Enhanced burst discharges in the CA1 area of the immature versus adult hippocampus: patterns and cellular mechanisms

Abstract

Burst discharges in the immature brain may contribute to its enhanced seizure susceptibility. The cellular mechanisms underlying burst discharges in the CA1 area of the immature versus adult hippocampus were investigated with simultaneous whole-cell and field-potential recordings. When GABA_A receptors were blocked pharmacologically, bursts in CA1 were either graded or all-or-none (or mixed) as a function of electrical stimulation intensity. Most CA1 minislices from immature rats displayed all-or-none or mixed bursts, whereas the slices from adult rats predominantly elicited graded bursts. The frequency and amplitude of spontaneous excitatory postsynaptic currents (sEPSCs) were greater in CA1 pyramidal cells from the immature than the adult slices. The developmental differences in CA1 bursting were also detected in slices adjusted for maturational changes in brain volume (i.e., 350 µm thick for immature vs. 450 µm thick for adult rats). Neither N-methyl-d-aspartate (NMDA) nor group I metabotropic glutamate (mGlu1) receptor antagonists blocked the network-driven bursts in immature CA1, but an α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor blocker abolished them. Robust excitatory postsynaptic potentials (EPSPs) occurred after bursts in some immature CA1 slices (23%) but never in slices from the adult. The input-output (amount of current injected vs. number of action potentials generated) relationship was markedly greater in CA1 pyramidal cells in the immature compared with the adult hippocampus. These data suggest that the CA1 area of the immature brain is capable of generating network-driven bursts, which declines in adult rats. The increased propensity of burst generation in immature CA1 appears to involve a greater AMPA receptor-mediated synaptic network and an increased intrinsic spike-generating ability.

NEW & NOTEWORTHY Burst discharges in the developing brain can provide valuable insights into epileptogenesis. We show that the immature hippocampal CA1 area is capable of generating all-or-none (i.e., network) bursts, which transitions to graded (i.e., nonnetwork) bursts in the mature brain via both synaptic and intrinsic mechanisms. Our results provide new clues to help understand possible mechanisms that may be shared in the immature and epileptic brain and how the normal brain becomes seizure prone (i.e., epileptogenesis).

INTRODUCTION

Several studies have suggested that the immature brain has an increased susceptibility for seizures in both humans (1) and animals (2–4). In vitro experiments in brain slices have shown that the immature brain has a greater propensity for generating prolonged burst discharges, specifically in neocortex (5, 6) and in the CA3 area of the hippocampus (7–9). Although most epilepsies in the pediatric population are thought to have a genetic basis with generalized seizures, temporal lobe epilepsy typically involves the hippocampus. The CA1 region of the hippocampus from rats with kainate (KA)-induced epilepsy has been shown to have increased propensity to generate burst discharges (10–15), but the underlying mechanisms are not well understood. We hypothesize that the generation of burst discharges in the immature and the epileptic brain may share some common cellular mechanisms or substrates. Therefore, further study of neuronal bursting behavior in the immature CA1 may provide useful information for understanding the mechanisms for generation of the seizures and interictal discharges associated with temporal lobe epilepsy.

Several factors may contribute to an increased seizure propensity in the immature brain, such as delayed onset of GABAergic inhibition after a period with an excitatory effect of GABA (16–19), unique features of N-methyl-d-aspartate (NMDA) receptors (20–22), and more abundant axonal branches and synaptic connectivity (23). In addition, it has been reported that the group I metabotropic glutamate receptors (mGluR1s) may be important for the generation of prolonged burst discharges (24, 25). In a previous study (7), we showed that the generation of prolonged burst discharges in the CA3 area of the developing hippocampus involves a more extensive α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/KA receptor-mediated synaptic network and an increased intrinsic firing but does not require the activation of NMDA, mGlu1 receptors, or excitatory action of GABA.

In this study, we further explored network bursting behavior in the CA1 area of the immature versus adult hippocampus. Our results show that the immature CA1 area has an increased propensity for network bursts, which declines in the adult CA1. The increased bursting in the immature CA1 area also appears to involve a more pronounced AMPA/KA receptor-mediated synaptic network and increased intrinsic firing, and the difference in burst generation does not depend on activation of GABA, NMDA, or mGlu1 receptors.

METHODS

Brain Slice Preparation

Procedures used in this study were approved by the Animal Care and Use Committees of Colorado State University and the University of Utah. Experiments were performed in brain slices prepared from Sprague-Dawley rats. In the first set of experiments (i.e., bicuculline-induced burst discharges; see Figs. 1 and 2), the immature group contained rats 14–19 days of age (n = 10), and the adult group contained rats 90–123 days of age (n = 10). In addition, four rats 22–25 days of age and another four rats 36–73 days of age were also used in additional experiments to determine whether the developmental changes in network bursting occur gradually or abruptly. In this set of experiments, slices were cut at the same thickness for both immature and mature groups (400 µm). The second set of experiments (i.e., Figs. 3–5) were conducted in rats of 18–20 days (for immature group, n = 10) and 85–99 days (for adult group, n = 9). In these experiments, hippocampal slices were cut at different thicknesses (350 and 450 µm for immature and mature, respectively, to control for developmental differences in neuronal density; see next paragraph). Conventional procedures were used for slice preparation. Briefly, rats were decapitated after being anesthetized with halothane, and their brains were dissected out and placed in ice-cold oxygenated artificial cerebrospinal fluid (aCSF) containing the following (in mM): 124 NaCl, 3 KCl, 26 NaHCO₃, 1.4 NaH₂PO₄, 2 CaCl₂, 2 MgSO₄, and 11 glucose. Coronal hippocampal slices (i.e., mostly from dorsal hippocampus) were cut with a vibroslicer (Lancer series 1000; Vibratome, St. Louis, MO). These hippocampal slices were further cut between the CA1 and CA3 and between CA1 and subiculum to isolate the CA1 area from other regions of the hippocampus (i.e., CA1 minislices). The reason for using CA1 minislices was to avoid any possible effects from the upstream area of CA3, a region known to generate bursts. Slices were stored in a submerged storage chamber preheated to 32–34°C. Recordings usually started 2 h after slice preparation, when slices were transferred to a ramp-type interface recording chamber perfused with oxygenated aCSF at 32–34°C.

Figure 1.

Patterns of burst discharges in CA1 minislices from immature and adult rats: experimental observation and diagrammatic illustration. Responses of CA1 minislices to Schaffer Collateral stimulation were examined when GABAergic transmission was blocked. The patterns of responses are shown on right, and their respective hypothetical concepts are shown in diagrams on left. Three patterns of burst discharges (A1–A3) were observed. The first type of response (A1, right) was classical responses to pathway stimulation, i.e., the responses increased with the incremental increase in stimulus intensity. The hypothetic mechanism of this pattern of responses is illustrated in the diagram (A1, left). When the stimulus intensity increases (represented by small, medium, and large arrows in a, b, and c), more axons in the pathway (represented by red lines) are activated and more pyramidal cells (represented by triangles) are excited, thus generating graded responses. The second type of response (A2, right) showed a sudden onset of a full-blown burst (i.e., the same low-intensity stimulation evoked either no response in a or a full burst in b), and, once induced, the burst did not significantly increase with an increase in stimulus intensity (c). This type of response is a characteristic of a network burst and is illustrated in the diagram (A2, left). In this case, a local neuronal network is formed by the connection of neurons (triangles) with excitatory synapses (curves). Threshold pathway stimulation of the same intensity (a and b, small arrows) may activate a few neurons that may either spread through the whole network to generate a full burst or fail to spread through and result in no response. An increased stimulation (c, large arrow) activates more neurons and more easily spreads to the entire network to generate a full burst but no longer increases the burst. The third type of response (A3, right) also had a sudden onset (a and b); however, the amplitude and/or the number of population spikes/action potentials in the burst increased significantly with the increase of stimulus intensity (c). Thus, these bursts are considered “seminetwork” bursts or “mixed” network and graded bursts. One explanation is that some of the neurons are connected to form a smaller size of neuronal network (A3, left). Threshold stimulation of the same intensity (a and b, small arrows) may activate (or fail to activate) the reduced network and generate a short burst, but a further increase in stimulus intensity (c, large arrow) recruits more axons and activates the rest of the neurons outside the network, thus increasing the responses. B: whereas all 3 types of responses occurred in slices from immature rats, most of the slices showed network and seminetwork bursts (71% total, i.e., 43% “all-or-none” and 28% mixed bursts, left). In contrast, the responses in slices from adult rats were predominantly graded bursts (92%, right), suggesting a fundamental developmental difference in bursting behavior in the CA1 area. EC, extracellular; IC, intracellular.

Figure 2.

Different stimulation-response relationships of the 3 types of bursting. A: pooled data from 9 slices showing all-or-none responses. The responses were quantified by the area (mV·ms) of each burst and normalized to the maximal response (%). Once the stimulus intensity reached threshold (T), the slices responded with bursts that were near the maximum. Further increase in intensity up to 20 T did not significantly increase the responses, and there was no regression between stimulus intensity and responses (R = 0.24, P = 0.15). B: pooled data from 20 slices showing graded responses. In contrast to all-or-none bursts, these responses continued to increase as stimulation intensity was increased (up to 25 T) and displayed a clear regression (R = 0.74, P < 0.001). C: pooled data from 10 slices showing mixed responses. Threshold stimulation (1 T) evoked sudden-onset initial bursts ∼25–60% of the maximal responses. These bursts then significantly increased with the increase of stimulus intensity (R = 0.67, P < 0.001). Note that x-axis (stimulus intensity) is on logarithm scale for all 3 panels.

Figure 3.

Spontaneous excitatory postsynaptic currents (sEPSCs) occurring in CA1 pyramidal cells were more frequent and larger in slices from immature than adult rats. A: most of the CA1 pyramidal cells from immature rats (left) and some from adult rats (right) were observed to have sEPSCs. Each panel shows a 20-s continuous recording divided into 8 traces. B: histograms of logarithmically transformed interevent interval for the immature and adult groups. Both histograms fit a Gaussian distribution (black and red curves), and the distribution of sEPSC intervals of the immature group (black) is shifted to the left compared with that in the adult group (red), suggesting a shorter interval (thus a higher frequency) of sEPSCs in the immature group. C: cumulative probability of the interevent intervals revealed a clear difference between the immature and adult groups [P < 0.01, Kolmogorov–Smirnov (KS) test]. D: histograms of sEPSC amplitude for the immature (black) and mature (red) groups. Some of the large-amplitude events (1.7–2.3 in log value, or 65–200 pA) in the immature group were not present in the adult group. E: cumulative distributions confirmed a significant difference in amplitude between the 2 groups (P < 0.01, KS test). Note that the values of interevent interval and amplitude were log transformed for a direct comparison.

Figure 5.

“Reverberating” excitatory postsynaptic potentials (EPSPs) after bursts in the CA1 area of the immature rats. A: example showing that in CA1 minislices from the immature rats, threshold stimulation evoked either full-blown bursts (open arrow, action potentials are truncated for clarity) or no bursts (open triangle, bottom 2 traces). Multiple reverberating EPSPs occurred and persisted for hundreds of milliseconds (asterisk) after epileptiform bursts but not after subthreshold responses (i.e., no action potential). B: the postburst reverberating EPSPs were never observed in CA1 minislices from adult rats. Therefore, these data provide additional evidence for a synaptically connected excitatory network in the CA1 area of immature rats.

Because the brain volume, but not the number of neurons, increases during development, the density of neurons is lower in the adult brain. Therefore, it is possible that a slice of a particular thickness from an immature rat contains more neurons and synaptic circuits than a slice of the same thickness from a mature rat, which in theory could alone result in a greater bursting activity. To compensate for the lower number of neurons and reduced circuit density in adult brain and thus to minimize the potential for a false-positive effect, the slices from the immature group were cut thinner (350 µm) than those from the adult group (450 µm). The rationale for and calculation of the two thicknesses was explained in detail in a previous paper (7). Briefly, the density of pyramidal cells in the hippocampus at the second postnatal week is estimated to be roughly 1.5 times greater than that of adults (23). Therefore, to make the number of neurons comparable between the immature and adult brain, the thickness of the immature slices ought to be reduced to two-thirds of that of the adult slices. Because the superficial 50 μm of tissue on both sides of the slices is most likely damaged during sectioning, a 450-μm-thick adult slice would actually have healthy tissue of 350 μm. Thus, the immature slices ought to be two-thirds of the viable tissue in adult slices plus 50 μm of damaged tissue on both sides (i.e., 350 μm × 2/3 + 50 μm × 2 = 333 μm). For convenience, all slices for immature rats were cut at 350 μm.

Electrophysiology

Concurrent field-potential and whole-cell current-clamp recordings were conducted in the stratum pyramidale layer of the CA1 minislices, with a combination of an Axoprobe-1A amplifier and an Axopatch-1D amplifier (Axon Instruments/Molecular Devices, Foster City, CA), and were DC coupled. Thick-wall glass pipettes [outer diameter (OD) 1.65 mm, inner diameter (ID) 1.2 mm; Garner Glass, Claremont, CA) were pulled with a P-87 Flaming-Brown puller (Sutter Instruments, Novato, CA) and used for both extracellular and whole-cell recordings. Electrodes for field-potential recordings were filled with aCSF, and electrodes for whole-cell recordings were filled with intracellular solution containing the following (in mM): 120 K-gluconate, 1 NaCl, 5 EGTA, 10 HEPES, 1 MgCl₂, 1 CaCl₂, 2 ATP, and 5 biocytin. The pH was adjusted to 7.2 with 5 M KOH. A calculated liquid junction potential between the intracellular and the bath solution was +12 mV at 32°C (calculated with pClamp 8; Axon Instruments/Molecular Devices, Union City, CA) and was not corrected. Synaptic stimulation of the CA1 pyramidal cells was delivered through a bipolar stimulating electrode that was custom made from two Teflon-coated platinum wires (wire diameter = 25 µm, separation between two wires = 50 μm) and placed in the middle of stratum radiatum of CA1 area ∼150–200 μm lateral to the recording site. Rectangular current pulses with fixed duration (200 µs) and variable intensities (4–400 µA) were used to evoke synaptic responses. To examine intrinsic membrane properties of the CA1 pyramidal cells, synaptic transmission was eliminated with a mixture of GABA_A, GABA_B, AMPA/KA, and NMDA receptor antagonists. These experiments were performed in a submerged chamber at 34–35°C, with an infrared differential interference contrast (IR-DIC) visualization system (Zeiss microscopy, Zeiss Axioskop, Germany; CCD camera, Hamamatsu, Japan) and a MultiClamp700A amplifier (Axon Instruments/Molecular Devices, Union City, CA). To determine the input-output relationship, a series of hyperpolarizing-depolarizing current pulses (15-s long) were injected into the CA1 pyramidal neurons through recording pipettes. All signals were acquired at 10 kHz and low-pass filtered at 2 kHz through a Digidata-1320A digitizer (Axon Instruments/Molecular Devices, Union City, CA).

Data Analysis

Qualitative and quantitative data analyses were conducted with pClamp 8.0 (Clampfit; Molecular Devices, Union City, CA), MiniAnalysis 5.0 (Synaptosoft, Inc., Leonia, NJ), SigmaPlot (SPSS Inc., Chicago, IL), and Origin 8 (OriginLab Corporation, Northampton, MA). The spontaneous excitatory postsynaptic currents (sEPSCs) were detected with MiniAnalysis and visually checked to minimize errors. An epoch of 3–5 min recording per cell was used for measuring sEPSC frequency and amplitude. Neurons that only had occasional sEPSCs (i.e., ≤0.2 Hz) were considered to have no sEPSCs for both the immature and adult groups. A fixed number (versus a fixed time period of recording) of sEPSCs from each neuron (i.e., the first 50 sEPSCs) was pooled for the distribution (histograms) for sEPSC amplitude and interval, which minimized potential sampling bias due to the difference in sEPSC frequency among neurons [i.e., if sampled on a fixed time basis (e.g., 3 min), a histogram would contain more “events” from the neurons with higher frequency of sEPSCs and thus be skewed toward the higher-frequency neurons]. Also, because of the highly variable nature of the events (e.g., the sEPSC interval varied from a few milliseconds to a few thousands of milliseconds), values of all sEPSC intervals and amplitudes were logarithmically transformed in the histograms. The Kolmogorov–Smirnov (KS) test was used to compare the cumulative probability of sEPSC intervals and amplitudes between the immature and adult groups. The Student’s t test and Fisher’s exact test were used for comparisons of quantitative data and ratios, respectively, between the immature and adult groups. Linear regression was employed to determine the relationship between stimulus intensity and responses and to assess the trend of burst pattern changes during development. For all the tests, statistical significance is defined as P < 0.05.

Pharmacological Agents

Antagonists for GABA_A receptors (bicuculline methiodide, 30 µM or SR-95531, 10 µM), GABA_B receptors (SCH50911, 10 µM), NMDA receptors [dl-2-amino-5-phosphonopentanoic acid (AP-5), 50 µM], AMPA/kainate receptors [6,7-dinitroquinoxaline-2,3-dione (DNQX), 50 µM], and mGlu1 receptors [1-aminoindan-1,5-dicarboxylic acid (AIDA), 500 µM] were used to pharmacologically block the respective components of the synaptic events. SCH50911 was from Tocris (Ellisville, MO), and the other blockers were from Sigma (St. Louis, MO).

RESULTS

Distinct Patterns of Bicuculline-Induced Burst Discharges in CA1 Minislices from Immature vs. Adult Rats

The initial experiments were conducted in CA1 minislices (400 µm thick) bathed with the GABA_A receptor antagonist bicuculline (30 µM) to examine the propensity for and properties of burst discharges during development. Three types of bursting behavior (Fig. 1) were observed in response to Schaffer Collateral stimulation. The first type of responses was regarded as “graded bursts.” Namely, stimulation at the threshold level evoked one or more minimal field-potential population spikes and/or action potentials (Fig. 1A1a, right). Increases in stimulus intensity increased both the number of action potentials and the number and amplitude of the population spikes (but not the duration) of the burst (Fig. 1A1, b and c, right). When quantified by measuring the area of each field-potential burst (mV·ms) and normalized to the maximal response (%), these bursts displayed a clear linear regression with the stimulus intensity up to 20 times of threshold (T) level [Fig. 2B; R = 0.74, F(1,86) = 101, P < 0.001]. This pattern of graded responses is consistent with the classical activation of neurons by stimulation of a synaptic pathway (i.e., the larger the stimulus intensity, the more axons are excited and the more postsynaptic neurons are activated) (Fig. 1A1, left).

The second type of response was regarded as an “all-or-none” burst. That is, minimal stimulation at the threshold level evoked either no burst (Fig. 1A2a, right) or a sudden-onset, full-blown burst with a long and variable latency (i.e., typical characteristics of a network response; Ref. 26) (Fig. 1A2b, right). These bursts contained multiple population spikes/action potentials with a negative field-potential shift/depolarization and lasted from a few hundred milliseconds to up to 6.8 s (median: 770 ms, n = 6 slices). Once evoked, additional increases in stimulus intensity no longer significantly increased the number of action potential/population spikes or burst duration (Fig. 1A2c, right). Further quantitative analysis showed that these responses (as assessed by burst area relative to the maximum) did not constitute a linear regression relationship with stimulus intensity (Fig. 2A; R = 0.24, P = 0.15). This pattern of discontinuous, largely all-or-none burst responses was consistent with the activation of an interconnected neuronal network, where threshold stimulation may elicit electrical activity that spreads (or fails) to the entire network and generates a network response (burst). For all-or-none burst discharges, further increases in stimulus intensity no longer increase the output (Fig. 1A2, left).

A third type of response appeared to be a “mixed” graded and all-or-none response (Fig. 1A3, right). Specifically, the initial responses to low-intensity stimulation were either no burst (Fig. 1A3a) or a sudden short burst with a long and variable latency (Fig. 1A3b). However, further increases in stimulus intensity significantly increased both the amplitude and/or number of population spikes/action potentials as well as burst duration (Fig. 1A3c). These bursts lasted from several tens of milliseconds to a few hundred milliseconds. Quantitative analysis showed that the initial all-or-none bursts at threshold level (1 T) were ∼25–60% of the maximal response, which then increased at a near-linear rate as stimulus intensity increased (Fig. 2C; R = 0.67, P < 0.001). The sudden onset and long latencies of these bursts resemble the all-or-none bursts, whereas the gradual increase in population spikes resembles graded bursts. Based on our understanding of graded and all-or-none bursts, we hypothesize that this was a transitional pattern of responses, where some neurons were interconnected and activated by threshold stimulation and generated an initial short burst and the rest of the neurons were successively activated with the increase in stimulus intensity and produced more robust bursts (Fig. 1A3, left).

Interestingly, the CA1 minislices from immature and adult rats showed distinctly different patterns of responses. Nearly half of the minislices from 14- to 19-day-old immature rats (43%, n = 14 from 10 rats) responded to Schaffer Collateral stimulation with all-or-none bursts, 28% of the slices with mixed bursts, and 29% of the slices with graded bursts (Fig. 1B, left). In contrast, 92% of the slices from 3- to 4-mo-old adult rats (n = 13 from 10 rats) displayed graded responses, 8% of the slices showed mixed responses, and none of the slices showed all-or-none bursts (Fig. 1B, right). The ratio of slices exhibiting all-or-none or graded bursts is clearly different between immature and adult brain (P = 0.003, Fisher exact test). Additional experiments were made in rats at postnatal days (P)22–25 (7 slices from 4 rats) and P36–73 (5 slices from 4 rats). In slices from the 22–25 day rats, 43% of them displayed all-or-none bursts and the same portion (43%) of the slices displayed graded bursts; in slices from 36–73 day rats, none showed all-or-none burst, 40% showed mixed bursts, and 60% showed graded bursts. These observations show that the ratio of network bursts (all or none and mixed) gradually declines and the ratio of graded bursts increases during development [linear regression, R = 0.98, F(1,2) = 84.4, P = 0.012], suggesting a maturational transition in the patterns of the bursts. Altogether, these data suggest that the CA1 minislices from immature rats generate mostly network or seminetwork (mixed) bursts, whereas slices from adult rats produce predominantly graded bursts. This developmental switch in bursting behavior appears to occur gradually.

CA1 Pyramidal Cells Had More Frequent and More Large-Amplitude sEPSCs in Minislices from Immature than Adult Rats

In bicuculline, sEPSCs were observed in 77% of recorded pyramidal cells (n = 13 in 11 slices) from immature rats (Fig. 3A, left) and 40% of cells (n = 20 in 14 slices) from adult rats (Fig. 3A, right). Although the immature group tended to have more cells exhibiting sEPSCs than the adult group, this difference was not statistically significant (P = 0.072, Fisher’s exact test). The sEPSCs in CA1 pyramidal cells from immature rats occurred at nearly 2 Hz (1.9 ± 0.6 Hz, n = 10; Fig. 3A, left), whereas those in CA1 pyramidal cells from adult rats were typically <1 Hz (0.8 ± 0.3 Hz, n = 8; Fig. 3A, right). Histograms of interevent intervals showed that the peak of the histogram in the adult group was shifted to the larger side compared with the immature group (i.e., lower frequency; Fig. 3B), and their cumulative distribution was clearly different (P < 0.01, KS test; Fig. 3C). Moreover, CA1 pyramidal neurons from immature rats appeared to have more large-amplitude sEPSCs (i.e., those of 65–200 pA or 1.8–2.3 in log values; Fig. 3D) than those from adult rats. Consistently, the cumulative distribution further confirmed a distinct difference in the amplitude of sEPSCs between the immature and adult groups (P < 0.01, KS test; Fig. 3E). These data show that immature CA1 pyramidal cells exhibit frequent and large-amplitude sEPSCs, which are decreased in adult neurons. These data suggest an increased degree of local excitatory synaptic connectivity between and increased firing rate in immature CA1 pyramidal cells, which might mediate the increased propensity for network bursts in the immature CA1 slices.

The Developmental Difference in Network Bursting in CA1 Persisted in Slices Adjusted for Developmental Brain Volume Changes

Bursting behavior in CA1 was further examined in slices adjusted for developmental brain volume change. In the thinner (350 µm) CA1 minislices from immature rats, Schaffer Collateral stimulation evoked all-or-none, graded, and mixed bursts in 50%, 22%, and 28% of the slices, respectively, similar to what was observed in 400-µm-thick slices in previous experiments (i.e., 43%, 29%, and 28%, respectively; Fig. 1B, left). In contrast, in the thicker (i.e., 450 µm) slices from adult rats, Schaffer Collateral stimulation elicited graded, mixed, and all-or-none bursts in 82%, 12%, and 6% of the slices, respectively, which was also similar to what occurred in the 400-µm-thick slices (Fig. 1B, right). Thus, the developmental difference in epileptiform bursting between immature and adult CA1 was still present in slices adjusted in thickness for an increase in brain volume during development.

Developmental Difference in Bursting Was Independent of GABA_A, GABA_B, NMDA, and mGlu1 Receptor-Mediated Synaptic Transmission

To determine whether GABA_A, GABA_B, and NMDA receptor-mediated transmission was necessary for burst generation, experiments were conducted in the presence of GABA_A, GABA_B, and NMDA receptor antagonists gabazine (10 µM), SCH50911 (10 µM), and AP-5 (50 µM). Under these conditions, all-or-none network bursts were consistently elicited in half of the CA1 minislices from the immature rats (Fig. 4A). In this example, 10 consecutive stimulations at the same threshold intensity evoked either an excitatory postsynaptic potential (EPSP) only (i.e., the bottom 3 traces in Fig. 4A, top and bottom) or a burst with a long-and-variable latency (i.e., the other 7 traces in Fig. 4A). Interestingly, the same threshold stimulations often produced network bursts of short (i.e., tens to hundreds of milliseconds) and prolonged (i.e., hundreds of milliseconds to seconds) duration (Fig. 4A). Further increase of stimulus intensity tended to generate short bursts (data not shown). In slices from adult rats under the same experimental conditions but of greater thickness (450 µm), the responses to Schaffer Collateral stimulation were mostly graded with stimulus intensity (Fig. 4B). Furthermore, the role of mGlu1 receptors in the generation of network bursts in the immature CA1 was also examined by adding a specific blocker of mGlu1 receptor, AIDA (500 µM), to the medium in some experiments. Blockade of mGlu1 receptors did not block the network bursts in the immature CA1 (n = 4 slices, data not shown). In contrast, these network bursts were completely blocked by the AMPA/KA receptor antagonist DNQX (50 µM) (data not shown). Altogether, the data suggest that the generation of network bursts in the CA1 area of the immature rats and the developmental difference in epileptiform bursting are independent of the GABA_A, GABA_B, NMDA, and mGlu1 components of synaptic transmission.

Figure 4.

The developmental difference in burst discharges was still present after the thickness of the slices was adjusted for increased brain volume during development and was independent of developmental changes in GABA_A, GABA_B, and N-methyl-d-aspartate (NMDA) receptors. Slices of different thickness were made for immature and adult rats (350 and 450 µm, respectively) to compensate for the developmental increase in brain volume. Bursting behavior was tested in these slices. A: whole-cell (top) and simultaneous field-potential (bottom) recordings showing that the majority of the 350-µm-thick slices from immature rats expressed network bursts. B: in contrast, the 450-µm-thick slices from adult rats displayed predominately graded bursts. In addition, these bursts were evoked in the presence of GABA_A, GABA_B, and NMDA receptor antagonists [gabazine, SCH50911, and dl-2-amino-5-phosphonopentanoic acid (AP-5), respectively] to show that these receptor mechanisms do not mediate bursting. Note that in the example shown in A, the slice responded to 10 consecutive threshold stimulations with bursts (7 of 10, 3 prolonged, 4 short; top 7 traces in A and B) with long and variable latency or no burst (3 of 10, bottom 3 traces in A and B). Arrows indicate the stimulation, and dashed lines represent the baseline voltage levels.

Reverberating EPSPs after Bursts

Interestingly, some CA1 minislices (3/13, 23%) from the immature rats exhibited “reverberating” EPSPs after an epileptiform burst. As shown in Fig. 5A, threshold stimulations evoked either full-blown bursts with action potentials (Fig. 5A, arrow) or no burst (Fig. 5A, triangle, bottom 2 traces). Notably, after epileptiform bursts multiple EPSPs appeared, lasting for a few hundred milliseconds (Fig. 5A, asterisk), but they did not occur after subthreshold responses (i.e., without action potential) and were never observed in CA1 minislices from adult rats (Fig. 5B). This phenomenon suggests that 1) these EPSPs were driven by the bursts and 2) some neurons that participated in the bursts had excitatory connections to the recorded neuron. The fact that the EPSPs were multiple and long-lasting strongly suggests that an interconnected neuronal network was responsible for the polysynaptic responses. Therefore, these data provide additional evidence for a synaptically connected neuronal network in the CA1 area of the immature rats.

Immature CA1 Pyramidal Cells Were More Capable of Intrinsic Generation of Action Potentials than Adult Pyramidal Cells

In addition to the synaptic mechanisms, the intrinsic membrane properties of individual neurons can affect network activity. The possibility of differences in action potential generation between the CA1 pyramidal cells in the immature and adult hippocampus was tested through current injection, in the absence of synaptic transmission (i.e., blocked by GABA_A, GABA_B, AMPA/KA, and NMDA receptor antagonists). The membrane potential was held at approximately −55 mV, and a series of hyperpolarizing and depolarizing current pulses (15-s duration) was injected into the pyramidal cells. The resting membrane potential of the CA1 pyramidal neurons was similar for the immature and adult rats (−56.4 ± 0.9 mV, n = 5 vs. −56.5 ± 1.6 mV, n = 6, P > 0.05), whereas the input resistance was greater in neurons from immature than mature rats (314.4 ± 76.8 MΩ, n = 5 vs. 129.8 ± 13.3 MΩ, n = 6, P < 0.05). During the depolarizing current injection, CA1 pyramidal cells from the immature hippocampus were capable of firing action potentials repetitively for the entire duration of the current pulse period (i.e., 15 s), at relatively low current intensities (Fig. 6A). In contrast, CA1 pyramidal cells from adult hippocampus required much larger depolarizing currents to fire action potentials (Fig. 6B). Even with the much larger depolarizing currents, the number of action potentials generated was fewer than that in the immature cells. Therefore, the input-output relationship (Fig. 6C) was lower for CA1 pyramidal cells from adults. These data suggest that CA1 pyramidal cells are more capable of firing action potentials in immature than adult rats, which may contribute to the enhanced propensity for network bursting in the immature CA.

Figure 6.

The CA1 pyramidal cells from immature rats had a greater intrinsic spike-generating capacity than those from adults. Whole-cell current-clamp recording was performed to examine the intrinsic property for action potential generation in pyramidal cells from immature and mature rats. A series of current injections over a 15-s period was tested. A: depolarizing current pulses at relatively low intensity generated robust action potential firing in CA1 pyramidal cells from immature hippocampus. B: in contrast, much larger depolarizing currents were required to excite CA1 pyramidal cells from adult hippocampus, which produced fewer action potentials. C: summarized data showing the difference in the input-output relationship of CA1 pyramidal cells between the immature and adult groups.

DISCUSSION

The present study investigated the underlying mechanisms for generation of different types of burst discharges in the CA1 area of the immature versus adult hippocampus after pharmacological blockade of GABA_A receptors. The main results of the study were as follows: 1) the CA1 area of the immature hippocampus generated mostly network-driven bursts, whereas the CA1 area from adult hippocampus produced predominantly graded bursts; 2) the developmental differences in bursting were still observed in slices adjusted in thickness for the increase in brain volume during development and the higher neuronal density of the immature brain; 3) the frequency and amplitude of the sEPSCs were greater in the immature than the adult group; 4) the network-driven bursts and the developmental differences in bursting were independent of developmental changes in GABA_A, GABA_B, NMDA, or mGlu1 receptor-mediated synaptic transmission; and 5) the input-output (amount of current injected vs. the number of generation of action potentials generated) relationship was markedly greater in CA1 pyramidal cells from the immature than the adult hippocampus. Together, these data showed that the CA1 area of the immature hippocampus possesses a high propensity to generate network-driven bursts, which decreases substantially in adults. The increased propensity of bursting in the immature CA1 appears to involve a greater degree of AMPA receptor-mediated synaptic interactions and a higher intrinsic spike-generating ability.

Patterns of Bursting in the CA1 Area of the Immature vs. Adult Hippocampus

In a previous paper (7), we showed two types of burst discharges in CA3 [i.e., prolonged (i.e., ∼10 s) network bursts with repetitive afterdischarges in the immature CA3 and brief (i.e., <0.2 s) network bursts in adult CA3]. The burst patterns in CA1 are distinct from those in the CA3. The duration of network bursts in the immature CA1 is relatively short. The network bursts in the immature CA1 lasted for a few hundred milliseconds to a maximum of a few seconds (Fig. 1A2 and Fig. 4A); therefore, these bursts were shorter than those in the immature CA3. Second, in the CA1 minislices from immature rats, threshold stimulation of the same intensity often elicited network bursts with a wide range of durations (Fig. 4A). Stimulation at higher intensity tended to evoke bursts of shorter duration (data not shown). Third, some of the CA1 slices from the immature (and to a lesser extent, in the mature) rats displayed mixed or “seminetwork” bursts (Fig. 1B). Finally, adult CA1 slices exhibited predominantly graded bursts (i.e., the classical responses to stimulation of a synaptic pathway, Fig. 1A1 and Fig. 4B).

The features of the network bursts in the immature CA1 are somewhat different from the classical all-or-none pattern, as seen in CA3 (7). The traditional explanation for the all-or-none behavior is that in a neural network of interconnected pyramidal cells when a threshold number of neurons are excited, the activity can spread (or fail to spread) along synaptic pathways to trigger a chain reaction of neuronal discharge until a synchronous burst occurs (26–28). Possibly, the neuronal network in the immature CA1 is less densely connected than that in the immature CA3; thus the initial activation of a small number of neurons by threshold stimulation has a greater chance to fail to spread through the whole network, either at the beginning (i.e., no burst) or soon after the stimulus (i.e., short bursts). Stronger stimulation may recruit more axon collaterals in the synaptic pathway and thus may activate more CA1 neurons at the beginning, which could recruit the network more quickly, thus generating short bursts. CA1 pyramidal cells are densely packed, and their activity may also be synchronized through nonsynaptic mechanisms (29, 30), particularly by electrical stimuli that activate many neurons. In contrast, CA1 minislices of the adult hippocampus showed a pattern of classical pathway stimulation; that is, the responses increased incrementally with increased stimulus intensity (i.e., more neurons are incrementally activated; Fig. 1A1 and Fig. 4B), which is consistent with the hypothesis that CA1 pyramidal cells of the adult hippocampus are only sparsely interconnected (31, 32). Some of the slices exhibited mixed responses of the network and graded bursts (mixed bursts; Fig. 1A3), which may reflect a transition from the immature to mature brain. Presumably, the local excitatory connections between CA1 pyramidal cells are pruned and recurrent excitation among CA1 pyramidal cells reduced. Thus, threshold stimulation may excite the neurons in the network with less recurrent excitation and thus generate a short all-or-none burst, whereas a further increase in stimulus intensity would activate the rest of the neurons and further intensify the burst in a graded manner.

Parallel Changes in sEPSCs

Several lines of evidence from our analysis of sEPSCs further support the hypothesis of a more robust local excitatory synaptic network of the immature versus adult CA1 area. First, more CA1 pyramidal cells tended to express sEPSCs in slices from immature than adult rats; second, the frequency of sEPSCs was higher in immature than mature pyramidal cells (Fig. 3); and third, CA1 pyramidal cells from the immature group contained more large-amplitude sEPSCs (Fig. 3). The increased frequency of sEPSCs might have resulted from an increased degree of local excitatory connectivity of the immature CA1 pyramidal neurons, an increased firing of the presynaptic neurons due to enhanced intrinsic firing capacity of the immature CA1 pyramidal neurons, or both. The large-amplitude sEPSCs in the immature group likely reflect the action potential-mediated synaptic activity from local synapses between CA1 pyramidal cells, because the projected axons from neurons in remote areas were cut during preparation of the CA1 minislices. Only axons with intact presynaptic neurons (i.e., neighboring CA1 pyramidal cells) would be expected to generate spontaneous action potentials, thus producing larger sEPSCs. An increased size of the synaptic network, together with an enhanced intrinsic property for action potential generation (see below), may contribute to the enhanced propensity for epileptiform bursting in the CA1 area of the immature hippocampus.

Role of AMPA/KA vs. NMDA, mGlu1, and GABA Receptor-Mediated Synaptic Network in Burst Discharges

Virtually every synaptic component (i.e., AMPA/KA, NMDA, GABA_A, GABA_B, and mGlu1 receptor mediated) undergoes developmental changes (33, 34), which may contribute to the developmental differences in bursting in CA1. In early development, the action of GABA is proposed to be excitatory (16), NMDA receptors have unique features favoring excitation (20–22), and mGlu1 receptors could prolong the burst discharges (24, 25). Our data show that bursts were consistently elicited in the CA1 area of immature hippocampus when GABA_A receptor-mediated synaptic transmission was blocked by either bicuculline or gabazine (Fig. 1A and Fig. 4A), thus arguing that from the third postnatal week the action of GABA is predominantly inhibitory and normally prevents bursts. Our data also show that network bursts were consistently elicited in the developing CA1 when NMDA receptors were blocked with AP-5 (Fig. 4A), suggesting that the activation of NMDA receptor-mediated synapses is not required for the generation of the network bursts, although NMDA receptors likely contribute to network bursts (35–37). The burst discharges in the immature CA1 were also not blocked by AIDA, suggesting that mGlu1 receptor-mediated synaptic excitation does not significantly contribute to the generation of these bursts. Similarly, in an earlier study (38), mGluR1s were not involved in burst discharges in the CA1 area from guinea pig. Therefore, these data suggest that although multiple components of excitatory synaptic transmission may contribute to burst discharges, the AMPA/KA receptor-mediated synaptic network is the predominant contributor to the increased propensity for bursting in the immature CA1 area.

Differential Intrinsic Capacity for Action Potential Firing in Immature vs. Adult CA1

The spreading of neuronal activity through an excitatory multisynaptic network and generation of network bursts relies on reliable action potential firing of the participating neurons in the neuronal network. The CA1 pyramidal cells from the immature hippocampus had a superior spike-generating capability, which could facilitate multisynaptic transmission and favor the generation of network bursts.

Implications for Acquired Epileptogenesis in the Temporal Lobe

Previous studies (10–15) have shown that during epileptogenesis after status epilepticus the CA1 pyramidal cells from adult rats become interconnected with excitatory synapses through axonal sprouting and are more susceptible to all-or-none epileptiform bursts (i.e., similar to the immature CA1). Therefore, the underlying mechanisms of acquired epileptogenesis (i.e., normal adult brain becomes seizure prone) may involve a reversal in the course of maturation. The hypothesis that acquired epileptogenesis after a brain injury involves a reversion to a more immature developmental state may also provide important insights into acquired epileptogenesis.

In summary, our data suggest that the CA1 area of the immature hippocampus has an increased propensity for network bursting, which is lost in adults. The increased bursting in immature CA1 involves a more extensive AMPA/KA receptor-mediated synaptic network among the CA1 pyramidal cells and an increased intrinsic spike-generating capacity of the pyramidal neurons. Furthermore, NMDA and mGlu1 receptor-mediated excitatory synaptic transmission and the excitatory action of GABA are not required for the generation of network bursting in the immature CA1, although they may contribute to the more pronounced bursting that can occur in the CA1 area of the hippocampus. In addition, other mechanisms may also shape the bursting behavior in the immature CA1, such as low voltage-activated Ca²⁺ current (39), dendritic Ca²⁺ signaling (40), tonic GABAergic inhibition (41), Na⁺-K⁺-ATPase (42), and endocannabinoids (43).

GRANTS

This work was supported by National Institute of Neurological Disorders and Stroke Grant NS016683.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

L.-R.S. and F.E.D. conceived and designed research; L.-R.S. performed experiments; L.-R.S. analyzed data; L.-R.S. and F.E.D. interpreted results of experiments; L.-R.S. prepared figures; L.-R.S. drafted manuscript; L.-R.S. and F.E.D. edited and revised manuscript; L.-R.S. and F.E.D. approved final version of manuscript.