of the same neuron revealed the initial 1-4 ictal bursts were recorded first in the apical dendrite and only later at the soma. In contrast, the later bursts in the seizure were observed first at the soma and only later at the apical dendrite. These findings indicated the site of origin of ictal bursts shifted during the seizure. Similar to inter-ictal PDS responses the initial ictal PDS bursts originated in the apical dendrites, while the later bursts of the seizure originated in the soma-basal region.

The most likely mechanisms to explain the shift in the origin of ictal bursts during the seizure is a change in the activation pattern or relative activation timing of excitatory inputs innervating apical and basal dendrites. Namely, whereas at the early stages of the seizure, the inputs innervating the apical dendrites are activated, at the later phase of the seizure, the inputs innervating the basal dendrites are activated. Other mechanisms may also contribute to the shift in the origin of ictal bursts such as gradual attenuation of calcium-mediated regenerative responses in apical dendrites during the seizure or increased excitability of the basal dendrites or soma.