The Journal of Neurophysiology Vol. 80 No. 5 November 1998, pp. 2688-2698
Copyright ©1998 by the American Physiological Society

Neurotoxicity Mediated by Aberrant Patterns of Synaptic Activity Between Rat Hippocampal Neurons in Culture

John R. McLeod Jr., Maoxing Shen, Daniel J. Kim, and Stanley A. Thayer

Department of Pharmacology, University of Minnesota Medical School, Minneapolis, Minnesota 55455

	ABSTRACT

Abstract Introduction Methods Results Discussion References

McLeod, John R., Jr., Maoxing Shen, Daniel J. Kim, and Stanley A. Thayer. Neurotoxicity mediated by aberrant patterns of synaptic activity between rat hippocampal neurons in culture. J. Neurophysiol. 80: 2688-2698, 1998. Reducing the extracellular Mg²⁺ concentration ([Mg²⁺]_o) to 0.1 mM evoked an aberrant pattern of glutamatergic activity in the synaptic network formed by rat hippocampal neurons grown in primary culture. This treatment resulted in a significant increase in neuronal death when maintained for 20-24 h; 0.1 mM [Mg²⁺]_o elicited a stable and repetitive series of intracellular Ca²⁺ concentration ([Ca²⁺]_i) spikes as indicated by indo-1-based microfluorimetry. Fura-2-based digital imaging experiments found that the [Ca²⁺]_i spikes were synchronized for all the neurons in a given field. Thus electrophysiological recordings from individual cells were reasonable representations of the field as a whole, enabling correlation of electrical activity to viability. Underlying each [Ca²⁺]_i spike was an intense burst of action potentials. Whole cell voltage-clamp experiments showed that a burst was composed of fast action currents superimposed on a slow inward current. The N-methyl-D-aspartate (NMDA) receptor antagonist CGS19755 (10 µM) blocked [Ca²⁺]_i spiking, the slow inward current, and the cell death induced by low [Mg²⁺]_o. The L-type Ca²⁺ channel antagonist nimodipine (10 µM) blocked [Ca²⁺]_i spiking, all synaptic activity, and the cell death induced by low [Mg²⁺]_o. The non-NMDA receptor antagonist 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, 10 µM) exerted variable effects on [Ca²⁺]_i spiking and blocked the slow inward current only when the cells were held at a relatively negative holding potential. CNQX did not afford any protection from 0.1 mM [Mg²⁺]_o-induced neurotoxicity. [Ca²⁺]_i imaging experiments showed that CNQX inhibited [Ca²⁺]_i spiking in a subset of neurons within an active network. Thus, the neurons that were insensitive to CNQX appear to be those that were destined to die. We characterized an in vitro model that allowed us to correlate specific electrophysiological components of glutamatergic synaptic activity to the subsequent viability of the network. A slow NMDA receptor-mediated inward current was required to elicit [Ca²⁺]_i spiking and neurotoxicity. Non-NMDA receptors did not contribute to synaptically mediated cell death in this model. An L-type Ca²⁺ channel antagonist was neuroprotective when used at concentrations that blocked synaptic activity, suggesting that dendritic L-type Ca²⁺ channels present a useful target for neuroprotective drugs.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

The excitotoxicity that results from excessive activation of glutamate receptors is thought to underlie a number of neurodegenerative disorders (Choi 1994; Meldrum 1993; Rothman and Olney 1995). Increased extracellular glutamate was measured after ischemic insult (Benveniste et al. 1984), although a significant portion of neuronal death, resulting from ischemia as well as from other more slowly developing neurodegenerative processes, may be mediated synaptically (Obrenovitch and Urenjak 1997). For example, removing glutamatergic input will protect the striatum from hypoglycemia-induced damage (Linden et al. 1987). Beal et al. (1993) hypothesized that neurodegenerative disorders such as Huntington's disease result from metabolic compromise with a subsequent depolarization that sensitizes the postsynaptic cell to excitotoxicity by relieving the voltage-dependent block of the NMDA receptor by Mg²⁺.

Several in vitro models of excess excitatory synaptic activity are also based on a reduced Mg²⁺ block of NMDA receptors. Reduced [Mg²⁺]_o will elicit epileptic discharges in brain slice preparations (Kohr and Heinemann 1989) and intense synaptic activity in brain cultures (Robinson et al. 1993). The spontaneous glutamatergic activity that increases as the network develops is thought to limit the viability of hippocampal neurons in culture (Peterson et al. 1989). Inducing this synaptic activity by bathing the culture in [Mg²⁺]_o-free media leads to paroxysmal neuronal firing and subsequent cell death (Abele et al. 1990; Rose et al. 1990). Drugs that inhibit this activity may prove useful in certain neurodegenerative disorders.

We previously reported that by reducing [Mg²⁺]_o to 0.1 mM rather than omitting it altogether a stable pattern of repetitive [Ca²⁺]_i spiking was produced that relied on glutamatergic synaptic transmission and thus could be used to study pharmacological agents that modify excitatory neurotransmission (Shen et al. 1996). In this report, we characterize this model further by using pharmacological agents to relate specific electrophysiological components to the subsequent viability of the synaptic network. Only those agents that inhibited an NMDA receptor-mediated current were neuroprotective.

	METHODS

Abstract Introduction Methods Results Discussion References

Materials

Materials were obtained from the following companies: indo-1 and fura-2 from Molecular Probes, Eugene, OR; NMDA, cis-4-phosphonomethyl-2-piperidine carboxylic acid (CGS19755), and CNQX from RBI, Natick, MA; media and horse serum from GIBCO, Grand Island, NY; and all other reagents from Sigma, St. Louis, MO.

Cell culture

Rat hippocampal neurons were grown in primary culture as previously described (Wang et al. 1994) with minor modifications. Fetuses were removed on embryonic day 17 from maternal rats anesthetized with CO₂ and killed by decapitation. Hippocampi were dissected and placed in Ca and Mg-free N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES) buffered Hank's salt solution (HHSS), pH 7.45. HHSS was composed of the following (in mM): 20 HEPES, 137 NaCl, 1.3 CaCl₂, 0.4 MgSO₄, 0.5 MgCl₂, 5.0 KCl, 0/4 KH₂PO₄, 0.6 Na₂HPO₄, 3.0 NaHCO₃, and 5.6 glucose. Cells were dissociated by trituration through a 5-ml pipette and then a flame-narrowed Pasteur pipette. Cells were pelleted and resuspended in Dulbecco's modified magle's media (DMEM) without glutamine and supplemented with 10% fetal bovine serum and penicillin/streptomycin (100 U/ml and 100 µg/ml, respectively). Dissociated cells were then plated at a density of 50,000 cells/well onto 25-mm-round cover glasses that were coated with poly-D-lysine (0.1 mg/ml) and washed with H₂O. The neurons were grown in a humidified atmosphere of 10% CO₂-90% air (pH 7.4) at 37°C. The media was replaced 24 h after plating with DMEM supplemented with 10% horse serum and penicillin/streptomycin and fed every 7 days by exchange of 70% of the media. Cells used for neurotoxicity, and optical experiments were cultured for 12 days and were not treated with mitotic inhibitors.

Toxicity

Rat hippocampal neurons were plated on microetched coverslips (Belco) as described previosuly. Approximately 100 neurons were counted on each coverslip. Coverslips were then treated with the appropriate control or reduced [Mg²⁺]_o solutions. After 20-24 h, the same fields of cells were recounted. Viable neurons were identified based on morphological criteria; they were phase bright and had rounded somata and extended long fine processes. Cell death was determined by comparing the number of viable neurons before and after treatment. Experiments in which >40% of the controls died were excluded from the data set. Paired two-tailed Student's t-test was used to determine significance.

[Ca²⁺]_i Measurement

[Ca²⁺]_i was determined with Ca-sensitive fluorescent chelators with a previously described dual emission microfluorimeter (Werth and Thayer 1994) to monitor indo-1 (Grynkiewicz et al. 1985) in photometry experiments or a dual-excitation fluorescence imaging system (Jin et al. 1994) to monitor fura-2 in digital imaging experiments. Cells were loaded with indicator by incubation in 2 µM indo-1 or fura-2 for 45 min at 37°C in HHSS containing 0.5% bovine serum albumin. Loaded cells were placed in a flow-through chamber (Thayer et al. 1988), and experiments were performed at room temperature. The chamber was mounted on an inverted microscope, and cells were superfused with HHSS at a rate of 1-2 ml/min for 15 min before starting an experiment. Superfusion solutions were selected with a multiport valve coupled with several reservoirs.

Photometry

For excitation of the indo-1, the light from a 75-W Xe arc lamp was passed through a 350/10 nm band-pass filter (Omega Optical, Brattleboro, VT). Excitation light was reflected off of a dichroic mirror (380 nm) and through a ×70 phase-contrast oil immersion objective (Leitz, numerical aperture 1.15). Emitted light was sequentially reflected off of dichroic mirrors (440 and 516 nm) through band-pass filters (405/20 and 495/20 nm, respectively) to photomultiplier tubes operating in photon counting mode (Thorn EMI, Fairfield, NJ). Cells were illuminated with transmitted light (580 nm long pass) and visualized with a video camera placed after the second emission dichroic. Recordings were defined spatially with a rectangular diaphragm. The 5-V photomultiplier output was integrated by passing the signal through an eight-pole Bessel filter at 2.5 Hz. This signal was then input into two channels of an A/D converter (Indec Systems, Sunnyvale, CA) sampling at 1 Hz.

After completion of each experiment, cells were wiped from the coverslip with a cotton-tipped applicator, and then background light levels were determined (typically <5% of minimal 380 nm signal). Autofluorescence from cells that were not loaded with dye was not detectable. Records were later corrected for background, and the ratios were recalculated. Ratio values were converted to [Ca²⁺]_i by the equation [Ca²⁺]_i = K_d(R  R_min)/(R_max  R), in which R is the 405/495-nm fluorescence ratio. The K_d used for indo-1 was 250 nM, and  was the ratio of the emitted fluorescence at 495 nm in the absence and presence of calcium. R_min and R_max were determined in ionomycin (10 µM)-permeabilized cells in calcium-free [1 mM ethylene glycol-bis(-aminoethyl ether)-N-N-N'-N'-tetraacetic acid] and 5 mM Ca buffers, respectively. The system was recalibrated after any adjustments. Values of R_min, R_max, and  ranged from 0.35 to 0.38, 4.23 to 4.34, and 3.0 to 3.95, respectively.

Digital imaging

The chamber containing fura-2-loaded cells was mounted on the stage of an inverted microscope (Nikon Diaphot) and alternately excited at 360 (the isobestic point for fura-2) or 380 nm by rapidly switching optical filters (10 nm band pass) mounted in a computer-controlled wheel (Sutter Instrument) placed between a 75-W Xe arc lamp and the epifluorescence port of the microscope; 360-nm images were collected at the beginning and end of each recording; 380-nm images were collected every 500 ms. Excitation light was reflected from a dichroic mirror (400 nm for fura-2) through a 70× objective (Leitz, numerical aperture 1.15). Fluorescent images, 510 (40) nm for fura-2, were projected (×0.5) onto a cooled charge-coupled device camera (Photometrics, 384 × 576 binned to 192 × 288 pixels) controlled by a computer.

Background images were collected at the conclusion of each experiment after removing cells from the coverslip. Autofluorescence from cells not loaded with the dye was <5% and thus not corrected. Images were corrected for background, and cells were delimited by producing a mask that contained pixel values above a threshold applied to the 380-nm image for fura-2; 360-nm images were calculated for each time point by making a linear extrapolation between pixel values collected in the first and last image; 360 nm/380 nm ratio images were calculated, and the value converted to [Ca²⁺]_i as described in Photometry. Values for R_min, R_max, and  were 1.04, 3.87, and 3.66, respectively.

Electrophysiology

Whole cell recordings were obtained from cultured neurons with pipettes (3-5 M resistance) pulled from borosilicate glass (Narashige, Greenvale, NY) on a Sutter Instruments (Novato, CA) P-87 micopipette puller. Pipettes were filled with a solution containing (in mM) 130 K-gluconate, 10 KCl, 10 NaCl, 10 HEPES, 10 glucose, 5 MgATP, and 0.10 indo-1 pentapotassium salt (combined recordings). The osmolarity of the pipette solution was adjusted to 315 mosmol/kg with sucrose. Whole cell recordings were established in an extracellular solution containing (in mM) 140 NaCl, 5 KCl, 10 CaCl₂, 0.9 MgCl₂, 5 glucose, 0.001 glycine, and 10 HEPES, pH 7.4 with NaOH. After the gigaohm seal was formed, the external solution was changed to one containing (in mM) 140 NaCl, 5 KCl, 1.3 CaCl₂, either 0.9 or 0.1 MgCl₂, 5 glucose, 0.01 glycine, and 10 HEPES, pH 7.4 with NaOH. All extracellular solutions were adjusted to 325 mosmol/kg with sucrose.

Whole cell currents were recorded with an Axopatch 200A patch-clamp amplifier and the BASIC-FASTLAB interface system (Indec systems). For combined electrophysiology and microfluorimetry, membrane potential recordings were filtered at 10 Hz (4-pole Bessel low-pass filter) and sampled at a frequency of 50 Hz.

	RESULTS

Abstract Introduction Methods Results Discussion References

Reducing [Mg²⁺]_o results in a series of [Ca²⁺]_i spikes followed by neuronal death in synaptically connected rat hippocampal neurons in culture.

Reducing the [Mg²⁺]_o bathing cultured CNS neurons elicits an excitatory pattern of electrical activity that results in synaptically mediated neuronal death (Abele et al. 1990; Rose et al. 1990). As shown in Fig. 1A, bathing 11- to 19-day-old hippocampal cultures in Mg-free (no added Mg) media resulted in a significant increase in neuronal death. Cell viability was determined by counting the number of viable neurons before and 20-24 h after treatment. Viable neurons were identified based on morphological criteria; they were phase bright and had rounded somata and extended long fine processes. The same cells were counted after treatment by noting their location on the microetched coverslip on which they were grown. In some experiments viability was confirmed by demonstrating that cells identified as viable also excluded propidium iodide (2 µg/ml). We found that pairing pretreatment with posttreatment cell counts provided a more reproducible assessment of the relatively modest degree of cell death resulting from this treatment. In control cultures (media exchange only) 25 ± 2% of the neurons died. This value is in good agreement with the previous observation that, as CNS cultures mature in vitro, increased synaptic activity parallels spontaneous cell death (Peterson et al. 1989). Removing [Mg²⁺]_o increased neuronal death to 45 ± 2% (n = 21), a significant increase relative to untreated cultures from the same plating (P < 0.001). Complete removal of [Mg⁺]_o evoked complex electrical activity that often resulted in a sustained elevation in [Ca²⁺]_i (Abele et al. 1990). We found that reducing [Mg²⁺]_o to 0.1 mM resulted in a stable repetitive pattern of [Ca²⁺]_i spiking activity more suitable to electrophysiological characterization (Fig. 1B) (Shen et al. 1996). Prolonged exposure to 0.1 mM [Mg²⁺]_o resulted in 40 ± 3% (n = 12) cell death (Fig. 1A), a significant increase relative to control (P < 0.001). A brief 15-min exposure to 0.1 mM [Mg²⁺]_o media did not elicit cell death when assayed 24 h later (n = 3). Thus the initial intense activity seen in Fig. 1B is not sufficient to produce toxicity. This observation contrasts with the results of Abele et al. (1990), which demonstrated significant cell death after treatment with [Mg²⁺]_o-free media for 15 min. Reducing [Mg²⁺]_o to 0.1 mM rather than removing it altogether produces toxicity that relies on sustained synaptic activity.

View larger version (30K): [in this window] [in a new window]   FIG. 1. Reducing [Mg2+]o results in repetitive [Ca2+]i spikes followed by cell death. A: bathing cultured rat hippocampal neurons in 0.1 mM [Mg2+]o (0.1 Mg2+) or nominally Mg2+-free media (Mg2+-free) resulted in a significant increase in neuronal cell death (20-24 h) relative to control (0.9 mM [Mg2+]o). Viability was assessed by cell counting. B: 0.1 mM [Mg2+]o (horizontal bar) resulted in a stable repetitive pattern of [Ca2+]i spiking as indicated by indo-1-based single cell microfluorimetry. *P < 0.001 paired Student's t-test.

Because virtually every individual neuron from which we recorded [Ca²⁺]_i responded to 0.1 mM [Mg²⁺]_o with [Ca²⁺]_i spiking activity (Fig. 1B; n = 64/71), we suspected that reduced [Mg²⁺]_o might excite the entire network of neurons that forms in culture. Indeed, reduced [Mg²⁺]_o has been shown to evoke a repetitive pattern of synchronized bursts of action potentials in recordings from neural networks grown on planar electrode arrays (Jimbo et al. 1993; Robinson et al. 1993). In Fig. 2, we show that our single cell observations can be extended to the network of synaptically connected hippocampal neurons. [Ca²⁺]_i was recorded from a field of seven cells with fura-2-based digital imaging. When [Mg²⁺]_o was reduced to 0.1 mM, [Ca²⁺]_i spiking was recorded that, at least within the time resolution of our digital imaging system, was synchronized among the eight cells. This experiment is representative of 13 experiments that included 67 cells. Thus reducing [Mg²⁺]_o to 0.1 mM elicited a stable pattern of [Ca²⁺]_i spiking activity that was uniform across large fields of neurons, was sufficiently stable for electrophysiological characterization, and elicited cell death. We concluded that this model was suitable for evaluating the relationship between aberrant patterns of synaptic activity and the neurotoxicity that apparently results from it.

View larger version (45K): [in this window] [in a new window]   FIG. 2. 0.1 mM [Mg2+]o induced synchronized [Ca2+]i oscillations in synaptic networks that form in hippocampal cultures. Reducing [Mg2+]o to 0.1 mM elicited a synchronized pattern of [Ca2+]i spiking activity that was uniform across large fields of neurons as indicated by fura-2-based digital imaging. Pseudocolor images were scaled as shown in the plot and collected at the times indicated by the frame numbers annotating the trace. [Mg2+]o was reduced from 0.9 to 0.1 mM at the time indicated by the horizontal bar. Images were collected every 500 ms. Neuronal processes were not resolved in the focal plane used for this recording.

Bursts of action potentials underlie [Ca²⁺]_i transients

We used whole cell patch-clamp in combination with indo-1-based microfluorimetry to simultaneously record electrical activity and [Ca²⁺]_i. In Fig. 3A, current-clamp recordings are shown that indicate that underlying each [Ca²⁺]_i increase was an intense burst of action potentials. In physiological [Mg²⁺]_o (0.9 mM) bursts of action potentials were infrequent (Fig. 3A, left), although spontaneous action potentials were often recorded. Some action potentials were truncated due to the slow sampling rate. Individual action potentials did not produce a detectable increase in the somatic [Ca²⁺]_i in contrast to the bursts that always produced a large increase in [Ca²⁺]_i. Reducing [Mg²⁺]_o to 0.1 mM evoked repetitive bursts of action potentials (Fig. 3A, right). Distinguishing features of the action potential bursts were the high-frequency train of action potentials and a slow depolarization on which the action potentials were superimposed (Fig. 3A, inset).

View larger version (31K): [in this window] [in a new window]   FIG. 3. Synchronized bursts of electrical activity underlie [Ca2+]i transients. Whole cell patch clamp in combination with indo-1-based microfluorimetry was used to simultaneously record electrical activity (black trace) and [Ca2+]i (shaded trace). A: in physiological [Mg2+]o (0.9 mM) bursts of action potentials were infrequent (left). Decreasing the [Mg2+]o to 0.1 mM evoked repetitive bursts of action potentials and an associated increase in [Ca2+]i (right). Underlying each [Ca2+]i spike was a train of action potentials superimposed on a slow depolarization (inset). B: in cells voltage-clamped at 80 mV in physiological [Mg2+]o (0.9 mM), brief inward action currents were observed (left). Reducing [Mg2+]o to 0.1 mM evoked periodic inward currents (right). Bursts of rapidly inactivating inward currents were riding on a sustained inward current (inset). [Ca2+]i remained at basal levels in voltage-clamped cells (trace omitted).

When we switched from current to voltage clamp we saw inward currents that paralleled the action potential waveforms (Fig. 3B). In contrast to Ogura et al. (1988) and in agreement with Robinson et al. (1993) we did not observe changes in the somatic [Ca²⁺]_i in cells held in whole cell voltage clamp. In physiological [Mg²⁺]_o brief action currents were recorded from cells held at 80 mV (Fig. 3B, left). Reducing [Mg²⁺]_o to 0.1 mM evoked periodic bursts of rapidly inactivating inward currents riding on a sustained inward current (Fig. 3B, right). In this study we tested the hypothesis that this slow inward current induced by reduced [Mg²⁺]_o was the electrical activity responsible for neurotoxicity. To test this hypothesis we used selective ion channel antagonists to block particular components of the electrical activity that we subsequently related to neuroprotective efficacy.

Pharmacological characterization of 0.1 mM [Mg²⁺]_o-induced [Ca²⁺]_i spiking

We determined the sensitivity of low [Mg²⁺]_o-induced [Ca²⁺]_i spiking to the NMDA receptor antagonist CGS19755, the non-NMDA receptor antagonist CNQX, and the L-type Ca²⁺ channel antagonist nimodipine (Fig. 4). Treatment with 10 µM CGS19755 decreased spike amplitude by 76 ± 10% and reduced spiking frequency by 68 ± 13% (n = 10). Superfusion of 10 µM nimodipine onto spiking cells decreased spike amplitude by 98 ± 2% and reduced the frequency by 96 ± 5% (n = 11). Nitrendipine was also an effective inhibitor of low [Mg²⁺]_o-induced [Ca²⁺]_i spiking (n = 5). The effects of CNQX (10 µM) were variable. In some cells such as the one shown in Fig. 4A CNQX completely blocked [Ca²⁺]_i spiking activity (11/15 recordings), whereas other cells such as the one shown in Fig. 4B were less sensitive (4/15 cells exhibited incomplete block). [Ca²⁺]_i increases evoked by exposure to Mg-free media were not affected by CNQX (n = 4), consistent with the findings of Abele et al. (1990), suggesting that the non-NMDA receptor-mediated component is needed to relieve a Mg block of the NMDA receptor-gated channel. Thus it appears that the [Ca²⁺]_i spiking activity requires activation of NMDA receptors that are sometimes dependent on activation of non-NMDA receptors. We next determined the contribution of each of these pharmacologically defined components to the electrical activity and neurotoxicity that results from prolonged exposure to 0.1 mM [Mg²⁺]_o.

View larger version (25K): [in this window] [in a new window]   FIG. 4. Pharmacological characterization of 0.1 mM [Mg2+]o-induced [Ca2+]i spiking. [Ca2+]i was recorded from single hippocampal neurons with indo-1-based microfluorimetry. [Mg2+]o was reduced to 0.1 mM, and drugs were added by superfusion at the times indicated by the horizontal bars. A: treatment with 10 µM CGS-19755 decreased [Ca2+]i spike amplitude by 76 ± 10% and reduced spiking frequency by 68 ± 13% (n = 10). Superfusion of 10 µM nimodipine onto spiking cells decreased amplitude by 98 ± 2% and reduced the frequency by 96 ± 5% (n = 11). The effects of 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, 10 µM) were variable. In 73% of the recordings CNQX completely blocked [Ca2+]i spiking. B: trace is representative of the 27% of recordings in which CNQX had partial effects.

NMDA receptor antagonists block the slow inward current and protect from excitoxicity

Cell viability was assessed as described in METHODS and Fig. 1. Drugs were added to the 0.1 mM [Mg²⁺] media before application to the cells. Experiments in which >40% of the neurons died under control conditions were excluded from this data set (7 experiments). As shown in Fig. 5A, 10 µM CGS19755 was an effective neuroprotective agent. In this series of experiments (n = 5) exposure to 0.1 mM [Mg²⁺]_o resulted in 49 ± 2% neuronal death. In the presence of CGS19755 only 26 ± 4% of the neurons died, a value comparable with that seen in the control wells (25 ± 2%) and significantly less than that seen in 0.1 mM [Mg²⁺]_o (P < 0.01).

View larger version (38K): [in this window] [in a new window]   FIG. 5. CGS19755 inhibits the slow depolarization, inward current, and neurotoxicity induced by 0.1 mM [Mg2+]o. A: 10 µM CGS19755 significantly reduced the toxicity induced by 0.1 mM Mg2+; 49 ± 2% of the cells died when exposed to 0.1 mM [Mg2+]o. The toxicity was reduced to 26 ± 4% (n = 5) in the presence of 10 µM CGS19755 (P < 0.01). B: superfusion of a cell held in whole cell current clamp with 0.1 mM [Mg2+]o induced [Ca2+]i spikes (shaded trace) and bursts of action potentials (black trace). Application of 10 µM CGS19755 (hatched bar) significantly inhibited the [Ca2+]i increase as well as the slow depolarization associated with the periodic bursts of action potentials in whole cell current clamp. C: in voltage clamp recordings (80 mV), periodic slow inward currents with superimposed rapid transient currents were observed. The slow inward current was blocked by superfusion with 10 µM CGS19755 (hatched bar).

Repetitive bursts of action potentials and corresponding increases in [Ca²⁺]_i were elicted by 0.1 mM [Mg²⁺]_o, as shown in a whole cell current-clamp recording displayed in Fig. 5B. Application of 10 µM CGS19755 (hatched bar) inhibited the [Ca²⁺]_i increase as well as the slow depolarization associated with the periodic bursts of action potentials (n = 5/5). Action potentials were still observed in the presence of CGS19755, although no periodicity was apparent. The effect was readily reversible. The inhibition of burst firing but not of fast action potentials is in good agreement with a computer model developed by Traub et al. (1994) to describe epileptiform activity in the hippocampal slice. The model is complex, although a key element required for synchronized bursting is an enhanced NMDA-receptor-mediated conductance on the dendrites of pyramidal cells. In four whole cell voltage-clamp recordings held at 80 mV, periodic slow inward currents with superimposed rapid transient currents were observed. CGS19755 completely blocked the slow inward current in three of four recordings (Fig. 5C). These data led us to hypothesize that the slow inward current was required to evoke synaptically mediated excitotoxicity in this model.

Nimodipine blocks electrical activity and prevents excitotoxicity induced by 0.1 mM [Mg²⁺]_o

Nimodipine (10 µM) reduced significantly the toxicity induced by 0.1 mM [Mg²⁺]_o (P < 0.05, Fig. 6A). In this series of experiments (n = 4) 31% of the cells died in the untreated control wells. The death rate increased to 53 ± 3% in the presence of 0.1 mM [Mg²⁺]_o. In the presence of nimodipine this increase was reduced to 35 ± 2% (n = 4), a significant reduction relative to 0.1 mM [Mg²⁺]_o alone (P < 0.05). Our hypothesis, that the slow inward current mediates toxicity, predicted that nimodipine would block the slow depolarization, which it did. However, we were surprised to find that 10 µM nimodipine completely blocked all synaptically mediated activity (Fig. 6B). In six recordings, nimodipine completely blocked the bursts of action potentials and the increases in [Ca²⁺]_i. Similarly, as shown in Fig. 6C, when cells were held at 80 mV, both the slow and fast inward currents were completely blocked (n = 4).

View larger version (31K): [in this window] [in a new window]   FIG. 6. Nimodipine blocks electrical activity and prevents excitotoxicity induced by 0.1 mM [Mg2+]o; 10 µM nimodipine significantly protected hippocampal cultures from excitotoxicity; 53 ± 3% of the cells died when exposed to 0.1 mM [Mg2+]o. The toxicity was reduced to 35 ± 2% (n = 4) in the presence of 10 µM nimodipine (P < 0.05). B: in combined whole cell current-clamp microfluorimetry recordings, 10 µM nimodipine (striped bar) completely blocked the bursts of action potentials and [Ca2+]i spikes. C: in cells voltage clamped to 80 mV, nimodipine completely blocked the slow and fast inward currents.

CNQX failed to prevent excitotoxicity and had variable effects on 0.1 mM [Mg²⁺]_o-induced electrical activity

The cell death induced by 0.1 mM [Mg²⁺]_o was not significantly attenuated when 10 µM CNQX was included in the bathing medium (Fig. 7A). In this series of experiments 29% of the cells died in the untreated control wells (n = 7). When treated with 0.1 mM [Mg²⁺]_o for 20-24 h in the absence and presence of 10 µM CNQX, 48 ± 3% and 44 ± 4% of the cells died, respectively. Treatment with 10 µM CNQX alone (no reduction in [Mg²⁺]_o) was not itself toxic (29 ± 2% in control vs. 26 ± 2% in CNQX).

View larger version (36K): [in this window] [in a new window]   FIG. 7. CNQX has variable effects on 0.1 mM [Mg2+]o-induced electrical activity and fails to prevent excitotoxicity. CNQX failed to protect from 0.1 mM [Mg2+]o-induced excitotoxicity; 48 ± 3 and 44 ± 4% of the cells died when exposed to 0.1 mM [Mg2+]o in the absence and presence of 10 µM CNQX, respectively. B: in some cells, 10 µM CNQX inhibited the bursts of action potentials and associated [Ca2+]i spiking evoked by bathing the cell in 0.1 mM [Mg2+]o. C: in contrast, other cells were relatively insensitive to CNQX. B and C: 10 µM CNQX was superfused at the time indicated by the hatched bar.

Although CNQX consistently failed to protect from 0.1 mM [Mg²⁺]_o-induced excitotoxicity, the effects of the drug on the electrical activity induced by 0.1 mM [Mg²⁺]_o were less straightforward. Two separate observations were made in current-clamp recordings that parallel the differences noted in microfluorimetry recordings (Fig. 4). In two of five recordings, 10 µM CNQX inhibited the bursts of action potentials and associated [Ca²⁺]_i spiking evoked by bathing the cell in 0.1 mM [Mg²⁺]_o (Fig. 7B). In contrast, some cells were relatively insensitive to CNQX (3/5 recordings). For example in Fig. 7C a recording is shown from a cell in which CNQX failed to block the slow inward current. The drug did however, decrease the frequency of depolarizing bursts and reduce the amplitude of the [Ca²⁺]_i increases that coincided with the bursts. These mixed effects of CNQX might result from the variable degree to which the NMDA receptor-mediated, slow depolarization is contingent on prior activation of non-NMDA receptors to relieve the Mg²⁺ block of the NMDA-gated ion channel. This idea is consistent with the observation that CNQX has no effect on the [Ca²⁺]_i increases observed in Mg²⁺-free media (Abele et al. 1990). Thus bursting activity in cells with relatively depolarized membrane potentials would be predicted to be less sensitive to CNQX. Indeed, when 10 µM CNQX was applied to cells held at 80 mV, the drug completely blocked all bursting activity (Fig. 8A, n = 2). Presumably, the negative holding potential enhanced the Mg²⁺ block of the NMDA-gated ion channel. In contrast, when cells were held at 50 mV, CNQX reduced the frequency of bursts but clearly left both the fast action currents as well as the slow inward current intact (Fig. 8B, n = 2). These data demonstrate that sensitivity of this pattern of synaptic activity to CNQX is influenced by membrane potential. Because we have not assessed the effectiveness of the space clamp in the dendrites we cannot make quantitative comparisons.

View larger version (20K): [in this window] [in a new window]   FIG. 8. CNQX inhibition of inward current was dependent on holding potential. In a cell voltage clamped at 80 mV, 10 µM CNQX completely blocked all bursting activity (n = 4/4). B: CNQX reduced the frequency of bursts from a cell held at 50 mV but clearly left both the fast action currents and the slow inward current intact (n = 2/2).

It appears that the electrical component responsible for low [Mg²⁺]_o-induced toxicity is sensitive to both CGS19755 and nimodipine. Curiously, although treatment with CNQX sometimes blocked the Ca spiking activity, this drug showed no protective effects in this assay, suggesting that those cells that are sensitive to CNQX are not the population that will subsequently die. One prediction of this hypothesis is that the population of CNQX-insensitive cells should approximate the percentage of cells that die when treated with 0.1 mM [Mg²⁺]_o. To test this hypothesis we employed fura-2-based digital imaging to measure [Ca²⁺]_i spiking activity in large fields of hippocampal neurons treated with 0.1 mM [Mg²⁺]_o. In Fig. 9, A-C, a field of 20 neurons bathed in 0.1 mM [Mg²⁺]_o was shown to be spiking in synchrony. When the cells were treated with 10 µM CNQX, basal [Ca²⁺]_i was not affected (Fig. 9D), although [Ca²⁺]_i spikes were significantly inhibited in some cells (Fig. 9E). A plot of [Ca²⁺]_i versus time for two cells representing a CNQX-sensitive (green arrow) and a CNQX-insensitive cell (red arrow) are shown in Fig. 9, C and F. In 9 experiments including 152 cells, spiking frequency was reduced by 62%. Application of 10 µM CNQX completely blocked [Ca²⁺]_i spiking in 55% of the cells, in good agreement with the 60% of the cells that survive 20-24 h treatment with 0.1 mM [Mg²⁺]_o. CNQX failed to produce a reduction in the amplitude of the [Ca²⁺]_i spike in 28% of the cells and partially reduced the [Ca²⁺]_i spike amplitude in 17% of the cells. Clearly, the percentage of cells insensitive to CNQX could account for the cells that are destined to die in this model of excitotoxicity.

View larger version (22K): [in this window] [in a new window]   FIG. 9. Differential effects of CNQX on networked hippocampal neurons in culture. A-C: fura-2-based digital imaging of a field of 20 hippocampal neurons showed that the neurons were spiking in synchrony in 0.1 mM [Mg2+]o. D-F: application of 10 µM CNQX completely blocked Ca2+ spiking in 55% of cells (n = 152; green arrow indicates one such cell). In 45% of cells, CNQX failed to completely block [Ca2+]i spiking (red arrow indicates one such cell). The amplitude of the [Ca2+]i spike did not change in 28% of the cells.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

We developed a model system to study the effects of putative neuroprotective agents on the electrophysiological properties of synaptically mediated excitotoxicity. A reduction in [Mg²⁺]_o to 0.1 mM elicited a pattern of glutamatergic synaptic activity that was sufficiently stable to enable the pharmacological characterization of the electrophysiological components and relate these pharmacologically defined currents to the subsequent viability of the culture; 0.1 mM [Mg²⁺]_o evoked a repetitive pattern of bursts of action potentials. The burst could be dissected into fast rapidly inactivating action potentials superimposed on a slow depolarization that presumably resulted from a slow inward current of similar duration. The NMDA receptor antagonist CGS19755 blocked the slow inward current and protected from excitotoxicity. This observation is consistent with the idea that prolonged seizure-like activity mediated by glutamate receptor activation produces neuronal death in hippocampal cultures (Furshpan and Potter 1989). We used other drugs to further relate particular electrical components to the subsequent viability of the neuronal network.

The L-type Ca channel blocker nimodipine also protected from excitotoxicity, although it was effective only at a concentration that completely blocked synaptic activity. Other dihydropyridine drugs were found to be neuroprotective in similar models (Abele et al. 1990). Interestingly, dihydropyridine drugs were generally found to protect in neurotoxicity assays that develop slowly or have a significant synaptic component (e.g., Abele et al. 1990; Weiss et al. 1990), suggesting that the global increase in [Ca²⁺]_i produced by somatic voltage-gated Ca²⁺ channels is not in itself toxic. Indeed, recent evidence indicates that there is a source specificity to Ca²⁺-triggered gene expression (Bading et al. 1993) and cell death (Sattler et al. 1997). Perhaps a local increase in [Ca²⁺]_i, mediated by postsynaptic L-type Ca²⁺ channels and NMDA receptors, triggers cell death processes in a manner analogous to the selective activation of transcription factors by a pattern of synaptic activity with the same pharmacological profile (Deisseroth et al. 1996).

The complete block of electrical activity, including the slow inward current, produced by 10 µM nimodipine was consistent with the hypothesis that the slow depolarization was somehow related to the neurotoxicity induced by 0.1 mM [Mg²⁺]_o. We were unable however to precisely determine the site of action for nimodipine. There are several explanations for the effects of this drug. At a concentration of 10 µM, nimodipine may be acting nonselectively. Some dihydropyridine drugs were found to inhibit NMDA receptors, although this action was not attributed to nimodipine (Skeen et al. 1993). Nimodipine might inhibit non-L-type Ca²⁺ channels or possibly Na⁺ channels, although nimodipine was generally found selective for L-type currents (Bean and Mintz 1994). Another possibility is that, because the action potential waveform determines the relative contributions of the various Ca²⁺ channel subtypes (McCobb and Beam 1991), the pattern of electrical activity elicited by the 0.1 mM [Mg²⁺]_o may preferentially recruit L-type channels. Therefore application of an L-type channel antagonist may exert a greater effect on the propagation of electrical activity through the synaptic network than predicted from the 24% inhibition of whole cell Ca²⁺ currents previously described for cultured hippocampal neurons (Piser et al. 1995). In most neurons, L-type channels are primarily localized to the soma and proximal dendrites (Westenbroek et al. 1990). However, in the CA3 region of the hippocampus, L-type _1C subunits are present in clusters on dendrites (Hell et al. 1993) and localized in the postsynaptic density of excitatory synapses (Hell et al. 1996). Nimodipine may act on these dendritic L-type channels to inhibit network electrical activity and prevent neurotoxicity. In neocortical pyramidal neurons, voltage-gated Ca²⁺ channels located in the dendrites amplify local glutamatergic input (Schwindt and Crill 1997). Clusters of L-type Ca²⁺ channels were also localized to neuritic branch points in hippocampal cultures (Shitaka et al. 1996). As a result of this distribution, L-type channels may be required for the propagation of the depolarization through dendritic branch points, enabling nimodipine to block synaptic activity. This interpretation suggests that nimodipine decreases somatic [Ca²⁺]_i transients indirectly by acting on the dendrites rather than by inhibiting Ca²⁺ influx via L-type channels on the soma.

The non-NMDA receptor antagonist CNQX did not afford any protection from 0.1 mM [Mg²⁺]_o-induced toxicity, although it did block the aberrant pattern of synaptic activity in a subset of cells. The excitatory effect of reduced [Mg²⁺]_o media may bypass non-NMDA receptors, although in current-clamp experiments this was not always the case. These mixed results may depend on the degree of development and activity of the network. In a very active network, reducing the [Mg²⁺]_o to 0.1 mM, when combined with spontaneous activity, might be sufficient to activate NMDA receptors directly, hence circumventing the need for non-NMDA receptors. However, there were also cells in which electrical activity was totally blocked by CNQX. In contrast to the varied effects of CNQX on synaptic activity, under no circumstances did blocking non-NMDA receptors protect the network from 0.1 mM [Mg²⁺]_o-induced excitotoxicity. One theory to explain these results is that the population of cells that were insensitive to CNQX were the cells destined to die. Consistent with this theory was the finding that cells held at a depolarized membrane potential (50 mV) were insensitive to CNQX, whereas cells held at a normal resting potential (80 mV) were sensitive to the drug. We conclude that CNQX was least effective in those cells that were most likely to die, possibly because of a relatively depolarized membrane potential. This observation is consistent with in vivo studies in which systemic administration of mitochondrial poisons, presumably depolarizing the metabolically compromised cell, results in neuronal loss in brain regions that receive glutamatergic input (Schulz et al. 1996). This form of toxicity is attenuated by NMDA receptor antagonists.

In summary, we refined a previously described model of epileptic activity (Furshpan and Potter 1989) and excitotoxicity (Abele et al. 1990; Rose et al. 1990) to elicit a stable pattern of aberrant synaptic activity and reproducible neuronal death. We studied the effects of three drugs shown previously to be neuroprotective in other models. Each drug altered the electrical properties of the network differently, raising intriguing possibilities about the underlying currents responsible for synaptically mediated neurotoxicity. The NMDA receptor antagonist CGS19755 protected from excitotoxicity and inhibited a slow inward current, suggesting that the synaptic influx of Ca²⁺ triggers subsequent neuronal death. The lack of neuroprotective efficacy of CNQX highlights the importance of NMDA receptors in neurodegenerative processes in which the postsynaptic cell is sensitized to glutamate by metabolic compromise. Finally, the neuroprotection produced by nimodipine suggests that targets for neuroprotective drugs might be found between the postsynaptic density and the soma.

	ACKNOWLEDGEMENTS

  This work was supported by the National Institute on Drug Abuse Grants DA-07304 and DA-09293, and National Science Foundation Grant IBN9723796. M. Shen was support by Training Grant DA-07097.

	FOOTNOTES

  Address for reprint requests: S. A. Thayer, Department of Pharmacology; University of Minnesota Medical School, 3-249 Millard Hall, 435 Delaware Street, SE, Minneapolis, MN 55455.

  Received 13 February 1998; accepted in final form 17 July 1998.

	REFERENCES

Abstract Introduction Methods Results Discussion References

• ABELE, A. E., SCHOLZ, K. P., SCHOLZ, W. K., MILLER, R. J. Excitotoxicity induced by enhanced excitatory neurotransmission in cultured hippocampal pyramidal neurons. Neuron 4: 413-419, 1990.[Medline]
• BADING, H., GINTY, D. D., GREENBERG, M. E. Regulation of gene expression in hippocampal neurons by distinct calcium signaling pathways. Science 260: 181-186, 1993.[Abstract/Free Full Text]
• BEAL, M. F., HYMAN, B. T., KOROSHETZ, W. Do defects in mitochondrial energy metabolism underlie the pathology of neurodegenerative diseases? Trends Neurosci. 16: 125-131, 1993.[Medline]
• BEAN, B. P., MINTZ, I. M. In: Handbook of Membrane Channels. San Diego, CA: Academic, 1994, p. 199-210.
• BENVENISTE, H., DREJER, J., SCHOUSBOE, A., DIEMER, N. H. Elevation of the extracellular concentrations of glutamate and aspartate in rat hippocampus during transient cerebral ischemia monitored by intracerebral microdialysis. J. Neurochem. 43: 1369-1374, 1984.[Medline]
• CHOI, D. W. Glutamate receptors and the induction of excitotoxic neuronal death. Prog. Neurobiol. 100: 47-51, 1994.
• DEISSEROTH, K., BITO, H., TSIEN, R. W. Signaling from synapse to nucleuspostsynaptic CREB phosphorylation during multiple forms of hippocampal synaptic plasticity. Neuron 16: 89-101, 1996.[Medline]
• FURSHPAN, E. J., POTTER, D. D. Seizure-like activity and cellular damage in rat hippocampal neurons in cell culture. Neuron 3: 199-207, 1989.[Medline]
• GRYNKIEWICZ, G., PEONIE, M., TSIEN, R. Y. A new generation of Ca²⁺ indicators with greatly improved fluorescence properties. J. Biol. Chem. 260: 3440-3450, 1985.[Abstract/Free Full Text]
• HELL, J. W., WESTENBROEK, R. E., BREEZE, L. J., WANG, K., CHAVKIN, C., CATTERALL, W. A. N-methyl-D-aspartate receptor-induced proteolytic conversion of postsynaptic class C L-type calcium channels in hippocampal neurons. Proc. Natl. Acad. Sci. USA 93: 3362-3367, 1996.[Abstract/Free Full Text]
• HELL, J. W., WESTENBROEK, R. E., WARNER, C., AHLIJANIAN, M. K., PRYSTAY, W., GILBERT, M. M., SNUTCH, T. P., CATTERALL, W. A. Identification and differential subcellular localization of the neuronal class-C and class-D L-type calcium channel-alpha 1 Subunits. J. Cell Biol. 123: 949-962, 1993.[Abstract/Free Full Text]
• JIMBO, Y., ROBINSON, H.P.C., KAWANA, A. Simultaneous measurement of intracellular calcium and electrical activity from patterned neural networks in culture. IEEE Trans. Biomed. Eng. 40: 804-810, 1993.[Medline]
• JIN, W. Z., LEE, N. M., LOH, H. H., THAYER, S. A. Opioids mobilize calcium from Inositol 1,4,5-trisphosphate-sensitive stores in NG108-15 cells. J. Neurosci. 14: 1920-1929, 1994.[Abstract]
• KOHR, G., HEINEMANN, U. Effects of NMDA antagonists on picrotoxin-, low Mg²⁺- and low Ca²⁺- induced epileptogenesis and on evoked changes in extracellular Na⁺ and Ca²⁺ concentrations in rat hippocampal slices. Epilepsy Res. 4: 187-200, 1989.[Medline]
• LINDEN, T., KALIMO, H., WIELOCH, T. Protective effect of lesion to the glutamatergic cortico-striatal projections on the hypoglycemic nerve cell injury in rat striatum. Acta Neuropathol. 74: 335-344, 1987.[Medline]
• MCCOBB, D., BEAM, K. Action potential waveform voltage-clamp commands reveal striking differences in calcium entry via low and high voltage-activated calcium channels. Neuron 7: 119-127, 1991.[Medline]
• MELDRUM, B. S. Excitotoxicity and selective neuronal loss in epilepsy. Brain Pathol. 3: 405-412, 1993.[Medline]
• OBRENOVITCH, T. P., URENJAK, J. Altered glutamatergic transmission in neurological disordersfrom high extracellular glutamate to excessive synaptic efficacy. Prog. Neurobiol. 51: 39-87, 1997.[Medline]
• OGURA, A., MIYAMOTO, M., KUDO, Y. Neuronal death in vitro: parallelism between survivability of hippocampal neurones and sustained elevation of cytosolic Ca²⁺ after exposure to glutamate receptor agonist. Exp. Brain Res. 73: 447-458, 1988.[Medline]
• PETERSON, C., NEAL, J. H., COTMAN, C. W. Development of N-methyl-D-aspartate excitotoxicity in cultured hippocampal neurons. Dev. Brain Res. 48: 187-195, 1989.[Medline]
• PISER, T. M., LAMPE, R. A., KEITH, R. A., THAYER, S. A. -Grammotoxin SIA blocks multiple, voltage-gated, Ca²⁺ channel subtypes in cultured rat hippocampal neurons. Mol. Pharmacol. 48: 131-139, 1995.[Abstract]
• ROBINSON, H.P.C., KAWAHARA, M., JIMBO, Y., TORIMITSU, K., KURODA, Y., KAWANA, A. Periodic synchronized bursting and intracellular calcium transients elicited by low magnesium in cultured cortical neurons. J. Neurophysiol. 70: 1606-1616, 1993.[Abstract/Free Full Text]
• ROSE, K., CHRISTINE, C., CHOI, D. Magnesium removal induces paroxysmal neuronal firing and NMDA receptor-mediated neuronal degeneration in cortical cultures. Neurosci. Lett. 115: 313-317, 1990.[Medline]
• ROTHMAN, S., OLNEY, J. Excitotoxicity and the NMDA receptor- still lethal after eight years. Trends Neurosci. 18: 57-58, 1995.[Medline]
• SATTLER, R., CHARLTON, M. P., HAFNER, M., TYMIANSKI, M. Source-specificity of Ca²⁺ neurotoxicity revisited: the route of Ca²⁺ influx not the degree of Ca²⁺ loading, determines excitotoxic neuronal death. Soc. Neurosci. Abstr. 23: 2291, 1997.
• SCHULZ, J. B., MATTHEWS, R. T., HENSHAW, D. R., BEAL, M. F. Neuroprotective strategies for treatment of lesions produced by mitochondrial toxinsimplications for neurodegenerative diseases. Neuroscience 71: 1043-1048, 1996.[Medline]
• SCHWINDT, P. C., CRILL, W. E. Local and propagated dendritic action potentials evoked by glutamate iontophoresis in rat neocortical pyramidal neurons. J. Neurophysiol. 77: 2466-2483, 1997.[Abstract/Free Full Text]
• SHEN, M., PISER, T. M., SEYBOLD, V. S., THAYER, S. A. Cannabinoid receptor agonists inhibit glutamatergic synaptic transmission in rat hippocampal cultures. J. Neurosci. 16: 4322-4334, 1996.[Abstract/Free Full Text]
• SHITAKA, Y., MATSUKI, N., SAITO, H., KATSUKI, H. Basic fibroblast growth factor increases functional L-type Ca²⁺ channels in fetal rat hippocampal neuronsimplications for neurite morphogenesis in vitro. J. Neurosci. 16: 6476-6489, 1996.[Abstract/Free Full Text]
• SKEEN, G. A., TWYMAN, R. E., WHITE, H. S. The dihydropyridine nitrendipine modulates N-methyl-D-aspartate receptor channel function in mammalian neurons. Mol. Pharmacol. 44: 443-450, 1993.[Abstract]
• THAYER, S. A., STUREK, M., MILLER, R. J. Measurement of neuronal Ca²⁺ transients using simultaneous microfluorimetry and electrophysiology. Pflügers Arch. 412: 216-223, 1988.[Medline]
• TRAUB, R. D., JEFFERYS, J. G., WHITTINGTON, M. A. Enhanced NMDA conductance can account for epileptiform activity induced by low Mg²⁺ in the rat hippocampal slice. J. Physiol. (Lond.) 478: 379-393, 1994.[Medline]
• WANG, G. J., RANDALL, R. D., THAYER, S. A. Glutamate-induced intracellular acidification of cultured hippocampal neurons demonstrates altered energy metabolism resulting from Ca²⁺ loads. J. Neurophysiol. 72: 2563-2569, 1994.[Abstract/Free Full Text]
• WEISS, J. H., HARTLEY, D. M., KOH, J., CHOI, D. W. The calcium channel blocker nifedipine attenuates slow excitatory amino acid neurotoxicity. Science 247: 1474-1477, 1990.[Abstract/Free Full Text]
• WERTH, J. L., THAYER, S. A. Mitochondria buffer physiological calcium loads in cultured rat dorsal root ganglion neurons. J. Neurosci. 14: 348-356, 1994.[Abstract]
• WESTENBROEK, R. E., AHLIJANIAN, M. K., CATTERALL, W. A. Clustering of L-type Ca²⁺ channels at the base of major dendrites in hippocampal pyramidal neurons. Nature 347: 281-284, 1990.[Medline]

0022-3077/98 $5.00 Copyright ©1998 The American Physiological Society