Calcium-dependent Epileptogenesis in an In Vitro Model of Stroke-induced “Epilepsy”

Authors


Address correspondence and reprint requests to Dr. R. J. DeLorenzo at Box 980599, Medical College of Virginia, Virginia Commonwealth University, Richmond, VA 23298, U.S.A. E-mail: rdeloren@hsc.vcu.edu

Abstract

Summary:  Purpose: Stroke is the most common cause of acquired epilepsy. The purpose of this investigation was to characterize the role of calcium in the in vitro, glutamate injury–induced epileptogenesis model of stoke-induced epilepsy.

Methods: Fura-2 calcium imaging and whole-cell current clamp electrophysiology techniques were used to measure short-term changes in neuronal free intracellular calcium concentration and long-term alterations in neuronal excitability in response to epileptogenic glutamate injury (20 μM, 10 min) under various extracellular calcium conditions and in the presence of different glutamate-receptor antagonists.

Results: Glutamate injury–induced epileptogenesis was associated with prolonged, reversible elevations of free intracellular calcium concentration during and immediately after injury and chronic hyperexcitability manifested as spontaneous recurrent epileptiform discharges for the remaining life of the cultures. Epileptogenic glutamate exposure performed in solutions containing low extracellular calcium, barium substituted for calcium, or N-methyl-d-aspartate (NMDA)-receptor antagonists reduced the duration of intracellular calcium elevation and inhibited epileptogenesis. Antagonism of non–NMDA-receptor subtypes had no effect on glutamate injury–induced calcium changes or the induction epileptogenesis. The duration of the calcium elevation and the total calcium load statistically correlated with the development of epileptogenesis. Comparable elevations in neuronal calcium induced by non–glutamate receptor–mediated pathways did not cause epileptogenesis.

Conclusions: This investigation indicates that the glutamate injury–induced epileptogenesis model of stroke-induced epilepsy is calcium dependent and requires NMDA-receptor activation. Further, these experiments suggest that prolonged, reversible elevations in neuronal free intracellular calcium initiate the long-term plasticity changes that underlie the development of injury-induced epilepsy.

Epilepsy is one of the most common neurologic disorders, affecting an estimated 40 to 50 million people worldwide (1). Approximately 50% of all epilepsy cases have a known cause and are termed acquired epilepsy (2). Epileptogenesis is the process by which normal central nervous system tissue is transformed into brain tissue prone to the manifestation of spontaneous recurrent seizures (3). Neuronal injury in the form of stroke is the most common factor associated with epileptogenesis and acquired epilepsy (4). Despite the important role of stroke in the development of epilepsy, little is known regarding the mechanisms by which neuronal injury initiates epileptogenesis in the surviving tissue.

Our laboratory recently developed and characterized an in vitro model of injury-induced epileptogenesis to replicate the development of epilepsy as a consequence of stroke (5). In this model, injury produced by exposure of cultured hippocampal neurons to glutamate resulted in a mixed population of injured neurons that developed excitotoxic neuronal death and a larger population of neurons that survived, analogous to the ischemic penumbra (6). These surviving neurons manifested spontaneous, recurrent, epileptiform discharges (SREDs) in synchronized neural networks for the remaining life of the cultures. This in vitro model of glutamate injury–induced epileptiform activity provides a powerful tool to evaluate the molecular mechanisms mediating stroke-induced epileptogenesis.

The calcium (Ca2+) hypothesis of epileptogenesis suggests that alterations in neuronal Ca2+ homeostasis play an essential signaling role in the pathogenesis of epilepsy (7,8). Whereas excessive activation of the N-methyl-d-aspartate receptor (NMDAR) subtype of glutamate receptors and irreversible elevations in free intracellular calcium concentration ([Ca2+]i) have been implicated in excitotoxic neuronal death (9), NMDAR activation and prolonged but reversible elevations in [Ca2+]i have been implicated in the induction of epilepsy (8,10–14). This study was initiated to test the hypothesis that prolonged elevations in [Ca2+]i and NMDAR activation are required for glutamate injury–induced epileptogenesis in cultured hippocampal neurons. To this end, we used fluorescent Ca2+ imaging and electrophysiological techniques to characterize the ability of injury induced by glutamate exposure to cause short-term changes in neuronal [Ca2+]i and prolonged changes in neuronal excitability in the presence of different Ca2+ conditions and various glutamate-receptor antagonists.

METHODS

Materials

Unless otherwise noted, reagents were purchased from Sigma Chemical Co. (St. Louis, MO, U.S.A.). Sodium pyruvate, minimum essential medium (MEM) containing Earle's salts, fetal bovine serum, and horse serum were obtained from Gibco-BRL (Gaithersburg, MD, U.S.A.). 2-Amino-5-phosphonovalerate (APV), 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), 5-methyl-10,11-dihydro-5-H-dibenzocyclohepten-5,10-imine maleate (MK-801), and S-α-methyl-4-carboxyphenylglycine (MCPG) were purchased from Tocris Cookson, Inc. (Ellisville, MO, U.S.A.).

Hippocampal cell culture

Primary mixed hippocampal cultures were prepared by a modified method of Banker and Cowan (15). In brief, hippocampal tissue was dissected from the brains of 2-day-postnatal Sprague–Dawley rats (Harlan, Frederick, MD, U.S.A.), stripped of meninges and vasculature, and treated with 0.25% trypsin at 37°C for 30 min. After enzymatic treatment, mechanical trituration of the tissue was performed through a fire-polished Pasteur pipette. The resultant single-cell suspension was diluted to a final concentration of 1 × 105 cells/ml in a neuronal feeding medium [MEM with Earle's Salts (MEM), 25 mM N-2-hydroxyethylpiperazine-N′-2-ethane sulfonic acid (HEPES), 2 mMl-glutamine, 3 mM glucose, 100 μg/ml transferrin, 5 μg/ml insulin, 100 μM putrescine, 3 nM sodium selenite, 200 nM progesterone, 1 mM sodium pyruvate, 0.1% ovalbumin, 0.2 ng/ml triiodothyroxine, and 0.4 ng/ml corticosterone] supplemented with 5% horse serum, and plated on confluent glial support layers grown on poly-l-lysine–coated (0.05 mg/ml) 35-mm plastic culture dishes or Lab-Tek coverglass chambers (Nunc, Naperville, IL, U.S.A.) for 14 days at 37°C in a 5% CO2/95% air atmosphere, and fed with glial feeding medium (MEM, 2 mMl-Glutamine, 3 mM glucose, and 10% fetal bovine serum). Twenty-four hours after plating, cultures were treated with 5 mM cytosine arabinoside for 48 h to inhibit nonneuronal growth. This technique prevents overgrowth of the glial elements in the cultures. Cultures were maintained at 37°C in a 5% CO2/95% air atmosphere and fed twice weekly with neuronal feed supplemented with 20% conditioned medium. Neuronal feed in this modified preparation was supplemented with 20% conditioned medium (neuronal feed supplemented with 5% horse serum harvested after 48-h exposure to confluent glial beds). Using conditioned media optimized neuronal culture stability and provided long-term survival in control and experimental cultures. These mixed cultures were used for experiments from 13 days in vitro (DIV) through the life of the cultures (∼21 DIV).

Glutamate injury protocol

At 13 DIV, neuronal feed was replaced with a physiologic recording solution (145 mM NaCl, 2.5 mM KCl, 10 mM HEPES, 10 mM glucose, 2 mM CaCl2, 1 mM MgCl2, 2 μM glycine, pH 7.3, osmolarity adjusted to 325 mOsm with sucrose) in control cultures. This osmolarity was used as described previously (5) and is commonly used in tissue-culture electrophysiology. Epileptogenic glutamate injury was produced by exposing cultures to recording solution supplemented with 20 μM glutamate for 10 min. This treatment produced a neuronal injury, previously defined (5) as neurons surviving the glutamate treatment, but manifesting several signs of reversible or transient injury, manifested by reversible cell swelling, prolonged depolarization, and altered input resistance. These transiently injured neurons developed epileptogenesis (5).

This glutamate treatment was found to produce ∼40% cell death. The surviving 60% of the neurons developed epileptic discharges. These results were essentially identical to the 5 μM glutamate treatment for 30 min of exposure reported previously (5). In our current cultures, this higher glutamate concentration for a shorter duration was less influenced by neuronal density and endogenous glutamate levels in the culture media. The seizure patterns, cell death, and characterization of epileptic discharges were the same as described previously (5). All exposures were performed at 37°C in a 5% CO2/95% air atmosphere and terminated by three washes with recording solution and the addition of fresh neuronal feed. Because 13 DIV is a normally scheduled feeding day, these cultures maintained the same feeding protocol as naïve cultures.

High extracellular potassium solutions were prepared identically to physiologic recording solutions except that KCl was increased from 2.5 to 50 mM and NaCl was reduced from 145 to 97.5 mM. Treatment of the neurons with this solution was used to depolarize neurons and induce calcium influx through non–glutamate-receptor pathways.

Electrophysiology

Whole-cell current-clamp recordings were performed on pyramidal neurons by using methods previously described in our laboratory (5). In brief, cell-culture medium was replaced with recording solution (described earlier), and cultures were transferred to a heated stage (Brook Industries, Lake Villa, IL, U.S.A.) on a Nikon Diaphot inverted microscope (Garden City, NY, U.S.A.). Patch microelectrodes of 3–7 MΩ resistance were filled with an internal solution of 140 mM K+ gluconate, 1 mM MgCl2, 10 mM HEPES, 1.1 mM ethylene glycol-bis (β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid, 4 mM Na2 adenosine triphosphate (ATP), 15 mM Tris phosphocreatine, pH 7.2, with osmolarity adjusted to 310 mOsm with sucrose.

Recordings were obtained in the whole-cell current-clamp configuration (16) by using an Axopatch 200A amplifier (Axon Instruments, Foster City, CA, U.S.A.) in current-clamp mode. Data were digitized and stored on videotape by using a Neuro-corder DR-890 (Neurodata Instruments Corp., New York, NY, U.S.A.) and Sony VCR. Data later were played back for analysis to a Dash IV chart recorder (Astro-Med Inc., Warwick, RI, U.S.A.).

SREDs were defined as bursts of spike firing at a frequency of ≥3 Hz for durations of ≥20 s (5), analogous to electrographic seizures (17). Neurons were categorized as “epileptic” on manifestation of two or more SREDs; otherwise, neurons were categorized as “nonepileptic.” Based on previous experiments, neurons were monitored for a recording period of ≥10 min for the expression of SREDs (5).

Measurement of neuronal [Ca2+]i

Dye loading

Changes in neuronal [Ca2+]i were measured by using the ratiometric, high-affinity (Kd≈ 224 nM) fluorescent calcium indicator, Fura-2 (18). Neurons were loaded with 1 μM Fura-2 AM (Molecular Probes, Eugene, OR, U.S.A.) dissolved in recording solution (0.1% DMSO) for 1 h at 37°C in a 5% CO2/95% air atmosphere. Dye loading was terminated with three washes with recording solution and incubated an additional 15 min to allow complete cleavage of Fura-2 AM.

Microfluorometry

Cultures grown on Lab-Tek coverglass chambers (Nunc) were visualized on an inverted microscope (Olympus IX 70; Olympus America, Melville, NY, U.S.A.) by using a ×20, 0.7 numerical aperture fluorite water-immersion objective maintained at 37°C with a heated stage (Harvard Apparatus Inc.). Fura-2 was excited with a 75-W xenon arc lamp (Olympus Optical Co., Shinjuku-ku, Tokyo, Japan) with alternating wavelengths of 340 and 380 nm filtered through a Sutter Filter Wheel (Sutter Instruments Co., Novato, CA, U.S.A.). Fluorescent emission at 510 nm was captured through a Fura filter cube (Olympus America) with a dichroic at 400-nm emission by using a cooled digital CCD camera (LSR AstroCam Limited, Cambridge, England).

Ca2+ calibration curves

An in situ Ca2+ calibration curve was developed to convert fluorescent ratios to Ca2+ concentrations, as described by Pal et al. (19) with the equation

image

where R is the 340/380 ratio at any time; Rmax is the maximal measured ratio in saturating Ca2+ solutions; Rmin is the minimal measured ratio in Ca2+-free solutions; Sf2 is the absolute value of the corrected 380-nm signal at Rmin; Sb2 is the absolute value of the corrected 380-nm signal at Rmax; and the Kd is 224 nM(18).

The Temporal Module of the Perkin Elmer Life Sciences Imaging Suite, Version 4.0 (Gaithersburg, MD, U.S.A.), was used to control image acquisition and processing. Image pairs were captured once per minute, and background correction for nonspecific fluorescence was achieved by subtracting images acquired from non–indicator-loaded cultures. A region of interest was designated for each pyramidal neuron in the microscope field (five to eight neurons). Ratios for each neuron were measured from the corresponding region of interest and calibrated. Individual neurons from multiple experiments were pooled to determine mean ± SEM. A criterion of [Ca2+]i >400 nM sustained 120 min after glutamate application was used to define neuronal inability to restore resting [Ca2+]i. which is known to correlate with delayed excitotoxic neuronal death (20).

Statistical analysis

Data are reported as mean ± SEM. One-way analysis of variance (ANOVA) was used to compare the peak percentage increase in neuronal [Ca2+]i. Repeated measures analysis of variance (RM ANOVA) was performed on every second data point in Ca2+ imaging experiments. The Tukey test was used for all post hoc multiple comparisons. Fisher's Exact test was applied to categoric data comparing the percentage of neurons made epileptic with or without drug treatment. The Pearson product–moment correlation was used for all analysis of correlation. Statistical analysis was performed by using SigmaStat 2.0 (Jandel Corp., San Rafael, CA, U.S.A.). p < 0.05 was considered statistically significant for all data analysis.

RESULTS

Glutamate injury–induced epileptogenesis was inhibited by reduced [Ca2+]e

Control neurons displayed spontaneous action potentials, excitatory postsynaptic potentials (EPSPs), and inhibitory post-synaptic potentials (IPSPs) typical of normal neuronal activity (n = 37; Figs. 1A and 2A) and never manifested SREDs (Fig. 3). In contrast, 90% of neurons surviving an epileptogenic glutamate injury (20 μM, 10 min) manifested SREDs ≥24 h after glutamate exposure (p < 0.001; n = 30; Figs. 1B and 3). As described previously by our laboratory (5), SREDs were characterized by paroxysmal depolarizing shifts (PDSs) and high-frequency spike firing characteristic of electrographic seizures (Figs. 1B and 2B)(17).

Figure 1.

Glutamate injury–induced epileptogenesis in cultured hippocampal neurons. A: Representative current-clamp recording demonstrating normal spike activity of a control neuron (–64 mV membrane potential). Aa: Portion of the trace depicted in A (black bar) displayed at a faster time scale. Note the spontaneous action potentials, excitatory postsynaptic potentials (EPSPs) and inhibitory PSPs (IPSPs). B: Representative current-clamp recording indicative of the epileptiform activity manifested in neurons recorded 24 h after glutamate injury (5 μM, 30 min). This neuron (–62 mV membrane potential) manifested three spontaneous, recurrent, epileptiform discharges (SREDs) during a 20-min recording period. The three individual SREDs (indicated by black bars) of 1.4 min (Ba), 0.72 min (Bb), and 1.3 min (Bc) displayed at a faster time scale. SREDs were characterized by abruptly developing, paroxysmal depolarizing shifts of membrane potential with high-frequency spike firing.

Figure 2.

Whole-cell current-clamp recordings ≥24 h after epileptogenic glutamate injury performed in the presence of various extracellular Ca2+ conditions and glutamate receptor antagonists. A: Control neurons exhibited spontaneous postsynaptic potentials, which occasionally reached threshold for the generation of action potentials. B: Neurons surviving glutamate injury (20 μM, 10 min) manifested spontaneous, recurrent, epileptiform discharges (SREDs) characterized by paroxysmal depolarizing shifts (PDSs) and high-frequency spike firing. C, D: Neurons injured by glutamate in low-[Ca2+]e (0.2 mM) solutions (C) or Ba2+ substituted for extracellular Ca2+ solutions (D) never manifested SREDs. E: Glutamate injury performed in the presence of 2-amino-5-phosphonovalerate (APV; 50 μM) did not induce epileptogenesis. F, G: The non–N-methyl-d-aspartate receptor antagonists, 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; 2 μM; F) and S-α-methyl-4-carboxyphenylglycine (MCPG; 250 μM; G) did not inhibit glutamate injury–induced epileptogenesis.

Figure 3.

Glutamate injury–induced epileptogenesis in the presence of various extracellular Ca2+ conditions and glutamate receptor antagonists. Of neurons surviving an epileptogenic glutamate injury (GLU; n = 30), 90% manifested spontaneous, recurrent, epileptiform discharges (SREDs) not observed in controls (n = 37). Glutamate injury performed in low-[Ca2+]e (low Ca2+, n = 11), Ba2+ substitution (barium, n = 11), and in the presence of the competitive 2-amino-5-phosphonovalerate (APV, n = 12) and noncompetitive 5-methyl-10,11-dihydro-5-H-dibenzocyclohepten-5,10-imine maleate (MK-801, n = 10) N-methyl-d-aspartate receptor antagonists inhibited the induction of epileptogenesis. Glutamate injury–induced epileptiform activity was observed in 80% of neurons injured in the presence of 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; n = 10), as well as in S-α-methyl-4-carboxyphenylglycine (MCPG; n = 10). *p < 0.001, Fisher's Exact test versus control. Data are represented by percentages.

To test the hypothesis that glutamate injury–induced epileptogenesis was mediated through a Ca2+-dependent process, we injured neurons with the same epileptogenic glutamate exposure (20 μM glutamate, 10 min) in a reduced [Ca2+]e solution (2.0 mM[Ca2+]e reduced to 0.2 mM). Neurons surviving glutamate injury in this low-[Ca2+]e environment exhibited normal levels of activity similar to that of controls (Fig. 2C) and never manifested SREDs (p = 1.0; n = 11; Fig. 3).

Ca2+ is an established second messenger in neurons (21,22). To determine whether Ca2+ had a significant role as a second-messenger signal in glutamate injury–induced epileptogenesis, epileptogenic glutamate exposure was performed in Ca2+-free, barium (Ba2+)-substituted solutions. Although Ba2+ carries a divalent positive charge across the neuronal membrane through conventional routes of Ca2+ entry (23,24), Ba2+ does not activate Ca2+-sensitive second-messenger pathways (25,26). Extracellular solutions with a 2.0 mM Ba2+ concentration blocked glutamate injury–induced epileptogenesis (p = 1.0; n = 11; Figs. 2D and 3). These results indicate that glutamate injury–induced epileptogenesis required extracellular Ca2+ and activation of intracellular Ca2+ pathways, leading to the development of persistent neuronal hyperexcitability.

Glutamate injury–induced epileptogenesis was inhibited by antagonism of the NMDA receptor

To test the hypothesis that NMDAR activation was required for glutamate injury–induced epileptogenesis, we exposed cultured hippocampal neurons to epileptogenic glutamate concentrations in the presence of the competitive NMDAR antagonist, APV (27). Neurons injured in the presence of APV (50 μM) exhibited activity similar to that of control neurons (Fig. 2E), never manifesting SREDs (p = 1.0; n = 10; Fig. 3). Likewise, the noncompetitive NMDAR antagonist, MK-801 (10 μM) (28) significantly blocked glutamate injury–induced epileptogenesis (p = 1.0; n = 10; Fig. 3). Therefore activation of the NMDAR subtype of glutamate receptors was required for glutamate injury–induced epileptogenesis.

Antagonism of non-NMDAR subtypes of glutamate receptors did not inhibit glutamate injury–induced epileptogenesis

Certain configurations of the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)- and kainate (KA)-receptor subtypes of glutamate receptors also are permeable to Ca2+(29). Therefore we tested the role of AMPA and KA receptors in glutamate injury–induced epileptogenesis by exposing cultured hippocampal neurons to epileptogenic concentrations of glutamate (20 μM, 10 min) in the presence of the competitive AMPA- and KA-receptor antagonist, CNQX (2 μM) (30). Of neurons surviving epileptogenic glutamate injury in CNQX 80% manifested SREDs (Fig. 2F), statistically greater than in controls (p < 0.001; n = 10; Fig. 3). This rate of epileptogenesis was not statistically different compared with glutamate injury in the absence of CNQX (p = 0.584; n = 10).

G protein–coupled metabotropic glutamate receptors (mGluRs) alter neuronal [Ca2+]i through the modulation of intracellular Ca2+ stores, ligand-gated ion channels, and voltage-gated ion channels (31). Therefore we blocked mGluRs with the competitive antagonist, MCPG (250 μM) (32), to investigate the potential role of mGluR-mediated Ca2+ signaling during epileptogenic glutamate injury. MCPG had no effect on glutamate injury–induced epileptogenesis, with 80% of neurons recorded manifesting SREDs ≥24 h after glutamate injury, statistically greater than in controls (p < 0.001; n = 10; Figs. 2G and 3). This rate of epileptogenesis was not statistically different compared with glutamate injury in the absence of MCPG (p = 0.584, n = 10).

Epileptogenic glutamate injury caused prolonged, reversible increases in neuronal [Ca2+]i

To determine whether the glutamate injury–induced epileptogenesis was associated with changes in neuronal [Ca2+]i, we monitored neuronal [Ca2+]i before, during, and after epileptogenic glutamate exposure by using the ratiometric, high-affinity, calcium indicator Fura-2 (18). Control neurons that were exposed to solution changes without glutamate did not manifest elevations in [Ca2+]i, as demonstrated by the representative neuron in Fig. 4A and the time course in Fig. 5A. In contrast, [Ca2+]i increased rapidly on exposure to epileptogenic glutamate concentrations (Fig. 4B). On washout, the majority of neurons slowly restored basal Ca2+ levels (Figs. 4B and 5B). A subset of neurons (six of 21), however, still maintained [Ca2+]i >400 nM, even after 120 min of recording (data not shown). Neurons with sustained elevations in [Ca2+]i >400 nM for durations of ≥120 min were categorized as having an inability to restore resting [Ca2+]i (IRRC). Because neurons with IRRC undergo delayed excitotoxic neuronal death (20), this subset of neurons was excluded from statistical analysis. In the remaining neurons, [Ca2+]i was statistically elevated over basal levels from the time of exposure through 76 min of recording (66 GLU; p < 0.05; n = 15, Fig. 5B). Thus glutamate injury–induced epileptogenesis was associated with prolonged, reversible increases in neuronal free calcium.

Figure 4.

Representative pseudo-color digital images of Fura-2–loaded neurons during epileptogenic glutamate injury in the presence of various extracellular Ca2+ conditions and glutamate receptor antagonists. A: The representative control neuron did not undergo changes in [Ca2+]i during the recording period. B–F: In contrast, neuronal [Ca2+]i increased on treatment with glutamate in all other experimental conditions, as indicated by the change in color from blue to red. The representative neurons treated in nonepileptogenic conditions of low [Ca2+]e (C) and N-methyl-d-aspartate receptor antagonism (D) restored basal neuronal [Ca2+]i within 60 and 30 min, respectively. In contrast, neuronal [Ca2+]i measured in epileptogenic conditions of glutamate alone (B) and in the presence of antagonists of AMPA/KA receptors (E) and mGluRs (F) were elevated for ≥90 min.

Figure 5.

Alterations in neuronal [Ca2+]i in response to epileptogenic glutamate injury in the presence of various extracellular Ca2+ conditions and glutamate receptor antagonists. A: The [Ca2+]i in control neurons did not change during the recording period (n = 11). B: In contrast, application of epileptogenic glutamate (20 μM, 10 min, black bar) caused a significant and persistent increase in neuronal [Ca2+]i that was statistically greater than basal values for 76 min (n = 15). C, D: Nonepileptogenic glutamate injury (black bar) in low [Ca2+]e (C, n = 19) and 2-amino-5-phosphonovalerate (APV; D, n = 16) caused significant elevations in neuronal [Ca2+]i persisting for only 16 and 28 min, respectively. E, F: Similar to glutamate injury alone (B), epileptogenic glutamate injury in the presence of 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; E, n = 16) and S-α-methyl-4-carboxyphenylglycine (MCPG; F, n = 13) caused significant and persistent elevations of neuronal [Ca2+]i for 90 and 100 min, respectively. *p < 0.05 repeated measures analysis of variance and Tukey post hoc test. Data are represented by mean ± SEM.

Prolonged, reversible increases in neuronal [Ca2+]i associated with glutamate injury–induced epileptogenesis were reduced in low-[Ca2+]e environments

To test the hypothesis that the prolonged [Ca2+]i elevations observed during the epileptogenic glutamate injury played a critical role in the development of the epileptic condition, we measured the [Ca2+]i response of neurons in low-[Ca2+]e conditions known to inhibit glutamate injury–induced epileptogenesis (Fig. 2C). Neuronal [Ca2+]i rapidly increased in response to glutamate despite the low-[Ca2+]e environment (Figs. 4C and 5C). However, [Ca2+]i was rapidly restored in low [Ca2+]e, and neurons did not manifest IRRC. [Ca2+]i was statistically elevated over baseline for only 16 min of recording (6 min after GLU; p < 0.05; n = 19; Fig. 5C). Thus extracellular Ca2+ contributed to the elevations in [Ca2+]i observed during glutamate injury–induced epileptogenesis.

Prolonged, reversible increases in neuronal [Ca2+]i associated with glutamate injury–induced epileptogenesis were reduced by antagonism of the NMDA receptor

To test the hypothesis that the NMDA receptor mediated the prolonged, reversible increases in neuronal [Ca2+]i during epileptogenic glutamate injury, we monitored the neuronal [Ca2+]i response to epileptogenic glutamate in the presence of APV (50 μM). Similar to the low-[Ca2+]e experiments, [Ca2+]i rapidly increased in response to glutamate despite antagonism by APV, as demonstrated by the representative neuron in Fig. 4D. Similar to the low-[Ca2+]e experiments, basal [Ca2+]i also was rapidly restored in all neurons recorded (n = 16). [Ca2+]i was significantly elevated over basal values for only 28 min (18 min after GLU; p < 0.05; n = 16; Fig. 5D). Experiments performed in the presence of MK-801 (10 μM) demonstrated similar results (data not shown). In MK-801, [Ca2+]i was significantly elevated over basal values for only 12 min (2 min after GLU; p < 0.05; n = 17; data not shown), and all neurons restored resting [Ca2+]i. Thus activation of the NMDA receptor was required for the prolonged elevations in [Ca2+]i observed during glutamate injury–induced epileptogenesis.

Prolonged, reversible increases in neuronal [Ca2+]i associated with glutamate injury–induced epileptogenesis were not reduced by antagonism of non-NMDAR subtypes

Antagonism of AMPA and KA receptors, as well as antagonism of class I and II mGluRs subtypes had no inhibitory effect on glutamate injury–induced epileptogenesis (Fig. 3). Likewise, pharmacologic blockade of AMPA and KA receptors by CNQX (2 μM; Fig. 4E) and pharmacologic blockade of mGluRs by MCPG (250 μM; Fig. 4F) had no inhibitory effect on the prolonged elevations in neuronal [Ca2+]i measured in response to epileptogenic glutamate exposure. When exposed to glutamate and CNQX, neurons manifested prolonged elevations in neuronal [Ca2+]i(Fig. 5E). Although a subset of neurons manifested IRRC (four of 20), the majority of neurons buffered the alterations in [Ca2+]i over time. Neurons were statistically elevated over basal levels, with exclusion of neurons manifesting IRRC, for 98 min (88 min after GLU; p < 0.05; n = 16; Fig. 5E).

Similarly, neurons exposed to glutamate and MCPG manifested prolonged elevations in neuronal [Ca2+]i(Fig. 4F). Of these neurons, 28% manifested IRRC (five of 18). The remaining neurons were statistically elevated over basal level for 100 min (90 min after GLU; p < 0.05; n = 13; Fig. 5F). Thus inhibition of non-NMDAR subtypes of glutamate receptors did not reduce the duration of prolonged [Ca2+]i elevations in glutamate injury–induced epileptogenesis.

The induction of epileptogenesis correlated with the duration of the [Ca2+]i elevation and the total [Ca2+]i load

The experiments described demonstrated that extracellular Ca2+ and NMDAR activation were required for the induction of glutamate injury–induced epileptogenesis. Further, Ca2+ imaging experiments revealed that epileptogenic glutamate injury was associated with prolonged elevations of neuronal [Ca2+]i. Various parameters of injury-induced [Ca2+]i changes during different experimental treatments are summarized in Table 1, including the duration of statistically elevated [Ca2+]i (minutes), the total [Ca2+]i load as reflected by the area under the curve (nanomoles per liter × minutes), and the peak increase in [Ca2+]i (percentage of basal). The total [Ca2+]i load associated with the epileptogenic treatments of glutamate alone (50,890 ± 9,262 nM× min; n = 15), glutamate with CNQX (25,666 ± 5,074 nM× min; n = 16), and glutamate with MCPG (31,731 ± 6,058 nM× min; n = 13) was statistically greater than the total [Ca2+]i load of the nonepileptogenic treatments of controls (62 ± 34 nM× min; n = 11), glutamate in low [Ca2+]i (4,560 ± 827 nM× min; n = 19), glutamate with APV (4,416 ± 444 nM× min; n = 16), and glutamate with MK-801 (520 ± 144 nM× min; n = 17). The peak increase in neuronal [Ca2+]i was largest in the normal epileptogenic glutamate treatment (983 ± 166%; n = 15). Treatment conditions of controls (18 ± 10%; n = 11), low [Ca2+]e (476 ± 80%; n = 19), MK-801 (162 ± 45%; n = 17), and CNQX (358 ± 102%; n = 16) were significantly less than those in the glutamate injury (p < 0.05; Table 1). In contrast, the peak increase in the presence of APV (781 ± 123%; n = 16) and MCPG (521 ± 154%; n = 13) was not statistically different from glutamate alone. A statistically significant correlation was observed between the induction of epileptogenesis and the duration of statistically elevated [Ca2+]i (r = 0.95; p < 0.002). In addition, the induction of epileptogenesis and the total change in neuronal [Ca2+]i demonstrated a statistically significant correlation (r = 0.94; p < 0.002). In contrast, no statistical correlation was observed between epileptogenesis and the peak increase in neuronal [Ca2+]i (r = 0.46; p = 0.305).

Table 1.  Quantification of [Ca2+]i changes during and after glutamate injury and the induction of epileptogenesis in various treatment conditions
TreatmentDuration of [Ca2+]i
elevation (min)
Total [Ca2+]i load
(nM× min)
Peak [Ca2+]i
(% of basal)
Epilepsy (%, n)
  • *

    p < 0.05 one way ANOVA and Tukey post-hoc test vs. Control.

  • p < 0.05 one way ANOVA and Tukey post-hoc test vs. Glutamate.

  • ††

    p < 0.05 Fisher Exact Test vs. Control.

Control062 ± 6418 ± 10No (%, 37)
Glutamate (20 μM)7650890 ± 9262*983 ± 166*Yes (90%, 30)††
+Low [Ca2+]e164560 ± 827476 ± 80No (0%, 11)
+APV (50 μM)284416 ± 444781 ± 123*No (0%, 12)
+MK-801 (10 μM)12520 ± 144162 ± 45No (0%, 10)
+CNQX (2 μM)9825666 ± 5074*358 ± 102Yes (80%, 10)††
+MCPG (250 μM)10031731 ± 6058*521 ± 154Yes (80%, 10)††

Prolonged, reversible increases in neuronal [Ca2+]i induced by non–glutamate-mediated mechanisms were not associated with the induction of epileptogenesis

The previously described results in this study indicated that prolonged, reversible elevations in the NMDAR neuronal [Ca2+]i pathway were associated with the induction of epileptogenesis. To test the hypothesis that equal non–NMDAR-mediated [Ca2+]i elevations did not initiate SREDs, we produced a prolonged, reversible elevation in neuronal [Ca2+]i by using high extracellular potassium solution to induce neuronal depolarization that was of similar duration and equal to or greater than the epileptogenic glutamate-mediated calcium response. Exposure of neurons to high extracellular potassium (50 mM, 90 min) induced a statistically significant elevation in neuronal [Ca2+]i for the duration of the treatment (Fig. 6A). The peak increase in [Ca2+]i during high potassium exposure (958 ± 194%; n = 13) was not statistically different from epileptogenic glutamate treatments (GLU alone, GLU + CNQX, or GLU + MCPG). Likewise, the total calcium load during high extracellular potassium exposure (35,508 ± 4,085 nM× min; n = 13) was not statistically different from the epileptogenic conditions of glutamate treatment alone and glutamate in the presence of CNQX or MCPG. Although these parameters of peak increase and total calcium load for epileptogenic glutamate treatments and non-NMDA, potassium-induced calcium entry were not statistically different, and potassium-induced elevations in neuronal [Ca2+]i were statistically increased for a similarly prolonged duration of 90 min, this treatment condition did not induce epileptogenesis (Fig. 6B; none; n = 10). This result demonstrated that the equal non–NMDAR-mediated [Ca2+]i elevations did not produce SREDs.

Figure 6.

Prolonged elevations in [Ca2+]i by exposure to high extracellular potassium did not induce epileptogenesis. A: Cultured hippocampal neurons (n = 13) exposed to a high-potassium (50 mM, 90 min, black bar) solution underwent statistically significant, reversible elevations in [Ca2+]i. B: A representative whole-cell current-clamp recording from neurons 24 h after exposure to high-potassium solutions manifested spontaneous postsynaptic potentials and occasional action potentials similar to those of control neurons. Neurons exposed to high potassium never manifested spontaneous, recurrent, epileptiform discharges (SREDS; n = 10). *p < 0.05 repeated measures analysis of variance and Tukey post hoc test. Data are represented by mean ± SEM.

DISCUSSION

The Ca2+ hypothesis of epileptogenesis suggests that prolonged elevations in [Ca2+]i play a role in mediating some of the long-term plasticity changes associated with epileptogenesis and the persistent manifestation of seizure activity (7,8). In addition, activation of the Ca2+-permeable NMDAR has been associated with epileptogenesis (10–14) and has been implicated as a major source of the epileptogenic Ca2+ signal (8). The results of this investigation strongly support this Ca2+ hypothesis of epileptogenesis as an initiating mechanism in glutamate injury–induced epileptogenesis. Glutamate injury (20 μM, 10 min) causing epileptogenesis in 90% of neurons (n = 30; Figs. 1B, 2B, and 3) was associated with prolonged elevations in neuronal [Ca2+]i(Figs. 4B and 5B). This elevation was statistically significant for >1 h (Fig. 5B; Table 1). Similar elevations in neuronal [Ca2+]i were observed when glutamate injury was performed in the presence of the AMPA- and KA-receptor antagonist, CNQX, and in the presence of the mGluR antagonist, MCPG (Fig. 5E and F; Table 1). Of neurons recorded after epileptogenic glutamate injury in CNQX or MCPG, 80% exhibited epileptiform activity (Figs. 2F and G and 3). The results provide direct evidence that persistent elevations in [Ca2+]e after glutamate exposure contribute to the development of persistent epileptiform discharges in the in vitro, glutamate injury–induced epileptogenesis model of stroke-induced “epilepsy.”

In contrast, experimental treatment conditions that reduced the duration of statistically elevated [Ca2+]i likewise inhibited the development of epilepsy. In low-[Ca2+]e solutions (2.0 mM reduced to 0.2 mM), neuronal [Ca2+]i was statistically elevated over basal levels for only 16 min (Fig. 5C). Antagonism of the NMDAR with APV (Fig. 5D) and MK-801 reduced the duration of statistically elevated [Ca2+]i to 28 and 12 min, respectively (Table 1). Twenty-four hours after glutamate injury in these experimental conditions, neuronal activity was similar to that of controls (Fig. 2A, C, and E), and the induction of epileptogenesis was completely blocked (Fig. 3). Thus the prolonged elevations in [Ca2+]i associated with epileptogenesis required extracellular Ca2+ and NMDAR activation. These data, taken with the positive correlation between the induction of epileptogenesis and the duration of the [Ca2+]i elevation, as well as the positive correlation between the induction of epileptogenesis and the total [Ca2+]i load, indicate that Ca2+ influx through the NMDAR and prolonged elevations in [Ca2+]i were required for the induction of epileptogenesis by glutamate injury.

As described previously (5), glutamate injury–induced epileptogenesis required an injury sufficient to produce plasticity changes, but not so severe that it caused cell death. In the present study, we demonstrated that glutamate injury–induced epileptogenesis required parameters of total calcium load and duration of significantly elevated calcium sufficient to produce long-lasting hyperexcitability, but not sufficient to produce neuronal death. Various combinations of increased duration or total load of [Ca2+]i could produce a reversible injury sufficient to produce a reversible injury that induced epileptogenesis without reaching the threshold for the induction of IRRC and delayed excitotoxic neuronal death (20). In addition to the increase in [Ca2+]i, the pathway of Ca2+ entry was shown to be important (Fig. 6). An elevation in [Ca2+]i induced by exposure to high extracellular potassium equal in duration and load to the glutamate-induced elevations did not produced epileptogenesis, demonstrating that the route of Ca2+ entry was also important.

These data suggest a continuum of NMDAR–Ca2+ transduction during neuronal injury ranging from complete recovery of function, to neuronal survival with persistently altered function and, finally, to neuronal death. Experimental conditions that inhibited the NMDAR–Ca2+ transduction pathway, as indicated by short-duration elevations in [Ca2+]i (low [Ca2+]e, APV, and MK-801; Fig. 5C and D), manifested neither altered neuronal function in the form of SREDs (Fig. 2C and E) nor neuronal death in the form of IRRC, a short-term hallmark of excitotoxicity (20). In contrast, experimental conditions that permitted activation of the NMDAR–Ca2+ transduction pathway (glutamate alone, CNQX, and MCPG), as evidenced by prolonged, but reversible alterations in Ca2+ homeostasis, induced epileptogenesis. In addition, a small subset of neurons in these epileptogenic experimental conditions of glutamate, CNQX, and MCPG (29, 20, and 28%, respectively) manifested NMDAR–Ca2+-dependent IRRC, and underwent delayed excitotoxic neuronal death, died (20).

Ca2+ is an important second messenger in neurons that triggers a number of Ca2+-dependent signaling pathways including the modulation of enzyme systems and gene transcription (21,33,34) that may be altered during epileptogenesis. For example, Ca2+/calmodulin kinase II activity is depressed in a number of whole animal and in vitro models of epilepsy (35–39). Alterations in the levels of many messenger RNAs and proteins also have been characterized in models of epileptogenesis (40,41). The data in this investigation demonstrate that glutamate injury performed in Ca2+-free, 2.0 mM extracellular Ba2+ solutions did not induce epileptogenesis (Figs. 2D and 3) and indicate that the Ca2+ influx through the NMDAR during epileptogenic glutamate injury initiated some Ca2+-sensitive second-messenger system necessary for the induction of epilepsy.

Because NMDAR activation, cytoplasmic Ca2+ signals, and nuclear Ca2+ signals can regulate gene transcription through multiple Ca2+-activated enzyme and transcription factor pathways (42–44), the prolonged elevations in neuronal [Ca2+]i and activation of the NMDAR–Ca2+ transduction pathway may underlie potential alterations in enzyme regulation and gene expression during glutamate injury–induced epileptogenesis. Long-term modulation of gene expression in epilepsy has implicated as a molecular mechanism mediating the long-term plasticity changes associated with epileptogenesis (40,45,46). Persistent increases in DNA binding and expression of serum response factor and ΔfosB have been associated with long-term plasticity changes in the pilocarpine model of epilepsy (45,46). Because these transcription factors are activated by the NMDA–Ca2+ transduction pathway, it is reasonable to suggest that the NMDAR-mediated Ca2+ changes observed in the glutamate injury model in the present study induce epileptogenesis through the long-term modulation of gene expression. Future studies will investigate the role of Ca2+-activated gene expression in glutamate injury–induced epileptogenesis.

In summary, brain injuries like stroke have been associated with glutamate excitotoxicity wherein excessive activation of the NMDAR leads to an irreversible overload of calcium and neuronal death (9). This investigation demonstrated that a less severe glutamate exposure that induced epileptogenesis in surviving neurons also causes prolonged elevations in neuronal [Ca2+]i that are dependent on the presence of extracellular Ca2+ and activation of the NMDAR. In this situation, rather than triggering excitotoxic neuronal death pathways, calcium activates Ca2+-regulated pathways that initiate long-term plasticity changes in neurons that ultimately lead to the development of epilepsy.

Acknowledgment: This study was supported by NIH grants RO1-NS23350 (R. J. DeLorenzo), P50-NS25630 (R. J. DeLorenzo), NS07288 (D. A. Sun), and a GLENN/American Federation for Aging Research Grant (D. A. Sun).

Ancillary